Data Management and Access Plan

Data Management And Access Plan

The NIHR accepts no responsibility for the content of the form, the award holder/author is responsible for the content of the form. The NIHR checks the form for completion; however, any queries about the form should be directed to the award holder/author.

The NIHR promotes openness and transparency in research through the NIHR position on the sharing of research data, which applies to all studies where all the research costs are funded by the NIHR. For the purposes of sharing research data, the NIHR recommends and supports the use of Good Practice Principles for Sharing Individual Participant Data from Publicly Funded Clinical Trials, COMET Initiative Core outcomes set, Guidelines for the Content of Statistical Analysis Plans in Clinical Trials, SPIRIT 2013 checklist for clinical trial protocols and related documents and Science Europe Practical Guide to the International alignment of Research Data Management. Where relevant, these are to be considered in the preparation of this form.

The NIHR is mindful of the balance between researcher effort alongside avoiding waste in research and the benefits of supporting wider availability and reuse of data. Therefore when completing this form, please highlight where any sections do not apply to your study, with a brief explanation where possible.

Please note as the Chief Investigator of this NIHR project you are responsible for the submission of this Data Access Management Plan.

[{"setting":"Private mental healthcare in Malaysia","dataOutcomes":[{"area":"Comet - Life impact","domain":["Social functioning","Role functioning","Emotional functioning/wellbeing"],"types":"Qualitative","otherDetails":""},{"area":"Other","domain":[],"types":"Qualitative","otherDetails":"Discourse data collected from the public domain (policies, newsmedia, social media posts) related to SOGIESC minorities and mental health"},{"area":"Other","domain":[],"types":"Qualitative","otherDetails":"Focus-group recordings and transcripts"},{"area":"Other","domain":[],"types":"Survey data","otherDetails":"Surveys with SOGIESC minorities on mental health services in Malaysia"},{"area":"Other","domain":[],"types":"Qualitative","otherDetails":"Recorded consultations between SOGIESC minority clients and mental health professionals"}],"collection":"1. Explain which methodologies or software will be used if new data are collected or produced.\nNew data will be collected for all three work packages: WP1.1, WP1.2, and WP1.3.\n\nWP1.1 (Corpus-based discourse analysis):\nWe will use corpus-assisted critical discourse analysis to collect and analyze textual data from public archives, including news articles, policies, guidelines, and social media content (with ethical considerations applied). Textual data will be extracted manually and entered into a structured data extraction sheet, with all identifying information from social media anonymised. After extraction, the data will be cleaned for consistency and anonymisation and imported into the corpus analytic software LancsBox. Collection and preparation of corpus data for WP1.1 are now complete.\n\nWP1.2 (Stakeholder perspectives):\nTwo data collection methods are employed: Focus Group Discussions (FGDs) and online survey questionnaire. FGDs will be recorded (audio at minimum), transcribed verbatim, and anonymised. These transcripts will be imported into MAXQDA for qualitative analysis. This focus group data is ready for analysis. On the other hand. surveys will be disseminated online and through trusted networks. No identifiable information will be required. Open-ended responses will be imported into MAXQDA, and close-ended responses will be processed using Microsoft Excel and tentatively a statistical analysis software (e.g., SPSS or JASP). Survey data is also ready for analysis.\n\nWP1.3 (Therapeutic session data):\nReal-life mental healthcare sessions will be recorded in audio and video formats. Verbatim transcripts will be created and further transcribed at the micro-analytic level using DOTE, a transcription software suited for conversation analysis. All identifying information will be anonymised. The micro-level transcription is ongoing and not yet ready for analysis.\n\n2. State any constraints on the re-use of existing data if there are any.\nRe-use of existing data is not feasible due to the highly specific nature of the data we are collecting—particularly data involving sexual and gender minorities within Malaysian mental healthcare settings. Additionally, existing datasets (if any) often come with significant ethical and privacy constraints that limit accessibility and data-sharing, especially where participant vulnerability and context-specific stigma are involved.\n\n3. Explain how data provenance will be documented.\nData provenance will be documented systematically across all work packages using:\n- A data extraction sheet for WP1.1, detailing source URLs (when applicable), dates of collection, data types (e.g., social media post, news article), and cleaning or anonymisation steps taken.\n- Metadata logs for WP1.2 and WP1.3 that include the date of data collection, participant group, collection method (e.g., FGD, survey, recording), transcription status, anonymisation status, and storage location.\n- Version-controlled documents will be maintained to track any updates to raw data, transcripts, and coding across all qualitative software tools (MAXQDA, DOTE).