Family Practice Vol. 18, No. 3, 309-313
© Oxford University Press 2001
Shared decision making in hypertension: the impact of patient preferences on treatment choice
Division of Primary Health Care, University of Bristol, Canynge Hall, Whiteladies Road, Bristol BS8 2PR and
a Bristol Royal Infirmary, Marlborough Street, Bristol BS2 8HW, UK.
Montgomery AA, Harding J and Fahey T. Shared decision making in hypertension: the impact of patient preferences on treatment choice. Family Practice 2001; 18: 309–313.
Received 22 May 2000; Revised 2 October 2000; Accepted 8 January 2001.
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
Background.Recent guidelines for treatment of hypertension advocate a multifactorial approach based on absolute risk of a cardiovascular event. However, this does not take any account of individual patient values or preferences for health outcomes that result from having hypertension.
Objective.Our aim was to investigate the impact of patient preferences on treatment recommendations for hypertension using individual decision analysis.
Methods.We carried out an observational study based on interviews with 52 hypertensive patients. Patient preferences were measured using the standard gamble method. Associations between outcome of the individual decision analyses (recommendation to accept or decline antihypertensive medication) and treatment guidelines based on blood pressure and absolute cardiovascular risk were investigated. Adherence to medication during the 6 months following the interview was also assessed.
Results.Individual patient preferences have a substantial impact on the proportion of patients for whom drug treatment would be recommended. In 52 patients interviewed, decision analysis indicated that 29 [56%, 95% confidence interval (CI) 41–70] should be treated, compared with 27 (52%, 38–66) using a cardiovascular risk of 
10% over 5 years and 19 (37%, 24–51) using a systolic blood pressure of 160 mmHg. There was marked disagreement between the decision analysis and these recommendations (kappas 0.18 or less). There was no relationship between outcome of the decision analysis and adherence to medication [chi-square (1 d.f.) = 0.5, P = 0.5].
Conclusions.Quantifying patients' preferences and using decision analysis as a shared decision-making aid appears to have an impact on whether patients would be recommended for antihypertensive medication. Further evaluation of this method as a shared decision-making tool is warranted.
Keywords. Decision analysis, hypertension, treatment.
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Risk stratification based on absolute risk of a cardiovascular event is recommended for aiding treatment decisions in both high blood pressure and hyperlipi-daemia.1 However, basing treatment recommendations on absolute risk thresholds has been criticized as arbitrary.2 Furthermore, these recommendations do not take account of the values which individual patients place on the consequences of treatment.
A possible solution is to elicit and incorporate patient values or preferences for different health states which result from accepting or declining drug treatment. Individuals' values can then be combined with probability of experiencing these health states in a decision analysis.3 Decision analysis as an aid to shared decision-making has been shown to influence vaccination up-take,4 but its potential for aiding decision making in hypertension is unknown. Research suggests that treatment outcome, including medication adherence and achievement of blood pressure goal, may be improved if patients perceive greater control over their treatment choice.5
There were two aims of this study: (i) to evaluate the impact of incorporating patient values in a decision analysis and compare the proportion of subjects who would be candidates for drug treatment when systolic blood pressure or absolute cardiovascular risk are used as the criteria; and (ii) to examine the relationship between the outcome of individual decision analysis and subsequent adherence to blood pressure-lowering medication.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Subjects
A random sample of 100 subjects was selected from the patient list of a single general practice in Bristol. All subjects had a diagnosis of hypertension recorded in their computer notes and had been prescribed antihypertensive medication in the previous 12 months. No restrictions were placed on age. Subjects were sent an invitation to participate, and those who did not reply were telephoned until either agreeing or declining to take part in the study. Ethical approval for the study was obtained from the local research ethics committee.
