In descriptive epidemiology, the measures of disease frequency, association, and effects are used to study the occurrence and distribution of disease in a defined population in a specific period. In analytic epidemiology, measures are used to investigate the causes of the health outcome and the associated risk factors.
Measures of association and effects are based on an appropriate comparison between exposure and population groups to set hypotheses about an exposure-outcome relationship.
In this context, exposure is a condition in which the population groups are subjected to different environmental conditions, insects, people with communicable diseases or sexually transmitted diseases, toxic chemicals as well as the characteristics of people including age, race, sex, immune status, marital status, occupation, etc. or surroundings in they live including access to health services and socioeconomic status.
Measures of association are used to assess and quantify the strength of relationships between the exposure i.e., risk factors, and health outcomes i.e., disease, in both relative and absolute terms. These are the statistics that estimate the magnitude and direction of relationships between variables in clinical medicine and epidemiological research. It relies on the data collection method, study design, and the statistical approach used for analysis.
To identify factors that may play an etiological role in the onset of health outcomes, these measures are used to describe the comparison between two or more populations, particularly those with varying exposure or status of health outcomes. Measures of association such as risk ratio and odds ratio are calculated as relative terms while attributable risk and attributable rate are calculated as absolute terms.
Interesting Science Videos
Relative Risk (Risk Ratio)
The ratio of the probability of occurrence of a health outcome in a population group with exposure to the risk of occurrence of the outcome in a non-exposed group is referred to as relative risk or risk ratio.
If the value of the risk ratio is greater than 1, it indicates an increased risk for the exposed group. Similarly, if the risk ratio is less than 1, it indicates a decreased risk for the group in the numerator suggesting that exposure protects against disease occurrence. If the risk ratio is 1, it indicates an equal risk of disease occurrence in both groups.
For example, assuming that 20% of the population with high blood pressure (exposed group) suffer from stroke and 1% with normal blood pressure (non-exposed group) suffer from stroke, then the relative risk of stroke in people with high blood pressure versus people with low blood pressure is calculated as 20%/1% = 20. This suggests that the risk of suffering from the disease is 20 times greater in the exposed group as compared to the non-exposed group. Similarly, the probability of developing lung cancer in smokers when divided by the likelihood of lung cancer in non-smokers, the relative risk of lung cancer is obtained.
Often used in cohort studies, relative risk provides comparative information on the probability of an event based on the exposure in populations with differing disease prevalence. However, it gives a relative value and is not an absolute measure of the occurrence of an event.
Rate Ratio
The ratio that compares the rates (incidence/person-to-time/mortality) of two different groups is known as the rate ratio. The incidence rate ratio is obtained when the incidence rate of the event of interest in the exposed group is divided by the incidence rate of the same event in the unexposed group.
The value of the rate ratio is interpreted in the same way as the risk ratio.
Odds Ratio (OR)
Used in case-control or cross-sectional studies, the odds ratio is defined as the ratio of odds of exposure in the case group to the odds of exposure in the control group. It measures the association between exposure to two categories and health outcomes.
If the value of the Odds ratio is equal to 1, it indicates no association between exposure and health outcomes in the given population. If the value is less than 1, it indicates a negative association which means exposure is protective and the probability of disease occurrence is less likely in case groups. Similarly, if the value is greater than one, it indicates a positive association meaning the exposure is a risk factor and the probability of disease occurrence is less likely in control groups.
In some situations when the prevalence of a health outcome is low (<10%) or rare, the odds ratio closely approximates the risk ratio or rate ratio. However, in instances where the prevalence of a health outcome is common, the OR provides more extreme estimates than the RR.
Attributable Risk (Risk Difference)
An absolute measure of effect, risk difference, or attributable risk gives information about the association of effects with the given treatment. It is obtained by calculating the difference in risk of a health outcome in the exposed group and the risk of the outcome in the unexposed group.
Attributable Risk (AR) = Risk of disease occurrence in the exposed group – Risk of disease occurrence in the unexposed group
The idea of attributable risk provides a valuable method for epidemiological studies that examine the influence of exposure factors on a population level concerning the risk of a health outcome. It is a function of Relative Risk and the prevalence of exposure, therefore, AR can arise as a result of common exposure with low relative risk, or a rare exposure with high relative risk.
Attributable Fraction
The proportion of an event as a result of exposure in the exposed group is expressed as an Attributable Fraction (AF). It is obtained by dividing the attributable risk by the probability of occurrence of an event in the exposed group.
Population Attributable Risk (PAR)
The proportion of risk of a health outcome in both exposed and unexposed populations due to exposure is called Population Attributable Risk (PAR). It is obtained by subtracting the incidence of a health outcome in the exposed population from the total population (sum of exposed and unexposed).
PAR = Incidence of a health outcome in the total population – Incidence in the exposed population
Population Attributable Fraction (PAF)
Similarly, Population Attributable Fraction (PAF) refers to the proportion of cases of a specific health outcome in a defined population that can be attributed to a particular risk factor. It is calculated by dividing the population-attributable risk by the incidence of a health outcome in the total population.
References
- Tenny, S., Hoffman, M.R. (2023). Relative Risk. StatPearls Publishing; Available from: https://www.ncbi.nlm.nih.gov/books/NBK430824/
- Roberts, M.R., Ashrafzadeh, S. & Asgari, M.M (2019). Research Techniques Made Simple: Interpreting Measures of Association in Clinical Research. J Invest Dermatol. 139(3):502-511. DOI: 10.1016/j.jid.2018.12.023. PMID: 30797315; PMCID: PMC7737849.
- https://archive.cdc.gov/www_cdc_gov/csels/dsepd/ss1978/lesson3/section5.html
- https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7737849/
- Tripepi, G., Jager, K.J., Dekker, F.W., & Zoccali, C. (2010). Measures of Effect in Epidemiological Research. Nephron Clinical Practice. 115(2): c91–c93. DOI: https://doi.org/10.1159/000312870
- https://sph.unc.edu/wp-content/uploads/sites/112/2015/07/nciph_ERIC3.pdf
- https://www.populationmedicine.org/sites/default/files/2022-03/CEPH%201_Session%202_Frequency%20and%20Association%20Key%20Points.pdf
- Uter, W., & Pfahlberg, A. (1999). The Concept of Attributable Risk in Epidemiological Practice. Biometrical Journal, 41(8), 985–993. DOI:10.1002/(sici)1521-4036(199912)41:8<985::aid-bimj985>3.0.co;2-l \
- Askari, M., Namayandeh, S.M. (2020). The Difference between the Population Attributable Risk (PAR) and the Potential Impact Fraction (PIF). Iran J Public Health. 49(10):2018-2019. DOI: 10.18502/ijph.v49i10.4713. PMID: 33346214; PMCID: PMC7719653.