Policy, Financing, and Regulatory Barriers to Adopting AI-Powered Electrocardiography Interpretation Clinical Decision Support System in Ethiopia: A Qualitative Study

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

While the study addresses an important topic and the qualitative design is generally well-conceived, the manuscript lacks the level of analytic depth needed to fully support its conclusions. In particular, the results section does not provide sufficient analytical detail regarding the relationships between stakeholder groups and the emergent themes, nor does it explore how stakeholders’ perspectives relate to the identified system‑level determinants. As presented, the findings remain largely descriptive.

To strengthen the rigor and relevance of the study, the authors should consider refining their methodological approach to ensure that the analytic procedures align with the research objectives. This may include (but is not limited to):

• Incorporating structured qualitative analytic strategies (e.g., matrix analyses, cross‑case comparisons, or stakeholder‑specific thematic mapping) that allow for meaningful examination of relationships across groups.
• Presenting clearer outcomes that demonstrate how stakeholder assessments converge or diverge in relation to system‑level determinants.
• Providing analytic justification and transparent documentation of how themes were derived and how they relate to the overarching conceptual framework.

Overall, the study has the potential to make a valuable contribution, but the authors will need to enhance the methodological detail and present their analytic outcomes in a way that offers stronger interpretive relevance and supports more actionable conclusions.

Author Response

Reviewer 1

Comment:  While the study addresses an important topic and the qualitative design is generally well-conceived, the manuscript lacks the level of analytic depth needed to fully support its conclusions. In particular, the results section does not provide sufficient analytical detail regarding the relationships between stakeholder groups and the emergent themes, nor does it explore how stakeholders’ perspectives relate to the identified system‑level determinants. As presented, the findings remain largely descriptive.

Response: We thank the reviewer for the comment. Thank you for this thoughtful and constructive comment. We appreciate the reviewer’s concern regarding the analytical depth of the results.

We would like to clarify that this study was designed as a formative qualitative inquiry, with the primary aim of generating in-depth, context-specific insights from key stakeholders involved in or affected by the integration of AI-powered ECG technology in real-world clinical practice within the health system. Formative qualitative research is particularly appropriate for understanding complex systems, capturing diverse stakeholder perspectives, and identifying contextual factors that shape the feasibility and acceptability of emerging health technologies. Accordingly, our analysis intentionally emphasized rich, descriptive accounts of participants’ experiences and perspectives to generate actionable evidence that can inform policy and programmatic decision-making in the Ethiopian context.

Regarding the reviewer’s specific concern, we would like to note that the relationships between stakeholder perspectives, emergent themes, and system-level determinants were explored within the analytical framework, though not always presented in an explicitly comparative manner. In particular, policy and regulatory perspectives were analyzed and presented as a distinct thematic domain, while system-level determinants (e.g., infrastructure, workforce capacity, governance, and trust) were also examined as cross-cutting themes emerging from the data. We acknowledge that the presentation of findings may appear primarily descriptive; however, this reflects the formative purpose of the study, which prioritizes contextual depth and practical relevance over extensive abstraction. At the same time, to address the reviewer’s concern without substantially altering the structure of the results section, we have made targeted revisions to improve clarity. Specifically, we have added a paragraph that clarified the linkages between stakeholder views and key system-level determinants.

These refinements enhance the interpretive clarity of the findings while preserving the original structure and intent of the results. We believe the revised manuscript more clearly conveys how stakeholder perspectives and system-level factors interact to influence the integration of AI-powered ECG technology in the Ethiopian health system.

Comment: To strengthen the rigor and relevance of the study, the authors should consider refining their methodological approach to ensure that the analytic procedures align with the research objectives. This may include (but is not limited to):

• Incorporating structured qualitative analytic strategies (e.g., matrix analyses, cross‑case comparisons, or stakeholder‑specific thematic mapping) that allow for meaningful examination of relationships across groups.
• Presenting clearer outcomes that demonstrate how stakeholder assessments converge or diverge in relation to system‑level determinants.
• Providing analytic justification and transparent documentation of how themes were derived and how they relate to the overarching conceptual framework.

