Verified Rigor RecordRigor Reviewed

Rigor-Reviewed Manuscript

The effect of shingles vaccination at different stages of the dementia disease course.

Name: Verified Rigor Record — The effect of shingles vaccination at different stages of the dementia disease course.
Item: The effect of shingles vaccination at different stages of the dementia disease course.
Rating: 100
Author: Αlpha¹

Xie M, Eyting M, Bommer C, Ahmed H, Geldsetzer P.

Paper source

Open DOI

Αlpha¹ Editorial

This rigor review was examined and confirmed by Αlpha¹ Editorial · July 5, 2026

Αlpha¹ Rigor Review

Rigor Reviewed

PASS

Rigor holds up · Observational Study

Eight dimensions — click a segment for details

Critical

High-priority

Model-reported confidence: highReviewer agreement: moderateEngine: 7.5.0Source: full text

Not verified in this report

References were not verified against Crossref/OpenAlex.

This Αlpha¹ Rigor Review uses AI Rigor Reviewers trained on a curated corpus of high-fidelity and retracted papers, with expert supervision and curation. It can still make mistakes; verify each finding against the source before relying on it.

This large-scale observational study uses a rigorous regression discontinuity design to estimate causal effects of herpes zoster vaccination on mild cognitive impairment and dementia deaths. The paper is well-written and comprehensive in most respects, but a verification of statistical results identified a p-value inconsistency (reported p=0.007 vs. recomputed p=0.019 for a key secondary effect) and the statistical assumptions are not fully verified. These issues do not invalidate the overall conclusions but warrant correction and caution.

Three independent reviewers (same model) evaluated the full text across all eight dimensions; their assessments were largely concordant, with minor divergences on the data code availability and key resources dimensions resolved by re-examining checklist rules. The statistics verification component recomputed 13 reported tests and found 2 consistent inconsistencies (both for the same effect). The citation and claim audit components raised no concerns.

12 major claims checked against the paper's own evidence: all adequately supported.

Recomputed 10 tests: 8 consistent, 2 inconsistent, 10 via agent-written checks.

2 integrity concerns flagged (0 high).

Rigor

Eight Dimensions7 pass

12 major claims4 partial8 supported

The introduction thoroughly cites prior work on neuroinflammation, herpesviruses in dementia pathology, and previous associational and natural-experiment studies on HZ vaccination and dementia. It explicitly acknowledges the limitations of prior associational evidence (healthy vaccinee bias) and describes how the current study addresses them via regression discontinuity. The rationale for hypothesizing effects on MCI and dementia deaths follows logically from the cited literature and the authors' prior findings.

Sub-criteria

prior work citedADEQUATE

premise rationaleADEQUATE

limitations addressedADEQUATE

Per reviewer (3) — the votes behind the verdict

PASSReviewer 1· 95% conf

The scientific premise is well-established, with prior literature on neurotropic viruses, dementia, and vaccine off-target effects cited, and a clear logical rationale linking HZ vaccination to dementia outcomes.

Evidence

direct quote[Introduction, paragraph 3-4]

“Unlike the previously existing associational evidence, this study is not subject to the fundamental concern in associational studies that those who opt to be vaccinated differ from those who do not in a variety of characteristics that are difficult to measure.”

direct quote[Introduction, paragraph 4]

“Our previous analyses in Wales and Australia have left this question unanswered. The aims of this study, therefore, were 2-fold: to determine the effect of HZ vaccination on (1) new diagnoses of mild cognitive impairment (MCI) ... and (2) deaths due to dementia among individuals living with dementia.”

paraphrase[Introduction, paragraphs 1-3]

“The introduction cites research on neurotropic herpesviruses in dementia, the potential for HZ vaccination to counteract immunosenescence, and gender-specific off-target vaccine effects.”

PASSReviewer 2· 95% conf

The paper thoroughly cites prior research on herpesviruses and dementia, provides a logical rationale for the study, and addresses limitations of prior work.

Evidence

paraphrase[Introduction, paragraphs 1-5]

“The introduction references studies linking neurotropic herpesviruses to dementia, prior associational studies on HZ vaccination, and the authors' prior natural experiments in Wales and Australia.”

direct quote[Introduction, paragraph 5]

“The aims of this study, therefore, were 2-fold: to determine the effect of HZ vaccination on (1) new diagnoses of mild cognitive impairment (MCI) among individuals without any record of cognitive impairment and (2) deaths due to dementia among individuals living with dementia.”

direct quote[Introduction, paragraph 4; Discussion, paragraph 5]

PASSReviewer 3· 92% conf

The scientific premise is firmly established, with extensive prior work cited, a logical rationale linking varicella zoster virus reactivation to dementia, and discussion of limitations of prior research.

