A Canadian medical thesis research proposal has eight sections: background, research question and hypothesis, objectives, literature gap, study design, ethics and REB considerations, statistical plan, and significance with timeline. Below is a full worked example built around a fictional clinical study on point-of-care ultrasound for suspected deep vein thrombosis, with a reading after each part explaining why it works.
A note before you read further. Every figure, statistic, and finding below is a placeholder marked [figure] or [cite]. This is a shape to copy, not a clinical claim to submit — verify every named test, statute, or figure you actually use against a source you open yourself.
What does a Canadian medical thesis proposal need to contain?
A clinical or clinical-epidemiology master’s or MD thesis proposal is structured around eight questions in a fixed order; its overall length is set by your program.
- Background — what is the clinical problem, and why does current practice leave it unresolved?
- Research question and hypothesis — what, precisely, will the study test?
- Objectives — the primary and secondary outcomes the study is powered or designed to detect.
- Literature gap — what has already been published, and what specifically is missing.
- Study design — the design type and why it fits the question.
- Ethics and REB considerations — approval body, consent process, and any trial-registration requirement.
- Statistical plan — the analysis and, for a quantitative design, the sample-size justification.
- Significance and timeline — why the answer changes practice, and when the work will be done.
If you have not yet chosen your design, the full designs-by-field comparison is in how to write a methodology chapter; the proposal below assumes a design has already been selected.
What does the background section look like?
“Deep vein thrombosis (DVT) is a common emergency-department presentation, and the diagnostic pathway typically requires a formal compression ultrasound performed by radiology, which is not always available outside regular hours [cite]. Point-of-care ultrasound (POCUS), performed by the treating emergency physician at the bedside, has been proposed as a faster alternative, but its diagnostic accuracy compared with formal radiology-performed ultrasound in the Canadian emergency-department context, and its effect on actual time-to-diagnosis, is not well established [cite]. This proposal examines whether POCUS reduces time-to-diagnosis for suspected DVT without a meaningful loss of diagnostic accuracy, in a Canadian tertiary emergency department.”
Why it works: the paragraph states the clinical workflow problem in one sentence, names the proposed solution, and states plainly what is not yet known — accuracy and time-to-diagnosis in this specific setting — rather than implying POCUS is untested everywhere, which it is not.
What does the research question and hypothesis look like?
“Research question: Among adult emergency-department patients with suspected lower-limb DVT, does point-of-care ultrasound performed by a trained emergency physician reduce time-to-diagnosis compared with formal radiology-performed compression ultrasound, without a clinically meaningful reduction in diagnostic accuracy? Primary hypothesis: median time-to-diagnosis will be shorter in the POCUS arm than the radiology-ultrasound arm. Secondary hypothesis: POCUS sensitivity and specificity, against the radiology scan as reference standard, will fall within a pre-specified non-inferiority margin of [figure] percentage points.”
Why it works: stating a non-inferiority margin up front — rather than simply hoping accuracy “looks similar” — is what turns a vague comparison into a testable hypothesis a statistician can actually power the study against, and it is a common gap in first-draft clinical proposals. For more worked hypothesis examples across other fields, see research questions and hypotheses worked examples.
What do the objectives look like?
- Primary objective: to compare median time-to-diagnosis between POCUS and formal radiology-performed compression ultrasound for suspected DVT.
- Secondary objective 1: to compare POCUS sensitivity and specificity against radiology-performed ultrasound as the reference standard.
- Secondary objective 2: to describe emergency-department length-of-stay for patients diagnosed via each pathway.
Why it works: each objective maps to exactly one outcome the results chapter will report, in the same order, so a reader moving from proposal to eventual thesis sees no objective introduced late and no result appear that was not promised.
What does the literature-gap section look like?
“POCUS diagnostic accuracy for proximal DVT has been studied in several emergency-medicine settings internationally, with pooled sensitivity and specificity estimates reported in a systematic review [cite]. However, that review’s included studies were drawn predominantly from academic centres with dedicated POCUS training programs, and none were conducted in a Canadian tertiary emergency department where after-hours radiology access, physician POCUS certification requirements, and patient population may differ meaningfully [cite]. The student’s own search found no published Canadian study measuring the effect of POCUS on actual time-to-diagnosis, as distinct from accuracy alone [cite the search strategy], which is the operationally relevant outcome for department flow and, plausibly, for downstream anticoagulation decisions. This proposal addresses both gaps: Canadian-setting accuracy data and a time-to-diagnosis comparison in this setting.”
Why it works: the gap is stated as a precise intersection — accuracy data exists, but not from this setting, and time-to-diagnosis has not been measured in this specific way — rather than a vague “more research is needed,” which is what a clinical proposal committee is trained to distrust.
What does the study-design section look like?
“Design: a prospective, single-centre diagnostic-accuracy and time-motion study with a non-inferiority framework for the accuracy comparison. Setting: the emergency department of [named tertiary hospital], which has both after-hours radiology access and POCUS-certified emergency physicians. Population: adult patients ([figure]+ years) presenting with clinical suspicion of lower-limb DVT, excluding patients with [exclusion criteria — prior DVT in the same limb, contraindication to either scan]. Intervention/comparison: POCUS performed by a certified emergency physician, compared with formal radiology-performed compression ultrasound as the reference standard, performed on the same patient in randomized order to avoid a sequence effect on time measurement. Primary outcome: time from ultrasound order to result availability, compared between pathways. Secondary outcomes: POCUS sensitivity/specificity against radiology ultrasound; emergency-department length of stay.”
