Services
Evidence synthesis and research translation
We find out what is already known. Reviews and gap maps tell you what the evidence says, where it is thin, and what to do with it.
Questions this answers
- What does the evidence say about this approach?
- Where is the evidence thin or out of date?
- How do we turn findings into a decision?
The kinds of evidence work we run
Most engagements combine two or three of these.
Systematic reviews
Every relevant study found through a stated search strategy, screened by more than one person, and combined using a documented method. The most thorough, and the slowest.
Scoping reviews
A map of what kinds of evidence exist on a topic, and how much, before deciding whether a full systematic review is worth commissioning.
Rapid evidence reviews
The same discipline as a systematic review, on a shortened timeline, with the shortcuts stated plainly rather than hidden.
Evidence gap maps
A visual matrix of interventions against outcomes, showing where evidence exists and where it does not, adapted from the method 3ie and the Campbell Collaboration use.
Policy briefs and evidence-to-policy workshops
Turning a review into two or three pages a decision-maker will actually read, and, where useful, a session to work through what follows.
Learning agendas and dissemination events
A standing list of what your organisation needs to learn next, and events that put findings in front of the people who will use them.
How we synthesise evidence
The method changes how much the conclusion can be trusted. We state ours in every proposal.
The search strategy is stated before we start.
What databases, what search terms, what date range — fixed before the results could influence the search, not after.
Two people screen, not one.
A second reviewer checks which studies were included and excluded, and disagreements are resolved and recorded, not settled by one person alone.
The certainty is graded, not just summarised.
Every conclusion carries a certainty rating, not only a description of what the studies found.
What is missing is reported alongside what is found.
A gap in the evidence is a finding in itself, and we say so rather than staying silent where the evidence runs out.
How we grade the certainty of a conclusion
We use GRADE, the system that Cochrane, the World Health Organization and more than 120 other organisations use to rate how much a conclusion can be trusted, not only what it says.
High
We are very confident the true effect is close to what the evidence shows.
Moderate
The true effect is probably close to what the evidence shows, but it could be substantially different.
Low
Our confidence is limited; the true effect may be substantially different.
Very low
We have very little confidence: the true effect is likely to be substantially different.
Why one promising study is not the whole picture
Published studyA cash transfer trial in Zomba, Malawi, found that unconditional cash cut marriage and pregnancy among adolescent girls who had left school. When the same question was tested in the national government cash transfer programmes of Malawi and Zambia, both randomised trials found strong effects on poverty and school attendance, but limited effects on early marriage and pregnancy. A Kenyan programme had shown no effect on early marriage, and a South African trial none on pregnancy. Read together, the studies say cash reliably reduces poverty and raises school attendance, while its effect on early marriage depends on the programme and the setting. No single study could have told you that.
From our own work
Real work of ours that used this kind of research, not an invented example.
Probatio Africa
Real exampleOpen evidence questions
83 questions across our five sectors that the published evidence cannot yet answer, each with the reason it is still open. Built and kept up to date by our team.
Reading an evidence gap map
Illustrative| Intervention | Studies found | Certainty |
|---|---|---|
| Cash transfers → school attendance | 14 studies, example region | High |
| Cash transfers → early marriage | 5 studies, example region | Low |
| School feeding → learning outcomes | 2 studies, example region | Very low — evidence gap |
The third row is the useful finding: not a conclusion, but a gap. Two studies is not enough to say anything with confidence, and a gap map says so rather than reporting a number anyway.
Questions worth asking an evidence synthesis provider
We built five questions from PRISMA, the standard reporting guideline for systematic reviews, and from GRADE, and we answer all five in every proposal.
What is the search strategy, and was it fixed before the results came in?
Stated in the proposal: the databases, the terms, the date range.
Did more than one person screen the studies?
Yes, with disagreements recorded, not resolved silently by one reviewer.
What certainty rating does each conclusion carry?
Stated using GRADE’s four levels, not left as an unrated summary.
What did the search not find?
Reported as a finding, not omitted because it is less satisfying than a positive result.
How is a rapid review different from a full one, in this case?
The shortcuts taken to meet a shorter timeline are stated plainly, so you know what was, and was not, done.
What you receive
A review, a short brief for decision-makers and, if you want it, a workshop to work through what follows.
The review
What was searched, what was found, and the certainty rating for each conclusion.
The brief
Two or three pages, written for a decision-maker, not a literature review committee.
The workshop
A session to work through what the findings mean for your decision, if you want one.
Questions people ask about evidence synthesis specifically
What’s the difference between a systematic review and a rapid review?
The same discipline — a stated search, screening by more than one person, a documented method — but on a shortened timeline, with the shortcuts stated rather than hidden.
How long does a full systematic review take?
Usually six to eighteen months, depending on how much has been published on the topic. Reviews registered with PROSPERO took 67 weeks on average from start to publication. A scoping review first can tell you whether a full review is worth the time.
Can you do this quickly for an urgent decision?
Yes. A rapid evidence review usually takes one to three months. It keeps the same discipline on a shorter timeline, with the shortcuts and trade-offs stated plainly rather than quietly.
What if the evidence is genuinely contested or mixed?
We report the range of findings and say so, rather than picking a side. Where sources disagree, the review says who reports what.
Need to know what the evidence actually says before you commit?
Tell us the approach or the question. We will say what kind of review answers it, and how long it would take.