Front Cover

KEY

ACambridge University Hospitals NHS Foundation Trust
BBrighton & Sussex University Hospitals NHS Trust
CCardiff & Vale NHS Trust
DCentral Manchester & Manchester Children's University Hospitals NHS Trust
EGreat Ormond Street Hospital for Children NHS Trust
FGuy's & St. Thomas' NHS Foundation Trust
GHull & East Yorkshire Hospitals NHS Trust
HKing's College Hospital NHS Trust
ILeeds Teaching Hospitals NHS Trust
JThe Lewisham Hospital NHS Trust
KNewcastle upon Tyne Hospitals NHS Foundation Trust
K1Newcastle General Hospital
K2Newcastle Freeman Hospital
K3Newcastle Royal Victoria Infirmary
LUniversity Hospital of North Staffordshire NHS Trust
MNottingham University Hospitals NHS Trust
NOxford Radcliffe Hospitals NHS Trust
ORoyal Brompton & Harefield NHS Trust
PRoyal Liverpool Children's NHS Trust
QSheffield Children's NHS Foundation Trust
Q1Sheffield Children's Hospital (NICU)
Q2Sheffield Children's Hospital (PICU)
RSouthampton University Hospitals NHS Trust
SSouth Tees Hospitals NHS Trust
TSt. George's Healthcare NHS Trust
USt. Mary's NHS Trust
VBirmingham Children's Hospital NHS Trust
WUnited Bristol Healthcare NHS Trust
XUniversity Hospitals of Leicester NHS Trust
X1Leicester Glenfield Hospital
X2Leicester Royal Infirmary
YNHS Lothian - University Hospitals Division

Published in the UK by the Paediatric Intensive Care Audit Network (PICANet). This work is copyright. Apart from any use as permitted under the Copyright, Designs and Patents Act 1988, no part may be reproduced by any process without permission from PICANet. Requests and enquiries concerning reproduction rights should be directed to PICANet at:

PICANet
Paediatric Epidemiology Group
Centre for Epidemiology and Biostatistics
The Leeds Institute of Genetics, Health and Therapeutics
University of Leeds
30 Hyde Terrace
Leeds
LS2 9LN 
0113 343 4856
picanet@leeds.ac.uk

In all cases PICANet must be acknowledged as the source when reproducing or quoting any part of this publication. Please use the following format when citing this report:

Paediatric Intensive Care Audit Network National Report 2004 - 2006 (published June 2007): Universities of Leeds and Leicester. ISBN 978 0 85316 264 3.

1 CONTENTS

2 ACKNOWLEDGEMENTS

We are acutely aware that the success of this national clinical audit is highly dependent on the hard work and commitment of a large number of individuals working within the paediatric intensive care community. We are very grateful to all the audit clerks, secretaries, nurses and doctors who support and contribute to the Paediatric Intensive Care Audit Network (PICANet) from their own paediatric intensive care units (PICUs).

PICANet was established in collaboration with the Paediatric Intensive Care Society (PICS) and their active support continues to be a key component of our successful progress. The PICANet Steering Group (SG) has patient, academic, clinical, government and NHS members all of whom are thanked for their continuing assistance and advice. Members of our Clinical Advisory Group (CAG) are PICANet's formal interface with clinical care teams and their valuable support and contribution is gratefully acknowledged.

PICANet is funded by the Department of Health (DOH), Health Commission Wales Specialised Services, Royal Hospital for Sick Children, Edinburgh and the Pan Thames PICU Commissioning Consortium.

The organisation and functioning of PICANet is dependent on IT programming and development from Martin Perkins (University of Leicester), who we thank for his essential contributions.

3 FOREWORD

PICANet is showing the way both internationally in paediatric intensive care and, within the UK, to other areas of health care. This report demonstrates what can be achieved when clinicians and health services researchers work together. The combination of, on the one hand, clinical knowledge and experience, and on the other hand epidemiologists, statisticians and information technologists has resulted in the development of one of the finest clinical databases in the UK.

This report includes many examples of how such data can be used to shed light on the clinical management of severely ill children, the organisation of paediatric intensive care and the quality of care in intensive care units. Without such data, improvements in care would be seriously limited.

Increasingly there is a tendency to believe that routine data collected largely for administrative purposes are sufficient to audit care and provide a base for conducting research studies. Such a view ignores the shortcomings of such data. This report demonstrates why we need sophisticated, complex specialised databases. Rather than bemoan the fact, we should celebrate the multiple purposes and versatility of databases such as PICANet. It can underpin not only clinical and organisational audits, but also research, management and planning of services, individual patient care and the training needs of clinicians. These diverse uses are reflected in the contents of this report, such as contributions on the use of the database for commissioning care and developing financing mechanisms.

With increasing recognition by policy makers of the need for accurate information on the outcomes of care, PICANet can and should make a crucial and valued contribution over the coming years. Its quality is a tribute to the health services researchers and clinicians who have developed and lead this important work.

Nick Black
Professor of Health Services Research
London School of Hygiene & Tropical Medicine 
and Chair, PICANet Steering Group

4 EXECUTIVE SUMMARY

  1. PICANet is a clinical audit of paediatric intensive care (PIC) activity in England and Wales aiming to improve patient outcomes through providing information on delivery of care to critically ill children and an evidence base for clinical governance. PICANet was established in 2001 and functions in close collaboration with members of the PIC clinical community.
  2. The specific objectives of PICANet are to identify best practice, monitor supply and demand, monitor and review outcomes of treatment episodes, facilitate strategic health care planning, quantify resource requirements and study the epidemiology of critical illness in children.
  3. The national PICANet dataset continuously records details of admission, discharge, diagnoses (coded using Clinical Terms 3 (The Read Codes)), medical history, physiology, interventions and outcome. The outcome information is adjusted by 'case mix' to provide reliable evidence on patients' outcomes for clinicians, managers, patients. From 2006 the casemix adjustment tool is the updated Paediatric Index of Mortality 2.
  4. Rigorous data quality procedures, incorporating iterative feedback loops between PICANet and the units, continue to ensure the dataset is of high quality.
  5. PICANet are developing and expanding the core dataset in response to changes in the infrastructure and funding streams of the NHS. PICANet will provide the software for units to record the Paediatric Critical Care Minimum Dataset (PCCMDS) to support the Paediatric Critical Care Healthcare Resource Groups (HRGs) and Payment by Results (PbR). The flexibility for the collection of unit specific additional items will remain, whilst additional modules, such as that on retrievals, are under construction.
  6. Data are presented on 42,221 paediatric intensive care admissions to 24 NHS trusts in England and Wales and the Royal Hospital for Sick Children, Edinburgh over the 3 year period January 2004 to December 2006. Detailed tables present information nationally, by Strategic Health Authority/Health Board (SHA), Primary Care Organisation (PCO) and named individual NHS trust. For the first time, data are available for downloading from the Web in spreadsheet format.
  7. Children under 1 year comprise 48% of all admissions with an overall excess of boys (57%) compared to girls (43%). The majority of admissions (54%) are unplanned. Retrievals of 75% of children are by specialist paediatric intensive care teams.
  8. Invasive ventilation procedures are recorded for 67% of admissions. This varies by trust between 6% and 95% over the three years.
  9. A total of 242,997 bed days were delivered between 2004 and 2006. Length of stay has been calculated to the minute and presented as numbers of admissions by length of stay category ranging from less than an hour (0.8%) to 7 days or longer (16%). A 'bed census' has been calculated for children actually occupying a bed at 10 minutes past midnight on each day to provide a more accurate assessment of daily occupancy in the PIC service.
  10. It is extremely rare for a child to die in paediatric intensive care and 95% are discharged alive. Risk-adjusted performance of all trusts fell within acceptable limits in each individual year and aggregated across the three year period.
  11. The re-organisation of the NHS into Primary Care Organisations in 2006 is reflected in this report. Maps by SHA and PCO illustrate considerable variation in the geographical distribution of the volume of patients receiving paediatric intensive care and the percentage of children invasively ventilated.
  12. PICANet acknowledge that data on status 30-day post discharge is incomplete for 57% of children discharged alive.
  13. PICANet remains responsive to the needs of the clinical community and service providers and a number of new features are incorporated into this report. Clinicians and commissioners have contributed chapters on specific topics. These include a clinician's commentary, information on the PCCMDS, the retrievals dataset, health informatics, PICU staffing and a commissioner's perspective. These all add information on the context and environment within which PICANet operates.
  14. Twelve recommendations arising from this report are outlined in the next section.

5 RECOMMENDATIONS

PICANet recommend

  1. That high quality clinical audit data on children receiving intensive care in England, Wales and Scotland should continue to be collected to optimise the delivery of care, to facilitate future planning, permit ongoing audit and describe the epidemiology of critically ill children.
  2. Complete coverage of the UK to incorporate data from the PICU in Northern Ireland to enable the diversity of clinical practice to be characterised at a national level.
  3. That optimal outcome measures are developed for paediatric intensive care to facilitate the improvement of professional practice and quality of PIC services.
  4. That links with the clinical community and professional organisations, such as the Paediatric Intensive Care Society Study Group, continue to be strengthened and expanded via collaborative audit and research using the PICANet dataset.
  5. That links with PIC commissioners are enhanced to facilitate the planning of PIC services.
  6. The PICANet dataset should be used for future calibration of risk-adjustment algorithms in paediatric intensive care.
  7. That Trusts provide support for the collection of child status at 30 days following discharge from PIC especially in those trusts with little or no follow-up data.
  8. That Trusts share their experiences of the collection of NHS numbers to improve this data collection to a level in excess of 95%.
  9. Continued efforts to capture complete national data on children admitted to adult intensive care units.
  10. Further investigation of the differences in risk adjusted mortality and the prevalence of paediatric intensive care and invasive ventilation by Strategic Health Authorities and Primary Care Organisations to determine which factors might explain this variation.
  11. Further exploration of the patterns of admission for individual children, as one of the key functions of PICANet is to investigate patterns of re-admission to PICUs for children across the UK.
  12. International collaborations should be established to enable the development of large-scale audit comparisons between countries that will inform clinical practice.

6 BACKGROUND

PICANet was established in 2002, following a tender in 2000 by the DOH for a national paediatric intensive care database that would allow core data to be collected in a standardised way throughout all PICUs in the country.

Since November 2002, all NHS PICUs within England and Wales outside the Pan Thames region have been collecting data on consecutive admissions to their units. The Pan Thames units began data collection in March 2003, whilst the PICU at the Royal Hospital for Sick Children, Edinburgh began in December 2004. A full list of participating PICUs can be found in Appendix A.

PICANet receives support and advice from a Clinical Advisory Group consisting of doctors and nurses working within the speciality. A Steering Group (SG), comprising professionals from Health Services Research, the Royal Colleges of Paediatrics & Child Health, Nursing and Anaesthetics, and user groups such as Action for Sick Children, monitors PICANet and offers additional support and advice. Appendices B and C provide a full list of CAG and SG members. Additional support from the clinical community is provided through the Paediatric Intensive Care Society.

7 INTRODUCTION AND AIMS

This is the fourth national report produced by PICANet on data submitted by participating PICUs in the UK. This year, the report has been published in three formats:

  1. As a .pdf document, downloadable from http://www.picanet.org.uk/.
  2. As a web document with tables and figures available for download in Microsoft Excel format, again, available from http://www.picanet.org.uk/.
  3. A limited number of printed copies.

This year we are pleased to include a number of chapters from independent contributors. The views represented in these chapters are those of the authors and do not necessarily represent the views of PICANet.

We have decided to limit the print run for environmental and cost reasons. The downloadable format means that individuals can select specific sections of the report to print if necessary and the tables and figures can be manipulated and used in presentations and reports. Please ensure that PICANet is acknowledged as the source of this information using the format given on the inside cover.

In collaboration with participating units, PICANet remains committed to achieving the following objectives:

Since data collection commenced in 2002, one of the main aims of PICANet has been to provide a national database of paediatric intensive care activity of a consistently high quality, in order to help achieve the above objectives. The data collected allows comparisons of activity at a local level with nationwide benchmarks. PICANet therefore provides an important evidence base on paediatric intensive care outcomes, processes and structures, permitting planning for future practice, research and interventions.

PICANet is a resource available to clinicians and service providers, amongst others, and is being used for research, audit and commissioning (Appendix D). The provision of comprehensive, routinely available information to such parties is extremely important and is a powerful tool for supporting clinical governance. PICANet is also used to provide data to provide baseline information for clinical trials.

8 THE PICANet DATASET

8.1 Development and description of the current dataset

The PICANet dataset was established in consultation with members of the PICANet CAG, representing the paediatric intensive care community, and the Department of Health. The overriding criteria for inclusion of specific variables were that they provided key information on activity, case mix, demographics and outcome at a national and local level, that they were feasible to collect and that the wider paediatric intensive care community supported their inclusion in the national database.

The current PICANet dataset consists of 94 variables (including five address elements and the option for a second family name). These variables and their definitions are given in the PICANet Data Definitions Manual, obtainable from http://www.picanet.org.uk/. The data collection form is included in Appendix E. This dataset will be expanded from summer 2007 when PICANet software will enable the collection of the Paediatric Critical Care Minimum Dataset.

8.2 The Paediatric Critical Care Minimum Dataset

The Paediatric Critical Care Minimum Dataset (PCCMDS) has been developed by the Information Centre for health and social care (IC) under the guidance of the Paediatric Critical Care Expert Working Group (PCCEWG) and was issued as an NHS dataset change notice (DSCN) in January 2007. The PCCMDS has been developed to support the new Paediatric Critical Care Healthcare Resource Groups (HRGs) and Payment by Results (PbR). This dataset has many common elements with the PICANet dataset but collects information on interventions and treatment on a daily basis as opposed to an episode summary. This dataset has been mandated from October 2007.

With the support of the CAG, PICANet has agreed to enable collection of the PCCMDS using its software. The current intervention fields will be populated using the new data items. This will ensure comparability with historical PICANet data and will reduce duplication of data collection effort. In the future, PICANet will also have more detailed information on daily activity which will provide better information for clinical audit and commissioning. The software will also enable PICUs to export the PCCMDS for processing by their trust to enable accurate returns for PbR. The additional burden of data collection is estimated at 1 minute 45 seconds per patient per day based on the pilots carried out to develop the PCCMDS. PICANet will not be responsible for completing data returns for PbR from the central database The processes involved in developing the PCCMDS are described in detail in this report by Dr Kevin Morris, a member of the PCCEWG.

8.3 Retrievals dataset

PICANet has not collected detailed information on retrievals of critically ill children in the past, concentrating on their experience in PICU. With the support of PICANet, the Clinical Advisory Group and the Paediatric Intensive Care Society, Dr Allan Wardhaugh has developed detailed proposals for a dataset that will capture information on this important sub-population of children during the retrieval process. These proposals are outlined in a separate chapter in this report.

8.4 Data collection and validation

PICANet has developed a paper data collection form and bespoke data entry software to enable a consistent national dataset to be assembled. Those units who use their own or commercial data collection software have been provided with an export file specification to enable data to be imported by the PICANet software. Training sessions were organised over two days to familiarise clinical and data entry staff with data definitions, data collection issues and software. Since the original training sessions, ad hoc training has been provided by the PICANet team for new staff concerned with data collection and entry.

The PICANet software performs internal logical consistency and range checks as data are entered and provides an on-screen summary of outstanding validation checks on the completion of a record. Units importing data from their own databases are provided with an import log, detailing which records have been imported and any outstanding validation issues. Central validation and data quality issues are dealt with in the section on data quality.

8.5 Clinical coding

Clinical diagnoses and procedures are coded using Clinical Terms 3 (The Read Codes) referred to as CT3. CT3 encompasses a huge range of diagnostic, procedural and context-dependent clinical codes designed to reflect all aspects of clinical care in the population in general. The long-term strategy of the NHS is to use SNOMED CT® for clinical coding of diagnostic information (see http://www.connectingforhealth.nhs.uk/ for further details). PICANet will migrate to SNOMED CT® when the appropriate support architecture is in place but will continue to use CT3 in the meantime. There are plans to develop a SNOMED subset for PICU, an initiative supported by Connecting for Health. This issue is being taken forward by representatives of the Paediatric Intensive Care Society Study Group Health Informatics Group, with the support of PICANet, and is described in this report by Drs Padmanabhan Ramnarayan and Krishnan Thiru.

8.6 Confidentiality

PICANet collects patient identifiable information including name, address, date of birth and NHS number. With this information, PICANet can identify multiple admissions for the same individual, making the dataset person and episode-based. Personally identifiable information held by PICANet has been linked with death registration details, obtained from the Office for National Statistics (ONS), to assess long-term mortality in children admitted to paediatric intensive care. National census and other geographical data have been linked with validated postcodes of individual children to enable PICANet to assess the association between social class, population density and other geo-demographic and environmental information and paediatric intensive care admissions.

To comply with the provisions of the Data Protection Act, 1998, PICANet has implemented stringent confidentiality and data protection arrangements. The Patient Information Advisory Group (PIAG) has granted PICANet exemption from gaining signed parental consent under Section 60 of the Health and Social Care Act, 2001. This class support enables PICANet to collect and process patient identifiable information for the purpose of auditing, monitoring and analysing patient treatments, to ensure that adequate and appropriate paediatric intensive care services are available for all children admitted for paediatric intensive care. Exemption was given under specified conditions in December 2002 and is due for review in June 2007.

Posters providing information about PICANet are displayed in PICUs, and information leaflets for parents / guardians and children are available (see Appendix F for a copy of the information leaflet).

8.7 Data transmission

The PICANet data entry software includes the facility to transmit data electronically via the NHS intranet if local IT infrastructure can be configured appropriately. The data are first encrypted using public key encryption and then placed on the server. The uploaded data is regularly moved to a secure holding area, decrypted and uploaded onto the central server database.

Where local IT departments have been unable or unwilling to configure their systems and firewalls to allow electronic transfer, the data is encrypted and placed in a local folder and then sent as an email attachment.

9 DATASET DEFINITIONS FOR THIS REPORT

  1. This report covers the three year period January 2004 - December 2006. During this time, there were 43,140 admissions to participating PICUs.
  2. There are 25 participating NHS trusts (located in England, Wales and Scotland), 24 of whom collected data for the entire reporting period. The Royal Hospital for Sick Children, Edinburgh did not join PICANet until December 2004.
  3. Trusts are identified in this report, with agreement from all participating trusts' Chief Executives.
  4. A key enabling identification of each trust can be found on the inside cover.
  5. The main focus of this report are admissions aged 0 - 15 years of which there were a total of 42,221 over the three year period. In addition there were 919 admissions aged 16 years and above.
  6. Unless stated otherwise, the proportions in tables throughout the report are row percentages, except in the total column where they are column percentages.
  7. 'Unknown' includes cases where the unit have specifically recorded 'not known' and also cases where a required value has been left blank.

