HCM 440 Module 6 Analysis and Ethics Short Paper Example

From Data to Decisions, Responsibly: Analysis Plan and Ethical Safeguards for the Fairhaven Reminder Study

[Student Name]

Southern New Hampshire University

HCM 440: Healthcare Research and Evaluation Methodologies

Module Six Short Paper

[Instructor Name]

[Date]

The organization, setting and figures below are a composite written as a model document. No real employer, client, colleague or patient is described.

From Data to Decisions, Responsibly: Analysis Plan and Ethical Safeguards for the Fairhaven Reminder Study

The Fairhaven reminder study now has a design and measures. Two questions remain before the full proposal: how the numbers and transcripts will become answers, and what safeguards surround everyone whose information or time the study uses. This paper matches each analysis to its data type and question, explains how results will be judged for practical importance and then addresses ethical review, the Belmont principles, privacy and the particular risks of interviewing colleagues.

Describing the Data First

Analysis begins with description. For each clinic, the team will report the number of eligible appointments, patient age as a mean and standard deviation, the share preferring Spanish and the payer mix. Monthly no-show rates will be plotted for pilot and comparison clinics across the full period, with medians and ranges reported by specialty. Plots often reveal problems, such as a sudden jump in one clinic caused by a coding change, that formal models would hide.

Checking Trends Before the Change

The comparison between pilot and phone clinics is only fair if their no-show rates moved in step before March 2026. The team will plot the two groups' monthly rates over the preceding 24 months and fit a model testing whether the slopes differ. If the pilot clinics were already improving faster, the main analysis will adjust for their separate prior trend, and the report will say plainly that the comparison is weaker.

Part One: Controlled Segmented Regression

Wagner et al. (2002) describe segmented regression, which estimates the level and slope of a series before an intervention and the change in each afterward. For part one, monthly no-show counts in each clinic will be modeled with the number of eligible appointments as an offset, so results are expressed as rates. The model will include time, a post-change indicator, time since the change, the same terms interacted with pilot status and month indicators for seasonality. Lopez Bernal et al. (2017) recommend count models such as Poisson or negative binomial regression for this kind of outcome and checking residuals for autocorrelation, both of which the plan adopts.

Part Two: Stepped Wedge Analysis

In the stepped wedge, each appointment's outcome is a yes-or-no event, and appointments are clustered within clinics. Hemming et al. (2015) recommend models that account for this clustering and for time, since later periods have more clinics using the new practice. The analysis will use mixed-effects logistic regression with a random effect for clinic, fixed effects for each two-month period and an indicator for whether the clinic had switched. Adjusting for period separates the program's effect from background change across the group.

Secondary and Subgroup Analyses

The refill and late cancellation rates will be analyzed with the same models. Subgroup questions will be tested by adding interaction terms between reminder type and age group, preferred language and specialty. Because testing many subgroups raises the chance of false findings, only these three were chosen in advance, and results will be described as exploratory. Covariates identified in the measurement plan, such as lead time and prior no-shows, will be included in appointment-level models.

Analyzing the Survey

A chi-square test will show whether satisfied patients are more common in text clinics than in phone clinics, and mean scores on the eight-item scale with an independent samples t-test, after checking that score distributions are roughly symmetric. Cronbach's alpha will be reported for the final scale. Spanish and English responses will be summarized separately as well as together, so any difference in experience by language is visible rather than averaged away.

Analyzing the Interviews

Braun and Clarke (2006) set out six phases for thematic analysis: reading transcripts until they feel familiar, tagging passages with short codes, grouping codes into candidate themes, testing those themes against the transcripts, settling each theme's name and scope and finally writing up. Two coders, the analyst and a nurse manager from outside the study clinics, will code the first four transcripts independently, compare codes and agree on a codebook before coding the rest. Themes will be checked against the whole data set, and quotations will show both typical and contrasting views.

Judging Practical Importance

A statistically significant result is not automatically worth acting on. Results will be reported as estimated differences with 95% confidence intervals, not only p-values. Before seeing the data, leadership agreed that a reduction of at least 2 percentage points in weekday no-shows would justify the $140,000 annual cost, since it would recover roughly 4,800 visits a year. If the confidence interval lies entirely above that threshold, the case is strong; if it straddles it, the decision will weigh costs and uncertainty openly.

Table 1. Analysis Plan by Question

Note. Prespecified by the author before data extraction.

Quality Improvement or Research?