Overall, the study has the potential to make a valuable contribution, but the authors will need to enhance the methodological detail and present their analytic outcomes in a way that offers stronger interpretive relevance and supports more actionable conclusions.

Response: Thank you for this detailed and constructive feedback. We appreciate the reviewer’s emphasis on strengthening methodological rigor, analytical clarity, and interpretive relevance.

We would like to clarify that the study was designed as a formative qualitative inquiry aimed at generating context-specific, actionable insights to inform the integration of AI-powered ECG technology within the Ethiopian health system. In line with this objective, we employed a thematic analysis approach that prioritized depth of understanding and practical relevance of stakeholder perspectives across multiple levels of the health system.

Regarding the first and second points, we acknowledge the importance of structured analytic strategies and clearer articulation of convergence and divergence across stakeholder groups. Although our original analysis incorporated these elements, they were not always explicitly presented. In response, and in alignment with the reviewer’s earlier comment, we have refined the results section to more clearly demonstrate how stakeholder perspectives (e.g., clinicians, administrators, and policymakers) converge and diverge in relation to key system-level determinants such as infrastructure, workforce capacity, governance, and trust in AI. These enhancements improve the analytical depth while maintaining the formative and nuanced structure of the results.

Regarding the third point, we would like to clarify that the manuscript already provides detailed documentation of the analytic procedures. In the methods section, we explicitly describe the processes of coding, iterative codebook development, and theme generation. In addition, at the beginning of the results section, we outline how themes were derived and organized in relation to the study’s conceptual focus. To further address the reviewer’s concern, we have revised these sections to improve clarity. Overall, these revisions enhance the transparency, rigor, and interpretive value of the analysis without substantially altering the structure of the manuscript. We believe the updated version more clearly demonstrates how the analytic approach aligns with the study objectives and provides stronger, actionable insights for policy and practice in the integration of AI-powered ECG technology in low-resource health systems.

Reviewer 2 Report

Comments and Suggestions for Authors

This study focuses on the systemic barriers to the adoption of AI-powered electrocardiography interpretation technology in Ethiopia. The research topic aligns with the practical needs of medical technology implementation in low- and middle-income countries, providing valuable empirical evidence for relevant policy formulation and practice advancement. However, certain aspects can be further improved through text supplementation, logical optimisation, and other non-experimental adjustments.

 

1. The article mentions that Ethiopia lacks specific policies for AI in healthcare. It is recommended to supplement explanations in the policy section regarding the alignment between AI-powered electrocardiography technology and existing policies, so as to reduce redundant policy development costs and enhance the feasibility of the recommendations.

 

2. While core regulatory bodies such as the Ministry of Health and the EFDA have been identified, the specific responsibility boundaries of each entity in processes, including technical approval, quality supervision, and follow-up monitoring, are not clarified. Based on the existing regulatory system, a concise description of responsibility division, can be added to make the regulatory recommendations clearer.

 

3. The article addresses financing mechanisms. It is advisable to include successful financing cases of similar technologies, e.g., AI diagnostic tools from other low- and middle-income countries, providing practical and replicable references for Ethiopia.

 

4. Current recommendations focus on principled requirements. Simple and actionable operational suggestions can be supplemented, such as referencing existing electronic medical record data security standards and clarifying permission classification criteria for AI data usage. These adjustments align with Ethiopia’s current data management foundation and enhance the implementability of the recommendations.

 

5. The limitations section notes gender imbalance in the sample and the lack of perspectives from frontline personnel. In the discussion section, supplementary explanations can be added regarding the potential impacts of these limitations on the study’s conclusions, e.g., potential differences in perceptions of financing mechanisms between male decision-makers and female practitioners.

