# Antidepressant discontinuation × the claim that depression inevitably relapses: evidence checklist

Verdict 2937 · Evidence Check

Stored-material review date: Not recorded  
Original evidence cut-off: 2026-08-26  
Review-criteria version: 1.0  
Public page/material preparation date: 2026-10-10

Clinical papers were not read directly; existing review summaries were checked. No new clinical-literature search was performed. Unverified does not mean that research is absent or that there is no effect.

## Assessment scope

- Population: Adults in UK primary care with remission and long-term use who were considering stopping; not extended to every individual or setting
- Ingredient, formulation and route: Original ANTLER allocation strategy of tapering to placebo; actual restarting and adherence distinguished from complete discontinuation
- Dose: Not recorded
- Comparator: Original arm results for maintenance versus tapering to placebo; individual inevitability is a separate question
- Measured outcome: Existing question ‘discontinuation inevitably causes relapse’, original wording on 44% without relapse within 52 weeks, and the gap in individual prediction
- Duration: Original observation over 52 weeks; not extended to lifelong freedom from relapse or safe discontinuation for an individual
- Questions outside this scope: Extending the original reduction in relapse within 52 weeks to complete prevention for every person or period, or inevitable relapse on discontinuation; Extending the original 44% without relapse to lifelong freedom from relapse, safe discontinuation for an individual, or complete discontinuation without restarting; Calculating correspondence or overlap by assuming the discontinuation group's 52% adherence, 39% restarting and 44% without relapse refer to the same people; Calculating new exclusion proportions, denominators, events or individual classifications from 478 randomized, 468 in the time-to-event analysis and 10 with missing timing (6/4); Recalculating or converting the original 92/238, 135/240, 39%/56%, 17%p, HR 2.06/CI/P or DESS means and difference into reciprocals, NNT or individual probabilities; Double-counting NEJM and HTA as independent trials, or adding participants across the two sheets, arms, time points or the 478/8890 participant figures; Using the 40-study meta-analysis as a new independent replication under ANTLER conditions, or certification of inevitability or individual prediction; Treating possible withdrawal misclassification or awareness of allocation as the cause of all results, a newly quantified bias, or invalidation of the trial; Extending 9/8 participants with serious events and 0 deaths or suicide attempts to zero risk, long-term safety or drug selection; Treating the absence of reported company donations in the original account as absence of support, employee authors or conflicts of interest, or certification of independence; Inheriting the parent's B70, axes, historical cross-checking, public funding or original-material review statements as subclaim values, review of content in this work or external reading; New searches or numerical calculations; hypothetical quantitative examples; new effects, CIs, grades or certification of independence; guidance on maintenance, discontinuation, tapering, dosing, drug changes or treatment

## Checklist items

### N01 · What is the existing verdict on ‘discontinuation inevitably causes relapse’?

Verification status: Partly verified  
Reading scope: Summary read; external material unread

- Current findings: The selected existing question is ‘discontinuation inevitably causes relapse’, originally grade D, with a note that 44% of the discontinuation group did not relapse within 52 weeks. Subclaim score, status and axes are absent and separate from the parent's B70 for average relapse reduction.
- Why unverified: Only relevant preserved-summary information has been recorded; detailed original materials remain unread and existing content review is incomplete. Subclaim score, status and axes were originally absent; no new verdict is written. Review of existing content remains incomplete.
- Reason for missing information: Subclaim score, status and axes were originally absent; no new verdict is written.
- Required materials and conditions: Distinguishing the subclaim's D grade/note from parent B70 and the absence of score, status and axes
- Materials to obtain first: Additional materials for the selected question or user feedback
- Further research: Proposal: Separate source records for the inevitability claim and the parent's average effect are proposed; no new investigation, calculation or trial was conducted.
- Conditions for verification: Relevant original materials and analysis scope for Distinguishing the subclaim's D grade/note from parent B70 and the absence of score, status and axes must be supplied and unverified matters distinguished. Preserved wording is linked during this work; review of existing content remains incomplete.
- Why this condition is needed: Production item to distinguish Distinguishing the subclaim's D grade/note from parent B70 and the absence of score, status and axes and link the materials required for the selected question to currently available materials.
- Application exceptions: Limited to the stored summary for the selected formulation, population, endpoint and period. It does not substitute for numbers from other formulations/trials/endpoints or unread external material.
- Remaining uncertainty: This record links preserved-summary statements; it does not establish external-paper comparison, a new efficacy verdict or completed content review.
- Referenced methods: Question-clarification principles in M03 §3.1, §3.2.1–3.2.4; detailed separation rules are our own design
- Related sources: S01, S02

### N02 · What period, people and exposure does ‘inevitably’ refer to?

Verification status: Partly verified  
Reading scope: Summary read; external material unread

- Current findings: The original selected wording is ‘discontinuation inevitably causes relapse’. ANTLER actually compared tapering to placebo in adults with long-term use and remission who were considering stopping, using first-relapse data over 52 weeks. The 52-week period is not extended to a lifetime or every form of discontinuation.
- Why unverified: Only relevant preserved-summary information has been recorded; detailed original materials remain unread and existing content review is incomplete. Materials establishing inevitability for every person, period or discontinuation method are absent from this preserved scope. Review of existing content remains incomplete.
- Reason for missing information: Materials establishing inevitability for every person, period or discontinuation method are absent from this preserved scope.
- Required materials and conditions: Distinguishing an all-person or inevitability claim from the actual 52-week taper-to-placebo conditions
- Materials to obtain first: Direct materials specifying period, population and actual exposure, if supplied
- Further research: Proposal: Materials separating the claim's time, population and exposure scope are proposed; no new investigation, calculation or trial was conducted.
- Conditions for verification: Relevant original materials and analysis scope for Distinguishing an all-person or inevitability claim from the actual 52-week taper-to-placebo conditions must be supplied and unverified matters distinguished. Preserved wording is linked during this work; review of existing content remains incomplete.
- Why this condition is needed: Production item to distinguish Distinguishing an all-person or inevitability claim from the actual 52-week taper-to-placebo conditions and link the materials required for the selected question to currently available materials.
- Application exceptions: Limited to the stored summary for the selected formulation, population, endpoint and period. It does not substitute for numbers from other formulations/trials/endpoints or unread external material.
- Remaining uncertainty: This record links preserved-summary statements; it does not establish external-paper comparison, a new efficacy verdict or completed content review.
- Referenced methods: M03 §3.1, §3.2.1–3.2.4: principles for clarifying the question; detailed separation rules are Chamgap's own design; M03 §3.2.1; detailed population conditions are Chamgap's own design; M03 §3.2.2–3.2.3; reverse labeling and test adaptations are Chamgap's own design; M03 §3.2.1–3.2.4; dose, duration and effect-modification details are Chamgap's own design
- Related sources: S01, S02

### N03 · Is the 44% without relapse stated directly in the original text?

Verification status: Partly verified  
Reading scope: Summary read; external material unread

- Current findings: The value that 44% of the discontinuation group did not relapse within 52 weeks is directly stated in the existing reader answer and subclaim note. It was not generated by calculating a new complementary proportion from 56%, and is not changed to absence of relapse over another period.
- Why unverified: Only relevant preserved-summary information has been recorded; detailed original materials remain unread and existing content review is incomplete. Individual adherence, restarting, censoring, clinical characteristics and subsequent course among those without relapse remain unverified. Review of existing content remains incomplete.
- Reason for missing information: Individual adherence, restarting, censoring, clinical characteristics and subsequent course among those without relapse remain unverified.
- Required materials and conditions: Preserving the original no-relapse wording, period and unit of the proportion
- Materials to obtain first: Materials on the no-relapse group, follow-up, actual exposure and timing
- Further research: Proposal: Materials linking observed absence of relapse with individual courses are proposed; no new investigation, calculation or trial was conducted.
- Conditions for verification: Relevant original materials and analysis scope for Preserving the original no-relapse wording, period and unit of the proportion must be supplied and unverified matters distinguished. Preserved wording is linked during this work; review of existing content remains incomplete.
- Why this condition is needed: Production item to distinguish Preserving the original no-relapse wording, period and unit of the proportion and link the materials required for the selected question to currently available materials.
- Application exceptions: Limited to the stored summary for the selected formulation, population, endpoint and period. It does not substitute for numbers from other formulations/trials/endpoints or unread external material.
- Remaining uncertainty: This record links preserved-summary statements; it does not establish external-paper comparison, a new efficacy verdict or completed content review.
- Referenced methods: M03 §3.2.4 and M15 §15.3.1–15.3.3; Question-specific outcome adaptation is Chamgap design; M08 Table 8.2.a, M15 §15.3 and M19 §19.1.2–19.2.4; Design-specific adaptation is Chamgap design; M03 §3.2.1–3.2.4; Dose, duration and effect-modifier details are Chamgap design
- Related sources: S01, S02

