Rapid Responses to:
|
|
Rapid Responses: Rapid Responses published:
-
![[Read Rapid Response]](/icons/misc/sectionDOWN.gif)
The eulogy of a quality measure: a perspective from the UK.
- Jose M Valrderas, Oxford Road. Williamson Building. Manchester. M13 9PL (2 November 2007)
-
healthcare measurements must be collaborative measurements
- Arun K Chopra (8 November 2007)
-
Outcome Performance Indicators and Among Institution Data.
- Anthony P Morton (10 November 2007)
|
|
|||
|
Jose M Valrderas, Clinical Lecturer NIHR School for Primary Care Research. University of Manchester., Oxford Road. Williamson Building. Manchester. M13 9PL
Send response to journal:
|
In their analysis on the measurement of quality through performance[1], Majeed A et al. suggest that some performance indicators may not necessarily be unique to each country, making international comparisons possible. The use of performance measurements in different countries can indeed provide important clues on the pathways to successful improvement of the quality of health care. After observing in the United States both widespread high levels of achievement and very low variability across health plans for one of such indicators, the percentage of patients with acute myocardial infarction who receive a prescription for beta-blockers within 7 days of hospital discharge, the National Committee for Quality Assurance will no longer support its use[2]. It has been suggested that one of the drivers of this success has been a positive environment, rather than the availability of evidence of the effectiveness of the treatment[2]. In England and Wales, similar patterns have been observed for the prescription of both beta- blockers and statines[3,4]. Considering that available evidence for the benefits of secondary prevention is far more recent for statins than for beta-blockers[5], the use of statins, might well have benefited from the positive environment created around secondary prevention of coronary heart disease (halo effect), but it might have itself played a role in raising the attention on secondary prevention of ischemic heart disease in general. Even before greater standardisation of clinical data and performance indicators paves the ground for meaningful international comparisons, evidence from other countries can provide relevant information for the implementation of successful strategies for quality improvement. 1. Majeed A. Lester H, Bindman AB. Improving the quality of care with performance indicators. BMJ. 2007;335:916-918. 2. Lee TH. Eulogy for a quality measure. N Engl J Med. 2007 Sep 20;357(12):1175-7. 3. Walker L, Birkhead J. Weston C. Pearson J. Quinn T. How the NHS manages heart attacks. Royal College of Physicians. London. 2007. ISBN 978 -1-86016-310-4 4. Myocardial Infarction National Audit Project. How hospitals manage heart attacks. Royal College of Physicians. London. 2002. Available at: http://www.rcplondon.ac.uk/pubs/books/minap/HowHospitalsManageHeartAttacks12Nov2002.pdf 5. Ward S, Lloyd Jones M, Pandor A, Holmes M, Ara R, Ryan A, Yeo W, Payne N. A systematic review and economic evaluation of statins for the prevention of coronary events. Health Technol Assess. 2007 Apr;11(14):1- 178. Competing interests: None declared |
|||
|
|
|||
|
Arun K Chopra, Special Lecturer Nottingham University, Department of Psychiatry, Queens Medical Centre,NG7 2UH
Send response to journal:
|
Majeed et al's article regarding the use of quality indicators in healthcare is timely. Whilst many healthcare professionals will agree that the assessment and publication of performance league tables can lead to improvements in heathcare, it is important to mention the strong possiblity that in the absence of sophisticated markers which take into consideration the psychosocial and collaborative nature of medicine, such markers run the risk of misinforming the public. Patients are different. They are not 'raw materials'. Socio-economic differences,different healthcare beliefs and different levels of adherence and self-agency will affect healthcare outcomes. These psychosocial determinants need to be built into any measurement model to ensure that like is compared with like when healthcare league tales are published. Although we might thnk that people will read the small print about the patient-related factors when glancing at such league tables, the reality might be quite different. In an era where choice reigns supreme and information seeking is related to educational background, performance measures without this degree of sophistication, might have disastrous effects on heathcare provision in areas of social deprivation. Competing interests: None declared |
|||
|
|
|||
|
Anthony P Morton, medical quality improvement statistician Princess Alexandra Hospital, Woolloongabba, Brisbane, Australia 4102
Send response to journal:
|
There seems to be no clear idea about the use of outcome indicator data.1 There is increasing advocacy for their use to improve safety2 in spite of continuing lack if evidence that they are useful for this purpose.3 It is necessary for us to understand why we collect these data before they are collected so that relevant data are collected and they are analyzed productively. We suggest that it is always useful to collect system and process data.3 However, there must be a strong emphasis on the prior implementation of evidence-based systems of patient care.4 Local sequential monitoring of within institution outcome data can provide a useful early warning of unexpected adverse occurrences. Therefore, we propose that there should be a distinction between local sequentially analyzed within institution outcome data and among institution data that are frequently collected by central authorities, often aggregated for example by years, and used to compare institutions. We maintain that the latter process is flawed. Among institution outcome data need to be collected for reasons of accountability and transparency but their chief potential usefulness is to enable us to understand how to devise better evidence based systems using health services research.3 In order to be useful for the latter function, thought needs to be given to the data collected. Are the same data needed for this research as those that are collected to perpetuate the myth that safety is thereby improved? We are unaware of any controls used when improvement is claimed to be due to using among institution outcome data and we believe there are institutions that have yet to be subjected to this approach that do excellent work and have lowered postoperative mortality due to learning curves. There have been notable successes with the use of stochastic mathematical models to improve understanding of transmission of hospital-acquired bacteria.5 However, such studies need to involve many institutions and to be ongoing. We propose that central authorities institute Bayesian networks to analyze among institution data and that these be self-perpetuating. Setting up such networks would require a great deal of high-level statistical expertise but, when running, they could probably be automated. Initially, much of the required prior information would need to come from experts in their fields. However, as each new piece of information became available, the models would be updated producing new posterior knowledge (and improved evidence) that in turn would become the prior for the next batch of data. For such a mechanism to succeed, the among institution data collected would need to be those that are most likely to answer specific questions about gaps and weaknesses in existing evidence-based systems. There is little evidence that the existing approach achieves its stated aim of improving safety.3 Furthermore, without careful consideration, the data that are collected may prove useless for the worthwhile aim of devising better evidence-based systems. References. 1. Majeed A, Lester H and Bindman A “Improving the quality of care with performance indicators” BMJ 2007;335:916-918. 2. Sullivan E, Baker R, Jones D, Blackledge H, Rashid A, Farooqi A and Allen J “Primary health care teams’ views on using mortality data to review clinical policies” Quality and Safety in Health Care 2007;16:359-362. 3. Lilford R, Brown C and Nicholl J “Use of process measures to monitor the quali ty of clinical practice” BMJ 2007;335:648-650. 4. Berwick D, Calkins D, McCannon J and Hackbarth A “The 100000 lives campaign” JAMA 2006;295:324-327. 5. Cooper B “Confronting models with data” Journal of Hospital Infection 2007;65:88-92. Competing interests: None declared |
|||