Benefit-harm analysis and charts for individualized and preference-sensitive prevention: example of low dose aspirin for primary prevention of cardiovascular disease and cancer

Abstract

BackgroundClinical practice guidelines provide separate recommendations for different diseases that may be prevented or treated by the same intervention. Also, they commonly provide recommendations for entire populations but not for individuals. To address these two limitations, our aim was to conduct benefit-harm analyses for a wide range of individuals using the example of low dose aspirin for primary prevention of cardiovascular disease and cancer and to develop Benefit-Harm Charts that show the overall benefit-harm balance for individuals.MethodsWe used quantitative benefit-harm modeling that included 16 outcomes to estimate the probability that low dose aspirin provides more benefits than harms for a wide range of men and women between 45 and 84 years of age and without a previous myocardial infarction, severe ischemic stroke, or cancer. We repeated the quantitative benefit-harm modeling for different combinations of age, sex, and outcome risks for severe ischemic and hemorrhagic stroke, myocardial infarction, cancers, and severe gastrointestinal bleeds. The analyses considered weights for the outcomes, statistical uncertainty of the effects of aspirin, and death as a competing risk. We constructed Benefit-Harm Charts that show the benefit-harm balance for different combinations of outcome risks.ResultsThe Benefit-Harm Charts (http://www.benefit-harm-balance.com) we have created show that the benefit-harm balance differs largely across a primary prevention population. Low dose aspirin is likely to provide more benefits than harms in men, elderly people, and in those at low risk for severe gastrointestinal bleeds. Individual preferences have a major impact on the benefit-harm balance. If, for example, it is a high priority for individuals to prevent stroke and severe cancers while severe gastrointestinal bleeds are deemed to be of little importance, the benefit-harm balance is likely to favor low dose aspirin for most individuals. Instead, if severe gastrointestinal bleeds are judged to be similarly important compared to the benefit outcomes, low dose aspirin is unlikely to provide more benefits than harms.ConclusionsBenefit-Harm Charts support individualized benefit-harm assessments and decision making. Similarly, individualized benefit-harm assessments may allow guideline developers to issue more finely granulated recommendations that reduce the risk of over- and underuse of interventions. The example of low dose aspirin for primary prevention of cardiovascular disease and cancer shows that it may be time for guideline developers to provide combined recommendations for different diseases that may be prevented or treated by the same intervention.

Systematic review with meta‐analysis: the comparative effectiveness of aspirin vs. screening for colorectal cancer prevention

“…Although it may be premature to change practice, a head‐to‐head comparison trial on once‐only flexible sigmoidoscopy vs. aspirin vs. a combination of both strategies seems highly motivated. Furthermore, such a trial would be of importance for general policy making because primary prevention with aspirin of both cancer and cardiovascular disease is not uncontroversial; some benefit‐harm assessments indicate that it would be beneficial, others state it being harmful finally some support individualised assessments to improve treatment and prevention . Finally, in clinical practice, a future question is how aspirin is best combined with other preventive strategies and if chemoprevention with aspirin indeed affects the performance of screening strategies.…”

A proposed approach for quantitative benefit‐risk assessment in diagnostic radiology guideline development: the American College of Radiology Appropriateness Criteria Example

“…The exclusion of other QBRA methods does not imply that excluded methods cannot assist in guideline development. For example, meta‐analytic methods have been used jointly with MCDA and probabilistic simulation modeling commonly accompanies INHB . The 2 broad methods, INHB and MCDA, can be viewed as a starting place for guideline developers exploring quantitative benefit‐risk analysis to which other methods may be appended, as time and budget constraints allow.…”

Effect of prestroke antiplatelets use on first-ever ischaemic stroke severity and early outcome

“…Clinical practice guidelines provide separate recommendations for different diseases that may be prevented or treated by the same intervention. However, there are suggestions that recommendations should be addressed for individuals, and not for entire populations .…”

“…Clinical practice guidelines provide separate recommendations for different diseases that may be prevented or treated by the same intervention. However, there are suggestions that recommendations should be addressed for individuals, and not for entire populations (18). Some authors suggested that prestroke use of AP may be associated with reduced risk and severity of incident ischaemic strokes (7), also in patients with no prior history of stroke (6,19) or transient ischaemic attack (TIA) (19).…”