What Should You Do When Only a Few Relevant Studies Exist?

You searched carefully, but only found a few relevant studies. What now?

You expected dozens of studies. After database searching, screening, reference checking, and applying your eligibility criteria, perhaps only three, five, or eight studies remain. That can feel like a methodological problem, particularly when you planned a systematic review, thesis literature review, or evidence synthesis around a question that seemed important.

The temptation is often to fix the small number. You might loosen the eligibility criteria, include studies that are only partly relevant, expand the population or outcomes, or start treating almost-related research as direct evidence.

But the number of included studies is not something you should optimize. The more important question is why so few studies remain. A sparse evidence base can result from an incomplete search, an unnecessarily narrow question, restrictive eligibility criteria, a genuinely under-researched topic, or some combination of these. Those possibilities require different responses.

Sometimes, finding very little research is itself an important result.

Do not broaden the review simply to increase the study count

A small evidence base can mean several different things

Before changing your review, separate two questions that are easy to conflate: Have I failed to find the evidence? and Does the evidence actually exist? A low study count alone cannot answer either question.

First, check whether the search is the problem

A sparse result should trigger a search audit before it triggers a change in the research question. Systematic searches are expected to identify eligible studies as comprehensively as reasonably possible, and relevant records can be missed because terminology varies, indexing is inconsistent, database coverage differs, or the search strategy is overly restrictive.

Return to a few known relevant studies and inspect how their titles, abstracts, keywords, subject headings, interventions, populations, and outcomes are described. Ask whether your search retrieves those records. If it does not, the search strategy may need revision.

Also consider whether you searched the databases and information sources appropriate to the field. Reference-list checking and citation searching can identify records missed by database searches. Trial or study registers, dissertations, conference materials, institutional repositories, regulatory sources, and other forms of grey literature may also matter, depending on the research question. When evidence appears unusually scarce, a more extensive search for relevant grey literature may be especially useful because publication status can shape what becomes visible in conventional databases.

Then examine whether your eligibility criteria are narrower than your actual question

Sometimes the evidence appears scarce because the review criteria contain restrictions that are not essential to the question. Perhaps a population was defined more narrowly than necessary, a publication date restriction was imposed without a substantive reason, or only one terminology variant was effectively captured.

Revisit each major eligibility criterion and ask what scientific or methodological purpose it serves. Removing an arbitrary restriction can improve a review. Removing a defensible restriction merely because too few studies remain is different.

This distinction matters particularly when a protocol or preregistration already exists. Changes may still be justifiable, but they should be documented as amendments rather than quietly rewritten after seeing the search results. Otherwise, eligibility decisions can become influenced by the evidence that happens to be available.

Do not confuse few studies with weak evidence

The number of studies is only one feature of an evidence base. Five large, rigorous, directly relevant studies do not carry the same implications as five tiny studies with serious methodological limitations. Conversely, twenty studies do not automatically provide strong evidence if they share substantial risk of bias or answer a different question.

For intervention evidence assessed using GRADE, certainty is considered through domains such as risk of bias, inconsistency, indirectness, imprecision, and publication bias. A small amount of evidence can contribute to concerns about imprecision, particularly when there are few participants or events and wide confidence intervals, but the study count itself should not be treated as a standalone measure of certainty.

Broadening the evidence base changes the question you can answer

If direct evidence remains sparse, broader evidence may sometimes be informative. The key word is informative, not merely available.

For example, researchers might consider evidence from a wider population, additional outcomes, related conditions, similar interventions, different settings, or other study designs. Each expansion introduces a reasoning problem: how confidently can findings from the broader evidence be transferred back to the original question?

If the target population is extremely specific, you might consider whether evidence from a broader but clinically or conceptually related population can contribute useful information. Similarly, broadening the set of outcomes considered relevant may help in some reviews, provided those outcomes still address the underlying research purpose.

