This manuscript reports the development and content validation of a Focus Group Discussion (FGD) guide intended to explore the self-management needs, preferences and barriers of people with hypertension, with the ultimate aim of informing the development of a contextually appropriate mHealth application. The authors describe a development process involving literature review, initial question generation, face-validity assessment by five experts, content validation by ten experts, translation and back-translation, pilot testing with five individuals with hypertension, and subsequent refinement of a 19-item FGD guide. The content-validity assessment uses the Item-level Content Validity Index (I-CVI), Scale-level Content Validity Index (S-CVI), Content Validity Ratio (CVR), and modified kappa statistic.
The topic is relevant and potentially valuable. Developing an FGD guide systematically, rather than using an untested set of questions, is methodologically worthwhile, particularly when the resulting qualitative data are intended to inform an mHealth intervention. The availability of the underlying expert-rating data and supplementary FGD materials is also a strength.
However, I identified several issues that should be addressed before the manuscript can be considered scientifically sound. Most importantly, there are internal inconsistencies in the reported content-validity statistics, uncertainty regarding the conceptual distinction between I-CVI and CVR in the present dataset, insufficient detail to reproduce the literature-informed guide-development process, discrepancies in the reporting of the pilot FGD, and overinterpretation of modified kappa as evidence of reliability. The supplementary source data also indicate that several FGD questions introduce predetermined app features, which should be acknowledged and managed as a potential source of framing or leading-question bias.

1. Is the work clearly and accurately presented and does it cite the current literature?
Response: Partly
The study aim is clear and the manuscript generally follows a recognisable methodological structure. The Introduction provides an appropriate rationale for hypertension self-management, mHealth and the importance of systematically developing qualitative data-collection tools.
Nevertheless, the presentation requires substantial improvement in accuracy and consistency. Several statistical values differ between the tables, Results, Discussion and Conclusion. There are also inconsistencies in methodological terminology, including the number of development steps, use of the terms “questionnaire” and “FGD guide”, and interpretation of validity and reliability.
The literature base also requires updating. Although the manuscript includes relevant foundational literature on content validity and mHealth, a number of clinical and digital-health references are relatively old for an article published in 2026. For example, several core mHealth references cited in the manuscript date from 2013 to 2018. Contemporary hypertension guidelines are now available, including the 2024 ESC Guidelines for the Management of Elevated Blood Pressure and Hypertension and the 2025 AHA/ACC multisociety guideline. Recent systematic-review evidence on digital and mHealth interventions for hypertension is also available.
Recommended revisions
The authors should:

1. undertake careful language and technical editing throughout;
2. ensure that all numerical values are consistent across the Abstract, tables, Results, Discussion and Conclusion;
3. use the terms “FGD guide”, “item”, “question”, “I-CVI”, “S-CVI”, “CVR” and “modified kappa” consistently;
4. update the Introduction and Discussion with contemporary hypertension and mHealth literature;
5. clarify the specific methodological contribution of the present guide relative to existing qualitative approaches to mHealth development.

This point requires revision, although it is not independently fatal to the study.

Major comments that must be addressed
2. Clarify the conceptual distinction between I-CVI and CVR
This is an essential revision.
The manuscript states that experts rated each item using a three-point scale:
1 = Essential
2 = Useful but not essential
3 = Not necessary
This is consistent with the Lawshe approach for calculating the Content Validity Ratio, which evaluates item essentiality. However, the manuscript subsequently states that I-CVI was calculated as the proportion of experts rating each item as “essential”.
The underlying spreadsheet provided with the article confirms that the same expert-agreement pattern appears to have been used for I-CVI, CVR and modified kappa. For example, where nine of ten experts agreed:

• I-CVI = 9/10 = 0.90
• CVR = (9 − 5)/5 = 0.80
• modified kappa ≈ 0.899

Therefore, in this dataset:
CVR = 2(I-CVI) − 1
Consequently, I-CVI and CVR are mathematical transformations of the same underlying agreement judgement rather than independent assessments of different dimensions of content validity.
The authors must clarify whether experts actually completed separate relevance and essentiality ratings. If they did, the two rating procedures and the corresponding raw data should be reported separately. If they did not, the manuscript should clearly acknowledge that the indices derive from the same essentiality judgement and should not portray them as independent confirmations of validity.
The statement that calculating CVR, I-CVI, S-CVI and kappa provides a more comprehensive assessment should therefore be revised unless the authors can demonstrate that these statistics were derived from genuinely distinct expert judgements.

