How to assess quality in primary care

BMJ 2015; 351 doi: https://doi.org/10.1136/bmj.h5950 (Published 06 November 2015) Cite this as: BMJ 2015;351:h5950
  1. Robert Grant, senior lecturer in health and social care statistics 1,
  2. Jim Parle, professor of primary care2
  1. 1Kingston University and St George’s, University of London
  2. 2University of Birmingham, Birmingham, UK
  1. Correspondence to: R Grant robert.grant{at}sgul.kingston.ac.uk

The Health Foundation has produced a useful review of indicators

As patients in the UK we are entitled to a general practitioner. But how should we choose which one is right for us? How do we find out about the quality of general practices we might consider joining? Other stakeholders—commissioners, government, Monitor, the taxpayer—are at least as interested in the answer to this question as patients.

To help answer these questions the Health Foundation has produced, at the instigation of the health secretary, a review on indicators of quality in English general practice. What does it tell us? And what does it leave out?

Quality is an essential but slippery concept, reflected imperfectly in each of the many ways we can assess structure, process, and outcomes in primary care. Despite the difficulties, judgments about primary care are possible, and guidance has been available for some time.1 2

The Health Foundation has reviewed current sources of evidence (and there are many, of varying provenance and accessibility) and considered the risks inherent in choosing, presenting, and combining indicators. But importantly there is no mention of uncertainty. All measures of quality have uncertainties, and we should tackle the challenge of understanding these and communicating them effectively. We do not believe that any data are better than no data. Data need to be good enough to allow judgments (be they about resources, access, choice of practice, or whatever), even if they don’t need to be perfect.

The report includes a strongly worded caution against composite indicators, which we welcome. We have seen too much reduction of multifaceted healthcare into traffic lights and league tables, and the complexity of the system needs to be acknowledged. We hope that the current health secretary’s fondness for composite indicators in league tables will be tempered by reading this report.3 Likewise, a minimal set of “sentinel indicators” cannot meaningfully be ranked—but undoubtedly someone will. It may be best to avoid both composite and sentinel indicators, make a wide range of indicators available, and allow consumers to choose what matters to them. If you are over 75, for example, you might wish to know the practice’s data for that group; ditto, if you have a longstanding condition. Individuals have different wants and needs so a wide range of data is required to satisfy everybody’s requirements. Similarly, a commissioner may want to look at, say, life expectancy compared with practices in a similarly deprived area, and so on.

Interactive visualisations of data offer a beacon of hope; the theory goes that if consumers of information have many different needs, a single source is not enough.4 Any website should be layered, allowing users to dip in to the level that matters for them. The US website State of Obesity (http://stateofobesity.org) is an exemplar.

The Health Foundation report does not take sufficiently seriously the possibility of unintended consequences: indicators are known to skew behaviours, especially when linked to commissioning and payment. Consider, for example, gaming around the four hour emergency department limit or general practitioners’ refusal to book patients more than 48 hours ahead to escape penalties.5 Sentinel indicators could also skew behaviour, just as practices responded (rationally) to the Quality and Outcomes Framework by focusing on included indicators with inevitable consequences for other activity.6 7

The report recommends rationalising the current smorgasbord of indicators, reports, and websites from a wide range of providers while preserving the virtues of the current systems (for example, the ease of use and popularity of NHS Choices). There is a need for uniform reporting and sharing of know-how across the various groups collecting such data and to involve, as the National Institute for Health and Care Excellence already does, a range of stakeholders in developing indicators.

The Health Foundation’s report could bring clarity and value to the consumers of data about primary care and, with the caveats above, show a clear path to improvement in the relevance, validity, and accessibility of quality indicators. We hope that its messages win out with the health secretary over the seduction of simple composites, targets, and league tables.


Cite this as: BMJ 2015;351:h5950


  • Competing interests: We have read and understood BMJ policy on declaration of interests and declare that RG is a board member of the National Advisory Group on Clinical Audit and Enquiries, has provided training on data analysis software for Monitor, and statistical analysis consultancy for the National Clinical Guideline Centre and the Nursing and Midwifery Council.

  • Provenance and peer review: Commissioned; not externally peer reviewed.


View Abstract

Log in

Log in through your institution


* For online subscription