10 DESCRIPTION OF TABLES AND FIGURES

A brief description of the data contained in the tables and figures is given below, together with hyperlinks to the beginning of each section. In the .pdf version of this report, the hyperlink will bring you to the first page of the section. In the web document, the hyperlink will take you to an Excel spreadsheet that contains links to all the tables and figures in the section. These are all downloadable for use by individuals and organisations but please acknowledge the source of this data as indicated on the inside of the front cover. In some cases, individual figures are not described separately, as they clearly relate to the data in the tables on the same worksheet.

The PICANet dataset is dynamic and updated regularly. This means that overall admission figures have changed for 2004 and 2005 since the publication of the third national report. The data in this report are those supplied to PICANet up to March 2nd, 2007, when the dataset was frozen.

11 ADMISSION DATA

11.1 Admission numbers by age, sex, month and year of admission, NHS trust and diagnostic group

Tables 1 - 9 give numbers of admissions by age, sex, month of admission, NHS trust and diagnostic group. The primary diagnosis for the whole admission has been categorised into 13 diagnostic groups to enable a simple comparison between NHS trusts. The classification is based on CT3 (The Read Codes). Within these mutually exclusive thirteen groups:

Read codes are five characters in length and can be made up of numbers, letters, or periods. The ordering of the individual characters does not indicate the hierarchy (e.g. patent ductus arteriosus (P70..) is a subset of congenital abnormality of ductus arteriosus (Xa6aC)). Table 8 and figure 8 focus on admissions for respiratory conditions by year and month.

11.2 Admissions by Strategic Health Authority (SHA) / Health Board (HB)

Tables 10a and 10b give numbers of admissions by SHA / HB, prior to and following the July 2006 NHS reorganisation. These were obtained by linking the validated home address of children admitted to PICU to SHA / HB via the National Statistics Postcode Directory (NSPD) (http://www.statistics.gov.uk/geography/nspd.asp). These tables present column percentages. Of the total number of admissions 97.5% had addresses which were validated. The remaining 2.5% included foreign addresses (2.3%) and missing addresses (0.2%). Figures 10a and 10b identify the SHA / HB boundaries pre and post reorganisation together with their names; figure 10c overlays the primary care structure.

11.3 Admissions by mortality risk category

Table 11 gives numbers of admissions by mortality risk group by NHS trust. The expected probability of mortality was estimated using the paediatric index of mortality (PIM)1, using recalibrated coefficients supplied by UK PICOS2. The categorization into <1%, 1-<5%, 5%-<15%, 15-<30% and 30% plus expected probability of mortality reflects those used by the Australian and New Zealand Intensive Care Society (ANZPICS)3 for comparability.

11.4 Admissions by admission type

Tables 12 - 15 present numbers by admission type overall and by trust and year and a breakdown of the source of admission and care area admitted from by trust and year for emergency admissions (see below).

We have used the following definitions for type of admission:

NB: Surgery is defined as undergoing all or part of a procedure or anaesthesia for a procedure in an operating theatre or anaesthetic room. Patients admitted from the operating theatre where surgery is not the main reason for admission (e.g. a patient with a head injury who is admitted from theatre after insertion of an ICP monitor) are not included here. In such patients the main reason for admission is head injury and thus the admission type would be 'unplanned - other'.

11.5 Admissions by primary diagnostic group

Tables 16 - 22 present a breakdown of admissions by diagnostic group, overall, by trust and year and further by trust and year for each of the admission types listed above.

Tables 23 - 25 present the twenty most common Read Codes returned to PICANet for primary reason for admissions overall (these represent 15,274 (36%) of all admissions) and for unplanned admissions (after surgery and 'other') by sex without any attempt to group them further.

PICANet has not imposed an arbitrary grouping of codes but present the raw data for the top 20 codes. The level of precision in the coding method makes interpretation of these data difficult without some form of aggregation. However, PICANet has allowed the flexibility to code very specifically to enable prospective audit to focus on particular conditions; for example, respiratory syncytial virus (RSV) positive bronchiolitis. Some units have chosen to code diagnoses in more detail to allow them to use this information locally, others have coded a single diagnosis at a general level. For most reporting purposes, the broad diagnostic groups used in this report are sufficient. Further disaggregation is not always possible due to the variation in coding practice between individual units.

11.6 References

  1. Shann F, Pearson G, Slater A, Wilkinson K, Paediatric index of mortality (PIM): a mortality prediction model for children in intensive care. Intensive Care Med 1997; 23:201-207.
  2. Brady AR, Harrison D, Black S, Jones S, Rowan K, Pearson G, Ratcliffe J, Parry GJ, on behalf of the UK PICOS Study Group. Assessment and Optimization of Mortality Prediction Tools for Admissions to Pediatric Intensive Care in the United Kingdom. Pediatrics 2006; 117: 733-742.
  3. Australian and New Zealand Intensive Care Society. Report of the Australian and New Zealand Paediatric Intensive Care Registry 2005. ISBN: 1876980184 [Online] [Accessed 23/02/2007] Available from the World Wide Web at http://www.anzics.com.au/uploads/2005ANZPICRReport.pdf.

12 RETRIEVAL DATA

Tables 26 - 28 present retrieval data by team type and age, by diagnostic group for non­specialist team retrievals (see below) and by team type and trust.

Data are collected on whether or not a child was retrieved / transferred into the PICU. We have used the following definitions:

In the majority of PICUs, doctors and nurses who work on the unit undertake retrieval of critically ill children. Within London, there are two specific transport teams, the Children's Acute Transfer Service (CATS) and the South Thames retrieval team. CATS is based at Great Ormond Street Hospital (GOSH), and is staffed separately from the intensive care units at GOSH. For PICANet, any child retrieved by CATS into a PICU at GOSH is recorded as 'other specialist team (PICU)'. The South Thames retrieval team is based at Evelina Children's Hospital and is staffed by doctors and nurses from within the PICU. For PICANet, any child retrieved by the South Thames team into the PICU at Evelina Children's Hospital is classed as 'own team'.

The Central Manchester and Manchester Children's University Hospitals NHS Trust has two sister hospitals (Booth Hall and the Royal Manchester Children's Hospital). For local reporting reasons, hospital transfers between the two hospitals are classed as internal admissions (admissions from the 'same hospital') but as the hospitals are 6 miles apart, any transfer requires a 'retrieval' by ambulance and crew.

13 INTERVENTION DATA

Tables 29 - 31 present summary data relating to interventions carried out on PICU. Most of the interventions described are available in all PICUs, although a few specialist interventions (such as extra corporeal membrane oxygenation (ECMO) or left ventricular assist device to support cardiac function (LVAD)) are only available in a PICU where invasive cardiac procedures are routinely performed. Note that table 30 contains aggregated data for 2004 - 2006. This, however, does not include Birmingham Children's Hospital as no intervention data was returned for 2005.

Length of ventilation was calculated in whole days. Any ventilation during the period

00:00 to 23:59 was counted as one complete day of ventilation (e.g. a child intubated and ventilated at 23:45 on 7 March, and extubated at 02:30 on 8 March, would count as two days of ventilation). Intubation and extubation times are not recorded in the PICANet dataset.

Figures 31a - 31c map the percentage of children receiving invasive ventilation by SHA pre and post- the July 2006 NHS reorganisation and by primary care organisation (PCO) post October 2006 reorganisation for 2004 and 2006. Data for 2005 are not mapped as, no intervention data were returned by Birmingham Children's hospital in 2005. The proportion of children invasively ventilated has been used as a very rough proxy for level of care.

14 BED ACTIVITY AND LENGTH OF STAY

Tables 32 - 33 present data on total bed days delivered by age and sex overall and by age by trust. The total number of bed days delivered is calculated as the sum of children receiving intensive care in a PICU each day. Tables 34 - 35 and their associated figures present summary data by year and month and by trust and year on a 'bed census': the number of children present in a PICU bed at 10 minutes past midnight. Tables 36 - 37 present data we describe as 'bed activity' by month and by trust, where a bed is counted as occupied if a child was present on a unit for any part of a the day. This inevitably results in higher figures than the bed census data as a bed may have more than one child occupying it in any one day. Tables 38 - 39 present summary data on length of stay by trust and age group and trust and diagnostic group. Table 40 groups the number of admissions by length of stay by trust, calculated to the minute in categories ranging from less than 1 hour to over 1 week. Children admitted prior to the report period, but discharged during it, are counted from 00:00 on 1 January 2004 until their discharge (or until 24:00 on 31 December 2006 if not discharged). Children admitted during the report period but discharged in 2007 (or who are still on the PICU) are counted from their admission date until 24:00 on 31 December 2006.

The number of bed days, bed census, bed activity and length of stay data are summarised by median and interquartile range (IQR). Median daily bed census figures and daily bed activity are plotted using a box and whisker graph by month and year, and by NHS trust. This type of graph indicates the median by a line within the coloured box, the ends of which give the IQR. The 'whiskers' indicate values beyond the IQRs, although extreme outside values are not plotted.

15 OUTCOME DATA

PICU mortality data are described in terms of unit discharge status by age and sex for England, Wales and Edinburgh combined, and by trust in tables 41 - 45 and also using unadjusted and risk-adjusted standardized mortality ratios (SMRs). Table 46 describes the discharge destination of children discharged alive from PICU. Unadjusted SMRs are calculated by dividing the expected number of deaths, based on the national data by the observed number of deaths in each trust. In addition, risk-adjusted SMRs are calculated by dividing the expected number of deaths predicted by PIM1 by the observed number of deaths in each trust. We have used the original version of PIM with revised coefficients supplied by UK PICOS2 that give a better calibration as these coefficients are based on a more recent dataset. We have also produced SMRs using PIM 23 for 2006.

Unadjusted and risk-adjusted SMRs are presented by trust and year for 2004, 2005, 2006 and combined years in tables 47 - 49. PICU mortality funnel plots for the same periods are presented in figures 47a - 50b to provide a visual means of comparing unadjusted and adjusted SMRs between trusts, without imposing the ranking observed in league tables. Figure 49c presents risk-adjusted mortality using PIM 2.

The SMRs are plotted on the y-axis against the number of admissions to the trust on the x-axis. Higher mortality rates are represented by points plotted above the line of unity, with those appearing outside the upper control limit indicating an unusual excess mortality. Lower mortality rates are represented by points plotted below the line of unity and those falling below the lower control limit indicate unusually low mortality. In order to satisfy the condition, that if the overall distribution of the mortality ratios is random, there exists an approximately 5% chance of a unit falling outside the control limits, then the upper and lower control limits constructed at an individual unit level must represent not 95% confidence intervals, but 99.9% confidence intervals around a mortality ratio of one by number of admissions.4 This is analogous to increasing the confidence interval (or significance level) when correcting for multiple comparisons in data containing numerous groups. This means that the funnel plots are drawn in such a way that there is an approximately 5% chance of a unit falling outside the control limits if the distribution of SMRs is random.

In figures 50c and 50d, risk-adjusted SMRs by SHA, pre and post the July 2006 NHS organisation, have been produced by allocating children to the SHA in which they were living based on their address at admission. These ratios have then been expressed as a percentage and mapped to illustrate the range of variability in SMRs between SHAs. It should be noted that these ratios have not been subject to any spatial smoothing and confidence intervals are relatively wide in areas of low population. Tables 51 - 55 present 30-day follow-up data by age, sex and trust.

15.1 References

  1. Shann F, Pearson G, Slater A, Wilkinson K, Paediatric index of mortality (PIM): a mortality prediction model for children in intensive care. Intensive Care Med 1997; 23:201-207.
  2. Brady AR, Harrison D, Black S, Jones S, Rowan K, Pearson G, Ratcliffe J, Parry GJ, on behalf of the UK PICOS Study Group. Assessment and Optimization of Mortality Prediction Tools for Admissions to Pediatric Intensive Care in the United Kingdom. Pediatrics 2006; 117: 733-742.
  3. Shann F, Slater A, Pearson G. PIM 2: a revised version of the Paediatric Index of mortality. Intensive Care Med 2003; 29:278-285
  4. Spiegelhalter D. Funnel plots for institutional comparison. Quality and Safety in Health Care 2002;11(4):390-391.

16 DATA ON INDIVIDUAL CHILDREN

In all other chapters of this report, PICU activity is presented for episodes of care or admissions. This chapter describes activity related to 31,320 individual patients representing the 42,221 admissions (0 - 15 years) during 2004 - 2006.

Firstly, Table 56 summarises admissions by the source of their previous admission (same or other trust or single admission only). Table 57 reports the number of children having repeat admissions by trust and Table 58 the number of children admitted by diagnostic group. Table 59 summarises the number of children admitted by diagnostic group either once to a single trust, more than once to the same trust or more than once to more than 1 trust.

17 PREVALENCE FOR ADMISSION

Age and sex specific prevalence for admission to PICUs in England and Wales has been calculated with 95% Poisson confidence intervals using population counts from the 2001 Census1 (table 60). Age-sex standardised prevalence for the childhood population (less than 16 years) by SHA and HB (pre and post the October 2006 NHS reorganisation - tables 61a and 61b). These are mapped in figures 61a and 61b respectively.

Children were allocated to an SHA / HB using their residential address at admission. Addresses were validated using AFD Postcode Plus2 address validation software to obtain a correct postcode. Using the National Statistics Postcode Directory (http://www.statistics.gov.uk/geography/nspd.asp), postcodes were then linked to SHA / HB.

We have also presented age-sex standardised prevalence by 2006 primary care organisation (PCO) in figure 61c.

Prevalence for Scotland is not presented as PICANet currently only receives data from the Royal Hospital for Sick Children, Edinburgh.

  1. Office for National Statistics. 2001 Census : Census Area Statistics (England and Wales) [computer file]. ESRC/JISC Census Programme, Census Dissemination Unit, MIMAS (University of Manchester).
  2. AFD Refiner Q.2/07. AFD Software Ltd, Lough House, Approach Road, Ramsey, ISLE OF MAN, IM8 1RG, UK, 2007.

18 CHILDREN RECEIVING CARE IN ADULT INTENSIVE CARE UNITS

Data on children (under 16 years) treated in adult intensive care units (AICUs), including age in months, sex, date of admission and discharge, outcome and discharge location and admission diagnosis, were provided by the Intensive Care National Audit & Research Centre (ICNARC) and the South West Audit of Critically Ill Children (SWACIC). These data are summarised in tables 62 - 67. Analysis is restricted to 2005. ICNARC receives data from 74% of AICUs in England.

Signed consent was obtained from the unit director of each AICU. One AICU providing data to SWACIC did not give explicit permission for PICANet to receive their data.

19 DATA QUALITY
Dr Krish Thiru

PICANet has now embarked on its sixth year of data collection. Its does so with the knowledge that it has established one of the highest quality national core datasets in paediatric medicine within the UK.

Considerable effort has been made by both PICU staff and the PICANet team to ensure that the data is of the highest quality. During previous years, the PICANet team visited individual units to review a sample of records to cross check that the data submitted to PICANet corresponded to that data held in the unit's paper records and clinical information systems. Validation visits were suspended due to staff shortages but will resume this year.

This chapter details improvements in data quality during last year and highlights areas needing attention. The results are presented by NHS Trust as well as by unit to acknowledge the importance of unit level data management.

19.1 Data quality assurance processes

  1. At input, internal logical, consistency and range checks are carried out at input by the PICANet software with an on-screen summary of outstanding validation checks on completion of a record. Units importing data from their own databases or commercial software are provided with an import log detailing which records have been imported and outstanding validation issues.
  2. Data transmitted to the PICANet central server in Leeds are subject to a series of additional validation checks (including address and postcode validation and clinical coding verification). Data validation reports (DVRs) are returned via email (Appendix G).
  3. Units are provided with monthly admission reports (Appendix H) and asked to cross check these with local patient registers (e.g. unit admission book).
  4. Units are provided with error status reports (Appendix I) which highlight particular dimensions of data quality that require attention, these include the number of missing values returned.

Full details of the PICANet data quality control and assurance processes are provided in the PICANet National Report 2003 - 2004.

The completeness level for all data items collected by PICANet are given in Table DQ1, showing 94.7% completeness of the data items. Table DQ2 details the completeness of the data by month by year for the last 3 years, while table DQ3 provides a breakdown by individual unit for the combined 3 years. The PICANet dataset contains 4.7% of exception values (i.e. data collected as 'not recorded' or 'not known') and with 0.6% left blank. Figure DQ1 highlights twelve data items found to have the largest number of exception or blank values.

