Skip to main content
NCBI home page
The BMJ logoLink to The BMJ
. 2000 Jul 1;321(7252):49. doi: 10.1136/bmj.321.7252.49/b

Behavioural counselling in general practice about risk of CHD

Study was grossly underpowered

F D Richard Hobbs 1
PMCID: PMC1127693  PMID: 10939818

Editor—Steptoe et al draw unreliable conclusions from their randomised controlled trial of a brief behavioural counselling intervention, led by nurses, to promote healthy behaviour among adults at increased risk of coronary heart disease.1 Because of considerable difficulties in recruitment and retention the study is grossly underpowered, with only 316 intervention patients and 567 control patients recruited against the required target of 2000. The authors cannot therefore report that “brief counselling on the basis of systematic applications of behavioural principles is more efficacious in stimulating lifestyle modification than conventional counselling.”

The authors have further overinterpreted these unreliable data, since the only changes in behaviour were self reported reductions in dietary fat intake and number of cigarettes smoked and increases in physical activity. Objective measurements, such as body mass index, weight, blood pressure, and smoking cessation (validated by cotinine assay), did not change. Given the unreliability of self reporting as a primary outcome, it is inappropriate to draw positive conclusions.

Furthermore, the authors conclude that “there may be an important role for this counselling” among hard pressed service practitioners and that “more extended counselling . . . may be required.” These seem extraordinary assertions given the negative findings from the study. The authors also ignore their own findings that this “brief” intervention was actually rather substantial: nurse training took four days, and counselling sessions lasted up to 20 minutes on two or three occasions, with one or two follow up telephone calls. This is a considerable time commitment, and the researchers were able to get nurses to recruit only one third of patients needed in intervention practices. In addition, it is inappropriate for researchers to make recommendations on the implications for service practice without conducting any sort of economic analysis.

Given the wealth of unequivocally evidence based interventions that help to reduce coronary heart disease, busy practitioners would have been served better had the authors been more cautious in their conclusions from this negative trial. The paper made an interesting contrast with another randomised controlled trial published in the same issue of the BMJ, which used the stages of change model2 for smoking prevention among schoolchildren.3 That virtually fully powered study (8352 recruited of 8500 needed) produced a reliable negative result.

Acknowledgments

Competing interests: None declared.

References

  • 1.Steptoe A, Doherty S, Rink E, Kerry S, Kendrick T, Hilton S. Behavioural counselling in general practice for the promotion of healthy behaviour among adults at increased risk of coronary heart disease: randomised trial [commentary by S Day] BMJ. 1999;319:943–948. doi: 10.1136/bmj.319.7215.943. . (9 October.) [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Prochaska JO, DiClemente GC, Norcross JC. In search of how people change. Am Psychol. 1992;47:1102–1114. doi: 10.1037//0003-066x.47.9.1102. [DOI] [PubMed] [Google Scholar]
  • 3.Aveyard P, Cheng KK, Almond J, Sherratte E, Lancashire R, Lawrence T, et al. Cluster randomised controlled trial of expert system based on the transtheoretical (“stages of change”) model for smoking prevention and cessation in schools. BMJ. 1999;319:947–953. doi: 10.1136/bmj.319.7215.948. . (9 October.) [DOI] [PMC free article] [PubMed] [Google Scholar]
BMJ. 2000 Jul 1;321(7252):49.

Study had several methodological flaws

Gary Frost 1,2, Caroline Doré 1,2

Editor—Steptoe et al's paper seems to show for the first time that behavioural counselling using the stages of change model in primary care leads to sustained improvements in dietary fat intake, regular exercise, and the number of cigarettes smoked.1-1 But several methodological flaws in the study raise doubts about the validity of these conclusions.

Firstly, the target sample size was 2000, with 10 intervention and 10 control practices and 100 patients per practice. The sample size achieved was only 883 patients. The study had insufficient power to detect the improvements in biological risk factors that the authors considered to be clinically important.