### N04 · Does 56% relapse in the discontinuation group mean ‘inevitably’?

Verification status: Partly verified  
Reading scope: Summary read; external material unread

- Current findings: ANTLER's original discontinuation result of 135/240 (56%), maintenance result of 92/238 (39%) and absolute difference of 17%p are average between-arm comparisons. They are not transferred to relapse in every individual who discontinues; the subclaim's D grade/note is preserved unchanged.
- Why unverified: Only relevant preserved-summary information has been recorded; detailed original materials remain unread and existing content review is incomplete. Individual probabilities, outcomes for all people and inevitability over other periods or populations remain unverified. Review of existing content remains incomplete.
- Reason for missing information: Individual probabilities, outcomes for all people and inevitability over other periods or populations remain unverified.
- Required materials and conditions: Distinguishing original 135/240 (56%), 92/238 (39%) and 17%p from inevitability wording
- Materials to obtain first: Individual follow-up or prediction, or corresponding direct materials
- Further research: Proposal: Separate materials on arm comparisons and claims about every person or individual are proposed; no new investigation, calculation or trial was conducted.
- Conditions for verification: Relevant original materials and analysis scope for Distinguishing original 135/240 (56%), 92/238 (39%) and 17%p from inevitability wording must be supplied and unverified matters distinguished. Preserved wording is linked during this work; review of existing content remains incomplete.
- Why this condition is needed: Production item to distinguish Distinguishing original 135/240 (56%), 92/238 (39%) and 17%p from inevitability wording and link the materials required for the selected question to currently available materials.
- Application exceptions: Limited to the stored summary for the selected formulation, population, endpoint and period. It does not substitute for numbers from other formulations/trials/endpoints or unread external material.
- Remaining uncertainty: This record links preserved-summary statements; it does not establish external-paper comparison, a new efficacy verdict or completed content review.
- Referenced methods: M03 §3.2.4, M15 §15.3.1–15.3.3; question-specific outcome adaptations are Chamgap's own design; M03 §3.2.4; Chamgap's own limits on extension
- Related sources: S01, S02

### N05 · Do the population and recruitment conditions include every patient with depression?

Verification status: Partly verified  
Reading scope: Summary read; external material unread

- Current findings: Adults from 150 UK general practices felt well enough to consider stopping. They had at least 2 previous episodes or at least 2 years of treatment, with specific study conditions involving long-term use of four drugs.
- Why unverified: Only relevant preserved-summary information has been recorded; detailed original materials remain unread and existing content review is incomplete. Exact remission criteria, age, baseline risk, exclusions and the course without relapse under other population conditions remain unverified. Review of existing content remains incomplete.
- Reason for missing information: Exact remission criteria, age, baseline risk, exclusions and the course without relapse under other population conditions remain unverified.
- Required materials and conditions: Conditions involving 150 practices, long-term medication, remission and 2 episodes/2 years
- Materials to obtain first: Eligibility, exclusions, baseline tables and direct materials under other conditions
- Further research: Proposal: Materials linking eligibility conditions with individual outcomes are proposed; no new investigation, calculation or trial was conducted.
- Conditions for verification: Relevant original materials and analysis scope for Conditions involving 150 practices, long-term medication, remission and 2 episodes/2 years must be supplied and unverified matters distinguished. Preserved wording is linked during this work; review of existing content remains incomplete.
- Why this condition is needed: Production item to distinguish Conditions involving 150 practices, long-term medication, remission and 2 episodes/2 years and link the materials required for the selected question to currently available materials.
- Application exceptions: Limited to the stored summary for the selected formulation, population, endpoint and period. It does not substitute for numbers from other formulations/trials/endpoints or unread external material.
- Remaining uncertainty: This record links preserved-summary statements; it does not establish external-paper comparison, a new efficacy verdict or completed content review.
- Referenced methods: M03 §3.2.1; Detailed population conditions are Chamgap design; M03 §3.2.1–3.2.4; Dose, duration and effect-modifier details are Chamgap design
- Related sources: S01, S02

### N06 · Is study ‘discontinuation’ the same as complete or immediate stopping?

Verification status: Partly verified  
Reading scope: Summary read; external material unread

- Current findings: The original trial switched users of specified drugs to matching placebo after tapering over 1~2 months. This is not converted into abrupt-stopping exposure, every drug or dose, or outcomes of individuals who completely stopped. Study conditions are not made into tapering or dosing instructions.
- Why unverified: Only relevant preserved-summary information has been recorded; detailed original materials remain unread and existing content review is incomplete. Individual actual tapering, exposure, concomitant treatment and stopping or restarting times remain unverified. Review of existing content remains incomplete.
- Reason for missing information: Individual actual tapering, exposure, concomitant treatment and stopping or restarting times remain unverified.
- Required materials and conditions: Historical scope of tapering over 1~2 months, switching to placebo and actual exposure
- Materials to obtain first: Records of actual exposure, tapering and concomitant treatment
- Further research: Proposal: Materials linking allocation conditions with actual behavior are proposed; no new investigation, calculation or trial was conducted.
- Conditions for verification: Relevant original materials and analysis scope for Historical scope of tapering over 1~2 months, switching to placebo and actual exposure must be supplied and unverified matters distinguished. Preserved wording is linked during this work; review of existing content remains incomplete.
- Why this condition is needed: Production item to distinguish Historical scope of tapering over 1~2 months, switching to placebo and actual exposure and link the materials required for the selected question to currently available materials.
- Application exceptions: Limited to the stored summary for the selected formulation, population, endpoint and period. It does not substitute for numbers from other formulations/trials/endpoints or unread external material.
- Remaining uncertainty: This record links preserved-summary statements; it does not establish external-paper comparison, a new efficacy verdict or completed content review.
- Referenced methods: M03 §3.2.2–3.2.3; Reverse-direction labels and assessment adaptation are Chamgap design; M03 §3.2.3 and M08 §8.2.2–8.2.3; Chamgap application
- Related sources: S01, S02

### N07 · Do 39% restarting and 52% adherence identify those without relapse?

Verification status: Partly verified  
Reading scope: Summary read; external material unread

- Current findings: Original values state 52% adherence in the discontinuation group and that 39% restarted a clinician-prescribed antidepressant before the trial ended. These people are not assumed to be the same as the 44% without relapse; no new overlap is calculated.
- Why unverified: Only relevant preserved-summary information has been recorded; detailed original materials remain unread and existing content review is incomplete. Individual cross-classification of restarting or adherence with relapse or absence of relapse, timing and reasons remains unverified. Review of existing content remains incomplete.
- Reason for missing information: Individual cross-classification of restarting or adherence with relapse or absence of relapse, timing and reasons remains unverified.
- Required materials and conditions: Relationship between actual exposure or restarting in the discontinuation group and the no-relapse group
- Materials to obtain first: Individual linked materials on relapse, restarting, adherence and time
- Further research: Proposal: Materials separating absence of relapse in the allocated group from actual exposure pathways are proposed; no new investigation, calculation or trial was conducted.
- Conditions for verification: Relevant original materials and analysis scope for Relationship between actual exposure or restarting in the discontinuation group and the no-relapse group must be supplied and unverified matters distinguished. Preserved wording is linked during this work; review of existing content remains incomplete.
- Why this condition is needed: Production item to distinguish Relationship between actual exposure or restarting in the discontinuation group and the no-relapse group and link the materials required for the selected question to currently available materials.
- Application exceptions: Limited to the stored summary for the selected formulation, population, endpoint and period. It does not substitute for numbers from other formulations/trials/endpoints or unread external material.
- Remaining uncertainty: This record links preserved-summary statements; it does not establish external-paper comparison, a new efficacy verdict or completed content review.
- Referenced methods: M03 §3.2.3, M08 §8.2.2–8.2.3; internal application; M08 §8.2.2, M15 §15.3.1; internal application; M08 Table 8.2.a, M15 §15.3, M19 §19.1.2–19.2.4; design-specific adaptation is internally designed
- Related sources: S01, S02

### N08 · Can individual outcomes after actual stopping be separated from the allocation effect?

Verification status: Partly verified  
Reading scope: Summary read; external material unread

