The nature of medical evidence and its inherent uncertainty for the
clinical consultation: qualitative study
Frances Griffiths, Eileen Green, Maria Tsouroufli
Objective To describe how clinicians deal with the uncertainty
inherent in medical evidence in clinical consultations.
Design Qualitative study.
Setting Clinical consultations related to hormone replacement
therapy, bone densitometry, and breast screening in seven
general practices and three secondary care clinics in the UK
Participants Women aged 45-64.
Results 45 of the 109 relevant consultations included sufficient
discussion for analysis. The consultations could be categorised
into three groups: focus on certainty for now and this test, with
slippage into general reassurance; a coherent account of the
medical evidence for risks and benefits, but blurring of the
uncertainty inherent in the evidence and giving an impression
of certainty; and acknowledging the inherent uncertainty of the
medical evidence and negotiating a provisional decision.
Conclusion Strategies health professionals use to cope with the
uncertainty inherent in medical evidence in clinical
consultations include the use of provisional decisions that allow
for changing priorities and circumstances over time, to avoid
slippage into general reassurance from a particular test result,
and to avoid the creation of a myth of certainty.
Clinicians have access to a growing body of good clinical
research evidence informing them about the effectiveness of
many medical interventions. However robust the research, clini-
cians face the dilemma of applying this evidence to individual
patients.1This is the uncertainty inherent in the nature of medi-
cal evidence. For example, epidemiology tells us that smoking is
a risk factor for heart attack, but it does not tell us which
individuals will be affected.2Randomised controlled trials of hor-
mone replacement therapy3report on the number of extra
breast cancers identified in a large number of women receiving
treatment compared with those not receiving treatment,but they
cannot tell us which women will develop the extra cancers. This
dilemma between the nature of medical evidence and individual
patient care is central to medicine’s history and will not
disappear, as they are essential to each other. Diseases always
manifest themselves in patients’ bodies and minds, and in
seeking to understand, treat, and predict the outcome of disease,
clinicians need to move their focus from the individual to more
Clinicians recognise this dilemma and have reflected on this
in relation to their clinical practice2and the need for research
methods that give more attention to the particular rather than to
the general.5The importance of this dilemma is discussed within
related disciplines, including medical philosophy, ethics, and
health policy.6–9Few studies, however, have examined what clini-
cians actually say to patients.10Studies have considered how
clinicians communicate clinical evidence to patients, taking
account of their preferences11and maintaining the clinician-
patient relationship.12Studies have also acknowledged the
difficulty of communicating about the risks and benefits of inter-
ventions.13These studies do not, however, examine communica-
tion in relation to the inherent uncertainty in the evidence. We
examined how health professionals talk to patients about this
uncertainty, and we provide a framework for reflecting on how
they handle the dilemma of applying clinical evidence to
We examined consultations with health professionals in both
primary and secondary care where there was discussion of one
or more of the interventions of hormone replacement therapy,
bone densitometry, or breast screening. Our study included
healthcare sites in contrasting socioeconomic contexts in the
Midlands and north east England. The collection of these data
was part of a larger study, reported elsewhere.14 15
All women aged 45-64 attending one of seven general prac-
tices or three specialist clinics in the UK NHS were invited to
participate in our study. After consent was obtained, the health-
care professional audiotaped the consultations. These were
reviewed for their relevance to our study. We discarded those
with no mention of the relevant interventions, and we retained
all the others regardless of the extent of the discussion of the
interventions. Table 1 lists the details of the clinics and surgeries
and consultations recorded. The details of the research process,
including analysis, are on bmj.com.
Overall, 109 consultations were relevant: 73 from general
practice and 36 from specialist clinics.Most women attending the
clinics agreed to be recorded, whereas in general practice the
consent rate was lower (20% in some practices).
A key emergent theme was uncertainty and how it is
discussed between health professionals and women, particularly
the uncertainty inherent in medical evidence when it is applied
to particular patients. The data included 64 consultations with
only a brief mention of the interventions. For example, a woman
discusses with the practice nurse those symptoms she thinks are
due to the menopause, and hormone replacement therapy is
Details of the research process are on bmj.com
Cite this article as: BMJ, doi:10.1136/bmj.38336.482720.8F (published 31 January 2005)
BMJ Online First bmj.com
page 1 of 7
mentioned only briefly (also see bmj.com). Owing to insufficient
data, we did not include these consultations in subsequent analy-
sis. Through a process of discussion and comparison of data, we
developed categories for how uncertainty was dealt with in the
remaining 45 consultations, which were recorded by 25 different
health professionals (nine had more than one consultation in
this dataset and of these, three had more than two). The catego-
ries were developed as a tool for understanding and reflecting on
what was taking place in the consultations. The results of the
analysis were presented to three university based focus groups—
two of doctors and one of patients—which provided feedback on
the validity of the categories from their own experience. In
further comparative analysis we explored links between how
uncertainty was dealt with and the healthcare issues and context.
