| Course | HS840 Advanced Health Research and Analysis Methods |
|---|---|
| Unit | Unit 5 |
| Paper type | Sampling plan |
| Length | About 870 words, 6 pages double-spaced |
| Format | APA 7 student paper |
| School | Purdue University Global |
| Program | Doctor of Health Science (DHSc) |
| Updated | October 2026 |
Free sample paper for HS840 Unit 5
Twelve Clinics, Every Request: A Unit 5 Sampling Plan and Stepped-Wedge Power Calculation for a Same-Day Contraception Study
[Student Name]
Purdue University Global
HS840: Advanced Health Research and Analysis Methods
Unit 5 Sampling Plan
[Instructor Name]
[Date]
The network and its local figures are composite and shown as placeholders; the power method and its sources are real, and the calculation was run as described.
Level One: Clinics
All twelve clinics in the network are included. No sampling of clinics is needed, because the rollout reaches every clinic and the design depends on all of them crossing over. To allow a replication elsewhere to select comparable sites, the criteria that define these clinics are stated: general primary care clinics offering family medicine or internal medicine, with at least two clinicians credentialed to perform procedures, an on-site supply room, and a shared electronic health record. Clinics range from [4] to [11] clinicians and see between [6,000] and [21,000] patients a year.
Level Two: Patients
The patient population is every request, not a recruited sample. Because the primary outcome comes from records, the study does not depend on clinic staff to invite patients, which avoids the recruitment bias that can arise when clinics that know their allocation choose whom to enroll.
Inclusion. A patient enters the study at a request: a visit or message in which the patient asks for, or accepts a clinician's offer of, an implant or IUD, documented in the record by a structured order or a coded visit reason. Patients aged 15 to 49 at the time of request are included.
Exclusions, each with its reason:
1. Requests for removal or replacement only, since the question concerns new insertions.
2. Patients who are known to be pregnant at the request visit, since insertion is contraindicated.
3. Requests made at a postpartum or post-abortion visit in another service and referred to primary care, since those pathways follow separate protocols.
4. Patients whose only visit is with a visiting clinician from outside the network, since their follow-up cannot be traced.
A patient who makes a second request more than 90 days after the first is counted as a new request. Repeat requests within 90 days are folded into the first.
Counting every request from records, rather than recruiting patients, removes the one decision that clinic staff who know their allocation would otherwise make.
Who Is Told What
Clinicians are informed of the study and its schedule through clinic meetings led by the research team, separately from the randomization of crossover order, which is generated by a statistician with no role in clinics. Clinicians are told when their clinic will cross over two months in advance, to allow training. Patients receive routine care; the Unit 9 ethics memo addresses notification and the waiver of individual consent for records-based outcomes.
Size: Inputs and Their Sources
The design has 12 clinics, seven two-month periods, one baseline period and six steps of two clinics each, following the model of Hussey and Hughes (2007). Power was computed with their closed-form variance for a stepped-wedge design, treating the binary outcome with its binomial variance. Each input has a source.
1. Baseline proportion completing insertion within 30 days: [55] percent, from the network's records for the prior year, consistent with the 54 percent completion reported under a two-visit protocol in another setting (Bergin et al., 2012).
2. Target proportion: 70 percent, a 15-point increase that leadership considers the smallest change worth the training cost.
3. Requests per clinic per two-month period: [9], from the network's scheduling data, about [650] a year.
4. Intracluster correlation: [0.05], estimated from the prior year's clinic-level completion rates.
5. Two-sided significance level of 0.05.
Result
Under these inputs, power to detect a 15-point increase is about 0.76, just under the conventional 0.80. Power reaches about 0.81 for a 16-point increase and about 0.86 for a 17-point increase. Adding an eighth period at the end, so all clinics contribute another two months under the intervention, raises power for 15 points only to about 0.78, because those added periods contain no contrast between exposed and unexposed clinics.
Power was also checked across plausible intracluster correlations. For a 15-point increase, it ranges from about 0.83 at an intracluster correlation of 0.01 to about 0.75 at 0.10; in a stepped-wedge design, within-clinic comparisons protect power against larger correlations, so the result is not very sensitive.
Decision
The plan proceeds with the fourteen-month schedule and reports a minimum detectable difference of about 16 percentage points at 80 percent power. A smaller true effect might be missed, and the methods chapter will say so. Two options to recover power were considered and rejected: lengthening every period, which would delay the rollout leadership has scheduled, and adding clinics from outside the network, which would introduce different records systems. The simpler calculation omits correlation that decays over time within clinics, which more recent methods can model (Hemming et al., 2020), and the analysis plan will revisit it.
Limits
All local inputs are estimates from one prior year, and request volume could change once same-day service is advertised. The plan will recompute power after the baseline period with observed values, without changing the design.
