Measuring impact (Part 2)

Measuring of Association vs Measuring of Impact

In public health and epidemiology, measures of association and measures of impact are two distinct concepts used to understand relationships between exposures and outcomes, as well as their effects on populations.

A measure of association quantifies the strength and direction of the relationship between an exposure and an outcome, providing insight into whether and how strongly a specific factor increases the risk of a particular disease. It answers the question, “Is there a relationship between the exposure and outcome, and how strong is it?”

For example, studies may examine whether smoking increases the risk of developing lung cancer, often expressed through a risk ratio (RR) or odds ratio (OR). If the risk ratio is 2, this indicates that individuals who are exposed (e.g., smokers) are twice as likely to develop the disease compared to those unexposed. Measures of association are primarily used in etiological studies to identify risk factors and establish causal relationships, ultimately guiding research and informing clinical decision-making.

In contrast, measures of impact assess the potential public health effect of reducing or eliminating a specific risk factor, addressing the question, “What is the public health impact of this exposure on the population?” These measures are particularly important in public health planning, as they estimate the burden of disease that could be prevented through targeted interventions. For instance, if 50% of lung cancer cases are attributable to smoking, eliminating smoking could theoretically prevent half of these cases. Such estimates, including Population Attributable Risk Percent (PAR%), are critical for policymaking, resource allocation, and prioritizing prevention strategies to achieve the greatest impact.

While measures of association highlight relationships, measures of impact emphasize opportunities for reducing disease burden and improving population health outcomes. Figure 1 shows the difference in risk levels between exposed and non-exposed groups, highlighting the impact of background risk and exposure.

Figure 1

For example, a study on the benefits of blood pressure lowering after a stroke demonstrated that the treatment group had a lower stroke rate (10%) compared to the placebo group (14%) over a mean follow-up of 3.9 years. The absolute risk reduction (ARR) is calculated by subtracting the stroke rate of the treatment group from that of the placebo group, resulting in a 4% reduction. This means the risk of stroke was reduced by 4% in the treatment group. The relative risk (RR) is determined by dividing the stroke rate of the treatment group (10%) by the placebo group (14%), yielding an RR of 0.71.

To calculate the relative risk reduction (RRR), the relative risk is subtracted from 1, resulting in a 29% reduction. This indicates that individuals in the treatment group are 29% less likely to develop a stroke compared to those in the placebo group, demonstrating the treatment’s effectiveness. While the relative risk reduction of 29% appears more impressive than the absolute risk reduction of 4%, it is important to consider all measures, including the number needed to treat (NNT), to fully understand the benefit of the treatment.

Figure 2

Research has shown that enthusiasm for treatment varies depending on how the data is presented, with relative risk reduction often perceived as more persuasive.

Figure 3

Measures of impact, such as attributable risk, risk difference, and absolute risk reduction, are crucial for understanding the effect of an exposure or treatment. Attributable risk refers to the proportion of disease incidence that can be directly attributed to a specific exposure, indicating the potential for prevention if the exposure is eliminated.

It is calculated by comparing the incidence in exposed and unexposed groups and is often used for risk factors, whereas absolute risk reduction applies to protective interventions. Recognizing the differences between these measures ensures a more accurate interpretation of results and supports better decision-making in clinical practice and public health.

Figure 4

In public health, measures of association and measures of impact play complementary roles in understanding and addressing health outcomes. Measures of association focus on the relationship between an exposure and an outcome, providing insights into the strength and direction of this connection, which is crucial for identifying risk factors and understanding disease causation.

In contrast, measures of impact emphasize the public health significance of an exposure, highlighting the potential benefits of reducing or eliminating it, such as the proportion of disease cases that could be prevented. Together, these measures are essential for designing effective prevention strategies, guiding research, and informing policy decisions to improve population health outcomes.

References:

• (2008). Population Attributable Risk (PAR). In: Kirch, W. (eds) Encyclopedia of Public Health. Springer, Dordrecht. https://doi.org/10.1007/978-1-4020-5614-7_2685

• ROBERT H. RIFFENBURGH, Statistics in Medicine (Second Edition), Academic Press, 2006, Pages 241-279, ISBN 9780120887705

• Celentano, D. (2018) Gordis Epidemiology. 6th Edition, Elsevier, Amsterdam.