Effectiveness of a Dyadic Pain Management Program for Community-Dwelling Older Adults with Chronic Pain: A Cluster Randomized Controlled Trial

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Dear authors,

The manuscript addresses a clinically relevant and high-impact issue in community and geriatric nursing. However, before the manuscript can be considered for publication, key aspects related to the scientific framing (research question and hypotheses), methodological coherence, statistical analysis, and the interpretative moderation of the results need to be strengthened. These issues primarily affect the study's internal validity rather than its relevance.

Abstract

Comment 1.
The abstract does not clearly identify the trial's primary outcome, and multiple clinical and psychological outcomes are reported together. In a cluster randomised controlled trial, this lack of prioritisation makes it difficult to identify the evaluation's main focus. The primary outcome should be stated more explicitly, with other outcomes presented as secondary (lines 21–30).

Comment 2.
The wording used to describe the results in the abstract implies a strong causal relationship between the intervention and the observed effects (e.g., “showed,” “implied the effectiveness”). The tone should be slightly moderated to better reflect the design limitations and avoid overly definitive causal claims (lines 31–36).

Introduction

Comment 3.
The introduction does not conclude with an explicit research question, despite the study's experimental nature. Including a clearly formulated research question specifying the intervention, population, and comparison would improve conceptual clarity and coherence (lines 96–103).

Comment 4.
No explicit testable hypotheses (primary or secondary) are stated. While general aims are described, a cluster randomized controlled trial assessing multiple outcomes would benefit from clearly articulated hypotheses to guide interpretation and justify the analytical approach (lines 96–103).

Comment 5.
The conceptual role of the digital component (WhatsApp) is not clearly defined. It remains unclear whether it is considered an integral part of the intervention or a supportive tool to enhance adherence. This distinction should be clarified in the introduction to facilitate interpretation of the results (lines 65–104).

 

Methods – Study Design and Sample Size

Comment 6.
The phrase “intra-cluster cluster of 0.1%” used in the sample size calculation is methodologically unclear and not standard terminology. It should be clarified whether this refers to the intra-cluster correlation coefficient (ICC), and its value and role in sample size calculation should be explained clearly (line 111).

Methods – Randomisation and Blinding

Comment 7.
The manuscript states that participants and caregivers were blinded after group allocation. Given the nature of the intervention (face-to-face sessions and participation in a WhatsApp group), this claim appears inconsistent with the study design. This statement should be revised or removed, and the type of blinding actually achieved should be clearly specified (lines 124–125).

 

Methods – Inclusion Criteria

Comment 8.
Informal caregivers were required to meet criteria for non-cancer pain and a minimum pain intensity. The rationale for this requirement is not explained and may limit the generalizability of the findings. This decision should be justified or reconsidered (lines 147–149).

Methods – Statistical Analysis

Comment 9.
Statistically significant baseline differences between groups are reported for key variables, including pain self-efficacy and activities of daily living. The manuscript does not clearly state whether or how these imbalances were accounted for in the main analyses. This requires clarification (lines 299–300; 318–319).

Comment 10.
Although multilevel regression and GEE models are mentioned, it is not explicitly stated whether baseline outcome values were included as covariates. Given the baseline imbalances, this aspect should be clearly described to improve analytical transparency (lines 219–233).

 

Results

Comment 11.
Some findings are described as improvements even though they did not reach statistical significance (e.g., pain knowledge). The language should be revised to ensure that interpretations are strictly supported by the data (lines 309–312).

Comment 12.
Baseline differences between groups are reported in the results but are not sufficiently integrated into the interpretation of intervention effects. This limits the reader’s ability to assess the net impact of the intervention (lines 299–322).

Discussion

Comment 13.
In several sections, the discussion uses language that suggests a direct causal effect of the intervention on the outcomes. Given the lack of blinding and baseline imbalances, a more cautious, qualified interpretation is recommended (lines 345–351).

Comment 14.
The potential influence of baseline differences in pain self-efficacy and activities of daily living on post-intervention outcomes is not explicitly discussed. This issue should be addressed to provide a more balanced interpretation of the findings (lines 356–367).

Limitations

Comment 15.
The limitations section does not adequately address key methodological issues, including baseline group imbalance, lack of true participant blinding, and potential selection bias related to digital literacy. These limitations should be explicitly acknowledged (lines 404–413).

Conclusions

Comment 16.
The conclusions present relatively strong statements about the program's effectiveness but do not fully reflect the methodological limitations discussed. The language should be moderated to better align conclusions with the level of evidence provided (lines 414–423).

Best Regards

Author Response

Review 1 (In yellow)

 

Comment 1.

The abstract does not clearly identify the trial's primary outcome, and multiple clinical and psychological outcomes are reported together. In a cluster randomised controlled trial, this lack of prioritisation makes it difficult to identify the evaluation's main focus. The primary outcome should be stated more explicitly, with other outcomes presented as secondary (lines 21–30).