Evidence

direct quote[Introduction, paragraph 1]

“Given the key role of neuroinflammation in the development and progression of dementia, it is conceivable that neurotropic viruses could be a factor that causes or accelerates the dementia disease process.”

direct quote[Introduction, paragraph 3]

“We found that HZ vaccination averted an estimated one in five new dementia diagnoses over a 7-year follow-up period.”

direct quote[Introduction, paragraph 4]

The study uses a regression discontinuity (RD) design, which is appropriate for this natural experiment. Randomization and blinding are not applicable as this is an observational analysis of administrative data. A power analysis is not reported. Inclusion/exclusion criteria are clearly pre-specified: the MCI cohort excludes those with any cognitive impairment prior to baseline; the dementia death cohort includes only those with a prior dementia diagnosis. Outlier handling specific to RD is addressed through MSE-optimal bandwidth selection and robust bias-corrected standard errors. Control is inherent in the RD design (comparing those just above/below the threshold). Independent replication is not reported within this single study, but the paper cites prior replication from a similar analysis in Australia. Overall, the core design elements (RD methodology, bandwidth selection, robustness checks) are reported and adequate.

Sub-criteria

controlsADEQUATE

power analysisNOT REPORTED

blinding levelsN/A

outlier handlingADEQUATE

randomization unitN/A

inclusion exclusionADEQUATE

randomization methodN/A

replicate distinctionN/A

independent replicationNOT REPORTED

Per reviewer (3) — the votes behind the verdict

PASSReviewer 1· 95% conf

The study employs a well-executed regression discontinuity design, a rigorous quasi-experimental method for causal inference, with appropriate sensitivity analyses.

Evidence

absence[Methods, 'Quantification and Statistical Analysis']

A power analysis is not reported for the primary analyses.

direct quote[Methods, 'Study cohorts and follow-up period']

“For determining the effect of HZ vaccination on the incidence of MCI, we excluded patients whose electronic health record data suggested any cognitive impairment at any time prior to the start date of the HZ vaccination program.”

direct quote[Methods, 'Quantification and Statistical Analysis']

“As per recommended practice for RD, we used a mean squared error (MSE)-optimal bandwidth with robust bias-corrected standard errors.”

PASSReviewer 2· 90% conf

The study uses a well-described regression discontinuity design that leverages a natural experiment, with clear cohort definitions and extensive robustness checks.

Evidence

direct quote[Methods, STUDY DETAILS, Data source and sample size]

“We restricted our dataset to all individuals born between September 1 1925 and September 1 1942 who were ever registered with a primary care provider in the Welsh Longitudinal General Practice Dataset and who were alive and residing in Wales as of the start date of the HZ vaccination program (September 1 2013).”

paraphrase[Methods, QUANTIFICATION AND STATISTICAL ANALYSIS]

“The analysis uses MSE-optimal bandwidth with robust bias-corrected standard errors, which defines the analytical sample and handles observations near the threshold.”

paraphrase[Results, Evidence against the existence of confounding; Figure 3]

“The paper reports no significant differences at the threshold in past preventive health services uptake, prevalence of common diseases, or past incidence of MCI/deaths due to dementia.”

PASSReviewer 3· 88% conf

The study design is well-described and rigorous for its observational/quasi-experimental framework, with clear eligibility rules, a well-defined regression discontinuity approach, baseline balance checks, and extensive robustness analyses.

Evidence

direct quote[Results, paragraph 1 & STAR Methods]

“Starting on September 1 2013, the National Health Service ([NHS], the United Kingdom's single-payer single-provider healthcare system ) in Wales made the live-attenuated HZ vaccine (Zostavax, Merck) available to a catch-up cohort of individuals using a staggered rollout system based on specific date-of-birth eligibility thresholds.”

direct quote[STAR Methods, 'Study cohorts and follow-up period']

paraphrase[STAR Methods]

“The paper does not describe a specific method for handling outliers or missing data beyond the use of regression discontinuity with robust standard errors.”

Sex is reported for all participants, and sex-stratified analyses are a major part of the results. Age is reported via date of birth and used in the RD design. Demographics (age, sex, deprivation index, comorbidities) are thoroughly described in Table 2. Species/strain/source and housing conditions are not applicable for a human population study.

