PU555 Unit 4 Secondary Data Source Review Example

Four Candidate Files, Two Variables That Matter: A Unit 4 Secondary Data Source Review Weighing a National Examination Survey, a Telephone Survey, a Registry and One Clinic's Records for a Kidney Awareness Study

[Student Name]

Purdue University Global

PU555: Foundations of Public Health Research

Unit 4 Secondary Data Source Review

[Instructor Name]

[Date]

The composite health center is invented; the four data sources, their contents and the research cited are real.

The Two Variables the Question Requires

1. A measured laboratory value from which filtration can be estimated, obtained the same way for every participant.

2. A statement from the person about whether anyone ever told them they have kidney disease.

A source holding only the first can describe the disease. A source holding only the second can describe belief. Only a source holding both can measure the gap between them, which is the whole question.

Source One: The National Health and Nutrition Examination Survey

1. Has variable one: yes. Serum creatinine is measured in a mobile examination center, not self-reported, with documented laboratory methods and calibration across cycles.

2. Has variable two: yes. A kidney condition item asks the participant whether any clinician ever said so.

3. Sample: a few thousand examined adults per two-year cycle, nationally representative with weights, so the 30 to 59 band holds a few hundred people per cycle and pooling is needed.

4. Cost and access: free, public, documented. Analysis must use the survey weights and the complex design, or the standard errors will be wrong.

5. Verdict: the only source that answers the question as written. The classic national estimates of awareness come from exactly this file (Plantinga et al., 2008).

Source Two: The Behavioral Risk Factor Surveillance System

1. Has variable one: no. There is no laboratory measurement; kidney disease status is self-reported.

2. Has variable two: yes, in effect, since the kidney item is a self-report of having been told.

3. Sample: very large, with state and often county estimates, which is its real advantage.

4. Limits: telephone sampling, self-report throughout, and response rates that vary by state; comparisons across populations with different reporting behavior need caution (Pierannunzi et al., 2013).

5. Verdict: cannot measure unawareness, because the people who were never told are invisible in a data set where the only kidney variable is what people say. Useful later for geography, not for the main estimate.

A survey that asks only what people believe cannot count the people whose problem is that they believe nothing is wrong.

Source Three: The National Registry of Treated Kidney Failure

1. Has variable one: yes, clinically, but only for people already in treatment for kidney failure.

2. Has variable two: no.

3. Strength: near-complete coverage of dialysis and transplant, with long follow-up, which makes it the authority for incidence and outcomes at the end of the disease.

4. Verdict: the wrong population. This study is about people in the middle of the disease who have not been told, and the registry begins where that story has already ended.

Source Four: One Composite Health Center's Records

1. Has variable one: yes, from ordinary care laboratory results, with the irregularity that comes with it; values exist only for patients whose clinicians ordered the test.

2. Has variable two: partly and unreliably. A diagnosis code or problem-list entry records what the clinician wrote, not what the patient was told or understood.

3. Strength: local, specific, and linkable to the care process the study wants to criticize.

4. Limits: selection by who was tested, no denominator for the untested, and a variable that answers a different question. Research using clinical markers has shown that awareness tracks more than filtration alone, which is another reason a chart proxy is not a substitute for asking the person (Tuot et al., 2011).

5. Verdict: unusable for the prevalence estimate; valuable as the sampling frame for interviews, which is how the final proposal uses it.

What the Review Decides

The national examination survey supplies the estimate, pooled across cycles with weights and with unweighted counts reported. The health center records supply the interview sample. The telephone survey is kept for a secondary geographic comparison, labeled as self-report throughout. The registry is cited for context and used for nothing. This division of labor is the output of the review, and it is more useful than a ranking of the four sources would have been.

What the Division of Labor Costs

Using three sources for three purposes is not free, and the review states the price. The estimate and the interviews will describe different populations, one national and one from a single clinic in one county, so the interviews illustrate the national finding rather than explaining it. The geographic comparison will rest on self-report while the main estimate rests on laboratory values, so the two cannot be placed on the same axis in any figure. And a study drawing on three sources needs three sets of documentation, three data use agreements or their equivalents, and three descriptions in the methods section, which is a real cost in a one-term project. The alternative, a single source answering everything, does not exist here, so the cost is accepted and written down.

Three Limits of Using Any Secondary Source Here

First, none of these files was collected to answer this question, so every variable is an approximation of what the study means. Second, the awareness item's wording matters enough that the pooled synthesis across studies found it driving much of the variation between published estimates, so the analysis must quote the item verbatim (Chu et al., 2021). Third, the laboratory equation changed in 2021 to remove the race coefficient, which shifts who falls inside the 30 to 59 band and makes comparisons with older published estimates imperfect rather than impossible.

References

Chu, C. D., Chen, M. H., McCulloch, C. E., Powe, N. R., Estrella, M. M., Shlipak, M. G., & Tuot, D. S. (2021). Patient awareness of CKD: A systematic review and meta-analysis of patient-oriented questions and study setting. Kidney Medicine, 3(4), 576-585. https://doi.org/10.1016/j.xkme.2021.03.014

Pierannunzi, C., Hu, S. S., & Balluz, L. (2013). A systematic review of publicly available population health surveys: The Behavioral Risk Factor Surveillance System, 2004-2011. BMC Medical Research Methodology, 13, 49. https://doi.org/10.1186/1471-2288-13-49

Plantinga, L. C., Boulware, L. E., Coresh, J., Stevens, L. A., Miller, E. R., Saran, R., Messer, K. L., Levey, A. S., & Powe, N. R. (2008). Patient awareness of chronic kidney disease: Trends and predictors. Archives of Internal Medicine, 168(20), 2268-2275. https://doi.org/10.1001/archinte.168.20.2268

Tuot, D. S., Plantinga, L. C., Hsu, C. Y., Jordan, R., Burrows, N. R., Hedgeman, E., Yee, J., Saran, R., & Powe, N. R. (2011). Chronic kidney disease awareness among individuals with clinical markers of kidney dysfunction. Clinical Journal of the American Society of Nephrology, 6(8), 1838-1844. https://doi.org/10.2215/CJN.00730111