The extract in box 1 provides an example of how uncertainty
owing to the nature of medical evidence was managed within the
consultations; the doctor knows what should make a difference
to bone density based on medical research,but he does not know
what has made a difference for this particular woman.
The three approaches to the uncertainty inherent in medical
evidence in the consultations were certainty for now, the coher-
ent story of certainty, and acknowledgment of the uncertainty.
Approaches to uncertainty inherent in medical evidence
Certainty for now
The health professionals talked about certainty for now, or for
this test—for example, the result of ultrasonography at the time
of the procedure. However, they also slipped into general
Coherent story of certainty
The health professionals wove a coherent account of the medical
evidence for risks and benefits—for example, a great deal of
detail, including estimates of the size of risk, was included in a
discussion of hormone replacement therapy for osteoporosis.
The way in which this detail was delivered, however, gave an
impression of certainty,even though the health professional may
have used words implying uncertainty.
The uncertainty of outcome from using an intervention was
acknowledged,including the inherent uncertainty of the medical
evidence when applied to individuals. A strategy used to cope
with this uncertainty was negotiating a provisional decision.
Most consultations included elements of each of the three
categories. In all but four consultations, however, a dominant
approach to uncertainty was identified. Of the nine health
professionals who had more than one consultation, all except
one (specialist registrar) used more than one approach to the
uncertainty inherent in medical evidence.
Certainty for now
Health professionals talked of certainty in relation to the results
of the test they had carried out or were planning. Reassurance
was given before the results were available, but with the proviso
that the results were needed to be absolutely sure. For example,
in two consultations women told their general practitioner about
changes in their breasts. The women were examined and
reassured that their breasts seemed “normal.” The women were
referred to the breast clinic for further certainty from tests (see
box 2, extract 1).
A doctor in the breast clinic (consultation 032) emphasised
the need for certainty by saying “obviously we need to know for
sure” and arranged a biopsy to try and achieve that. He followed
this by saying that “often we biopsy things to prove that they’re
nothing ... we get so many surprises, we’re sort of duty bound to
offer you the . . . chance of biopsy.” The type of certainty being
talked about is a test result for the here and now—a particular
Table 1 Number of consultations recorded between health professionals
and women at midlife in which hormone replacement therapy, bone
densitometry, or breast screening was mentioned
Setting, health professional
General practice 1:
General practice 2:
General practice 3:
General practice 4:
General practice 5:
General practice 6;
General practice 7:
Hormone replacement clinic:
No of health professionalsNo of consultations
Box 1: The uncertainty of medical evidence
A follow up bone density reading shows that the patient is
“holding her own”—that is, her bone density is not decreasing.
Patient: I’m still on the Didronel, should I continue with it, I, I
thought possibly that you might have said come off it now,
because I understood that my level was sort of normal for my age
Doctor: For your age, that’s correct.
Patient: Umm, so I wondered possibly if that’s why I was coming
to see you today. You’d maybe say I had to come off it, but if you
feel that I should continue with it I’m quite happy to do that.
Doctor: Umm, as long as there’s no problems with it.
Patient: If necessary, I don’t have any problems whatsoever.
Doctor: Umm, okay. My view would be take a belts and braces
approach. By that I mean you’ve changed your diet, you’re doing
more exercise, those two things are good for you. Err, taking the
Didronel we know now is allowed on a long term basis.
Doctor: Err and I am a little uncertain as to which of these three
strands, the diet, the exercise or the medication, is making the
difference, but something is. (Bone clinic, consultation 054)
page 2 of 7
BMJ Online First bmj.com
piece of tissue at this time. The mention of surprises indicates
uncertainty, but only until the results of the biopsy are known.
