Clicky

Medical & Clinical Research Journal

Medical & Clinical Research Journal — Peer-reviewed health analysis, supplement safety reviews, and clinical research summaries.

clinical_research

Correlation and Causation in Health Headlines: Five Questions Before Sharing a Claim

By MedClinRes.org Clinical Research Team. This article has not been peer-reviewed or clinically reviewed, and it is general education, not medical advice.

A headline says a food, supplement, or habit is “linked to” a longer life, a lower cancer risk, or a slimmer waistline. Before treating that link as proof of cause and effect, five questions separate a headline worth sharing from one worth skepticism: What was the study design? Is there a real comparison group? Could confounding explain the result? How large is the effect, really? And does the study population look anything like you? Answering these takes a few minutes and applies to almost any “X is linked to Y” claim in health news.

Why “Linked To” Is Not the Same as “Causes”

Most health headlines are built on observational research — studies that track what people already do, rather than assigning them to a treatment and a control. Observational studies are useful for spotting patterns, but a pattern is not a mechanism. The National Institutes of Health's National Center for Complementary and Integrative Health (NCCIH) has walked through this exact problem using a study that reported a 29 percent decreased risk of death among older women who drank chamomile tea. Even after the researchers statistically adjusted for factors like age and health status, NCCIH noted that “the women who use chamomile may differ in many ways from those who do not,” and that “statistically correcting for other factors only captures the impact of measured differences.” Its conclusion applies far beyond chamomile: “there may be a big gap between correlation and causation.”

The Five-Question Check

Use these five questions on any claim that a product, food, or behavior is responsible for a health outcome. None of them requires a statistics background — they just require reading past the headline.

1. What was the study design?

Randomized controlled trials assign people to a treatment or a comparison condition, which spreads unmeasured differences evenly between groups. Observational studies simply track what people already do. As the NIH's overview of outcome measures in field trials explains, “the most powerful way to minimize bias in the assessment of the impact of an intervention is through the conduct of a double-blind randomized trial.” A headline based on an observational study is describing an association, not a proven effect. Our guide to primary and secondary outcomes covers what a well-designed trial commits to measuring before it starts.

2. Is there a real comparison group?

An effect only means something relative to what would have happened otherwise. The CDC's epidemiology training materials define a measure of association as a figure that “quantifies the relationship between exposure and disease among the two groups,” arrived at “by comparing the observed group with another group that represents the expected level.” If a claim does not say what the comparison group was — people who didn't use the product, ate differently, or received a placebo — there is no baseline to judge the result against.

3. Could confounding explain the result?

A confounding variable is something that affects both the “exposure” and the outcome, making the two look connected when the real driver is something else entirely. In the chamomile example above, NCCIH pointed out that tea-drinkers and non-drinkers “may differ in many ways” that a survey doesn't capture — diet, activity level, or general health-consciousness could all be doing the real work. Before accepting a causal claim, ask what else tends to travel alongside the behavior in question.

4. How large is the effect, really?

Effect size is usually expressed as a risk ratio (also called relative risk): the risk in one group divided by the risk in another. Per CDC training materials, “a risk ratio of 1.0 indicates identical risk among the two groups,” while one real example in that material describes a risk ratio of 6.1 as “6.1 times as likely,” and a risk ratio of 0.28 as “only approximately one-fourth as likely.” A “40 percent higher risk” headline sounds dramatic but can describe a change from a very small baseline risk to a slightly less small one. Our guide to confidence intervals covers the related question of how much uncertainty surrounds that number in the first place.

5. Does the study population match you?

A result generated in one group of people does not automatically transfer to another. The NIH's guidance on field trial outcome measures notes, as one concrete example, that requiring written informed consent to enroll “will select a subgroup of the population who accept to sign such a form and participate in the study, generating a potential selection bias” — meaning who chose to take part can already skew a study away from the general population. More broadly, a result measured in older adults, a specific sex, a narrow age range, or a particular country may not describe your own risk. Our guide to inclusion and exclusion criteria walks through how to check who a study actually enrolled.