Table DQ1 Data completeness

FIELDEligibleCompleteIncomplete
ValidExceptionsTotalInvalidBlankTotal
n%n%n%n%n%n%
ADDATE4314043140(100.0)0(0.0)43140(100.0)0(0.0)0(0.0)0(0.0)
ADDRESS14314043119(100.0)0(0.0)43119(100.0)0(0.0)21(0.0)21(0.0)
ADNO4314043139(100.0)0(0.0)43139(100.0)0(0.0)1(0.0)1(0.0)
ADTIME4314043135(100.0)0(0.0)43135(100.0)0(0.0)5(0.0)5(0.0)
ADTYPE4314043023(99.7)116(0.3)43139(100.0)0(0.0)1(0.0)1(0.0)
APDIAG4314043140(100.0)0(0.0)43140(100.0)0(0.0)0(0.0)0(0.0)
BASEEXCESS3307027101(82.0)5969(18.0)33070(100.0)0(0.0)0(0.0)0(0.0)
BGFIRSTHR2420623192(95.8)1010(4.2)24202(100.0)0(0.0)4(0.0)4(0.0)
BPSYS4314036854(85.4)6191(14.4)43045(99.8)0(0.0)95(0.2)95(0.2)
CAREAREAAD4251641118(96.7)1396(3.3)42514(100.0)0(0.0)2(0.0)2(0.0)
CASENO4314043139(100.0)0(0.0)43139(100.0)0(0.0)1(0.0)1(0.0)
DELORDER13421141(85.0)201(15.0)1342(100.0)0(0.0)0(0.0)0(0.0)
DISPALCARE4091440297(98.5)617(1.5)40914(100.0)0(0.0)0(0.0)0(0.0)
DOB4313143131(100.0)0(0.0)43131(100.0)0(0.0)0(0.0)0(0.0)
DOBEST4314043132(100.0)6(0.0)43138(100.0)0(0.0)2(0.0)2(0.0)
DOD26762668(99.7)0(0.0)2668(99.7)0(0.0)8(0.3)8(0.3)
ECMO4314041985(97.3)1153(2.7)43138(100.0)0(0.0)2(0.0)2(0.0)
ETHNIC4314043138(100.0)0(0.0)43138(100.0)0(0.0)2(0.0)2(0.0)
FAMILYNAME4314043131(100.0)0(0.0)43131(100.0)0(0.0)9(0.0)9(0.0)
FIO23183125080(78.8)5861(18.4)30941(97.2)0(0.0)890(2.8)890(2.8)
FIRSTNAME4314043130(100.0)0(0.0)43130(100.0)0(0.0)10(0.0)10(0.0)
FU30DISSTATUS3953619598(49.6)19857(50.2)39455(99.8)0(0.0)81(0.2)81(0.2)
FU30LOCATION1918516803(87.6)2381(12.4)19184(100.0)0(0.0)1(0.0)1(0.0)
FU30LOCHOSP32323125(96.7)107(3.3)3232(100.0)0(0.0)0(0.0)0(0.0)
GEST2496616487(66.0)8477(34.0)24964(100.0)0(0.0)2(0.0)2(0.0)
HEADBOX3183130008(94.3)1475(4.6)31483(98.9)0(0.0)348(1.1)348(1.1)
ICPDEVICE2420623468(97.0)735(3.0)24203(100.0)0(0.0)3(0.0)3(0.0)
INTTRACHEOSTOMY4314041865(97.0)1273(3.0)43138(100.0)0(0.0)2(0.0)2(0.0)
INTUBATION3183130913(97.1)580(1.8)31493(98.9)0(0.0)338(1.1)338(1.1)
INTUBDAYS2828(100.0)0(0.0)28(100.0)0(0.0)0(0.0)0(0.0)
INTUBEVER4314043140(100.0)0(0.0)43140(100.0)0(0.0)0(0.0)0(0.0)
INVVENT4312741848(97.0)1278(3.0)43126(100.0)0(0.0)1(0.0)1(0.0)
INVVENTDAY2870628532(99.4)172(0.6)28704(100.0)0(0.0)2(0.0)2(0.0)
LVAD4314041981(97.3)1157(2.7)43138(100.0)0(0.0)2(0.0)2(0.0)
MECHVENT4314042596(98.7)538(1.2)43134(100.0)0(0.0)6(0.0)6(0.0)
MEDHISTEVID4314042630(98.8)503(1.2)43133(100.0)0(0.0)7(0.0)7(0.0)
MULT4314033469(77.6)9668(22.4)43137(100.0)0(0.0)3(0.0)3(0.0)
NHSNO4314032268(74.8)1620(3.8)33888(78.6)0(0.0)9252(21.4)9252(21.4)
NONINVVENT4314041719(96.7)1419(3.3)43138(100.0)0(0.0)2(0.0)2(0.0)
NONINVVENTDAY52875267(99.6)19(0.4)5286(100.0)0(0.0)1(0.0)1(0.0)
PAO23307023077(69.8)9991(30.2)33068(100.0)0(0.0)2(0.0)2(0.0)
POSTCODE4314043103(99.9)0(0.0)43103(99.9)0(0.0)37(0.1)37(0.1)
PREVICUAD4314042538(98.6)602(1.4)43140(100.0)0(0.0)0(0.0)0(0.0)
PRIMDIAG4314042963(99.6)0(0.0)42963(99.6)37(0.1)140(0.3)177(0.4)
PRIMREASON2420623596(97.5)598(2.5)24194(100.0)0(0.0)12(0.0)12(0.0)
PUPREACT4314039136(90.7)3999(9.3)43135(100.0)0(0.0)5(0.0)5(0.0)
RENALSUPPORT2420623482(97.0)721(3.0)24203(100.0)0(0.0)3(0.0)3(0.0)
RETRIEVAL4314042954(99.6)178(0.4)43132(100.0)0(0.0)8(0.0)8(0.0)
RETRIEVALBY1479614362(97.1)411(2.8)14773(99.8)0(0.0)23(0.2)23(0.2)
SEX4314043090(99.9)43(0.1)43133(100.0)7(0.0)0(0.0)7(0.0)
SOURCEAD4314042977(99.6)163(0.4)43140(100.0)0(0.0)0(0.0)0(0.0)
TIMEDTH22122212(100.0)0(0.0)2212(100.0)0(0.0)0(0.0)0(0.0)
UNITDISDATE4312743121(100.0)0(0.0)43121(100.0)0(0.0)6(0.0)6(0.0)
UNITDISDEST4091440562(99.1)351(0.9)40913(100.0)0(0.0)1(0.0)1(0.0)
UNITDISDESTHOSP3966135955(90.7)3706(9.3)39661(100.0)0(0.0)0(0.0)0(0.0)
UNITDISSTATUS4314043126(100.0)1(0.0)43127(100.0)0(0.0)13(0.0)13(0.0)
UNITDISTIME4312743113(100.0)0(0.0)43113(100.0)0(0.0)14(0.0)14(0.0)
VASOACTIVE4314041815(96.9)1323(3.1)43138(100.0)0(0.0)2(0.0)2(0.0)
Total20311401923860(94.7)95866(4.7)2019726(99.4)44(0.0)11370(0.6)11414(0.6)

Figure DQ1 Percentage of exception or blank values in the PICANet dataset

Figure DQ1 Percentage of exception or blank values in the PICANet dataset

Note: Full description of variables are provided in the PICANet Data Definitions Manual

Some of these data items are used in the calculation of the Paediatric Index of Mortality (PIM) 2. PICANet is investigating the impact of missing data on this risk adjustment index. Thirty-day follow-up status is a standard patient care outcome measure used across the NHS. Within PICANet, 30 day follow-up data is 99% complete, however 50% of this data is recoded as 'not known'. The distribution of 30 day follow-up data collection across PICANet units is detailed in figure DQ2.

Figure DQ2 Data completeness for 30-day follow-up information

Figure DQ2 Data completeness for 30-day follow-up information

The NHS Number is a unique patient identifier that provides a common link between patient records across the NHS. The number can be used by Trust Patient Administration Systems/Patient Information Systems to easily and reliably link to the PICANet database.

Table DQ2 Data completeness by year (all variables)

YearMonthEligibleCompletion
CompleteIncomplete
ValidExceptionsTotalInvalidBlankTotal
n%n%n%n%n%n%
 
200415868255491(94.6)2744(4.7)58235(99.2)4(0.0)443(0.8)447(0.8)
 25523152175(94.5)2645(4.8)54820(99.3)1(0.0)410(0.7)411(0.7)
 35833354873(94.1)2987(5.1)57860(99.2)1(0.0)472(0.8)473(0.8)
 45260649716(94.5)2515(4.8)52231(99.3)4(0.0)371(0.7)375(0.7)
 55191348991(94.4)2506(4.8)51497(99.2)2(0.0)414(0.8)416(0.8)
 65288850072(94.7)2421(4.6)52493(99.3)0(0.0)395(0.7)395(0.7)
 74990447047(94.3)2508(5.0)49555(99.3)1(0.0)348(0.7)349(0.7)
 84935046474(94.2)2512(5.1)48986(99.3)0(0.0)364(0.7)364(0.7)
 95050547714(94.5)2426(4.8)50140(99.3)1(0.0)364(0.7)365(0.7)
 105138548414(94.2)2549(5.0)50963(99.2)0(0.0)422(0.8)422(0.8)
 115515852056(94.4)2592(4.7)54648(99.1)4(0.0)506(0.9)510(0.9)
 125679953729(94.6)2667(4.7)56396(99.3)0(0.0)403(0.7)403(0.7)
2004 Total642754606752(94.4)31072(4.8)637824(99.2)18(0.0)4912(0.8)4930(0.8)
 
200515568651916(93.2)3340(6.0)55256(99.2)1(0.0)429(0.8)430(0.8)
 25256249017(93.3)3199(6.1)52216(99.3)1(0.0)345(0.7)346(0.7)
 35636052425(93.0)3601(6.4)56026(99.4)0(0.0)334(0.6)334(0.6)
 45206548634(93.4)3111(6.0)51745(99.4)0(0.0)320(0.6)320(0.6)
 55515051706(93.8)3150(5.7)54856(99.5)4(0.0)290(0.5)294(0.5)
 65810454620(94.0)3180(5.5)57800(99.5)1(0.0)303(0.5)304(0.5)
 75781054293(93.9)3167(5.5)57460(99.4)4(0.0)346(0.6)350(0.6)
 85381750577(94.0)2932(5.4)53509(99.4)0(0.0)308(0.6)308(0.6)
 95657053105(93.9)3147(5.6)56252(99.4)6(0.0)312(0.6)318(0.6)
 105483951642(94.2)2874(5.2)54516(99.4)1(0.0)322(0.6)323(0.6)
 116150557892(94.1)3300(5.4)61192(99.5)2(0.0)311(0.5)313(0.5)
 126113257278(93.7)3533(5.8)60811(99.5)4(0.0)317(0.5)321(0.5)
2005 Total675600633105(93.7)38534(5.7)671639(99.4)24(0.0)3937(0.6)3961(0.6)
 
200616515362395(95.8)2576(4.0)64971(99.7)0(0.0)182(0.3)182(0.3)
 25911456608(95.8)2336(4.0)58944(99.7)0(0.0)170(0.3)170(0.3)
 36311060567(96.0)2375(3.8)62942(99.7)0(0.0)168(0.3)168(0.3)
 45742254962(95.7)2280(4.0)57242(99.7)1(0.0)179(0.3)180(0.3)
 56015257785(96.1)2188(3.6)59973(99.7)0(0.0)179(0.3)179(0.3)
 65788755633(96.1)2049(3.5)57682(99.6)0(0.0)205(0.4)205(0.4)
 75657354386(96.1)1974(3.5)56360(99.6)0(0.0)213(0.4)213(0.4)
 85607053821(96.0)2051(3.7)55872(99.6)0(0.0)198(0.4)198(0.4)
 95483852655(96.0)1958(3.6)54613(99.6)0(0.0)225(0.4)225(0.4)
 105976557450(96.1)2126(3.6)59576(99.7)0(0.0)189(0.3)189(0.3)
 116177759424(96.2)2109(3.4)61533(99.6)0(0.0)244(0.4)244(0.4)
 126092558317(95.7)2238(3.7)60555(99.4)1(0.0)369(0.6)370(0.6)
2006 Total712786684003(96.0)26260(3.7)710263(99.6)2(0.0)2521(0.4)2523(0.4)
 
Total20311401923860(94.7)95866(4.7)2019726(99.4)44(0.0)11370(0.6)11414(0.6)

The distribution of NHS number recording in PICANet units is detailed in table DQ4 and in figure DQ3 below. 25% of patients within PICANet do not have NHS numbers.

Table DQ3 Data completeness by PICU

PICUEligibleCompleteIncomplete
ValidExceptionsTotalInvalidBlankTotal
n%n%n%n%n%n%
A6081353744(88.4)6495(10.7)60239(99.1)0(0.0)574(0.9)574(0.9)
B3486432734(93.9)1582(4.5)34316(98.4)0(0.0)548(1.6)548(1.6)
C4120640316(97.8)877(2.1)41193(100.0)0(0.0)13(0.0)13(0.0)
D8537083469(97.8)1819(2.1)85288(99.9)0(0.0)82(0.1)82(0.1)
E234799225582(96.1)8123(3.5)233705(99.5)0(0.0)1094(0.5)1094(0.5)
F159625151325(94.8)7457(4.7)158782(99.5)9(0.0)834(0.5)843(0.5)
G62866181(98.3)102(1.6)6283(100.0)0(0.0)3(0.0)3(0.0)
H4599642271(91.9)3140(6.8)45411(98.7)0(0.0)585(1.3)585(1.3)
I126319123715(97.9)2282(1.8)125997(99.7)0(0.0)322(0.3)322(0.3)
J1182510931(92.4)570(4.8)11501(97.3)0(0.0)324(2.7)324(2.7)
K14147639747(95.8)1449(3.5)41196(99.3)0(0.0)280(0.7)280(0.7)
K24820746105(95.6)1978(4.1)48083(99.7)0(0.0)124(0.3)124(0.3)
K33964037546(94.7)1946(4.9)39492(99.6)0(0.0)148(0.4)148(0.4)
L3944038133(96.7)966(2.4)39099(99.1)0(0.0)341(0.9)341(0.9)
M5523553394(96.7)1748(3.2)55142(99.8)0(0.0)93(0.2)93(0.2)
N4367441687(95.5)1608(3.7)43295(99.1)0(0.0)379(0.9)379(0.9)
O8779682105(93.5)4893(5.6)86998(99.1)0(0.0)798(0.9)798(0.9)
P151790146637(96.6)5048(3.3)151685(99.9)0(0.0)105(0.1)105(0.1)
Q11166811130(95.4)520(4.5)11650(99.8)0(0.0)18(0.2)18(0.2)
Q26701364263(95.9)2469(3.7)66732(99.6)0(0.0)281(0.4)281(0.4)
R9651194700(98.1)1359(1.4)96059(99.5)0(0.0)452(0.5)452(0.5)
S2559324299(94.9)1088(4.3)25387(99.2)0(0.0)206(0.8)206(0.8)
T5622752094(92.6)3506(6.2)55600(98.9)0(0.0)627(1.1)627(1.1)
U5546951274(92.4)3424(6.2)54698(98.6)0(0.0)771(1.4)771(1.4)
V139071122461(88.1)15027(10.8)137488(98.9)35(0.0)1548(1.1)1583(1.1)
W9619189415(93.0)6049(6.3)95464(99.2)0(0.0)727(0.8)727(0.8)
X17382969980(94.8)3829(5.2)73809(100.0)0(0.0)20(0.0)20(0.0)
X25311948135(90.6)4911(9.2)53046(99.9)0(0.0)73(0.1)73(0.1)
Y4208840487(96.2)1601(3.8)42088(100.0)0(0.0)0(0.0)0(0.0)
Grand Total20311401923860(94.7)95866(4.7)2019726(99.4)44(0.0)11370(0.6)11414(0.6)

Table DQ4 Data completeness for NHS number by NHS trust

NHS trustEligibleValidExceptionsInvalidBlank
n%n%n%n%
A13280(0.0)755(56.9)0(0.0)573(43.1)
B763404(52.9)0(0.0)0(0.0)359(47.1)
C851848(99.6)0(0.0)0(0.0)3(0.4)
D17801772(99.6)8(0.4)0(0.0)0(0.0)
E49933900(78.1)0(0.0)0(0.0)1093(21.9)
F34112667(78.2)0(0.0)0(0.0)744(21.8)
G132131(99.2)0(0.0)0(0.0)1(0.8)
H979431(44.0)0(0.0)0(0.0)548(56.0)
I26782356(88.0)0(0.0)0(0.0)322(12.0)
J25320(7.9)0(0.0)0(0.0)233(92.1)
K27452475(90.2)45(1.6)0(0.0)225(8.2)
L844505(59.8)0(0.0)0(0.0)339(40.2)
M11591153(99.5)0(0.0)0(0.0)6(0.5)
N912714(78.3)0(0.0)0(0.0)198(21.7)
O18261029(56.4)0(0.0)0(0.0)797(43.6)
P31463080(97.9)0(0.0)0(0.0)66(2.1)
Q16971677(98.8)7(0.4)0(0.0)13(0.8)
R19941683(84.4)0(0.0)0(0.0)311(15.6)
S549537(97.8)0(0.0)0(0.0)12(2.2)
T1241614(49.5)0(0.0)0(0.0)627(50.5)
U1175408(34.7)0(0.0)0(0.0)767(65.3)
V29811715(57.5)0(0.0)0(0.0)1266(42.5)
W20351311(64.4)0(0.0)0(0.0)724(35.6)
X27872762(99.1)0(0.0)0(0.0)25(0.9)
Y88176(8.6)805(91.4)0(0.0)0(0.0)
Total4314032268(74.8)1620(3.8)0(0.0)9252(21.4)

In the absence of the NHS Number it is difficult to definitively link patients with additional data repositories. PICANet is establishing a linked data set with Hospital Episode Statistics data. The NHS number is a crucial item of data which will enable long term follow-up and outcomes study of PICU patients.

Figure DQ3 Data completeness for NHS number

Figure DQ3 Data completeness for NHS number

Over the coming year, PICANet will be implementing the collection of the Paediatric Critical Care Minimum Data Set. The consequences will be a greater volume of data for units and PICANet but the importance of quality assurance processes will remain. A collaborative approach to data quality, with regular and timely feedback from PICANet to units, will ensure that the PICANet dataset remains of the highest standard.

20 A CLINICIAN'S COMMENTARY
Dr Gale Pearson

When PICANet was first conceived in the early 1990's the available information suggested that only 40% of intensive care admissions of children in the United Kingdom were to the 28 paediatric intensive care units. The other 60% were looked after in adult / general units or on ordinary paediatric wards. This data came from surveys by the then British Paediatric Association (now the Royal College of Paediatrics and Child Health). In those days (not so long ago) intensive care was delivered to children in so many different sites that it was estimated that these surveys captured at best only 80% of child ICU 'admissions'. The data suggested that the average PICU had 3.6 staffed beds and had only 236 admissions per year. 22% of PICUs had no consultant in their employ with a special interest in the care of critically ill children. Outcomes were not monitored and not included in the surveys so no risk adjustment model could be applied to the data. A later study that did use risk adjustment (published in 1997)2 compared a representative sample of the British system (in Trent) with a more centralised system (Victoria, Australia).3 This study made the strong suggestion that such fragmentation of the intensive care service for children was associated with a prohibitive excess mortality. At the time most doctors wishing to train in paediatric intensive care spent the most significant parts of their training abroad.

By the time PICANet started collecting data, the team had validated the available risk adjustment models against British data and much more of the clinical care was already being provided in designated paediatric intensive care units by specialists in paediatric intensive care. The centralisation of care had largely been achieved by the expansion of 'lead centres' as a result of the recommendations of a national coordinating group. The clinical advisors to this group had also strongly endorsed the formation of PICANet. As a consequence PICANet data postdates the shift in service provision of the late 1990's and PICANet cannot report on any effect that the changes may have had. However I would assert that PICANet has been part of a substantial improvement in paediatric intensive care standards and that it has an enormous potential for future contribution. This is firstly because audit is an essential component of good clinical practice. Units that don't audit their performance are arguably not providing a good quality of care. The PICANet audit is as scrupulous as it can be (within budget) in terms of data collection, validation and analysis. Thanks to PICANet we now work in an era where the public, commissioners, providers and patients can all be reassured that the risk adjusted performance of the PICUs in England and Wales is monitored and has remained acceptable (within statistical confidence limits) for the past three years.