            Response 1:

Thank you to the reviewer for pointing this out; we have revised the abstract to state the trial’s primary outcome explicitly and to present the remaining outcomes as secondary. (lines 21–29)

 

Comment 2.

The wording used to describe the results in the abstract implies a strong causal relationship between the intervention and the observed effects (e.g., “showed,” “implied the effectiveness”). The tone should be slightly moderated to better reflect the design limitations and avoid overly definitive causal claims (lines 31–36).

            Response 2:

Thank you to the reviewer for this helpful suggestion. We agree that the original abstract wording could be read as overly definitive, and we have therefore revised the Results and Conclusions to use more neutral, observational language (e.g., “were observed,” “was associated with,” “suggest”) while still accurately reporting statistically significant between-group differences. (lines 29–39)

 

Introduction

Comment 3.

The introduction does not conclude with an explicit research question, despite the study's experimental nature. Including a clearly formulated research question specifying the intervention, population, and comparison would improve conceptual clarity and coherence (lines 96–103).

            Response 3:

Thank you to the reviewer for their thoughtful and constructive feedback, which helped us improve the clarity and presentation of the manuscript. (lines 96–111)

 

Comment 4.

No explicit testable hypotheses (primary or secondary) are stated. While general aims are described, a cluster randomized controlled trial assessing multiple outcomes would benefit from clearly articulated hypotheses to guide interpretation and justify the analytical approach (lines 96–103).

            Response 4:

Thank you to the reviewer for this valuable comment.  We agree that, given the cluster randomized controlled trial design and the assessment of multiple outcomes, explicitly stated testable hypotheses (primary and secondary) improve interpretability and better justify the analytical approach. (lines 96–111)

 

Comment 5.

The conceptual role of the digital component (WhatsApp) is not clearly defined. It remains unclear whether it is considered an integral part of the intervention or a supportive tool to enhance adherence. This distinction should be clarified in the introduction to facilitate interpretation of the results. (lines 65–104)

            Response 5:

Thank you to the reviewer for highlighting this important point; we agree that clarifying the conceptual role of the WhatsApp component strengthens interpretation of the findings. (lines 114–125)

 

Methods – Study Design and Sample Size

Comment 6.

The phrase “intra-cluster cluster of 0.1%” used in the sample size calculation is methodologically unclear and not standard terminology. It should be clarified whether this refers to the intra-cluster correlation coefficient (ICC), and its value and role in sample size calculation should be explained clearly (line 111).

            Response 6:

Thank you to the reviewer for pointing this out; we agree that the phrase “intra-cluster cluster of 0.1%” is unclear and not standard terminology. (lines 131–134)

 

Methods – Randomisation and Blinding

Comment 7.

The manuscript states that participants and caregivers were blinded after group allocation. Given the nature of the intervention (face-to-face sessions and participation in a WhatsApp group), this claim appears inconsistent with the study design. This statement should be revised or removed, and the type of blinding actually achieved should be clearly specified (lines 124–125).

            Response 7:

            This sentence has now been removed. Thank you for pointing this out.

 

Methods – Inclusion Criteria

Comment 8.

Informal caregivers were required to meet criteria for non-cancer pain and a minimum pain intensity. The rationale for this requirement is not explained and may limit the generalizability of the findings. This decision should be justified or reconsidered (lines 147–149).

            Response 8:

Thank you to the reviewer for this important comment; we agree that requiring informal caregivers to meet pain-related eligibility criteria should be explicitly justified, as it may affect generalizability. (lines 158–164).

 

Methods – Statistical Analysis

Comment 9.

Statistically significant baseline differences between groups are reported for key variables, including pain self-efficacy and activities of daily living. The manuscript does not clearly state whether or how these imbalances were accounted for in the main analyses. This requires clarification (lines 299–300; 318–319).

            Response 9:

Thank you to the reviewer for this important observation; we agree that the handling of baseline imbalances should be stated explicitly in the manuscript. (lines 262–270)

 

Comment 10.

Although multilevel regression and GEE models are mentioned, it is not explicitly stated whether baseline outcome values were included as covariates. Given the baseline imbalances, this aspect should be clearly described to improve analytical transparency (lines 219–233).

            Response 10:

Thank you to the reviewer for this helpful suggestion; we agree that explicitly stating how baseline outcome values were handled improves analytical transparency, particularly given the baseline imbalances. (line 278 – 289)

 

Results

Comment 11.

Some findings are described as improvements even though they did not reach statistical significance (e.g., pain knowledge). The language should be revised to ensure that interpretations are strictly supported by the data (lines 356–360).

            Response 11:

Thank you to the reviewer for this helpful comment; we agree that the wording should distinguish clearly between statistically significant effects and non-significant trends. (line 385 – 388)

 

Comment 12.

