Protocol

A protocol is a detailed, prespecified plan for conducting a research study or systematic review, written in advance of data collection to define the research question, study design, methods, eligibility criteria, data collection procedures, and statistical analysis plan. It serves as a binding blueprint that guides the study team and enables external scrutiny of the research process.

Protocols are essential for scientific transparency, reproducibility, and integrity. By documenting methods before the study begins, protocols guard against selective reporting, outcome switching, and post hoc analytical decisions that can inflate false positive rates, and they make such departures detectable when they occur. For clinical trials, the International Committee of Medical Journal Editors (ICMJE) recommends that medical journal editors require registration in a public trial registry (such as ClinicalTrials.gov) at or before the time of first patient consent for enrollment as a condition of consideration for publication, a policy that has applied to trials starting enrollment after July 1, 2005. For systematic reviews, prospective registration on PROSPERO (which accepts systematic reviews with a health-related outcome) or publication of the protocol documents the review methods before results are known. Protocols also facilitate coordination in multicenter studies, helping all sites follow identical procedures.

A study protocol typically includes the following elements: the research question or objective, the study design, the population and eligibility criteria, the intervention and comparator (for intervention studies), the primary and secondary outcomes with their measurement methods and time points, the sample size justification (including power analysis), the randomization and blinding procedures (for trials), the data collection methods and schedule, the statistical analysis plan, procedures for handling missing data, ethical considerations including informed consent procedures, and plans for data management and monitoring.

For systematic reviews, the protocol specifies the review question (often structured using PICO), the databases to be searched, the search strategy, the eligibility criteria, the study selection process, the data extraction procedures, the risk of bias assessment tool, the planned synthesis methods (narrative, meta analytic, or both), and any planned subgroup or sensitivity analyses.

Protocol amendments are sometimes necessary as new information emerges during the study. These should be documented with their rationale and timing, and any changes to the primary outcome or analysis plan should be transparently reported. Comparing published results against the registered protocol is a standard method for detecting outcome switching and selective reporting.

The SPIRIT statement (Standard Protocol Items: Recommendations for Interventional Trials), updated in 2025 as SPIRIT 2025 with a checklist of 34 minimum items, provides guidance for randomized trial protocols, and PRISMA-P (2015, 17 items) provides analogous guidance for systematic review protocols. Following these guidelines helps make the protocol complete, so that reviewers can evaluate the planned methods.

Every research study, from a small observational study to a large multicenter trial, should have a written protocol. Journals that follow the ICMJE recommendations expect clinical trials to be registered before the first participant consents to enrollment. In the US, FDAAA 801 (42 CFR Part 11) requires the responsible party for an applicable clinical trial to register it on ClinicalTrials.gov no later than 21 days after the first participant is enrolled; in the EU, sponsors of trials of medicinal products must apply for authorization through the Clinical Trials Information System (CTIS) under Regulation (EU) No 536/2014 before the trial can start, and most of the trial information is then made public there. Registration is strongly recommended for systematic reviews. The protocol should be finalized before data collection or literature searching begins. Even for observational studies and qualitative research, a prespecified protocol improves rigor and transparency, though formal registration requirements may be less stringent.

A common mistake is writing the protocol after the study has begun or after preliminary data have been reviewed, which defeats the purpose of prespecification. Another error is registering a vague or incomplete protocol that allows too much flexibility in analytical decisions, effectively enabling post hoc modifications without accountability. Researchers sometimes amend protocols without documenting the changes or their rationale, making it impossible for readers to assess whether the amendments were justified. Finally, some teams write detailed protocols but fail to follow them during the study, introducing discrepancies between planned and executed methods that may introduce bias.

A multicenter randomized trial of a new wound dressing develops a 42 page protocol specifying a two arm parallel design, 400 patients with chronic venous leg ulcers, complete wound healing at 24 weeks as the primary outcome, seven secondary outcomes measured at 4, 12, and 24 weeks, and an analysis plan with intention to treat as the primary analysis and a per protocol analysis as a sensitivity analysis. The protocol is registered on ClinicalTrials.gov six weeks before enrollment begins. During the trial, an amendment adds an 8 week visit, changing the follow up schedule from three to four time points, after the investigators find that early dressing problems are being missed between visits. The amendment and its rationale are documented, approved by the ethics committee before it is implemented, and updated in the registry; the primary outcome and analysis plan are unchanged.