Reading a Real Headline Through the Five Questions

Applying the checklist to the chamomile example above: the design was observational, not a randomized trial; a comparison group existed (women who drank little or no chamomile tea) but the two groups likely differed in unmeasured ways; NCCIH itself flagged that the two groups “may differ in many ways” that adjustment couldn't fully capture; the headline number (29 percent lower mortality) is an effect size worth noting but not treating as precise; and the population studied was a specific older-adult survey sample, which may not represent every reader. None of that means the finding is worthless — it means the finding is a hypothesis, not a verdict.

What This Checklist Does Not Tell You

These five questions help evaluate whether a single study supports a causal claim. They cannot tell you whether a specific product, supplement, or treatment is right for your own health situation, and they are not a substitute for reading the full study rather than a headline summary of it. Readers managing an existing health condition, taking medication, or considering a new supplement should discuss the decision with a physician or pharmacist who knows their full medical history, as noted in this site's medical disclaimer.

What You Can Do Next

Before sharing or acting on a health headline, look for the original study rather than relying on a press release or social summary. Check the study design, the comparison group, the effect size, and the population studied — all five are usually stated in the abstract. If a headline claim leaves any of the five questions unanswered, treat it as a lead worth investigating rather than a settled fact. If the coverage also mentions who paid for the study, our guide to study funding and conflicts of interest covers what that disclosure does and doesn't tell you.

Sources

Supplement Study Dosage: How to Compare the Research Amount With a Label

This article has not been peer reviewed or clinically reviewed and is not individual medical advice. Checked against the cited sources on September 30, 2026. See the Medical Disclaimer.

Short Answer

A dose reported in a study can be compared with a supplement label only after four items line up: the ingredient form, the amount per serving as the label defines it, how the ingredient was prepared, and who was studied and for how long. Even when every number matches, that match does not show the product will work, and it does not show the product is safe for a particular person. The NIH Office of Dietary Supplements (ODS) notes that the manufacturer suggests the serving size, but a health care provider might decide a different amount is more appropriate.

What “Amount Per Serving” Means on a Label

ODS explains that a Supplement Facts label lists the active ingredients, the amount per serving (the dose), and other ingredients such as fillers, binders, and flavorings (ODS, Dietary Supplements: What You Need to Know). Three points from FDA's labeling guide affect any comparison with a study:

  • A serving is defined as the maximum amount recommended on the label per eating occasion. FDA's example: if directions say to take 1-3 tablets with breakfast, the serving size is 3 tablets. A listed amount can therefore describe the top of a recommended range, not the smallest quantity.
  • Products with different servings, such as one tablet in the morning and two at night, may show additional columns labeled by quantity, for example “Amount per 1 Tablet” and “Amount per 2 Tablets.”
  • Percent Daily Value (%DV) is the share of an established Daily Value in one serving. It is not the amount used in a study, and some ingredients have no Daily Value established.

Source: FDA Dietary Supplement Labeling Guide, Chapter IV.

Ingredient Form, Units, and Preparation

A study amount and a label amount can use the same number and still describe different things. FDA's guide illustrates why:

  • Salt versus the nutrient: When calcium carbonate supplies calcium, the label lists the weight of the calcium, not the weight of the whole salt.
  • Source and plant part: A Supplement Facts panel may list the source of a dietary ingredient, and it must include the plant part from which an ingredient is derived.
  • Liquid extracts: These are listed by the volume or weight of the total extract and the condition of the starting material, and the solvent must be identified on the label.
  • Dried extracts: These are listed by the weight of the dried extract.
  • Proprietary blends: The guide describes listing the blend's total weight, with the other dietary ingredients in the blend named in descending order by weight. That description does not call for an individual amount for each of those ingredients, so a study amount for one of them may not be checkable against the label.
  • Units: FDA says vitamins and minerals should use the units given for their Daily Values, for example vitamin C in milligrams. Check that the study and the label use the same unit before comparing. Do not convert between units yourself unless the source gives the conversion.

ODS adds that a supplement's safety depends on many things, including its chemical makeup, how it works in the body, how it is prepared, and the amount taken. Preparation therefore belongs on your comparison, not just the milligram figure.

