| Course | PU570 Chronic Disease Epidemiology |
|---|---|
| Unit | Unit 2 |
| Paper type | Measure interpretation exercise |
| Length | About 1,000 words, 6 pages double-spaced |
| Format | APA 7 student paper |
| School | Purdue University Global |
| Program | Master of Public Health (MPH) |
| Updated | October 2026 |
Free sample paper for PU570 Unit 2
One Cohort, Six Measures, and Only Two of Them Answer the Question Anybody Asked: A Unit 2 Measure Interpretation Exercise Reading Risk, Risk Ratio, Odds, Attributable Fraction and Number Needed From the Same Two Percentages
[Student Name]
Purdue University Global
PU570: Chronic Disease Epidemiology
Unit 2 Measure Interpretation Exercise
[Instructor Name]
[Date]
The cohort of 16,000 women and its percentages are composite; the measures, the formulas and the research cited are real.
The Data
Cohort of 16,000 women, 6,000 with body mass index of 30 or above and 10,000 between 18.5 and 24.9, followed twelve years with complete ascertainment assumed. Endometrial cancer cases: 108 among women with obesity, 70 among women at a healthy weight.
Measure One: Cumulative Incidence, Each Group
1. With obesity: 108 divided by 6,000, which is 1.8 percent over twelve years.
2. At a healthy weight: 70 divided by 10,000, which is 0.7 percent over twelve years.
What it answers: the probability that a woman in each group was diagnosed during the follow-up period. What it cannot do: say anything about rate over time, since it treats twelve years as one interval and ignores when cases occurred and who left the study.
Measure Two: Risk Ratio
1.8 divided by 0.7 equals 2.57. What it answers: how many times greater the twelve-year risk was in the exposed group. What it cannot do: tell a woman what her chance is. A ratio of 2.57 is compatible with a risk of 90 percent against 35 percent and with 1.8 percent against 0.7 percent, and only one of those should alarm anybody.
Measure Three: Risk Difference
1.8 minus 0.7 equals 1.1 percentage points over twelve years. What it answers: the extra cases per hundred women over twelve years, which is the measure a woman deciding whether to care about this should be given.
Measure Four: Odds Ratio
Odds with obesity: 108 divided by 5,892, which is 0.01833. Odds at a healthy weight: 70 divided by 9,930, which is 0.00705. Odds ratio 2.60. What it answers: the ratio of odds, which here is almost identical to the risk ratio because the outcome is rare in both groups. What it cannot do: substitute for the risk ratio when an outcome is common, where the odds ratio runs further from one and is routinely misread as a risk ratio in abstracts. The odds ratio and the risk ratio agree here only because endometrial cancer is rare. Nothing about the formula guarantees that, and the agreement should be stated rather than assumed.
Measure Five: Attributable Fraction, Two Versions
1. Among the exposed: the risk ratio minus one, divided by the risk ratio, which is 1.57 divided by 2.57, or 61 percent. Read as: among cases occurring in women with obesity, 61 percent would not have occurred had their risk equaled that of the unexposed group, if the association is causal and unconfounded.
2. In the population: with 37.5 percent of the cohort exposed, the population attributable fraction is about 37 percent of all cases in this cohort.
What it cannot do: survive a causal assumption that fails. Every attributable fraction is a counterfactual claim, and the phrase if causal belongs in the sentence rather than in a footnote.
Measure Six: Number Needed to Prevent
The inverse of the risk difference, 1 divided by 0.011, which is 91. Read as: if 91 women with obesity in this cohort had carried the risk of the healthy-weight group across twelve years, one endometrial cancer would have been avoided. What it cannot do: describe any actual intervention, because nothing here tested weight change, and a measure built from an observed contrast is not an estimate of what a program would achieve.
Which Two Measures Answer the Question Anybody Asked
A woman asking whether this matters to her needs the risk difference and the absolute risks behind it, 1.8 against 0.7 percent. A health department deciding where to spend needs the population attributable fraction, since that single figure folds in how strong the association is together with how widespread the exposure happens to be. The risk ratio is the measure most likely to be quoted and the least likely to answer either question on its own, a preference for relative measures that has been part of the field's habits for a century (Poole, 2010).
What a Rate Would Have Added
The exercise is built from cumulative incidence because that is what the two percentages are, and it is worth saying what a rate would have given instead. An incidence rate uses person-time in the denominator, so it accommodates women who died, left the state or were diagnosed at different points in the twelve years. With 6,000 and 10,000 women followed for a mean of 10.4 and 11.1 years respectively, the rates would be about 173 and 63 cases per 100,000 person-years, and their ratio, 2.74, differs slightly from the risk ratio of 2.57. The difference is not an error in either figure. It arises because the exposed group contributed less follow-up time per woman, which a risk calculation hides and a rate calculation exposes. In a cohort of older adults, where loss to competing causes is substantial, the rate is the measure most registry reports use and the one this exercise would recommend for publication.