- Current findings: The allocated comparison of 478 participants and original HR are study results for maintenance versus a taper-followed-by-placebo strategy. No new estimate is made of a pure effect among people who completely stopped, inevitable relapse, or the proportion without relapse and without restarting.
- Why unverified: Only relevant preserved-summary information has been recorded; detailed original materials remain unread and existing content review is incomplete. Causal or temporal analysis by actual exposure, crossover, rescue treatment and denominators remains unverified. Review of existing content remains incomplete.
- Reason for missing information: Causal or temporal analysis by actual exposure, crossover, rescue treatment and denominators remains unverified.
- Required materials and conditions: Distinguishing the randomized comparison's estimand from individual and actual-exposure effects
- Materials to obtain first: Materials on exposure, crossover and event timing, and the prespecified analysis plan
- Further research: Proposal: Materials distinguishing allocation effects from actual-exposure effects are proposed; no new investigation, calculation or trial was conducted.
- Conditions for verification: Relevant original materials and analysis scope for Distinguishing the randomized comparison's estimand from individual and actual-exposure effects must be supplied and unverified matters distinguished. Preserved wording is linked during this work; review of existing content remains incomplete.
- Why this condition is needed: Production item to distinguish Distinguishing the randomized comparison's estimand from individual and actual-exposure effects and link the materials required for the selected question to currently available materials.
- Application exceptions: Limited to the stored summary for the selected formulation, population, endpoint and period. It does not substitute for numbers from other formulations/trials/endpoints or unread external material.
- Remaining uncertainty: This record links preserved-summary statements; it does not establish external-paper comparison, a new efficacy verdict or completed content review.
- Referenced methods: M08 §8.2.2, M15 §15.3.1; our own application; M03 §3.2.3, M08 §8.2.2–8.2.3; our own application
- Related sources: S01, S02

### N09 · How does primary first relapse differ from symptom worsening and withdrawal?

Verification status: Partly verified  
Reading scope: Summary read; external material unread

- Current findings: ANTLER's primary endpoint is time to first depression relapse within 52 weeks, and the parent describes relapse assessed using a symptom scale. Withdrawal symptoms, other transient worsening and recurrent episodes are not merged into the same event as relapse or inevitability.
- Why unverified: Only relevant preserved-summary information has been recorded; detailed original materials remain unread and existing content review is incomplete. Details of the relapse instrument, threshold, assessment schedule, withdrawal distinction, first event and censoring remain unverified. Review of existing content remains incomplete.
- Reason for missing information: Details of the relapse instrument, threshold, assessment schedule, withdrawal distinction, first event and censoring remain unverified.
- Required materials and conditions: Event definitions for first relapse within 52 weeks, symptom instruments and withdrawal
- Materials to obtain first: Relapse and withdrawal definitions, assessment and adjudication materials, and SAP
- Further research: Proposal: Materials distinguishing events from symptoms and recurrent outcomes are proposed; no new investigation, calculation or trial was conducted.
- Conditions for verification: Relevant original materials and analysis scope for Event definitions for first relapse within 52 weeks, symptom instruments and withdrawal must be supplied and unverified matters distinguished. Preserved wording is linked during this work; review of existing content remains incomplete.
- Why this condition is needed: Production item to distinguish Event definitions for first relapse within 52 weeks, symptom instruments and withdrawal and link the materials required for the selected question to currently available materials.
- Application exceptions: Limited to the stored summary for the selected formulation, population, endpoint and period. It does not substitute for numbers from other formulations/trials/endpoints or unread external material.
- Remaining uncertainty: This record links preserved-summary statements; it does not establish external-paper comparison, a new efficacy verdict or completed content review.
- Referenced methods: M03 §3.2.4, M15 §15.3.1–15.3.3; outcome adaptation to each question is internally designed; M08 Table 8.2.a, M03 §3.2.4; internal application; M08 Table 8.2.a, M04 §4.6.3; internal application
- Related sources: S01, S02

### N10 · How are 478 randomized, 468 in time-to-event analysis and 10 with missing timing related?

Verification status: Partly verified  
Reading scope: Summary read; external material unread

- Current findings: The preserved wording states 478 randomized and 468 in the primary time-to-event analysis, with 10 lacking timing (6 maintenance/4 discontinuation) excluded from the primary analysis but included in absolute relapse counts. Missing-data participants are not newly classified as all having relapsed or all having had no relapse.
- Why unverified: Only relevant preserved-summary information has been recorded; detailed original materials remain unread and existing content review is incomplete. Exact timing, reasons for missingness, individual exposure and censoring for the 10 participants, and follow-up details for those without relapse remain unverified. Review of existing content remains incomplete.
- Reason for missing information: Exact timing, reasons for missingness, individual exposure and censoring for the 10 participants, and follow-up details for those without relapse remain unverified.
- Required materials and conditions: 238 maintenance/240 discontinuation, missing timing in 6/4, and denominator-specific scope
- Materials to obtain first: Flow diagram and records of events, missing timing and follow-up
- Further research: Proposal: Materials linking relapse or no-relapse observations by denominator are proposed; no new investigation, calculation or trial was conducted.
- Conditions for verification: Relevant original materials and analysis scope for 238 maintenance/240 discontinuation, missing timing in 6/4, and denominator-specific scope must be supplied and unverified matters distinguished. Preserved wording is linked during this work; review of existing content remains incomplete.
- Why this condition is needed: Production item to distinguish 238 maintenance/240 discontinuation, missing timing in 6/4, and denominator-specific scope and link the materials required for the selected question to currently available materials.
- Application exceptions: Limited to the stored summary for the selected formulation, population, endpoint and period. It does not substitute for numbers from other formulations/trials/endpoints or unread external material.
- Remaining uncertainty: This record links preserved-summary statements; it does not establish external-paper comparison, a new efficacy verdict or completed content review.
- Referenced methods: M08 Table 8.2.a, M15 §15.3 and M19 §19.1.2–19.2.4; Design-specific adaptation is Chamgap design; M08 §8.2.2 and M15 §15.3.1; Chamgap application
- Related sources: S01, S02

### N11 · Do best- and worst-case missing-data assumptions test relapse in everyone?

Verification status: Partly verified  
Reading scope: Summary read; external material unread

- Current findings: The original statement that best- and worst-case missing-data analyses remained in the same direction concerns the arm comparison. It is not converted into relapse in everyone, individual discontinuation outcomes, or resolution of all missing-data bias.
- Why unverified: Only relevant preserved-summary information has been recorded; detailed original materials remain unread and existing content review is incomplete. Actual values, CIs and denominators under each assumption, pathways without relapse and temporal materials remain unverified. Review of existing content remains incomplete.
- Reason for missing information: Actual values, CIs and denominators under each assumption, pathways without relapse and temporal materials remain unverified.
- Required materials and conditions: Distinguishing the original direction of sensitivity analyses from the inevitability question
- Materials to obtain first: Sensitivity tables, assumptions and prespecified plan
- Further research: Proposal: Separate materials on sensitivity of arm comparisons and individual inevitability are proposed; no new investigation, calculation or trial was conducted.
- Conditions for verification: Relevant original materials and analysis scope for Distinguishing the original direction of sensitivity analyses from the inevitability question must be supplied and unverified matters distinguished. Preserved wording is linked during this work; review of existing content remains incomplete.
- Why this condition is needed: Production item to distinguish Distinguishing the original direction of sensitivity analyses from the inevitability question and link the materials required for the selected question to currently available materials.
- Application exceptions: Limited to the stored summary for the selected formulation, population, endpoint and period. It does not substitute for numbers from other formulations/trials/endpoints or unread external material.
- Remaining uncertainty: This record links preserved-summary statements; it does not establish external-paper comparison, a new efficacy verdict or completed content review.
- Referenced methods: M08 Table 8.2.a, M15 §15.3 and M19 §19.1.2–19.2.4; Design-specific adaptation is Chamgap design; M08 §8.2.2 and M15 §15.3.1; Chamgap application; M15 §15.3.1–15.3.3; Chamgap application
- Related sources: S01, S02

### N12 · Is the discontinuation-versus-maintenance HR 2.06 a probability of inevitability?

Verification status: Partly verified  
Reading scope: Summary read; external material unread

- Current findings: The original discontinuation-versus-maintenance HR 2.06 (1.56~2.70), P<.001 is a time-to-event comparison. It is not changed into a value meaning 100% probability of relapse or relapse in a particular person. No maintenance-versus-discontinuation reciprocal, new RR or effect is calculated.
- Why unverified: Only relevant preserved-summary information has been recorded; detailed original materials remain unread and existing content review is incomplete. Model, censoring, proportional hazards, individual prediction and CIs for absolute outcomes remain unverified. Review of existing content remains incomplete.
- Reason for missing information: Model, censoring, proportional hazards, individual prediction and CIs for absolute outcomes remain unverified.
- Required materials and conditions: Separate units for HR, CI, P, event proportions and individual probabilities
- Materials to obtain first: Original HR model and SAP, and individual prediction materials
- Further research: Proposal: Materials linking comparative estimands and individual prediction are proposed; no new investigation, calculation or trial was conducted.
- Conditions for verification: Relevant original materials and analysis scope for Separate units for HR, CI, P, event proportions and individual probabilities must be supplied and unverified matters distinguished. Preserved wording is linked during this work; review of existing content remains incomplete.
- Why this condition is needed: Production item to distinguish Separate units for HR, CI, P, event proportions and individual probabilities and link the materials required for the selected question to currently available materials.
- Application exceptions: Limited to the stored summary for the selected formulation, population, endpoint and period. It does not substitute for numbers from other formulations/trials/endpoints or unread external material.
- Remaining uncertainty: This record links preserved-summary statements; it does not establish external-paper comparison, a new efficacy verdict or completed content review.
- Referenced methods: M03 §3.2.4, M15 §15.3.1–15.3.3; question-specific outcome adaptations are Chamgap's own design; M08 §8.2.2, M15 §15.3.1; Chamgap's own application; M15 §15.3.1–15.3.3; Chamgap's own application
- Related sources: S01, S02