3. Recalculate and correct the content-validity statistics
This is an essential revision for scientific soundness.
The numerical reporting is internally inconsistent.
Table 1 shows:

• 13 items with I-CVI = 1.00
• 6 items with I-CVI = 0.90

Therefore:
Sum of I-CVI = (13 × 1.00) + (6 × 0.90) = 18.40
and:
S-CVI/Ave = 18.40 / 19 = 0.9684, approximately 0.97
The underlying source spreadsheet independently confirms an S-CVI/Ave of 0.968421.
However:

• Table 1 reports S-CVI/Ave = 0.96;
• the Results report S-CVI/Ave = 0.93;
• the Results state that the sum of I-CVIs was 17.8.

The value 17.8 corresponds instead to the sum of the CVR values.
Similarly, Table 2 shows:

• 13 items with CVR = 1.00
• 6 items with CVR = 0.80.

Therefore:
Mean CVR = [(13 × 1.00) + (6 × 0.80)] / 19 = 17.80 / 19 = 0.9368, approximately 0.94.
The source spreadsheet confirms approximately 0.937.
However, the manuscript reports a CVR range of 0.86 to 1.00 and an average of 0.9. These values are inconsistent with both Table 2 and the underlying dataset.
The S-CVI/UA calculation does appear internally reproducible:
13 / 19 = 0.6842, approximately 0.68.
Required action
The authors should recalculate the entire analysis directly from the deposited expert-level dataset and amend all relevant locations, including:

• Abstract;
• Methods;
• Table 1;
• Table 2;
• Table 3, if necessary;
• Results;
• Discussion;
• Limitations;
• Conclusion; and
• supplementary files.

A consistent rounding convention should also be specified.

4. Correct the interpretation of the CVR threshold
The Discussion states that expert ratings were evaluated against a CVR threshold of 0.75.
However, the underlying worksheet itself states that the critical CVR for a panel of ten experts is 0.80. The manuscript should therefore use a consistent critical threshold and provide the appropriate source and decision rule.
This is important because six items have a CVR of exactly 0.80. The Results currently refer to values below 0.9 as “non-essential”. This interpretation is inappropriate if 0.80 is the prespecified critical value.
The authors should report:

• number of experts;
• critical CVR used;
• rationale/reference for this threshold;
• minimum number of experts required to classify an item as essential;
• whether all 19 items met the prespecified criterion.

5. Claims regarding “reliability” should be revised
This is an essential interpretative revision.
The Abstract concludes that the FGD guide demonstrates “acceptable reliability”. However, no test-retest assessment, reproducibility study, inter-moderator assessment or other direct evaluation of reliability of the FGD guide is reported.
The modified kappa statistic used here corrects expert agreement on content-validity judgements for chance. It should not automatically be interpreted as establishing reliability of the FGD guide when subsequently used to collect qualitative data.
This issue becomes particularly apparent because the final Conclusion appropriately states that “further evaluation of reliability and additional validity measures is recommended”. This conflicts with the stronger reliability claim in the Abstract.
I recommend replacing statements about established reliability with wording such as:
“The guide demonstrated high item-level expert agreement and evidence of content validity within the study context.”
The instrument should preferably be described as a systematically developed and content-validated FGD guide, rather than as a comprehensively validated and reliable instrument.

6. The literature-search and item-generation process is insufficiently reproducible
This is an essential methodological revision.
The literature-search stage is currently described as a review conducted using “standard search engines”. This is insufficient to allow another researcher to reproduce the development process or determine how the 19 questions were generated.
The authors should provide, at minimum:

• databases and/or search platforms searched;
• date on which the search was conducted;
• search period;
• search terms and Boolean strategy;
• inclusion and exclusion criteria;
• screening process;
• number and types of publications informing development;
• method used to identify self-management domains;
• process by which evidence was converted into individual FGD questions;
• individuals involved in drafting and refining the initial questions.

A supplementary content-development matrix would substantially improve transparency, for example:
FGD item | Underlying domain | Evidence/source | Expert feedback | Modification | Final item
This is particularly important because the Conclusion identifies medication management, blood pressure monitoring, DASH diet, physical activity, weight management, stress reduction, smoking/alcohol and health education as key self-care domains. The Methods should demonstrate how these domains were generated.

