Factors involved in deciding to start preventive treatment: qualitative study of clinicians' and lay people's attitudesBMJ 2003; 327 doi: https://doi.org/10.1136/bmj.327.7419.841 (Published 09 October 2003) Cite this as: BMJ 2003;327:841
- David K Lewis (), GP fellow in public health1,
- Jude Robinson, deputy director2,
- Ewan Wilkinson, consultant in public health1
- 1 Department of Public Health, Central Liverpool Primary Care Trust, Hamilton House, Liverpool L3 6AL
- 2 Health and Community Care Research Unit, Thompson Yates Building, University of Liverpool, Liverpool L69 3GB
- Correspondence to: D K Lewis, Vauxhall Primary Health Care, Liverpool L5 8XR
- Accepted 28 July 2003
Objectives To explore the views of clinicians and lay people about the minimum benefit needed to justify drug treatment to prevent heart attacks, and to explore the rationale behind treatment decisions.
Design Qualitative study using semi-structured interviews.
Participants 4 general practitioners, 4 practice nurses, and 18 lay people.
Setting 8 general practices and 6 community settings across Liverpool.
Results Participants varied widely in the minimum acceptable benefits chosen. Most people found the concepts difficult initially, and few appreciated that increased length of treatment should increase absolute benefits. Lay people usually wanted to make decisions for themselves, and clinicians supported this. Participants wanted to consider adverse effects and costs of treatment. Dislike of drug taking was common, and many people preferred lifestyle change to an imperfect treatment. Quality of life and personal views were more important than an individual's age.
Conclusions Evidence based guidelines make assumptions about people's preferences, and, by using 10 year estimates of risk, inflate the apparent benefits of treatment. It is unlikely that guidelines could incorporate the wide range of people's preferences, and true dialogue is necessary between clinicians and patients before starting long term preventive treatment.
It is becoming accepted that patients' preferences and values should be explicitly incorporated into decisions about treatments, but the rationale behind decisions is poorly understood.1 2 Most studies have examined groups of patients who are taking drugs, so excluding the views of those who have declined treatment,3 4 and how general practitioners' beliefs might influence prescribing decisions has not been explored.2
The decision to treat an asymptomatic risk factor involves balancing expected benefits with expected harms.5 Individual values and priorities, rather than facts, will determine the choice made,6 7 and patients' values are at least as valid as professionals'.2 Lipid lowering and blood pressure lowering drugs reduce the risk of ischaemic cardiovascular events.8 9 The absolute benefit (proportion of people treated who benefit) is highest in people at greatest risk of having an event, but adverse effects and costs of treatment are similar whatever the baseline risk.
Little is known about the level of coronary risk at which either doctors or patients want treatment, yet current UK guidelines arbitrarily recommend treatment for those whose 10 year risk of coronary heart disease is at least 30%.8 9 Assuming a relative risk reduction of about a third, this equates to an absolute benefit of about 10% over 10 years (or 5% over five years), though some argue for treating those at lower risk.10 11
We explored the views of primary care clinicians and lay people about the minimum benefit they thought would justify drug treatment to prevent heart attacks. Doctors, nurses, and lay people are likely to have different perceptions, values, and levels of knowledge when making treatment decisions, and doctors tend to have more power within the clinical relationship.12 This can result in doctors unwittingly excluding patients' preferences from treatment decisions and pressing them to take treatments that they would decline if fully informed, or not offering treatments that they might choose. We used the same scenario to elicit responses from all groups and to learn about the factors influencing preferences for preventive treatment.
Participants and methods
We recruited four doctors from a randomised list of general practitioners in Liverpool, by contacting the first two male doctors and first two female doctors by telephone. As one male doctor declined, we approached the next man on the list. The practices of the next four general practitioners on the randomised list were contacted by telephone to invite their practice nurse to participate, and all four nurses agreed. No participants were known to the researcher (DKL), and all worked at different practices across Liverpool.
We chose to interview people outside a healthcare environment. We recruited 18 lay people from the following settings across Liverpool (number of people in brackets): two meetings of the British Cardiac Support Group (5), two lunch clubs run by Age Concern (6), a course run by a Tenants' Association (6), and a university department (2). A key contact from each organisation asked for volunteers and passed their contact details to the researcher. Lay participants were told that the interviewer (DKL) was a researcher from public health, so that respondents would feel able to comment freely; they were told that he was a general practitioner only if they asked. Sample size was pragmatic, but we recruited lay people until no new themes emerged.