 

6. The discussion section can be enriched by including the insights of the following articles:

 

[1] Al-Khatib SM, Singh JP, Ghanbari H, McManus DD, Deering TF, Silva JN, Mittal S, Krahn A, Hurwitz JL. The potential of artificial intelligence to revolutionize health care delivery, research, and education in cardiac electrophysiology. Heart Rhythm. 2024 Jun 1;21(6):978-89. doi:10.1016/j.hrthm.2024.04.053

[2] Arends BK, McCormick JM, van Der Harst P, Heus P, van Es R. Barriers, facilitators and strategies for the implementation of artificial intelligence‐based electrocardiogram interpretation: A mixed‐methods study. European Journal of Clinical Investigation. 2025 Apr;55:e14387. doi:10.1111/eci.14387

Author Response

Reviewer 2

Comment: - This study focuses on the systemic barriers to the adoption of AI-powered electrocardiography interpretation technology in Ethiopia. The research topic aligns with the practical needs of medical technology implementation in low- and middle-income countries, providing valuable empirical evidence for relevant policy formulation and practice advancement. However, certain aspects can be further improved through text supplementation, logical optimization, and other non-experimental adjustments.

1. The article mentions that Ethiopia lacks specific policies for AI in healthcare. It is recommended to supplement explanations in the policy section regarding the alignment between AI-powered electrocardiography technology and existing policies, so as to reduce redundant policy development costs and enhance the feasibility of the recommendations.

Response: Thank you for this constructive and insightful comment. We agree that clarifying the alignment between AI-powered electrocardiography (ECG) technologies and existing policy frameworks would strengthen the practical relevance and feasibility of our recommendations.

In response, we have expanded the policy part on the discussion section to better articulate how AI-enabled ECG interpretation can be integrated within Ethiopia’s current health and digital policy landscape.  Specifically, we have highlighted its alignment with existing health sector priorities, particularly in the areas of digital health transformation and NCD management and broader health system strengthening initiatives. We also discuss how ongoing efforts in digital transformation, data governance, and medical device regulation can be leveraged to accommodate AI-based tools, thereby minimizing redundancy in policy development.

Furthermore, we have clarified that a phased and adaptive policy approach building on existing structures while addressing AI-specific considerations such as accountability, validation, and data privacy may offer a more feasible and resource-efficient pathway for implementation in low-resource settings. These revisions aim to enhance the coherence, applicability, and policy relevance of the manuscript, in line with the reviewer’s suggestion.

Comment: - 2. While core regulatory bodies such as the Ministry of Health and the EFDA have been identified, the specific responsibility boundaries of each entity in processes, including technical approval, quality supervision, and follow-up monitoring, are not clarified. Based on the existing regulatory system, a concise description of responsibility division, can be added to make the regulatory recommendations clearer.

Response:  Thank you for this important and constructive comment. We appreciate the need for clearer articulation of responsibility boundaries among the key regulatory bodies.

We would like to clarify that our manuscript already outlines the general division of responsibilities within the existing regulatory framework. Specifically, the discussion highlights the Ministry of Health (MoH) as the lead body for overall governance, including policy development, strategic direction, and ethical guidance, while the Ethiopian Food and Drug Authority (EFDA) is identified as responsible for regulation, certification, and quality assurance of AI-powered ECG technologies. In addition, the manuscript emphasizes a multi-stakeholder governance model involving institutions such as INSA and EAII to address cross-cutting areas like data security, innovation, and capacity building.

However, in response to the reviewer’s suggestion, we have further refined the relevant section to make these role distinctions more explicit and concise. In particular, we now clearly indicate that EFDA’s role encompasses technical approval (pre-market evaluation), quality supervision (compliance and standards enforcement), and follow-up monitoring (post-market surveillance), while the MoH maintains a stewardship and coordination role across the system. We have also streamlined the description of how supporting institutions contribute to specific functional areas. These revisions improve the clarity of responsibility division and strengthen the practical relevance of our regulatory recommendations.

Comment: 3. The article addresses financing mechanisms. It is advisable to include successful financing cases of similar technologies, e.g., AI diagnostic tools from other low- and middle-income countries, providing practical and replicable references for Ethiopia.

Response: Thank you for this helpful and constructive suggestion. We agree that incorporating examples of successful financing mechanisms from similar low- and middle-income country (LMIC) contexts would enhance the practical relevance and applicability of our discussion.

In response, we have revised the financing section to include illustrative examples of how AI-based diagnostic technologies have been funded and implemented in comparable settings. Specifically, we highlight models such as public–private partnerships, donor-supported pilot programs, and integration of digital health innovations into existing health financing schemes. We also reference experiences where AI diagnostic tools were introduced through phased implementation supported by international partners, with gradual transition to domestic financing to ensure sustainability.

