Measuring impact (Part 1)

Understanding Impact Measurement in Public Health

In public health, one of the primary goals is to measure and communicate the impact of interventions, exposures, or risk factors. Whether we’re assessing disease prevention programs, evaluating treatments, or identifying high-risk behaviours, quantifying the effect is essential. Several measures help us achieve this: Attributable Risk, Absolute Risk Reduction (ARR), Population-attributable risk (PAR), and Population-attributable risk fraction (PAR%). These metrics offer insights into the scale of impact and guide decisions for individuals, healthcare providers, and policymakers.

1. Why Measure Impact?

Let’s understand why impact measurement matters:

• Quantify Effect: Determine how much a particular factor increases or decreases risk.

• Prioritize Interventions: Decide which strategies should receive resources based on their measurable impact.

• Communicate Benefits: Help stakeholders (patients, providers, and the public) understand the practical value of prevention or treatment.

For example, consider a new drug that reduces the risk of heart disease. Knowing the exact risk reduction helps compare it with other drugs and decide if it’s worth implementing widely.

2. What is Attributable Risk?

Attributable Risk (AR) measures the proportion of disease cases that can be attributed to a specific exposure within an exposed group. It helps us determine the “burden” of a disease caused by a specific risk factor.

The attributable risk (AR) is a measure of association that provides information about the absolute effect of the exposure or excess risk of disease in those exposed compared with those unexposed, assuming that the risk is causal. The risk or rate difference estimates the excess risk caused by exposure in the exposed group, that is, the risk attributable to the factor being investigated (e.g. smoking, alcohol).

• Attributable risk (AR) = incidence risk among exposed – incidence risk among unexposed

The risk in the exposed group refers to the probability of disease or outcome among those exposed to a certain factor. The risk in the unexposed group serves as a baseline comparison.

Example

Imagine we study smoking and lung cancer:

• Among smokers, 20 out of 100 develop lung cancer (Risk = 20%).

• Among non-smokers, 5 out of 100 develop lung cancer (Risk = 5%).

The attributable risk is: 20% − 5% = 15%.

This means that 15% of the cases of lung cancer among smokers can be attributed to smoking.

Why is Attributable Risk Important?

• AR tells us how much disease burden could be prevented by eliminating the exposure (e.g. if no one smoked).

• It’s particularly useful for designing prevention programs and setting public health priorities.

3. Risk Difference or Absolute Risk Reduction

Risk Difference (RD), also called Absolute Risk Reduction (ARR), measures the difference in risk between two groups: typically, an intervention group and a control group.

• Risk Difference (RD) or ARR = Risk in Control Group − Risk in Treatment or Intervention Group

Unlike attributable risk, ARR focuses on comparing two interventions or treatments. Usually used for a protective effect.

Example

Suppose we study a new drug to prevent heart attacks:

• In the control group (no drug), 10 out of 100 people had a heart attack (Risk = 10%). • In the treatment group (with the new drug), only 4 out of 100 had a heart attack (Risk = 4%).

The ARR is: 10% − 4% = 6%.

This means the new drug reduces the absolute risk of heart attack by 6%.

Why is ARR Important?

• It helps determine the practical benefit of an intervention. A small relative effect can sometimes look large if the baseline risk is low, but ARR gives a clearer picture of actual risk reduction.

• It’s especially useful in clinical trials and patient communication because it conveys the absolute benefit of a treatment.

4. Understanding Attributable Risk and ARR in Context

While attributable risk and absolute risk reduction both measure impact, they focus on different contexts:

• Attributable Risk focuses on exposures (e.g., smoking and lung cancer). It’s used in public health to estimate preventable disease burden.

• Absolute Risk Reduction focuses on interventions (e.g., new drugs or treatments). It’s used to measure treatment effectiveness.

Both measures are important but serve different purposes. Attributable Risk highlights prevention opportunities, while ARR quantifies the benefit of treatments.

5. Attributable Risk Percent (AR%) Sometimes, it is more useful to express attributable risk as a percentage or proportion. This is known as the Attributable Risk Percent (AR%) or the proportion of risk attributable to the exposure. It can be calculated by subtracting the risk in the unexposed group from the risk in the exposed group, dividing the result by the risk in the exposed group, and then multiplying it by 100 to express it as a percentage.