Box 1:Scenario presented to participants and prompts used during interview
”I would like you to think about a drug, a tablet, which can prevent heart attacks. It is not perfect inthat it does not prevent them completely. What that means is that if, say, 100 or 1000 people take the tablets everyday for five years, then some will be saved from a heartattack—they would have had a heart attack without thetablet, but the tablet has prevented it. Some will have aheart attack anyway, even though they take the tablet.And the rest will not have a heart attack, but wouldnever have had one anyway even without the tablets. Sothe ones who are saved from a heart attack will have beenhelped by the drug. The rest would have been just thesame if they had not taken the tablets at all. The problem is no one can say exactly who will be better and who willnot, so they all had to take the tablets. [clarify] “WhatI want to know is—of, say, 100 similar people, all takingthe drug for five years, how many have to be saved from aheart attack for it to be worth while all of them takingthe drug every day?”
Prompts used in all interviews
“Why did you choose that number and not, say, one a bit higher [or lower]?”
“How about if everybody has to take the tablet for 10 years instead of five? Would that make any difference?”
“Do you think it makes any difference how old the people taking the tablets are? How would that change your answer?”
“Is there anything else about the people taking the tablets which might alter your answer?”
“Let's say that the tablets are mostly safe, but that you have to have a blood test at least every year and that the long term side effects are not known”
At the end of interview
“Thank you. Would you still choose the same number nowas you did at the beginning?”
[If someone said that one person benefiting out of 100would be worth while] “OK. Let's say that 1000 people have to take it every day for just one to be better. Wouldthat be worth it?” [increasing the number until they saidthat it would not be worth it]
Participants were given the opportunity to read an information sheet and to ask questions, and then invited to take part in a confidential, face to face interview at a time and place convenient to themselves. They were presented with a hypothetical drug to prevent heart attacks, and asked for their views on the minimum benefit needed to justify people taking the drug, and the reasons for their choices (see box 1). It was stressed that there were no right or wrong answers. Interviews lasted 15-30 minutes and were audiotaped and later transcribed verbatim for thematic analysis of content. One of us (DKL) read the transcripts, identified themes, and re-read every transcript to search for each theme as it was identified. A second author (JR) read all transcripts and validated these themes.13 Key findings were discussed by all authors.
The table describes the participants' background characteristics.
Understanding the scenario
Most participants, including the health professionals, found the concepts difficult to grasp at first (for example, stating they were confused) and gave different numerical answers through the interview, some of which were self contradictory. For example, one doctor changed the minimum absolute benefit from 20% to 5% without seeming to realise. All the doctors, half the nurses, and two of the 18 lay people seemed confident of their choice. Three clinicians initially said that a third should benefit from the drug because there were three categories of patients in the scenario. Some balanced an ideal against an acceptable benefit: one lay person (lay interview 10) wanted 99% of people to benefit but also thought that any benefit at all was worth while.
Minimum benefits chosen
Overall, doctors chose lower minimum absolute benefits (5%-10%) than nurses (10%-25%). Some lay people had high expectations of treatment, with minimum benefits varying from 10% to 99% (see box 2). One doctor, two nurses, and seven lay people felt that any benefit would justify a large number of people taking treatment, even up to a million treated for one to benefit, although many required the treatment to be guaranteed cost free and risk free.
Benefits over time
Even those who gave numerically consistent answers usually wanted the same benefit over 10 years as over five; just one doctor, one nurse, and one lay person stated that the benefit should be greater if the drug was taken for a longer time. Guidelines use 10 year estimates of risk and benefit, although they are based on treatment trials usually lasting five years.8 9 Absolute benefit roughly doubles over double the time, but almost nobody we interviewed appreciated this. Rather, 10 and five years were viewed as “a long time.”
Shared decision making (see box 2)
Most lay people wanted to make decisions for themselves, based on information provided by health professionals. One thought that most people aged over 60 years, like herself, would not want treatment but that “they should be given a chance” (lay interview 1). This view was supported by clinicians (three doctors and three nurses), who wanted to help patients make their own decisions, but one doctor acknowledged that professional values might influence patient choices: “It depends on the way you sell it” (general practitioner 1).
Only two lay people wanted health professionals to make the decision.