Decision analysis
The treatment alternatives and their potential consequences were mapped by means of a decision tree (Fig. 1
).
|
Patient preference interviews were carried out by either AM or JH. The terms ‘preference’, ‘value’ and ‘utility’ are often used interchangeably, and in this study are taken to mean a numerical value between 0 (equal to death) and 1 (equal to perfect health) that indicates the individual's preference for an intermediate health state. Utility assessment was carried out as follows: the health states (outcomes) from the decision tree were described in bullet-point format on laminated cards, and the patient was asked to rank them in order of preference. Utilities for each health state were then elicited using the standard gamble method.3 The standard gamble method of utility assessment requires the subject to make a hypothetical choice between the certainty of a lifetime in the intermediate health state of interest or a gamble with the outcomes of perfect health (probability = P) and death (probability = 1 – P). P is varied until the subject is unable to choose between the two, at which point P is equal to the utility value for the health state in question.
Individual patient data (gender, age, systolic blood pressure, smoking, total and HDL cholesterol, diabetes and left ventricular hypertrophy) were collected to allow calculation of the 5-year absolute risk of a fatal or non-fatal cardiovascular event (CVE, defined as newly diagnosed angina, myocardial infarction, coronary heart disease, stroke or transient ischaemic attack).6 Missing total and HDL cholesterol were assigned values of 5.7 and 1.3 mmol/l for men, and 5.8 and 1.5 mmol/l for women, respectively.7 The following probabilities were obtained from the literature: relative risk reduction of CVE with antihypertensive drug treatment = 0.33;8 risk of treatment side effects = 0.1;9 risk of death after CVE = 0.3;10,11 and risk of being affected if surviving the CVE = 0.3.10 The probabilities and utilities were assigned to each individual's decision tree. These were then multiplied and summed to give the expected utility values for the two main branches of the tree (treatment/ no treatment).3 Following this, if the expected utility of ‘treatment’ exceeds that of ‘no treatment’, the recommendation for the patient would be to accept treatment. However, neither the patient nor the GP were informed of the outcome of the decision analysis as the purpose of the study was to examine the effect of incorporating patient preferences for drug treatment on established treatment recommendations based on blood pressure level and absolute cardiovascular risk.
Data analysis
Outcome of the decision analysis was compared with treatment recommendations based on systolic blood pressure 160 mmHg12 and 5-year absolute risk of a fatal or non-fatal CVE 10%.13 Comparisons were performed using both crude and chance-corrected (kappa) agreement between decision analysis and the two other treatment recommendations. Adherence with medication during the 6 months following the utility assessment interview was calculated from repeat prescribing records and was defined as taking at least 80% of prescribed medication.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Sixty-four patients attended, of whom 52 were able to complete the utility assessment successfully. Patients unable to complete the utility assessment were older (mean age 64 years versus 57 years, 95% CI for difference –0.3 to 14 years, P = 0.06). Characteristics of these 52 subjects are shown in Table 1
.
|
According to the decision analysis, 29 of the 52 patients would prefer treatment with antihypertensive medication (56%, 95% CI 41–70%). The corresponding figures were 28 (54%, 39–68%) for 5-year absolute cardiovascular risk
10% and 19 (37%, 24–51%) for systolic blood pressure 160 mmHg.
Comparisons between decision analysis and treatment recommendations based on absolute cardiovascular risk and systolic blood pressure are shown in Table 2.
|
Of the 28 patients whose 5-year absolute risk is
10%, 14 would decline treatment if their preferences were taken into account. In contrast, 15 out of 24 patients below this threshold would accept treatment if their preferences were taken into account. Of 19 patients whose systolic blood pressure exceeds 160 mmHg, six would decline treatment if their preferences were taken into account. In contrast, 16 out of 33 patients below this blood pressure threshold would accept treatment if their preferences were taken into account. As would be expected, correcting for chance agreement using kappa statistics simply emphasizes the level of disagreement (absolute cardiovascular risk kappa = –0.13; systolic blood pressure kappa = 0.18). There was no evidence that adherence to blood pressure-lowering medication was associated with outcome of the decision analysis [chi-square (1 d.f.) = 0.5, P = 0.5]. |
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
This study demonstrates that individuals vary widely in their preferences for the possible outcomes of treating high blood pressure. When combined with the estimated probabilities of these outcomes actually happening, individual preferences have a considerable effect on the proportion of patients for whom treatment would be recommended. This is illustrated by the marked disagreement between treatment recommendations based on decision analysis and those based on either absolute risk of a CVE or systolic blood pressure.
