Skip navigation
Link to Back issues listing | Back Issue Listing with content Index | Subject Index

What patients think

Much preventive prescribing is based on good evidence. Just like antiplatelets, for many interventions we have solid meta-analyses of good trials demonstrating clear statistical efficacy. We calculate NNTs or number of patients benefiting out of every 1,000 patients taking such-and-such medicine for three or five years. At a population level there is clear benefit.

What about the individual patient? What about their views concerning the amount of benefit that makes taking a medicine worthwhile for them? Even more, what about the balance of benefit and harm? Bandolier has long been aware that many people taking statins or anti-hypertensives believe that doing so prevents them having the heart attack or stroke they would have if they didn't take them. That's why they put up with, sometimes, quite awful adverse effects that have a big negative impact on the quality of their lives.

So a look at patient expectations [1] is a welcome breath of fresh air.


Subjects were randomly selected from three groups:

1 Group 1 had just been discharged from the coronary care unit (CCU) with a diagnosis of myocardial infarction.
2 Group 2 had no recent history of myocardial infarction but were taking preventive cardiovascular drugs.
3 Group 3 had no known cardiovascular disease and were on no preventive cardiovascular drugs.

Out of 550 subjects, 308 (56%) agreed to participate, approximately 100 in each group. They were given an explanation of the study and a questionnaire. Subjects were asked to imagine that their blood cholesterol was higher than normal, putting them at an increased risk of heart attack in the next five years. They were told that a new, safe, drug was available which would reduce this risk, but would not benefit everyone. Some would not benefit because they would not have a heart attack anyway; some would not benefit because the drug was not strong enough to prevent a heart attack in them.

Subjects were asked to mark a visual analogue chart expressing benefit in a semilogarithmic fashion.


There were some differences between the groups in demographics (Table 1), but about four out of five would want to know the chance of benefiting from treatment. Only a minority of patients would take a drug if they thought that they had a 5% chance or less of benefiting over five years (NNT about 20).

Table 1: Main results from three groups of patients - demographics and responses to questionnaires

Group 1

Group 2

Group 3


Just discharged from CCU

On preventive drugs but no MI

No preventive drugs, no MI

Mean age (years) 62 64 58
Percent men 76 53 45
Percent poor or very poor health 20 25 14
Smoker (%) 20 21 24
Against taking drugs 12 13 36
Want to know chance of benefiting (%) 79 72 84
Acceptable ARR reduction (median) 20% 20% 30%
Prolongation of life expected to be worth taking preventive medicine (month) 12 12 18
Percent prepared to take preventive drug with at least 5% chance of benefit over 5 years 32 29 21
Same, but if drug recommended by doctor 69 74 56

Half of the patients would take a drug if the chance of them benefiting over five years was 20% (NNT 5). Recommendation by their doctor increased the percentage of patients willing to take a preventive drug if the benefit was 5% or less over five years.


This is an interesting and imaginative paper that tells us what patients think. Half were happy to take a preventive drug if the hypothetical five year absolute risk reduction was 20%, or an NNT of 5. Table 2 shows some of the absolute risk reductions and NNTs from commonly used cardiovascular medicines. There is a clear discrepancy. Few preventive medicines for preventing heart attacks would seem to meet patient expectation.

Table 2: Absolute risk reductions and approximate NNTs for a number of common interventions



Number of patients


Untreated event rate (%)

Absolute risk reduction (%)


Pravastatin for secondary prevention 73 9014 Death 14.1 3.1 32
Pravastatin for primary prevention 59 6595 Coronary event 7.9 2.3 43
Ramipril in high risk patients 60 9297 Any MI 12.3 2.4 42
Antiplatelet post MI 27 20006 Vascular event 17 3.5 29
Hypertension DBP 80-109 mmHg 59 17354 Stroke 1.3 0.6 167
Warfarin in atrial fibrillation 22 571 Stroke 7.2 5.6 18

And what is really important here is that in this paper, patients were told that the hypothetical treatment was safe. Safe has many connotations, but we know that many preventive treatments carry a significant burden of adverse effects. It would be interesting to repeat the study giving subjects a slightly harder task, choosing acceptable levels of benefit against different descriptions of likely harm. It is probable that the gap between expectation of benefit and delivery would grow even wider.

The saving grace in this study was the power of the doctor to advise. If their doctor recommended it, more than twice as many subjects would take the medicine. This, though, imposes a significant burden on doctors properly to inform their patients. Much less attention has been paid to how patients think about their own versus population benefit, and especially how the information is presented. An area where there is scope for more research, perhaps.

These results for what patients thought about heart attacks seem to be different from what patients thought about strokes. There appears to be a different attitude to stroke, which most seem to consider to be an outcome that is much more important to avoid.

It is interesting to speculate idly about a league table of patient wants and expectations concerning preventive interventions, matching intervention with patients' acceptance.


  1. PN Trewby et al. Are preventative drugs preventive enough? A study of patients' expectation of benefits from preventive drugs. Clinical Medicine 2002 2: 527-533.
previous story in this issue