In the second extract in box 2, the doctor talks about
certainty provided by the ultrasound result for the breast tissue at
this time and then goes on to explain to the woman the limited
nature of this certainty. Other consultations in this category did
not include such explanation. The health professionals took care
to tell the women that the particular tissue examined was
normal, but followed this up with a reassuring phrase which was
rather general—for example,“it’s perfectly normal,you’re alright”
Coherent story of certainty
In some consultations, the health professional wove an account
or explanation for the woman that was coherent, almost as a
story. The intention seemed to be to provide information and
explanation so that the woman could make her own decisions,
although the overall tenor of the consultations was in favour of
the intervention. In some of the consultations a great deal of
detailed information was provided, including numerical esti-
mates of risk and explanations of uncertainty. From the way
women responded, however, it seems this formed an unfocused
backdrop for their decisions.
In box 3,extract 1,both the doctor and the woman seemed to
struggle with the uncertainty inherent in medical evidence. The
doctor actually contradicts himself in the process of trying to
provide a coherent account of the risk of osteoporosis. The
woman also struggles to understand how the evidence applies to
her. At one point the doctor links his explanation to the experi-
ence of the woman’s mother, a reality they both know about.
However, most of what the doctor says is drawn from evidence
based on populations (much of this detail has been removed for
brevity). The impression this creates is one of certainty about
how the evidence applies to this particular woman despite the
doctor using words and phrases that include uncertainty and
probability. The doctor creates a myth about the certainty of the
evidence for this woman.
Consultations in general practice tended to be shorter than
those in specialist clinics, with less detail given of the risk and
benefits. Some general practitioners expressed certainty about
the effect of hormone replacement therapy. For example, in dis-
cussing hormone replacement therapy (consultation 008) a
patient says “I don’t really want to come off it,if it’s not doing any
harm.” To which the doctor replies “Not, not any harm at all.” In
box 3, extract 2, a different general practitioner gives quite a lot
of information about the risks and benefits of hormone replace-
ment therapy and the different factors to be weighed up for dif-
ferent individuals. However, the tenor of the consultation is of
weaving a coherent account that indicates that it is possible for
each individual to work out what is best for them with some cer-
Approaches to acknowledging uncertainty
In box 4, extract 1, the woman is concerned about the new
evidence abouthormone replacement
concluded that the risks are small.The general practitioner backs
up the woman’s assessment of the risk and also explains the dif-
ficulty of applying population evidence to an individual:“It’s very
difficult to know whether if something happens to you whether
it’s this or more likely whether it would have happened anyway.”
It then becomes clear that for the woman having energy for her
“young lad”is important to her and given priority over the medi-
cal risk. A provisional plan is made whereby hormone
replacement therapy will be used for now but then reviewed. It is
through this provisional approach that the woman and doctor
have achieved some integration of future risk from the interven-
tion including the uncertainty inherent in the medical evidence,
with how things are for the woman in the current time and place.
In another consultation (box 4, extract 2) there is agreement
of a provisional plan for a reduction in the dosage of hormone
replacement therapy, a suggestion that came from the woman.
This plan integrates the concern about future risk from the
therapy with the woman’s experience of symptoms, so linking
across the gap between the medical evidence and the woman’s
In a consultation with a practice nurse (box 4, extract 3) the
risks of hormone replacement therapy are discussed and the
woman describes feeling well. The nurse explains the risk of
breast cancer, weaving a coherent story of the risks and benefits.
The woman introduces the idea of a provisional decision “by
then I might be okay we’ll just have to wait and see.” They agree
on continuing the therapy for now, aware of the potential risk
and of the good quality of life for the woman.
In another consultation (005) the doctor tells a woman who
has been receiving hormone replacement therapy for six years
for relief of symptoms, has a family history of breast cancer, and
has annual mammography,that her risk of breast cancer is going
up: it is about “weighing the two up,” “it becomes personal
choice.” The woman says “Will anybody sort of say ‘hey’ at a cer-
tain point? Or will that be up to me?” The doctor says “I think
what you’ll find is that there’ll be conversations like this once in a
while,” indicating that the decision is a provisional one.