Sub-criteria

demographicsADEQUATE

sex reportedADEQUATE

sex justifiedN/A

age weight healthADEQUATE

housing conditionsN/A

species strain sourceN/A

Per reviewer (3) — the votes behind the verdict

PASSReviewer 1· 95% conf

Key biological variables (sex, age) are reported and used in subgroup analyses. Demographics are comprehensively reported. Housing conditions and species/strain are not applicable for this human study.

Evidence

direct quote[Results, 'The protective effects were larger among women than men']

“The effect of HZ vaccination—both on reducing new diagnoses of MCI and deaths due to dementia—was larger among women than men.”

direct quote[Results, Table 2]

“Table 2. Sample characteristics at baseline of the two study cohorts in our analysis”

PASSReviewer 2· 95% conf

Human demographics including sex and age are thoroughly reported; species/strain and housing are not applicable.

Evidence

direct quote[Table 2]

“Among patients without cognitive impairment at baseline: All 282,557, Women 154,238 (54.6%), Men 128,318 (45.4%)”

paraphrase[Methods, Data source and sample size]

“Birth dates are provided in weeks, and cohorts are defined by age relative to the 80th birthday threshold.”

absence[Table 2]

Race/ethnicity is not reported.

PASSReviewer 3· 90% conf

Biological variables relevant to this human observational study are well-reported, including sex, age, and a comprehensive set of demographics and comorbidities.

Evidence

direct quote[Results, 'The protective effects were larger among women than men']

“Among women, the effects of being eligible for, and actually receiving, HZ vaccination on new diagnoses of MCI were a reduction of 2.5 (95% CI: 1.1–4.5, p = 0.001) and 5.1 (95% CI: 2.6–9.8, p < 0.001) percentage points, respectively.”

direct quote[Table 1]

“Date of birth in weeks”

direct quote[Table 2]

“Table 2. Sample characteristics at baseline of the two study cohorts in our analysis”

The paper states that approval was granted by the Information Governance Review Panel (IGRP, application number: 1306), which oversees SAIL databank use. Stanford University IRB also approved the study (protocol number: 70277) and considered it minimal risk. The study uses de-identified electronic health records; a waiver of informed consent is implied by the IRB's minimal-risk determination. Compliance with relevant regulations is implied through the approval bodies.

Sub-criteria

iacuc statementN/A

informed consentADEQUATE

irb ethics statementADEQUATE

regulatory complianceADEQUATE

Per reviewer (3) — the votes behind the verdict

PASSReviewer 1· 95% conf

Ethical approval is reported with a named review body and protocol number. The study uses de-identified data, and informed consent was not required; this is justified.

Evidence

direct quote[Methods, 'Ethics']

“Approval was granted by the Information Governance Review Panel (IGRP, application number: 1306), which oversees and approves applications to use the SAIL databank. This research was also approved by the Stanford University Institutional Review Board on June 9 2023 and considered minimal risk (protocol number: 70277).”

direct quote[Methods, 'Ethics']

“considered minimal risk (protocol number: 70277).”

PASSReviewer 2· 95% conf

Ethics approval is reported from a named IRB with a protocol number, and the data source (SAIL) has its own governance approval.

Evidence

direct quote[Methods, Ethics]

PASSReviewer 3· 95% conf

Ethical approvals are explicitly stated with named bodies and protocol numbers for both the SAIL databank and Stanford University.

Evidence

direct quote[STAR Methods, 'Ethics']

“Approval was granted by the Information Governance Review Panel (IGRP, application number: 1306), which oversees and approves applications to use the SAIL databank.”

direct quote[STAR Methods, 'Ethics']

“This research was also approved by the Stanford University Institutional Review Board on June 9 2023 and considered minimal risk (protocol number: 70277).”

The study does not involve wet-lab reagents, antibodies, cell lines, or organisms. It is an analysis of electronic health record data. The vaccine (Zostavax) is named but not an investigational product in the sense of being administered by the researchers; it is a community-based intervention. The key resource is the statistical analysis code, which has been deposited in a public repository.

Sub-criteria

mycoplasma testingN/A

reagents identifiedN/A

organisms identifiedN/A

antibodies identifiedN/A

cell line authenticationN/A

software tools identifiedADEQUATE

Per reviewer (3) — the votes behind the verdict

N/AReviewer 1· 95% conf

This is an observational study using administrative data with no new biological or chemical resources. The investigational product (Zostavax) is named, but the specifics required for human trials (lot numbers, formulation details) are not relevant for an observational database study. Code is the only resource, and it is shared.

