Q J Med 2000; 93: 67-73
© 2000 Association of Physicians
Review |
A practical guide to continuous population-based data collection (PACE): a process facilitating uniformity of care and research into practice
From the Department of Haematology, Royal Victoria Infirmary, Newcastle upon Tyne, UK
Address correspondence to Professor S.J. Proctor, Department of Haematology, Royal Victoria Infirmary, Newcastle upon Tyne NE1 4LP. e-mail:s.j.proctor{at}ncl.ac.uk
 |
Introduction |
|---|
 Top Introduction PACE: the process and...Communication: the key elementCan it work for...Developing the systemAmalgamations and collaborationsPracticalities of financing and...Building the process into...PACE in actionIndividual advice on patient...ConclusionReferences |
|---|
In an editorial on clinical databases in the British Medical Journal, Professor Black indicated that in this era of assessment and accountability in medicine, there is a tendency to cling to traditional research methodology, in spite of its well-recognized limitations.1 He points out that classical studies are expensive, of limited duration and in limited clinical arenas. A major problem in trials is that study populations are unlike the average population seen by doctors in their practice, and the lack of introduction of positive research findings into practice continues to be of major concern. In his article, Black discusses the potential value of new processes of intervention in health care, using as the starting point high-quality clinical databases which potentially allow everyday practise to come closer to research.
Over the past 20 years progress has been made in this area in the UK, following the pioneering work of the surgeons throughout Lothian in Scotland during the 1970s2 and obstetricians in the Thames region in the 1980s.3 More emphasis needs to be placed on finding alternative methodologies of investigation, particularly where a randomized controlled trial is impractical. In the area of clinical pharmacology, in a review of the studies related to self-poisoning, Buckley introduced the concept of ‘toxico-epidemiology’ (a population-based assessment) as the only way he perceived progress could occur in this difficult area, not amenable to the RCT.4
The present paper is a preliminary practical guide to a process evolved by the Northern Regional Haematology Group (NRHG) in the UK, entitled PACE (population-adjusted clinical epidemiology). In essence, PACE has several components, the most important of which relate to communication with all the individual doctors interested in a given disease area over a defined geographical area. In its original form, the core process is the maintenance of a comprehensive regional register for all the patients with malignant haematological problems, including detailed demographic treatment and outcome data, on a geographically-based census population of incidence cases. Such an approach provides ‘reality-based’ information for which co-operating physicians can then rationally adjust therapeutic options. The register is thus at the heart of the overall research and audit strategy for the Region, and provides a broad population-based context for new questions and investigations which present themselves once such a database is established.
|
PACE: the process and how to do it |
|---|
|
TopIntroductionPACE: the process and...Communication: the key elementCan it work for...Developing the systemAmalgamations and collaborationsPracticalities of financing and...Building the process into...PACE in actionIndividual advice on patient...ConclusionReferences |
|---|
In 1982, a change in personnel marked the beginning of a new era for management of adult haematological malignancies in the former Northern Health Region of England (3 million population) (Figure 1
). A long-term regional plan commenced which contained three elements: (i) the establishment of a regular meeting of all doctors diagnosing and treating patients with haematological malignancy in the Northern Region, to facilitate co-operation through communication in a non-stressful environment (haematologists, medical oncologists and clinical oncologists); (ii) the creation of a robust registry system where detailed information of presentation data, treatment and outcome could be recorded on all cases of haematological malignancy, as part of a long-term programme to improve patient care; and (iii) the development of a ‘menu’ of treatment packages based on mutually-agreed best practice (from regional research outcome and the published literature) which would be used by all the physicians to provide uniformity of care across the region.
|
Subsequently, audited clinical research was encouraged in order to advance further progress in treatment. The emphasis was on optimizing treatment outcome across the Region without the need for centralization of patient care.
|
Communication: the key element |
|---|
|
TopIntroductionPACE: the process and...Communication: the key elementCan it work for...Developing the systemAmalgamations and collaborationsPracticalities of financing and...Building the process into...PACE in actionIndividual advice on patient...ConclusionReferences |
|---|
In haematology, PACE relies on the continuing co-operation of all 28 haematology consultants in the region. Regular communication between the doctors is essential for the process. There was and remains a weekly haematology meeting open to (and organized by) all the regional consultants (and trainees) held centrally in Newcastle upon Tyne. The meetings are universally recognized as an open forum for debate on policy issues and management of difficult patients. This opportunity to be actively involved in policy and trial development means that even the most geographically distant members of the group find regular attendance worthwhile.