More children receive intensive care now, than at the time of the British Paediatric Association surveys. In the years for which data are presented in this report there has also been a small but perceptible continuing increase in the numbers of patients treated in paediatric intensive care units. It is not clear if the threshold for PICU admission across the PICUs is falling, whether there is a continued gravitation of patients to PICUs from general ICUs in referring units or whether intensive care provision is beginning to reach a group of children who always needed it but in the past did not receive it, as was the accusation in 1993. A workload of the order of 14000 admissions per annum in England & Wales is currently distributed between 25 NHS trusts representing 29 PICUs some of which still report very small volumes of activity and others high refusal rates. Nevertheless these figures translate to an average of 563 admissions per trust across the three years (median 443; IQR: 297 - 868). This volume of activity is proving sufficient to provide credible training in the specialty without overseas travel.

PICANet is a clinically conceived initiative, supported by the Department of Health with funding that may soon be channelled through the Health Care Commission. Clinicians (like their public health colleagues, commissioners, government and any patient's advocate) are very interested in bench marking their units and comparing the variety of clinical practice and performance. They need reassurance over the equity of service provision and the opportunity to improve clinical practice through thorough and rigorous research. PICANet provides an unparalleled resource in all these respects and has opportunities to further improve the service that it provides.

20.1 Inter unit comparison

Inter unit comparison has become much easier since PICANet dropped the practice of anonymising units in deference to the Freedom of Information Act1. Just scanning through the tables in this report we can see that amongst children, the age distribution of patients within our units is largely comparable between units and between years. Although over­age (16yrs plus) admissions and admissions in the higher mortality risk groups are more common in the larger / higher volume units. Both trends may reflect different attributes of specialist services in those units such as cardiac surgery. Smaller units also appear to experience much greater fluctuation in the numbers of admissions month by month.

One dramatic demonstration of a clinical difference between our units is the great variation in the proportion of patients receiving invasive ventilation, larger / higher volume units having a greater percentage of invasively ventilated patients. This may be related to the selected provision of specialist services but it could also reflect different relationships between supply and demand in intensive care in different regions. The location of high dependency services also has to be taken into account. If one has spare intensive care unit capacity it is not difficult to envision it being used to provide high dependency care (in which case the effect on the performance analysis is to improve the risk-adjusted outcome). PICANet could compensate for this latter effect by comparing risk adjusted outcome data for invasively ventilated patients as a separate group in the future.

PICANet usefully presents two types of standardised mortality, in funnel plot format, against the number of admissions per unit per year. The first is unadjusted by severity of illness, the standardisation being merely against the average mortality in the data. Some units (including the one where I work) appear close to or outside the confidence limits in this respect. In the second plot these figures are adjusted to account for the severity of illness at presentation. Pleasingly risk adjustment corrects for apparent outlying behaviour. Nevertheless where these effects persist year upon year it remains for these units to reassure us with clinical explanations as to how their case-mix ends up with greater proportions of high risk patients and for PICANet to look for other explanations within the data. PICANet has formal procedures for these sort of 'quality assurance' enquiries that have been tted in earnest at least once since it was formed.

20.2 Equity of provision

PICANet only collects information on children who are admitted to the units participating in the audit. This includes all the PICUs in England and Wales and the PICU in Edinburgh. In the future the addition of data from the unit in Glasgow will provide more comprehensive cover of activity in Scotland. However significant numbers of children still receive care in adult / general intensive care units. Some of these contribute outline data to PICANet but despite our best efforts these units do not yet supply risk adjusted data to the audit. There have also been observations during the PICANet era that large numbers of children have been turned away from the PICU of first referral in some regions (such as the West Midlands where I work). Certainly the data cannot yet reassure us that all those children who require intensive care in fact receive it, at all or in a timely fashion. The presentation of data with geographical denominators rather than inter-unit based comparison is the way forward in this respect, especially the presentation of patient flows. The real opportunity to pin this subject down will come from the analysis of the 'Referral and retrieval dataset' which is now being introduced as an addition to the core data collection.

20.3 Research

PICANet enjoys a close and productive association with the Paediatric Intensive Care Society Study Group. Notable examples being the epidemiological analyses of traumatic brain injury, ethnicity and the analysis of seasonal respiratory admissions in the under ones (which was used to advise the joint committee on vaccination and immunisation on the timing of active immunisation programmes against respiratory syncitial virus). PICANet is a well tapped resource, providing services which include baseline data and denominators, power calculations for clinical trials and the opportunity to act as a vehicle for comprehensive data collection. This latter facility has also been used to allow units the facility to easily collect the paediatric critical care minimum dataset (part of 'payment by results') which will be necessary from October 2007. PICANet also contributes to the academic field of risk adjustment in health care services audit with notable current projects being latent class analysis to evaluate the impact of missing data on apparent performance and the development of techniques which will allow faster feedback to participating units on their performance within clinically useful timeframes. In common with the Australia and New Zealand Paediatric Intensive Care Society these faster feedback approaches will use CUSUM based techniques such as sequential probability ratio testing.

In summary I think PICANet is a tremendous achievement and should act as a great reassurance to patients and their parents in an environment where public faith in the NHS is continually shaken. The audit has enormous clinical support and is a great resource that is just starting to realise its potential. I look forward to a productive relationship with the Health Care Commission.

Gale Pearson
Chair of the Clinical Advisory Group of PICANet

20.4 References

  1. Freedom of Information Act 2000. [Online] [Accessed 05/06/2007] Available from the World Wide Web at http://www.legislation.hmso.gov.uk/acts/acts2000/20000036.htm.
  2. Shann F, Pearson G, Slater A, Wilkinson K, Paediatric index of mortality (PIM): a mortality prediction model for children in intensive care. Intensive Care Med 1997; 23:201-207.
  3. Pearson G, Shann F, Barry P, Vyas J, Thomas T, Powell C, Field D. Should paediatric intensive care be centralised? Trent versus Victoria. Lancet 1997; 349: 1213-1217.

21 THE PAEDIATRIC CRITICAL CARE MINIMUM DATASET (PCCMDS), HEALTHCARE RESOURCE GROUPS (HRGS) AND PAYMENT BY RESULTS (PBR)
Dr Kevin Morris

Healthcare Resource Groups (HRGs) have been used for a number of years as a way of classifying diseases and interventions, in relation to the amount of healthcare resources that they consume. HRG Version 3.5 has been in use since October 2003. The latest update (Version 4) is very important in assisting the Department of Health in implementing Payment by Results (PbR). An Expert Working Group (EWG), under the Chairmanship of Dr Nick Griffin, working on Version 4 HRGs for Paediatrics, identified the need to consider how Paediatric and Neonatal Intensive Care could be included. The NHS Information Centre established two further EWGs to take this work forward.

The Paediatric Critical Care EWG was established in 2004, chaired by Dr Nick Griffin, with Ian Hughes as Project Manager. Membership (Appendix L) included medical and nursing PIC and HDU representatives, commissioners, an NHS Finance director, PICANet, casemix consultants and Professor Stuart Tanner, representing the Department of Health. The EWG was charged with defining a Paediatric Critical Care Minimum Dataset that would in turn define a number of HRGs. In contrast to most HRGs, which are applied to a complete hospital episode, it was acknowledged that a daily HRG would be appropriate for all critical care episodes, whether neonatal, paediatric or adult. The EWG was also asked to include HDU levels of care, as distinct from the adult critical care dataset, which specifically does not include HDU or Level 1 patients.

Early discussions identified the lack of an existing system that was validated and could be easily adopted. A number of candidate approaches were discussed which included:

  1. use of a Therapeutic Index Scoring System (TISS) or the recently published 'improved TISS', called the Nursing Activity Score (NAS), which in adult critical care has been shown to be a useful predictor of resource use,
  2. a system based on further development and refinement of the existing PICS Levels of Care,
  3. a system based on the number of organ system supports used, the system adopted for adult critical care HRGs.

The PICS Levels of Care have proved to be of value in defining PICU patient dependency, but have not been evaluated as a tool for measuring resource use and in their current form are not precise enough to reliably assign a level of care.

As 75-85% of costs associated with critical care are staff costs, it was agreed that it would be useful to collect information relating patient factors to use of staff resources. An observational study was undertaken in 10 PICUs, to evaluate the three models discussed above against the staff resource consumed by each patient. Observers were present for six hours and recorded the PICU staff present at each bedspace every 10 minutes. They also recorded the seniority of each staff member, in order to assign a cost, and included nursing staff, medical staff, technical support staff and physiotherapists. They did this for three days in succession and completed three blocks of three days separated by 3-4 weeks. For the same observational period of three months, shift by shift data was collected on the number and types of intervention, diagnosis, the number of organ supports and the Nursing Activities Score. The NHS Casemix team undertook a detailed statistical analysis of each approach and how well it predicted staff resource use. Alongside the observational study, each PICU provided detailed bottom-up costing for the financial year 2005/2006. The 10 PICUs involved were selected to include a mixture of large, intermediate and smaller units and those with single multidisciplinary units and those with split site working.

21.1 What were the findings of the observational study and the costings exercise?

The costings study found that 83% of PICU costs were related to staff costs, with the single largest cost being nursing costs (Figure PCCMDS1).

Figure PCCMDS1 Breakdown of costs for each PICU

Figure PCCMDS1 Breakdown of costs for each PICU

Note: Data shown is mean value for 10 units

The Nursing Activities Score was found to be a poor predictor of staff resource use, as was the adult HRG model based on the number of organ systems supported.

A decision was taken to build up a system of HRGs based on a further refinement of the 'Levels of Care' approach. Allocation of a particular intervention to an HRG level was informed by the information available from the observational study - an intervention shown to be associated with a high staff resource being placed in a higher HRG category.

Considerable discussion took place regarding HDU levels of care. Whilst there are a number of existing systems that have been developed for HDU, a number of the data items that are included lack objectivity and cannot be measured in all patients. An absolute requirement for any item to be included in the minimum dataset is that it must be objective and measurable. An example would be the need for greater than 60% oxygen. This cannot be reliably quantified in an infant receiving supplemental oxygen via nasal cannulae. It was therefore necessary, in some situations, to modify a data item to make it acceptable for inclusion.

A key attribute of HRGs is that they are setting independent, that is to say, they apply whether the child is in a PICU at the time or in HDU or in a ward area. If a ward area collects the Paediatric Critical Care Minimum Dataset and identifies that an episode of critical care has occurred, then the relevant HRG will apply to that episode and in the future will be reimbursed under PbR.

21.2 What system of HRGs was chosen?

A system based on 7 HRGs was proposed:

HRG1 - High Dependency

HRG2 - High Dependency Advanced

HRG3 - Intensive Care Basic

HRG4 - Intensive Care Basic Enhanced

HRG5 - Intensive Care Advanced

HRG6 - Intensive Care Advanced Enhanced

HRG7 - Intensive Care - ECMO / ECLS

To define these levels, a Paediatric Critical Care Minimum Dataset of 32 items is necessary. To take account of the additional staff costs associated with nursing a patient in an isolation cubicle, this is included in the dataset. A list of medical conditions that define the need for isolation is also needed (based on ICD 10 codes). The HRG level assigned to a patient increases by one level if the patient is recorded as nursed in an isolation cubicle and having a relevant ICD 10 diagnosis that justifies isolation.

Further detail on the HRGs with the interventions that map to each level is shown in Appendix M.

Across the 10 PICUs in the study, the breakdown of cases over a three month period of activity data collection is shown in table P1.

Table P1 Breakdown of cases over 3 month period, according to HRG level

HRG levelHRG categoryApproximate percentage of patient activity
HRG 7Intensive Care - ECMO / ECLS1
HRG 6Intensive Care, advanced enhanced5
HRG 5Intensive care, advanced10
HRG 4Intensive care, basic enhanced20
HRG 3Intensive care, basic40
HRGs 1 & 2HDU20
—No HRG categorya3

a These cases would not attract a tariff under PbR in the future

An analysis was undertaken of the staff resource costs associated with each HRG. These are expressed as a cost ratio with Intensive Care Basic as the reference HRG with a value of 1.00.

HRG1 - High Dependency0.75
HRG2 - High Dependency Advanced0.91
HRG3 - Intensive Care Basic1.00
HRG4 - Intensive Care Basic Enhanced1.22
HRG5 - Intensive Care Advanced1.40
HRG6 - Intensive Care Adv Enhanced2.12
HRG7 - Intensive Care - ECMO / ECLS3.06

It should be stressed that the HDU level data was obtained predominantly within an ICU setting, which could have impacted on the nurse:patient ratio, with a higher nursing input than that delivered to the same patient in an HDU or ward setting.

21.3 How many HRGs will be allocated to a patient?

Each patient will be allocated a single 'parent' or 'core' HRG to cover the episode of hospitalisation. If more than one diagnostic code applies to a patient, the HRG with the highest tariff will apply. For example, a child admitted for cardiac surgery who develops a post-operative pneumonia will be allocated the HRG related to the cardiac surgery.

In addition, a number of aspects of care are 'unbundled' from the parent HRG. These include:

  1. chemotherapy
  2. radiotherapy
  3. interventional radiology
  4. diagnostic imaging eg MRI
  5. rehabilitation
  6. renal dialysis
  7. critical care (including PICU)
  8. specialist palliative care
  9. high cost drugs

The list of all eligible high cost drugs will be updated on a regular basis, but the current list includes many drugs that will be used in PICU (Appendix N).

If the above patient admitted for cardiac surgery spends four days in PICU, is treated with Sildenafil and requires an MRI scan of the brain, then the number of HRGs will be:

Core HRG
+
4 PICU HRGs (depending on level(s) required)
+
High cost drug HRG for Sildenafil
+
Diagnostic imaging HRG for MRI scan

21.4 What about patient transport services?

Any internal transport of a PICU patient within an institution will not receive a separate HRG and will be covered within the daily HRG.

Currently, retrieval is not included within the HRG classification, but work is ongoing to look at the costs associated with both neonatal and paediatric transport. It is quite likely that transport will form another unbundled category of care to add to the list of nine items shown above.

21.5 How will the Paediatric Critical Care Minimum Dataset be collected?

It will be up to individual units and Trusts to decide on how to collect this data. The quality of the data will be important and if no patient data is collected, it will not be possible to be reimbursed under PbR in the future.

The publication of the Dataset Change Notice (DSCN) by the NHS Information Standards Board mandates the providers of IT development, under the umbrella of Connecting for Health, to provide Trusts will the ability to collect the PCCMDS. However, many Trusts are not yet covered by these developments and will need to seek a local solution.

The PCCMDS will be incorporated into the PICANet software, allowing participating units to collect data by this route, if they choose.

Systems must be developed within a unit to ensure complete and high quality data. If an intervention occurs at any time within a 24 hour period, it should be recorded even if it was only started at 23:50 at night. A patient who is discharged at 08:00 should have data collected, as they are eligible for a critical care HRG for that 24 hour period.

Greater difficulty is likely to be experienced in the collection of accurate data in patients who are cared for in ward areas. Cohorting of sicker ward patients into a limited number of HDU areas should both optimise their care and facilitate PCCMDS data collection.

21.6 Should a pre-term neonate looked after in PICU or a ward area have the Neonatal Critical Care Minimum Dataset collected rather than PCCMDS?

No, the PCCMDS will be collected for all patients in a hospital environment that predominantly looks after children. Equally, if a 25 year old is admitted to PICU, the PCCMDS should be collected.

21.7 What are the key milestones over the next few years?

Collection of the PCCMDS is mandatory from October 2007. The data that is collected will be analysed alongside costings information provided by Trusts as part of the Reference Costs exercise, to inform the setting of tariffs by the Payment by Results team of the Department of Health. It is envisaged that PICU will enter the PbR arena in April 2009 with tariffs from that point. There is considerable anxiety about the destabilising effect that full implementation could have, with the income of some units potentially cut drastically, so a phased implementation is likely to be considered, with part of a unit's income based on a block contract and part based on HRGs and PbR.

21.8 Will we be stuck with the current HRGs or can they be modified?

HRGs can be modified but there is a process that must be gone through. We are told that the HRG Expert Working Groups will continue to function with this remit. A further 'full' version of the HRGs is undertaken every five years or so.

21.9 How else can we use the PCCMDS data?

PICU Audit

For the first time, all units will be collecting data on patients that is focussed on a day rather than a complete PICU episode. This will provide additional information about the epidemiology of critical care, the frequency with which interventions are used, and allow a more meaningful comparison of workload across different units. This information can also be used to allow improved modelling of staff requirements within a unit.

Cost information

With the publication of a tariff for each HRG, units will be provided with much more detailed and transparent information about the income that the Trust is receiving for PICU activity and make it easier to breakdown income into certain categories eg private / NHS, in region / out of region. In addition, trials that are examining PICU costs or the cost effectiveness of a particular intervention will be able to use the HRG system and tariffs to calculate PICU costs, which will be a considerably easier methodology.

21.10 Further information

http://www.ic.nhs.uk/casemix
http://www.dh.gov.uk/en/Publicationsandstatistics/Publications/PublicationsPolicyAndGuidance/DH_062914

22 DEVELOPMENT OF THE RETRIEVALS DATASET
Dr Allan Wardhaugh

PICANet have now developed an extended dataset to allow collection of information regarding the retrieval and inter-hospital transfer of critically ill children. This will include collecting data regarding refused retrievals.

Each child who is referred to a PICU participating in PICANet in whom it is agreed PICU is appropriate will have data collected. In the event of a patient being refused admission because of lack of beds or lack of staff, a minimal amount of data will be collected, and the episode recorded as a refused admission. If the child is then referred to another unit, a new episode will be generated. The same child may generate several PICANet entries. The example below demonstrates an outline of the data collection.

The data required for a refused admission is minimal - simple patient identifiers, referring unit, outcome of referral. This should take a minimum of time to collect, as the referring unit are unlikely to be keen to spend too much time giving information over the phone about a patient they have not yet found a bed for.

Information will be collected relating to the transfer process, including the type of personnel involved, interventions undertaken during stabilisation and the transport itself, and critical incidents.

PICANet will be able to describe in detail where, when and by whom PICU retrieval is delivered in the United Kingdom. Individual units will receive information on how often, and for what reason, they have to turn away patients referred to them. This will be useful information for future commissioning of services. It may also provide insight into the ways different retrieval service configurations perform in terms of interventions and critical incidents.