These examples are used to demonstrate feasible and context-sensitive financing pathways that could be adapted to Ethiopia, while acknowledging differences in health system structure and resource availability. The additions aim to provide more practical, evidence-informed guidance for policymakers and stakeholders considering the scale-up of AI-powered ECG technologies.

Comment: Current recommendations focus on principled requirements. Simple and actionable operational suggestions can be supplemented, such as referencing existing electronic medical record data security standards and clarifying permission classification criteria for AI data usage. These adjustments align with Ethiopia’s current data management foundation and enhance the implementation of the recommendations.

Response: Thank you for this insightful and practical suggestion. We agree that complementing the principled recommendations with more concrete and operational guidance would enhance the implementation of the manuscript.

In response, we have revised the relevant section to incorporate more actionable details. Specifically, we now reference existing electronic medical record (EMR) data security practices and standards currently in use within Ethiopia’s digital health system as a foundation for governing AI-related data. In addition, we have clarified the need for defining permission and access classification criteria for AI data usage, including distinctions between identifiable and de-identified data, as well as role-based access levels for different users.

We also highlight how these operational elements can be aligned with Ethiopia’s ongoing digital health and data governance initiatives, thereby avoiding the need for entirely new systems while strengthening current frameworks. These additions aim to improve the clarity, feasibility, and practical relevance of our recommendations.

 Comment: The limitations section notes gender imbalance in the sample and the lack of perspectives from frontline personnel. In the discussion section, supplementary explanations can be added regarding the potential impacts of these limitations on the study’s conclusions, e.g., potential differences in perceptions of financing mechanisms between male decision-makers and female practitioners.

Response: Thank you for this valuable suggestion. We agree that further elaboration on the implications of these limitations would strengthen the interpretation of our findings.

While the manuscript already acknowledges the overrepresentation of high-level stakeholders and the existing gender imbalance, we have now expanded the discussion section to more explicitly describe how these factors may have influenced the study’s conclusions. In particular, we clarify that the dominance of perspectives from senior, predominantly male decision-makers may have shaped the findings toward system-level priorities such as policy, financing, and infrastructure—while potentially underrepresenting operational and practice-level concerns.

We also note that frontline healthcare providers, including a greater proportion of female practitioners, may have different perspectives on key issues such as feasibility, usability, workload implications, and access to resources. For example, perceptions of financing mechanisms and implementation priorities may differ between decision-makers and frontline staff, which could influence how acceptable and practical proposed solutions are in real-world settings. These additions aim to provide a more nuanced interpretation of the findings and to highlight the importance of including more diverse and representative stakeholder groups in future research.

Comment:6. The discussion section can be enriched by including the insights of the following articles:

[1] Al-Khatib SM, Singh JP, Ghanbari H, McManus DD, Deering TF, Silva JN, Mittal S, Krahn A, Hurwitz JL. The potential of artificial intelligence to revolutionize health care delivery, research, and education in cardiac electrophysiology. Heart Rhythm. 2024 Jun 1;21(6):978-89. doi:10.1016/j.hrthm.2024.04.053

[2] Arends BK, McCormick JM, van Der Harst P, Heus P, van Es R. Barriers, facilitators and strategies for the implementation of artificial intelligence‐based electrocardiogram interpretation: A mixed‐methods study. European Journal of Clinical Investigation. 2025 Apr;55:e14387. doi:10.1111/eci.14387

Response: Thank you for this helpful recommendation. We appreciate the suggested literature and its relevance to strengthening the discussion.

In response, we carefully reviewed both articles. We found the second reference (Arends et al., 2025) to be highly aligned with our study focus, particularly in its detailed examination of barriers, facilitators, and implementation strategies for AI-based ECG interpretation. Accordingly, we have incorporated this study into the discussion to enrich our analysis, especially in areas related to implementation challenges, stakeholder engagement, and system readiness. The inclusion of this reference allows us to better situate our findings within the broader empirical literature and to highlight converging evidence across different contexts. These revisions strengthen the discussion by linking our findings with relevant and recent empirical evidence on AI-enabled ECG implementation.