By the end of the 1980s, a mutually agreed ‘menu’ of treatments for the various diseases treated by the group had been introduced, compiled as a result of an ongoing research/audit strategy and regular review of current literature.5–7 The ‘menus’ consist of a mixture of accepted ‘gold standard' therapies and national or regional trial protocols or studies. This ‘menu book’ has facilitated the introduction of new pilot studies in experimental treatments. The ongoing data collection facilitates the research approach, by recording patient selection biases.8 We have found that the process maximizes recruitment into trials and studies across the Region: e.g. over 75% of eligible patients with poor-risk Hodgkin's disease have been offered a randomized controlled trial of intensive treatment8 and over 80% of patients were recruited to a study on the feasibility of collecting Philadelphia negative cells from the peripheral blood stem-cell harvests of patients with chronic myeloid leukaemia.9
|
Can it work for other disciplines? |
|---|
|
TopIntroductionPACE: the process and...Communication: the key elementCan it work for...Developing the systemAmalgamations and collaborationsPracticalities of financing and...Building the process into...PACE in actionIndividual advice on patient...ConclusionReferences |
|---|
Following the publication of the original paper describing the process and other papers using the haematological PACE data, we have been surprised by the interest shown from other specialties. We anticipated that the cancer specialties, in particular, might be interested in this approach, and presently there is an ongoing feasibility study for the introduction of a programme based on the existing haematological model for breast cancer in our region, as part of the Calman process. What was rather more surprising has been the interest expressed from areas of medicine as diverse as psychiatry, primary care and care of the elderly. Following discussions with doctors in psychiatry, who have commenced a PACE project, we are now convinced that the approach should translate well to most medical specialties, both to provide uniformity of care and as a vehicle for research into practice and practice into research.
The remainder of this article aims to give a preliminary, practical guide to the evolution of the process, shown as a flow diagram in Figure 2 and described point-by-point below.
|
|
Developing the system |
|---|
|
TopIntroductionPACE: the process and...Communication: the key elementCan it work for...Developing the systemAmalgamations and collaborationsPracticalities of financing and...Building the process into...PACE in actionIndividual advice on patient...ConclusionReferences |
|---|
- 1. In areas of clinical practice where randomized controlled trials (RCTs) have been difficult to conduct, clinicians interested in a given disease process arrange to pool resources over a defined geographical area and patient population.
2. All clinicians in the geographical area with an interest for a given disease/treatment should be made aware of the programme and involved from the outset. Opinions should be widely sought on what should be included in the study programme, and consultants must be persuaded that the collection of the data can become a continuous part of their job. We estimate that the maximum size of group that can co-operate in such a way would be 30–40 doctors. From the outset, concepts of trust and ownership of the data are of paramount importance (Figure 3
), as the imposition of such a process from an outside source usually results in negative responses. Some clinicians are lukewarm or negative on the idea initially. However, if approached diplomatically and reassured their clinical choices are not being compromised, all eventually see the value of the approach, and many see PACE as the ideal physician-led vehicle for compliance with new Clinical Governance directives. 3. The next stage is an educational one in which the concepts of the PACE process are explained and the strengths of the idea of uniform registration discussed.
4. Not all the doctors caring for patients with the conditions being investigated can be expected to have an interest in the day-to-day management of the process, so it is important to establish a Speciality Working Group and a regular forum for meetings. The membership of this Working Group must be representative of the peer group. Members should be reminded that the data is owned collectively (Figure 3).
5. At this point it may be useful to engage the Trust management team, Health Authorities and Hospital Audit Groups to gain their support and (hopefully) some degree of financial backing. Audit departments can be potentially very helpful with the regular follow-up of cases, once the patient population has been identified and registration forms are in place.
6. The Speciality Working Groups should, as a first task, begin work on the data collection system. This would usually include a very simple registration form and a much more detailed secondary form, according to the needs of the speciality, to record treatment details. Treatment details are best collected once treatment programmes have been completed and therefore ‘intention to treat’ and actual treatment can both be registered more accurately, e.g. on an annual follow-up form.
7 and 8. The first year is run as a pilot, to enable defects to become apparent and changed. Following feedback from the doctors involved in the process and the finalization of data collection forms, the second year collection system can begin. Simultaneously, the organization and computerization of the data and the collection of core information data will have been agreed. A ‘neutral’ central co-ordinator (preferably medical) is essential here (Figure 4), to reassure colleagues about the collective ownership of the data.
9. Once the population of doctors have seen the process going for two years and have been involved in regular feed-back sessions, the climate is then right for ‘functional audit’. Functional audit we define as an audit in which >90% of interested physicians meet and produce an agreed therapeutic strategy or intervention which can be documented and put into practise immediately. This is the mechanism whereby research findings immediately enter regional clinical practice: involvement and personal communication are the key factors. (Figure 3).