Study Population and Duration

The people studied and the length of the study are part of the dose. An amount tested in one group over a set period does not automatically apply to someone with a different age, health status, or medicine list, or to longer use. ODS notes that many supplements have not been well tested for safety in children and in women who are pregnant or nursing. Record who took part, how long they took the amount, and whether the ingredient was given alone or with others, then ask whether that describes the person who would use the product.

Why a Matching Amount Does Not Settle Safety or Effect

  • FDA does not determine whether dietary supplements are effective before they are marketed (ODS).
  • Side effects are most likely at high doses, when supplements replace prescribed medicines, or when many supplements are taken together (ODS).
  • Manufacturers may add vitamins, minerals, and other ingredients to foods such as cereals and beverages, so total intake can be higher than the label of one product suggests (ODS).
  • Some supplements interact with medicines. ODS gives examples including vitamin K with warfarin and St. John's wort with many medicines.
  • FDA's guide states that an ingredient specifically added to a product must be present at 100% of the declared amount, while a naturally occurring ingredient must be present at 80% of the declared value.
  • Independent quality seals can indicate a product was properly manufactured and contains the listed ingredients, but ODS states these seals do not guarantee a product is safe or effective.

Research-to-Label Worksheet

Copy this into a note or print it. Fill in one worksheet per ingredient. Where a detail is not stated in the study or on the label, write “not stated” instead of guessing.

Part A: The Study

  1. Study title, year, and where it was published: ________
  2. Ingredient name exactly as the study reports it: ________
  3. Form or preparation reported (extract, powder, salt, solvent, plant part): ________
  4. Amount given, with its unit, and how often it was given: ________
  5. Was the amount per dose, per day, or per body weight? ________
  6. Who took part (age range, health status, whether they took medicines): ________
  7. How long the amount was taken: ________
  8. Was the ingredient given alone or with other ingredients? ________
  9. Any side effects or dropouts the study reports: ________

Part B: The Label

  1. Product name and lot or expiration information: ________
  2. Serving Size (the exact wording and count of capsules, tablets, or scoops): ________
  3. Servings Per Container: ________
  4. Ingredient name as printed, and any source or plant part shown: ________
  5. Amount per serving, with its unit: ________
  6. Is the ingredient inside a proprietary blend? If yes, is its individual amount listed? ________
  7. Directions on the label (how many servings, when, with food or not): ________
  8. Other ingredients and any warnings printed on the label: ________

Part C: Side-by-Side Check

  • Do the ingredient form and preparation match, or are they different? ________
  • Are the units the same? If not, who confirmed the conversion? ________
  • Is the study amount per dose or per day, and is the label amount per serving or per day? ________
  • If the label lists a range in the directions, which end does the Serving Size reflect? ________
  • Do the study participants resemble the person who would take the product? ________
  • Does the study duration match the intended use? ________

Part D: Questions for a Pharmacist or Clinician

ODS suggests using its information sources and talking to your health care providers to answer these questions before taking any dietary supplement:

  1. What are its potential benefits for me?
  2. Does it have any safety risks?
  3. What is the proper dose to take?
  4. How, when, and for how long should I take it?

Add these from your own worksheet: Does this study population resemble me? Could this ingredient interact with my medicines or upcoming procedures? Is the amount on this label appropriate for me, given everything else I take, including fortified foods?

Who Should Get Professional Advice First

Ask a physician, pharmacist, or other qualified professional before starting a supplement if any of the following applies:

  • You take prescription medicines, or plan to take a supplement instead of, or together with, a prescribed medicine. ODS advises getting your provider's approval first.
  • You are taking a supplement to treat a health condition.
  • You are scheduled for surgery.
  • You are pregnant or nursing. ODS advises caution about supplements beyond a standard prenatal supplement.
  • The product is for a child. ODS says to be careful about giving supplements to a child unless their health care provider recommends it.