The dataset has been developed by Allan Wardhaugh in consultation with members of the National Paediatric Retrieval Group over a period of several months, and was ratified by the group in the Paediatric Intensive Care Society meeting in November 2006. We are now working centrally to develop user-friendly forms and a database extension to allow this data to be collected, and over the course of the next year should be able to write to all PICUs asking them to participate in the collection of this data.

22.1 Example

Gareth Jones is a 2 month old with bronchiolitis admitted to Gloucester Royal Infirmary.

He deteriorates and requires intubation and ventilation. The consultant paediatrician refers him to Birmingham Children's Hospital - they are full. Bristol PICU are contacted - they are full, but are able to retrieve. Cardiff PICU is contacted, and accepts the patient, with the Bristol team retrieving.

On day 2 in Cardiff PICU he deteriorates further and is referred to Great Ormond Street for ECMO. They have a bed, and ask North Thames Children's Acute Transfer Service (CATS) to retrieve him.

After his ECMO run, he is referred back to Cardiff, still ventilated, on day 14. He is repatriated by a team from Great Ormond Street.

He is discharged from Cardiff on day 18 to his local hospital, spontaneously ventilating.

22.2 Forms completed

Birmingham Children's PICU1 formNot admitted - no staffed bed
Bristol Children's PICU1 formNot admitted - no staffed bed
1 formReferral and Retrieval
Cardiff PICU1 formReferral and admission
CATS1 formReferral and retrieval
GOSH1 formReferral and admission
1 formTransfer out*
Cardiff PICU**1 formReferral and admission

*This transfer generates a form because he is still requiring critical care.
**No transport form for final discharge, as he is not requiring critical care.

This patient with a complicated, but not unprecedented journey, has 8 forms completed and entered into the PICANet database. Under the current system, he would have had only 3 forms - one for the original Cardiff admission, one for his GOS admission, and one for his re-admission to Cardiff following ECMO.

The episodes will appear in the reports issued to each participating service: in Birmingham's report, he appears as a refused admission due to lack of a staffed bed; in Bristol's report, he appears as a refused admission due to a lack of a staffed bed, and as a retrieval; in Cardiff's report, he appears as a referral and admission retrieved by another service on the first occasion, and again as an admission retrieved by another service; at CATS, he appears as a referral and retrieval; at GOS, he appears as a referral and admission, then as a transfer out.

The full dataset has been designed to cover a variety of contingencies reflecting the various service configurations and journey types known to occur, and full discussion of these is beyond the scope of this introduction, but more detailed information will be provided to all PICANet centres closer to the time of introduction.

When the forms are completed and the database extended, the process will be piloted in a few regions, and hopefully then extended nationally after any glitches are sorted out.

The full dataset can be found at http://www.cardiffpicu.com/pages/nprg.html.

Figure RET1 Referral, retrieval and PICU admission, the data collection pathway

Figure RET1 Referral, retrieval and PICU admission, the data collection pathway

23 PICU HEALTH INFORMATICS
Dr Padmanabham Ramnarayan & Dr Krish Thiru

Medical Informatics is the name given to the application of information technology and data processing techniques in healthcare. Also referred to as health informatics and bioinformatics, the discipline deals with how we acquire, store, retrieve and use information, data and knowledge. Although primarily concerned with the flow of information, with the advancement of computer sciences, it has become synonymous with the use of computers in health care. Health informatics plays a particularly vital role in information-rich specialities such as PICU.

With the increase in demand for accurate and timely information by clinicians, managers, commissioners and the Department of Health, UK PICU has been extensively computerised over the last decade. Ever increasing pressure for standardised accurate data is being placed upon PICU (e.g. Payment by Results). In order to ensure that developments in PICU health informatics are systematic and coherent, efforts have been made to establish a National PICS Health Informatics Group under the auspices of the Paediatric Intensive Care Society Study Group.

23.1 National PICS Health Informatics Group: News update

The goal of shared and collaborative work in informatics across the UK PICU community moved closer to reality with the formation of a Health Informatics Group at the PICS meeting last November. Although the group is still in its infancy, membership is rapidly growing. The Group plans to have its first formal meeting at the 2007 PICS conference at Nottingham where current and planned work will be presented.

23.2 PICU Clinical Information Systems Survey

Coming soon to an Inbox near you!

This survey aims to find out how units are collecting, storing and using information. The results will help PICUs in developing a coherent strategy for the implementation of information systems, within and outside the National Program for IT. A web-based survey has been piloted in pan-Thames units. Now the survey is being rolled out to all UK PICUs. You can respond to the survey using one of three response methods. (1) a web-based questionnaire (2) electronic PDF response (3) Printed PDF response to be posted back. Please choose the easiest method for you.

23.3 SNOMED PICU Subset Development Project

The National Programme for Information Technology (NPfIT) has been hailed as the largest civilian IT project in the world, and it is going to change your life (or ruin it, depending on who you speak to)! The Informatics Group has been actively engaged in raising the profile of PICU with Connecting for Health. Our first piece of collaborative work is the SNOMED subset development project.

23.4 What is SNOMED-CT?

Clinical terminologies such as Read Codes (used in PICANet to collect data on diagnoses, procedures etc.) evolved out of the need to ensure standardised recording of electronic information and accurate retrieval of stored clinical data for audit and research.

The Systematised Nomenclature of Medicine- Clinical Terms (SNOMED-CT or SCT) is an evolving clinical terminology which aims to be the most comprehensive terminology for medicine. SCT which encompasses all current Read codes, is intended to be the de facto standard within NPfIT, which means that all coded information in the future will be need to be recorded as SCT terms rather than Read Codes or any other classification system. As a result, SCT will have significant implications on how we collect information for the Paediatric Critical Care Minimum Dataset (Payment by Results), PICANet, and any NHS IT systems developed for PICU.

23.5 What is a SNOMED subset?

SCT contains >400,000 unique concepts, and 1 million synonyms. It will be time-consuming, and almost impossible, for PICU staff to browse and find the right SCT term(s) to record clinical information. A SNOMED subset is a "cut-down" version of SCT relevant for each speciality. A PICU subset will only contain SCT terms relevant for PICU, making it easier to find the right term to represent each piece of clinical information. The informatics group has proposed a study methodology for subset development to ensure that the subset is comprehensive for PICU needs, and overlaps with related fields such as cardiology. For further details on the methodology (presented at the PICS SG recently) and how PICANet data is being used to drive subset development, please visit http://www.informatics.nhs.uk/.

23.6 Why develop a subset?

This is a significant project with far-reaching implications for our community. We need an effective and efficient subset of terms so that data collection in the future is easy and accurate. An incomplete or poorly conceived subset will make accurate data collection and retrieval in the future difficult and ineffective. We are actively looking for clinical input into this project and would welcome interested parties.

If you are interested in becoming involved in any of the activities detailed above or are thinking of developing a research and development study in this area of work, please contact Dr P Ramnarayan (RamnaP@cats.nhs.uk), Krish Thiru (ThiruK1@gosh.nhs.uk), or Stuart Rowe (stuart.rowe@nhs.net).

24 UK PICU STAFFING STUDY
Dr Dawn Coleby & Ms Namita Srivastava

24.1 Background

The UK PICU Staffing Study aims to look at the patterns of staff working in Paediatric Intensive Care Units, and how this impacts on staff wellbeing and patient outcomes. We are undertaking fieldwork in 12 PICUs across the UK, focusing on extended nursing roles and comparing PICUs where such roles are more or less well-developed.

Objectives include exploring extended nursing roles in PIC, and testing the impact of extended roles in relation to direct patient care time, quality of care and staff wellbeing. Additionally units with higher and lower extended roles are also being compared to explore workforce context, human resources management strategy (HRMS), and staffing costs.

24.2 The 3 different phases of the study

The UK PICU Staffing Study is organised into 3 phases.

Phase 1 was completed in December 2005, and involved a survey of all PICUs. This provided information on the skillmix and the types of tasks undertaken by nurses within each PICU. The information was then used to categorise all units and randomly select 12 to participate in the next phases of the study (6 with well-developed extended nursing roles and 6 with less well-developed extended nursing roles).

Phase 2 involves visiting the 12 participating units, collecting unit profile information and observing nursing staff. Information on direct patient care time is collected by observing nursing staff as they work and by asking the nursing staff to complete a summary shift diary. This requires the nurses to estimate and record the percent of time they spend on certain tasks, during the observed shift. This part of the study also involves staff interviews to explore their experiences and views on staffing issues. HRMS, staffing costs and unit context are determined by staff interviews, unit profile questionnaires, and further data from phase 3.

Phase 3 is the final year- long prospective stage of the project.

This phase requires the shift leaders in each unit to complete a regular workload log. The workload log records twice-daily the number and skill-mix of staff on duty to provide clinical care and the number and illness severity of patients in the unit. Patient outcomes will be ascertained from PICANet data and cases of probable ventilator associated pneumonia (VAP) will be collected from clinical notes by research link-nurses. VAP is an important new outcome of the study because it is an infection that is acquired inside the PICUs. This data is collected in each participating unit by a trained research link-nurse or clinical research fellow.

Finally, parent/child interviews and a staff well-being questionnaire are also undertaken during this phase. The user interviews are designed to capture the views of the parents and some children with regards to the care they received in the PICU. The staff well-being questionnaires are designed to capture staff views of their workplace unit and the processes in place for ensuring health and safety of staff and supporting their wellbeing.

24.3 Progress phase 2

Observation data

Observational data is collected on nurses shadowed by our researcher for 2 hours during a typical 12 hour shift. The researcher records the tasks that the nurse undertakes. 4 or 5 nurses are observed at each unit (mostly band 5 or 6). Any other staff that attend the patient's bedside and that are involved in providing care are also noted. The majority of the observations are undertaken during the day, but one observation per unit is undertaken during a night shift. Because casemix of the patients being cared for by the observed nurses clearly may influence nursing activity, diagnosis of those patients is also recorded. Additionally, all nurses on shift during observations are asked to self-report activity by completing a simple end-of-shift summary diary. Once completed the observational data and summary shift diaries are used to estimate the amount of direct patient care time the nurses undertake in each unit.

Interviews with staff

In order to explore a broad range of perspectives on staffing issues within PICUs, staff are selected in each unit representing different levels of seniority. The aim is to interview, within each unit, the Clinical Director, Nursing Manager, one staff nurse (band 5/6) and one junior doctor (SHO or Registrar). Some appointments with senior staff are made prior to the site visit, while most junior staff are approached during the site visit, and interviews are conducted as convenient to both parties. There have been relatively few problems so far in obtaining consent for and then conducting staff interviews within units. Staff are generally willing to discuss issues openly, and an interesting set of diverse opinions and experiences are being explored.

Unit profile

The unit profile questionnaire is given to the nursing managers before the site visit, so that the researcher can take the completed questionnaire back to the study centre after the visit. However, if the unit profile is not completed during the researchers site visit, units can fax the completed form to the study centre in Leicester at a later date.

24.4 Progress phase 3

Ventilator-associated-pneumonia (VAP) link-nurse recruitment and training

Nursing managers at each unit were asked to identify a suitable member of nursing staff to undertake the VAP data extraction. Once the research link-nurses were identified, two training sessions were arranged to show the link-nurses how to complete the data extraction form. Thus ensuring consistent completion between units. The research link-nurses were asked to start data collection on 1st March 2007. All link-nurses attended one of the training sessions and a training manual was supplied.

Interviews with parents and children

The study aims for a total of 36 interviews with parents with a subset of 10 parent-child paired interviews. We also aim to obtain a cross-section of children's reasons for admission and length of stay within the sample. Parents are approached during the researcher's site visit - in practice this is on an ad hoc basis depending on parent availability and also stability of the child which can determine suitability of parents for interview.

24.5 Outstanding work

Phase 2 is nearly complete. The staff at unit visits have been extremely welcoming and helpful, contributing to successful and informative site visits. Thank you to all the staff who have given us their time. All the information we are collecting in phase 2 provides important information on the nursing roles and how the team works together in each unit.

Phase 3 is now underway in all 12 PICUs and will be ongoing until March 2008. The research link-nurses have begun extracting information to detect probable Ventilator Associated Pneumonia from clinical notes on behalf of the project. Because VAP is an important outcome of the study it is important that this data is collected accurately and consistently between units. The study is therefore indebted to the research link-nurses for collecting these data on our behalf. Progress so far shows that the VAP data extraction is well underway in most units, with forms completed and returned regularly to the study centre.

In addition to the VAP data collection, participating units are also completing workload logs. The workload logs are being returned to the study centre regularly with the VAP forms by the link research nurses. We are pleased that most sites are returning logs that have been completed fully.

The staff wellbeing postal survey will be taking place in summer and autumn of 2007.

Acknowledgements
Investigators
Dr. Janet Tucker
Dr. Elizabeth Draper
Dr. Dawn Coleby
Ms. Namita Srivastava
Prof. Lorna McKee
Dr. Diane Skatun
Dr. Mark Darowski
Dr. Gareth Parry

The project is funded by NHS R&D Service Delivery and Organisation Programme (Grant number SDO-96-2005) and is being undertaken by researchers from the Universities of Aberdeen, Leicester, Harvard, USA and Leeds Teaching Hospital Trust in collaboration with PICANet.

Contacts: Dr Dawn Coleby (dc55@le.ac.uk) and Ms Namita Srivastava (ns161@le.ac.uk), University of Leicester, 22-28 Princess Road West, Leicester. LE1 6TP. Tel: 0116 2523200.

Project Website: http://www.abdn.ac.uk/dugaldbairdcentre/projects/PICUStaffing.shtml.

25 SPECIALISED COMMISSIONERS PERSPECTIVE ON PICANet
Dr Corinne Camilleri-Ferrante and Ms Roz Jones

25.1 Introduction

In order to gain an understanding of the use of PICANet to inform commissioning decisions, Specialised Commissioners from the North West and the East Midlands were asked to answer the following questions:

  1. Is PICANet useful to Commissioners?
  2. How could PICANet be developed in the future?

In capturing the Commissioners' perspective on PICANet, the two commissioners sought views from Paediatric Intensive Care Commissioners within England to collate this section of the annual report.

25.2 The Context of Commissioning Paediatric Intensive Care

Paediatric Intensive Care (PIC) services are commissioned under Specialised Commissioning arrangements and are defined under the Specialised Services Definition Set, second edition, number 23, 'Specialised Services for Children'. The definition covers all activity relating to:

In response to the recommendations of the 'Review of Commissioning Arrangements for Specialised Services' (the Carter Review) (May 2006)2 these definitions are currently under review.

Over the last decade, national policy initiatives have been introduced, such as payment by results (PbR), Choice, the establishment of new Foundation Trusts, Independent Sector Treatment Centres and Practice Based Commissioning. In addition access waiting times, particularly the introduction of the 18 week target, are now key performance measures. All of these influence commissioning decision-making processes and priorities.

These initiatives will have a limited effect on the commissioning of PIC services; the majority of paediatric intensive care is unplanned and not influenced greatly by patient or GP choice. The impact will be felt in those areas, such as elective surgery, requiring PIC support where limitations in the numbers of PIC beds available could have an adverse effect on waiting times and other parameters. This can be a particular problem at times of high activity where pressure on PIC beds can be acute.

The Carter Review made a number of recommendations on commissioning arrangements for specialised services. The main recommendations which may have an impact on Paediatric Critical Care services are as follows:

The Carter Review states that 'activity at undesignated providers should not be funded by commissioners.'2 In preparation for payment by results, standardised data have been identified by way of a minimum data set for paediatric critical care;3 these have been developed by Information for Health and Social Care to support the operation of new paediatric health resource groups (HRGs) which will form the component parts of the packages of care to payment by results (PbR). This data set must be collected in the Commissioning Data sets from 1st October 2007 (with optional collection from 1st April 2007) in preparation for paediatric critical care service inclusion within PbR from 2008/09. The rationale for this policy is that critical care is a high cost, low volume service whose case mix and activity levels are not necessarily related directly to normal commissioned activity. There are features of HRG version 4 that include the concept of 'unbundling' high cost elements; this should allow better representation of activity and cost of specialised services in order to ensure recognition of priority areas.

As well as developing HRGs, the data from the Paediatric Critical Care Minimum Data Set (PCCMDS) will inform service delivery. It is, therefore, fundamental that there should be no duplication with PICANet data collections. PICANet are developing the necessary software to collect PCCMDS for PICUs; it should then be possible to export a file for the Trust's use. This will reduce the burden of data collection, but will leave the responsibility for making the return to the individual Trust. Some additional funds may be requested to assist in this.

25.3 Commissioners views on the usefulness of PICANet

Overall Commissioners reported PICANet information to be extremely useful. Annually, the data collated provide Paediatric Intensive Care services with a clearer understanding of the size and nature of the service, by way of information based on professional standards and agreed definitions. They provide an insight into changes in case mix by geography and time, together with a national benchmark to compare local service provision. From a local and national perspective the data, when analysed in line with professional guidance and research, provide a strategic overview to inform service viability and future planning intentions.

As always, data are only as useful as they are timely and accessible. Using the data also helps to ensure that they are accurate. Commissioners would find direct access to PICANet data most useful, without the need for permissions from the clinical leads of the various units. Our need for comparative analysis means that we often wish to compare with units not directly within our sphere of influence, and this is aided by a more open attitude. We applaud the fact that all units were identified in the last report and would urge that we have open access to data with a minimum of red tape.

An example on the use of PICANet data to inform planning decisions is shown below in boxes 1 and 2:

Box 1 - Paediatric Intensive Care Review within the East Midlands and South Yorkshire Region
The review of PIC undertaken by East Midlands and South Yorkshire was completely underpinned by PICANet data. Commissioners requested, and received, timely and comprehensive data on the three units which were being reviewed. Commissioners were able to identify some errors in the data relating to the number bed days reported. PICANet reviewed this and it was found that two patients had no discharge dates recorded which skewed the data. Comparative analysis of the data from the three units showed interesting differences which have underpinned recommendations and planning assumptions.

Box 2 - Paediatric Intensive Care Review within the North West Region
Within the North West, a review of PIC within the Royal Manchester Children's Hospital and Royal Liverpool Children's Hospital was undertaken to assess the current utilisation of resources against best practice and to make recommendations regarding future capacity plans. The PICANet service provided a rapid response to requests and was easy to access. Information with regard to case mix and activity throughput was gained from annual report data, but the majority of information within the report was sought via specialist requests separately commissioned:
  • Bed census in comparison to bed day information to identify patient throughput/efficiency
  • Patient inflows (Non North West) and outflows (North West) by specialist centre (PICU) to assess specialist referrals, general emergencies or unmet need. In order to identify the names of the PICUs, authorisation was sought from the relevant lead clinicians. Of the PICUs accessed, all but one unit provided approval.
The review data collection would have been smoother if the PICANet service held data with regard to services that are interdependent on PIC such as high dependency, long-term ventilation and level 3 burn care. These data were collected manually from the two providers to ensure a better understanding of step up/down care and to identify any blockages within current patient pathways.