\n- A secure, shared folder structure with timestamps and researcher initials will support transparency and traceability throughout the research process.\n\n4. Briefly state the reasons if the re-use of any existing data sources has been considered but discarded.\n\nRe-use of existing data sources was considered but ultimately discarded due to:\n- Lack of existing datasets specific to the intersection of SOGIESC (sexual orientation, gender identity, and expression, and sex characteristics) and mental healthcare in Malaysia.\n- Sensitivity of the topic, which often results in data being either unpublished, inaccessible, or held under strict ethical restrictions.\n- Ethical concerns regarding participant confidentiality and safety in contexts where LGBT+ identities are criminalized or heavily stigmatized.\n\nHence, collecting new, ethically sound data is necessary and more productive for achieving the project’s aims.","management":"This project will generate several types of data across WP1.1, WP1.2, and WP1.3, including textual corpus data, qualitative data such as interview transcripts and open-ended survey responses, and quantitative data from structured surveys. We will use appropriate qualitative and quantitative software tools such as LancsBox, MAXQDA, Excel, and DOTE to store and analyse its data for their functionality, compatibility, and long-term accessibility. Data sharing decisions are guided by the sensitivity of the data and the ethical imperative to protect participants, especially concerning the safety and dignity of LGBT+ participants. While anonymised data from WP1.1 (excluding social media), WP1.2 focus groups, and surveys will be openly shared, the social media corpus and real-life therapy session data will remain restricted due to their potential for re-identification and the need to preserve participant confidentiality and safety.\n\nFor WP1.1, we will collect textual discursive data from public archives, such as news articles, policy documents, official guidelines, and public social media content that relate to LGBT+ issues and mental health. These data will be extracted and entered into an extraction sheet, cleaned for consistency, and anonymised where necessary, especially in the case of social media posts. The cleaned corpus will then be imported into corpus linguistic software, tentatively LancsBox, which is suitable for linguistic pattern analysis and visualisation. The estimated volume of this corpus will range from 300,000 to 1,000,000 words. Textual data from official publications and publicly available policy materials will be openly shared. However, social media data, despite being publicly accessible, will not be openly shared due to ethical concerns. These concerns include the risk of traceability and potential re-identification, especially in a socio-political context where LGBT+ individuals are vulnerable. Only aggregate and anonymised findings drawn from the social media data will be disseminated publicly.\n\nFor WP1.2, we will collect both qualitative and quantitative data. Two focus group discussions will be conducted and recorded, one with mental health professionals (n=35) and/or representatives of professional bodies and NGOs (n=5), and another with LGBT+ persons who have experience as mental healthcare clients (n=30). These audio recordings will be transcribed verbatim, anonymised, and imported into MAXQDA for qualitative coding and analysis. Alongside this, an online survey with LGBT+ persons in Malaysia (n=100) will be conducted, comprising both open-ended and closed-ended questions. The open-ended responses will also be analysed using MAXQDA, while the closed-ended responses will be entered into Excel and potentially a statistical analysis software for descriptive and inferential analysis. The survey is expected to produce about 100 completed responses. Anonymised transcripts and survey datasets from WP1.2 will be publicly shared through a secure data repository, pending informed consent and appropriate redaction of any identifying details.\n\nFor WP1.3, the project will collect audiovisual data from real-life mental healthcare sessions between mental health professionals (n=3) and LGBT+ clients (n=10) in Malaysia. These sessions will be recorded with participants’ consent during naturally occurring consultations, with no researcher present in the room to preserve the therapeutic relationship. A total of 160 sessions, equating to approximately 80 to 160 hours of recordings, will be collected. The recordings will first be transcribed verbatim, and then transcribed again at a micro-analytic level using the DOTE transcription software for conversation analysis. This software is selected because it allows for highly detailed transcription of talk, including pauses, intonation, and overlap. The resulting transcripts will contain rich detail necessary for interactional and discourse analysis. Due to the highly sensitive and personal nature of this data, these recordings and full transcripts will not be publicly shared. Even when anonymised, the content could pose a risk of indirect identification or emotional harm if misused. Therefore, only carefully selected transcript excerpts, with all identifying details removed, will be included in academic publications or shared for educational purposes under strict ethical control."}]