7. Resolve inconsistencies in the pilot FGD
This must be addressed.
There is a direct discrepancy in the reported duration of the pilot session.
The Methods state: approximately 60 minutes
The Results state:
approximately 1.5 to 2 hours
The authors should confirm the actual duration and correct both sections.
The pilot study also requires considerably more detail, including:

• eligibility criteria;
• recruitment procedure;
• participant characteristics;
• age and sex distribution;
• hypertension characteristics where relevant;
• smartphone use and prior mHealth experience;
• language used;
• location;
• moderator identity, qualifications and training;
• relationship between the moderator and participants;
• role of the note-taker;
• transcription procedure;
• name/version of qualitative-analysis software;
• analytical approach;
• who undertook the analysis;
• how disagreements were handled;
• specific findings that resulted in changes to the guide.

COREQ provides a recognised 32-item reporting framework specifically for interviews and focus groups and would be useful for improving reporting of these elements.

8. Clarify participant consent and ethical procedures for the pilot FGD
This should be addressed before approval.
The ethics section states that written informed consent was obtained from “all expert”. However, five people with hypertension also participated in an audio-recorded pilot FGD.
The supplied Participant Information Sheet and consent form establish general study consent but do not appear to contain a separate explicit statement concerning consent for audio recording.
I am not assuming that recording consent was absent, because it may have been obtained through an additional approved procedure that is not included in the supplied material. The authors should nevertheless clarify:

1. whether the five pilot participants provided written informed consent;
2. whether explicit consent for audio recording was obtained;
3. how confidentiality of audio recordings and transcripts was protected;
4. whether the pilot FGD was covered by IEC approval 587/2021;
5. whether pilot data were included in or excluded from subsequent main-study qualitative analysis.

The grammar in the current ethics statement should also be corrected from “all expert” to the appropriate description of all relevant participants.

9. Clarify the relationship between this methodological study and the parent trial
The manuscript states only near the Conclusion that this methodological study formed part of a larger randomised controlled trial.
The supplied Participant Information Sheet clarifies that the parent project is a community-based cluster randomised trial evaluating an mHealth application for remote monitoring and hypertension self-management.
This context should be introduced in the Study Design section. The authors should explain clearly:

• that development of the FGD guide preceded development of the mHealth intervention;
• where the present methodological study sits within the parent project;
• which elements were conducted before randomisation/intervention delivery;
• why the parent CTRI registration is cited for this methodological work.

This would prevent readers from incorrectly interpreting the present study itself as an RCT.

10. Correct the number and sequence of development stages
The Study Design section states that the process was conducted in four sequential steps. The subsequent Methods describe six numbered steps, including pilot testing and final refinement.
This should be corrected.
A clearer structure would be:

1. literature review and domain identification;
2. development of the initial FGD guide;
3. face-validity assessment;
4. expert content validation;
5. translation and cultural/linguistic adaptation;
6. pilot testing and final refinement.

A flow diagram illustrating this development pathway would improve readability.

11. Provide greater transparency regarding expert selection and panel composition
The manuscript reports that the content-validation panel included physicians from General Medicine and Cardiology together with Health Information Management professionals.
The deposited demographic data provide useful additional information:

• three cardiologists;
• four General Medicine clinicians;
• three Health Information Management professionals;
• five male and five female experts;
• mean professional experience approximately 19 years.

These characteristics should be reported in the main manuscript.
The professional experience of the panel is a strength. However, the panel is relatively medically focused for a guide concerned with broad self-management and mHealth design. The proposed mHealth domains include diet, physical activity, medication adherence, stress, digital interaction and health behaviour. The authors should therefore discuss the limited or absent representation of disciplines such as behavioural science/psychology, allied health, pharmacy, dietetics, nursing, usability or human-computer interaction, implementation science and patient/public contributors.
I would not necessarily recommend increasing the panel solely to obtain a larger number. Rather, the relevant limitation is the breadth and diversity of expertise represented.

12. Several FGD questions may be leading because predetermined app features are introduced
This is an important methodological issue identified from the supplementary guide.
The stated purpose is to identify patients’ needs and expectations so that these can inform development of a customised mHealth application. However, several follow-up questions present participants with predetermined solutions and ask whether they would like these included.
Examples include questions concerning:

• a DASH diet feature;
• daily salt-intake guidance;
• a workout plan;
• step and sleep tracking;
• a medication pill organiser;
• medication alerts;
• weekly blood-pressure tracking;
• reminders and notifications;
• visual education through images/charts/videos;
• e-consultation;
• stress-management features;
• recording pre-existing health conditions.

