HD520 Unit 2 Design Rationale Paper Example

A Waitlist the Agency Already Had: A Unit 2 Design Rationale for Assessing Whether a Refugee Health Orientation Works

[Student Name]

Purdue University Global

HD520: Health Education Program Assessment

Unit 2 Design Rationale Paper

[Instructor Name]

[Date]

The agency, the funder, the class and all figures are composite; the design texts, the trial-ethics and interpreter studies cited are real.

Purpose of the Assessment

The funder's renewal letter asks one question: "Can graduates of the health orientation use primary care independently within their first [90] days in Calder City?" Restated as a decision, the agency must choose whether to continue the class as it is, revise it or replace it with one-to-one coaching by caseworkers. An assessment design is useful here only if its results could move that decision. A design that shows participants liked the class, for example, would leave all three options open.

The class runs six sessions over three weeks and covers finding a clinic, booking an appointment through a telephone interpreter, bringing the right documents, understanding a prescription label and knowing when to use urgent care rather than the emergency department. Newly arrived adults are referred by caseworkers in their first week.

Candidate Designs

Five designs were considered. They are compared below on what each could establish, what it would cost the agency and whether it is acceptable for this population.

Design A, post-class questionnaire: establishes satisfaction only; low cost; acceptable. Design B, single-group pretest and posttest: establishes change over the class period; moderate cost; acceptable. Design C, historical comparison with adults who arrived before the class existed: establishes a difference between eras; low cost; acceptable but weak. Design D, randomized assignment to immediate or delayed class: establishes a causal effect; moderate cost; not acceptable here. Design E, comparison of adults who start immediately with adults who wait for a seat because classes are full: establishes a difference under stated assumptions; moderate cost; acceptable.

Why Four Designs Were Rejected

Design A was rejected first. The agency's current survey already shows how little a post-class rating tells a funder: it measures how participants felt about the class, collected by the person who taught it, and it cannot show whether anyone can book an appointment. Asking better questions would not fix the core problem, which is that the decision depends on ability, not opinion.

Design B looks stronger, since it measures participants before and after. But newly arrived adults learn quickly from everything around them in their first weeks: caseworkers, neighbors, relatives already in the city and their first clinic visit. A single group's improvement over three weeks cannot be separated from what they would have learned anyway. Shadish et al. (2002) describe this as maturation and history operating together, and in a resettlement setting both are strong.

Design C would compare current graduates with adults who arrived two years earlier, before the class began. Those groups differ in country of origin, in the clinics open to them and in the interpreter services the city funded at the time. Any difference between eras would mix the class's effect with all of those changes.

Design D, randomizing adults to start the class at once or after a delay, would give the cleanest estimate. It was refused. Deliberately delaying an orientation for people with pressing health needs, solely to create a comparison, would ask participants to carry a cost so the agency can learn something, and research ethics guidance warns against assigning people to a condition expected to be worse when equipoise is doubtful (Freedman, 1987). Several agency staff also said that newly arrived families would struggle to see such a delay as anything other than unfair.

The Chosen Design

Design E uses a delay the agency already has. Classes are limited to twelve adults so that interpreters can keep up, and during busy months new arrivals wait about six weeks for a seat. That waitlist exists whether or not an assessment is run. Every adult takes the same performance task within ten days of arrival and again on day [42], when early starters have finished the class and those waiting have not yet begun.

The comparison costs participants nothing they would not already have lost, because the delay comes from the agency's capacity and would happen whether or not anyone measured it.

The outcome is a simulated task rather than a self-report. Each participant calls a clinic scheduler, played by a trained assessor, through the agency's contracted telephone interpreter line, and asks for a new-patient appointment. A ten-item checklist scores the call. Self-report was rejected as the main outcome for the same reason Design A was: people tend to say they can do what they were just taught. Interpreter use is built into the task because professional interpreters improve communication and care for patients with limited English proficiency (Karliner et al., 2007), so using the line is a skill the class teaches and the funder cares about.

Threats to Validity

Design E has weaknesses that a randomized design would not. The two groups are not formed by chance, so they may differ in ways that affect the outcome. Three threats are named here and carried into later units.

Selection by arrival timing. Who waits depends on when they arrive, and arrivals from one country often come in groups. If one language group arrives mainly during busy months, the waiting group will differ in language and background. Response: compare the groups at baseline and adjust for language.

Testing. Both groups take the task twice, and the first attempt may teach something. Because both groups are tested on the same schedule, this effect should affect them equally, but it could add to the gains in both.

History. Something outside the class, such as a clinic changing its scheduling line, could affect one group more than the other if it happens during a busy month. Response: record such events in an assessment log.

Conclusion

The chosen design cannot prove that the class causes better use of primary care, but it can tell the agency whether adults who finished the class perform better than comparable adults who have not taken it, at the same point after arrival. That answer bears directly on the decision to continue, revise or replace the class. If results show a large gain, the agency keeps the class and may add seats so fewer adults wait; if the gain is small, it will test caseworker coaching against the class the following year. Either way, the funder receives an answer about the ability to use care, the question it actually asked.

References

Freedman, B. (1987). Equipoise and the ethics of clinical research. New England Journal of Medicine, 317(3), 141-145. https://doi.org/10.1056/NEJM198707163170304

Karliner, L. S., Jacobs, E. A., Chen, A. H., & Mutha, S. (2007). Do professional interpreters improve clinical care for patients with limited English proficiency? A systematic review of the literature. Health Services Research, 42(2), 727-754. https://doi.org/10.1111/j.1475-6773.2006.00629.x

Shadish, W. R., Cook, T. D., & Campbell, D. T. (2002). Experimental and quasi-experimental designs for generalized causal inference. Houghton Mifflin.