The main comment from Commissioners concerned the restriction in local access to PICANet clinical data. At the moment apart from the annual report, or local permission from lead clinician to see individual unit reports, the process to gain data is as follows:

  1. Commissioner completes data request form
  2. Lead Clinician from PICU is asked for authorisation to release data
  3. If authorised, PICANet prepare data
  4. If additional analysis is required funding is requested to undertake this work

This raises the question of who should own the data?

The service is funded by the Department of Health who established a legally binding agreement with PICANet. This outlined that the PICANet data and intellectual property rights are owned by PICANet. The forum for steering the scope of the work that PICANet undertakes is through the National PICANet Steering Group and Clinicial Advisory Group which is represented by Lead Clinicians and Commissioners.

From the perspective of Commissioners, there needs to be assurance that aggregate (not patient identifiable) data from all parts of the country will be available in a timely manner. This will ensure that Commissioners are able to make sensible and consistent planning decisions based on population need not on historical accident. PCT level analysis does not require patient identifiers to be included in reports, but those reports can only be produced by someone who has access to patient identifiers. The issue of future access and ownership of patient identifiers needs to be clarified and very clearly understood.

Consistency within data definitions across units is extremely important and is not being achieved at present. Examples of this include definitions of planned care (where National Confidential Enquiry into Patient Outcome and Death (NCEPOD) definitions could be used consistently across the country) and assessing complexity of care within units. This last point is very subjective, with a number of different methods in use (e.g. staff to patient ratios or organ dependency via the augmented care dataset). A decision needs to be taken as a matter of some urgency exactly which tools will be used and their exact definitions. Consistency of use across the country then needs to be audited.

As well as levels of care, Commissioners would like PICANet to include GP practice codes within the data set so that the commissioning PCT can be identified. This information is fundamental in grouping HRG activity in order to apply the payment by results charging mechanism.

Within the current data collection, it is very difficult to identify activity through the whole of the patient pathway. Current data collections do not account for the full journey of the patient. Data are shown for the PIC episode of care, with the result that delays in discharge, or bottlenecks within the step up/down pathway, are difficult to identify. This is important as current information systems do not link the interdependent services, such as High Dependency Care, Long Term Ventilation and level 3 burn care, to ensure appropriate access and maximum utilisation of PIC resources. Furthermore, data links with primary care and local authority child services would inform further intelligence on correlation between PIC admissions and deprivation.

Over the last decade, lead centres have experienced an increase in referrals from local District General Hospitals (DGHs). This is in part due to changes in the anaesthetic guidelines, which resulted in a change in surgical practice where all children under three years old are referred to the specialist centres for surgery. If the PIC is full, this will increase cancellation rates for surgery and thus have an impact on the Trusts' requirement to meet the 18 week target. An example of the impact of this surgical shift within the Pan Thames Consortium area is shown below in box 3:

Box 3 - Surgical Shifts within the Pan Thames area
As a consequence of the changes in national anaesthetic guidelines affecting the level of experience required to care for children under three years old on clinical governance grounds, a number of DGHs in North Thames decided it was unsafe to continue performing surgery on younger children. This required patients to be transferred to the tertiary centre for surgery. In addition, due to the local centre not being able to provide high dependency care a number of patients were transferred to the lead centre by a transport team for stabilisation overnight and then back to their local DGH. A subsequent analysis of activity flows identified that out of 22 overnight stays, 11 could have remained in their local DGH if sufficient HDU cover was available at this time.

Data on outcomes are extremely useful in assessing the service delivery and needs of a particular population. Inevitably, however, these data are limited. Mortality data are a blunt instrument, with limited usefulness. Morbidity outcome data would be extremely useful, but collecting them in a meaningful way presents many problems. Some thought to surrogate outcome measures which could be used and which would not be too onerous to collect would be most welcome.

Finally, current data with regard to PIC transport is very limited, only covering the fact that a retrieval/transfer has taken place, from what source and who undertook the transfer. It would be helpful to understand which geographical area and Trust patients have been transferred from, and whether it was a planned specialist referral, emergency referral or other. This would help to inform PICU capacity.

25.4 Commissioners thoughts on how PICANet could be utilised more efficiently and developed in the future

Commissioners provided the following suggestions on how PICANet could be developed to aid in decision-making and to work in line with National Commissioning guidance:

25.5 Access to data

Commissioners would like to be able to access data regionally in order to undertake local analysis. This could be achieved where access is approved for one person from each Specialised Commissioning Group; data would be password protected with read only access that could be downloaded. Access to detailed local data would be extremely useful.

25.6 Quarterly Reporting

Commissioners would find quarterly reports useful. Key performance indicators could be agreed and reported showing the local position in comparison to the national picture. In addition, exception reports showing national variances in trends would be useful.

25.7 Strategic planning from a National perspective

Commissioners need to make more use of PICANet data to inform national or sub-national level strategic planning. This could lead to the development of a national database of bed state which if supported could provide ongoing information on national bed capacity which would in turn inform local transport provision. Consideration should be given to a real time bed state system which links to the PICANet data set. This should be considered by the National Specialised Services Commissioning Groups.

25.8 Commissioning the patient pathway

In line with national guidance, Commissioners would like PICANet to incorporate paediatric high dependency care, long term ventilation and level 3 burns. Ideally, this would include data from outside tertiary centres, although the difficulties of data collection are recognised.

The planning process for PICANet to date has often been focussed on dealing with the complexities involved with the collection and storage of the audit data.The new PCCMDS and HRGs will allow the collection of a more complex range of information by specialty to be collected. Now that this stage of the development of the system is well established, this presents an opportunity to focus more on the analysis and future use of the information that is being collected.

25.9 Integration of PICANet information with Connecting for Health

At the moment PICU units provide PICANet information via a database. Information is then exported to a secure server behind the Leeds Teaching Hospitals Trust firewall and accessed via a secure connection by the PICANet team at Leeds University. If data were included within the Connecting for Health IT infrastructure wider information searches could be carried out and there would be less need for individual, different analyses.

25.10 Capturing data on PIC Transport

Commissioners recognise the difficulties associated with capturing data on PIC transport. However, information that allows a better national, or at least regional, picture would be most helpful. Currently, most PICs collect their own data but the system often does not have information on what happens to children who are refused. Collecting the data over wider areas would help to identify pockets of poor provision. This has been recognised within the PICANet work programme where a retrieval data set has been developed by Dr Allan Wardhaugh in consultation with members of the National Paediatric Retrieval Group. This was then ratified at the Paediatric Intensive Care Society in November 2006. This will be piloted in 2007 within a number of agreed sites. Further information can be found within Dr Wardhaugh's chapter in this report

25.11 Conclusions

25.12 Recommendations

25.13 References

  1. Department of Health, Health Services Directorate, July 1997, Paediatric Intensive Care "Framework for the Future"
  2. Review of Commissioning Arrangements for Specialised Services, May 2006. An independent review requested by the Department of Health
  3. DSCN Notice: 01/2007 Version 3.0 [Online] [Accessed 05/06/2007] Available from the World Wide Web at http://www.connectingforhealth.nhs.uk/dscn/dscn2007.

26 USES AND DISSEMINATION OF PICANet DATA

PICANet was established in collaboration with clinical colleagues from all participating NHS trusts, with a view to providing timely and accurate national and local information on PICU activity for those who deliver the service and those who plan the delivery of care. In common with all datasets the use of the data inevitably improves its quality. No data are ever provided or presented which allows an individual to be identified. In this, we act in accordance with the guidelines provided by ONS.

Information on PICANet is available to clinical care teams and parents through posters that are displayed in units and leaflets that are produced in 'parent packs'. The PICANet website address is given in this material and provides a further source of general information and copies of the national reports. Newsletters on progress are distributed regularly to lead nurses and consultants in each unit.

PICANet is pleased to report an increasing number of requests for data and information (Appendix D). Some requests have only requested aggregated, anonymised data from the entire dataset. For other requests, for example those that identify individual PICUs, PICANet always ensures that lead clinicians are informed and seeks permission for their data to be used.

Requests have been received from individual clinicians, groups of researchers and NHS commissioners. Some of the reports produced have required complex data processing and analyses and this has incurred additional costs which have been charged accordingly.

Dissemination of information from PICANet has been of prime importance to the team and Appendix K details specific talks given at various venues, a list of abstracts that have been presented at conferences and papers published by members of the PICANet team on PICANet and related topics. We welcome the opportunity to present data in these forums: please contact one of the team if you would like us to speak at local or national meetings.

27 TABLES AND FIGURES

Excel Tables and Figures

APPENDIX A PARTICIPATING NHS TRUSTS AND HOSPITAL CHARACTERISTICS

NHS TrustParticipating HospitalUnit / WardNumber of ITU bedsNumber of HDU bedsType of unit
Birmingham Children's Hospital NHS TrustBirmingham Children's HospitalPICU190General & Cardiac
Brighton & Sussex University Hospitals NHS TrustRoyal Alexandra Hospital for Sick ChildrenLydia Ward1a1General
Cambridge University Hospitals NHS Foundation TrustAddenbrooke's HospitalPICU62General
Cardiff & Vale NHS TrustUniversity Hospital of WalesPICU70General
Central Manchester & Manchester Children's University Hospitals NHS TrustRoyal Manchester Children's HospitalPICU150General
Great Ormond Street Hospital for Children NHS TrustGreat Ormond Street Hospital for ChildrenCCCU14-16b0Cardiac
Great Ormond Street Hospital for ChildrenPICU & NICU210General & Neonatal Unit
Guy's & St. Thomas' NHS Foundation TrustEvelina Children's HospitalPICU150General & Cardiac
Hull & East Yorkshire Hospitals NHS TrustHull Royal Infirmary PICU beds on AITU04Adult ICU providing General PICU
King's College Hospital NHS TrustKing's College Hospital PICU60General & Hepatic & Neurosurgical
Leeds Teaching Hospitals NHS TrustLeeds General InfirmaryWards 2 & 416c0General & Cardiac
St. James's University HospitalPICU16c0General
Newcastle Upon Tyne Hospitals NHS Foundation TrustNewcastle General HospitalPICU10d6dGeneral
Royal Victoria InfirmaryWard 310d6dSurgical ICU
Freeman HospitalPICU Freeman7e0Cardiothoracic surgery & ECMO
NHS Lothian - University Hospitals DivisionRoyal Hospital for Sick Children, EdinburghPICU6f6fGeneral
Oxford Radcliffe Hospitals NHS TrustThe John Radcliffe HospitalPICU72General & Cardiac
Nottingham University Hospitals NHS TrustQueen's Medical CentrePICU64General (plus regional neurosurgical, spinal and cleft lip & palate services)
Royal Brompton & Harefield NHS TrustRoyal Brompton HospitalPICU104Cardiac & Respiratory
Royal Liverpool Children's NHS TrustRoyal Liverpool Children's HospitalPICU210General & Cardiac
Sheffield Children's NHS Foundation Trust Sheffield Children's HospitalPICU92General
Sheffield Children's Hospital Neonatal Surgical Unit20Neonatal Surgical Unit
Southampton University Hospitals NHS TrustSouthampton General HospitalPICU9g0General & Cardiac
South Tees Hospitals NHS TrustJames Cook University HospitalPICU40General
St. George's Healthcare NHS TrustSt. George's HospitalPICU50General
St. Mary's NHS TrustSt. Mary's HospitalPICU82General
The Lewisham Hospital NHS TrustUniversity Hospital, LewishamPICU12hGeneral & Surgery
United Bristol Healthcare NHS TrustBristol Royal Hospital for ChildrenPICU130General & Cardiac
University Hospitals of Leicester NHS TrustLeicester Royal Infirmary CICU62General
Glenfield HospitalPICU50Cardiac
University Hospital of North Staffordshire NHS TrustUniversity Hospital of North StaffordshirePICU61General

Notes:aUpon moving to the new Children's hospital in June 2007, the unit will run at 1 ITU bed, 2 medical HDU beds and 2 surgical HDU beds initially
 bThe actual figure depends on the number of ECMO patients and HDU patients.
 cNurses / beds used flexibly across the sites
 dTotal bed numbers split between two hospital sites
 eMay become 8 beds, 2007
 fITU / HDU beds used flexibly (e.g. 6 ITU + 6 HDU, 9 ITU + 3 HDU, 11 ITU +1 HDU)
 g3 additional beds may be opening shortly
 hFlexed by a further 2 beds to support winter pressures

APPENDIX B CLINICAL ADVISORY GROUP MEMBERSHIP

NamePositionNHS Trust / HospitalPeriod served
Dr Paul BainesConsultant in Paediatric Intensive CareRoyal Liverpool Children's NHS Trust2002 - present
Alder Hey Hospital
Ms Corenna BowersSisterCardiff & Vale NHS Trust2002 - 2004
University Hospital of Wales
Dr Peter DavisConsultant in Paediatric Intensive CareUnited Bristol Healthcare NHS Trust2006 - present
Bristol Royal Hospital for Children
Dr Andrew DurwardConsultant in Paediatric Intensive CareGuy's & St Thomas' NHS Foundation Trust2002 - present
Evelina Children's Hospital
Ms Georgina GymerResearch NurseNottingham University Hospitals NHS Trust2005 - 2006
Queen's Medical Centre
Dr James FraserConsultant in Paediatric Intensive CareUnited Bristol Healthcare NHS Trust2002 - 2006
Bristol Royal Hospital for Children
Dr Hilary KloninConsultant in Paediatric Intensive CareHull & East Yorkshire Hospitals NHS Trust2002 - present
Hull Royal Infirmary
Ms Christine MackernessSisterNewcastle Upon Tyne Hospitals NHS Foundation Trust2002 - present
Newcastle General Hospital
Ms Tina McClellandAudit SisterRoyal Liverpool Children's NHS Trust2006 - present
Alder Hey Hospital
Dr Jillian McFadzeanConsultant in Paediatric Intensive CareNHS Lothian - University Hospitals Division2005 - present
Edinburgh Royal Hospital for Sick Children
Ms Victoria McLaughlinAudit NurseCentral Manchester & Manchester Children's University Hospitals NHS Trust2002 - present
Royal Manchester Children's Hospital
Dr Roddy O'DonnellConsultant in Paediatric Intensive CareCambridge University Hospitals NHS Foundation Trust2002 - present
Addenbrooke's Hospital
Ms Geralyn OldhamInformation Support ManagerGreat Ormond Street Hospital for Children NHS Trust2002 - present
Great Ormond Street Hospital for Sick Children
Dr Gale Pearson (Chair)Consultant in Paediatric Intensive CareBirmingham Children's Hospital NHS Trust2002 - present
Birmingham Children's Hospital
Dr Damian PryorConsultant in Paediatric Intensive CareCardiff & Vale NHS Trust2002 - 2004
University Hospital of Wales
Dr Allan WardhaughConsultant in Paediatric Intensive CareCardiff & Vale NHS Trust2004 - present
University Hospital of Wales
Ms Debbie WhiteSisterCambridge University Hospitals NHS Foundation Trust2002 - present
Addenbrooke's Hospital

APPENDIX C STEERING GROUP MEMBERSHIP

NamePositionOrganisationRepresentationPeriod Served
Mrs Pamela BarnesChair of Action for Sick ChildrenAction for Sick ChildrenLay Member2002 - present
Professor Nick Black (Chair)Head of Health Services Research UnitLondon School of Hygiene and Tropical MedicineHealth Services Research / Public Health2002 - present
Mr William Booth Clinical Nurse ManagerUnited Bristol Healthcare NHS TrustRoyal College of Nursing 2002 - present
Bristol Royal Hospital for Children PICU
Ms Bev BottingChild Health and Pregnancy StatisticsOffice for National StatisticsOffice for National Statistics (data protection)2002 - 2003
Dr Jean Chapple Consultant in Perinatal Epidemiology / Public Health Westminster Primary Care Trust PICNET founder 2002 - 2006
Dr Bill Chaudhry Consultant Paediatrician Newcastle Upon Tyne Hospitals NHS TrustClinical IT 2002 - 2003
Newcastle General Hospital PICU
Dr Mark DarowskiConsultant Paediatric AnaesthetistLeeds Teaching Hospitals NHS TrustRoyal College of Anaesthetists2002 - present
Leeds General Infirmary PICU
Mr Noel DurkinDepartment of HealthChild Health Services DirectorateDepartment of Health2002 - present
Dr Ian JenkinsConsultant in Paediatric Intensive CareUnited Bristol Healthcare NHS TrustPaediatric Intensive Care Society2006 - present
Bristol Royal Hospital for Children PICU
Dr Steve KerrConsultant in Paediatric Intensive CareRoyal Liverpool Children's NHS TrustChair of PICS2003 - present
Alder Hey Hospital PICU
Ms Helen Laing Clinical AuditHealthcare CommissionHealthcare Commission2004 - 2006
Mr Ian LangfieldAudit Co-ordinatorNational Assembly of WalesNational Assembly of Wales2002 - 2003
Dr Michael MarshConsultant in Paediatric Intensive CareSouthampton University Hospitals NHS TrustRoyal College of Paediatrics and Child Health2002 - present
Southampton General Hospital PICU
Dr Jillian McFadzean
Ms Laura Reekie		Consultant in Anaesthesia & Intensive Care
PA		NHS Lothian - University Hospitals DivisionEdinburgh Royal Hospital for Sick Children2005 - present
Edinburgh Royal Hospital for Sick Children
Dr Roddy McFaulMedical AdvisorChild Health Services DirectorateDepartment of Health2002 - 2003
Dr Kevin MorrisConsultant in Paediatric Intensive CareBirmingham Children's Hospital NHS TrustClinical Lead for the West Midlands Medicines for Children Local Research Network2006 - present
Birmingham Children's Hospital PICU
Professor Jon NichollDirector of Medical Care Research UnitSchool of Health and Related ResearchHealth Services Research / Statistics2002 - 2006
University of Sheffield
Dr Gale PearsonConsultant in Paediatric Intensive CareBirmingham Children's Hospital NHS TrustChair of PICANet CAG 2002 - present
Birmingham Children's Hospital PICU
Ms Tanya RalphNursing Research LeadSheffield Children's NHS Foundation TrustPICS2002 - 2006
Sheffield Children's Hospital PICU
Dr Kathy Rowan (on sabbatical 2004 - present, represented by Lucy Scott)Director ICNARCIntensive Care National Audit & Research Centre2002 - present
Mr Stuart Rowe PCT CommissionerCommissioning DepartmentPCT Commissioner (Pan-Thames)2003 - present
Hammersmith & Fulham PCT
Ms Dominique SammutAudit Co-ordinatorHealth Commission WalesHealth Commission Wales2003 - present
Dr Jennifer SmithMedical AdvisorOffice Project TeamCommission for Health Improvement2002 - 2004
Dr Charles Stack Consultant in Paediatric Intensive CareSheffield Children's NHS Foundation TrustPICS 2002 - 2006
Sheffield Children's Hospital PICU
Professor Stuart TannerMedical Advisor in Paediatrics and Child HealthChild Health Services DirectorateDepartment of Health2003 - 2006
Department of Health
Dr Robert TaskerLecturer in PaediatricsDepartment of PaediatricsPICS SG2004 - present
University of Cambridge Clinical School
Dr Edward WozniakMedical Advisor in Paediatrics and Child HealthChild Health Services DirectorateDepartment of Health2006 - present
Department of Health

