Turning a clinical curiosity into an answerable research question

START HERE

The essential idea

Translate a clinical curiosity into a bounded question and estimand. Operationalize an outcome without changing the original construct. Produce a feasible methods outline and identify a claim it cannot support.

BUILD THE REASONING · 01

Keep the important question while making it answerable

A clinical curiosity often begins broadly: “Why do people disappear after an intake?” That question may contain several different problems. Some people cannot find an appointment, some decide the service does not fit, some obtain help elsewhere, and some encounter fees or transportation barriers. Calling all of these “dropout” can create a measure that is easy to count but impossible to interpret. First describe the decision the research is intended to inform and whose experience needs to be understood.

A bounded quantitative question could ask whether offering a choice of appointment times, compared with the current scheduling process, changes attendance at a first follow-up within 30 days among new adult clients at a specified clinic. A qualitative question could ask how clients experience the transition from intake to ongoing care. The first estimates a contrast; the second investigates meaning and process. Neither replaces the other, and combining them requires a reason and an integration plan rather than simply collecting more information.

For an intervention question, a population–intervention–comparison–outcome–time structure can expose missing elements. It is a drafting aid, not a rule that every study must be an experiment. For an exploratory qualitative question, forcing an intervention/comparator structure may distort the inquiry. Choose the structure that makes the intended claim visible, then examine whether the proposed evidence can actually support that claim.

BUILD THE REASONING · 02

Define the variable before collecting it

Operationalization links a construct to a procedure. “Engagement” might mean attendance, active participation, perceived fit, completion of an agreed task, or continuity of care. Those are not interchangeable outcomes. In the scheduling study, attendance is observable but cannot by itself show whether the client benefited or felt respected. Name it accurately as attendance, and add a separate experience or functioning outcome only when its purpose and collection burden are justified.

Write a codebook with the unit, source, permitted values, time window, and handling of ambiguity. If a client chooses another provider before follow-up, that is not necessarily a failure of engagement; retain a transfer category when available and authorized. If the appointment is canceled by the clinic, do not code it as a client no-show. Preserve dates needed to determine opportunity for attendance, while avoiding collection of unnecessary personal detail.

Identify the estimand: for example, the difference in the proportion attending within 30 days under an offer of scheduling choice versus the usual offer, among eligible enrolled clients. This is different from the effect among only people who accepted a particular time or completed the study. Define how interruptions, transfers, and missing outcomes enter the question before analysis. A clear estimand prevents an apparently small exclusion from changing what the study answers.

BUILD THE REASONING · 03

Make the design feasible without weakening the claim silently

If ethically and operationally appropriate, random assignment to two acceptable scheduling processes can improve the comparison. Both conditions must preserve necessary care and the institution’s standards. If the whole clinic changes process on one date, a simple before/after comparison may be feasible but introduces seasonal, staffing, and referral-mix explanations. Repeated observations and an appropriate comparison service may help, but their assumptions must be stated rather than hidden under the word evaluation.

Plan recruitment and power with the outcome and a meaningful difference in mind. A clinic that sees 12 eligible clients each month may not support a precise small-effect estimate within a short student project. A feasibility study can instead examine whether the offer can be delivered, how clients understand it, and whether outcome collection works. That is a legitimate narrower question when explicitly chosen; it should not be described afterward as a definitive efficacy trial.

Prepare a literature matrix before asserting a gap: study question, setting/population, design, intervention, outcome, key result, limitations, and relevance to this proposal. An article about reminder texts is related to scheduling choice but does not test the same mechanism. Explain what existing work suggests and what remains uncertain. Search and report enough detail for another reader to understand why the question follows from the evidence.

BUILD THE REASONING · 04

Write an executable methods paragraph

A methods paragraph should let a reader identify who can enter, who invites them, what happens in each condition, what data are collected and when, who scores/extracts outcomes, and how the comparison will be analyzed. Add a participant-flow plan and a missing-data plan. An intention-to-treat approach generally seeks to preserve the assigned comparison, but naming it does not make unobserved outcomes disappear; describe how missingness will be investigated and how sensitive conclusions are to assumptions.

Address consent, privacy, power, and oversight before implementation. A clinician recruiting their own clients can create pressure even when the intervention seems minor. Ask the institution for the required determination and safeguards, and do not infer that routine records are freely available for research. Limit the collected fields and define access, retention, and reporting protections. This lesson supplies a fictional proposal exercise, not approval to collect actual client information.

End the proposal with a decision rule and its limits: what finding would justify continuing the process, revising it, or obtaining a larger study? For the scheduling example, feasibility and acceptability may be prerequisites before testing an attendance effect. If appointment capacity is inadequate, the study may reveal that offering choice cannot solve the underlying supply problem. A good research question remains connected to a real decision even when the answer is that the original intervention targets the wrong mechanism.

FROM CONCEPT TO DECISION

From “dropout” to a testable scheduling question

1. 1

   Clarify the decision and mechanism

   Bea wants to decide whether appointment choice improves continuity. Combining three changes would make a positive result difficult to attribute to choice, so the proposal first defines the scheduling offer and documents usual care.

2. 2

   Define attendance and opportunity

   The study records whether a feasible slot was offered, scheduled, attended, canceled by the clinic, or replaced by a transfer. The primary 30-day attendance outcome remains distinct from those explanatory process categories and from satisfaction.

3. 3

   Choose a feasible comparison

   The clinic considers assignment between two acceptable scheduling procedures after oversight review. If that is infeasible, the design and causal claim must change together; a before/after dataset cannot simply retain the trial’s strongest causal wording.

4. 4

   Specify what would revise the intervention

   If most nonattendance occurs because no acceptable slots exist, the next task concerns capacity rather than reminders. If choice is offered inconsistently, implementation must be repaired before judging the proposed mechanism.

MAKE THE REASONING YOURS

Work through a new situation

READ WITH A PURPOSE

Best references

These readings deepen the specific models, evidence, and decisions in this lesson.

1. 01
   APA: Journal Article Reporting Standards

   Use design-specific reporting requirements to identify what a question, method, analysis, and limitations section must make inspectable.

2. 02
   CONSORT 2025: Reporting Randomised Trials

   Read the original guidance on trial questions, allocation, outcomes, participant flow, and analysis before using a randomized design label.

3. 03
   HHS OHRP: Human-Subjects Regulations

   Locate the official framework for institutional oversight and participant protections; obtain the appropriate determination for a real project.

4. 04
   AHRQ: Elements of an Evaluation Plan

   Connect a practical decision to answerable questions, criteria, feasible data collection, and intended use of findings.