What You Can Do Next

  1. Fill in Parts A and B of the worksheet from the actual study and the actual label in front of you.
  2. Look up the ingredient in the ODS Dietary Supplement Fact Sheets, which cover health effects, safety, recommended amounts, and interactions with medicines.
  3. Record every supplement and medicine you take, with dose, frequency, and reason, using the ODS form My Dietary Supplement and Medicine Record.
  4. Bring the completed worksheet and the record to a pharmacist or clinician and go through Part D together.
  5. If you think you had a bad reaction, tell your health care provider and report it to FDA at 800-FDA-1088 or through FDA's online form. ODS also advises reporting it to the manufacturer using the contact details on the label.

Limits of This Article

FDA's labeling guide is dated April 2005, contains nonbinding recommendations, and refers readers to Title 21, Part 101 of the Code of Federal Regulations for requirements, so check current regulations for anything that affects a decision. The ODS consumer fact sheet cited here was last updated in January 2023. This article explains how to compare figures. It does not recommend an ingredient, an amount, or a product, and it contains no affiliate links. Med Clin Res may earn compensation from some links elsewhere on the site, as described on our About and Editorial Standards pages. Our Editorial Team and Review Process page explains how content is prepared and what we do not claim.

Sources

Clinical Trial Status Labels: Recruiting, Active, Completed, and Withdrawn

A status label on a trial registry record describes where a study is in its lifecycle. It does not tell you whether a treatment works, whether you would qualify, or whether you should take part. Those questions depend on the study's eligibility criteria, the study team's screening, and advice from your own clinician. This guide explains what each ClinicalTrials.gov label means, what it does not mean, and how to check whether a record is current.

Editorial note: this article is a plain-language explainer. It has not been peer reviewed or clinically reviewed, and it is not medical advice. See the site's About and Editorial Team and Review Process pages.

The Short Answer

  • Recruiting means the study is currently seeking participants. It does not mean you are eligible or that a spot is open near you.
  • Active, not recruiting means the study is continuing but is not currently enrolling new participants.
  • Completed means the study ended normally. It does not mean the treatment worked, and it does not mean results have been posted.
  • Withdrawn means the study stopped before enrolling its first participant.

ClinicalTrials.gov's protocol definitions describe these labels in nearly these terms (Protocol Registration Data Element Definitions).

What Each Status Label Means

Not yet recruiting

Participants are not yet being recruited. The study may be planned, but enrollment has not opened. The registry applies this wording at the level of individual facilities (see the PBRC ClinicalTrials.gov PRS Guide (2024), hosted by LA CaTS), so confirm the study-wide definition on the registry itself.

Recruiting

Participants are currently being recruited, whether or not anyone has been enrolled yet. The label describes the study's recruitment effort, not your personal eligibility.

Enrolling by invitation

Participants are being selected, or will be selected, from a population the researchers decided on in advance.

Active, not recruiting

The study is continuing. Participants are receiving an intervention or being examined, but new participants are not currently being recruited or enrolled.

Completed

The study has concluded normally and participants are no longer receiving an intervention or being examined; the last participant's last visit has occurred. “Completed” describes how the study ended. It says nothing about whether the intervention was safe or beneficial. Posted results are a separate matter: a record has them only if summary results were submitted to the results database.

Suspended

The study halted early but may resume.

Terminated

The study halted early and will not resume. Participants are no longer being examined or receiving the intervention. For studies that are suspended, terminated, or withdrawn, the registry's definitions call for a brief explanation of why the study stopped, so look for it.

Withdrawn

The study stopped before enrolling its first participant, meaning no one was enrolled.

Unknown

The record's last known status was not yet recruiting, recruiting, or active, not recruiting; its completion date has passed; and the information has not been confirmed within the past two years. ClinicalTrials.gov considers unknown-status studies closed. Do not assume an unknown-status study is still enrolling.

Build a Registry-Status Timeline You Can Verify

Instead of relying on one label, write a short timeline for the study you are reading. Fill each step from the record itself. If a field is blank, write “not stated” rather than guessing. Field names on the website may be worded slightly differently from the descriptions below.