Secondly, patients in the intervention arm had much greater contact with the practice nurses than did controls, with the counselling sessions and telephone contacts to consolidate them. The results would have been more convincing if patients had had equal contact with the practice nurses.

Thirdly, the authors had planned to recruit equal numbers in the intervention and control groups. They achieved only 316 patients in the intervention practices and 567 in the control practices. This difference will also have had an adverse effect on the power of the study.

Fourthly, 626 of the 883 patients completed the four month assessment and only 520 completed the 12 month assessment. Patients lost to follow up tended to be younger and were more likely to be smokers—for example, only 40 of the 124 smokers in the intervention arm completed the 12 month assessment. The authors do not perform any sensitivity analyses to investigate the impact of these differential dropout rates on their conclusions.

Claiming a lifestyle change without any concurrent change in biological risk factors in a self selected group is questionable, as it is known that many people underreport dietary intake and overreport exercise frequency.1-2 We recognise that carrying out large education based studies in primary care is difficult. Because of the problems listed here, however, this study adds little to the literature on this topic. We do not consider that a convincing case has been made for use of the stages of change model in preference to other systems for providing health education information.

Acknowledgments

Competing interests: None declared.

References

  • 1-1.Steptoe A, Doherty S, Rink E, Kerry S, Kendrick T, Hilton S. Behavioural counselling in general practice for the promotion of healthy behaviour among adults at increased risk of coronary heart disease: randomised trial [commentary by S Day] BMJ. 1999;319:943–948. doi: 10.1136/bmj.319.7215.943. . (9 October.) [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 1-2.MacDiarmid J, Blundell J. Assessing dietary intake: who, what and why of under-reporting. Nutrition Research Reviews. 1998;11:231–251. doi: 10.1079/NRR19980017. [DOI] [PubMed] [Google Scholar]
BMJ. 2000 Jul 1;321(7252):49.

Non-attendance for follow up distorts results and shows that people don't like counselling

George Davey Smith 1,2, Shah Ebrahim 1,2, Rachel Bennett 1,2

Editor—Attempts to modify the risk of cardiovascular disease through individually based health education activities have proved disappointing. 2-1 The report of the apparently successful trial of behavioural counselling seems to provide just what is needed to improve the efficacy of health promotion.2-2

A paragraph in This Week in the BMJ (issue of 9 October) provides the paper's “take home” message—that brief behavioural counselling resulted in reductions in risk factors and could be a useful strategy. Unfortunately, this conclusion does not follow from the study's findings. The follow up rates were low and were related to the group to which the patients were assigned.

By far the most dramatic result—more so than the results highlighted in the paper—is the influence of counselling on people not attending follow up. The odds ratio of non-attendance among baseline smokers in the intervention compared with the control group at 12 months is 2.35 (95% confidence interval 1.48 to 3.76; P=0.0001). This suggests that a large proportion of smokers assigned to the intervention did not attend follow up, probably because they had failed to stop or reduce their smoking. Thus comparison of smoking behaviour in only those who attended follow up is misleading. A similar effect is seen with physical activity: a smaller percentage of the people engaging in little activity at baseline returned in the intervention than the control group.

Trials with such major losses to follow up are difficult to interpret. Analytical options include assuming that all those not returning to follow up continued to smoke or that smoking rates were similar to those in the control group. Simply examining the effects of counselling among those attending for follow up will give a biased result. Randomised controlled trials of health education interventions that achieve reasonable follow up (closer to 90% among smokers than the authors' 30%) show much less influence on risk factors than this trial.2-1

The poor recruitment and follow up do not support the trial's efficacy but do show that many high risk patients and primary care teams do not want brief behavioural counselling. Although the counselling methods used may be considered promising, before extended counselling and support are tested in randomised controlled trials further work is needed to find out why brief counselling failed. Better means of achieving change in mass lifestyle behaviours may be found by focusing on the larger forces (the food industry, tobacco promotion, transport policy) that shape the way we live.