Why it works: randomizing the order of the two scans on the same patient — rather than simply comparing historical time data from before and after a POCUS program started — controls for the most obvious confound (a POCUS program that started later might simply coincide with a generally faster department), which is exactly the kind of design decision a methodology committee checks for. If you have not yet worked out your target sample size, see how many participants you need for a thesis before finalizing this section.
What does the ethics and REB section look like?
“This study requires Research Ethics Board approval under TCPS 2 (2022), as it involves human participants and prospective data collection beyond standard clinical care (the research-order compression ultrasound, when not otherwise clinically indicated). Approval will be sought from [named hospital/university] REB prior to enrolment. Given the diagnostic nature of the comparison and that participants undergo two scans rather than the usual one, informed consent will be obtained from all participants or their substitute decision-maker before enrolment, and the additional scan’s minimal-risk nature (no radiation, no contrast, non-invasive) will be stated plainly in the consent form. As this is an observational diagnostic-accuracy study of an already-available imaging method rather than a trial of an investigational drug or device, no Health Canada clinical-trial authorization is anticipated [cite the applicable regulatory guidance and confirm with the REB]; the study will nonetheless be registered on a public trial registry (e.g., ClinicalTrials.gov) for transparency.”
Why it works: the paragraph states explicitly why the extra scan is minimal risk (a real ethics-board question for any design adding a procedure beyond standard care) and clarifies the regulatory and trial-registration question rather than leaving it ambiguous, which is one of the more common oversights in a first-draft clinical proposal.
What does the statistical plan look like?
“Sample size was calculated for the non-inferiority accuracy comparison: assuming radiology-ultrasound sensitivity of [figure]%, a non-inferiority margin of [figure] percentage points, 80% power, and one-sided α = .025, a sample of [figure] patients is required. Time-to-diagnosis will be compared using [named non-parametric test, appropriate for skewed time data] given the expected right-skew of time measurements. Sensitivity and specificity will be reported with 95% confidence intervals against the pre-specified non-inferiority margin. All analyses will follow an intention-to-diagnose principle, retaining patients in their randomized scan-order group regardless of protocol deviations.”
Why it works: naming the specific test for skewed time data, rather than defaulting to a t-test, shows the statistical plan was built around the actual shape of the expected data rather than copied from a template — and stating the analysis principle (intention-to-diagnose) up front, before any data exists, is what keeps the eventual results chapter honest about protocol deviations rather than quietly excluding them.
What five things do proposal committees flag in a medical thesis proposal?
| What committees flag | The fix |
|---|---|
| An accuracy comparison with no pre-specified non-inferiority margin | State the margin and its clinical justification before data collection begins |
| A before/after design presented as if it controls for confounds a randomized comparison would | Randomize where possible; if not, name the confound explicitly and how it will be addressed |
| No stated position on regulatory authorization or trial registration | State explicitly why authorization or registration is or is not required, and cite the applicable rule |
| A statistical test chosen without checking the outcome’s expected distribution | Name the outcome’s likely distribution and pick the test that fits it |
| A consent form that does not explain the added risk or burden of a research-only procedure | State plainly what is beyond standard care and why its risk is minimal |
Ready to turn your clinical question into a full proposal instead of a blank page? Start free with Tesify and work through each section above as your checklist.
Frequently asked questions
Does every medical thesis proposal need a sample-size calculation?
Any proposal with a quantitative hypothesis test does; a purely qualitative or case-based clinical thesis instead needs a stated sampling and saturation rationale rather than a power calculation.
How is a medical thesis proposal different from a grant application?
A thesis proposal is evaluated by your supervisory committee for scientific and methodological soundness; a grant application additionally requires a budget and must persuade a funder the work merits its specific funding priorities — the two documents share content but serve different readers.
Do I need Health Canada approval for every clinical thesis?
No. Health Canada authorization is aimed at clinical trials of investigational drugs, biologics or medical devices; many observational or diagnostic-accuracy studies fall outside it, but confirm what applies to your design with your REB. Researchers do not register studies in Health Canada’s own trial listing — its Canadian Clinical Trial Search Portal, which replaced the older Clinical Trials Database in July 2026, draws on international registries such as ClinicalTrials.gov.
What is a non-inferiority margin, and who decides it?
It is the largest acceptable difference between two methods that would still be considered clinically equivalent; it is set based on clinical judgment and prior literature, ideally in consultation with a biostatistician, before any data is collected.
Can the study design change after the proposal is approved?
Minor refinements are expected as recruitment begins, but a change to the primary outcome, sample size, or comparison group usually requires a formal amendment to both your supervisory committee’s approval and your REB approval.
How detailed does the ethics section need to be at the proposal stage?
Detailed enough to show you have identified every element beyond standard clinical care and its associated risk; the full REB application will go further, but the proposal should already anticipate its major questions. See the general TCPS 2 requirements guide for the baseline process every design must clear.
What is the difference between sensitivity/specificity and a p-value in this context?
Sensitivity and specificity describe how well a diagnostic test performs against a reference standard; a p-value in this design would test whether an observed difference between arms is unlikely to be due to chance — both are reported, but they answer different questions.