### N13 · What gaps do withdrawal symptoms and awareness of allocation leave in an inevitability interpretation?

Verification status: Partly verified  
Reading scope: Summary read; external material unread

- Current findings: The preserved wording gives 12-week DESS means of 3.1 versus 1.3, the original difference of 1.9 (1.5~2.3), and greater awareness of allocation in the discontinuation group. Some withdrawal may have been misclassified as relapse, but the magnitude is undetermined. No new conclusion is made that all relapse was withdrawal.
- Why unverified: Only relevant preserved-summary information has been recorded; detailed original materials remain unread and existing content review is incomplete. Misclassified cases, original model and denominators, allocation-guessing figures, and temporal relationships between withdrawal and relapse remain unverified. Review of existing content remains incomplete.
- Reason for missing information: Misclassified cases, original model and denominators, allocation-guessing figures, and temporal relationships between withdrawal and relapse remain unverified.
- Required materials and conditions: Original scope of 12-week modified DESS, possible misclassification and allocation awareness
- Materials to obtain first: Withdrawal and relapse items, original tables, masking and adjudication records
- Further research: Proposal: Materials distinguishing withdrawal from relapse and recording allocation awareness are proposed; no new investigation, calculation or trial was conducted.
- Conditions for verification: Relevant original materials and analysis scope for Original scope of 12-week modified DESS, possible misclassification and allocation awareness must be supplied and unverified matters distinguished. Preserved wording is linked during this work; review of existing content remains incomplete.
- Why this condition is needed: Production item to distinguish Original scope of 12-week modified DESS, possible misclassification and allocation awareness and link the materials required for the selected question to currently available materials.
- Application exceptions: Limited to the stored summary for the selected formulation, population, endpoint and period. It does not substitute for numbers from other formulations/trials/endpoints or unread external material.
- Remaining uncertainty: This record links preserved-summary statements; it does not establish external-paper comparison, a new efficacy verdict or completed content review.
- Referenced methods: M08 Table 8.2.a, M03 §3.2.4; internal application; M19 §19.1.2–19.2.4; finer distinctions are internally designed; design-specific harm questions in M19 §19.2; details of temporal relationships are internally designed
- Related sources: S01, S02

### N14 · Is 39% relapse during maintenance a different question from inevitability after discontinuation?

Verification status: Partly verified  
Reading scope: Summary read; external material unread

- Current findings: The parent and related wording that 39% relapsed despite maintenance concerns the complete-prevention question. It is not transferred to the denominator of ‘discontinuation inevitably causes relapse’ or to relapse outcomes for other individuals.
- Why unverified: Only relevant preserved-summary information has been recorded; detailed original materials remain unread and existing content review is incomplete. Individual pathways during maintenance or discontinuation and original subgroup prediction materials remain unverified. Review of existing content remains incomplete.
- Reason for missing information: Individual pathways during maintenance or discontinuation and original subgroup prediction materials remain unverified.
- Required materials and conditions: Boundary between the other arm's result and the related complete-prevention question
- Materials to obtain first: Individual courses by arm and direct materials for the separate question
- Further research: Proposal: Materials separating average reduction, complete prevention and inevitability claims are proposed; no new investigation, calculation or trial was conducted.
- Conditions for verification: Relevant original materials and analysis scope for Boundary between the other arm's result and the related complete-prevention question must be supplied and unverified matters distinguished. Preserved wording is linked during this work; review of existing content remains incomplete.
- Why this condition is needed: Production item to distinguish Boundary between the other arm's result and the related complete-prevention question and link the materials required for the selected question to currently available materials.
- Application exceptions: Limited to the stored summary for the selected formulation, population, endpoint and period. It does not substitute for numbers from other formulations/trials/endpoints or unread external material.
- Remaining uncertainty: This record links preserved-summary statements; it does not establish external-paper comparison, a new efficacy verdict or completed content review.
- Referenced methods: Question-clarification principles in M03 §3.1, §3.2.1–3.2.4; detailed separation rules are internally designed; M04 §4.6.3, M15 §15.3.1–15.3.3; status separation is internally designed
- Related sources: S01, S02

### N15 · Does absence of relapse within 52 weeks certify lifelong freedom from relapse or safe discontinuation?

Verification status: Partly verified  
Reading scope: Summary read; external material unread

- Current findings: The 44% without relapse is original observational wording within 52 weeks. It is not extended to lifelong freedom from relapse, safe discontinuation for every person or a guarantee of individual treatment choice. The original account states that the trial did not identify who could discontinue safely.
- Why unverified: Only relevant preserved-summary information has been recorded; detailed original materials remain unread and existing content review is incomplete. Long-term follow-up, individual risk, withdrawal, restarting and other safety outcomes remain unverified. Review of existing content remains incomplete.
- Reason for missing information: Long-term follow-up, individual risk, withdrawal, restarting and other safety outcomes remain unverified.
- Required materials and conditions: Actual observation period and gaps beyond observation or in individual safety
- Materials to obtain first: Follow-up beyond observation and individual prediction or safety materials
- Further research: Proposal: Separate materials on gaps beyond the time or individual scope are proposed; no new investigation, calculation or trial was conducted.
- Conditions for verification: Relevant original materials and analysis scope for Actual observation period and gaps beyond observation or in individual safety must be supplied and unverified matters distinguished. Preserved wording is linked during this work; review of existing content remains incomplete.
- Why this condition is needed: Production item to distinguish Actual observation period and gaps beyond observation or in individual safety and link the materials required for the selected question to currently available materials.
- Application exceptions: Limited to the stored summary for the selected formulation, population, endpoint and period. It does not substitute for numbers from other formulations/trials/endpoints or unread external material.
- Remaining uncertainty: This record links preserved-summary statements; it does not establish external-paper comparison, a new efficacy verdict or completed content review.
- Referenced methods: M03 §3.2.1–3.2.4; dose, duration and effect-modification details are our own design; M19 §19.1.2–19.2.4; own application
- Related sources: S01, S02

### N16 · Can those who will remain without relapse be identified beforehand?

Verification status: Unverified  
Reading scope: Summary read; external material unread

- Current findings: The preserved wording states that ANTLER reported average relapse reduction and did not distinguish people who could discontinue safely. No individual prediction rule is created from original drug, history or duration conditions alone.
- Why unverified: The information currently required remains unverified in the preserved summary. Validated individual models, prespecified subgroups and external validation in other populations remain unverified in these authoring materials. Review of existing content remains incomplete.
- Reason for missing information: Validated individual models, prespecified subgroups and external validation in other populations remain unverified in these authoring materials.
- Required materials and conditions: Distinguishing group comparisons from validated individual prediction and subgroup information
- Materials to obtain first: Individual prediction, prespecified subgroup or external-population materials, if supplied
- Further research: Proposal: Separate materials linking average effects and individual prediction are proposed; no new investigation, calculation or trial was conducted.
- Conditions for verification: Relevant original materials and analysis scope for Distinguishing group comparisons from validated individual prediction and subgroup information must be supplied and unverified matters distinguished. Preserved wording is linked during this work; review of existing content remains incomplete.
- Why this condition is needed: Production item to distinguish Distinguishing group comparisons from validated individual prediction and subgroup information and link the materials required for the selected question to currently available materials.
- Application exceptions: Limited to the stored summary for the selected formulation, population, endpoint and period. It does not substitute for numbers from other formulations/trials/endpoints or unread external material.
- Remaining uncertainty: This record links preserved-summary statements; it does not establish external-paper comparison, a new efficacy verdict or completed content review.
- Referenced methods: M03 §3.2.1–3.2.4; dose, duration and effect-modification details are our own design
- Related sources: S01, S02

### N17 · Are the NEJM and HTA citations independent trials?

Verification status: Partly verified  
Reading scope: Summary read; external material unread