Box 2: Participants' views on shared decision making, side effects, effectiveness, and cost
Shared decision making
“You've just got to allow people to make an informeddecision and leave it up to them”—General practitioner 4
“I think it should be their choice”—Nurse 3
“I've got a particularly good doctor, very very gooddoctor, and I would put my whole trust in her…. If sheadvised it, I'd take it but not otherwise”—Lay interview 7
“I would like to rely on… whatever it is that approves these drugs… I would want to be very influencedby the experts”—Lay interview 17
Side effects and guaranteed effectiveness
“Yes, if you were sure you reduce your risk of gettinga heart attack by 1 in a million—you could be the one ina million. If there is no cost, financial or medical risk, then it's worth it”—Lay interview 17
“I'd take it. But if you'd say ‘It could have theseside effects,’ then I'd say no. Only if it was guaranteedand didn't do any harm, then it'd be OK”—Lay interview 15
“You should have a tablet that is going to cure 99% ofpeople before you start giving it out. With technologythe way it is, the idea of who should and shouldn't begoing to have heart attacks should be betterforecasted”—Lay interview 12
“You can't put a figure on it unless you are aware ofthe costs and side effects”—General practitioner 2
“At the final analysis, it becomes a financialdecision. You know, do we keep people alive indefinitelyuntil they are 150, and can the country, the world, afford to do that?”—Lay interview 16
“Why go on giving me these pills for a dubious qualityof life, when… the same money spent on, I don't know, on reducing teenage pregnancy could have an enormousimpact on 60 years of somebody's life? The best you cando for me is to give me an extra three or four years at85; well I don't rate that as highly as improving, say, child nutrition”—Lay interview 18
Side effects and guaranteed effectiveness (see box 2
All of the doctors, half of the nurses, and five of the lay people wanted to know about side effects before making a decision about the benefit required, so that they could balance benefits against possible harms. One lay person would not take a new treatment if the risk of side effects was unknown; some said they would stop taking tablets if they noticed any side effects, whereas others were happy to tolerate mild inconvenience if this were balanced with greater effectiveness. Six lay people wanted a perfect drug—either that there should be no possible adverse effects, or that it should definitely prevent all heart attacks. Some felt that only those patients certain to suffer a heart attack should take preventive treatment.
Cost (see box 2)
Cost of the treatment was mentioned spontaneously by all of the doctors, three of the nurses, and a third of the lay people; they felt it was inappropriate to choose the minimum acceptable benefit without considering costs. One person balanced the opportunity costs of treatment.
Pill taking (see box 3)
Seven lay people did not like the idea of taking medication—for example, “Tablets are a terrible thing” (lay interview 9)—but one person was not bothered by pill taking. Others were reluctant to take a drug without understanding its purpose. Three of the nurses and one doctor commented that patients or they themselves disliked medication, and half of the clinicians pointed out that there was no point prescribing drugs that would not be consumed.
Lifestyle (see box 3)
Three of the practice nurses found it frustrating to give drugs to people who were not prepared to adopt a healthier lifestyle, although they would not withhold treatment. Some lay people made similar comments, and two suggested that people who refused to alter their lifestyle (such as by stopping smoking) should not receive preventive treatment from the NHS. Many people we interviewed said they would prefer lifestyle change to an imperfect treatment.
One lay person described the detrimental effect on the family of labelling someone with an illness: “If it's your mother or father, we know they have got something that they are taking the tablet for, so it is making us wary and more watchful of them” (lay interview 12).
Age (see box 4)
Opinion was divided on the relevance of pill takers' age (prompt 3, box 1). One doctor, two nurses, and eight lay people said that age would not influence a person's choice to take preventive treatment, but their expected quality of life should always be taken into account. One lay person felt her own quality of life did not justify the treatment, but continued to take it on the advice of her doctor. Nine others felt that the benefit should be greater to justify treatment with increasing age. Clinicians' reasons for this included that a heart attack when young was more significant and that older people were likely to die of other causes before benefiting from treatment. Lay reasons included that, over a certain age, people would not want to be bothered with this type of thing (two included themselves) or that older people could not cope with side effects, as well as that older people would die before benefiting.
Box 3:Participants' views on pill taking and lifestyle
“The ethos of taking tablets I find difficult”—Nurse 2
“But then they don't enlighten you what they're for. When you start off with them.”
Researcher: “Do you think they should?”
“Yes I do. Yes—last time I went in hospital they said, ‘You've got to take these. There's four boxes there; takeone every morning.’ What they're for, God only knows.So… I stopped them”—Lay interview 9
“I am very sceptical about wonder pills. I mean, Ithink… heart attacks probably fall in that group ofthings which… preventative things other than takingpills are likely to be effective and cheaper—you know, like walking the dog, like running up the stairs, and noteating fat and not smoking…. People who don't want to have heart attacks, I think, know what they have to do…and I certainly wouldn't give the pill to anybody whosmoked, particularly if it was expensive—if they smoked, drank more than, say, a little, and took no exercise. I'dsay, ‘Well try those three first, and once you've gotthose going, you know, come back and we can give you thepill’”—Lay interview 18
“I just found it really strange… there's this manprepared to take a statin that could damage his liver, and he's still gaily having bacon butties, you know, fried in lard”—Nurse 2
BBox 4: Ageand quality of life
Two lay people, aged 78 and 88, thought that having towait years before seeing benefit from treatment was notworth it, but would consider treatment if there was achance of imminent benefit:
“Five years ahead that's, at our age, let's be honest, that's way out”—Lay interview 11
“Well, as you get older you are not able to look afteryourself, and there is always one left. I'm left; I've been 25 years a widow now. No, if the doctor took mytablets off me tomorrow that would not worry me. No, itwouldn't worry me… Well, you know, I take them. I willadmit I still take them.”