Use of the different approaches
Analysis of the consultations by role of the health professional
and type of healthcare setting indicates a link between the
approach used for the uncertainty inherent in medical evidence
and the healthcare site (table 2).Certainty “for now”was found in
the breast clinic. Weaving a coherent story of certainty predomi-
nated in the hormone replacement therapy clinic and bone
clinic. General practice used all three approaches. The pattern of
approach became clearer when explored in relation to the health
concern discussed in the consultations (table 3). In all
consultations where there was concern about a breast problem,
health professionals used the approach of certainty for now with
slippage into general reassurance. Where the result of bone den-
sitometry and subsequent management was discussed, which in
some consultations included use of hormone replacement
Box 2: Certainty for now and this test, with slippage into
Extract 1: Woman mentions changes in her breasts
Patient: I just kept putting it to the back of my mind and then it
was just, I thought well its not, it doesn’t feel right you know it
was like pulling and I thought hmmm.
Doctor: I’ll sort you out a review at the breast clinic and then
they’ll be able to reassure you fully I’m, I’m sure . . . (General
practice, consultation 094)
Extract 2: Woman has ultrasonography of her breasts
Doctor: Here it is looking very clear that it is an innocent kind of,
er, thing. That’s why we don’t need to do any biopsy.
Doctor: The thing is, it doesn’t exclude you to getting something
else some other place . . . that’s the thing. I can tell about
what—what is happening today, and about these ones, which look
innocent. (Breast clinic, consultation 003)
BMJ Online First bmj.com
page 3 of 7
therapy, most of the consultations used a coherent story of
certainty.In the one consultation on this health issue that did not
use this approach, further test results were awaited. A coherent
story of certainty was also used for consultations where hormone
replacement therapy was initiated for other reasons. The health
issues were discussed in specialist clinics and in general practice
and by both doctors and nurses.
When reviewing the use of hormone replacement therapy or
restarting therapy after a break, acknowledging uncertainty pre-
dominated.Some health professionals,however,wove a coherent
story of certainty (see table 3). The consultations on this health
issue were all recorded in general practice. No pattern was
apparent linking the category of the consultation and whether
the review was initiated by the woman or by the health
To achieve good communication between health professionals
and patients,health professionals need strategies for coping with
the dilemma of applying medical evidence to individual patients.
These strategies could include using provisional decisions that
allow for changing priorities and circumstances over time,avoid-
ing slippage into general reassurance from a particular test
result, and avoiding the creation of a myth of certainty.
Box 3: Weaving a coherent account of the medical evidence for risks and benefits, but with blurring of the uncertainty
inherent in the evidence and an impression of certainty
Extract 1: Consultation after bone densitometry
Doctor: Your bone mineral density is following the course you would, we would expect.
Doctor: It is going down, you would expect that at this point in the menopause.
Patient: So it’s not abnormal then or anything?
Doctor: It’s not abnormal.
(The woman’s mother has osteoporosis. The doctor explains:)
A woman with a close female relative has 30% chance of having osteoporosis just ’cos you know they’re related . . .
(The doctor then suggests she considers taking calcium and vitamin D and taking hormone replacement therapy. The woman says “I’ve
never really been very keen on HRT.” The doctor then examines her and continues:)
With the constant, bone loss starts just round the very beginning that the hormones start to change, what we call the perimenopause and
then you’re likely to lose bone well totally predictably to lose bone for about 10 years after the menopause so it will start to gradually come
down. At the moment the results are normal, you have normal bone mineral density but err after about 10 years it’s going to drop into the
below normal range, you can’t be certain, but it’s predictable, err, and it’s obviously what’s happened to your mum . . .
(A further detailed explanation followed of the role of hormone replacement therapy, its benefits and risks, including numerical
expressions of risk, with the woman saying little until the doctor says:)
Effectively the choice is yours.
Doctor: Err, it doesn’t suit everybody, really the only way to know if it’s going to suit you is to try for a time.
Patient: Mmmh, do you really think that I need to be on it then?
Patient: Do you think that if I don’t go on it I’m going to end up more with osteoporosis.
Doctor: I think you’ll continue, you will continue to lose bone, it’s quite a difficult decision to take because you’re decision now, really you’re
trying to take a decision now to improve your health when you’re in your 70s and 80s with osteoporosis.
(The doctor explains further. The consultation ends with the doctor saying:)
Anyway the choice is yours.
Patient: All right thank you for your time. (Bone clinic, consultation 001)
Extract 2: Woman mentions tiredness
(The doctor inquires about menopausal symptoms and after some discussion the woman asks:)
With HRT, can’t you only go on that for so long, and then they take you off? Am I wrong?
Doctor: What happens with HRT is . . .
