NUR 520 Module 7 Milestone Two Example

Milestone Two: An Analysis Plan for COPD in Coös County, With Measures of Association Worked Through

[Student Name]

Southern New Hampshire University

NUR 520: Epidemiological and Biostatistical Applications in Healthcare

Milestone Two

[Instructor Name]

[Date]

Milestone Two: An Analysis Plan for COPD in Coös County, With Measures of Association Worked Through

Milestone One described COPD among adults in Coös County as a population health problem with smoking at its center and access barriers making it worse. This milestone plans how the final project will analyze that problem. It works through the measures of association the analysis will rely on, using practice datasets because individual-level county data are not publicly available, and it states which real data sources and methods the final project will use. The plan shows that relative risk and attributable fractions are the measures that best support prevention decisions, while odds ratios from case-control data are the practical tool for testing whether a local exposure, mill work, adds to the county's burden. All numbers in the practice datasets below are invented for teaching and are labeled as such.

Practice Dataset 1: A Cohort of Smokers and Nonsmokers

Imagine a practice cohort of 1,000 adults aged 40 to 70, all without COPD at the start, followed for ten years. Two hundred smoke and 800 do not. Over ten years, 40 of the smokers and 32 of the nonsmokers develop COPD. These figures are practice data, not observations from Coös County.

Table 1

Practice Cohort: Ten-Year COPD Risk by Smoking Status

Note. Invented teaching data. Not drawn from any real population.

Measures From the Practice Cohort

Relative risk divides the exposed group's risk by the unexposed group's risk: 20.0% divided by 4.0% equals 5.0; its 95% interval runs from roughly 3.2 up to 7.8. In plain terms, smokers in the practice cohort were five times as likely as nonsmokers to develop COPD over ten years. The risk difference, 20.0% minus 4.0%, is 16 percentage points: for every 100 smokers, 16 more developed COPD than would have if they had the nonsmokers' risk. The attributable fraction among the exposed, calculated as the relative risk minus one divided by the relative risk, is 80%, meaning that in this practice cohort about four in five COPD cases among smokers are attributable to smoking. The population attributable fraction, which also depends on how common smoking is, is about 44%: if smoking were removed from this practice population and the causal assumptions held, overall ten-year risk would fall from 7.2% to 4.0% (Celentano & Szklo, 2019).

Practice Dataset 2: A Case-Control Sample for Mill Work

For a question about past occupational exposure, a cohort would take decades. A case-control design is faster. Imagine a practice sample of 100 adults with COPD and 200 without, drawn from the same clinic population. Among the cases, 45 report working in a paper mill for at least five years; among controls, 50 do. The odds of mill work are 45 to 55 among cases and 50 to 150 among controls, giving an odds ratio of 2.45 with a 95% confidence interval of about 1.48 to 4.08. In this practice sample, people with COPD had about two and a half times the odds of past mill work. Because the interval stays above 1.0, random variation is an unlikely explanation on its own, though it could still reflect confounding by smoking, which a real analysis would need to adjust for.

Comparing Practice Results With Published Evidence

The practice results were built to be broadly consistent with published research, which is where the final project's causal claims will come from. In a 25-year Danish cohort, Lokke et al. (2006) put the 25-year absolute risk for people who kept smoking throughout at a quarter or more, and that quitting substantially reduced that risk. In a case-control study from a U.S. managed care population, Blanc et al. (2009) found an adjusted odds ratio of 2.11 for COPD among people who said their longest job exposed them to airborne irritants, and an odds ratio of 14.1 for those with both occupational exposure and smoking. The final project will cite these real findings, not the practice figures, when estimating how much disease prevention could avert.

Interpreting Attributable Fractions Carefully

Population attributable fractions are persuasive because they translate risk into potential prevention, but they are easy to misuse. Rockhill et al. (1998) cautioned that the fraction depends on the assumption that the exposure causes the disease, that estimates must be adjusted for confounding, and that fractions for different exposures can add to more than 100% because diseases often have several contributing causes. For Coös, a statement such as smoking accounts for a large share of COPD should therefore be supported by adjusted published estimates and phrased as the share of cases that could be prevented if smoking were eliminated, assuming causation.

Confounding and Effect Modification

Two further issues shape the plan. Confounding occurs when a third factor is related to both the exposure and the outcome. Age is the obvious confounder for smoking and COPD, and smoking itself is a likely confounder for mill work, since many mill workers also smoked. A real analysis would handle these by stratifying, for example by calculating the odds ratio for mill work separately among smokers and nonsmokers, or by adjusting in a regression model. Effect modification is different: it occurs when the effect of one exposure depends on another. The very high odds ratio reported for combined smoking and occupational exposure suggests that the two may act together, which would matter for a prevention program aimed at people with both exposures.

The Analysis Plan for the Final Project

The final project will use four kinds of data. CDC PLACES 2025 estimates, with confidence intervals, will describe prevalence of COPD and smoking in Coös compared with other counties, using age-adjusted figures for comparisons of risk and crude figures for estimating the number of people affected. Published cohort and case-control studies will supply relative risks and odds ratios for smoking and occupational exposure. National surveillance will provide the trend context. If a local clinic partner can share de-identified data, the project will describe COPD admissions and rehabilitation referrals by age group and use them as baseline measures for any proposed intervention. Differences will be reported with confidence intervals, and any evaluation of an intervention will compare before and after periods, ideally with a comparison county.

Conclusion

This milestone works through the measures the final project will depend on: relative risk and risk difference for smoking, attributable fractions for prevention potential and odds ratios for local exposures, all calculated here from labeled practice data. It separates those teaching figures from the real published estimates that will support the project's conclusions, and it sets out an analysis plan that matches each question to a data source and a method.

References

Blanc, P. D., Iribarren, C., Trupin, L., Earnest, G., Katz, P. P., Balmes, J., Sidney, S., & Eisner, M. D. (2009). Occupational exposures and the risk of COPD: Dusty trades revisited. Thorax, 64(1), 6-12. https://doi.org/10.1136/thx.2008.099390

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

Lokke, A., Lange, P., Scharling, H., Fabricius, P., & Vestbo, J. (2006). Developing COPD: A 25 year follow up study of the general population. Thorax, 61(11), 935-939. https://doi.org/10.1136/thx.2006.062802

Rockhill, B., Newman, B., & Weinberg, C. (1998). Use and misuse of population attributable fractions. American Journal of Public Health, 88(1), 15-19. https://doi.org/10.2105/AJPH.88.1.15