10. At the outset of the programme it may be difficult to define in great detail which particular research/audit questions need to be addressed. Once information becomes available, the questions that need answering become self-evident and potential trials and studies can be readily identified. The data collection forms can then be modified or additional forms added to enable specific questions to be answered.
|
|
|
Amalgamations and collaborations |
|---|
|
TopIntroductionPACE: the process and...Communication: the key elementCan it work for...Developing the systemAmalgamations and collaborationsPracticalities of financing and...Building the process into...PACE in actionIndividual advice on patient...ConclusionReferences |
|---|
Once the above process has been in place for sometime, it is possible to arrange various amalgamations of the data collected from other similar registration groupings, as long the data collection system has also encompassed entire patient populations. For instance, in collaboration with colleagues from the whole of Scotland, we contribute to a data collection system for 90% of patients with malignant non-Hodgkin's lymphoma and Hodgkin's disease, providing details of diagnosis, treatment and outcome across a population of 8.5 million people.
Additional collaborations have been possible between the Haematology groups in the former Northern, North West, Mersey, Oxford, West Midlands and South East Thames Regions. Here again there is an agreed ‘core’ database to collect information on unselected populations of acute lymphoblastic,10 acute myeloid leukaemia11,12 and chronic myeloid leukaemias.9 A recent study demonstrated that there is no advantage in survival terms from centralizing care in young adults with acute leukaemia.13 The total population from which these are drawn is 25 million people, and gives fascinating information on patterns of care and realistic overall survival data.13
|
Practicalities of financing and data collection |
|---|
|
TopIntroductionPACE: the process and...Communication: the key elementCan it work for...Developing the systemAmalgamations and collaborationsPracticalities of financing and...Building the process into...PACE in actionIndividual advice on patient...ConclusionReferences |
|---|
The response of most individuals listening to an account of this approach is to predict that it must be extremely labour-intensive and very expensive. Most individuals when introduced to the concept think immediately of computer hardware and software: we believe that these issues are much less important than the basic idea that the enthusiasm of the personnel involved in the process is the key to its success.
There are, we believe, two key posts for a given disease group in a geographical area. Our experience of implementing this process in haematology, psychiatry and breast cancer have suggested that a medically-qualified person, working part-time or whole time as the ‘neutral’ co-ordinator, is a critical post (Figure 4
). Such a person requires knowledge of the area of medicine studied, but most importantly must be a person that all participants can trust. The medical qualification for this key individual enables the person to debate and discuss with consultant colleagues the clinical detail on equal terms.
The second post is a part-time or whole-time research secretary or research associate to work with the co-ordinator and the speciality group. The degree of additional data management would largely be dependent on the volume of patients and the size of the geographical area to be covered. In general, we estimate the introduction of a PACE programme for a 3 million population with some 30 consultants would require a funding base of some £50 000 per annum.
|
Building the process into practice |
|---|
|
TopIntroductionPACE: the process and...Communication: the key elementCan it work for...Developing the systemAmalgamations and collaborationsPracticalities of financing and...Building the process into...PACE in actionIndividual advice on patient...ConclusionReferences |
|---|
The transformation of this process from the level of a project into a continuous evaluation which is part of the everyday practise of the consultants concerned, is what makes PACE different from classical audit studies. Once the whole body of interested consultants have been engaged in the programme, the key individuals are the consultants' secretaries. The secretaries can be brought together to the central data-collecting department, meet the individuals concerned and be educated about the importance of the whole programme. It is key to the success of the programme that total capture of patient data is maintained, and since all the paperwork for the patients concerned will go through the hands of this group of people, enthusiasm and commitment at this level is critical. Once this group of individuals has been identified and met, various processes can be put in place for checking and cross-checking data, for example in the case of cancer, diagnoses can be checked through histopathology and cancer registries.