  1. Identify the record. Copy the study's NCT number and title. ClinicalTrials.gov assigns every registered study a unique identification code.
  2. Record the overall status label. Note the exact wording shown and compare it with the definitions above.
  3. Record when the information was last confirmed. The registry tracks the most recent date the sponsor or investigator confirmed the information as accurate. If a study labeled recruiting, not yet recruiting, or active, not recruiting goes two years without confirmation, its status is shown as unknown (ClinicalTrials.gov FAQ).
  4. Record when the record was last updated. An old update date beside a “Recruiting” label is a reason to verify before relying on it.
  5. Record the start and completion dates and whether they are anticipated or actual. Registry definitions use “anticipated” for estimated dates (see this NIAID registration training deck).
  6. Record the reason for stopping, if any. This applies to suspended, terminated, and withdrawn studies.
  7. Record site-level information. Recruitment status can also be listed for each participating facility, so a location may differ from the study-wide label.
  8. Record contact information, if shown. ClinicalTrials.gov does not list contacts for studies labeled active not recruiting, completed, terminated, enrolling by invitation, suspended, or withdrawn (How to Read a Study Record).
  9. Record whether results are posted. A completed study may or may not have posted results.

Warning: a status label is not a treatment recommendation. A “Recruiting” label says nothing about whether an intervention is safe or effective, and a “Completed” or “Withdrawn” label is not evidence for or against one. Registry records describe research plans and their progress. They are not clinical guidance.

Why the Label Can Be Wrong or Out of Date

Study teams supply and update their own record information. A published cross-sectional analysis that examined ClinicalTrials.gov statuses as of 1 January 2017 concluded that recruitment status information was often outdated or wrong (Discrepancies between ClinicalTrials.gov recruitment status and actual trial status). That analysis reflects 2017 data, so treat it as a caution about the limits of self-reported status rather than a current error rate.

For people who work with registry data programmatically, the ClinicalTrials.gov API is refreshed daily Monday through Friday, generally by 9 a.m. Eastern Time, and provides a data timestamp so you can confirm a refresh has finished (ClinicalTrials.gov API). Even fresh data only reflects what the study team has reported.

Who Should Get Professional Advice

  • Anyone considering joining a study should discuss it with their own physician or care team first, and ask the study team directly about risks, procedures, and what participation involves.
  • Anyone thinking about changing, stopping, or delaying current treatment because of a trial listing should speak with their clinician before doing anything.
  • Caregivers reading a record on someone else's behalf should confirm details with the study contact rather than relying on the label alone.

Eligibility is decided by the study team using the criteria in the record. A “Recruiting” label cannot tell you whether you qualify.

What You Can Do Next

  1. Open the record for the study you care about and fill in the timeline above.
  2. Read the eligibility section, compare it with your situation, and bring your notes to your clinician.
  3. If contact details are listed, ask the study team whether the record is current and whether the site near you is enrolling.
  4. If the status is unknown, terminated, or withdrawn, look for the stated reason and for any posted results before drawing conclusions.

Limits of This Guide

Definitions here follow ClinicalTrials.gov. Other registries may use different labels or wording, and registry text can change, so confirm current definitions on the registry itself. For how this site describes its sourcing and independence, see About Med Clin Res.

Sources

Study Funding and Conflicts of Interest: What to Look for Before You Rely on a Result

This article has not been peer reviewed or clinically reviewed. It is educational and is not individual medical advice.

The Short Answer

Funding and disclosed relationships may be reported in more than one place. Look for a funding statement, author disclosure statements or forms, a data-access statement, and, where one is cited, a protocol. Each answers a different question. None of them, on its own, tells you whether a result is right or wrong.

The International Committee of Medical Journal Editors (ICMJE) says that having a relationship or activity does not always indicate a problematic influence on a paper's content. It also says perceptions of conflict matter, and that readers must be able to judge for themselves whether a relationship is pertinent. That is a sound posture for a reader: treat disclosures as information to weigh, not as a verdict. (ICMJE, Section II.B)

Disclosure Locator: What Each Item Can and Cannot Establish

1. Funding Statement (Sometimes Labeled “Sources of Support”)

  • Can establish: who directly supported the work. ICMJE says a funding statement should include only direct support of the work described, and that support for an individual's time should be reported as such, separately from funding of the work itself.
  • Can also establish: the sponsor's stated role. ICMJE says source-of-support statements should name sponsors and explain any role in study design, data collection, analysis, interpretation, writing, or restrictions on submitting the report for publication, or state that the source had no such involvement or restrictions.
  • Cannot establish: whether a funder actually influenced the work. As a reader, remember that a statement is a report, not an independent check. A named funder is not proof of bias.