- Current findings: S01 NEJM PMID34587384 and S02 NIHR HTA PMID34842135 report the same ANTLER trial. They share the first evidence record's 478 and 468 participants and withdrawal, adherence and missing-data results. The two reports are not counted as 2 independent trials or two participant populations.
- Why unverified: Only relevant preserved-summary information has been recorded; detailed original materials remain unread and existing content review is incomplete. Original tables, analyses and correspondence between the two reports remain externally unread. Review of existing content remains incomplete.
- Reason for missing information: Original tables, analyses and correspondence between the two reports remain externally unread.
- Required materials and conditions: Relationship between the same ANTLER trial, the first evidence record, and the main and detailed reports
- Materials to obtain first: Linkage table for shared results and trial registration across the two reports
- Further research: Proposal: Materials on multiple-report relationships within the same trial are proposed; no new investigation, calculation or trial was conducted.
- Conditions for verification: Relevant original materials and analysis scope for Relationship between the same ANTLER trial, the first evidence record, and the main and detailed reports must be supplied and unverified matters distinguished. Preserved wording is linked during this work; review of existing content remains incomplete.
- Why this condition is needed: Production item to distinguish Relationship between the same ANTLER trial, the first evidence record, and the main and detailed reports and link the materials required for the selected question to currently available materials.
- Application exceptions: Limited to the stored summary for the selected formulation, population, endpoint and period. It does not substitute for numbers from other formulations/trials/endpoints or unread external material.
- Remaining uncertainty: This record links preserved-summary statements; it does not establish external-paper comparison, a new efficacy verdict or completed content review.
- Referenced methods: Distinction between studies and reports in M04 §4.6.1–4.6.2; cohort/outcome linkage is an internal application; data collection and selection principles in M04 §4.5–4.6; separation of access from reading and the hyperlink contract are internally designed
- Related sources: S01, S02

### N18 · Do average meta-analysis results directly answer ‘inevitably’?

Verification status: Partly verified  
Reading scope: Summary read; external material unread

- Current findings: Kato's maintenance result of 20.9% versus placebo 39.7%, OR .38 (.33~.43), P<.00001 synthesizes varied maintenance trials. An average relapse difference is not transferred to direct evidence that every person relapses after stopping. The preserved wording that does not count it as an independent replication of ANTLER's primary-care decision is retained.
- Why unverified: Only relevant preserved-summary information has been recorded; detailed original materials remain unread and existing content review is incomplete. Periods, populations, events, inclusion denominators, overlap, individual prediction and heterogeneity remain unverified. Review of existing content remains incomplete.
- Reason for missing information: Periods, populations, events, inclusion denominators, overlap, individual prediction and heterogeneity remain unverified.
- Required materials and conditions: Differences between Kato's 40 trials and 8890 participants, proportions or OR, and the actual original question
- Materials to obtain first: Inclusion and exclusion tables, trial characteristics, periods, denominators and individual materials
- Further research: Proposal: Materials separating pooled averages from claims about every person or individual are proposed; no new investigation, calculation or trial was conducted.
- Conditions for verification: Relevant original materials and analysis scope for Differences between Kato's 40 trials and 8890 participants, proportions or OR, and the actual original question must be supplied and unverified matters distinguished. Preserved wording is linked during this work; review of existing content remains incomplete.
- Why this condition is needed: Production item to distinguish Differences between Kato's 40 trials and 8890 participants, proportions or OR, and the actual original question and link the materials required for the selected question to currently available materials.
- Application exceptions: Limited to the stored summary for the selected formulation, population, endpoint and period. It does not substitute for numbers from other formulations/trials/endpoints or unread external material.
- Remaining uncertainty: This record links preserved-summary statements; it does not establish external-paper comparison, a new efficacy verdict or completed content review.
- Referenced methods: M03 §3.2.4 and M15 §15.3.1–15.3.3: question-specific outcome adaptation is in-house; M04 §4.6.1–4.6.2 and M15 §15.3.1: in-house application; M03 §3.2.4: in-house extension boundaries.
- Related sources: S03

### N19 · Do people and reports overlap across the review, ANTLER and related sheets?

Verification status: Partly verified  
Reading scope: Summary read; external material unread

- Current findings: This sheet shares the same ANTLER trial and review summary as the 52-week relapse-reduction sheet; NEJM and HTA report the same trial. The 40 trials with 8890 participants, 478 participants and two sheets or reports are not added as independent trials or total participants.
- Why unverified: Only relevant preserved-summary information has been recorded; detailed original materials remain unread and existing content review is incomplete. Unique clinical totals and review inclusion or overlap relationships are not established and remain unrecorded. Review of existing content remains incomplete.
- Reason for missing information: Unique clinical totals and review inclusion or overlap relationships are not established and remain unrecorded.
- Required materials and conditions: Shared preserved summary, two subclaim sheets, and report, trial and synthesis relationships
- Materials to obtain first: Trial, report, population and review inclusion lists, if supplied
- Further research: Proposal: Records of material families and overlap relationships are proposed; no new investigation, calculation or trial was conducted.
- Conditions for verification: Relevant original materials and analysis scope for Shared preserved summary, two subclaim sheets, and report, trial and synthesis relationships must be supplied and unverified matters distinguished. Preserved wording is linked during this work; review of existing content remains incomplete.
- Why this condition is needed: Production item to distinguish Shared preserved summary, two subclaim sheets, and report, trial and synthesis relationships and link the materials required for the selected question to currently available materials.
- Application exceptions: Limited to the stored summary for the selected formulation, population, endpoint and period. It does not substitute for numbers from other formulations/trials/endpoints or unread external material.
- Remaining uncertainty: This record links preserved-summary statements; it does not establish external-paper comparison, a new efficacy verdict or completed content review.
- Referenced methods: Study/report distinction in M04 §4.6.1–4.6.2; Cohort/outcome linkage is Chamgap application; M04 §4.6.1–4.6.2 and M15 §15.3.1; Chamgap application
- Related sources: S01, S02, S03

### N20 · Have registration and prespecified endpoints actually been read?

Verification status: Partly verified  
Reading scope: Summary read; external material unread

- Current findings: ANTLER's ‘ISRCTN15969819 · EudraCT2015-004210-26’ is preserved unchanged. Kato's ‘Not applicable’ is preserved review wording. Identifiers or wording alone do not establish that registration or plans were read, or that prespecified individual prediction was validated.
- Why unverified: Only relevant preserved-summary information has been recorded; detailed original materials remain unread and existing content review is incomplete. Registration amendment history, protocol, SAP, subgroup and missing-data plans remain unread; access was not attempted. Review of existing content remains incomplete.
- Reason for missing information: Registration amendment history, protocol, SAP, subgroup and missing-data plans remain unread; access was not attempted.
- Required materials and conditions: Original ISRCTN/EudraCT identifiers and unread registration or plans
- Materials to obtain first: Registration or plan copies, or user-supplied materials
- Further research: Proposal: Links between prespecified plans, versions and reports are proposed; no new investigation, calculation or trial was conducted.
- Conditions for verification: Relevant original materials and analysis scope for Original ISRCTN/EudraCT identifiers and unread registration or plans must be supplied and unverified matters distinguished. Preserved wording is linked during this work; review of existing content remains incomplete.
- Why this condition is needed: Production item to distinguish Original ISRCTN/EudraCT identifiers and unread registration or plans and link the materials required for the selected question to currently available materials.
- Application exceptions: Limited to the stored summary for the selected formulation, population, endpoint and period. It does not substitute for numbers from other formulations/trials/endpoints or unread external material.
- Remaining uncertainty: This record links preserved-summary statements; it does not establish external-paper comparison, a new efficacy verdict or completed content review.
- Referenced methods: M04 §4.5–4.6 material collection/selection principles; acquisition/reading separation and hyperlink specifications are internally designed; M04 §4.5, §4.6.3; limits of this material conversion follow the review scope and internal design; M08 Table 8.2.a, M04 §4.6.3; internal application
- Related sources: S01, S02, S03

### N21 · Do public support and unreported product or COI information certify independence?

Verification status: Partly verified  
Reading scope: Summary read; external material unread