APPENDIX D DATA/INFORMATION REQUESTS RECEIVED TO DATE

Excel Data and Information Requests

APPENDIX E DATA COLLECTION FORM

Acrobat PICANet Data Collection Form 2006

APPENDIX F INFORMATION LEAFLET

Acrobat PICANet Information Leaflet 2006

APPENDIX G DATA VALIDATION REPORT

Sample Data Validation Report

APPENDIX H MONTHLY ADMISSIONS REPORT

YearMonth12345689101112131415161718192021222324252627282931Total
200411092371333999563489133114204829421054192635183028344295451285
2923670352477563789143872250183945371933242433747208561210
38635504327684640104167106205328391258182325284331353222481278
487205137258755247814910223362727852113131232628748167381154
571125434157850317515110136444333445132837182528446232421143
67016543313776346841619231512923943142528143733654174391166
7721847392360513276160922653342954617183018262774113391099
87823452818665338741627522472823540182522214233853123281090
9822452441967411984158802841302794792232333716850213281111
107424504411723229701389725483134751182723182632974213431131
1190325744245752307914510527514043660152225213624460194391211
128530603530703936911501283731353534915312225282774421447231238
2004 Total996293661449268878594396993181711793175533723948259817429734326138034073614234454922314116
200517333553424793835911509522563336186419203120281765024543341233
273206439318135308798923143363554013172729362985924148371168
39213604522685845771331032739553496418243224262554624939421258
4742256312472433986132892946313455318242318261975816249331159
581236040206858301001297326372930134414232018262865724434291144
6781271342469363110112797385831279359313522363085521540351205
77516603925743230791531033665313011558262726291675322441281201
866959321654463275134882360352174412262722262476124636471119
98520593120664829781158527503430555203218283430107123240401185
106323603120763336911197520613439445112523163626116123333371135
117724583723763336961171133156345064819283024313196332461281275
128421533225884326731391193047364655024362133232255435551371263
2005 Total92123671542527487150339910341546113234061841941297597185312314280357297896882925051542714345
20061921566303777443410813710329543938568162727414229127028531321336
268295147308028351041131041845463565912223127332145919748351216
36823663530804232116152891747413974917274027402276726442481300
4881352271865493383134912550362774617323326412275131440391187
59019573925805129901388828643140749192522283617116419230381236
67917584020655231101142842855312353719154025252676227343331193
7991554372180422788155843252462425015202021371854621229291171
810623503522654822821407930723615342131932143423114925226391157
9822153362163462470143882652372354717163032312385330328251133
1092154548278861287812786266530375461423251936295593247391232
11101275332287842351011329032603535125114223026352765728241391271
12991754483010825359811711431394233104117282521321875532333351247
2006 Total106423465945430992953036511191630110032265545036974585190276355307422275906923183743843114679
Total2981763203513288512678162711603146499334119791826124111752531780549885101284811599122521994844132144588143140

APPENDIX I ERROR RATE REPORT

Error Rate Report

APPENDIX J POLICY FOR UNITS FALLING OUTSIDE THE CONTROL LIMITS
PICANet policy on PICUs lying outside the control limits of the mortality ratio funnel plots (PICANet November 2005)

Background - mortality ratios and funnel plots

PICANet is required by the Department of Health to report on the mortality outcomes of all children admitted for paediatric intensive care. The PICANet Clinical Advisory Group and Steering Group recommended that the mortality outcomes from each PICU be adjusted for the illness severity of the child at admission using the Paediatric Index of Mortality (PIM).1 PICANet reports the unadjusted mortality outcome from all PICUs and a mortality ratio based on the ratio of observed mortality in each PICU to the expected mortality calculated using PIM. From 2005, revised coefficients for PIM have been used derived from the recently completed United Kingdom Paediatric Intensive Care Outcome Study.2 PIM23 has been used for risk-adjustment in this report for 2006 only and will be used in future reports as the data become available.

Earlier work published by members of PICANet team4 has highlighted the problems of attempting to rank PICUs on their annual mortality, whether unadjusted or adjusted. PICANet, however, has also recognised the need to identify units which appear to have outcomes very different to other units. Consequently, PICANet has published a funnel plot of the observed to expected mortality ratio of individual PICUs. The funnel plots are constructed in such a way that there is an approximately 5% chance of a PICU falling outside the control limits, if the distribution of the mortality ratios is random.

The mortality ratio is calculated for each PICU by dividing the expected number of deaths calculated using the published PIM algorithm by the observed number of deaths for each PICU. The mortality ratio is then plotted on the y-axis against the number of admissions to the PICU on the x-axis. In order to satisfy the condition that if the overall distribution of the mortality ratios is random there exists an approximately 5% chance of a PICU falling outside the control limits, then the upper and lower control limits constructed at an individual PICU level must represent not 95% confidence intervals, but 99.9% confidence intervals around a mortality ratio of 1 by number of admissions.5 This is analogous to increasing the confidence interval (or significance level) when correcting for multiple comparisons in data containing numerous groups.

Data outliers

In these circumstances, PICANet will:

  1. Review the data to investigate whether there are data driven reasons for a PICU lying outside of the control limits (it is known that risk-adjustment tools can be unreliable when a PICU has a particularly high proportion of patients at either end of the bounds of the tool.)
  2. Review the data quality of the PICU. The quality of the data is the PICUs' responsibility. PICANet will provide feedback from PICU visits and central validation procedures. PICUs will be expected to check the quality of individual data items.
  3. Plot the data quality indicators over time to identify whether the anomaly can be traced to a certain data collection period.
  4. Plot the mortality ratio over time to identify whether the anomaly can be traced to a certain data collection period.
  5. Plot the observed mortality over time to identify whether the anomaly can be traced to a certain data collection period.
  6. Plot the expected mortality over time to identify whether the anomaly can be traced to a certain data collection period.
  7. Investigate the primary reason for admission to the PICU. If the PICU has a markedly high proportion of some primary reason of admission to the PICU compared with other PICUs this may suggest further refinements to the risk-adjustment method are required.
  8. Produce a brief summary report of the above to be forwarded to the lead clinician and Chief Executive at the PICU concerned, together with an invitation to meet in person to review the data with the PICANet team.

Where reference is made to the Chief Executive, it is accepted that they may be represented by their clinical governance lead.

NOTE: Excess mortality in particular sub-groups of patients or associated with other aspects of service provision may be identified using different statistical methods. The process outlined above will be implemented wherever anomalous results/outliers are identified.

References:

  1. Parry GJ, Gould CR, McCabe CJ, Tarnow-Mordi WO. Annual league tables of hospital mortality in neonatal intensive care: A longitudinal study. BMJ 1998; 316:1931-1935.
  2. Brady AR, Harrison D, Black S, Jones S, Rowan K, Pearson G, Ratcliffe J, Parry GJ, on behalf of the UK PICOS Study Group. Assessment and Optimization of Mortality Prediction Tools for Admissions to Pediatric Intensive Care in the United Kingdom. Pediatrics 2006; 117: 733-742.
  3. Shann F, Slater A, Pearson G. PIM 2: a revised version of the Paediatric Index of mortality. Intensive Care Med 2003; 29:278-285
  4. Shann F, Pearson G, Slater A, Wilkinson K, Paediatric index of mortality (PIM): a mortality prediction model for children in intensive care. Intensive Care Med 1997; 23:201-207
  5. Spiegelhalter D. Funnel plots for institutional comparison. Qual. Saf. Health Care, Dec 2002; 11: 390- 391.

APPENDIX K PUBLICATIONS/PRESENTATIONS

K.1 Presentations

Meeting/ConferenceVenueDatePresentation TitlePICANet Team Attendees
Presentation to Glasgow PICU teamGlasgow18/08/2003PICANetSam Jones & Tricia McKinney
NW Paediatric Intensive Care Seminar (North West Specialised Commissioning Group)Dunkenhalgh Hotel, Clayton-le-Moors, Lancashire23/06/2004PICANet: Results of national activitySam Jones & Roger Parslow
PICANet AGMLondon24/06/2004Presentation of National reportPICANet Team
Welsh National Commissioning Advisory Board MeetingRoyal Welsh Showground, Builth Wells28/07/2004PICANet: Presentation of National and Welsh reportLiz Draper & Nicky Davey
Strategic Issues in Health Care Management, Sixth International ConferenceUniversity of St Andrews02/09/2004Collection of personally identifiable information for a national clinical database: how feasible is it to obtain signed consent?Sam Jones
PICS SGCambridge University09/09/2004PICANet: How can it be used for research and audit?Nicky Davey, Sam Jones, Roger Parslow & Krish Thiru
Confidential Enquiry into Maternal and Child HealthLondon08/03/2005National Paediatric Intensive Care Database (PICANet)Liz Draper
Intensive Care National Audit & Research Centre (ICNARC): Eight Annual Meeting of the Case Mix ProgrammeSavoy Hotel, London13/04/2005Why is it important to include information on paediatric admissions in the new Case Mix Programme Dataset?Sam Jones
Pan Thames Report Update: Commissioning ConsortiumLondon06/05/2005PICANet: Update on Pan Thames data quality for commissioningKrish Thiru & Sam Jones
Paediatric Intensive Care Study DayRoyal Manchester Children's Hospital10/05/2005The epidemiology of critical illness in childrenRoger Parslow
Trent PIC commissionersQMC, Nottingham12/05/2005PICANet: Presentation of National report 2003-2004Liz Draper
Paediatric Intensive Care Trainee MeetingRoyal Liverpool Children's Hospital (Alder Hey)13/05/2005Role of PICANet and the relevance of the national audit to the clinical communityNicky Davey & Sam Jones
PICANet AGMLondon24/05/2005Presentation of National reportPICANet Team
NORCOM, TRENTCOM & LNR PIC commissionersLeicester13/06/2005PICANet in LNR, Trent & South Yorkshire PCTsLiz Draper
Health Protection Agency (HPA) annual conferenceWarwick12/09/2005Mortality, deprivation and ethnicity of critically ill children in England and Wales: preliminary findings from the Paediatric Intensive Care Audit Network (PICANet)Roger Parslow
Paediatric Critical Care Network Board (East leeds PCT)Leeds06/10/2005PICANet: Presentation of national data and relevance to commissioningTricia McKinney
Welsh National Commissioning Advisory Board MeetingLamb and Flag Hotel, Llanwenarth, Abergavenny11/10/2005PICANet: Presentation of National and Welsh ReportGareth Parry
PICANet AGMPerinatal Institute, Birmingham29/06/2006Presentation of the National ReportPICANet Team
Pan Thames Commissioners MeetingLondon28/07/2006Pan Thames PICANet Report 2004-2005Krish Thiru, Tricia McKinney
Paediatric Intensive Care Society Scientific MeetingGlasgow16 & 17/11/06PICU Health InformaticsK Thiru, P Ramnarayan, S Rowe on behalf of the pan Thames Health Informatics Group
University of Leicester, Department of Health Sciences. University of Leicester14/03/2007The UK Paediatric Traumatic Brain Injury StudyRoger Parslow
Paediatric Intensive Care Society Study GroupCambridge21 & 22/03/07PICU Health Informatics: Clinical Information SystemsK Thiru, P Ramnarayan, S Rowe on behalf of the pan Thames Health Informatics Group
Pan Thames Commissioners PbR RoadmapASIA House 14/06/2007PICANet and the PCCMDSRoger Parslow

K.2 Publications

JournalTitleAuthors
Pediatrics (2004) 113 1653-1657Trends in the incidence of severe retinopathy of prematurity in a geographically defined population over a 10-year periodHameed B, Shyamanur K, Kotecha S, Manktelow B, Woodruff G, Draper ES & Field D
Archives of Disease in Childhood (2005) 90 380-387Neuropsychological and educational problems at school age associated with neonatal encephalopathyMarlow N, Rose AS, Rands CE & Draper ES
Archives of Disease in Childhood (2005) 90 1182-1187Epidemiology of traumatic brain injury in children receiving intensive care in the UKParslow RC, Morris KP, Tasker RC, Forsyth RJ & Hawley C
British Medical Journal (2005) 330 43 (1 January)Paediatric cardiac surgical mortality after Bristol: details of risk adjustment tools were not given (letter)Parry GJ, Draper ES & McKinney P
British Medical Journal (2005) 330 877-879 (16 April)A feasibility study of signed consent for the collection of patient identifiable information for a national paediatric clinical audit databaseMcKinney PA, Jones S, Parslow R, Davey N, Darowski M, Chaudhry B, Stack C, Parry G, Draper ES for the PICANet Consent Study Group
European Journal of Obstetrics, Gynecology & Reproductive Biology (2005) 118 272-274Presentation of the European project models of organising access to intensive care for very preterm births in Europe (MOSAIC) using European diversity to explore models for the care of the very preterm babies.Zeitlin J, Papiernik E, Breart G, Draper E & Kollee L
Prenatal Diagnosis (2005) 25 286-291Population based study of the outcome following the antenatal diagnosis of cystic hygromaHowart ES, Draper ES, Budd JLS, Konje J, Kurinczuk JJ & Clarke M
Emergency Medical Journal (2006) 23 519-522Emergency access to neurosurgery in the United KingdomTasker RC, Morris KP, Forsyth RJ, Hawley CA, Parslow RC, on behalf of the UK Paediatric Brain Injury Study
Intensive Care Medicine (2006) 32(9) 1458Organ donation in paediatric traumatic brain injuryMorris KP, Tasker RC, Parslow RC, Forsyth RJ, Hawley CA
Intensive Care Medicine (2006) 32(10) 1606-1612Monitoring and management of intracranial pressure complicating severe traumatic brain injury in childrenMorris KP, Forsyth RJ, Parslow RC, Tasker RC, Hawley CA on behalf of the UK Paediatric Traumatic Brain Injury Study Group and the Paediatric Intensive Care Society Study Group
Lancet (2006) 367 1080-85Outcome after neonatal continuous negative-pressure ventilation: follow-up assessmentTelford K, Waters L, Vyas H, Manktelow BN, Draper ES, Marlow N
Pediatrics (2006) 117 733-742Assessment and optimisation of mortality prediction tools for admissions to paediatric intensive care in the United KingdomBrady AR, Harrison D, Black S, Jones S, Rowan K, Pearson G, Ratcliffe J, Parry GJ; UK PICOS Study Group
Archives of Disease in Childhood. Fetal and Neonatal Edition (2007) 92 19-24Outcome following neonatal continuous negative pressure ventilationTelford K, Waters L, Vyas H, Manktelow BN, Draper ES, Marlow N
Paediatric Intensive Care Medicine (2007) (In Press)Prediction of raised intracranial pressure complicating severe traumatic brain injury in children: implications for trial designForsyth RJ, Parslow RC, Tasker RC, Hawley CA, Morris KP. On behalf of the UK Paediatric Traumatic Brain Injury Study Group and the Paediatric Intensive Care Society Study Group (PICS SG)

K.3 Abstracts

AbstractTitleAuthors
Health Protection Agency (HPA) Annual Conference, 12-15 September 2005, Warwick (oral presentation)Mortality, deprivation and ethnicity of critically ill children in England and Wales: preliminary findings from the Paediatric Intensive Care Audit Network (PICANet)Parslow RC, Tasker RC, Chater T, Davey N, Draper ES, Jones S, Parry GJ & McKinney PA.
European Society for Paediatric and Neonatal Intensive Care (ESPNIC) annual conference, 15-17 September 2005, Antwerp (oral presentation)Mortality, deprivation and ethnicity of critically ill children in England and Wales: preliminary findings from the Paediatric Intensive Care Audit Network (PICANet)Parslow RC, Tasker RC, Chater T, Davey N, Draper ES, Jones S, Parry GJ, Thiru K & McKinney PA.
Developmental Medicine and Child Neurology (2005) 47 (Suppl 101) 4Design of randomized controlled trials of the management of raised intracranial pressure in paediatric traumatic brain injuryForsyth RJ, Morris K, Parslow RC, Hawley C & Tasker RC
5th World Congress on Pediatric Critical Care, 24-28 June 2007, Geneva, Switzerland (oral presentation)Infants admitted to paediatric intensive care with acute respiratory failure in England and WalesParslow RC, McKinney PA, Draper ES, O'Donnell R
5th World Congress on Pediatric Critical Care, 24-28 June 2007, Geneva, Switzerland (poster presentation)Collecting national data for clinical audit: The Paediatric Intensive Care Audit Network in Great BritainParslow RC, McKinney PA, Draper ES, Thiru K
5th World Congress on Pediatric Critical Care, 24-28 June 2007, Geneva, Switzerland (poster presentation)Admission to PICU with severe bronchiolitis and acute respiratory failure after preterm birth is associated with a longer duration of stay and a higher incidence of apnoeas but not mortalityO'Donnell DR, Parslow RC, McKinney PA, Draper ES
5th World Congress on Pediatric Critical Care, 24-28 June 2007, Geneva, Switzerland (poster presentation)Severe bronchiolitis is associated with the annual UK winter increase in PICU admissions and prolonged stay compared with other diagnosesO'Donnell DR, Parslow RC, McKinney PA, Draper ES
5th World Congress on Pediatric Critical Care, 24-28 June 2007, Geneva, Switzerland (poster presentation)Hyperglycaemia and insulin therapy in UK paediatric intensive care unitsNayak P, Morris KP, Parslow RC
5th World Congress on Pediatric Critical Care, 24-28 June 2007, Geneva, Switzerland (oral presentation)The effect of missing data on PIM-predicted SMREmsden S, Baines P, McClelland T, Parslow RC
5th World Congress on Pediatric Critical Care, 24-28 June 2007, Geneva, Switzerland (poster presentation)Clinical information system utilisation in paediatric intensive care: A UK perspectiveRamnarayan P, Thiru K, Rowe S on behalf of pan Thames Health Informatics Group
The 15th Annual Public Health Forum, Edinburgh International Conference Centre, 28-29 March 2007, Edinburgh, UK (poster presentation)Using Data to Inform Commissioning of Paediatric Intensive CareSidhu S, Rowe S & Thiru K