2. Author Disclosure Statements or Forms

  • Can establish: the relationships and activities the authors reported. ICMJE lists financial examples such as employment, consultancies, stock ownership or options, honoraria, patents, and paid expert testimony. It also says non-financial interests, such as personal relationships, rivalries, academic competition, and intellectual beliefs, may represent or be perceived as conflicts.
  • Cannot establish: whether a listed relationship affected the paper, or that every relevant relationship was listed. ICMJE places the responsibility for disclosing relationships on authors.

3. Acknowledgments

  • Can establish: who the authors chose to thank. The ICMJE page cited here does not address acknowledgments, so this article makes no claim about what journals require in that section.
  • Reader step: check whether an organization or person appears in the acknowledgments that does not appear in the funding statement or disclosures. If so, treat it as a question to raise, not a conclusion.
  • Cannot establish: the nature of any support, unless the text says so.

4. Data-Access Statement

  • Can establish: whether the authors say they had access to the study data. ICMJE says articles should declare this, explaining the nature and extent of access and whether it is ongoing.
  • Also worth knowing: ICMJE says editors may ask authors of a study sponsored by a funder with a proprietary or financial interest in the outcome to sign a statement confirming full data access and responsibility for the analysis. It also says authors must avoid agreements with sponsors that interfere with their access to the data or their ability to analyze, interpret, and publish independently.
  • Cannot establish: that the analysis was sound. Access to data is not the same as a correct analysis.

5. Protocol or Registration Record (When One Is Cited)

  • What we can say: the sources reviewed for this article do not describe what a protocol or registry record must contain, so we make no claim about that.
  • Reader step: if the paper cites a protocol or registration, open it. Check whether it names a sponsor or funder, and compare that with the paper's funding statement. Treat a mismatch as a question to ask, not a finding.
  • Cannot establish: anything the record does not contain. An absent detail is not evidence of a problem.

A Five-Step Way to Use the Locator

  1. Find the funding statement and note each named funder and any stated role.
  2. Read the author disclosures and note financial and non-financial relationships that relate to the topic studied.
  3. Check for a data-access statement and any statement about restrictions on publication.
  4. Compare the acknowledgments and any cited protocol against the funding statement for gaps or mismatches.
  5. Write down what is unclear. Then weigh the result using the study's design and data, with the disclosures as context.

What Not to Conclude

  • Do not assume industry or commercial funding makes a result false. A disclosed relationship is a reason to read more carefully, not to dismiss the work.
  • Do not assume a clean disclosure makes a result reliable. Design, sample, outcomes, and analysis still matter.
  • Do not treat a missing statement as proof of wrongdoing. It is a reason to look for the information elsewhere, or to ask the journal or the authors.

Limits and When to Seek Professional Advice

This article explains how to locate disclosures. It does not assess any particular study, and it cannot tell you whether a result applies to you. If a study is influencing a decision about your health, medication, or treatment, discuss it with a qualified clinician or pharmacist before acting. This site's Medical Disclaimer page explains the limits of its content.

What You Can Do Next

  • Pick one study you are relying on and complete the five steps above.
  • If a funding or disclosure statement is missing or unclear, contact the journal or corresponding author and ask.
  • To see how this site describes its sourcing and review limits, read Editorial Team and Review Process and About Med Clin Res.

Source

Systematic Review Protocols: Why a Plan Written First Matters

Guidance such as the Cochrane Handbook expects a systematic review to define its question and methods in a written protocol before the work of selecting studies begins. If a review does not show that plan, or shows one written after the results were known, you cannot easily tell which choices were made in advance and which were made after seeing the data. This article gives a short protocol check card you can apply to any systematic review you read.