What the Exercise Cannot Settle From These Numbers
Confounding, which is untouched here: the cohort's exposed and unexposed groups differ in parity, diabetes and physical activity, and the published meta-analytic estimates for body mass index and endometrial cancer come from analyses that adjust for them (Renehan et al., 2008). Duration of exposure, which a single baseline measurement cannot capture. Competing risk of death from other causes over twelve years, which cumulative incidence ignores and which matters in this age band. And the deeper point that each of these measures encodes a different question about causation rather than a different description of the same fact, which is the distinction the methodological literature makes central (Rothman & Greenland, 2005).
References
Poole, C. (2010). On the origin of risk relativism. Epidemiology, 21(1), 3-9. https://doi.org/10.1097/EDE.0b013e3181c30eba
Renehan, A. G., Tyson, M., Egger, M., Heller, R. F., & Zwahlen, M. (2008). Body-mass index and incidence of cancer: A systematic review and meta-analysis of prospective observational studies. The Lancet, 371(9612), 569-578. https://doi.org/10.1016/S0140-6736(08)60269-X
Rothman, K. J., & Greenland, S. (2005). Causation and causal inference in epidemiology. American Journal of Public Health, 95(S1), S144-S150. https://doi.org/10.2105/AJPH.2004.059204
What the PU570 Unit 2 assignment asks for
Unit 2 usually supplies figures and asks what each measure means. Compute every measure from the same raw numbers so a reader can see that the differences come from the questions rather than the data. For each one, write what it answers and what it cannot support, in that order. Say why an odds ratio and a risk ratio agree when an outcome is rare and diverge when it is common. Put the causal assumption inside every attributable fraction sentence rather than in a footnote. Treat a number needed figure derived from an observational contrast as a counterfactual, not a program estimate. Then name the measures that answer a patient's question and a health department's, and list what the data cannot settle.
Inside the PU570 Unit 2 example
The exercise here states the cohort and the case counts, then works through six measures in order, each with its arithmetic shown and a short note on what it cannot do. A highlighted line explains that the odds ratio agrees with the risk ratio only because the outcome is rare. A margin note points out that every measure derives from the same two percentages. A further section converts the risks into rates with person-time, showing a rate ratio of 2.74 beside the risk ratio of 2.57 and explaining why they differ. The closing sections name the two measures that answer real questions, flagged by a third note, and list confounding, duration and competing risk as unsettled.
PU570 Unit 2 rubric: what earns full points
Graders in this unit look past the arithmetic to the reading. Credit is heaviest where each measure is paired with the question it answers and the claim it cannot carry. Explaining the rare-disease relationship between odds and risk ratios earns marks, as does noticing that the risk ratio alone cannot tell an individual anything. Attributable fractions must carry their causal condition in the sentence to score full marks. A number needed figure described as a counterfactual rather than an intervention effect is rewarded. Points also go to distinguishing risk from rate, and to a closing list of what confounding, follow-up time and competing mortality leave open. APA 7 citation of the substantive literature is expected.
Common PU570 Unit 2 mistakes, and how to avoid them
The frequent mistake is correct arithmetic with no interpretation, which answers half the assignment. Write what each number means to somebody. The second error is quoting a relative measure as though it described individual risk; give the absolute risks alongside. Third, do not report an attributable fraction without its causal condition, because the number is a counterfactual and reads as a measurement. Watch the number needed figure, which describes a contrast rather than a program. And say where confounding sits, since an unadjusted comparison in a cohort like this one differs from published adjusted estimates. A worked version on your own figures takes two days.
Get a PU570 Unit 2 example written to your instructions
Upload the results table and the questions from the PU570 Unit 2 exercise, along with the grading criteria. Every measure is computed or read in the order asked, each paired with what it licenses and what it cannot support. Turnaround is 24-48 hours, and that first custom sample carries no charge. The paper above is an original model document written by our desk, not a submitted student paper and not an official Purdue University Global document.
More PU570 papers and MPH samples
- PU570 Unit 1: Hazard and Risk Discussion Post
- PU570 Unit 3: Study Design Critique
- PU570 Unit 4: Causal Reasoning Paper
- PU570 Unit 5: Confounding Analysis
- PU570 Unit 6: Seminar Reflection
- PU570 Unit 7: Surveillance Data Analysis
- PU570 Unit 8: Risk Factor Profile
- PU570 Unit 9: Prevention Strategy Brief
- PU570 Unit 10: Epidemiological Report
- PU565 Unit 2: Paradigm Comparison Memo
- PU555 Unit 2: Research Question Memo
- PU550 Unit 2: District Health Agency Profile
- PU545 Unit 2: Hazard Vulnerability Analysis
PU570 Unit 2 questions, answered
Where can I find a free PU570 Unit 2 sample paper?
Every measure, its arithmetic and its limits are on this page in full.
Which measure should a patient be given?
The absolute risks and the risk difference. A ratio without a baseline cannot tell anybody what their chance is.
Why do the odds ratio and risk ratio agree here?
Because the outcome is rare in both groups. When an outcome is common the odds ratio runs further from one and is often misread.
What does a number needed figure mean in a cohort?
A counterfactual: how many exposed people would have had to carry the unexposed risk to avoid one case. It is not a program estimate.
What does this exercise leave unsettled?
Confounding, exposure duration from one baseline measurement, and competing mortality over twelve years in this age band.