These may be useful questions if the objective is to assess acceptability of candidate features, but they are more problematic if presented as purely exploratory questions intended to identify user requirements inductively.
The authors should therefore distinguish between:

1. open-ended needs-assessment questions; and
2. questions assessing responses to candidate app features derived from previous evidence.

One possible approach would be to ask broad, neutral questions first and introduce specific candidate features only as secondary prompts after spontaneous needs have been elicited. This would reduce framing bias and allow genuinely unanticipated user priorities to emerge.
The Methods should also explain where each predetermined candidate feature originated.

13. Review multi-component and double-barrelled questions
Some items combine several distinct concepts.
For example, one question combines a proposed workout plan with tracking steps and sleep. These represent separate behavioural and digital functions. Another combines education concerning risk factors, consequences and treatment with several delivery formats, including charts, images, graphs and videos.
If a participant agrees or disagrees, it may therefore be unclear which component generated the response.
The authors should review the guide for:

• double-barrelled questions;
• overly complex wording;
• assumptions embedded in questions;
• potentially leading terminology;
• unnecessary clinical jargon.

Where appropriate, broader questions should be followed by discrete neutral probes.

14. Strengthen reporting of translation and back-translation
The manuscript states that the guide was translated into Kannada and back-translated into English. This is appropriate, but insufficiently detailed.
The authors should report:

• who completed the forward translation;
• their professional and linguistic qualifications;
• who completed the back-translation;
• whether translators worked independently;
• whether the back-translator was blinded to the original English version;
• how discrepancies were identified and resolved;
• whether cultural adaptation was undertaken in addition to literal translation;
• which language/version was used during pilot testing.

Given that the study emphasises contextual and cultural relevance, this process deserves fuller reporting.

15. Target-population characteristics should be presented
The supplied study documentation contains relatively specific eligibility criteria, including diagnosed primary hypertension, Android smartphone use, age 18 to 60 years and ability to understand English or Kannada.
The manuscript merely states that five individuals with hypertension were purposively selected for pilot testing.
The authors should specify whether the pilot participants met the stated parent-study eligibility criteria and provide their relevant characteristics.
This is especially important because digital access and experience are integral to an mHealth needs-assessment study. Smartphone familiarity, previous app use and digital literacy may materially influence participants’ understanding of proposed features and their expressed preferences.

16. Data availability and reproducibility require clarification
Response to source-data question: Partly
The data-availability statement is a strength. The manuscript states that anonymised expert ratings, I-CVI, S-CVI, modified kappa statistics, expert demographic information and the FGD guide are available through Figshare.
The supplied worksheet is sufficient to reproduce the major quantitative content-validity calculations. Indeed, recalculation from this source data reveals the numerical discrepancies noted above.
However, full reproduction of the developmental process is not currently possible from the supplied materials. In particular, I could not identify:

• the complete original pre-validation version of the guide with item-level change history;
• anonymised qualitative comments from individual experts;
• a structured record of decisions following face/content validation;
• pilot FGD transcript or appropriately de-identified qualitative source data;
• pilot coding/analysis;
• an item-level record showing how pilot feedback changed the final guide.

The authors should either provide these materials where ethically and practically possible or explain why particular qualitative data cannot be shared.

17. Interpretation of the S-CVI/UA should be more cautious
The S-CVI/UA of approximately 0.68 is correctly reproducible from 13 of 19 items achieving unanimous agreement.
The manuscript describes this as moderate/accepted content validity. The authors should provide a methodological source supporting the chosen interpretation threshold rather than simply labelling the value “accepted”.
The authors may also explain that universal agreement becomes increasingly stringent as the number of experts increases. However, that methodological consideration should be discussed explicitly rather than using an unexplained categorical label.

18. The Discussion should be made more analytical and less repetitive
The Discussion currently repeats much of the methodology and numerical results.
I suggest restructuring it around:

1. principal methodological finding;
2. implications of high item-level agreement;
3. interpretation of the difference between S-CVI/Ave and S-CVI/UA;
4. contribution of patient pilot testing;
5. implications for user-centred mHealth development;
6. cultural and linguistic context;
7. digital literacy and accessibility;
8. transferability outside the study population;
9. limitations of expert-based content validation;
10. next stages of evaluation.

