Skills

Systematic Review & Literature Synthesis

10 weeks · 0 milestones

Conduct a systematic literature search using PRISMA methodology on a real health question. Proof requires: (a) a PRISMA flow diagram showing search results at each stage (records identified, screened, assessed for eligibility, included); (b) a documented search strategy with database names, search terms, date ranges, and inclusion/exclusion criteria; and (c) a synthesis table summarising included studies (author, year, design, population, intervention, outcome, findings). Reviewed by a researcher or academic with systematic review experience. The health question must be specific and the search must be reproducible from the documented strategy.

Part of: Health & Medicine · Proves for: UG

Milestone map

3 milestones

3–4 weeks (protocol registration + search)

Choose a clinical question using PICO format for a systematic review. Register your review protocol with PROSPERO (the international prospective register of systematic reviews — free registration). Your protocol must include: PICO question, planned databases (minimum 3 — PubMed, Cochrane, and at least one specialist database such as CINAHL, PsycINFO, or Embase), search strategy with full search terms, inclusion and exclusion criteria, primary outcome, and planned data synthesis approach. Then conduct the full registered search and document the PRISMA flow diagram.

Proof required

Submit your PROSPERO registration confirmation (with registration number), your search strategy documentation, and your PRISMA flow diagram showing how you arrived at your included study set.

What gets checked

PROSPERO registration is completed — the registration number must appear in the proof.
Search covers at least 3 databases including Cochrane — single-database searches are not systematic reviews.
PRISMA diagram is accurate — each stage's number must be traceable from the search documentation.

Common mistakes

Review protocol not registered — registration is a defining feature of a systematic review; without it, you have conducted a literature search.
Less than 3 databases searched — CINAHL or PsycINFO is required alongside PubMed and Cochrane depending on your topic.
PRISMA diagram produced without conducting the screening stages in sequence.

Resources

Foundationstart here

Depthgo deeper

Masteryfor the dedicated

What a verifier looks for

Ask for the PROSPERO registration number — this is non-negotiable for this milestone.
Ask the student to justify the database selection for their specific clinical question.
Check that the PRISMA diagram numbers are traceable — ask them to explain the 'excluded at full-text' stage.

You'll sign in first, then come straight back here.

Screen studies and assess risk of bias

3–4 weeks (screening + bias assessment)

Complete the two-stage screening process for your registered systematic review: title and abstract screening of all identified records (against your inclusion/exclusion criteria), followed by full-text screening of all records that passed title/abstract screening. Then assess risk of bias for all included studies using a recognised tool (Cochrane RoB 2 for RCTs, ROBINS-I for non-randomised studies, or equivalent). Produce a risk of bias summary table for all included studies.

Proof required

Submit your screening decision log (showing number of records at each stage), your risk of bias summary table for all included studies (naming the tool used and the verdict for each domain), and a brief narrative explaining the most common bias risk across the included studies.

What gets checked

Screening was done in two stages — title/abstract screening before full-text is the PRISMA-required process.
Risk of bias tool is named and consistently applied — different tools for different study types must be explained.
Bias narrative identifies the most common risk across all studies, not just the worst individual study.

Common mistakes

One-stage screening (going directly to full-text without title/abstract stage) — this is a process deviation from PRISMA.
Risk of bias assessment that rates all studies as low risk — critical bias assessment finds real risks in most studies.
Bias summary table without a narrative synthesis of what the bias picture means for the review's conclusions.

Resources

Foundationstart here

Depthgo deeper

What a verifier looks for

Ask the student to explain the bias risk verdict for the included study with the most participants — tests understanding of bias in context of statistical power.
Check that the screening was two-stage — ask to see the title/abstract screening log.
Ask what the most common bias domain was across the included studies.

You'll sign in first, then come straight back here.

Synthesise the evidence and produce a review summary

3–4 weeks (synthesis + write-up)

Synthesise the evidence from your included studies: perform a narrative synthesis (or meta-analysis if ≥3 studies are sufficiently homogeneous) summarising the evidence on your PICO question. Apply the GRADE framework to assess the overall certainty of the evidence for your primary outcome. Produce a 4-page summary suitable for a clinical decision-maker audience: PICO question, number and type of included studies, what the evidence shows, how certain the evidence is (GRADE), what the clinical implications are, and where the evidence gaps remain.

