Data Sharing & Research Transparency Policy

The purpose of data transparency is to support:

• verification of published findings;
• reproducibility;
• responsible secondary analysis;
• research integrity;
• appropriate reuse;
• methodological transparency; and
• public confidence in the scholarly record.

This policy applies, where relevant, to:

• quantitative datasets;
• qualitative data;
• clinical datasets;
• genomic and other high-dimensional data;
• imaging data;
• survey data;
• statistical code;
• software and algorithms;
• study protocols;
• analysis plans;
• questionnaires and instruments;
• data dictionaries;
• supplementary materials; and
• other research outputs required to understand or verify published findings.

The statement should accurately describe one of the applicable situations, such as:

Publicly Available Data

Where data are deposited in an open repository, authors should name the repository and provide the DOI, accession number, or permanent link.

Example:

The dataset supporting the findings of this study is available in [repository name] at [DOI or permanent URL].

Available Upon Reasonable Request

Where public deposit is not appropriate but controlled sharing is possible:

The data supporting the findings of this study are available from the corresponding author upon reasonable request, subject to applicable ethical and institutional requirements.

Restricted Data

Where data cannot be made publicly available because of privacy, ethics, legal, contractual, security, or consent restrictions:

The data are not publicly available because of confidentiality and ethical restrictions. Access may be considered subject to appropriate institutional and ethical approval.

No New Data

Where the article did not generate or analyse a new dataset:

No new data were generated or collected for this study.

Authors should not state that data are publicly available when the cited repository, file, or access route does not actually provide the described material.

Appropriate repositories may include:

• discipline-specific repositories;
• institutional repositories;
• national repositories;
• funder-designated repositories;
• controlled-access biomedical repositories; or
• reputable general-purpose repositories.

General-purpose examples may include platforms such as Zenodo, Figshare, Dryad, or the Open Science Framework where these are appropriate for the material concerned.

Repository selection should consider:

• data sensitivity;
• participant consent;
• ethical approval;
• long-term accessibility;
• versioning;
• persistent identifiers;
• access controls;
• licensing;
• disciplinary standards; and
• applicable legal requirements.

Authors must not publicly release data where doing so would unreasonably risk:

• identification of participants;
• breach of informed consent;
• breach of clinical confidentiality;
• disclosure of protected health information;
• violation of ethical approval conditions;
• exposure of sensitive community information;
• breach of legal duties; or
• other material harm.

De-identification should be appropriate to the nature of the data. Removal of direct identifiers alone may not be sufficient where combinations of variables allow realistic re-identification.

Data requiring enhanced consideration may include:

• whole-genome or whole-exome sequence data;
• genotyping data;
• pharmacogenomic information;
• biobank-linked records;
• multi-omics data;
• rare-disease datasets;
• detailed longitudinal clinical records;
• high-resolution imaging;
• wearable-device data;
• geospatial health data; or
• datasets capable of identifying individuals through linkage.

Open public deposit is not automatically appropriate for these materials merely because data sharing is scientifically desirable.

Authors should use controlled-access mechanisms where open release would conflict with consent, ethics approval, privacy obligations, or reasonable participant expectations.

Where genomic data have implications for biological relatives or identifiable communities, these broader risks should also be considered.

Where data cannot be openly shared, authors should explain:

• why access is restricted;
• whether access may nevertheless be requested;
• who controls access;
• what approvals are required;
• whether a data-use agreement is needed; and
• any legitimate eligibility conditions.

Controlled access should not be described as "data unavailable" where a legitimate access mechanism actually exists.

Conversely, authors should not promise "available upon reasonable request" if there is no realistic mechanism through which requests can be considered.

Qualitative material may contain:

• unique life histories;
• occupational information;
• geographic information;
• family relationships;
• rare clinical circumstances;
• institutional identifiers;
• socially sensitive disclosures; or
• contextual details capable of identifying participants.

Authors should therefore balance transparency against the terms of consent and the realistic risk of re-identification.

Where raw transcripts cannot be shared, authors may still improve transparency through appropriate methodological detail, coding frameworks, interview guides, analytic procedures, or other non-identifying supporting materials.

Relevant materials may include:

• R scripts;
• Python code;
• SPSS syntax;
• Stata code;
• SAS programs;
• machine-learning pipelines;
• algorithm parameters;
• data-cleaning procedures;
• model configuration files;
• simulation code; or
• other executable or computational materials underlying the findings.

Where possible, code should be versioned and deposited in a stable repository with a persistent identifier.

Authors may use public version-control platforms during development, but a stable archived version associated with the publication is preferable where long-term reproducibility is important.

Where proprietary software or algorithms are used, authors should report, as relevant:

• software name;
• version;
• manufacturer or provider;
• key parameters;
• analytic settings;
• model assumptions;
• validation procedures;
• relevant preprocessing; and
• other information needed to understand the analysis.

A commercial licence restriction may justify withholding source code but does not justify an opaque description of the methodology.