 

Reviewer 3 Report

Comments and Suggestions for Authors

Dear authors,

General comment

Overall, the manuscript addresses a relevant topic for health systems, particularly in low- and middle-income countries, by examining the factors that influence the adoption of artificial intelligence–based technologies for ECG interpretation. The study stands out for considering political, regulatory, financial, and governance dimensions, which allows the implementation of these innovations to be understood beyond purely technical aspects. Likewise, the qualitative approach and the participation of different actors within the health system provide a contextual perspective that is valuable for decision-making in digital health. Taken together, the work represents a relevant contribution to the analysis of the integration of artificial intelligence into health systems. However, I have several observations that I would like the authors to consider.

 

Highlights

• I believe it would be preferable for the authors to provide three specific “highlights,” potentially one corresponding to each research question of interest (as six may be excessive).

   

  • Public health relevance —How does this work relate to a public health issue?  
  • Public health significance—Why is this work of significance to public health?
  • Public health implications —What are the key implications or messages for practitioners, policy makers and/or researchers in public health?

Introduction

• The text highlights the potential of artificial intelligence aplied to electrocardiograms to improve diagnosis in resource-limited settings such as Ethiopia. However, the argument remains at a general level and lacks references or evidence supporting the claims about global trends.
• Sources supporting the challenges mentioned for low- and middle-income countries are also lacking.
• The ethiopian context could be described in greater depth by including concrete examples of policies or digital health initiatives.
• Although the objective of the study is clearly stated, the section could be strengthened with greater empirical support.

 

Methods

• The description of the methods is excesively lengthy and somewhat repetitive in several sections, which afects clarity and conciseness.
• Although maximum variation purposive sampling is mentioned, the manuscript does not sufficiently explain how participants were identified and selected.
• The justification of the sample size is limited to a general reference to theoretical saturation.
• The tematic analysis process is described in general terms without clearly detailing how categories were developed.
• Overall, although the methodology is well structured, it could be strengthened through a more concise presentation and greater precision in some key analytical procedures.

 

Results

• The content is overly extensive and repetitive, reiterating similar ideas regarding policy, financiing, and data governance.
• Some sections adopt a descriptive or explanatory tone that resembles a conceptual review rather than the presentation of qualitative results.
• Few direct quotations from participants are included, and the link between findings and participants’ perspectives is not always clear.
• The structure could be simplified by grouping similar ideas and prioritizing the most relevant findings.
• This would improve clarity, syntesis, and analytical rigor within the section.

 

Discussion

• A considerable portion of the content reiterates previously presented results, with limited analytical depth or critical interpretation of the findings.
• Although global studies and experiences are mentioned, clear comparisons with previous research are not established to better contextualize the findings.
• Some sections maintain a descriptive and normative tone, closer to policy recommendations than to an evidence-based academic discussion.
• The discussion includes only 11 references (30% of the total references in the article), which limits engagement with the existing literature. Expanding the number of references and strengthening integration with previous studies would therefore be advisable.
• Overall, the section could be improved through a more critical analysis of the findings and a more concise presentation of the main arguments.

 

Conclusions

• The conclusions mainly reiterate the study results without analytically syntesizing the most relevant implications of the findings.
• Some statements remain descriptive and general in nature.
• The specific contribution of the study relative to the existing literature is not sufficiently highlighted.
• The section could benefit from a more concise formulation.
• It would also be appropriate to more clearly emphasize the practical implications and potetial directions for future research derived from the results.

Author Response

Reviewer 3

Comment: General comment

Overall, the manuscript addresses a relevant topic for health systems, particularly in low- and middle-income countries, by examining the factors that influence the adoption of artificial intelligence–based technologies for ECG interpretation. The study stands out for considering political, regulatory, financial, and governance dimensions, which allows the implementation of these innovations to be understood beyond purely technical aspects. Likewise, the qualitative approach and the participation of different actors within the health system provide a contextual perspective that is valuable for decision-making in digital health. Taken together, the work represents a relevant contribution to the analysis of the integration of artificial intelligence into health systems. However, I have several observations that I would like the authors to consider.

