Systematic review

What is a systematic review?

A systematic review uses explicit methods to find, select, assess and synthesize research addressing a defined question. The purpose is to reduce the chance that a conclusion reflects whichever papers were easiest to find or most convenient to the argument. Cochrane's introduction emphasizes a prespecified question, documented methods and appraisal of the evidence.

“I read several studies” is not the same claim. An ordinary reading list can be useful, but a systematic review should let another reader inspect how the evidence entered the review and why relevant-looking studies were excluded.

Define the question before collecting the answers

Suppose a community organization wants to know whether text reminders improve attendance at adult education classes. A review question could be: among adults enrolled in in-person courses, what is the effect of a text reminder, compared with the usual course communication, on attendance at the next scheduled class?

That wording excludes several tempting substitutions. Clicking a registration link is not attending. Schoolchildren are not the chosen population. A reminder combined with free transport is not evidence for the reminder alone. Those studies may inform other questions, but combining them without explanation would change this one.

What the review needs to show

1. A protocol: state the question, eligible designs, outcomes and planned synthesis before selecting results. Record later amendments.
2. A search someone can inspect: name the databases and other sources, give the search terms and dates, and seek relevant unpublished work when possible.
3. Study selection: show how records were screened, how disagreements were handled and why full-text candidates were excluded. Keep multiple reports of one study together.
4. Data extraction and appraisal: record participants, intervention, comparison, outcome, follow-up, estimate and uncertainty. Assess risks such as missing outcomes, poor allocation or selective reporting.
5. A synthesis matched to the evidence: explain which findings can be compared and what remains uncertain. Make access limits visible rather than silently treating unavailable results as favorable or null.

Does every systematic review need a meta-analysis?

No. A meta-analysis combines numerical estimates. A systematic review may instead need a structured narrative synthesis because studies answer different questions or do not supply compatible numbers. “Three significant studies and two nonsignificant studies” is not an adequate substitute: significance also depends on precision, so that count can misrepresent the pattern of effects.

In the reminder example, a table separating next-class attendance from end-of-term completion may be more useful than a single average across both. It shows the reader where evidence exists and where the hoped-for durable benefit has not been established.

Judge the review's reach

A systematic method can expose a weak evidence base; it cannot repair every included study. Check when the search ended, whether the question matches your decision, and whether missing evidence could change the conclusion. The review's label is a description of its method, not a guarantee that the intervention works.