Researcher: “Why do you take them?”
“Well I don't know; I suppose it's habit, being onthem so long”—Lay interview 6RETURN TO TEXT
Many of the preferences expressed by the clinicians and lay people in this study are at odds with recommendations in guidelines. The doctors made decisions more confidently than others but were not necessarily more consistent or logical. All our respondents found it difficult to discuss benefits occurring in 10 years' time; by using 10 year estimates, guidelines may artificially inflate the benefit of treatment (halving the number need to treat), and skew the opinions of clinicians and patients. In addition, health professionals often accept smaller benefits (larger numbers needed to treat) from preventive treatment than those being treated expect, though a few patients opt for treatment even when there is no anticipated benefit.14–17
In the concordant model of prescribing, doctors present the scientific evidence and elicit patients' views; patients consider the information and voice their preferences.2 Most people we interviewed seemed to follow this model, which includes asking a healthcare professional to decide, but one effect may be that people choose not to take treatment.4 18 19 Many patients want to be involved in decision making, but not all.1 20 Such negotiated decision making is believed to lead to more realistic and ultimately sustainable treatment decisions.5 Unfortunately, in practice doctors may mistakenly think they know what patients' preferences are, communication may flow only from doctors to patients but not in reverse, and many prescribing decisions are not shared.19 21
Whether someone views taking a pill every day as affecting their quality of life greatly influences their choice whether to accept treatment.22 Labelling asymptomatic people, by telling them that they need treatment, may cause a reduced sense of wellbeing, increased absenteeism from work, and a feeling of loss of personal control over health.3 21 23 24 Our study confirms the finding that many people are averse to taking drugs unless absolutely necessary.21 Even patients taking preventive treatment may feel that drugs are best avoided, but on balance choose to take them, and some patients we interviewed were ambivalent about their own treatment.3 Despite, this a combination of drugs to prevent heart attacks, the polypill, has been advocated for everyone over age 55.25
What isalready known on this topic
Most people want to be involved in decisions abouttheir medical treatment, but many prescribing decisionsare not shared. Patients' preferences are often differentto those of clinicians, and may lead to choices contraryto guidelines
Little is known about the factors which clinicians orlay people use to decide whether to start preventivetreatment
Guidelines are based on arbitrary judgments of whatbenefit is worthwhile
What this study adds
Understanding the numerical basis for decisions aboutpreventive treatment is difficult for both clinicians andlay people. Participants in this study wished to considercosts before deciding when people should start treatment
Many people dislike the concept of taking pills everyday, and might choose lifestyle change rather thanmedication, if possible. Some people want only a perfecttreatment, with guaranteed effectiveness and no possible adverse effects
The diversity of beliefs makes consensus unlikely, andguidelines should reflect this. True dialogue is required if patients' preferences are to be used to make decisions
These findings represent the views of a small number of general practitioners, nurses, and lay people using a hypothetical scenario rather than actual treatment decisions. It is possible that an alternative scenario, or the same one presented differently, would have elicited different responses. Whether people wish to take such a drug, or wish resources to be used in this way, requires public debate, but our study suggests that some would prefer risk to be reduced by improved lifestyle rather than polypharmacy. In the current enthusiasm for guidelines, there is little about the need to listen to patients' views and to ensure they understand that they may take tablets for years with no benefit to themselves.
Both clinicians and lay people in this study found it difficult to make logical decisions about preventive treatment, but most wanted to be involved in determining their own treatment. Many people would prefer lifestyle change to medication. Despite the political reluctance to discuss rationing and prioritising, we found a general acceptance that resources are finite and should be targeted where they are most effective.
There is a danger that increased pressure on general practitioners to prescribe some drugs (either political, such as national service frameworks, or financial, such as the new GP contract) may distort practice and marginalise patients' preferences. People's values are not predictable,1 17 21 22 so probably the best way to ascertain their preferences is to ask them. The wide range of preferences makes consensus unlikely, and we believe guidelines should reflect the importance of true dialogue between clinicians and patients before embarking on lifelong preventive treatment.
We thank the participants, who gave up their time to be interviewed, and Fiona Johnstone who advised on recruitment.
Contributors DKL had the idea for the study. DKL and EW designed the study. DKL recruited and interviewed participants and analysed the transcripts for themes. JR reviewed the data and validated the analysis. All authors discussed key themes and concepts. DKL wrote the paper, and all authors revised it. DKL is guarantor for the study.
Funding DKL was funded by a North West Regional Health Authority Public Health Fellowship.
Competing interests None declared.
Ethical approval The study was approved by the local research ethics committee.