Health checks and screening: what works in general practice?

I know that prevention is better than cure but it’s still the bit of general practice I have always enjoyed least. It’s hard to be motivated by something not happening. It’s easier to be motivated by screening, where the aim is to detect early-stage disease and there is often compelling evidence for the effectiveness of treatment. But GP enthusiasm for prevention has been dampened by ill-conceived NHS policy initiatives since at least 1990.¹ Dalton et al ² are not the first to ask whether NHS Health Checks are defensible when measured against World Health Organization screening criteria.³

MORE EVIDENCE AGAINST HEALTH CHECKS

Enthusiasm for health checks will be further dampened by two other studies in the BJGP. Caley et al ⁴ reported that they have surprisingly little impact on the recognition of undiagnosed diabetes, hypertension, chronic heart disease, chronic kidney disease, or atrial fibrillation. Koekkoek et al⁵ point out that the evidence favours targeted stepped interventions, which avoid investing scarce resources in the worried well. And the evidence from all three articles is consistent with the Cochrane meta-analysis which makes it quite clear that promoting health checks in unselected adult patients has limited impact on cardiovascular risk and no significant impact on cardiovascular mortality.⁶

However, it is important not to throw out the baby with the bathwater. Many of the constituent elements of health checks (for example, smoking advice, blood pressure management, and statin prescribing) are trial-proven effective interventions. The studies by Korhonen et al ⁷ and Gil-Guillen et al, ⁸ as well as Caley et al, ⁴ confirm that population screening in primary care can detect undiagnosed cardiovascular risk. So what is going on here? Why is the evidence inconsistent?

WHY IS THE EVIDENCE INCONSISTENT?

The first half of the answer to this question is straightforward. We have known for more than 20 years that those at highest risk of cardiovascular disease are the least likely to attend for health check screening.⁹ We have known for even longer that identifying risk does nothing but harm if you don’t go on to manage it effectively. Effective management is less likely in programmes that try to deal with multiple rather than individual risks; for example, smoking cessation advice is less effective when given in the context of multifactorial health checks.¹⁰

The second half of the answer is more nuanced: context is usually more important than content in determining the effectiveness of a complex intervention so trial evidence is time and context specific. The pioneering North Karelia community-based programme of cardiovascular disease prevention (begun in 1972) provides a good example. The initial trial evidence showed no significant reduction in smoking or weight and a significant reduction in cholesterol only in males and not females.¹¹ The programme was nevertheless rolled out to the rest of Finland and subsequently associated with a fall in coronary mortality of 80%.¹²

LESSONS FROM STANFORD AND NORTH KARELIA

This is not the only important lesson from North Karelia. The most effective element of the initial intervention was the one in which primary care played the greatest role: a 46% reduction in the number of people with dangerously raised blood pressure (defined as a diastolic blood pressure >100 or systolic blood pressure >175 mmHg).¹¹ However the subsequent fall in mortality has been attributed much more to changes in diet and tobacco use achieved through government legislation, fiscal policy, and mass education rather than personalised intervention; even the major reduction in blood pressure was achieved less by identifying and treating those at very high risk than a shift to the left of the population mean.¹² Similarly in the Stanford project in the US (the other major cardiovascular prevention project in the 1970s), the initial impact of personalised care by health practitioners was transient in the context of wider community education and major secular change and was not thought to have had sufficient impact to be included in the main five-cities programme.¹³

So what can we learn from this historical evidence? Personalised health interventions tend to be swamped by other factors influencing secular trends in health behaviour. This does not mean that GPs should play no role in primary prevention: patients notice that their GP doesn’t smoke and medical opinion as a whole is an important element of the political context, which makes legislative and other changes possible. But it does mean that giving lifestyle advice in primary care is seldom cost effective. We concluded in 1990 that GPs:

‘... should be careful not to absolve the government of its public health obligations by substituting unproved preventive interventions aimed at the individual patient.’⁹

This is as true now, as we consider how to respond to an epidemic of obesity, as it was when written 24 years ago.

WHEN GOVERNMENTS FAIL

However, governments often don’t act effectively and primary prevention fails. In this situation, we can’t escape from picking up the pieces through secondary prevention. As already cited, the North Karelia project demonstrated that primary care can do this task of secondary prevention very effectively.¹¹ But the key to effective secondary prevention is recognising that case finding and subsequent clinical management are essential but separate components. Both components need to be actively managed and quality-assured. In terms of case finding, Koekkoek et al ⁵ draw attention to the greater effectiveness of stepped screening programmes, with the first step being identification of patients most likely to benefit from a resource-intensive face-to-face appointment. They also echo Lindenmeyer et al¹⁴ in stressing the importance of proactive patient and community engagement to increase the attendance for screening of those at greatest risk. In terms of clinical care, Dalton et al² imply that when risk factors such as hypertension or hypergylycaemia are identified by screening we have a responsibility to manage them with the same attention and quality control as premorbid conditions identified in other national screening programmes such as breast cancer.

Lindenmeyer et al ¹⁴ give a practical example. One important reason for diabetic case finding is that blindness from micro-vascular complications is preventable. Diabetic retinopathy screening uptake apparently varies between practices from 55% to 95%. Although three contributory factors are beyond our control (social deprivation, ethnic diversity, and transport access), GPs can substantially improve screening uptake by contacting patients and encouraging them to attend, integrating screening with routine care, and facilitating good communication with regional screening teams. Preventing blindness seems a particularly compelling argument for action but these commonsense observations must apply to most other national screening activities, all of which are supported by trial evidence of their effectiveness. They also apply to vaccination. The variation in vaccine uptake between practices, and the key role GPs can and should play in maximising compliance, has been discussed in this journal many times before, most recently in the context of influenza vaccination for at-risk children.¹⁵

SO WHAT WORKS?

So to return to the title — what works in general practice? The answer is simple: secondary prevention. Secondary prevention certainly includes identification and management of patients with high blood pressure, high blood sugar, hyperlipidaemia, or renal insufficiency. It also includes advising people to stop smoking (it’s secondary prevention because it’s only necessary when failure of primary prevention means people start smoking in the first place). And in each case, it involves case finding, but not through NHS Health Checks. This NHS preventive flagship merits scuttling because it’s unfit for purpose. It’s inefficient at case finding, strays into primary prevention, and lacks an adequate quality-assurance mechanism to ensure subsequent treatment is effective.

In designing a better programme we should give thought to the damage we have caused by unnecessarily medicalising another important NHS prevention programme: the provision of contraception. Not everyone has the ability and motivation to self-care, but surely we should not only be involving patients in self-recognition of risk but should also be delegating to them as much responsibility as possible for its subsequent management. Patient self-monitoring may not only be more convenient for them, it may also be more effective.¹⁶ And ‘fire and forget’ may sometimes be a more efficient and cost-effective prescribing strategy than ongoing clinician monitoring in primary care. So let’s look forward to moving from health-checks to facilitated self-checks; and for cardiovascular disease, to a more targeted, quality-assured, and evidence-aware programme.

Provenance

Commissioned; not externally peer reviewed.