- Current findings: ANTLER's NIHR HTA 13/115/48, PRIMENT core support and UCL sponsor, and Kato's Japanese research grants are preserved unchanged. The historical wording that neither report described donated trial drug or placebo is not changed into actual absence or certification of independence in this work.
- Why unverified: Only relevant preserved-summary information has been recorded; detailed original materials remain unread and existing content review is incomplete. Support and product contracts, employee authors, conflicts of interest, study and manuscript roles, and support for included trials in the review remain unverified. Review of existing content remains incomplete.
- Reason for missing information: Support and product contracts, employee authors, conflicts of interest, study and manuscript roles, and support for included trials in the review remain unverified.
- Required materials and conditions: NIHR, PRIMENT, UCL, product provision, employees, conflicts of interest and review funding
- Materials to obtain first: Original materials on support, provision, authors, contributions and conflicts of interest
- Further research: Proposal: Materials on funding, support types and actual roles are proposed; no new investigation, calculation or trial was conducted.
- Conditions for verification: Relevant original materials and analysis scope for NIHR, PRIMENT, UCL, product provision, employees, conflicts of interest and review funding must be supplied and unverified matters distinguished. Preserved wording is linked during this work; review of existing content remains incomplete.
- Why this condition is needed: Production item to distinguish NIHR, PRIMENT, UCL, product provision, employees, conflicts of interest and review funding and link the materials required for the selected question to currently available materials.
- Application exceptions: Limited to the stored summary for the selected formulation, population, endpoint and period. It does not substitute for numbers from other formulations/trials/endpoints or unread external material.
- Remaining uncertainty: This record links preserved-summary statements; it does not establish external-paper comparison, a new efficacy verdict or completed content review.
- Referenced methods: Chamgap's own conditions for independence and attribution. The official methodological sections used in this example do not validate every detail of these items.
- Related sources: S01, S02, S03

### N22 · Are withdrawal, serious events and absence of relapse the same safety outcome?

Verification status: Partly verified  
Reading scope: Summary read; external material unread

- Current findings: Original serious adverse events of 9 maintenance participants (4%)/8 discontinuation participants (3%), 0 deaths or suicide attempts in both groups, and increased withdrawal at 12 weeks are recorded separately. Absence of relapse is not converted into no adverse effects, safe discontinuation or 0 long-term harm.
- Why unverified: Only relevant preserved-summary information has been recorded; detailed original materials remain unread and existing content review is incomplete. Event definitions, collection, safety denominators, exposure, follow-up, CIs and long-term or drug-specific harms remain unverified. Review of existing content remains incomplete.
- Reason for missing information: Event definitions, collection, safety denominators, exposure, follow-up, CIs and long-term or drug-specific harms remain unverified.
- Required materials and conditions: Scope of harm collection, 12-week withdrawal, 9/8 participants and 0 deaths or suicide attempts
- Materials to obtain first: Endpoint-specific safety, withdrawal, exposure, denominator and follow-up materials
- Further research: Proposal: Materials separating absence of relapse from other harm outcomes are proposed; no new investigation, calculation or trial was conducted.
- Conditions for verification: Relevant original materials and analysis scope for Scope of harm collection, 12-week withdrawal, 9/8 participants and 0 deaths or suicide attempts must be supplied and unverified matters distinguished. Preserved wording is linked during this work; review of existing content remains incomplete.
- Why this condition is needed: Production item to distinguish Scope of harm collection, 12-week withdrawal, 9/8 participants and 0 deaths or suicide attempts and link the materials required for the selected question to currently available materials.
- Application exceptions: Addresses gaps in withdrawal and safety materials for the selected relapse question. Relapse, absence of relapse or original safety wording is not used to assess overall safety or an individual's ability to discontinue, or to generate harm incidence or guidance on maintenance, discontinuation, tapering, dosing or treatment.
- Remaining uncertainty: This record links preserved-summary statements; it does not establish external-paper comparison, a new efficacy verdict or completed content review.
- Referenced methods: M19 §19.1.2–19.2.4; internal application; M19 §19.1.2–19.2.4; finer distinctions are internally designed; M19 §19.1.2–19.2.4; internal application; M19 §19.1.2–19.2.4; internal application; Chamgap's own interpretation boundary. M19 principles are referenced; no official causality tool was executed
- Related sources: S01, S02

### N23 · Does this answer questions about other drugs, abrupt stopping or new treatment settings?

Verification status: Partly verified  
Reading scope: Summary read; external material unread

- Current findings: The original materials' four drugs, specified doses and study taper over 1~2 months are historical exposure. No new verdict is made on inevitability after abrupt stopping, other drugs or doses, acute treatment, other diseases or ages, or every treatment setting.
- Why unverified: Only relevant preserved-summary information has been recorded; detailed original materials remain unread and existing content review is incomplete. Direct relapse and safety materials under conditions outside the scope remain unverified. Review of existing content remains incomplete.
- Reason for missing information: Direct relapse and safety materials under conditions outside the scope remain unverified.
- Required materials and conditions: Original four drugs, study taper, remission, long-term use and questions outside that scope
- Materials to obtain first: Direct materials under separate conditions, if supplied
- Further research: Proposal: Separate questions and material scopes by exposure and population are proposed; no new investigation, calculation or trial was conducted.
- Conditions for verification: Relevant original materials and analysis scope for Original four drugs, study taper, remission, long-term use and questions outside that scope must be supplied and unverified matters distinguished. Preserved wording is linked during this work; review of existing content remains incomplete.
- Why this condition is needed: Production item to distinguish Original four drugs, study taper, remission, long-term use and questions outside that scope and link the materials required for the selected question to currently available materials.
- Application exceptions: Limited to the stored summary for the selected formulation, population, endpoint and period. It does not substitute for numbers from other formulations/trials/endpoints or unread external material.
- Remaining uncertainty: This record links preserved-summary statements; it does not establish external-paper comparison, a new efficacy verdict or completed content review.
- Referenced methods: M03 §3.2.1; detailed population conditions are our own design; M03 §3.2.2–3.2.3; reverse-direction labels and test adaptations are our own design; M03 §3.2.1–3.2.4; dose, duration and effect-modification details are our own design
- Related sources: S01, S02

### N24 · What preserved summaries and external materials have actually been read?

Verification status: Partly verified  
Reading scope: Summary read; external material unread

- Current findings: ‘Summary read’ means the complete preserved Korean and English texts, original figures, bibliography and subclaims were used for production. External papers, registration, protocol, SAP, support and COI materials remain ‘not read’; historical statements about reading clinical originals or URLs are not inherited as new reading.
- Why unverified: Only relevant preserved-summary information has been recorded; detailed original materials remain unread and existing content review is incomplete. External access, link or click tests and comparison against original external materials were not performed. Review of existing content remains incomplete.
- Reason for missing information: External access, link or click tests and comparison against original external materials were not performed.
- Required materials and conditions: Preserved summary, original NEJM/HTA/Kato URLs and unread external materials
- Materials to obtain first: User-provided original materials and actual reading records.
- Further research: Proposal: Records of acquisition, reading and versions by material are proposed; no new investigation, calculation or trial was conducted.
- Conditions for verification: Relevant original materials and analysis scope for Preserved summary, original NEJM/HTA/Kato URLs and unread external materials must be supplied and unverified matters distinguished. Preserved wording is linked during this work; review of existing content remains incomplete.
- Why this condition is needed: Production item to distinguish Preserved summary, original NEJM/HTA/Kato URLs and unread external materials and link the materials required for the selected question to currently available materials.
- Application exceptions: Limited to the stored summary for the selected formulation, population, endpoint and period. It does not substitute for numbers from other formulations/trials/endpoints or unread external material.
- Remaining uncertainty: This record links preserved-summary statements; it does not establish external-paper comparison, a new efficacy verdict or completed content review.
- Referenced methods: M04 §4.5–4.6 material collection/selection principles; acquisition/reading separation and hyperlink specifications are internally designed; M04 §4.5, §4.6.3; limits of this material conversion follow the review scope and internal design
- Related sources: S01, S02, S03

### N25 · Are original subclaim D, parent B70, Korean/English texts, dates and empty values preserved?

Verification status: Partly verified  
Reading scope: Summary read; external material unread

- Current findings: Selected D grade/note and English wording, parent B70, Caution, Korean and English texts, axes with unrecorded precision, false, 2026-08-26, and empty correction and deliberation values are preserved unchanged. Absent subclaim score, status and axes are not filled from parent values.
- Why unverified: Only relevant preserved-summary information has been recorded; detailed original materials remain unread and existing content review is incomplete. The historical production did not translate the verdict text, create a new verdict or correction, or update external evidence. Review of existing content remains incomplete.
- Reason for missing information: The historical production did not translate the verdict text, create a new verdict or correction, or update external evidence.
- Required materials and conditions: Whole original preservation, absence, unrecorded precision, false values and correction history
- Materials to obtain first: Correction originals or user comments
- Further research: Proposal: Links between original versions, dates and the selected subclaim sheet are proposed; no new investigation, calculation or trial was conducted.
- Conditions for verification: Relevant original materials and analysis scope for Whole original preservation, absence, unrecorded precision, false values and correction history must be supplied and unverified matters distinguished. Preserved wording is linked during this work; review of existing content remains incomplete.
- Why this condition is needed: Production item to distinguish Whole original preservation, absence, unrecorded precision, false values and correction history and link the materials required for the selected question to currently available materials.
- Application exceptions: Limited to the stored summary for the selected formulation, population, endpoint and period. It does not substitute for numbers from other formulations/trials/endpoints or unread external material.
- Remaining uncertainty: This record links preserved-summary statements; it does not establish external-paper comparison, a new efficacy verdict or completed content review.
- Referenced methods: M03 §3.1, §3.2.1–3.2.4 question-clarification principles; detailed separation rules are internally designed; M04 §4.5, §4.6.3; limits of this material conversion follow the review scope and internal design
- Related sources: S01, S02, S03