Proof required

Submit your 4-page evidence synthesis summary including a GRADE table for the primary outcome and a plain-language interpretation of the GRADE rating.

What gets checked

GRADE framework is applied and the rating is explained — not just stated ('low certainty') but explained in terms of what the rating means for the clinical decision.
Synthesis distinguishes what is known with higher certainty from what is more uncertain — not treating all evidence as equal.
Clinical implications section is specific — not 'more research is needed' but what a clinician or commissioner can act on now, and what they should wait for.

Common mistakes

Narrative that summarises each study separately rather than synthesising across studies.
GRADE rating stated without explanation of what it means for the clinical question.
Evidence gap section that is a wish list for more research rather than specific unanswered questions.

Resources

Foundationstart here

Depthgo deeper

What a verifier looks for

Ask the student to explain the GRADE rating for their primary outcome in plain English.
Ask what the single most important implication of their review is for clinical practice.
Ask whether their review changed what was previously believed about the question — tests whether the synthesis adds to knowledge.

You'll sign in first, then come straight back here.

Part of

Prove Your Public Health Chops →

Related outcomes

Milestone map

3 milestones

3–4 weeks (protocol registration + search)

Proof required

Submit your PROSPERO registration confirmation (with registration number), your search strategy documentation, and your PRISMA flow diagram showing how you arrived at your included study set.

What gets checked

PROSPERO registration is completed — the registration number must appear in the proof.
Search covers at least 3 databases including Cochrane — single-database searches are not systematic reviews.
PRISMA diagram is accurate — each stage's number must be traceable from the search documentation.

Common mistakes

Review protocol not registered — registration is a defining feature of a systematic review; without it, you have conducted a literature search.
Less than 3 databases searched — CINAHL or PsycINFO is required alongside PubMed and Cochrane depending on your topic.
PRISMA diagram produced without conducting the screening stages in sequence.

Resources

Foundationstart here

Depthgo deeper

Masteryfor the dedicated

What a verifier looks for

Ask for the PROSPERO registration number — this is non-negotiable for this milestone.
Ask the student to justify the database selection for their specific clinical question.
Check that the PRISMA diagram numbers are traceable — ask them to explain the 'excluded at full-text' stage.

You'll sign in first, then come straight back here.

Screen studies and assess risk of bias

3–4 weeks (screening + bias assessment)

Proof required

What gets checked

Screening was done in two stages — title/abstract screening before full-text is the PRISMA-required process.
Risk of bias tool is named and consistently applied — different tools for different study types must be explained.
Bias narrative identifies the most common risk across all studies, not just the worst individual study.

Common mistakes

One-stage screening (going directly to full-text without title/abstract stage) — this is a process deviation from PRISMA.
Risk of bias assessment that rates all studies as low risk — critical bias assessment finds real risks in most studies.
Bias summary table without a narrative synthesis of what the bias picture means for the review's conclusions.

Resources

Foundationstart here

Depthgo deeper

What a verifier looks for

Ask the student to explain the bias risk verdict for the included study with the most participants — tests understanding of bias in context of statistical power.
Check that the screening was two-stage — ask to see the title/abstract screening log.
Ask what the most common bias domain was across the included studies.

You'll sign in first, then come straight back here.

Synthesise the evidence and produce a review summary

3–4 weeks (synthesis + write-up)

Proof required

Submit your 4-page evidence synthesis summary including a GRADE table for the primary outcome and a plain-language interpretation of the GRADE rating.

What gets checked

GRADE framework is applied and the rating is explained — not just stated ('low certainty') but explained in terms of what the rating means for the clinical decision.
Synthesis distinguishes what is known with higher certainty from what is more uncertain — not treating all evidence as equal.
Clinical implications section is specific — not 'more research is needed' but what a clinician or commissioner can act on now, and what they should wait for.

Common mistakes

Narrative that summarises each study separately rather than synthesising across studies.
GRADE rating stated without explanation of what it means for the clinical question.
Evidence gap section that is a wish list for more research rather than specific unanswered questions.

Resources

Foundationstart here

Depthgo deeper

What a verifier looks for

Ask the student to explain the GRADE rating for their primary outcome in plain English.
Ask what the single most important implication of their review is for clinical practice.
Ask whether their review changed what was previously believed about the question — tests whether the synthesis adds to knowledge.

You'll sign in first, then come straight back here.