Depending on the study, relevant information may include:

• data sources;
• training, validation, and test-set separation;
• feature selection;
• preprocessing;
• model architecture;
• hyperparameters;
• performance metrics;
• calibration;
• handling of missing data;
• external validation;
• bias and fairness assessment;
• code or model availability; and
• restrictions preventing full release.

Authors should not claim that a proprietary or inaccessible model is fully reproducible where independent reproduction is not realistically possible.

Where appropriate, dataset citations should identify:

• creator or creators;
• year;
• dataset title;
• repository;
• version where applicable; and
• DOI or other persistent identifier.

Citation of a dataset in the reference list is preferable where the dataset constitutes a distinct scholarly output supporting the article.

Relevant materials may include:

• clinical-trial protocols;
• systematic-review protocols;
• prospective observational-study protocols;
• statistical analysis plans;
• registered analysis plans;
• data-management plans; and
• documented amendments.

Material deviations from a prespecified protocol or analysis plan should be disclosed and explained where they affect interpretation of the published findings.

Authors should report:

• registry name;
• registration number;
• registration date;
• relevant protocol information; and
• material deviations from the registered plan where applicable.

A registration record should not be altered retrospectively in a way that misleadingly obscures substantive changes to prespecified outcomes or analyses.

Manuscripts should report sufficient information concerning:

• study design;
• setting;
• eligibility criteria;
• sampling;
• measurement methods;
• data preprocessing;
• missing-data handling;
• statistical analyses;
• model assumptions;
• sensitivity analyses;
• software and versions;
• protocol deviations; and
• other methods necessary to evaluate the findings.

A public dataset does not compensate for inadequate methodological reporting in the article itself.

Depending on the circumstances, JPMHR may request:

• de-identified raw data;
• processed datasets;
• analysis outputs;
• statistical code;
• original figures or images;
• audit trails;
• ethics approvals;
• consent documentation;
• protocols;
• registration records; or
• other materials necessary to assess a credible concern.

Such requests do not override legitimate ethical, legal, privacy, or contractual restrictions. Authors should explain applicable restrictions and, where possible, provide an alternative verification mechanism.

Concerns may arise from:

• inconsistent datasets;
• impossible or implausible values;
• duplicated data patterns;
• unexplained image duplication;
• inconsistent participant numbers;
• unverifiable analysis outputs;
• material discrepancies between data and reported results;
• failure to provide a reasonable explanation for requested supporting material; or
• other credible indicators of data-integrity problems.

An anomaly does not by itself prove misconduct. The journal assesses context and evidence and ordinarily gives affected authors an opportunity to respond.

Depending on the findings, action may include:

• request for clarification;
• additional review;
• correction;
• Expression of Concern;
• retraction; or
• referral to an appropriate institution or authority.

Relevant records may include:

• source data;
• analysis files;
• study protocols;
• consent records;
• ethics approvals;
• laboratory records;
• coding frameworks;
• original images;
• audit trails; and
• other documentation needed to support the published work.

JPMHR does not impose one universal retention period that overrides applicable institutional, ethical, regulatory, contractual, or legal requirements.

Restrictions may arise from:

• data-use agreements;
• commercial licences;
• national registries;
• clinical databases;
• government datasets;
• biobanks;
• institutional agreements;
• confidentiality obligations; or
• other lawful access conditions.

Authors should not redistribute third-party data where they lack authority to do so.

Where the data cannot be shared by the authors, the Data Availability Statement should identify the original data source and explain, where appropriate, how eligible researchers may seek access directly.

The CC BY 4.0 licence applied to a JPMHR article does not automatically determine the legal status of every independently deposited dataset, software package, or third-party research object.

Authors should select repository licences that reflect:

• ownership;
• participant consent;
• ethical constraints;
• third-party rights;
• software dependencies;
• funder requirements; and
• the intended form of reuse.

Legitimate restrictions may include:

• participant confidentiality;
• lack of consent for public sharing;
• high re-identification risk;
• legal or regulatory prohibitions;
• court restrictions;
• national security considerations;
• commercial confidentiality;
• third-party data licences;
• indigenous or community governance requirements;
• intellectual-property restrictions; or
• other legitimate limitations.

Restrictions should be stated transparently rather than concealed or described misleadingly.

JPMHR may:

• require clarification of a Data Availability Statement;
• request a repository link or persistent identifier;
• check whether a stated repository record exists;
• request correction of inconsistent data statements;
• seek supporting data during integrity assessment;
• link published articles with deposited datasets where appropriate;
• encourage code and protocol sharing;
• protect legitimate restrictions on sensitive data; and
• correct the scholarly record where data-access claims prove materially inaccurate.

JPMHR does not guarantee that an independently operated repository will remain permanently available and does not control third-party repository governance.

• ICMJE Recommendations
• EQUATOR Network
• COPE Guidance and Resources
• Principles of Transparency and Best Practice in Scholarly Publishing

Reference to external guidance does not imply membership, accreditation, certification, or endorsement unless such status is independently established.