A systematic review of randomized trials has shown that decision aids improve patient knowledge, comfort and participation in decision making without any increase in anxiety.14 The review included only one published trial of decision analysis, that showed a significant influence on individual clinical decision making.4 To our knowledge, the impact of decision analysis as an aid to shared decision making in hypertensive patients has not been evaluated. Although decision analysis is normally employed in policy decision making, with utilities being elicited at a group level,15 there is potential for decision analysis to assist patients in their individual decision making. This is particularly the case in hypertension, an asymptomatic condition that normally requires lifelong treatment. The decision an individual must make regarding the benefits and harms of treatment options is complex and value-laden. Cardiovascular risk assessment,16 utility assessment17 (see http://osler.wustl.edu/~utiter/index.html?and http://prefdev.ucsd.edu/ for demonstration and download) and decision analysis software18 is becoming increasingly easier to use, and may mean that individual decision analysis will be more common in the future.19
There are some limitations regarding the findings of the present study. (i) Over one-third of patients invited were either unwilling or unable to participate, and of those who attended, 12 (19%) were unable to complete the utility assessment interview. This response rate indicates that this type of decision aid may not be suitable for all patient groups; (ii) a balance must be achieved between a full decision tree containing all possible outcomes and keeping manageable the number of health states that the patient must assess. In this study, each patient had to complete eight standard gamble assessments. All side effects were represented together; the tree would have contained a further three health states had side effects been separated into ‘major’ and ‘minor’; (iii) all interviews were conducted face to face by either AM or JH and, although care was taken to standardize interviews as much as possible, it is not possible to exclude interviewer bias in such circumstances. This could be reduced in future studies by using utility assessment software requiring minimal input from researchers;17 (iv) patients interviewed in this study were already taking blood pressure-lowering treatment. Their preferences for health outcomes, particularly side effects from drugs, may be different once on treatment. Similarly, the absence of any association with adherence to medication may differ in newly diagnosed and treated hypertensive patients; (v) study constraints allowed only collection of repeat prescription data as a measure of medication adherence. The insensitivity of this measure is recognized, although moderate correlations with measures such as other compliance behaviours, serum drug levels and blood pressure control have been found.20
This observational study has demonstrated the impact of individual values regarding the benefits and risks of blood pressure treatment. Prospective research, in the form of a randomized trial, regarding use of decision aids in newly diagnosed hypertensive patients is warranted. This will allow examination of effects on persistence with choices, health outcomes and identification of patients who may benefit from simple and complex interventions to aid decision making.
|
Acknowledgments |
|---|
We would like to thank all 64 patients of Bradgate Surgery, Bristol who provided data for the study. The study was funded by the Scientific Foundation Board of the Royal College of General Practitioners.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
1 Working Party of the British Cardiac Society, British Hyperlipidaemia Association and British Hypertension Society. Joint British recommendations on prevention of coronary heart disease in clinical practice. Heart 1998; 80 (Suppl 2): S1–S29.
2 Wierzbicki A, Reynolds T, Jackson R, Chamberlain J, Fraser W. Sheffield risk and treatment table for primary prevention of coronary heart disease (letters). Lancet 1996; 348: 1039–1041.
3 Sox H, Blatt M, Higgins M et al. Medical Decision Making. Boston: Butterworths, 1988.
4 Clancy CM, Cebul RD, Williams SV. Guiding individual decisions: a randomised, controlled trial of decision analysis. Am J Med 1988; 84: 283–288.[Web of Science][Medline]
5 Legg England S, Evans J. Patients' choices and perceptions after an invitation to participate in treatment decisions. Soc Sci Med 1992; 34: 1217–1225.
