RISKY BUSINESS
Marshall Marinker has characterised the role of the GP as being to ‘marginalise danger’, contrasting this with the specialist, whose diagnostic role is to ‘marginalise uncertainty.1 In other words GPs have the often difficult task of identifying the minority of patients whose presenting symptoms represent serious illness from the majority who do not have something seriously wrong. The specialist is then required to use the skills and technologies of secondary care medicine to arrive at a precise diagnosis. This conceptualisation retains validity today, although the pressures to achieve greater diagnostic precision and for speedier diagnosis in primary care have increased greatly in fields such as cancer diagnosis, the assessment of febrile children, coeliac disease, inflammatory bowel disease, and infectious diseases. The intervening years have also seen a considerable research effort in fields such as the assessment of risk, using increasingly sophisticated scoring systems, the communication of risk, in which the optimum ‘depiction’ of risk for patients has been studied, and the incorporation of these into more transparent decision-making shared between patients and doctors. There is also a continuing research interest in the extent to which an individual’s willingness to tolerate risk or their risk-aversion influences their clinical behaviour, and how these might be modified.
In this issue of the BJGP we examine safety and risk from a number of perspectives, including the generation of prescribing-safety indicators, variation and subtherapeutic prescribing of penicillins to children, potentially hazardous over-the-counter sales of NSAIDs, the use of clinical prediction rules, and the effect of communicating risk and therapeutic effectiveness in different formats. Two excellent editorials set the scene and place this research in context. Carl de Wet points out:
‘Safety is often defined as a condition where “nothing goes wrong,” which is a fine description of a successful aeroplane journey or an incident-free day at a nuclear power plant. However, in the general practice context, this translational meaning of safety is improbable. Perhaps it is more realistic at the outset, therefore, to focus our safety improvement efforts on known issues that can be prevented, such as the preliminary list of “never events” reported in the BJGP last month.’2
Writing from Sydney, Australia, Lyndal Trevena comments on the:
‘ ... perceived tension between population-level quality frameworks and the underpinning principles of general practice ... none of which are easily captured or identifiable in most quality frameworks’.
One of the most powerful research approaches used in the measurement of risk of various kinds is the analysis of clinical data held in large databases, and many of the risk assessment tools in clinical use have been derived from large database research. Studies of matched cohorts of patients followed over time can provide valuable information on the chances of developing certain outcomes after consulting with a particular symptom, such as the incidence of colorectal cancer in patients with a primary complaint of rectal bleeding. These databases are also crucial tools in monitoring the adverse effects of drug treatments (pharmacovigilance) and their size gives them the statistical power to detect rare but important events. It is particularly unfortunate that NHS England’s Care.data project got off to such a bad start and fuelled, rather than calmed, concerns about data insecurity, commercialisation, and failures of anonymisation. This is especially ironic when the General Practice Research Database, now part of the CPRD (Clinical Practice Research Datalink), has been running for over 20 years without a single adverse event involving data leakage or disclosure of identity, as well as generating a multitude of significant and clinically important studies, thanks to its strong and effective operational, governance, and ethical procedures.
- © British Journal of General Practice 2014