Meta-analysis topic ideas in gastroenterology and hepatology

Gastroenterology and hepatology mix a few very large, well-reviewed questions (Helicobacter pylori eradication, proton pump inhibitor safety, biologics in inflammatory bowel disease) with a long tail of procedures, diets and liver conditions that are studied in small trials and cohort studies. The long tail is where new reviews usually come from.

Two features of the field shape a good topic. Many trials test a procedure or an endoscopic technique, where operator skill and centre matter, and many outcomes are symptom scores or endoscopic grades that are measured in different ways. Settling the definitions first is half of the work.

Find a gastroenterology and hepatology topic that has not been done

SynthGap searches the published literature, checks each candidate against existing systematic reviews, and proposes topics with the primary studies behind them. Start on the Free plan: no card needed, and every run shows every topic it finds.

Where gaps tend to hide in gastroenterology and hepatology

Newer biologics and small molecules

Each new drug for ulcerative colitis or Crohn's disease is tested in a handful of trials against placebo, and head-to-head comparisons are rare. Updates and network meta-analyses that add recent trials are common needs, and they go out of date within a few years.

Endoscopic techniques and devices

Comparisons of haemostasis methods, stent types, resection techniques and bowel preparation regimens are made in many small randomized trials from single centres. Pooling them by technique, with attention to operator experience, is a recurring source of topics.

Metabolic liver disease

Steatotic liver disease, as its name and definitions have changed, is studied in trials of diet, exercise, drugs and bariatric procedures with varying outcome definitions: histology, imaging fat fraction, liver enzymes. Reviews that fix one outcome definition and pool by it are useful.

Functional gut disorders

Irritable bowel syndrome, functional dyspepsia and constipation are treated with diets, probiotics, psychological therapies and neuromodulators in trials with a high placebo response. How the placebo response varies by trial design is a pooling question in itself.

Microbiome-based and dietary interventions

Probiotics, faecal microbiota transplantation and low-FODMAP diets are tested in many small trials with different strains, doses and delivery methods. Reviews that keep strains and delivery routes apart avoid the common mistake of treating them as one intervention.

Outcomes and effect measures that pool well in gastroenterology and hepatology

Clinical remission and response in inflammatory bowel disease: risk ratios, with the index used (Mayo score, Crohn's Disease Activity Index, Harvey-Bradshaw) and the week of assessment stated. Induction and maintenance phases should be pooled separately.
Sustained virological response and other binary treatment outcomes: risk ratios or odds ratios. With few events, check that the method you choose handles studies with zero events.
Bleeding, re-bleeding and mortality after endoscopic treatment: risk ratios over a fixed follow-up window, with the definition of re-bleeding written out.
Symptom scores for functional disorders: mean differences on one validated scale where possible, standardized mean differences only when scales differ; report the minimal important difference for the scale.
Liver fibrosis and steatosis measured by elastography or imaging: mean differences, with the device and cut-offs noted, since measurements from different machines are not interchangeable.

Common traps in gastroenterology and hepatology reviews

Pooling different probiotic strains, or different bowel preparation regimens, as if they were one intervention.
High placebo response in functional disorders and induction trials: an uncontrolled before-and-after change is not a treatment effect.
Endoscopic and surgical trials where the operator cannot be blinded and skill varies, which tends to flatter the technique the investigators developed.
Combining induction and maintenance trials, or biologic-naive and biologic-experienced patients, in one estimate.
Endoscopic and histologic scores read by one local reader, so the outcome varies by centre.

Check before you commit

Before committing months to a topic, check that nobody has beaten you to it. Search PubMed with its systematic review filter, the Cochrane Library (Cochrane Gut covers many gastroenterology questions), the PROSPERO register for reviews that are registered but not yet published, and Epistemonikos. Note the date of the last search in the newest review and whether new trials have appeared since, which matters most for drug classes where new agents are approved every year. Step-by-step guide.

Starting points to try in SynthGap

These are areas to explore, not claims that they are open. Some may already be well covered, and finding that out quickly is the point.

Faecal microbiota transplantation for conditions other than Clostridioides difficile infection, by delivery route
Endoscopic treatments for obesity and metabolic disease: weight and diabetes outcomes at one year
Psychological and diet-based treatments for functional dyspepsia

Questions

How do I check whether a gastroenterology meta-analysis already exists?

Search PubMed with the systematic review filter, the Cochrane Library, PROSPERO for registered reviews that are not yet published, and Epistemonikos. Check the date of the last search of the newest review and whether new trials have been published since.

Can I combine studies of different biologics?

Only for a question about the class, and then you should report each drug separately as well. For comparing drugs, a network meta-analysis is the usual design, because few trials compare them directly.

Is a network meta-analysis a good project for a first review?

It is a demanding one. You need a connected network, a plan for judging whether the trials are similar enough for indirect comparisons, and a statistician comfortable with the software. A pairwise meta-analysis of a focused question is a safer start.

Other specialties

This page is general methodological guidance, not medical advice, and it does not list specific unreviewed topics. SynthGap's suggestions are AI-generated from the published literature, so verify them before registering a protocol.