PU520 Unit 3 Freight Corridor Association Exercise Example

Fifty-Six of Eight Hundred Near the Freight Line: An Association Measures Exercise Reading Relative Risk, Odds Ratio, Risk Difference and Attributable Fractions From One Two-by-Two Table of New Hypertension Among Composite Residents Living Near and Far From a Freight Corridor

[Student Name]

Purdue University Global

PU520: Principles of Epidemiology

Unit 3 Association Measures Exercise

[Instructor Name]

[Date]

The cohort and every count are composite and bracketed for the writer's own data; the methods cited are real.

The Table

Exposed, within [500] meters: [56] developed hypertension, [744] did not; total [800]. Unexposed, farther away: [42] developed hypertension, [1,158] did not; total [1,200]. All residents: [98] cases among [2,000]. The design is a cohort followed forward for five years, so incidence in each group can be measured directly.

Part A: Risk in Each Group

Question: how often did hypertension develop in each group? Exposed: [56] / [800] = [7.0] per 100. Unexposed: [42] / [1,200] = [3.5] per 100.

Part B: Relative Risk

Question: how many times higher is the risk in the exposed? Relative risk = [7.0] / [3.5] = [2.0]. The 95 percent confidence interval, calculated on the log scale, runs from [1.35] to [2.95]. Because the interval stays above one, random variation is a poor account of the gap, though it says nothing about bias or confounding.

Part C: Odds Ratio

Question: what would the odds ratio show? Odds ratio = ([56] × [1,158]) / ([744] × [42]) = [2.08]. It runs slightly higher than the relative risk because odds exaggerate risk ratios as outcomes become more common; with an outcome as uncommon as here, the two are close, but the gap would widen if hypertension were frequent (Rothman, 2012). In a cohort, relative risk is the natural measure; the odds ratio earns its place in case-control designs, which sample on the outcome and so never observe risk directly. The odds ratio and relative risk agree here only because new hypertension was uncommon in five years among young adults.

Part D: Risk Difference

Question: how much additional risk does living near the corridor carry? Risk difference = [7.0] minus [3.5] = [3.5] per 100 over five years; its 95 percent interval spans roughly [1.4] to [5.6] per 100. Put another way, for about every [29] residents living near the corridor for five years, one additional case of hypertension occurred.

Part E: Attributable Fraction Among the Exposed

Question: what share of cases among the exposed is attributable to the exposure, if the association is causal? (Relative risk minus one) / relative risk = [50] percent.

Part F: Population Attributable Fraction

Question: what share of all cases in this population is attributable to living near the corridor? With [40] percent of residents exposed, the population attributable fraction is [28.6] percent. Equivalently, overall risk was [4.9] per 100 against [3.5] among the unexposed, so about [28.6] percent of cases would not have occurred without the exposure, again if the association is causal.

Part G: What the Measures Cannot Show

Each measure assumes the groups are comparable except for distance from the corridor. They may not be. Housing near freight lines is often cheaper, so residents may have lower incomes, less access to care, more stress and different diets, all linked to hypertension. Noise and air pollution are both possible pathways, and the table cannot separate them. Before treating the relative risk as causal, an analysis would adjust for income, body weight, smoking and family history, and look for a dose-response pattern with distance (Celentano & Szklo, 2019).

Part H: A Next Step the Table Suggests

Question: how could the analysis strengthen a causal reading? Split the exposed group by distance, for example within [200] meters and between [200] and [500], and see whether risk rises as distance falls. A gradient would be harder to explain by income alone, since income is unlikely to change sharply within a few hundred meters. Comparing residents near the corridor on the side facing prevailing winds with those upwind would help separate air pollution from noise. Neither step proves causation, but each makes alternative explanations less likely and tells a planning board whether a buffer of a given width would matter.

Choosing the Measure for Each Audience

For a scientific report: relative risk with its confidence interval, accompanied by the absolute risks so readers can judge scale (Szklo & Nieto, 2019). For a city planning board considering a buffer zone: the risk difference and the number of residents per additional case, which express the burden in people. For a health department estimating how many cases a policy might prevent: the population attributable fraction, with its causal assumption stated.

Summary Table

Relative risk [2.0], interval [1.35] to [2.95]; odds ratio [2.08]; risk difference [3.5] per 100; attributable fraction among exposed [50] percent; population attributable fraction [28.6] percent.

Common Errors to Avoid

Reporting an odds ratio as if it were a relative risk when the outcome is common. Interpreting attributable fractions without stating that they assume causation. Ignoring confounding because the confidence interval excludes one.

References

Celentano, D. D., & Szklo, M. (2019). Gordis epidemiology (6th ed.). Elsevier.

Rothman, K. J. (2012). Epidemiology: An introduction (2nd ed.). Oxford University Press.

Szklo, M., & Nieto, F. J. (2019). Epidemiology: Beyond the basics (4th ed.). Jones & Bartlett Learning.