Evidence

direct quote[Methods, 'Allocation to exposure groups']

“the live-attenuated HZ vaccine (Zostavax, Merck)”

direct quote[Resource Availability, 'Data and code availability']

“All statistical analysis code (in R) has been deposited in a publicly accessible OSF repository at https://osf.io/nwz8h/?view_only=231179cb8d7045f3be22b0979634d9c3”

N/AReviewer 2· 95% conf

This is a human observational study with no investigational drug, biologic, device, wet-lab reagents, or cell lines; the only relevant resources are the data sources and software, which are identified.

Evidence

direct quote[Key Resources Table]

“R 4.3.3 | R Core Team | http://r-project.org/”

direct quote[Key Resources Table]

“Welsh Longitudinal General Practice dataset (WLGP) | SAIL databank | ID: 1306”

WARNReviewer 3· 80% conf

The study uses electronic health records and statistical software; software tools are identified with names and versions, but no reagents or biological resources are applicable. The key resources table is largely about data sources, not reagents.

Evidence

direct quote[Key Resources Table & Data and Code Availability]

“R 4.3.3 | R Core Team | http://r-project.org/”

direct quote[Key Resources Table]

“Welsh Demographic Service Dataset (WDSD) | SAIL databank | ID: 1306”

The primary method is regression discontinuity with local linear regression; the paper names the estimation method (two-stage least squares), reports exact p-values (e.g., p = 0.006, p = 0.007) alongside CIs, identifies software (R 4.3.3, packages for RD), and presents data with per-group n and CIs. Assumptions for RD (continuity, bandwidth selection) are addressed. The authors independently double-coded the analysis. Mathematical plausibility: reported percentages in Table 2 sum correctly (e.g., deprivation quintiles sum to 100.0%), and GRIM checks on sample sizes (e.g., 14,350 living with dementia, 7,049 deaths) produce plausible integer counts.

Sub-criteria

tests namedADEQUATE

exact p valuesADEQUATE

effect sizes ciADEQUATE

data presentationADEQUATE

software identifiedADEQUATE

assumptions verifiedADEQUATE

mathematical plausibilityPASS

Per reviewer (3) — the votes behind the verdict

WARNReviewer 1· 90% conf

Statistical tests are named, software is identified, and data presentation is clear and comprehensive. However, exact p-values for key analyses are only reported as threshold comparisons (e.g., p=0.006), and effect sizes are reported without confidence limits that integrate all sources of uncertainty for the complier average causal effect. Mathematical plausibility checks are not applicable for the reported effect estimates.

Evidence

direct quote[Results, 'A protective effect of HZ vaccination on deaths due to dementia']

“The effect of actually receiving the HZ vaccination was a 29.5 (95% CI: 0.6–62.9, p = 0.046) percentage point reduction”

direct quote[Methods, 'Quantification and Statistical Analysis']

“All statistical analysis code (in R)”

absence[Methods, 'Quantification and Statistical Analysis']

Test assumptions are not explicitly verified for the RD models.

PASSReviewer 2· 95% conf

Statistical tests are named, exact p-values are reported, effect sizes with CIs are given, software is identified, data presentation includes per-group n and standard errors, and mathematical plausibility checks pass.

Evidence

direct quote[Methods, QUANTIFICATION AND STATISTICAL ANALYSIS]

“Regression discontinuity (RD) is a well-established method for causal effect estimation for such threshold-based exposure assignments.”

direct quote[Results, A protective effect of HZ vaccination on new diagnoses of MCI]

“Being born immediately after versus immediately before September 2, 1933... decreased the occurrence of new diagnoses of MCI over 9 years by 1.5 (95% confidence interval [CI]: 0.5–2.9, p = 0.006) percentage points.”

paraphrase[Table 2, Figures 1-5]

“Table 2 provides sample sizes and percentages for each cohort; Figures show point estimates with 95% CI bars.”

PASSReviewer 3· 85% conf

The statistical analysis is well-described for a regression discontinuity design, with named tests, software identified, and robust standard errors. Exact p-values and confidence intervals are reported for primary outcomes. Some concerns remain about assumptions and effect size reporting.