The statement that the questions were “sufficient for obtaining the insights” should also be moderated. A five-participant pilot can demonstrate feasibility and identify obvious problems with wording or flow, but it does not in itself establish that the guide will capture the complete range of relevant experiences in the main study.

19. Conclusions should be moderated
Response: Partly supported
The study does support the conclusion that the authors developed an FGD guide using structured expert assessment and pilot testing and that the retained items showed high levels of expert agreement.
However, the current language goes beyond what has been demonstrated. The Abstract describes the instrument as a validated tool with “acceptable reliability”.
A more defensible conclusion would be that:
The study developed and established preliminary content-validity evidence for a context-specific FGD guide intended to explore the hypertension self-management and mHealth requirements of the target population.
The authors should avoid implying that:

• reliability has been demonstrated;
• all forms of validity have been established;
• the guide is immediately generalisable to other populations, languages or health systems;
• high expert agreement guarantees comprehensiveness of subsequent qualitative data.

Further evaluation in more diverse populations and settings would be appropriate.

Minor comments

1. The abbreviation list contains “CVI” twice. The second entry appears to be intended as I-CVI.
2. Tables 1 to 3 refer to the instrument as a “questionnaire”. It should consistently be described as an FGD guide.
3. The phrase “hypertension individuals” should be replaced with “individuals with hypertension” or “people with hypertension”.
4. Patient-centred or person-centred terminology should be used consistently.
5. The Methods require grammatical revision, particularly the description of probe, follow-up and exit questions.
6. The phrase “how well things match or reflect a specific domain” in the statistical-analysis section is insufficiently scientific and should be replaced with precise methodological terminology.
7. The modified kappa formula should be typeset clearly and checked for missing mathematical notation.
8. Tables should define what an “x” denotes and whether blank cells represent ratings other than essential or missing observations.
9. The manuscript should clarify whether the five face-validity experts were independent of, or overlapped with, the ten content-validity experts.
10. The manuscript should state whether expert ratings were completed independently before any consensus discussion.
11. The authors should clarify how many questions were present before expert review and whether any items were actually removed. The text states that redundant or unclear items were revised or removed, but the precise changes are not shown.
12. “Generalisability” in the Limitations section could be complemented by discussion of transferability, particularly because the guide is contextually and linguistically adapted.
13. The opening sentence of the Abstract concerning perceptions of service quality does not connect clearly with the subsequent rationale for developing an mHealth FGD guide and should be reconsidered.
14. Contemporary hypertension guidance should be incorporated into the clinical rationale. The 2024 ESC guideline and 2025 AHA/ACC guideline are relevant current sources. These should be considered alongside contemporary mHealth evidence rather than replacing the methodological references already used.

Essential revisions required for scientific soundness
The following issues must be addressed before I would recommend an Approved status:

1. Recalculate all I-CVI, S-CVI/Ave, S-CVI/UA and CVR results directly from the underlying expert-level dataset and correct all inconsistencies.
2. Clarify whether I-CVI and CVR were calculated from separate relevance and essentiality assessments. If they were derived from the same essentiality ratings, this must be explicitly acknowledged and interpreted appropriately.
3. Correct and justify the CVR critical value and item-retention criterion.
4. Remove or substantially modify claims that reliability of the FGD guide has been established.
5. Provide a reproducible description of the literature-search, domain-identification and item-generation process.
6. Correct the discrepancy in pilot FGD duration and provide sufficient detail regarding pilot recruitment, participants, moderation, transcription, analysis and resulting changes to the guide.
7. Clarify ethical consent for the patient pilot participants, including consent for audio recording.
8. Clarify the relationship between the present methodological study and the parent cluster randomised trial.
9. Correct the inconsistent description of four versus six development stages.
10. Address the potentially leading and solution-oriented nature of the feature-specific questions in the FGD guide and explain how predetermined mHealth features were generated.
11. Review double-barrelled/multi-component questions and improve the neutrality and interpretability of individual questions.
12. Provide greater methodological detail concerning translation and back-translation.
13. Report expert-panel and pilot-participant characteristics adequately in the manuscript.
14. Moderate the conclusions so that they reflect evidence of content validity within the present study context rather than comprehensive validity and reliability.