Highlights

• I believe it would be preferable for the authors to provide three specific “highlights,” potentially one corresponding to each research question of interest (as six may be excessive).
• Public health relevance —How does this work relate to a public health issue?  
• Public health significance—Why is this work of significance to public health?
• Public health implications —What are the key implications or messages for practitioners, policy makers and/or researchers in public health?

Response: Thank you for this comprehensive and encouraging assessment of our manuscript. We appreciate the reviewer’s recognition of the study’s relevance, particularly its multidimensional perspective on the adoption of AI-based ECG technologies in low- and middle-income country contexts.

We also thank you for the constructive suggestion regarding the “Highlights” section. In response, we have revised this section to present three concise and focused highlights, aligned with the key dimensions suggested. Specifically, we now provide: (1) a statement on the public health relevance of AI-powered ECG technologies in addressing cardiovascular disease burden and diagnostic gaps; (2) a clear articulation of the study’s public health significance, emphasizing the importance of understanding systemic, non-technical barriers and enablers; and (3) key public health implications, outlining actionable insights for policymakers, practitioners, and researchers regarding governance, financing, and implementation strategies. This revision improves the clarity, focus, and policy relevance of the manuscript, in line with the reviewer’s recommendation.

Comment: Introduction

• The text highlights the potential of artificial intelligence aplied to electrocardiograms to improve diagnosis in resource-limited settings such as Ethiopia. However, the argument remains at a general level and lacks references or evidence supporting the claims about global trends.
• Sources supporting the challenges mentioned for low- and middle-income countries are also lacking.

 Response: Thank you for this important observation. We agree that strengthening the introduction with relevant references would improve the rigor and contextual grounding of the manuscript.

In response, we have revised the introduction to incorporate up-to-date and relevant literature supporting the global trends in the application of artificial intelligence to electrocardiography, particularly its potential to improve diagnostic accuracy and expand access in resource-limited settings. We have also added appropriate references to substantiate the challenges faced by low- and middle-income countries, including limited access to diagnostic technologies, shortages of specialized healthcare professionals, and infrastructure constraints. These additions enhance the credibility of our arguments and better position the study within the existing global and LMIC-focused evidence base.

Comment: The Ethiopian context could be described in greater depth by including concrete examples of policies or digital health initiatives.

• Although the objective of the study is clearly stated, the section could be strengthened with greater empirical support.

Response: Thank you for these valuable suggestions. We agree that providing a more detailed contextual background and strengthening the empirical grounding of the introduction would improve the manuscript.

In response, we have expanded the description of the Ethiopian context by incorporating concrete examples of existing policies and digital health initiatives, such as the national Digital Health Strategy, ongoing eHealth implementations, and efforts toward health system digitalization. These additions help better situate the study within the country’s current policy and implementation landscape. Furthermore, we have strengthened the section by adding relevant empirical references to support the study objective and the broader rationale for examining AI-powered ECG adoption in Ethiopia. These revisions enhance the clarity, contextual depth, and evidence base of the introduction.

Comment: Methods

• The description of the methods is excessively lengthy and somewhat repetitive in several sections, which affects clarity and conciseness.

Response: Thank you for this helpful observation. We agree that the Methods section could be streamlined to improve clarity and readability.

In response, we have carefully revised the section to reduce redundancy and eliminate repetitive descriptions across subsections. We have consolidated overlapping content, simplified wording, and ensured that each component of the methodology is presented clearly and concisely while retaining all essential details required for transparency and reproducibility. These revisions improve the overall flow and readability of the Methods section without compromising methodological rigor.

Comment: Although maximum variation purposive sampling is mentioned, the manuscript does not sufficiently explain how participants were identified and selected.

The justification of the sample size is limited to a general reference to theoretical saturation.

Response: Thank you for these valuable comments. We agree that additional detail on participant selection and sample size justification improves methodological transparency.

In response, we have revised the Methods section to provide a clearer explanation of the maximum variation purposive sampling process. Participants were identified and selected based on their roles, expertise, and relevance to AI-powered ECG adoption, including policymakers, regulators, digital health and innovation experts, health financing professionals, and hospital leaders. Geographic diversity was also considered, encompassing federal institutions, four administrative regions, and five tertiary hospitals, to ensure a broad representation of perspectives across the health system.