Why the plan comes first

The Cochrane Handbook explains that a well-formulated question guides most of the later work: eligibility criteria, the search for studies, data collection, and how results are grouped and presented (Cochrane Handbook, Chapter 2). The Handbook also lists a mandatory expectation for its own reviews: define the objectives, including population, interventions, comparators, and outcomes, in advance.

The same chapter says the review's population, intervention, comparison, and outcome (PICO) are planned at the protocol stage and set which studies are included or excluded. Deciding this before seeing study results limits the room to shape the question around a preferred answer.

What a protocol register does

PROSPERO is an international register of systematic review protocols. An editorial in PLoS Medicine announced it as a service of the Center for Reviews and Dissemination at the University of York and advised investigators to record the protocol for a planned or ongoing review of health care interventions even before screening studies for inclusion (PLoS Medicine editorial). A published description says the register aims to avoid duplicated effort, reduce reporting bias and promote transparency (PubMed record).

Registration is not the only route. A paper surveying options identified PROSPERO, the Research Registry, INPLASY, Open Science Framework Registries, and protocols.io. It also separates prospective registration from publishing a protocol as a manuscript (Where to prospectively register a systematic review).

Protocol check card

Use these six checks on any systematic review. Find the protocol or registry record first, then compare it with the published paper.

  • Question: Is there a clear primary objective naming the intervention, the health problem and the people studied? The Handbook suggests stating this in a single sentence. Look for the population, intervention, comparison, and outcome to be defined, not implied.
  • Eligibility rules: Are the inclusion and exclusion criteria written down before study selection? In the Handbook's framework, these come from the review's PICO. Check whether the paper applies the same criteria the protocol set.
  • Databases: Does the protocol name the sources to be searched? The Handbook notes the question guides searching. As an editorial suggestion, a named list lets you compare the plan with the methods actually reported.
  • Outcomes: Are the outcomes named in advance, and do they matter to the people the review is for? The Handbook expects outcomes to address issues important to consumers, health professionals, and policymakers, and expects potential adverse effects to be considered so summaries are not one-sided.
  • Analysis plan: Does the protocol say how results will be grouped and combined? The Handbook says the PICO for each synthesis is also planned at the protocol stage. It also says authors may build contingencies into the plan for sparse data, such as grouping elements more broadly.
  • Later changes: If the question or methods changed, does the paper say so and explain why? The Handbook says to document changes at the start of the Methods section, and it asks whether results from included studies could have influenced a change.

Changes are not automatically a problem

The Handbook says some refinement of a review question is expected as authors learn what evidence exists. It also warns that data-driven questions can produce false conclusions. What matters is whether authors disclose and justify changes, not whether the plan never moved. If a broad review is split into narrower ones, the Handbook says each component needs its own protocol documenting its eligibility criteria.

Limits of this check

  • A protocol or registry record shows a plan. On its own it does not show that the plan was sound or that it was followed. Comparing the two is the reader's job.
  • This check covers a review's planning, not the quality of the studies it includes. Other parts of the Handbook address risk of bias and certainty of evidence.
  • The Handbook chapter is written mainly for intervention reviews. Other review types may follow different conventions.
  • Registry rules, including who may register and when, can change. Confirm current requirements on the register itself.

Who should seek professional advice?

This article is educational and is not medical advice. If a review's conclusions might affect a treatment decision for you or someone in your care, discuss them with a qualified clinician who can weigh them against your situation. If you plan to conduct a systematic review, consider consulting an experienced review methodologist or an information specialist, and read the current registry guidance before you begin screening. See our Medical Disclaimer.

What you can do next

  1. Look in the paper for a registry number or a link to the protocol.
  2. Open the record, for example on the PROSPERO register, and read the question, eligibility rules, and outcomes.
  3. Compare them with the published paper using the six checks above.
  4. Note any differences and whether the authors explain them.
  5. If no protocol can be found, treat the review's methods with extra caution and look for other evidence.

For how this site handles sources and errors, see our Editorial Standards and Corrections Policy. This article has not been peer reviewed or clinically reviewed.

Sources