References
Bergin, A., Tristan, S., Terplan, M., Gilliam, M. L., & Whitaker, A. K. (2012). A missed opportunity for care: Two-visit IUD insertion protocols inhibit placement. Contraception, 86(6), 694-697. https://doi.org/10.1016/j.contraception.2012.05.011
Hemming, K., Kasza, J., Hooper, R., Forbes, A., & Taljaard, M. (2020). A tutorial on sample size calculation for multiple-period cluster randomized parallel, cross-over and stepped-wedge trials using the Shiny CRT Calculator. International Journal of Epidemiology, 49(3), 979-995. https://doi.org/10.1093/ije/dyz237
Hussey, M. A., & Hughes, J. P. (2007). Design and analysis of stepped wedge cluster randomized trials. Contemporary Clinical Trials, 28(2), 182-191. https://doi.org/10.1016/j.cct.2006.05.007
The Purdue Global HS840 Unit 5 instructions
In the fifth unit, learners usually describe whom the study will include at every level, give inclusion and exclusion rules with reasons, describes recruitment and justifies sample size with a power analysis. Five to seven pages is typical. For clustered designs, both levels need definitions: the clusters and the individuals within them. Instructors expect each power input to come from a stated source, such as prior data, a pilot or published studies, and the method of calculation to be named. Sensitivity checks across plausible values of key inputs are often required. If the calculation falls short, most prompts expect the learner to say so and explain the options considered rather than altering assumptions until the result looks acceptable.
How this HS840 Unit 5 example is built
Clinics are defined by type, staffing, procedure capacity and shared records so a replication could find similar sites. Patients enter through documented requests, with four exclusions and a rule for repeat requests within 90 days. The highlighted sentence links the records-based population to the recruitment bias found in the earlier appraisal. A section explains who tells clinicians about the study and when. Five numbered inputs follow, each with its source. Results report power at 15, 16 and 17 points, the small gain from adding a final period and how power varies across intracluster correlations. The decision section accepts a 16-point minimum detectable difference and explains why two ways of recovering power were rejected.
HS840 Unit 5 rubric: what earns full points
Sampling plans tend to be assessed on the clarity of population definitions, the justification of criteria, the recruitment procedure, the rigor and transparency of the sample size calculation, and writing. Population definitions earn points when they can be applied without the author present. Criteria score well when each exclusion has a reason linked to the question. Recruitment points reward procedures that reduce bias, especially in cluster designs where staff know the allocation. Sample size is often the largest category, and rubrics reward named methods, sourced inputs, sensitivity analysis and an honest statement of what the study can detect. Clear numbered inputs and a short limitations section help organization scores.
HS840 Unit 5 help: mistakes that cost points
The most frequent weakness is a power calculation that ignores clustering, which can overstate power badly in a clinic-level design. Another is inputs pulled from nowhere, with no source for the baseline rate or correlation. Some learners keep raising the assumed effect until power reaches 80 percent, which instructors spot quickly. Others define the population so loosely that two people would select different patients. Recruitment plans that let clinic staff choose whom to enroll invite bias. If your study uses another outcome or design, perhaps a continuous score in a parallel trial, send your question and available data, and we will prepare a plan whose calculation shows every input openly.
Get an HS840 Unit 5 example written to your instructions
Power depends on the design, so include the question and design already settled, the Unit 5 prompt, the rubric, and any pilot figures or prior-year data. Eligibility becomes criteria, recruitment stays separate from assignment, and the size is justified by script with every input sourced. First custom samples are free; turnaround is 24-48h. 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 HS840 papers and DHSc samples
- HS840 Unit 1: Research Question Discussion Post
- HS840 Unit 2: Research Question Critique Paper
- HS840 Unit 3: Cluster Trial Methods Appraisal
- HS840 Unit 4: Stepped-Wedge Design Justification
- HS840 Unit 6: Instrument Review
- HS840 Unit 7: Seminar Reflection
- HS840 Unit 8: Data Analysis Plan
- HS840 Unit 9: Ethics and Consent Memo
- HS840 Unit 10: Methods Chapter Draft
- HS830 Unit 5: Autism Prevalence Seminar Reflection
- HS820 Unit 5: Kerala and Uttar Pradesh Burden Review
- HS810 Unit 5: Policy Alternatives Memo
HS840 Unit 5 questions, answered
What if the calculation says I need more clinics than exist?
Then say so, and consider changing what can change: a longer study, more steps, a more precise outcome, or a larger detectable effect. If none suffices, the study may be underpowered for its primary outcome and should be framed as a feasibility or estimation study. Reporting that honestly is better than adjusting assumptions until the number fits.
Where can I find a free HS840 Unit 5 sample paper?
The complete Unit 5 sampling plan is on this page, defining clinics and patients for a same-day contraception trial and reporting stepped-wedge power with every input sourced.
Which software can run a stepped-wedge power calculation?
Several work, including R packages, Stata commands and the Shiny CRT Calculator. The sample uses the Hussey and Hughes closed-form formula and lists every input so any tool can reproduce it.
Where do the power inputs come from in this plan?
The baseline completion rate, request volume and intracluster correlation come from the network's prior-year records, and the target reflects the smallest rise leadership said would justify training.
Why not recruit patients directly?
Counting every documented request from records avoids letting clinic staff, who know their clinic's allocation, decide whom to enroll, a bias identified in the earlier appraisal.