Baseline differences between groups are reported in the results but are not sufficiently integrated into the interpretation of intervention effects. This limits the reader’s ability to assess the net impact of the intervention (lines 299–322).

            Response 12:

Thank you to the reviewer for this helpful comment; we agree that baseline imbalances should be explicitly incorporated into how intervention effects are interpreted. (line 364 – 368, 393 – 398)

 

Discussion

Comment 13.

In several sections, the discussion uses language that suggests a direct causal effect of the intervention on the outcomes. Given the lack of blinding and baseline imbalances, a more cautious, qualified interpretation is recommended (lines 345–351).

            Response 13:

Thank you to the reviewer for this important comment.  We agree that, given the lack of blinding and the observed baseline imbalances, the discussion should use more cautious language and avoid implying unequivocal causality. (line 440 – 448)

 

Comment 14.

The potential influence of baseline differences in pain self-efficacy and activities of daily living on post-intervention outcomes is not explicitly discussed. This issue should be addressed to provide a more balanced interpretation of the findings (lines 356–367).

            Response 14:

Thank you to the reviewer for this thoughtful comment; we agree that the discussion should explicitly address how baseline differences in pain self-efficacy and activities of daily living (ADL) may have influenced post-intervention outcomes. (line 452 – 470)

 

Limitations

Comment 15.

The limitations section does not adequately address key methodological issues, including baseline group imbalance, lack of true participant blinding, and potential selection bias related to digital literacy. These limitations should be explicitly acknowledged (lines 404 – 413).

            Response 15:

Thank you to the reviewer for this important suggestion; we agree that these methodological limitations should be stated more explicitly. (line 507 – 522)

 

Conclusions

Comment 16.

The conclusions present relatively strong statements about the program's effectiveness but do not fully reflect the methodological limitations discussed. The language should be moderated to better align conclusions with the level of evidence provided (lines 414 – 423).

            Response 16:

Thank you to the reviewer for this important comment; we agree that the conclusions should be phrased more cautiously to align with the study’s methodological limitations. (line 524 – 535)

Reviewer 2 Report

Comments and Suggestions for Authors

The study addresses an important and timely need for scalable, non-pharmacological pain management in older adults, yet the novelty of the dyadic plus WhatsApp-supported approach should be more clearly distinguished from existing mHealth and caregiver-assisted interventions.

The cluster randomized design is appropriate, but the randomisation and clustering structure are insufficiently described, particularly regarding the number of centers, intracluster correlation assumptions, and whether analyses fully accounted for clustering at all time points.

Several baseline imbalances are evident, including significant differences in pain self-efficacy and activities of daily living, and the manuscript should provide adjusted models or baseline-corrected estimates to ensure that intervention effects are not partly driven by pre-intervention differences.

The intervention is well detailed, but the digital component remains somewhat under-evaluated, and the authors should report engagement metrics such as message frequency, adherence to home exercises, and variation in WhatsApp participation across dyads.

The statistical analysis plan lacks clarity, as both multilevel regression and GEE are mentioned without specifying which outcomes used which model, and the manuscript would benefit from a more transparent primary analytic framework with explicit handling of missing data and intention-to-treat procedures.

The outcome reporting focuses heavily on p-values, while effect sizes and clinical relevance are less developed, and the results section should emphasize magnitude of change and minimally important differences for pain and psychological outcomes.

Baseline imbalances, limited reporting of clustering and engagement, and insufficient analytical transparency weaken the strength of the conclusions. With clearer methodological reporting, adjusted analyses, and more detailed evaluation of the digital component, the manuscript could make a meaningful contribution to the growing literature on scalable chronic pain interventions in aging populations.

Author Response

Reviewer 2 (in green)

Comment: The study addresses an important and timely need for scalable, non-pharmacological pain management in older adults, yet the novelty of the dyadic plus WhatsApp-supported approach should be more clearly distinguished from existing mHealth and caregiver-assisted interventions.

            Response:

Thank you for this comment. We agree that the manuscript should more explicitly differentiate the “dyadic + WhatsApp-supported” model from (i) patient-only mHealth pain self-management and (ii) caregiver-assisted programs without an asynchronous, low-burden messaging channel. We will revise the Introduction and Discussion to state the specific innovation as combining dyadic skill-building with structured, ongoing WhatsApp micro-support that targets adherence, troubleshooting, and caregiver co-regulation between sessions (not merely “digital delivery”). (line 127 – 135)

The cluster randomized design is appropriate, but the randomisation and clustering structure are insufficiently described, particularly regarding the number of centers, intracluster correlation assumptions, and whether analyses fully accounted for clustering at all time points.

Response:

Thank you for your valuable feedback on our study's randomization and clustering details. (line 204 – 218)

Comment: Several baseline imbalances are evident, including significant differences in pain self-efficacy and activities of daily living, and the manuscript should provide adjusted models or baseline-corrected estimates to ensure that intervention effects are not partly driven by pre-intervention differences.