The study uses routine data to improve operations, which sounds like quality improvement, yet it randomizes clinics and will be shared outside Fairhaven, which sounds like research. Lynn et al. (2007) argued that quality improvement carries real ethical obligations even when it is not formally research, and that oversight should be proportionate to risk rather than depend on a label. Fairhaven will therefore submit the proposal to an external institutional review board and ask it to determine the appropriate level of review. The team expects expedited review with a waiver of individual consent for appointment data, which are already collected and pose minimal risk.

Respect for Persons

The Belmont Report (National Commission, 1979) grounds research ethics in three principles, taken here one at a time. The first, respect for persons, requires informed, voluntary participation where feasible. Patients can already opt out of all reminders and will continue to be able to. Survey invitations will explain the study, state that participation is voluntary and will not affect care and provide a contact for questions. Staff interviews will require written consent.

Beneficence and Privacy

Beneficence requires minimizing harm. The main risk is privacy. A text on a shared or employer-owned phone could reveal that someone sees a psychiatrist. Messages will therefore name only "Fairhaven Medical Group," the date and time, never the specialty or clinician. The vendor has signed a business associate agreement under HIPAA, study extracts will use study numbers instead of names and only the analyst and statistician will hold the linking key. The vendor will have no role in analysis or reporting, avoiding a conflict of interest.

Justice

Justice concerns the fair distribution of burdens and benefits. Choosing the rollout order by lottery prevents managers from placing favored clinics first. Patients without mobile phones will continue to receive calls, and Spanish-language texts will be available from the start, so the study does not widen gaps for groups already more likely to miss visits. Subgroup results by language will be reported even if unflattering.

Protecting Staff Interviewed by a Colleague

The analyst works alongside many front-desk staff, which could make them reluctant to criticize the program or their managers. Interviews will be voluntary and scheduled outside supervisors' view, transcripts will be de-identified and managers will never see individual responses. Quotations will be attributed only by role category. As the COREQ checklist by Tong et al. (2007) advises, the report will describe the analyst's relationship to participants so readers can judge its influence.

Conclusion

Each question in the study now has a matching analysis, a prespecified threshold for practical importance and a plan for reporting uncertainty. The ethical plan treats a quality improvement project with the seriousness of research: independent review, respect for choice, careful message wording, fair rollout and protection for staff. Project Two will bring these pieces together into the complete proposal.

References

Braun, V., & Clarke, V. (2006). Using thematic analysis in psychology. Qualitative Research in Psychology, 3(2), 77-101. https://doi.org/10.1191/1478088706qp063oa

Hemming, K., Haines, T. P., Chilton, P. J., Girling, A. J., & Lilford, R. J. (2015). The stepped wedge cluster randomised trial: Rationale, design, analysis, and reporting. BMJ, 350, Article h391. https://doi.org/10.1136/bmj.h391

Lopez Bernal, J., Cummins, S., & Gasparrini, A. (2017). Interrupted time series regression for the evaluation of public health interventions: A tutorial. International Journal of Epidemiology, 46(1), 348-355. https://doi.org/10.1093/ije/dyw098

Lynn, J., Baily, M. A., Bottrell, M., Jennings, B., Levine, R. J., Davidoff, F., Casarett, D., Corrigan, J., Fox, E., Wynia, M. K., Agich, G. J., O'Kane, M., Speroff, T., Schyve, P., Batalden, P., Tunis, S., Berlinger, N., Cronenwett, L., Fitzmaurice, J. M., . . . James, B. (2007). The ethics of using quality improvement methods in health care. Annals of Internal Medicine, 146(9), 666-673. https://doi.org/10.7326/0003-4819-146-9-200705010-00155

National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research. (1979). The Belmont report: Ethical principles and guidelines for the protection of human subjects of research. U.S. Department of Health, Education, and Welfare.

Tong, A., Sainsbury, P., & Craig, J. (2007). Consolidated criteria for reporting qualitative research (COREQ): A 32-item checklist for interviews and focus groups. International Journal for Quality in Health Care, 19(6), 349-357. https://doi.org/10.1093/intqhc/mzm042

Wagner, A. K., Soumerai, S. B., Zhang, F., & Ross-Degnan, D. (2002). Segmented regression analysis of interrupted time series studies in medication use research. Journal of Clinical Pharmacy and Therapeutics, 27(4), 299-309. https://doi.org/10.1046/j.1365-2710.2002.00430.x