APPENDIX L MEMBERSHIP OF THE PAEDIATRIC CRITICAL CARE EXPERT WORKING GROUP

ChairNick Griffin (Chair)Consultant Paediatrician, Northampton General Hospital
Project ManagerIan HughesIC
Clinical RepresentativesKevin MorrisConsultant Paediatric Intensivist, Birmingham Children's Hospital
Pete BarryConsultant Paediatric Intensivist, University Hospitals of Leicester
Charles StackConsultant Paediatric Intensivist, Sheffield Children's Hospital
Andy DarbyshireNurse Consultant, Paediatric HDU, Alder Hey Hospital, Liverpool
William Booth Senior Nurse, Paediatric Intensive Care Unit, Bristol Royal Hospital for Children And Chair of the Royal College of Nursing Paediatric Intensive Care Nurses Forum
Ian MurdochClinical lead, Guys Hospital, London
Robert YatesConsultant Paediatric Intensivist, Manchester Children's Hospital
PICANetRoger ParslowSenior Research Fellow, PICANet
Department of HealthProfessor Stuart TannerDepartment of Health, Medical Adviser, Paediatrics & Child Health,
CommissioningStuart RowePan Thames PICU Commissioning Consortium
CasemixPaul SmithSenior Casemix Consultant, HSCIC
CostingSujit KoonerCosting Consultant, HSCIC
FinanceLee BondDirector of Finance, Sheffield Children's Hospital
Previous membersAndy GillSenior Casemix Consultant, IC
Lyvonne TumeLecturer Practitioner, Alder Hey Hospital, Liverpool

APPENDIX M MAPPING OF INTERVENTIONS TO DIFFERENT HRG LEVELS

HRGLabelCriteria
7Intensive Care - ECMO/ECLSExtracorporeal membrane oxygenation (ECMO) / Extracorporeal Life Support (ECLS) including VAD, or aortic balloon pump
6Intensive Care Advanced EnhancedInvasive Mechanical Ventilation (IMV) or Advanced Respiratory Support (ARS)
Plus one or more of:
  • Burns >79% BSA
  • >80 mls/kg volume boluses
OR
HRG 5 + Isolation
5Intensive Care AdvancedInvasive Mechanical Ventilation (IMV) or Advanced Respiratory Support (ARS)
Plus one or more of:
  • Haemofiltration
  • Haemodialysis
  • Peritoneal dialysis
  • Burns 50-79% BSA
  • Extracorporeal Liver Support (MARS)
  • Exchange transfusion
  • iNO
  • Surfactant
  • Plasmafiltration

OR
HRG 4 + Isolation
4Intensive Care Basic EnhancedInvasive Mechanical Ventilation (IMV)
Plus one or more of:
  • Vasoactive infusion
  • ICP monitoring
  • Burns 20-49% BSA
  • Intravenous thrombosis
  • CPR in last 24 hrs
OR
Advanced Respiratory Support (ARS) (Jet ventilation or High Frequency Oscillatory Ventilation (HFOV))
OR
HRG 3 + Isolation
3Intensive Care BasicInvasive Mechanical Ventilation (IMV)
OR
Non invasive ventilation / CPAP
Plus one or more of:
  • Burns >79% BSA
  • >80 mls/kg volume boluses
  • Haemofiltration
  • Haemodialysis
  • Peritoneal dialysis
  • Burns 50-79% BSA
  • Extracorporeal Liver Support (MARS)
  • Exchange transfusion
  • iNO
  • Surfactant
  • Plasmafiltration
  • Vasoactive infusion
  • ICP monitoring
  • Burns 20-49% BSA
  • Intravenous thrombolysis
  • CPR in last 24 hrs
OR
HRG 2 + Isolation
2High Dependency AdvancedNon invasive ventilation / CPAP
  • Arterial monitoring
  • Haemofiltration
  • "Acute" haemodialysis
  • "Acute" Peritoneal dialysis
  • Plasmafiltration
  • Exchange transfusion
  • Temporary pacing
  • Vasoactive infusion
  • Intravenous thrombolysis (tPA, streptokinase)
  • ICP monitoring
  • Intraventricular catheter / external ventricular drain
  • CPR in last 24 hrs
  • iNO
  • Surfactant
  • Extracorporeal Liver Support (MARS)
  • >80 mls/kg volume boluses
  • Apnoea Requiring Intervention in past 24 hrs (>3 stimulation or bag-mask)
OR
HRG 1 + Isolation
1High DependencyCVP monitoring
  • Continuous ECG monitoring
  • Oxygen Therapy plus Continuous Pulse Oximetry
  • Nasopharyngeal airway
  • Care of tracheostomy
  • Upper airway obstruction requiring nebulised adrenaline
  • Severe Asthma requiring intravenous bronchodilator, or continuous nebulisers
  • DKA requiring continuous insulin infusion

APPENDIX N PCCMDS: HIGH COST DRUGS WHICH ARE UNBUNDLED

High Cost DrugOPCS 4.3 CodeOPCS 4.3 Code LabelHRGHRG Label
SildenafilX821Pulmonary hypertension drugs Band 1XD01ZPrimary Pulmonary Hypertension drugs Band 1
BosentanX822Pulmonary hypertension drugs Band 2XD02ZPrimary Pulmonary Hypertension drugs Band 2
IloprostX823Pulmonary hypertension drugs Band 3XD03ZPrimary Pulmonary Hypertension drugs Band 3
EpoprostenolX824Pulmonary hypertension drugs Band 4XD04ZPrimary Pulmonary Hypertension drugs Band 4
Factor VIIa (recombinant)X831Blood products Band 1XD05ZBlood products Band 1
Recombinant activated protein CX832Blood products Band 2XD06ZBlood products Band 2
AlteplaseX833Fibrinolytic drugs Band 1XD07ZFibrinolytic drugs Band 1
ReteplaseX833Fibrinolytic drugs Band 1XD07ZFibrinolytic drugs Band 1
TenecteplaseX833Fibrinolytic drugs Band 1XD07ZFibrinolytic drugs Band 1
Nitric oxideX841Medical gases Band 1XD08ZMedical gases Band 1
Botulinum toxinX851Torsion dystonias and other involuntary Band 1XD09ZTorsion dystonias and other involuntary movements drugs Band 1
RiluzoleX852Amyotrophic lateral sclerosis drugs Band 1XD10ZAmyotrophic lateral sclerosis drugs Band 1
Amphotericin liposomalX861Anti-fungal drugs Band 1XD11ZAnti fungal drugs Band 1
CaspofunginX861Anti-fungal drugs Band 1XD11ZAnti-fungal drugs Band 1
FlucytosineX861Anti-fungal drugs Band 1XD11ZAnti-fungal drugs Band 1
VoriconazoleX862Anti-fungal drugs Band 2XD12ZAnti-fungal drugs Band 2
AdefovirX863Hepatitis B treatment drugs Band 1XD13ZHepatitis B treatment drugs Band 1
Interferon alfaX863Hepatitis B treatment drugs Band 1XD13ZHepatitis B treatment drugs Band 1
Peginterferon alphaX864Respiratory syncytial virus treatment and Hepatitis C treatment drugs Band 1XD14ZRespiratory syncytial virus treatment and Hepatitis C treatment drugs Band 1
RibavirinX864Respiratory syncytial virus treatment and Hepatitis C treatment drugs Band 1XD14ZRespiratory syncytial virus treatment and Hepatitis C treatment drugs Band 1
PalivizumabX865Respiratory syncytial virus prevention drugs Band 1XD15ZRespiratory syncytial virus virus prevention drugs Band 1
PegvisomantX871Growth hormone receptor antagonist drugs Band 1XD16ZGrowth hormone receptor antagonist drugs Band 1
SomatropinX872Growth hormone analogue drugs Band 1XD17ZGrowth hormone analogue drugs Band 1
TeriparatideX873Bone metabolism drugs Band 1XD18ZBone metabolism drugs Band 1
AlemtuzumabX891Monoclonal antibodies Band 1XD19ZMonoclonal antibodies Band 1
RituximabX892Monoclonal antibodies Band 2XD20ZMonoclonal antibodies Band 2
Beta interferonX893Immunomodulating drugs Band 1XD21ZImmunomodulating drugs Band 1
GlatiramerX893Immunomodulating drugs Band 1XD21ZImmunomodulating drugs Band 1
LanreotideX894Somatostatin analogues Band 1XD22ZSomatostatin analogues Band 1
OctreotideX894Somatostatin analogues Band 1XD22ZSomatostatin analogues Band 1
Darbopoetin alfaX901Hypoplastic haemolytic and renal anaemia drugs Band 1XD23ZHypoplastic haemolytic and renal anaemia drugs Band 1
Epoetin alfa and betaX901Hypoplastic haemolytic and renal anaemia drugs Band 1XD23ZHypoplastic haemolytic and renal anaemia drugs Band 1
Antilymphocyte globulinX902Hypoplastic haemolytic and renal anaemia drugs Band 2XD24ZHypoplastic haemolytic and renal anaemia drugs Band 2
FilgrastimX903Neutropenia drugs Band 1XD25ZNeutropenia drugs Band 1
LenograstimX903Neutropenia drugs Band 1XD25ZNeutropenia drugs Band 1
PegfilgrastimX903Neutropenia drugs Band 1XD25ZNeutropenia drugs Band 1
Total parenteral nutritionX904Intravenous nutrition Band 1XD26ZIntravenous nutrition Band 1
Cysteamine (mercaptamine)X911Metabolic disorder drugs Band 1XD27ZMetabolic disorder drugs Band 1
Sodium phenylbutyrateX912Metabolic disorder drugs Band 2XD28ZMetabolic disorder drugs Band 2
MiglustatX913Metabolic disorder drugs Band 3XD29ZMetabolic disorder drugs Band 3
Agalsidase beta (galactosidase)X914Metabolic disorder drugs Band 4XD30ZMetabolic disorder drugs Band 4
ImigluceraseX914Metabolic disorder drugs Band 4XD30ZMetabolic disorder drugs Band 4
LaronidaseX914Metabolic disorder drugs Band 4XD30ZMetabolic disorder drugs Band 4
AdalimumabX921Cytokine inhibitor drugs Band 1XD31ZCytokine inhibitor drugs Band 1
AnakinraX921Cytokine inhibitor drugs Band 1XD31ZCytokine inhibitor drugs Band 1
EtanerceptX921Cytokine inhibitor drugs Band 1XD31ZCytokine inhibitor drugs Band 1
InfliximabX921Cytokine inhibitor drugs Band 1XD31ZCytokine inhibitor drugs Band 1
RasburicaseX922Hyperuricaemia drugs Band 1XD32ZHyperuricaemia drugs Band 1
EfalizumabX951Immune response drugs Band 1XD33ZImmune response drugs Band 1
FlebogammaX961Immunoglobulins Band 1XD34ZImmunoglobulins Band 1
GammagardX961Immunoglobulins Band 1XD34ZImmunoglobulins Band 1
OctagamX961Immunoglobulins Band 1XD34ZImmunoglobulins Band 1
SandoglobulinX961Immunoglobulins Band 1XD34ZImmunoglobulins Band 1
SubcuviaX961Immunoglobulins Band 1XD34ZImmunoglobulins Band 1
SubgamX961Immunoglobulins Band 1XD34ZImmunoglobulins Band 1
VigamX961Immunoglobulins Band 1XD34ZImmunoglobulins Band 1
 X818Other specified high cost gastrointestinal drugsXD35ZOther specified high cost drugs
 X828Other specified high cost hypertension drugsXD35ZOther specified high cost drugs
 X838Other specified high cost other cardiovascular drugsXD35ZOther specified high cost drugs
 X848Other specified high cost respiratory drugsXD35ZOther specified high cost drugs
 X858Other specified high cost neurology drugsXD35ZOther specified high cost drugs
 X868Other specified high cost anti-infective drugsXD35ZOther specified high cost drugs
 X878Other specified high cost endocrinology drugsXD35ZOther specified high cost drugs
 X888Other specified high cost reproductive and urinary tract drugsXD35ZOther specified high cost drugs
 X898Other specified high cost immunosuppressant drugsXD35ZOther specified high cost drugs
 X908Other specified high cost haematology and nutrition drugsXD35ZOther specified high cost drugs
 X918Other specified high cost metabolic drugsXD35ZOther specified high cost drugs
 X928Other specified high cost musculoskeletal drugsXD35ZOther specified high cost drugs
 X938Other specified high cost ophthalmology drugsXD35ZOther specified high cost drugs
 X948Other specified high cost ear, nose and throat drugsXD35ZOther specified high cost drugs
 X958Other specified high cost dermatology drugsXD35ZOther specified high cost drugs
 X968Other specified high cost immunology drugsXD35ZOther specified high cost drugs
 X978Other specified high cost anaesthesia drugsXD35ZOther specified high cost drugs
 X819Unspecified high cost gastrointestinal drugsXD36ZUnspecified high cost drugs
 X829Unspecified high cost hypertension drugsXD36ZUnspecified high cost drugs
 X839Unspecified high cost other cardiovascular drugsXD36ZUnspecified high cost drugs
 X849Unspecified high cost respiratory drugsXD36ZUnspecified high cost drugs
 X859Unspecified high cost neurology drugsXD36ZUnspecified high cost drugs
 X869Unspecified high cost anti-infective drugsXD36ZUnspecified high cost drugs
 X879Unspecified high cost endocrinology drugsXD36ZUnspecified high cost drugs
 X889Unspecified high cost reproductive and urinary tract drugsXD36ZUnspecified high cost drugs
 X899Unspecified high cost immunosuppressant drugsXD36ZUnspecified high cost drugs
 X909Unspecified high cost haematology and nutrition drugsXD36ZUnspecified high cost drugs
 X919Unspecified high cost metabolic drugsXD36ZUnspecified high cost drugs
 X929Unspecified high cost musculoskeletal drugsXD36ZUnspecified high cost drugs
 X939Unspecified high cost ophthalmology drugsXD36ZUnspecified high cost drugs
 X949Unspecified high cost ear, nose and throat drugsXD36ZUnspecified high cost drugs
 X959Unspecified high cost dermatology drugsXD36ZUnspecified high cost drugs
 X969Unspecified high cost immunology drugsXD36ZUnspecified high cost drugs
 X979Unspecified high cost anaesthesia drugsXD36ZUnspecified high cost drugs

APPENDIX O CHANGES TO THE STRUCTURE OF NHS PRIMARY CARE IN ENGLAND ON 1ST OCTOBER 2006

ExcelChanges to the Structure of NHS Primary Care in England on 1st October 2006

APPENDIX P GLOSSARY

The following abbreviations / terms are used within the text of this report:

A&EAccident and Emergency Department
AICAdult Intensive Care
AICUAdult Intensive Care Unit
ANZPICSAustralian and New Zealand Paediatric Intensive Care Registry
CAGClinical Advisory Group
CATSChildren's Acute Transfer Service
CT3Clinical Terms 3
ECMOExtra corporeal membrane oxygenation
ENBEnglish National Board
GBGreat Britain
GOSHGreat Ormond Street Hospital
HBHealth Board
ICInformation Centre for health and social care
ICNARCIntensive Care National Audit & Research Centre
ICP deviceIntracranial pressure device
Invasive ventilationAny method of ventilation delivered via an endotracheal tube, laryngeal mask or tracheotomy tube
IQRInterquartile Range
IVvasoactive therapy Intravenous drug therapy to support blood pressure and heart rate
LVADLeft ventricular assist device to support cardiac function
NPfITNational Programme for Information Technology
NSPDNational Statistics Postcode Directory
NHSNational Health Service
NHSIANational Health Service Information Authority
NHSnetA secure wide area network connecting NHS organisations which enables units to transfer data electronically to PICANet
Non-invasive ventilationAny method of ventilation NOT given via an endotracheal tube, laryngeal mask or tracheostomy tube
PbRPayment by Results
PCCEWGPaediatric Critical Care Expert Working Group
PCCMDSPaediatric Critical Care Minimum Dataset
PCOPrimary Care Organisations
PIAGPatient Information Advisory Group
PICPaediatric Intensive Care
PICANetPaediatric Intensive Care Audit Network
PICNETPaediatric Intensive Care Network
PICSPaediatric Intensive Care Society
PICS SGPaediatric Intensive Care Society Study Group
PICUPaediatric Intensive Care Unit
PIMPaediatric Index of Mortality
PIM 2Paediatric Index of Mortality version 2
READ CodesClinical terminology used to describe clinical conditions, symptoms and observations
RSVRespiratory syncytial virus
SCTSee SNOMED CT®
SHOSenior House Officer
SGSteering Group
SNOMED CT®SNOMED CT® is a clinical terminology - the Systematised Nomenclature of Medicine. It is a common computerised language that will be used by all computers in the NHS to facilitate communications between healthcare professionals in clear and unambiguous terms
SMRStandardised mortality ratio
SHAStrategic Health Authority
SWACICSouth West Audit of Critically Ill Children
UK PICOSUnited Kingdom Paediatric Intensive Care Outcome Study

PICANet logo

http://www.picanet.org.uk/
picanet@leeds.ac.uk

University of LeedsUniversity of LeicesterPan-Thames Co-Ordinator
Patricia McKinney
Roger Parslow
Thomas Fleming
Angie Willshaw		Elizabeth Draper
Caroline Lamming		Krish Thiru
PICANet
Paediatric Epidemiology Group
Centre for Epidemiology & Biostatistics
The Leeds Institute of Genetics, Health and
Therapeutics
University of Leeds
30 Hyde Terrace
Leeds LS2 9LN		PICANet
Department of Health Sciences
University of Leicester
22-28 Princess Road West
Leicester LE1 6TP		PICANet
Cardiorespiratory & Critical Care Division
Room 8086, Level 8 - Nurses Home
Great Ormond Street Hospital for Children
Great Ormond Street
London WC1N 3JH
r.c.parslow@leeds.ac.uk
0113 343 4856		crl4@le.ac.uk
0116 252 5414		thiruk1@gosh.nhs.uk
020 7762 6713
PICS logo NHS logo