Response:

Thank you for this comment. We agree that baseline imbalances (e.g., pain self-efficacy, ADL) can bias unadjusted estimates. (line 264 – 270)

Comment: The intervention is well detailed, but the digital component remains somewhat under-evaluated, and the authors should report engagement metrics such as message frequency, adherence to home exercises, and variation in WhatsApp participation across dyads.

Response:

Thank you for this comment. We agree and will add a dedicated “digital engagement/fidelity” subsection plus a table summarizing WhatsApp participation and home-practice adherence. We will report engagement using objective metrics where available (e.g., number of messages exchanged, response rates, completion of prompts), consistent with common engagement measurement approaches in mHealth studies. (line 252 – 258)

Comment: The statistical analysis plan lacks clarity, as both multilevel regression and GEE are mentioned without specifying which outcomes used which model, and the manuscript would benefit from a more transparent primary analytic framework with explicit handling of missing data and intention-to-treat procedures.

Response:

Thank you for this comment. We agree the analysis section should specify one transparent primary framework. (line 278 – 289)

Comment: The outcome reporting focuses heavily on p-values, while effect sizes and clinical relevance are less developed, and the results section should emphasize magnitude of change and minimally important differences for pain and psychological outcomes.

Response:

Thank you for highlighting the need to emphasize effect sizes, magnitude of change, and clinical relevance alongside p-values; we agree this strengthens the interpretation of our findings. (line 413 – 431)

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

Dear Authors,

Thank you for the revised version of your manuscript. The paper has improved considerably, and I appreciate the care taken in addressing most of the reviewer’s concerns.

In particular, the following issues have been appropriately handled:

• The primary outcome is now clearly identified in the abstract.

• The language has been moderated to avoid overstatement of causal effects.

• Explicit primary and secondary hypotheses have been added.

• The role of the WhatsApp component has been conceptually clarified.

• Baseline imbalances and lack of blinding are now properly acknowledged in both Results and Discussion.

• The limitations section is more balanced and transparent.

These revisions significantly strengthen the manuscript.

That said, a few points still need clarification before the manuscript can be considered fully consistent methodologically:

1. Intra-cluster correlation coefficient (ICC)
   Two different ICC values (0.10 and 0.001) appear in the manuscript. This must be clarified, as it directly affects the sample size calculation in a cluster RCT. Please ensure consistency and briefly explain which value was ultimately used.

2. Adjusted analyses
   The Methods indicate that baseline-adjusted multilevel/GEE models were used. However, it is not fully clear in the Results tables whether the reported p-values come from adjusted models or simple comparisons. This should be explicitly stated (e.g., in table footnotes).

3. Digital engagement metrics
   Engagement indicators for the WhatsApp component are described in Methods but not reported in Results. Please either provide a brief summary or clarify whether these analyses are beyond the scope of the current paper.

Overall, this is a well-designed and relevant study, and with these clarifications, it should be ready for publication.

Kind regards,
Reviewer

Author Response

Dear Reviewer 1,

Review 1 (In blue)

Intra-cluster correlation coefficient (ICC)

Two different ICC values (0.10 and 0.001) appear in the manuscript. This must be clarified, as it directly affects the sample size calculation in a cluster RCT. Please ensure consistency and briefly explain which value was ultimately used.

            Response:

Thank you for noting the inconsistency. We have corrected the manuscript so that a single ICC value is used consistently throughout the sample size description. (line 206 – 207)

 

Adjusted analyses

The Methods indicate that baseline-adjusted multilevel/GEE models were used. However, it is not fully clear in the Results tables whether the reported p-values come from adjusted models or simple comparisons. This should be explicitly stated (e.g., in table footnotes).

            Response:

Thank you for this helpful suggestion. We clarified in the Results tables (footnotes) which p-values come from baseline-adjusted multilevel/GEE models and which come from unadjusted/simple comparisons. Specifically, the primary inference regarding intervention effects over time is based on the longitudinal multilevel/GEE models that include baseline values (and other prespecified covariates as applicable) as described in the Statistical Analysis section. (line 386 – 389, 406 – 410, 439 – 443)

 

Digital engagement metrics

Engagement indicators for the WhatsApp component are described in Methods but not reported in Results. Please either provide a brief summary or clarify whether these analyses are beyond the scope of the current paper.

            Response:

Thank you for highlighting this. We agree that engagement indicators for the WhatsApp component are described in Methods; however, they were not included in the current Results because the present paper focuses on clinical effectiveness outcomes of the overall multicomponent programme (face-to-face plus WhatsApp reinforcement), and the engagement analyses (e.g., message frequency/responsiveness/duration of active participation and their relationship to outcomes) were prespecified as secondary/process outcomes.