PU525 Unit 6 Statistical Output Interpretation Memo Example

What 5.8 Millimeters of Mercury Means: An Output Interpretation Memo Reading Software Output From a Composite County's Community Health Worker Hypertension Program for Its Director, Table by Table, in the Units the Director Cares About

[Student Name]

Purdue University Global

PU525: Biostatistics for Public Health

Unit 6 Output Interpretation Memo

[Instructor Name]

[Date]

The program, the output and every figure are composite and bracketed for the writer's own data; the research cited is real.

Memo

To: [Director], Hypertension Program From: [Analyst] Date: [date] Subject: Twelve-month blood pressure results, output attached

The Finding

At twelve months, program participants averaged [132.4] mmHg systolic, compared with [138.2] among comparison residents, a difference of [5.8] mmHg. The data are consistent with a true difference anywhere from about [2.6] to [9.0] mmHg. At baseline the two groups were similar, [145.1] and [143.9].

Walking Through the Output

Group statistics table. The first block lists each group's number of people, mean, standard deviation and standard error of the mean. Participants: [118] people, mean [132.4], standard deviation [12.5]. Comparison: [124] people, mean [138.2], standard deviation [12.7]. The standard deviations are similar, which matters for the next block. Test of equal variances. The second block tests whether the spread of blood pressure is the same in both groups. Its significance value, [0.41], is large, so the output's first row, which assumes equal variances, is the one to read. t-test row. The test statistic is [3.57] with [240] degrees of freedom. The two-tailed significance, shown as [.000], means the P value is below [0.001], not zero. The mean difference is [5.8] mmHg with a standard error of [1.62], and the 95 percent confidence interval runs from [2.6] to [9.0] mmHg. Every number in the output can be read in millimeters of mercury, and that is how the board will understand it.

What the Difference Means Clinically

A reduction of about six points in systolic pressure is meaningful at the population level. A large meta-analysis of trials found that each [10] mmHg reduction in systolic pressure was associated with roughly a fifth fewer major cardiovascular events (Ettehad et al., 2016). A [5.8] mmHg difference, if sustained and caused by the program, would be expected to prevent a share of strokes and heart attacks among participants. Community health worker programs for hypertension have shown improvements in blood pressure control in earlier reviews, so the size of the effect is plausible (Brownstein et al., 2007).

How the Data Were Collected

Participants were adults with uncontrolled hypertension who enrolled in the program at two clinics between [months]. Comparison residents were adults with uncontrolled hypertension at two similar clinics in the county that did not yet offer community health workers, matched on age group and sex. Blood pressure was measured by clinic staff using automated cuffs, two seated readings averaged after a rest period, at enrollment and twelve months later. People who moved or missed the twelve-month visit, [21] participants and [19] comparison residents, are not in the analysis; their baseline pressures were similar to those who remained, which reduces but does not remove concern about who dropped out.

Reading a P Value Shown as Zero

Software often prints a very small P value as [.000]. This does not mean there is no chance the result arose by chance; it means the value is smaller than the software displays. The correct report is that P is below [0.001]. More important than the P value is the interval, which shows the range of plausible differences in the units that matter.

What I Would Tell the Board

The program's participants ended the year with blood pressure about six points lower than similar residents without the program. That is a promising result for a program at this cost, and a second year of data, with a stronger comparison, will show whether it holds.

What the Output Cannot Tell You

The comparison group was not randomized. Residents who enrolled may have been more motivated or had better access to care, which could account for part of the difference. Baseline similarity is reassuring but does not rule this out. The analysis also compares follow-up values only; an analysis adjusting for each person's baseline, which I can run next, would give a more precise estimate.

Recommendation

Report the result as a [5.8] mmHg lower average systolic pressure among participants, with the interval, and describe it as encouraging rather than proof of effect. Continue the comparison for a second year and add the baseline-adjusted analysis.

Attachment Notes

The attached output is unedited. Highlighted rows correspond to the numbers in this memo. Rows not discussed, such as the second line of the t-test table that assumes unequal variances, give nearly identical results and can be ignored for this question (Rosner, 2016).

References

Brownstein, J. N., Chowdhury, F. M., Norris, S. L., Horsley, T., Jack, L., Zhang, X., & Satterfield, D. (2007). Effectiveness of community health workers in the care of people with hypertension. American Journal of Preventive Medicine, 32(5), 435-447. https://doi.org/10.1016/j.amepre.2007.01.011

Ettehad, D., Emdin, C. A., Kiran, A., Anderson, S. G., Callender, T., Emberson, J., Chalmers, J., Rodgers, A., & Rahimi, K. (2016). Blood pressure lowering for prevention of cardiovascular disease and death: A systematic review and meta-analysis. The Lancet, 387(10022), 957-967. https://doi.org/10.1016/S0140-6736(15)01225-8

Rosner, B. (2016). Fundamentals of biostatistics (8th ed.). Cengage Learning.