(The woman laughs)
Doctor: Right HRT . . . whilst you’ve got your own hormone you don’t need HRT, so your bones are being protected by your natural
hormone. Um, and HRT you get benefit from for your bones, for your heart point of view, from lots of different points of view. Now the
longer you’re on HRT from the bones point of view, the better. The problem is the longer you’re on it from your breast point of view,
people worry about the increase in breast cancer.
Doctor: And so what they try . . . it’s a balance of risks. So you take everybody individually. So somebody who has, a, a, wor . . . a concern
about breasts, maybe family history of breast cancer or something like that, you may be a bit more cautious on that side, but if somebody’s
got a dreadful history of thinning of the bones, and osteoporosis you sort of have to weigh that up, don’t you. So you’d say “oh well perhaps
you . . .” you know. So everybody’s individual, you weigh it up individually. The basic thing is that if you’re on HRT for, say, 10 years, say,
there is definitely an increase in risk of breast cancer. At five years, less so. Seven-and-a-half, it . . . what . . . up to five years is thought to be
fairly safe. So what . . . that, that’s where this business about “you can only be on it a certain length of time.”
Doctor: I’ve actually got ladies that have been on it 15 years. And are very very happy with it. I mean they wouldn’t stop it because it makes .
. . it keeps them well.
Doctor: So you, what you do is you balance up that good you’re getting from it, with the downside.
(The consultation continues, returning to consideration of the woman’s tiredness. Hormone replacement therapy is not prescribed, but the
woman is asked to think about it as a possibility for the future.) (General practice, consultation 025)
page 4 of 7
BMJ Online First bmj.com
We studied how health professionals and women have been
dealing with the dilemma of uncertainty inherent in medical evi-
dence in relation to medical interventions focused on women at
midlife. These interventions offer prevention, screening, and
relief of symptoms, so the results may inform other areas of
medicine where the type of evidence base is similar, such as pre-
vention and treatment of chronic diseases. Further research may
be needed to examine consultations about acute illness. The
recorded consultations include examples where the doctor was
attempting to communicate risk in ways that are known to be
unhelpful to patients,16particularly when weaving a coherent
story of certainty. Training in clinical communication, including
Box 4: Acknowledging the inherent uncertainty of the medical evidence and negotiating a provisional decision
Extract 1: Woman is concerned about taking hormone replacement therapy
Patient: I’ve been having ‘em, HRT patches and in the middle of the year there was a new finding.
Doctor: Right, the scare.
Patient: Right, so when they’ve finished I thought, I’d try to do without them.
Patient: And I’ve been considering it and considering it—what I want to know is do you think—what’s your opinion on it—when we talked
about—when we talked about it earlier we weighed up all the pros and cons.
Doctor: Yes. Yes.
Patient: Is there a history of cancer, is there a history of heart problems—no history of cancer—but a history of heart problems so we
decided it offered some a sort of protection to—but it seems to have taken a change—and then when I sort of thought about it later the
percentage is quite small really isn’t it.
Patient: When we, sort out how many people we’re talking about it isn’t large so I think that, I think that I’ll go ahead with some more. Is
that what, is that what you would advise, do you think it isn’t—it isn’t a big risk.
Doctor: No. It’s certainly not a big risk—how long were you been on HRT for?
Patient: Oh not long—less than a year.
Doctor: OK, that’s important because there’s also risks associated with time that you’re on HRT, so basically the longer you’re on, the risk
goes up, particularly if you’re looking at breast cancer, but having said that you’re absolutely right, the risk is still very small so any risk that
there is only affects a very tiny minority of women and of course it’s very difficult to know whether if something happens to you whether it’s
this or more likely whether it would have happened anyway.
Patient: And I was thinking of the quality of my life as well—my young lad I really need a bit more energy.
Doctor: Well that’s important too (laughing).
(The discussion continues and blood pressure is checked. Towards the end of the consultation the doctor says:)
So I’ll just give you some more now—and then what we do . . . if you’re happy with them you can either come and see one of us or see
(practice nurse) in six months for the next lot. (General practice, consultation 072)
Extract 2: Consultation to review hormone replacement therapy
Patient: Err my Estraderm patches, I’m getting a new prescription today, now the last time I saw the nurse, she said this would be my last
prescription and I wouldn’t be able to have any more.
Doctor: Did she mean because . . .