Section 1 - Research Details

Contracting Organisation

NIHR Funding Programme

NIHR Study ID

Research Title

Study Registration

Registration ID

If your study is not registered, or the registry is not listed above, please provide details or a brief explanation below.

Research Type

Research Methodology

Other Study Methodology/Trial - Please provide more details

Section 2 - Contact Information

Chief Investigator

ORCID ID

Additional Organisation Contacts

Relevant institutional, departmental or study policies on data sharing and security

Section 3 - Data Collection

What do we define as research data?

The NIHR supports the use of the COMET health outcome classification. Please see COMET's 38- category outcome scale. For further information on details and the development of the taxonomy please click here to read the paper. However, we recognise that your research outcomes/data may not fall into these health outcomes, and where they do not apply please add the other outcomes as free text as appropriate.

You can add multiple types of data to this management plan. Click the Add Another Type of Data button below to add another set of data.

Data Collection

Study Settings

Type of Data

Data Area

Data Type

Other

Section 4 - Data Description and Collection, or Re-Use of Existing Data

Q4a. How will new data be collected or produced and/or how will existing data be re-used?

Q4b. What data (e.g. the kinds, formats and volumes) will be collected or produced?

Section 5 - Data Quality Control

Q5a. What data quality control measures will be used?

To ensure data quality and consistency across all work packages, several control measures will be implemented throughout the data lifecycle. For WP1.1, all extracted corpus data—whether from policy documents, media articles, or social media posts—will be cross-verified by at least two researchers to ensure accuracy, relevance, and consistent formatting. A standardised data extraction template will be used, and all entries will be cleaned and checked for duplication, formatting inconsistencies, and anonymisation errors. In WP1.2, audio recordings from focus group discussions will be transcribed verbatim by trained research assistants and subsequently reviewed by a second team member for accuracy. The transcription will follow a consistent template to ensure reliability in subsequent coding. Survey data will undergo real-time logic checks during online submission (e.g., mandatory fields, branching logic), and once exported, will be screened for incomplete or inconsistent responses before analysis. Data entry into MAXQDA and Excel will follow a clear coding framework to minimise subjective interpretation, and intercoder reliability will be tested where applicable. For WP1.3, session recordings will be transcribed first at the verbatim level and then at a micro-analytic level by researchers trained in conversation analysis, using the DOTE software. These transcripts will be reviewed line-by-line by at least one other team member to ensure fidelity to the original recording, especially for fine-grained features like overlaps, pauses, and intonation. Additionally, all anonymisation processes will follow a clear protocol, and all sensitive data will be stored in encrypted, access-controlled environments to minimise any accidental disclosure.

Section 6 - Storage and Backup During the Research Process

Q6a. How will data be stored and backed up during the research process?

Throughout the research process, all data will be stored securely in two separate locations to ensure redundancy and protect against data loss. First, offline storage will be maintained on encrypted, password-protected hard drives accessible only to authorised project members. Second, an encrypted, end-to-end secure cloud storage system—tentatively ProtonDrive—will be used for online storage. This dual-location system ensures that data remain accessible in the event of hardware failure or unforeseen disruptions. Our data storage plans also take into consideration cybersecurity concerns and potential risks to the safety and privacy of both participants and researchers, particularly given the sensitivity of our study population and context. All personal or identifying information in the data will be anonymised through removal or masking, including the use of pseudonyms, facial blurring, and voice alteration where applicable. Access to any raw or processed data will be strictly limited to authorised project members. Data will be backed up to the offline storage once per month to maintain data integrity and ensure recovery capability. Transcripts, survey data, and analytic files will also be stored in encrypted formats and managed in accordance with strict confidentiality protocols.