Regarding sample size, we have elaborated that the 25 participants were recruited iteratively until thematic saturation was reached, meaning no new themes or insights emerged from additional interviews. This justification, aligned with established qualitative research standards, ensures that the sample size was sufficient to provide rich and actionable evidence while maintaining methodological rigor. These revisions strengthen the transparency and credibility of the study’s sampling strategy and sample size determination.

Comment: The thematic analysis process is described in general terms without clearly detailing how categories were developed.

• Overall, although the methodology is well structured, it could be strengthened through a more concise presentation and greater precision in some key analytical procedures.

Response: Thank you for these helpful and constructive comments. We agree that greater clarity in the description of the thematic analysis and improved conciseness would strengthen the Methods section.

In response, we have revised the manuscript to more explicitly describe how categories were developed during the analytic process. We now clarify that codes were generated inductively from the data and iteratively grouped into categories based on recurring patterns and conceptual similarities. These categories were then refined through constant comparison across transcripts and collaborative discussions among the research team, leading to the development of final themes. In addition, we have streamlined the overall presentation of the Methods section by removing redundancies and improving precision in the description of key analytical steps. These revisions enhance both the clarity and methodological rigor of the study.

Comment: Result The content is overly extensive and repetitive, reiterating similar ideas regarding policy, financing, and data governance.

Response: Thank you for this insightful comment. There might be some sections of the manuscript seems repetitive, but this reflects the inherently interconnected nature of these system-level domains, where policy frameworks, regulation, governance, and financing are closely interrelated and often overlap in shaping the adoption of AI technologies. To address the concern, we have revised the manuscript to better articulate these interrelationships while reducing unnecessary repetition. Specifically, we have consolidated overlapping content and clarified linkages between these concepts to present them more coherently and succinctly. These revisions improve the clarity and readability of the manuscript while preserving the integrated perspective that is central to our analysis.

Comment: Some sections adopt a descriptive or explanatory tone that resembles a conceptual review rather than the presentation of qualitative results.

Response: Thank you for this important observation. We agree that maintaining a clear distinction between empirical findings and broader conceptual discussion is essential. In response, we have revised the manuscript to ensure that the Results section more clearly reflects the qualitative nature of the study. Specifically, we have reduced overly descriptive or generalized statements and strengthened the presentation of findings by grounding them more explicitly in participants’ perspectives and themes derived from the data. At the same time, broader explanatory and interpretive content has been streamlined and appropriately repositioned within the Discussion section to maintain a clear separation between results and interpretation. These revisions enhance the methodological rigor and clarity of the manuscript by ensuring that the findings are presented as data-driven insights.

Comment: Few direct quotations from participants are included, and the link between findings and participants’ perspectives is not always clear.

• The structure could be simplified by grouping similar ideas and prioritizing the most relevant findings.
• This would improve clarity, syntesis, and analytical rigor within the section.

Response: Thank you for these valuable and constructive comments. We agree that strengthening the linkage between findings and participants’ perspectives, as well as improving structure and synthesis, would enhance the quality of the Results section. In response, we have revised the manuscript and tried to ensure that the findings are clearly grounded in the data. We have also improved the clarity of presentation by explicitly linking interpretations to participants’ views.

Furthermore, we have streamlined the structure of the Results section by grouping related ideas and prioritizing the most relevant findings. Redundant or less central content has been reduced to improve synthesis and readability. These revisions enhance the clarity, coherence, and analytical rigor of the findings.

Comment: A considerable portion of the content reiterates previously presented results, with limited analytical depth or critical interpretation of the findings.

Response: Thank you for this important and constructive comment. We agree that the Discussion section should go beyond restating results and provide deeper analytical insight.

In response, we have carefully revised the manuscript to reduce repetition of previously presented findings and strengthen the analytical depth of the discussion. Specifically, we have focused on interpreting the results in relation to existing literature, highlighting convergences and divergences, and drawing out their implications for policy and practice. These revisions enhance the interpretive rigor of the manuscript and ensure that the discussion provides meaningful insights beyond descriptive reporting of results.

Comment: Although global studies and experiences are mentioned, clear comparisons with previous research are not established to better contextualize the findings.