Patient: Because of my age or something—and I thought well I’ll come and see you, because I did funnily enough try to come off patches
myself, and I still got very flushed, so I thought I better just pop in and see you while I’m here anyway.
Doctor: Yes, I mean you’re 62 and therefore, sort of 10 years beyond a natural menopause but you had a pretty dramatic
menopause—you’ve had your ovaries taken out.
Patient: Oh I’ve had all sorts.
Doctor: I guess, she may have been thinking in terms of osteoporosis prevention, in that 10 years would be adequate for that and also as
you also will know, a longer term use of HRT is associated with breast cancer, however, if you feel that you’d rather carry on, bearing in
mind you know the increased risk of breast cancer.
Doctor: You know the big one, then I don’t have any particular problem with this.
Patient: What about after this six months I mean obviously it’s—would it—if I only say tried one a week instead of two how would that—or
don’t you do that with HRT.
Doctor: Well, or else what you could well. I’m just looking to see if they come in 25s—if you put one a week on, you’d be fine for the first half
of the week and then . . .
Patient: Sure enough.
Doctor: Yes, they come in 25s so one option might be to draw three months of the 25s to see how you get on.
Patient: Yes, yes.
Doctor: You might find that when you decide to stop you have no hot flushes or you know whatever you got when you last decided to stop.
But I think she probably just felt that that she would flag it up about breast cancer. (General practice, consultation 002)
Extract 3: Consultation with practice nurse
The woman and nurse have discussed the increase in breast cancer risk from taking hormone replacement therapy long term as shown by
the US study reported in the media. The woman is feeling well while receiving hormone replacement therapy.
Nurse: But there is still a risk of breast cancer—but there again there is a risk of breast cancer in this age group anyway, but it is increased
with long term use of . . .
Patient: Well when you say long term use of . . .
Nurse: Long term—10 years plus.
Patient: Oh, I’m getting up to that one now aren’t I—8 years isn’t it?
Nurse: Yes, that’s right—they advise five years, fine, up to 10 years is okay and then to rethink about it.
Patient: Well I mean by then I might be okay we’ll just have to wait and see.
Nurse: That’s right—blood pressure’s fine—but it is something that you’ve got to be aware of.
Patient: Oh yes, I realise that—yes. (General practice, consultation 083)
BMJ Online First bmj.com
page 5 of 7
how to communicate risk, is important. Many successful models
exist for such training. Our research does not suggest a new
model, but highlights the importance of including in existing
models an awareness of the dilemma involved in applying medi-
cal evidence to individual patients and strategies to cope with
The health professionals expressed an understanding of the
evidence about the risks and benefits of the interventions more
or less in line with the prevailing medical consensus at the time.
During data collection, however, new evidence on the risks of
hormone replacement therapy was published,3so the content of
some of the consultations would be different with less positive
accounts of hormone replacement therapy.17However, it is the
way the accounts of the medical evidence were interwoven that
produces the impression of certainty rather than the detail.
The data reveal a danger of creating a myth of certainty
around what is inherently uncertain through the way the medical
evidence is presented and discussed. This seems to be
particularly so when there is a test result, such as for bone densi-
tometry,or where an intervention such as hormone replacement
therapy is being initiated. This way of presenting evidence about
a medical intervention reinforces the idea of medicine as a pre-
cise science independent of context and people with the ability
to predict outcome, which has become incorporated into lay
models of illness.18Apparent certainty can be persuasive and can
lead to health professionals changing their understanding of the
evidence to fit the story they are presenting to the patient.Part of
learning to communicate well about risks and benefits of health
interventions, and so truly to include patients in decision
making, may be to fully recognise the uncertainties inherent in
clinical evidence and not to hide this from patients. Health pro-
fessionals would then stop reinforcing the myth of medicine as a
science of certainty and prediction and could work creatively
with its uncertainties alongside patients.
In consultations where hormone replacement therapy was
being reviewed or restarted, a provisional decision was often
agreed. This avoided the danger of further reinforcing the myth
of certainty. The women interpreted the medical evidence for
their current situation,19including their physical symptoms,
hopes and fears, social situation, and priorities.14They may have
been more able to do this at a review appointment as by then
they had some experience of hormone replacement therapy.