|
PACE in action |
|---|
|
TopIntroductionPACE: the process and...Communication: the key elementCan it work for...Developing the systemAmalgamations and collaborationsPracticalities of financing and...Building the process into...PACE in actionIndividual advice on patient...ConclusionReferences |
|---|
A total care package for Hodgkin's disease—PACE based
The Scotland & Newcastle Lymphoma Group (SNLG) began collecting data on all patients with Hodgkin's disease in 1979. By 1982, it became evident that 50% of patients with advanced disease appeared to be cured with available four-drug chemotherapy, leaving 50% who were not. The obvious question concerned the differences between the two groups and whether they could be identified at diagnosis by objective factors with prognostic implications for that disease. The data which had been prospectively collected was analysed for all known risk factors taken from available literature and a numerical prognostic index derived from 100 consecutive patients from one institution.14 This index was then validated on the group of patients from the Scottish Centres where similar information had been prospectively collected. The index was found to be sufficiently robust to provide more accurate predictions of outcome than existing classical staging for introducing a new aggressive form of treatment. Twenty-five per cent of the total Hodgkin's population fell into a ‘poor-risk’ group, i.e. patients predicted to die on ‘best available’ therapy in the mid-1980s. In 1988, following piloting of a new chemotherapeutic intensive eight-drug schedule, and a new form of auto transplant15 on relapsed patients, a prospective RCT was introduced to the ongoing population study. Patients who were not ‘poor risk’ but advanced stage were not eligible for the randomized trial, and physicians were encouraged to use four-drug combinations with established long-term toxicities in this group of patients. The existence of the database facilitated the evaluation of patients not entered into the trial. The existence of the prognostic index further cemented the data collection process which was itself enhanced by the addition of detailed audits on the efficiency of staging. From 1991 an audit of radiological investigations was introduced, as well as a central review of all pathological samples. Within a year, it was clear that only 12% of patients were being accurately staged by CT scan using existing guidelines. An audit the following year led to an improvement of radiological staging procedures and guidelines, strengthening the quality of the selection of patients for the RCT and overall care of patients within the Region.8 Seventy-five per cent of eligible patients for the RCT with advanced disease were entered into the trial.
|
Individual advice on patient care using PACE data |
|---|
|
TopIntroductionPACE: the process and...Communication: the key elementCan it work for...Developing the systemAmalgamations and collaborationsPracticalities of financing and...Building the process into...PACE in actionIndividual advice on patient...ConclusionReferences |
|---|
Request letter from consultant colleague
‘Re: Outcome measures for acute lymphoblastic leukaemia patients aged 50–60 years. I wondered whether you could easily lay your hands on the outcome data for patients on the North East acute lymphoblastic leukaemia protocol between the ages of 50–60? We have encountered a lot of problems in one of our patients with periods of prolonged pancytopenia and that is before we have even reached the intensification section of therapy. She suffered a grand mal convulsion possibly due to a small intracerebral haemorrhage in February. I wonder whether there is any evidence that increasing intensity of treatment has actually benefited patients in this age group?’
Reply
‘We have 34 unselected patients on the Northern Register within this age group. The majority of patients (27 out of 34) had treatment with curative intent. Complete remission rate was 85%; this is impressive considering that 11 of the patients had very poor prognosis disease based on cytogenetics. Most patients relapsed (median time to relapse 4 months, range 1–23) but 6 went into long-term remissions (1 died in remission of an unrelated cause at 13 years). Thus, of those treated with curative intent, 20% are destined to survive beyond five years. It would, therefore, seem reasonable to attempt maximum tolerable therapy in your patient but, if you really feel that she cannot tolerate further intensive treatment, then perhaps continuing maintenance therapy might be an option. I hope this information is of some help to you.’
Such relevant information for individual patient management is not readily available from other sources.
|
Conclusion |
|---|
|
TopIntroductionPACE: the process and...Communication: the key elementCan it work for...Developing the systemAmalgamations and collaborationsPracticalities of financing and...Building the process into...PACE in actionIndividual advice on patient...ConclusionReferences |
|---|
There is little doubt that the introduction of the PACE process for haematological malignancy has created uniformity of care for adults throughout a region of 3 million people. The whole process has become a part of routine care within the region, and all haematologists are not only actively involved but extremely enthusiastic and very proud of the mutual achievements. We believe that it provides a robust vehicle which has the capacity to make an impact on large areas of secondary care and potentially in the future may also be modified to enhance and facilitate research activity in primary care. Once established, the research potential is substantial and it is necessary and valuable to discuss evolving projects with colleagues in Health Service Research, epidemiology and medical statistics to ensure application of appropriate methodology and analysis.
|
Acknowledgments |
|---|
The authors wish to thank the following members of the Northern Region Haematology Group: Dr M. Abela, Dr N. Browning, Dr P. Carey, Dr R. Cartner, Dr C. Chapman, Dr J. Chandler, Dr P. Condie, Dr M. Dewar, Dr R. Finney, Dr M. Galloway, Dr D. Goff, Dr P. Hamilton, Dr A. Hendrick, Dr G. Jackson, Dr F. Keenan, Dr P. Kesteven, Dr Z. Maung, Dr I. Neilly, Dr H. O'Brien, Dr P. Saunders, Dr D. Stainsby, Dr G. Summerfield, Dr H. Tinegate, Dr J. Wallis, Dr N. West, Dr P. Williamson, Dr A. Youart.