Attributable risk % (AR%) = (risk in exposed group – risk in unexposed group)/ (risk in exposed group) x 100

Here’s an example of that calculation using the attributable risks from a study of deaths due to lung cancer and smoking.

Lung cancer AR % = (166-7)/166 x 100 = 96%

Heart Diseases AR % = (599-422)/599 x 100 = 30%

For lung cancer deaths, the attributable risk per cent is 96%, which represents the proportion of lung cancer deaths among smokers attributable to smoking. In contrast, the attributable risk per cent for heart disease deaths is 30%.

When comparing these attributable risk percentages, it becomes evident that eliminating smoking would reduce a larger proportion of lung cancer deaths among smokers than heart disease deaths. However, because heart disease is far more common in the general population compared to lung cancer, the absolute number of heart disease deaths prevented by eliminating smoking would be greater. This is true even though the relative risk of dying from heart disease in smokers compared to non-smokers is smaller.

5. Population Attributable Risk (PAR)

The PAR is a similar measure to the attributable risk (or risk difference) but is concerned with the rate in the total study population (exposed and unexposed) compared with the rate in the exposed group.

• PAR=Risk in the total population−Risk in unexposed group

The PAR provides a measure of the public health impact of exposure in the population (assuming that the association is causal).

Key Differences:

For example, if we know the death rate due to lung cancer in the general population and in non-smokers, we can also calculate the risk in the general population of lung cancer deaths attributable to smoking. To do this we take the lung cancer death rate in the general population and subtract the lung cancer death rate in the non-smokers.

The result is the population-attributable risk (or PAR). This rate tells us the number of lung cancer deaths that would be eliminated from the general population if the exposure, in this case, smoking, were eliminated.

Suppose the lung cancer death rate in the general population is 60 per 100,000 persons per year, while in non-smokers, it is 7 per 100,000 persons per year. The rate of lung cancer deaths in the general population attributable to smoking is the difference between these rates: 60 – 7 = 53 per 100,000 persons per year.

The Population Attributable Risk (PAR) is a useful measure for guiding public health decisions, as it helps prioritize resources where they can most effectively protect public health. Similar to the attributable risk, the PAR can also be expressed as a percentage, known as the Population Attributable Risk Percent (PAR%).

6. Population Attributable Risk Percent (PAR %)

Sometimes it is more useful to express attributable risk in terms of a per cent, or proportion. Attributable risk can be calculated as a percentage by subtracting the risk in the unexposed from the risk in the exposed and dividing the result by the risk in the exposed.

• Population Attributable risk % (PAR %) = (risk in exposed group – risk in unexposed group)/ (risk in exposed group) x 100

Population-attributable risk is helpful for guiding public health decisions about where to focus resources most effectively to protect the public’s health.

As with attributable risk, population attributable risk can be expressed as a per cent. We calculate it in a way similar to calculating attributable risk per cent: the incidence in the general population minus the incidence in the unexposed population divided by the incidence in the general population.

To calculate the PAR%, we subtract the incidence in the unexposed population (non-smokers) from the incidence in the general population and divide the result by the incidence in the general population, then multiply by 100:

Population Attributable risk % (PAR%) = (incidence in general population – incidence in unexposed group)/ (incidence in general population) x 100

Turning our smoking example into a population-attributable risk per cent, we can say that of the 60 per 100,000 deaths per year due to lung cancer in the general population, 60 minus 7 divided by 60, or 88%, can be attributed to smoking.

PAR% = (60 -7)/60 x 100 = 88%

7. Translating Risk Measures into Action

Understanding these measures isn’t just theoretical—it guides critical decisions: • For Policymakers: Attributable Risk helps prioritize interventions, such as smoking cessation programs or clean water initiatives. • For Clinicians: ARR helps evaluate the effectiveness of treatments, helping providers make evidence-based decisions for their patients. • For Patients: Presenting ARR instead of Relative Risk makes it easier for patients to understand the real benefit of a treatment.

Example

Imagine a vaccine trial where:

• Control group: 100 infections per 1,000 people (10%).

• Vaccinated group: 30 infections per 1,000 people (3%).

• ARR = 10% − 3% = 7%

• Relative Risk Reduction = (10% − 3%) ÷ 10% = 70%

While the vaccine reduces risk by 70% (RRR), the actual reduction is 7% (ARR), which is easier to interpret for practical decision-making.

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.