Acknowledgments

Competing interests: None declared.

References

  • 2-1.Ebrahim S, Davey Smith G. Systematic review of randomised controlled trials of multiple risk factor interventions for preventing coronary heart disease. BMJ. 1997;314:1666–1674. doi: 10.1136/bmj.314.7095.1666. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2-2.Steptoe A, Doherty S, Rink E, Kerry S, Kendrick T, Hilton S. Behavioural counselling in general practice for the promotion of healthy behaviour among adults at increased risk of coronary heart disease: randomised trial [with commentary by S Day] BMJ. 1999;319:943–948. doi: 10.1136/bmj.319.7215.943. . (9 October.) [DOI] [PMC free article] [PubMed] [Google Scholar]
BMJ. 2000 Jul 1;321(7252):49.

Authors' reply

Andrew Steptoe 1,2, Elizabeth Rink 1,2, Sally Kerry 1,2, Tony Kendrick 1,2, Sean Hilton 1,2

Editor—The authors of these letters make some legitimate points in commenting on our paper, although they are perhaps guiltier than us in overinterpreting what we actually claimed in reporting our findings.

Hobbs takes us to task for overinterpreting our positive results, when difficulties in recruitment left us with an underpowered trial. We accept the loss of power, and discuss it in the paper. This, however, is of less relevance to our significant findings in reported behaviour change than to the likelihood of a type II error. The loss of power may partly explain the non-significant changes in biological risk factors—a point made by Frost and Doré. We were not funded to carry out an economic analysis, although we agree that this would have been desirable.

Davey Smith et al refer to the “apparently successful trial”—not a claim made anywhere by us. We believe that our conclusions and comments on implications for primary care were appropriately cautions. We certainly do not disagree with Davey Smith et al about the need for policy changes, but even if these occur they are unlikely to remove the need to identify effective health promotion and disease prevention activities for primary care.

All authors comment on the apparent unacceptability of the behavioural counselling method to staff and patients. Recruitment to this study started around the time of publication of two influential studies on primary prevention of coronary heart disease in general practice,3-1,3-2 and we have little doubt that there was an atmosphere of scepticism regarding health promotion. Staff attitudes have been reported elsewhere and are less negative than asserted by Hobbs and Davey Smith et al.3-3

We do not believe that the time expected of practice nurses (if the method was shown unequivocally to be effective) is unrealistic. Much of it related to assessments for the analyses and would not translate to routine care. Training and practice time is no more than that for nurses offering respiratory care on a routine basis in recent years. We are addressing Frost and Doré's concern about disparate time allocation by practice nurses to the control and intervention groups in a separate trial.

We entirely accept the concerns expressed about the high dropout rate of smokers and that this might vitiate differences found in those followed up. We stand by our conclusion that, if appropriately targeted, this method of counselling may have an in important role in primary care prevention.

Acknowledgments

Competing interests: None declared.

References

  • 3-1.Imperial Cancer Research Fund OXCHECK Study Group. Effectiveness of health checks conducted by nurses in primary care: final results of the OXCHECK study. BMJ. 1995;310:1099–1104. [PMC free article] [PubMed] [Google Scholar]
  • 3-2.Family Heart Study Group. Randomised controlled trial evaluating cardiovascular screening and intervention in general practice: principal results of British family heart study. BMJ. 1994;308:313–320. [PMC free article] [PubMed] [Google Scholar]
  • 3-3.Steptoe A, Doherty S, Kendrick T, Rink EM, Hilton SR. Attitudes to cardiovascular health promotion among general practitioners and practice nurses. Fam Pract. 1999;16:158–163. doi: 10.1093/fampra/16.2.158. [DOI] [PubMed] [Google Scholar]

Articles from BMJ : British Medical Journal are provided here courtesy of BMJ Publishing Group

RESOURCES