Dr Penny Taylor, medical co-ordinator of Haematological PACE, has been funded variously by the Tyneside Leukaemia Research Association, the Northern & Yorkshire Regional Research & Development Grants and Northern Regional Audit Grants and as Associate Specialist by the Northern Regional Health Authority.
Margaret Graham and Jennifer Wilkinson have been involved as Research Secretary and Personal Assistant over a period of 15 years. Mrs Linda Burn is the Data Manager for Greater Tyneside. We would like to thank Jennifer Wilkinson for preparation of the manuscript. The Secretariat and members of the Therapy Working Party of the Scotland & Newcastle Lymphoma Group are involved in all aspects of the PACE lymphoma programme.
|
References |
|---|
|
TopIntroductionPACE: the process and...Communication: the key elementCan it work for...Developing the systemAmalgamations and collaborationsPracticalities of financing and...Building the process into...PACE in actionIndividual advice on patient...ConclusionReferences |
|---|
1. Black N. Developing high quality clinical databases. Br Med J 1997; 315:381–2.
2. Gruer R, Gordon DS, Gunn AA, Ruckley CV. Audit of surgical audit. Lancet 1986; 1:23–6.[Medline]
3. Paterson CM, Chapple JC, Beard RW, Joffe M, Steer PJ, Wright CS. Evaluating the quality of maternity services—a discussion paper. Br J Obstetr Gynaecol 1991; 98:1073–8.
4. Buckley NA. Poisoning and epidemiology: ‘Toxicoepidemiology’. Clin Exp Pharmacol Physiol 1998; 25:195–203.[Web of Science][Medline]
5.
Charlton B, Taylor PRA, Proctor SJ. The PACE strategy: a new approach to multicentre clinical research. Q J Med 1997; 90:147–51.
6. Engels EA, Spitz MR. PACE-setting research. Lancet 1997; 350:677–8.[Medline]
7.
Proctpor SJ. Why clinical research needs medical audit. Quality Health Care 1993; 2:1–2.
8.
Taylor PRA, Angus B, Owen JP, Proctor SJ. Hodgkin's disease: a population-adjusted clinical epidemiology study (PACE) of management at presentation. Q J Med 1998; 91:131–9.
9. Lennard AL, Storey N, Dickinson AM, Irving JA, Rowe E, Conn JS, Proctor SJ. Collection of Philadelphia-negative peripheral blood progenitor cells in unselected patients with chronic granulocytic leukaemia. Leukemia 1998; 12:746–52.[Medline]
10.
Taylor PRA, Reid MM, Bown N, Hamilton PJ, Proctor SJ, on behalf of the Northern Region Haematology Group. Acute lymphoblastic leukemia in patients aged sixty years and over: a population based study of incidence and outcome. Blood 1992; 80:1813–17.
11. Taylor PRA, Reid MM, Stark AN, et al. De-novo acute myeloid leukaemia in patients over 55 years old. A population based study of incidence, treatment and outcome. Leukemia 1995; 9:231–37.[Web of Science][Medline]
12. Proctor SJ, Taylor PRA, Stark A, et al. Evaluation of the impact of allogeneic transplant in first remission on an unselected population of patients with acute myeloid leukaemia aged 15–55 years. Leukemia 1995; 9:1246–51.
13. Stiller CA, Benjamin S, Cartwright R, et al. Patterns of care and survival for adolescents and young adults with acute leukaemia. Br J Cancer 1999; 79:658–65.[Medline]
14. Proctor SJ, Taylor P, Donnan P, et al. A numerical prognostic index for clinical use in identification of poor risk patients with Hodgkin's disease at diagnosis. Eur J Cancer 1991; 27:624–9.
15. Proctor SJ, Taylor P, Mackie MJ, et al. A numerical prognostic index for clinical use in identification of poor-risk patients with Hodgkin's disease at diagnosis. Leukemia Lymphoma 1992; 7(Suppl.):17–20.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  M B Muzaimi, J Thomas, S Palmer-Smith, L Rosser, P S Harper, C M Wiles, D Ravine, and N P Robertson Population based study of late onset cerebellar ataxia in south east Wales J. Neurol. Neurosurg. Psychiatry, August 1, 2004; 75(8): 1129 - 1134. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M R Velangi, M M Reid, N Bown, G H Jackson, G P Summerfield, S J Proctor, and P R A Taylor Acute lymphoblastic leukaemia of the L3 subtype in adults in the Northern health region of England 1983-99 J. Clin. Pathol., August 1, 2002; 55(8): 591 - 595. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||