Response: Thank you for this valuable comment. We agree that a clearer comparison with existing studies would strengthen the contextualization of our findings. In response, we have revised the discussion section to more explicitly compare our results with previous research. We now highlight key areas of convergence and divergence with global and LMIC-based studies, particularly regarding policy readiness, regulatory challenges, financing constraints, and data governance issues in AI adoption. These revisions improve the analytical depth of the manuscript and better situate our findings within the broader body of evidence.

Comment: Some sections maintain a descriptive and normative tone, closer to policy recommendations than to an evidence-based academic discussion.

Response: Thank you for this insightful comment. We agree that maintaining an appropriate academic tone grounded in evidence is essential.

In response, we have revised the relevant sections to reduce overly descriptive and normative language. We have reframed statements to more clearly reflect evidence-based interpretation of the findings, ensuring that claims are supported by data and relevant literature. Policy-oriented recommendations have been streamlined and more clearly distinguished from the analytical discussion. These revisions enhance the academic rigor and balance of the manuscript.

Comment: The discussion includes only 11 references (30% of the total references in the article), which limits engagement with the existing literature. Expanding the number of references and strengthening integration with previous studies would therefore be advisable.

Overall, the section could be improved through a more critical analysis of the findings and a more concise presentation of the main arguments.

Response: Thank you for these valuable and constructive comments. We agree that a stronger engagement with the existing literature and more critical analysis would improve the discussion section. In response, we have expanded the number of references to better reflect relevant global and LMIC-based evidence, particularly in areas related to AI adoption, digital health governance, and implementation challenges. This has allowed us to more effectively situate our findings within the broader body of research and to draw clearer comparisons with previous studies.

Additionally, we have revised the discussion to enhance critical analysis by moving beyond descriptive summaries and providing deeper interpretation of the findings, including their implications, contextual nuances, and areas of convergence and divergence with existing evidence. We have also streamlined the section to reduce redundancy and present the main arguments more concisely. These revisions strengthen both the analytical depth and the scholarly rigor of the manuscript.

Comment: The conclusions mainly reiterate the study results without analytically synthesizing the most relevant implications of the findings.

Response: Thank you for this important observation. We agree that the conclusion section should provide a more synthesized and analytical summary of the study’s key implications.

In response, we have revised the conclusion to move beyond a simple restatement of results and instead emphasize the most critical insights and their broader implications for policy, practice, and research. We now highlight how the identified policy, regulatory, financing, and governance factors collectively shape the adoption of AI-powered ECG technologies, and we articulate the key priorities for enabling their responsible and sustainable integration within the Ethiopian health system. These revisions improve the clarity, synthesis, and overall impact of the conclusion.

Comment: Some statements remain descriptive and general in nature.

The specific contribution of the study relative to the existing literature is not sufficiently highlighted.

Response: Thank you for these constructive comments. We agree that some statements in the manuscript were overly descriptive and that the study’s unique contribution to the literature could be emphasized more clearly.

In response, we have revised the manuscript to reduce general and descriptive statements and to present findings in a more analytical and interpretive manner. We have also explicitly highlighted the study’s contribution relative to existing literature, emphasizing that it provides one of the first context-specific, qualitative examinations of system-level determinants: policy, regulatory, financing, governance, and infrastructure affecting AI-powered ECG adoption in a low-resource, LMIC setting. These revisions clarify how the study advances understanding beyond technical considerations and provides actionable evidence for policymakers, regulators, and implementers.

Comment: The section could benefit from a more concise formulation.

• It would also be appropriate to more clearly emphasize the practical implications and potential directions for future research derived from the results.

 

Response: Thank you for this constructive feedback. We agree that the section could be more concise and should more clearly highlight practical implications and future research directions.

In response, we have revised the section to streamline the text, removing redundancies and presenting the key points in a more focused manner. We have also emphasized the practical implications of the findings for policymakers, regulators, and implementers, including actionable strategies for policy development, regulatory oversight, sustainable financing, data governance, and workforce capacity building. Additionally, we have outlined potential directions for future research, such as examining clinical efficacy, evaluating real-world impact, and exploring scalable implementation strategies. These revisions enhance the clarity, conciseness, and applied relevance of the section.

 

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

The authors have adequately addressed the issues. Well done!