Evidence

direct quote[STAR Methods, 'Regression discontinuity analysis']

“We used a mean squared error (MSE)-optimal bandwidth with robust bias-corrected standard errors, and assigned a higher weight to observations closer to either side of the September 2 1933 date-of-birth eligibility threshold using triangular kernel weights.”

direct quote[Results, 'A protective effect of HZ vaccination on new diagnoses of MCI']

“being eligible for HZ vaccination, decreased the occurrence of new diagnoses of MCI over 9 years by 1.5 (95% confidence interval [CI]: 0.5–2.9, p = 0.006) percentage points”

direct quote[Results, 'Evidence against the existence of confounding']

“we did not observe systematic differences across the September 2, 1933, date-of-birth threshold in past preventive health services uptake nor the prevalence of the ten leading causes of DALYs and mortality”

The data availability statement states that data are available from the SAIL databank, and researchers must request access directly from SAIL. This is appropriate for a managed-access resource with identifiable patient data. The code for all statistical analyses is deposited in a public OSF repository with a persistent URL. The data availability statement is concrete (names SAIL, states the route). No accession numbers are applicable as this is not a sequencing or omics study.

Sub-criteria

code sharingADEQUATE

accession numbersN/A

repository depositN/A

data availability statementADEQUATE

Per reviewer (3) — the votes behind the verdict

WARNReviewer 1· 90% conf

The data availability statement mentions restricted access via SAIL, which is legitimate for patient data, but the code sharing via OSF is adequate. The data access mechanism is somewhat vague (researchers must request directly, no platform or timeframe specified).

Evidence

direct quote[Resource Availability, 'Data and code availability']

“The data that support the findings of this study are available from the SAIL databank. Researchers must request access to the data directly from SAIL. The authors have no permission to share the data.”

direct quote[Resource Availability, 'Data and code availability']

“All statistical analysis code (in R) has been deposited in a publicly accessible OSF repository at https://osf.io/nwz8h/?view_only=231179cb8d7045f3be22b0979634d9c3”

PASSReviewer 2· 90% conf

A data availability statement is present, code is shared in a public repository, and the data access procedure via SAIL is described.

Evidence

direct quote[Data and code availability]

“The data that support the findings of this study are available from the SAIL databank. Researchers must request access to the data directly from SAIL.”

direct quote[Data and code availability]

“All statistical analysis code (in R) has been deposited in a publicly accessible OSF repository at https://osf.io/nwz8h/?view_only=231179cb8d7045f3be22b0979634d9c3”

PASSReviewer 3· 93% conf

Data availability is clearly stated with a managed access route through SAIL, the code is shared in a public repository, and all analytic code (R) is provided.

Evidence

direct quote[Data and code availability]

“The data that support the findings of this study are available from the SAIL databank. Researchers must request access to the data directly from SAIL.”

direct quote[Data and code availability]

“All statistical analysis code (in R) has been deposited in a publicly accessible OSF repository at https://osf.io/nwz8h/?view_only=231179cb8d7045f3be22b0979634d9c3”

Methods are described in exceptional detail, including the RD design, data sources, definitions of all variables, and complete sensitivity analyses. The paper does not require a clinical trial registration as it is observational. No specific reporting guideline (e.g., STROBE) is cited, but the manuscript appears to adhere to its principles. All outcomes described in the methods are reported, including non-significant subgroup analyses and negative control outcomes. Limitations are discussed in detail in a dedicated section. Conclusions are appropriately cautious, noting the need for confirmation and the wide CIs. Funding sources (implied from lead contact's institution) and conflicts of interest (none declared) are reported.

Sub-criteria

funding coiADEQUATE

trial registrationN/A

reporting guidelineNOT REPORTED

methods completenessADEQUATE

all outcomes reportedADEQUATE

limitations discussedADEQUATE

conclusions proportionalADEQUATE

Per reviewer (3) — the votes behind the verdict

PASSReviewer 1· 95% conf

The manuscript is exceptionally well-reported, with detailed methods, comprehensive results including null/negative findings, transparent limitations, proportional conclusions, and full funding/COI disclosure.

Evidence

direct quote[Discussion, 'Limitations of the study']

“This study has several important limitations.”

absence[Methods]

No reporting guideline is cited, but the methods are comprehensive.

direct quote[Author Contributions]

“P.G. conceived the overall project, acquired funding, conceived the study”

PASSReviewer 2· 90% conf

Methods are detailed, limitations are discussed, conclusions are proportional, and funding/COI are reported. Trial registration is not applicable, and no formal reporting guideline is referenced, but the paper is thorough.