### N26 · Do applied conditions and completed production mean existing content review is complete?

Verification status: Partly verified  
Reading scope: Summary read; external material unread

- Current findings: Preservation, freezing and application of conditions, case production and content review are separate records. Existing content review remains incomplete; counts of written records are not totals of unique clinical trials or participants.
- Why unverified: Only relevant preserved-summary information has been recorded; detailed original materials remain unread and existing content review is incomplete. Existing content review remains incomplete. Additional content review, testing, screen checks and link tests were not conducted during the original production. Review of existing content remains incomplete.
- Reason for missing information: Existing content review remains incomplete. Additional content review, testing, screen checks and link tests were not conducted during the original production.
- Required materials and conditions: Distinguishing preservation, freezing and application of conditions, case production and incomplete content review
- Materials to obtain first: Comments on existing content review and additional preserved materials
- Further research: Proposal: A history of production versions, conditions and user corrections is proposed; no new investigation, calculation or trial was conducted.
- Conditions for verification: Relevant original materials and analysis scope for Distinguishing preservation, freezing and application of conditions, case production and incomplete content review must be supplied and unverified matters distinguished. Preserved wording is linked during this work; review of existing content remains incomplete.
- Why this condition is needed: Production item to distinguish Distinguishing preservation, freezing and application of conditions, case production and incomplete content review and link the materials required for the selected question to currently available materials.
- Application exceptions: Limited to the stored summary for the selected formulation, population, endpoint and period. It does not substitute for numbers from other formulations/trials/endpoints or unread external material.
- Remaining uncertainty: This record links preserved-summary statements; it does not establish external-paper comparison, a new efficacy verdict or completed content review.
- Referenced methods: M04 §4.5–4.6 material collection/selection principles; acquisition/reading separation and hyperlink specifications are internally designed; M04 §4.5, §4.6.3; limits of this material conversion follow the review scope and internal design
- Related sources: S01, S02, S03

## Evidence and sources

### S01

Lewis G, Marston L, Duffy L, et al. Maintenance or Discontinuation of Antidepressants in Primary Care. N Engl J Med. 2021;385:1257-1267. PMID: 34587384. DOI: 10.1056/NEJMoa2106356.

[Publisher](https://doi.org/10.1056/NEJMoa2106356) · [PubMed](https://pubmed.ncbi.nlm.nih.gov/34587384/)

Materials actually read: Summary read; external material unread  
Findings recorded in the summary: Relapse: 92/238 (39%) versus 135/240 (56%); discontinuation versus maintenance HR 2.06 (95% CI 1.56~2.70), P<0.001; absolute difference 17%p. The existing reader answer and subclaim state that 44% of the discontinuation group had no relapse within 52 weeks and that 39% of the maintenance group relapsed. The 10 participants with missing timing were excluded from the primary analysis but included in absolute counts. Adherence of 70%/52%, restarting in 39% of the discontinuation group, and 12-week DESS means of 3.1/1.3 with the original difference of 1.9 (CI 1.5~2.3) are preserved separately. No new proportions, differences, individual probabilities, reciprocals, effects, CIs or grades are calculated.  
Record level: Outcome wording in the stored summary  
Independent replication: S01 NEJM/Lewis2021 PMID34587384 is the main ANTLER report. The first record in the existing evidence list describes both reports together. They share 478 randomized participants (238 maintenance/240 taper followed by placebo), 468 in the time-to-event analysis, 10 with missing timing (6/4), absolute relapse counts, withdrawal, adherence and restarting. Reports, related sheets, arms, time points and people are not added as new independent trials or participants. Counts of report or material slots are not totals of unique clinical trials, participants or independent replications. Unique clinical totals remain unrecorded. S02 NIHR HTA/Duffy2021 PMID34842135 is a detailed report of the same ANTLER trial. The first record in the existing evidence list describes both reports together. They share 478 randomized participants (238 maintenance/240 taper followed by placebo), 468 in the time-to-event analysis, 10 with missing timing (6/4), absolute relapse counts, withdrawal, adherence and restarting. Reports, related sheets, arms, time points and people are not added as new independent trials or participants. Counts of report or material slots are not totals of unique clinical trials, participants or independent replications. Unique clinical totals remain unrecorded. S03 Kato2021 PMID32704061 is a meta-analysis of 40 trials with 8,890 participants comparing maintenance with switching to placebo after various acute treatments. The original wording that does not count it as an independent replication of ANTLER's long-term primary-care discontinuation decision is preserved. Inclusion, overlap, timing and unique trial totals remain unverified; participants in the review, ANTLER and related sheets are not added together. Counts of report or material slots are not totals of unique clinical trials, participants or independent replications. Unique clinical totals remain unrecorded.

| Material | Public availability | Access status | Actually read | Notes |
|---|---|---|---|---|
| Stored summary | Stored file | Stored summary | Summary read | Only the stored JSON summary and bibliography were used for production. This does not constitute external verification or completed content review. |
| External paper | Public availability unverified; only bibliography links preserved | Unread; access not attempted | Not read | No external access or click testing was performed in the original preparation. URLs or registration identifiers alone do not upgrade acquisition/reading status. |
| Registration material | Public availability unverified; only bibliography links preserved | Unread; access not attempted | Not read | No external access or click testing was performed in the original preparation. URLs or registration identifiers alone do not upgrade acquisition/reading status. |
| Protocol | Public availability unverified; only bibliography links preserved | Unread; access not attempted | Not read | No external access or click testing was performed in the original preparation. URLs or registration identifiers alone do not upgrade acquisition/reading status. |
| Statistical analysis plan | Public availability unverified; only bibliography links preserved | Unread; access not attempted | Not read | No external access or click testing was performed in the original preparation. URLs or registration identifiers alone do not upgrade acquisition/reading status. |
| Funding, product-support and COI material | Public availability unverified; only bibliography links preserved | Unread; access not attempted | Not read | No external access or click testing was performed in the original preparation. URLs or registration identifiers alone do not upgrade acquisition/reading status. |

### S02

Duffy L, Clarke CS, Lewis G, et al. Antidepressant medication to prevent depression relapse in primary care: the ANTLER RCT. Health Technol Assess. 2021;25(69):1-62. PMID: 34842135. DOI: 10.3310/hta25690.

[Publisher](https://doi.org/10.3310/hta25690) · [PubMed](https://pubmed.ncbi.nlm.nih.gov/34842135/)

Materials actually read: Summary read; external material unread  
Findings recorded in the summary: Relapse: 92/238 (39%) versus 135/240 (56%); discontinuation versus maintenance HR 2.06 (95% CI 1.56~2.70), P<0.001; absolute difference 17%p. The existing reader answer and subclaim state that 44% of the discontinuation group had no relapse within 52 weeks and that 39% of the maintenance group relapsed. The 10 participants with missing timing were excluded from the primary analysis but included in absolute counts. Adherence of 70%/52%, restarting in 39% of the discontinuation group, and 12-week DESS means of 3.1/1.3 with the original difference of 1.9 (CI 1.5~2.3) are preserved separately. No new proportions, differences, individual probabilities, reciprocals, effects, CIs or grades are calculated.  
Record level: Outcome wording in the stored summary  
Independent replication: S01 NEJM/Lewis2021 PMID34587384 is the main ANTLER report. The first record in the existing evidence list describes both reports together. They share 478 randomized participants (238 maintenance/240 taper followed by placebo), 468 in the time-to-event analysis, 10 with missing timing (6/4), absolute relapse counts, withdrawal, adherence and restarting. Reports, related sheets, arms, time points and people are not added as new independent trials or participants. Counts of report or material slots are not totals of unique clinical trials, participants or independent replications. Unique clinical totals remain unrecorded. S02 NIHR HTA/Duffy2021 PMID34842135 is a detailed report of the same ANTLER trial. The first record in the existing evidence list describes both reports together. They share 478 randomized participants (238 maintenance/240 taper followed by placebo), 468 in the time-to-event analysis, 10 with missing timing (6/4), absolute relapse counts, withdrawal, adherence and restarting. Reports, related sheets, arms, time points and people are not added as new independent trials or participants. Counts of report or material slots are not totals of unique clinical trials, participants or independent replications. Unique clinical totals remain unrecorded. S03 Kato2021 PMID32704061 is a meta-analysis of 40 trials with 8,890 participants comparing maintenance with switching to placebo after various acute treatments. The original wording that does not count it as an independent replication of ANTLER's long-term primary-care discontinuation decision is preserved. Inclusion, overlap, timing and unique trial totals remain unverified; participants in the review, ANTLER and related sheets are not added together. Counts of report or material slots are not totals of unique clinical trials, participants or independent replications. Unique clinical totals remain unrecorded.

