Health: New approach quantifies evidence behind risks to human health

A standardized approach for quantifying the strength of evidence that supports associations between risk factors, such as smoking and eating unprocessed red meat, and given health outcomes is presented as part of a package of papers published in Nature Medicine. The package — called the ‘IHME Burden of Proof studies’ — includes a detailed methodology paper and four proof-of-concept studies testing the validity of the approach in assessing evidence for the effect on health of smoking, high blood pressure, and consumption of unprocessed red meat and vegetables.

Exposure to risk factors can influence the likelihood of developing or dying from certain diseases. Understanding and quantifying risk is therefore crucial for steering the direction of public policy — for example, in guiding public health practices, and allowing more-informed choices to be made at both the clinical and the personal level. However, current approaches can be subjective, with levels of uncertainty surrounding quantifying the strength of evidence, hindering accurate interpretation and clear messaging.

Peng Zheng, Christopher Murray, and colleagues present the burden of proof risk function (BPRF) —a meta-analytic approach to estimate a conservative measure of the elevated or reduced risk of a particular health outcome after exposure to a harmful or protective risk factor, respectively. This is based on the available evidence after incorporating disagreement between studies and corrections for study design bias. The BPRF can be interpreted as the smallest level of risk or protection that exposure to a lifestyle factor may offer based on the available data. The authors classify BPRF into an interpretable format, using five-star ratings ranging from one star (potentially no association) to five stars (very strong association).

In the first of four proof-of-concept papers, Xiaochen Dai and colleagues reveal that 29 of the 36 health outcomes studied were significantly associated with smoking, including five outcomes that received a five-star rating for association with increased risk (for example, lung and laryngeal cancer, and peripheral artery disease). In a separate study, Christian Razo and colleagues describe the significant harmful effect of high systolic blood pressure on the risk of ischemic heart disease. A systolic blood pressure of between 107.5 and 165 mmHg, for example, was found to raise a person’s risk by an average of 101.36% (equal to a five-star rating). Consuming unprocessed red meat is ranked on the two-star boundary (weak to no evidence) for an association with increased risk of ischemic heart disease, whereas vegetable consumption is weakly associated (two stars) with a lower risk of ischemic heart disease, according to studies by Haley Lescinsky and colleagues and Jeffrey Stanaway and colleagues, respectively.

The authors conclude that the BPRF approach could be used alongside existing methods to better inform the development of clinical and public health guidelines, and they intend to continue updating their analyses as new evidence becomes available.