Evidence

paraphrase[Methods, STUDY DETAILS, METHODS DETAILS, QUANTIFICATION AND STATISTICAL ANALYSIS]

“The methods describe data sources, cohort definitions, outcome definitions (MCI, deaths due to dementia), exposure (eligibility), and the RD analysis in detail.”

direct quote[Discussion, Limitations of the study]

“This study has several important limitations. First, our findings pertain to those age groups born near to the September 2, 1933, date-of-birth threshold...”

direct quote[Data and code availability; Author contributions]

“Requests for further information and resources should be directed to, and will be fulfilled by, the Lead Contact, Pascal Geldsetzer (pgeldsetzer@stanford.edu).”

PASSReviewer 3· 90% conf

The manuscript is transparent and well-reported, with comprehensive methods, limitations discussed, proportional conclusions, and funding/COI statements. Trial registration is not applicable.

Evidence

direct quote[Discussion, 'Limitations of the study']

“This study has several important limitations. First, our findings pertain to those age groups born near to the September 2, 1933, date-of-birth threshold...”

paraphrase[Author contributions/funding section]

“P.G. conceived the overall project, acquired funding...”

absence[Entire text]

No reporting guideline is referenced.

2 copyedit issues flagged: mostly typo, clarity.

Reproducibility

4 data/code links checked; 4 live.

Registration stated in text, but no registry ID was detected. No reporting guideline cited.

Αlpha¹ Rigor Action ItemsFinal pass before journal submission

In this post-publication audit, the paper is robust in design and reporting, but the identified p-value inconsistency is a concrete error that should be corrected via an erratum or clarified in a published comment. Readers should interpret the exact p-value for the effect of vaccine receipt on MCI (reported as p=0.007) with caution; the confidence intervals are likely more reliable. The missing formal test for manipulation of the running variable is a minor methodological oversight. Otherwise, the paper provides strong internally valid evidence.

1.
HIGHstatistics
Clarify the p-value inconsistency for the effect of receiving HZ vaccination on new MCI diagnoses: reported as p=0.007 in the Results but recomputed as p=0.019 from the 95% CI. Submit an erratum with the correct value or provide a rationale for the discrepancy (e.g., different test method).
An independently detected inconsistency undermines reader trust and may affect meta-analyses; correction is essential for scientific accuracy.
2.
HIGHcopyedit
Correct the typo 'September 2 1993' to 'September 2 1933' in the METHODS (Quantification and Statistical Analysis) and Figure 1 legend.
A date error in the eligibility threshold could confuse readers and is an obvious correction needed.
3.
HIGHreporting
Add a formal test for manipulation of the running variable (McCrary density test or similar) to the Methods section and report its result.
The regression discontinuity validity assumption of no manipulation near the threshold is not explicitly verified; this is a standard diagnostic that strengthens confidence.
4.
HIGHreporting
Report exact p-values for all subgroup and sensitivity analyses, not just the main results, or clarify instances where only threshold comparisons are given.
Exact p-values facilitate full transparency and allow readers to evaluate robustness across multiple comparisons.
5.
HIGHreporting
Include a statement confirming adherence to a formal reporting guideline for observational studies (e.g., STROBE) in the Methods section.
While the paper is thorough, citing a reporting guideline is a standard transparency practice and expected by many journals.
6.
HIGHreporting
Report specific grant numbers or funding source identifiers in the funding statement.
Current funding description is vague; specific identifiers allow funder attribution and accountability.
7.
MEDIUMreporting
Explicitly describe how missing data and outliers were handled in the regression discontinuity analysis, even if only to state that the RD method with robust standard errors is robust to these issues.
Transparency about missing data handling is important for reproducibility, especially in large administrative datasets.
8.
MEDIUMdata code
Provide a direct link or summary of the SAIL data access application process in the Data Availability section.
Adding a concrete reference (e.g., URL for the SAIL application portal) would facilitate independent replication.
9.
MEDIUMreporting
Consider adding a pre-registration or analysis plan statement, even for a natural experiment, to strengthen credibility.
Pre-registration of the analytic plan (even for retrospective analyses) reduces flexibility in data analysis and increases trust.
10.
MEDIUMstatistics
In the Methods, clarify that the 'effect of actually receiving the HZ vaccination' from two-stage least squares is a complier average causal effect (CACE) and discuss the assumption of no defiers.
Readers unfamiliar with instrumental variable methods may misinterpret the estimate; explicit labeling aids understanding.

Full text (119,793 chars · analyzed ~63%, head + tail)Analyzed: Jul 5, 2026

Αlpha¹ assesses methodological rigor, not the importance of the findings. See how we calibrate →