| Material | Public availability | Access status | Actually read | Notes |
|---|---|---|---|---|
| Stored summary | Stored file | Stored summary | Summary read | Only the stored JSON summary and bibliography were used for production. This does not constitute external verification or completed content review. |
| External paper | Public availability unverified; only bibliography links preserved | Unread; access not attempted | Not read | No external access or click testing was performed in the original preparation. URLs or registration identifiers alone do not upgrade acquisition/reading status. |
| Registration material | Public availability unverified; only bibliography links preserved | Unread; access not attempted | Not read | No external access or click testing was performed in the original preparation. URLs or registration identifiers alone do not upgrade acquisition/reading status. |
| Protocol | Public availability unverified; only bibliography links preserved | Unread; access not attempted | Not read | No external access or click testing was performed in the original preparation. URLs or registration identifiers alone do not upgrade acquisition/reading status. |
| Statistical analysis plan | Public availability unverified; only bibliography links preserved | Unread; access not attempted | Not read | No external access or click testing was performed in the original preparation. URLs or registration identifiers alone do not upgrade acquisition/reading status. |
| Funding, product-support and COI material | Public availability unverified; only bibliography links preserved | Unread; access not attempted | Not read | No external access or click testing was performed in the original preparation. URLs or registration identifiers alone do not upgrade acquisition/reading status. |

### S03

Kato M, Hori H, Inoue T, et al. Discontinuation of antidepressants after remission with antidepressant medication in major depressive disorder: a systematic review and meta-analysis. Mol Psychiatry. 2021;26:118-133. PMID: 32704061. DOI: 10.1038/s41380-020-0843-0.

[Publisher](https://doi.org/10.1038/s41380-020-0843-0) · [PubMed](https://pubmed.ncbi.nlm.nih.gov/32704061/)

Materials actually read: Summary read; external material unread  
Findings recorded in the summary: Maintenance 20.9% versus switching to placebo 39.7%, OR 0.38 (95% CI 0.33~0.43), P<0.00001. The original proportions, OR and direction are preserved within the scope of this synthesis. This estimand differs from the ANTLER HR and 17%p; no new reciprocal, pooled effect, period-specific incidence, total participant count or individual prediction is generated.  
Record level: Outcome wording in the stored summary  
Independent replication: S01 NEJM/Lewis2021 PMID34587384 is the main ANTLER report. The first record in the existing evidence list describes both reports together. They share 478 randomized participants (238 maintenance/240 taper followed by placebo), 468 in the time-to-event analysis, 10 with missing timing (6/4), absolute relapse counts, withdrawal, adherence and restarting. Reports, related sheets, arms, time points and people are not added as new independent trials or participants. Counts of report or material slots are not totals of unique clinical trials, participants or independent replications. Unique clinical totals remain unrecorded. S02 NIHR HTA/Duffy2021 PMID34842135 is a detailed report of the same ANTLER trial. The first record in the existing evidence list describes both reports together. They share 478 randomized participants (238 maintenance/240 taper followed by placebo), 468 in the time-to-event analysis, 10 with missing timing (6/4), absolute relapse counts, withdrawal, adherence and restarting. Reports, related sheets, arms, time points and people are not added as new independent trials or participants. Counts of report or material slots are not totals of unique clinical trials, participants or independent replications. Unique clinical totals remain unrecorded. S03 Kato2021 PMID32704061 is a meta-analysis of 40 trials with 8,890 participants comparing maintenance with switching to placebo after various acute treatments. The original wording that does not count it as an independent replication of ANTLER's long-term primary-care discontinuation decision is preserved. Inclusion, overlap, timing and unique trial totals remain unverified; participants in the review, ANTLER and related sheets are not added together. Counts of report or material slots are not totals of unique clinical trials, participants or independent replications. Unique clinical totals remain unrecorded.

| Material | Public availability | Access status | Actually read | Notes |
|---|---|---|---|---|
| Stored summary | Stored file | Stored summary | Summary read | Only the stored JSON summary and bibliography were used for production. This does not constitute external verification or completed content review. |
| External paper | Public availability unverified; only bibliography links preserved | Unread; access not attempted | Not read | No external access or click testing was performed in the original preparation. URLs or registration identifiers alone do not upgrade acquisition/reading status. |
| Registration material | Public availability unverified; only bibliography links preserved | Unread; access not attempted | Not read | No external access or click testing was performed in the original preparation. URLs or registration identifiers alone do not upgrade acquisition/reading status. |
| Protocol | Public availability unverified; only bibliography links preserved | Unread; access not attempted | Not read | No external access or click testing was performed in the original preparation. URLs or registration identifiers alone do not upgrade acquisition/reading status. |
| Statistical analysis plan | Public availability unverified; only bibliography links preserved | Unread; access not attempted | Not read | No external access or click testing was performed in the original preparation. URLs or registration identifiers alone do not upgrade acquisition/reading status. |
| Funding, product-support and COI material | Public availability unverified; only bibliography links preserved | Unread; access not attempted | Not read | No external access or click testing was performed in the original preparation. URLs or registration identifiers alone do not upgrade acquisition/reading status. |

## Review criteria and methods

These are Chamgap's own application conditions, informed by the relevant sections of official methodological guidance. They do not constitute official assessment-tool certification or a new efficacy score.

- M03 [Cochrane Handbook 6.5 (2024), chapter updated 2023-08](https://www.cochrane.org/authors/handbooks-and-manuals/handbook/current/chapter-03) · §3.1, §3.2.1–3.2.3, MECIR C5/C7/C13 · Clearly record the question, eligibility conditions and reasons for changes, and distinguish variations in interventions. · Preserved method HTML copies; no current method-content re-review or official tool execution
- M04 [Cochrane Handbook 6.5.1 (2025), chapter updated 2025-03](https://www.cochrane.org/authors/handbooks-and-manuals/handbook/current/chapter-04) · §4.5, §4.6.1–4.6.3, MECIR C42 · Distinguish studies from reports and link them; do not omit a study solely because results are unreported. · Preserved method HTML copies; no current method-content re-review or official tool execution
- M08 [Cochrane Handbook 6.5 (2024), chapter updated 2019-10](https://www.cochrane.org/authors/handbooks-and-manuals/handbook/current/chapter-08) · §8.1–8.2.3, Table 8.2.a · Define the outcome and effect of interest, then examine randomization, deviations, missing data, measurement and selective reporting. · Preserved method HTML copies; no current method-content re-review or official tool execution
- M15 [Cochrane Handbook 6.5 (2024), chapter updated 2023-08](https://www.cochrane.org/authors/handbooks-and-manuals/handbook/current/chapter-15) · §15.2.4, §15.3.1–15.3.3 · Distinguish effect estimates, intervals, importance and applicability. Sample size alone does not determine precision. · Preserved method HTML copies; no current method-content re-review or official tool execution
- M19 [Cochrane Handbook 6.5 (2024), chapter updated 2019-10](https://www.cochrane.org/authors/handbooks-and-manuals/handbook/current/chapter-19) · §19.1.2, §19.2.1–19.2.4 · Harm questions may require different designs, populations and follow-up from efficacy questions. Case reports do not establish incidence rates. · Preserved method HTML copies; no current method-content re-review or official tool execution

## What the statuses mean

- Met: Status for necessary evidence confirmed after existing content review. No completion is declared in the current presentation.
- Partly verified: Some information in the stored summary is linked into the table. External comparison/content review is not complete.
- Unverified: The article leaves this unverified, or required material is absent from the current stored summary. This is distinct from zero, absence or safety.
- Review incomplete: Existing content review incomplete.
- Not reported: Not reported within the stored material actually read; not extended to the entire primary paper.
- Inaccessible: Means an actual access failure. Access not attempted here is not assigned this state.
- Not applicable: Not used as a substitute for an evidence gap.
- Conflicting data: Requires an observed disagreement under the same comparison criteria; not automatically judged in this production.
- Insufficient evidence for the conditions: Marks a gap in additional material required by the selected conditions; not a new efficacy grade.

## Change history

2026-10-10 · Public page/materials 1.0. The reader view and materials were prepared from the existing 26 items. No new clinical search, verification-status changes or scoring were performed.