‘Do physicians understand screening statistics? A National Survey of primary Care Physicians in the United States’ (Ann Intern Med. 2012;156(5):340-349. http://annals.org/article.aspx?articleid=1090696) concluded ‘Most primary care physicians mistakenly interpreted improved survival and increased detection with screening as evidence that screening saves lives. Few correctly recognized that only reduced mortality in a randomized trial constitutes evidence of the benefit of screening’.
-Apparently, this is just part of a wider problem where physicians fail to understand numerical information. In 1996 it was demonstrated that GPs’ decisions were different when information was presented in Relative Risk terms rather than absolute risk terms. Something similar is true for patients.
Br J Gen Pract. 1996 Nov;46(412):661-3.
‘Same information, different decisions: the influence of evidence on the management of hypertension in the elderly.’ Cranney M, Walley T.
BACKGROUND: Evidence-based medicine requires general practitioners (GPs) to act upon the results of clinical trials. Clinical trial evidence may be difficult to understand and apply in practice.
AIM: To investigate whether GPs were unduly influenced in managing hypertension in the elderly by the ways in which benefits of trial results were presented, and to establish whether their current treatment of an elderly hypertensive patient was broadly in line with recent clinical trial evidence.
METHOD: Seventy-three GPs attending a refresher course were given a written questionnaire containing data from one clinical trial of treatment of hypertension in the elderly presented in four different ways (absolute risk reduction, relative risk reduction, difference in event-free patients, and number of patients who had to be treated in order to prevent one clinical event), as if from four different trials. The effect of each presentation on treatment preferences was assessed using Likert scales. The results were analysed to determine whether the method of presentation of results influenced decision making. A clinical scenario was presented to investigate their current treatment preferences in an elderly hypertensive.
RESULTS: All GPs returned completed questionnaires. Relative risk reduction was the only presentation which was significantly different from the others, and was the most likely to influence prescribing. In free-text comments, 75% of GPs admitted having problems understanding statistics commonly found in medical journals. More than 90% conformed with recent clinical trial evidence for the management of hypertension.
CONCLUSION: GPs were most influenced by relative risk reduction, and were unaware of how the presentation of research results could affect treatment decisions. Most GPs freely admitted to difficulty in comprehending medical statistics. Almost all of the GPs expressed treatment decisions which were broadly in line with clinical evidence.
Br J Gen Pract. 2001 Apr;51(465):276-9.
Patients’ responses to risk information about the benefits of treating hypertension.
Misselbrook D, Armstrong D.
BACKGROUND: The medical profession is often presented with information on the value of treatment in terms of likely risk reduction. If this same information was presented to patients–so enabling them to give proper informed consent–would this affect their decision to be treated?
AIM: To examine patients’ choice about treatment in response to different forms of risk presentation.
DESIGN OF STUDY: Postal questionnaire study.
SETTING: The questionnaire was sent to 102 hypertensive patients and 207 matched non-hypertensive patients aged between 35 and 65 years in a UK general practice.
METHODS: Patients were asked the likelihood, on a four-point scale, of their accepting treatment for a chronic condition (mild hypertension) on the basis of relative risk reduction, absolute risk reduction, number needed to treat, and personal probability of benefit.
RESULTS: An 89% response rate was obtained. Of these, 92% would accept treatment using a relative risk reduction model, 75% would accept treatment using an absolute risk reduction model, 68% would accept treatment using a number needed to treat model, and 44% would accept treatment with a personal probability of benefit model.
CONCLUSION: Many patients may prefer not to take treatment for mild hypertension if the risks were fully explained. However, given that the form of the explanation has a strong influence on the patient’s decision, it is not clear how decision-making can be fully shared nor what should constitute informed consent to treatment in this situation.