What prevention works in medical practice – a Canadian perspective

James A Dickinson

HK Pract 2020;42:25-27

The core business for Family Practice is treating disease and distress. But we are urged to emphasise prevention too. Prevention is sold as good business by some in the health sector. Yet there is complexity: for us to recommend a preventive activity, we must be confident that the benefits outweigh the potential harms.

Conventionally we think in ter ms of primary, secondary and tertiary prevention, but some have added primordial, and quaternary as well. ¹ Potentially the benefits are greatest for primordial and primary prevention through social determinants of health, both at a population level, and in the clinic, for individuals. Those without a stable home nor adequate food, have difficulty thinking beyond the next day or two, and thinking forward a few years is irrelevant. Helping them to stabilise their life is more important than checking their blood pressure.

In daily clinical practice, while we can urge patients to quit smoking, eat better and exercise, our efforts are rarely effective when societal conditions are set up to encourage poor life patterns. I deliberately do not use the word “choices” here, since that implies a conscious and deliberate choice by the individual. When many of the men around you are smokers, when everyone eats white rice or white bread, and when getting sweaty from exercise is regarded as socially unacceptable, it is difficult for people to make individual decisions against social norms, though the long term effects of these behaviour changes are more powerful than clinical interventions.

In terms of patient value for medical effort, we become more effective as we move towards tertiary and quaternary prevention. Even for hypertension, detection and treating in the secondary prevention setting requires over a hundred patient years of treatment to benefit one patient. For a patient who returns to the office after surviving a myocardial infarct, the chance of a recurrence within the next few years is high, and medical management can reduce that risk. This tertiary prevention pays off with much lesser numbers needed to treat for one to benefit. Quaternary prevention reduces the harms already being caused by medical interventions, such as excessive drug treatment. So in clinical practice, tertiary and quaternary prevention provide the greatest value.

The choice of preventive activities is complicated by the increasing recognition that such actions also cause harm. The spectrum of harms ranges from trivial, such as sore arms after immunisation, to really severe, such as septicaemia after prostate biopsy during follow-up of a raised Prostate Specific Antigen (PSA) test.

We invoke preventive actions for large numbers of people, in the hope that a few will benefit at a distant time, by not developing the disease, or at least not dying from that disease. But the harms often occur in the short term, and to a higher proportion of the participants than can ever gain from not getting the disease. For example, diabetes and antihypertensive drugs cause side effects for a proportion of people. Many find the costs of “preventive” medications are a problem too, especially when the effort of repeatedly returning for monitoring and review is factored in.

A cautious approach accounts for the difference between the Canadian Task Force on Preventive Health Care ² and many British recommendations, compared to many American recommendations, where only the benefits are sought, and the harms are largely ignored. ³ Even the United States Preventive Services Task force, which is the most cautious US body, still produces more positive recommendations than the Canadian group ⁴ , and these are overwhelming for clinicians.

The Canadian Task Force uses the GRADE categorisations: with four grades of recommendations. ⁵

1. Strong recommendation for an activity,
2. Conditional recommendation for,
3. Conditional recommendation against,
4. Strong recommendation against.

These are based on the quality of evidence available, and the strength of the effect observed. Surprisingly, the evidence is limited for many preventive activities, even those we in the medical profession have taken for granted over a long time. The evidence may show that the benefit is small or marginal. Unlike the US Task Force ⁶ , which adds a grade of “Insufficient evidence”, the four point system forces assignment of an action: as we need in clinical practice. If there is doubt about the value of an activity, the Canadian Task Force assigns a negative recommendation, on the grounds that it is more appropriate to use our energy on other activities that are proven valuable.

Many of the judgements about balance of harms and benefits are subjective, so individuals appropriately make different judgements about them, depending upon their personal values and how they weight the importance of different outcomes. This means that we as their physicians should not tell them what to do: we should inform them of the choices so they can make their own decisions. Note that here we should not call them “patients”, but people, because healthy people are not patients, and we should not needlessly change them to that state.

When there is a strong recommendation to do something, with strong evidence that it is worthwhile, we do not need to spend much time in discussion. Examples would be screening for hypertension in middle-aged people, or doing a cervical screening test in a woman aged 35 years who had been sexually active. For conditional recommendations in favour, discussion is recommended, to enable the person to choose the right action for their values. It is unreasonable to expect serious decisions that are potentially life changing to be made in the limited time of an office visit. It is better to inform people, give them materials to think about, and then return later, either after they have made a decision, or to inquire further before their decision.

To aid such discussions, “decision aids” are helpful. ⁷ These represent a new approach, still in development, that gives patients explicit information about the chance of benefit, against chance of harms usually in a numerical or probability form, to enable their choices. ⁸ A classic for m is the 1000 person diagram ⁹ , illustrating what would happen to 1000 similar people undertaking the procedure. Another form is the “Fact box” ¹⁰, or the “option grid” ¹¹. Many are surprised to see how small the potential benefit is, and how relatively common are the harms.

Being really critical about what prevention we do in practice sounds difficult, but it actually simplifies what we do. In our Calgary clinic, a spreadsheet that is posted in every examination room relating activities to the person’s age. ¹² This list is short, and even shorter by age group, so it is doable in practice. It starts with poverty screening, then behaviours, identifies family history, contraception and diet, urges dental care and immunisations. Mental health issues are listed, for awareness that mood disorders and safety in the home are important issues; though no formal screening has shown good effectiveness. Screening tests, which feature at the top of many prevention programs, are listed at the end, because the benefit for most people is small. We consider Chlamydia ur ine test for young people who are sexually active, testing for hepatitis among high-risk groups, and strongly recommend checking blood pressure from age 30, lipids and HbA1c from age 40, and ultrasound for Aortic Aneurysm for 60-year-old men who smoked. Cervical screening is given a strong grade, but mammograms are a conditional recommendation, for special-risk women from age 40 and standard-risk women from ages 50 to 75. ³ Lung cancer screening is a possible recommendation for 60-year-old men who smoked for at least 30 pack years. Colon screening may start at age 50, but is strongly recommended from age 60 to age 75.

The quality of the process is critical for any screening. Given the narrow balance between benefit and harm, if the test is done poorly, then the balance of benefit over harms may be lost. This applies whether we do the activity in our office (e.g. blood pressure), by a laboratory, or by a radiologist reading images. Consequently we must check our staff as they take measurements, ensure we refer only to quality laboratories, and ensure that our radiologists are participating in quality assurance. ¹³

The frequency of repetition matters. Harms arise from false positive results, and these are likely every time a test is repeated. Therefore, we need to leave an interval long enough that enough new disease has arisen for detection in the second test to be greater than the risk of harms. Thus sticking to the recommended screening intervals is important. For example many US physicians recommend annual mammograms, though the evidence suggests every two to three years provides benefit with less false positive s. Some disease will be missed by even the best test, and much will be missed by poorer tests. But the solution is not to test more often: we must ensure the quality, and recognise that in this imperfect world we cannot prevent everything.

So while everyone agrees that prevention is important, the most effective actions are outside our clinical role. Within our office, we must be highly selective. That is good: we can focus on the few valuable activities, and resist pressure to do many tests, since that may cause harm more than benefit.