Q6b. How will data security and protection of sensitive data be taken care of during the research?

This research involves the collection and analysis of highly sensitive data, including audio recordings from interviews and consultation sessions, open-ended survey responses, and public social media content that may express identity, health disclosures, or political views. Given the vulnerability of the participant population—particularly those from marginalised and LGBT+ communities—the project places a high priority on safeguarding participant privacy and ensuring strict data security throughout all stages of the research.

To mitigate risks of privacy breaches and unintended disclosure, all data collected will undergo anonymisation. This includes the removal or modification of personally identifiable information during transcription, such as names, locations, and affiliations. Where necessary, additional steps such as facial blurring in video recordings, voice alteration in audio recordings, and the use of pseudonyms will be applied to preserve confidentiality. Data will be stored in encrypted, password-protected locations with access restricted to essential, authorised research team members only. Devices used in the project will also be secured with password protection and encryption, and automatic backups will be performed on encrypted cloud platforms.

A data breach or loss incident could significantly impact both participants and the research team. In the event of data loss, the project will follow a clear incident response plan, which tentatively includes notifying affected participants (if re-identifiable), documenting the breach, and reporting it to the appropriate institutional authorities. To prevent such incidents, researchers will only use encrypted communication channels such as ProtonDrive  for data transfer.

When working with social media data, additional ethical concerns arise. Although the data is publicly accessible, there is still a risk of harm if content is traceable back to its original poster. This is especially critical in the Malaysian context, where LGBT+ identities are criminalised and heavily stigmatised. To address this, all metadata—such as usernames, timestamps, and hashtags—will be removed, and identifying phrases will either be paraphrased or omitted entirely to prevent search-based re-identification. The analysis will focus on discursive patterns rather than profiling individual users. While informed consent is not feasible for publicly available posts, the project will respect contextual privacy norms and ensure that no private or locked-group content is included in the analysis.

Protecting participant data is not only an ethical obligation but also essential for preserving trust among participants and the broader community. A failure to secure sensitive data could lead to reputational harm for both researchers and the institution, potential legal consequences, and long-term damage to future research engagement. To mitigate this, all team members will be trained in digital security practices and ethical data handling procedures aligned with institutional and international standards, including GDPR principles, as part of the Risk Training budgeted in the project.

Data will be retained for a minimum of seven years following publication, in compliance with institutional policies. This retention allows for potential audits or re-analysis. After the retention period, all paper-based data will be shredded, and digital files will be permanently deleted using secure erasure tools.

Section 7 - Data Sharing and Long-term Preservation

Q7a. How and when will data be shared? Are there possible restrictions to data sharing or embargo reasons?

Q7b. How will data for preservation be selected, and where will data be preserved long-term, for example a data repository or archive?

Q7c. What methods or software tools will be needed to access and use the data?

Section 8 - Data Management Responsibilities and Resources

Q8a. Who (for example role, position and institution) will be responsible for data management i.e. the data steward?

At a project level, the Principal Investigator (PI), based at the Faculty of Languages and Linguistics, Universiti Malaya, will serve as the data steward for this project. The PI will be supported by the project’s Research Assistants (RAs) and the Junior Project Manager (cum Research Coordinator) across the three work packages. Together, they will be responsible for ensuring secure data collection, proper storage, documentation, and eventual archiving or deletion of data in accordance with ethical and institutional guidelines. The PI will oversee compliance with data protection policies and ensure that all team members are trained on safe data handling practices. 
On an institutional level, data management responsibilities are managed according to the Universiti Malaya Research Data Management Policy https://openscience.um.edu.my/dmp-online-tool.  While there are Data Stewards identified under the UM Open Science platform and policy, for the abovementioned reasons, we do not intend to share our data on the Open Science platform.

Q8b. What resources (for example financial and time) will be dedicated to data management and ensuring that data will be FAIR (Findable, Accessible, Interoperable and Re-usable)?

Section 9 - Data Management and Access Plan Authorship

Data Access Plan Author

Please provide author details.

Telephone

Please provide telephone details.

Please provide a valid email address.

Current Version

Date and Time of Current Version

Details of Changes

Approved