They may also have sought information themselves about the
medical evidence, and through this process developed their abil-
ity to assess the evidence.20
Time is an important dimension in this analysis. The
clinicians in the breast clinic struggled to stay with the here and
now in their desire to reassure the women. Consultations at the
bone clinic and hormone replacement therapy clinic included
mention of review of treatment in three, four, or five years. Men-
tion of this time added to the impression of certainty rather than
implying something provisional. The use of time, by making
provisional plans, was the striking feature of the category of
acknowledging the uncertainty. This fits in with reality for
women, as their context, experience, and level of risk changes
over time. The consultations in this category may provide useful
examples of using time in health related decisions for use in the
teaching of communication skills,as they show how a conditional
decision can be reached and be a satisfactory outcome for a con-
Reassurance is appropriate where there are high levels of
anxiety, such as in breast clinics (see box 2, extract 1); however, it
is also possible to be clear about the temporary and tissue
specific nature of the test result. Patients may seek certainty from
Table 2 Categories of approaches to uncertainty inherent in nature of medical evidence by role of health professional and type of healthcare setting
Healthcare setting and professional
Hormone replacement therapy clinic:
Focus on certainty for now and this
test, with slippage into general
Weaving coherent account of medical
evidence for risks and benefits, but
with blurring of uncertainty inherent in
evidence and impression of certainty
Acknowledging uncertainty of outcome from
using intervention including inherent
uncertainty of medical evidence, and coping
with this uncertainty through negotiated
provisional decisionNot categorised
Table 3 Categories of approaches to uncertainty inherent in nature of medical evidence by health issue
Concern about breast lump or positive screening result
Bone densitometry result and subsequent management
Starting hormone replacement therapy
Review of hormone replacement therapy or restarting after break
Requesting information or referral for screening (mammography
or bone densitometry)
Focus on certainty for
now and this test, with
slippage into general
Weaving coherent account of medical
evidence for risks and benefits, but with
blurring of uncertainty inherent in evidence
and impression of certainty
Acknowledging uncertainty of outcome from using
intervention including inherent uncertainty of
medical evidence, and coping with this uncertainty
through a negotiated provisional decision
Table excludes four consultations that were not categorised.
page 6 of 7
BMJ Online First bmj.com
health professionals because they feel vulnerable at that time or
because they believe the myth of medical certainty. Health
professionals are in a position of influence with patients, so in
responding to a desire for certainty they should critically reflect
on the effect this may have on their patient now and in the future,
such as building an expectation of certainty of outcome from
medical interventions. The assessment of how much to
emphasise certainty or not for each patient should be explicit in
the training of medical communication skills.
In the consultations where a provisional decision was made,
negotiation was present between the health professional and
woman. How much it was guided by the woman and how much
by the health professional varied (see box 4).Data from the study
shows that women vary in their preference for involvement in
decision making with health professionals, and that this varies
according to their circumstances.15It is the provisional nature of
the decision, rather than the woman’s involvement in the
decision, that seems to allow the decision to sit comfortably with
acknowledging the uncertainty inherent in medical evidence.
In general practice in the United Kingdom, it is possible to
make provisional decisions with patients and to review them. It
provides continuity of care for individuals,21of which this
decision making process is one aspect. In contrast, specialists
may see patients only once or review their treatment only at
infrequent intervals making it more difficult to negotiate
provisional decisions.The challenge for health professionals is to
develop the skills to acknowledge uncertainty and to negotiate
provisional decisions, including when considering test results or
starting new interventions.
The major types of evidence used in clinical medicine cannot
be directly applied to an individual, so health professionals will
continue to face the dilemma this creates. Through the teaching
of training in communication skills and the design of healthcare
systems it is important to enable health professionals to make
provisional decisions with individual patients. This approach to
decision making has the most potential for a continuing
acknowledgment of the inherent uncertainty in medical
evidence, an uncertainty which will remain even with progress in
basing medical interventions on robust research evidence.
We thank the participants for their time, the Leicester Warwick Medical
School GP Lecturer Group, the University of Warwick Academic GP Regis-
trar Group, the University of Warwick Primary Care Research User Group
for their contribution to the study, and the reviewers for their suggestions.
Contributors: FG was principal investigator for the study. EG was
coapplicant on the study funding proposal, managed a study field site, and
contributed to each stage of the study development, process, analysis, and
coapplicants on the study funding proposal, contributed to the design of
the study, advised on the conduct of the study, and contributed to analysis
and reporting. Di Thompson and MT undertook the data collection and
contributedto analysisand reporting.
Lindenmeyer contributed to analysis.