Related conditions, interventions, or contexts can sometimes provide supporting evidence as well. Their inclusion should depend on a defensible relationship to the target question rather than superficial similarity. The further the evidence moves from the original population, intervention or exposure, comparator, outcome, or context, the more carefully you need to consider whether the evidence has become too indirect.

Different study designs may answer different parts of the problem

Scarcity can also make researchers reconsider which forms of evidence are useful. That does not mean every lower-level or alternative design suddenly becomes equivalent to stronger direct evidence.

For some questions, a case report or case series may provide information about unusual events, emerging phenomena, rare harms, or observations that larger comparative studies have not captured. The appropriate question is therefore not simply whether case reports can be included when stronger evidence is absent, but what claims those reports can reasonably support.

Likewise, a small qualitative study may provide rich evidence about experiences, implementation barriers, acceptability, or mechanisms even when it cannot estimate an intervention effect or prevalence. In a sparse literature, the contribution of small qualitative studies should be judged against the question they can answer rather than against the sample sizes expected of quantitative designs.

A meta-analysis is not required simply because several studies exist

Finding a few studies does not automatically mean they should be statistically pooled. Before meta-analysis, consider whether the studies are sufficiently comparable in their questions, populations, interventions or exposures, comparators, outcomes, designs, and effect measures.

With a very small number of studies, some statistical procedures also become less informative. For example, evidence about between-study heterogeneity may be limited, and methods used to investigate small-study effects or publication bias may have little power. A numerical pooled estimate can look impressively precise on a page while resting on a surprisingly fragile evidence base. Statistics, alas, do not become more sociable merely because only three studies turned up.

When pooling is inappropriate, a structured narrative or other appropriate synthesis may be more defensible. The synthesis method should follow the nature of the evidence rather than a desire to produce a forest plot.

Scarcity should change the strength of your conclusions, not the evidence

The final response to sparse evidence occurs during interpretation. A small or uncertain evidence base may support statements about what the available studies observed, but it may not support strong claims about effectiveness, absence of effect, generalizability, or what should happen in populations that were barely studied.

This is where wording becomes consequential. “We found no evidence that the intervention works” can easily be interpreted as evidence that it does not work, even when the actual problem is that very little research exists. The appropriate conclusion may instead be that the available evidence is insufficient to determine the effect with confidence.

Learning to limit conclusions to what a small evidence base can actually support is therefore part of the method, not merely cautious academic phrasing.

Suppose your review finds only four eligible studies

The small number of studies is therefore not repaired. It is investigated, explained, and incorporated into the interpretation. If the search was sufficiently comprehensive and the eligibility criteria remain defensible, the scarcity itself tells readers something important about the state of knowledge.

Common mistakes when the literature turns out to be sparse

Use a sequence of decisions rather than chasing more studies

When the literature is sparse, your task is to determine where the scarcity comes from and what kind of evidence remains defensible. That process can be approached sequentially.

The last possibility deserves particular emphasis. Researchers sometimes regard a sparse literature as an embarrassing outcome because it produces fewer tables, fewer comparisons, and perhaps no impressive meta-analysis. Yet a carefully established evidence gap may be precisely what researchers, funders, practitioners, and policymakers need to know. In some questions, the lack of evidence becomes a substantive finding rather than an inconvenience to hide.

That does not mean every small review proves that more research is needed. Recommendations for future research should identify what is missing and why filling that gap would matter. If several weak studies already ask essentially the same question, another small study of the same kind may add little. Evidence gaps are most useful when characterized rather than merely announced.

Before concluding that the evidence really is scarce

Questions researchers ask when very few studies qualify

Do not solve an evidence shortage by changing what counts as evidence

How to Cite This Guide

This guide is intended to be read, shared, and used in research, teaching, and academic work. If you draw on its ideas, explanations, or other content, please acknowledge the source by citing the guide. Doing so gives appropriate credit and helps your readers locate the original resource.