Rapid Evidence Synthesis Template

Rapid Evidence Synthesis Template

Purpose: Quickly gather, assess, and convert existing evidence into a concise, decision‑ready package while documenting limits and uncertainty. Use this template when time is limited but decision quality matters: experiments, pilots, vendor selection, product decisions, policy choices, and early research scoping.

Quick guidance

Timebox the work, keep transparency, and prioritise reproducibility for what matters most to the decision. Assign clear roles (lead, searcher, screener, synthesizer, reviewer, sponsor). Capture and store search records, screening logs, and included study extracts for follow‑up.

Core sections (use these headings in your review)

1. Review snapshot

One short paragraph (2–4 sentences) that answers: Who asked for this review, what decision is it informing, the timeframe, and the bottom line conclusion.

2. Research question(s)

State the question(s) using a simple PICO/PECO or similar frame where useful. Example: "In small manufacturers (P), does low‑cost sensor monitoring (I) compared with manual inspection (C) reduce unplanned downtime within 6 months (O)?"

3. Scope & limits

• Inclusion criteria (population, interventions, comparators, outcomes, study types, date/language limits)
• Exclusion criteria (what is out of scope)
• Planned search end date and timebox (see Time‑boxing below)

4. Search strategy summary

Record databases and sources searched, main search strings, filters, and dates. Example:

Databases: PubMed, Scopus, IEEE Xplore, Google Scholar (first 200 results)
Search date: 2026‑06‑01
Search string (example): ("sensor" OR "condition monitoring") AND ("downtime" OR "maintenance") AND (manufactur* OR factory)
Limits: English, 2015–present
Additional sources: vendor white papers, relevant standards, citation tracking

Keep a short search log as an appendix for reproducibility.

5. Screening & selection

Document screening steps and numbers (title/abstract screening; full text). Use a tiny PRISMA flow adapted for rapid reviews: records identified → screened → full text assessed → included. Capture reasons for exclusion at full‑text stage where relevant.

6. Extracted evidence (concise table)

For each included item, capture: citation, study type/design, population/sample, intervention/exposure, outcome(s), key result(s), limitations/notes, and a short confidence rating. A one‑line result summary helps end readers.

7. Synthesis & key findings

Provide a concise synthesis (3–6 bullet points). Map the evidence to the research question. Distinguish consistent findings, conflicting findings, and unanswered aspects.

8. Confidence assessment (rapid rubric)

Assign a confidence level (High / Moderate / Low / Very low) for each key finding using these quick criteria:

• Quantity: Are there multiple, independent sources?
• Consistency: Do results point in the same direction?
• Directness: Is the evidence directly about the population and outcome?
• Risk of bias: Are study designs and reporting trustworthy?
• Precision: Are effects large/consistent enough to matter?

Note: This is a pragmatic, transparent judgement—record the rationale for each rating.

9. Implications for hypotheses and decisions

Translate findings into implications for the specific decision or hypotheses. Example: "Evidence suggests sensors can reduce small‑plant downtime in similar settings—Moderate confidence—recommend a 3‑month pilot with defined KPIs."

10. Recommended experiments, pilots or actions

List recommended near‑term actions mapped to evidence strength, with success metrics, duration, and minimal viable experiment design. Include quick tests to reduce key uncertainties.

11. Evidence gaps and follow‑up research

List critical unknowns, priority questions for deeper review or primary research, and suggested data to collect during pilots.

12. Transparency & appendices

Attach or link to: full search log, screening log, extracted study table, full citations, short methodological note (how searches were run), and conflicts of interest. State who performed the work and the date.

One‑page stakeholder brief (copy/paste friendly)

Use this ready layout to communicate to decision makers. Aim for 300–450 words.

Suggested time‑boxing rules (examples)

• Rapid scan for situational awareness: 1–3 days (broad search, key grey literature, one reviewer)
• Focused rapid review for immediate decisions: 5–10 business days (two reviewers if possible; title/abstract screening + targeted full‑text)
• Decision‑grade rapid review (higher transparency): 2–4 weeks (clear documentation, dual screening or spot checks, structured extraction)

Adjust scope to match the decision’s risk and value. Use shorter timeboxes for lower‑risk operational choices and longer ones for high‑impact strategic choices.

Roles & minimum team

• Lead (owner of question and output)
• Searcher (runs searches and documents strategy)
• Screener(s) (screen titles/abstracts and full text)
• Synthesizer / writer (produces the 1‑page brief and synthesis)
• Reviewer / sponsor (decision maker who reviews conclusions)

Common pitfalls to avoid

• Selective citation: do not pick only studies that fit a favored conclusion; document reasons for exclusions
• Pretending precision: report uncertainty clearly and avoid exact quantification when evidence is weak
• Scope creep: if the decision changes, reset the question and timebox

Example deliverables

1. One‑page stakeholder brief
2. Appendix A — search log and search strings
3. Appendix B — screening log and PRISMA numbers
4. Appendix C — extracted study table

Template checklist (quick)

1. Question framed and sponsor assigned
2. Timebox agreed
3. Searches run and logged
4. Screening documented
5. Extracts created for included items
6. Synthesis and confidence judgments recorded
7. One‑page brief produced and reviewed
8. Appendices saved for follow‑up

Notes: Keep this template as a living artifact. For repeated use, consider turning the extraction table and the one‑page brief into an interactive form that stores each review (see Capability suggestions below).