Funding: Economicand Social
(L218252038); part of the innovative health technology programme.
Competing interests: None declared.
Ethical approval: Warwickshire local research ethics committee and Hartle-
pool and North Tees local research ethics committee.
Kathryn Backett Milburn were
Pamela Lowe andAntje
1Fox RC. Medical uncertainty revisited. In: Bendelow G, Carpenter M, Vautier C,
Williams S, eds. Gender, health and healing: the public/private divide. London: Routledge,
Willis J. The paradox of progress. Oxford: Radcliffe Medical Press, 1995.
Writing Group for the Women’s Health Initiative Investigators. Risks and benefits of
estrogen plus progestin in healthy postmenopausal women. JAMA 2002;288:321-33.
Rosenberger NR. The process of discourse: usages of a Japanese medical term. Soc Sci
McWhinney IR. An acquaintance with particulars . . . Fam Med 1989;21:296-8.
Gorovitz S, MacIntyre A. Toward a theory of medical fallibility. J Med Philos 1976;1:51-
Beresford E. Uncertainty and the shaping of medical decisions. Hastings Cent Rep
Tanenbaum S. What physicians know. N Engl J Med 1993;329:1268-71.
Tanenbaum S. Evidence and expertise: the challenge of the outcomes movement to
medical professionalism. Acad Med 1999;74:757-63.
10 Tanenbaum SJ. Knowing and acting in medical practice: the epistemological politics of
outcomes research. J Health Polit,Policy Law 1994;19:27-44.
11 Haynes RB, Devereaux PJ, Guyatt GH. Physicians’ and patients’ choices in evidence
based practice. BMJ 2002;324:1350.
12 Summerskill WS,Pope C.“I saw the panic rise in her eyes and evidence-based medicine
went out of the door.” An exploratory qualitative study of the barriers to secondary
prevention in the management of coronary heart disease. Fam Pract 2002;19:605-8.
13 Lewis D,Robinson J,Wilkinson E.Factors involved in deciding to start preventive treat-
ment: qualitative study of clinicians’ and lay people’s attitudes. BMJ 2003;327:841-7.
14 Green E, Thompson D, Griffiths FE. Narratives of risk: women at midlife, medical
“experts” and health technologies. Health Risk Soc 2002;4:243-86.
15 Griffiths FE, Green EE, Bendelow G, Backett-Milburn K. Innovative health technologies at
women’s midlife; theory and diversity among women and experts. Swindon: Economic and
Social Research Council, 2003.
16 Edwards A, Elwyn G, Mulley A. Explaining risks: turning numerical data into meaning-
ful pictures. BMJ 2002;324:827-30.
17 Minelli CAK, Sutton A, Cooper N. Benefits and harms associated with hormone
replacement therapy: clinical decision analysis. BMJ 2004;328.
18 Griffiths F, Green E. A normal biological process? Brittle bones, HRT and the patient-
doctor encounter. In: Williams SJ, Birke L, Bendelow GA, eds. Debating biology:sociologi-
cal reflections on health medicine and society. London: Routledge, 2003:210-22.
19 Massé R, Légaré F. The limitations of a negotiation model for perimenopausal women.
Soc Health Illness 2001;23:44-64.
20 Bond M, Bywaters P. Working it out for ourselves: women learning about hormone
replacement therapy. Womens Stud Int Forum 1998;21:65-76.
21 WONCA Europe. The European definition of general practice/family medicine, 2002.
www.euract.org/pap041.html (accessed 12 Jul 2004).
(Accepted 9 December 2004)
Centre for Primary Health Care Studies, University of Warwick, Coventry CV4
Frances Griffiths senior clinical lecturer
Centre for Social and Policy Research, University of Teesside, Middlesbrough TS1
Eileen Green professor
Institute for Society, Health and Ethics, University of Cardiff, Cardiff CF10 3AT
Maria Tsouroufli research fellow
Correspondence to: F Griffiths firstname.lastname@example.org
What is already known on this topic
Uncertainty about outcome for an individual patient is
intrinsic to the nature of medical evidence
This creates a dilemma that will always be present
Communicating evidence to patients is a key part of clinical
consultations, with a growing evidence base of how it is best
What this study adds
A dilemma for health professionals is creating a myth of
certainty around what is inherently uncertain
This may be avoided by negotiating provisional decisions
BMJ Online First bmj.com
page 7 of 7