6 Anderson KM, Odell PM, Wilson PWF, Kannel WB. Cardiovascular disease risk profiles. Am Heart J 1991; 121: 293–298.[Web of Science][Medline]
7 Department of Health. Health Survey for England. London: HMSO, 1998.
8 Collins R, Peto R, Swales JD (eds). Antihypertensive drug therapy: effects on stroke and coronary heart disease. In Textbook of Hypertension. Oxford: Blackwell Scientific Publications, 1994: 1156–1164.
9
Phillipp T, Anlauf M, Distler A, Holzgreve H, Michaelis J, Wellek S. Randomised, double blind, multicentre comparison of hydrochlorothiazide, atenolol, nitrendipine, and enalapril in antihypertensive treatment: results of the HANE study. Br Med J 1997; 315: 154–159.
10
Bamford J, Sandercock P, Dennis M, Burn J, Warlow C. A prospective study of acute cerebrovascular disease in the community: the Oxfordshire Community Stroke Project 1981–86. Incidence, case fatality rates and overall outcome at one year of cerebral infarction, primary intracerebral and subarachnoid haemorrhage. J Neurol, Neurosurg Psychiatry 1990; 53: 16–22.
11
Volmink JA, Newton JN, Hicks NR, Sleight P, Fowler GH, Neil HAW. Coronary event and case fatality rates in an English population: results of the Oxford myocardial infarction incidence study. Heart 1998; 80: 40–44.
12 Ramsay LE, Williams B, Johnston G et al. Guidelines for the management of hypertension: report of the third working party of the British Hypertension Society. J Hum Hypertension 1999; 13: 569–592.[Web of Science][Medline]
13 Core Services Committee. Guidelines for the Management of Mildly Raised Blood Pressure in New Zealand. Wellington: Ministry of Health, 1995.
14
O'Connor AM, Rostom A, Fiset V et al. Decision aids for patients facing health treatment or screening decisions: systematic review. Br Med J 1999; 319: 731–734.
15
Lilford RJ, Pauker SG, Braunholtz D, Chard J. Decision analysis and the implemenation of research findings. Br Med J 1999; 317: 405–409.
16
Hingorani A, Vallance P. A simple computer programme for guiding management of cardiovascular risk factors and prescribing. Br Med J 1999; 318: 101–105.
17 Nease RF, Tsai R, Hynes LM, Littenberg B. Automated utility assessment of global health. Qual Life Res 1996; 5: 175–182.[Web of Science][Medline]
18 Data 3.5 for Healthcare. TreeAge Software Inc. Williamstown, MA: TreeAge Software Inc., 1999.
19 Coulter A. Partnerships with patients: the pros and cons of shared clinical decision-making. J Health Serv Res Policy 1997; 2: 112–121.[Medline]
20 Steiner JF, Prochazka A. The assessment of refill compliance using pharmacy records: methods, validity and applications. J Clin Epidemiol 1997; 50: 105–116.[Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  N. Moumjid, A. Gafni, A. Bremond, and M.-O. Carrere Shared Decision Making in the Medical Encounter: Are We All Talking about the Same Thing? Med Decis Making, October 1, 2007; 27(5): 539 - 546. [Abstract] [PDF]
|
|
|
|
 |
|
 M.S. van Roosmalen, P.F.M. Stalmeier, L.C.G. Verhoef, J.E.H.M. Hoekstra-Weebers, J.C. Oosterwijk, N. Hoogerbrugge, U. Moog, and W.A.J. van Daal Randomized Trial of a Shared Decision-Making Intervention Consisting of Trade-Offs and Individualized Treatment Information for BRCA1/2 Mutation Carriers J. Clin. Oncol., August 15, 2004; 22(16): 3293 - 3301. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. K Lewis, J. Robinson, and E. Wilkinson Factors involved in deciding to start preventive treatment: qualitative study of clinicians' and lay people's attitudes BMJ, October 11, 2003; 327(7419): 841. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Austin, P. Oakeshott, and S. King Shared decision making in Hypertension Fam. Pract., June 1, 2002; 19(3): 310 - 310. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||