Digital Transformation and Sustainable Customer Value in Healthcare: Evidence from an AI-Based Diabetes Prognostic Service

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

1. The title of the paper is too long even though it does not accurately reflect important aspects of the paper such as the brand equity, digital quality, customer satisfaction … etc. Meanwhile, the is no component in the paper related to digital transformation or even sustainable customer value. Please kindly adjust the title to accurately reflect the paper content.
2. The abstract and title are inconsistent, for example the digital transformation and sustainable customer value should be highlighted in the abstract. Additionally, the abstract is too long and too detailed, and it does not indicate a clear contribution of the paper.
3. The introduction should provide details of the research gap and have a weak connection to the related work that must be strengthened.
4. The paper cites a large body of literature; however, it is not recent enough and the literature review has a shallow depth in several important aspects such as recent studies in digital service quality, recent quality equity models, post-pandemic digital transformation … etc.
5. The methodology provided clear description of the samples used and clear hypotheses, however, it needs several improvements such as:
   1. The sampling method should be detailed and proper justifications for the adopted sample size.
   2. Proper explanation of construct operationalization.
   3. The paper does not show if the measurement scale came from Aaker, Yoo & Donthu, or custom design.
   4. Several important tests should be provided such as Harman’s single-factor test, Marker variable approach … etc.
   5. Several important aspects about the control variables should be explained in this section such as the justifications behind selecting them, the number of control variables used, how they are measured, is there any standardization performed on them … etc.
   6. The detailed justification behind why PLS-SEM is chosen instead of CB-SEM, whether the model is formative or reflective, why including the moderating effect … etc.
6. The results section presents clear results, however, the comparative analysis in the results section is very weak and not sufficient. Actually, it relies in the comparison heavily on really old work. Additionally, it does not provide a comparison against recent work in AI-based healthcare adoption, digital service quality in healthcare, patient satisfaction in AI-based healthcare systems … etc. Moreover, detailed discussion of the contradiction found to mainstream theories should be further justified, elaborated, and discussed in details, for example, the environment uncertainty generally weaken loyalty, the digital quality often strengthen loyalty … etc. Additionally, theoretical interpretation of the findings based on modern approaches (e.g., Digital Experience Quality, Post-pandemic digital healthcare frameworks … etc.) should be provided.
7. The results section is mostly descriptive with limited analytical interpretation.
8. The paper provides several limitations in the conclusion (moderator model limitations, serial mediation model … etc.), however, there are other equally important limitations that should be discussed such as the generalizability of the findings, the limitation of collecting the data at certain point in time, where the some of the findings concerning the relationship between the satisfaction and loyalty might need longitudinal data, the evaluation of AI component such as the explainability, interaction between AI results and patients’ behavior … etc.
9. Several important factors are missing in the methodology (e.g., the method of selecting the participants, sampling procedure, randomization method used (if any), if self-selection bias is taken into consideration, if pilot testing is adopted or not … etc.) that makes the reproducibility of the methodology/result very limited.
10. More than 73% of the references are published before 2020, which can be considered a bit old. Meanwhile, around 55% of all references are dated before 2015, which can be considered old (except for the foundational ones). Please try to rely on more recent references.
11. Some references are non-peer reviewed such as [32].
12. Some references are not related to the context of the manuscript such as [15], [28], [43], [44], [57], [58], [79], [111], [113] … etc.
13. Some references are very weakly relevant and can be considered as out of context such as [102].
14. References [79] and [111] are the same reference.
15. Some references are not ranked/indexed such as [4], [5], [30], [32], [40], [49], [32] which can reduce the credibility of the manuscript.
16. The iThenticate similarity index shows a really large index (26%), please kindly consider reducing it to the minimum possible.

Comments on the Quality of English Language

1. “pc web” in Table A2 is repeated twice in many lines in the table, please kindly adjust.
2. “Availi,” in section 4.2.2, I think it should be “Availability”
3. “generally pleasant..” in Table A1, there is one extra period that should be removed.
4. Some sentences are too long and should be separated into multiple sentences for clear readability such as: “Third, digital quality did not exhibit a significant moderating effect on the satisfaction– loyalty …”, “After excluding invalid responses, 800 valid samples were analyzed ….” … etc.

Author Response

Comments 1: The title of the paper is too long even though it does not accurately reflect important aspects of the paper such as the brand equity, digital quality, customer satisfaction … etc. Meanwhile, the is no component in the paper related to digital transformation or even sustainable customer value. Please kindly adjust the title to accurately reflect the paper content.

Response 1: In response to the reviewer’s comments, we replaced it with “Digital Transformation and Sustainable Customer Value in Healthcare: Evidence from an AI-based Diabetes Prognostic Service”

 

Comments 2: The abstract and title are inconsistent, for example the digital transformation and sustainable customer value should be highlighted in the abstract. Additionally, the abstract is too long and too detailed, and it does not indicate a clear contribution of the paper.

Response 2: In response to the reviewer’s comments, we aligned the content and emphasized the study’s contributions, while shortening the abstract as follows:

“This study investigates how digital transformation in healthcare shapes sustainable customer value by analyzing the role of digital quality and its influence on satisfaction and loyalty within an AI-based diabetes prognostic service. Drawing on system, information, and service quality as core dimensions of digital quality, the study examines their direct effects on satisfaction and their contribution to loyalty formation relative to traditional service factors. Using survey data collected from over 1,000 users of a digital healthcare platform equipped with an AI-driven diabetes prognostic algorithm, 800 valid responses were analyzed through PLS-SEM in SmartPLS 4.0. The results show that both traditional service attributes and digital quality significantly enhance customer satisfaction, which in turn promotes loyalty. However, digital quality does not strengthen the satisfaction–loyalty linkage, indicating that its value lies in establishing baseline trust and usability rather than amplifying loyalty outcomes. Environmental uncertainty—captured as technological and market uncertainty—also positively affects loyalty. This study contributes to digital healthcare research by providing empirical evidence from an AI-based long-term prognostic service and clarifying that digital quality operates as a foundational hygiene factor essential for sustainable customer value, rather than as a competitive differentiator.”

The contribution-emphasizing sentence is as follows: “This study contributes to digital healthcare research by providing empirical evidence from an AI-based long-term prognostic service and clarifying that digital quality operates as a foundational hygiene factor essential for sustainable customer value…”

 

Comments 3: The introduction should provide details of the research gap and have a weak connection to the related work that must be strengthened.

Response 3: It has been revised. We clearly identified the limitations of prior studies and strengthened the linkage between previous research and the present study. In addition, the contributions of this study are now explicitly stated in the Introduction.

 

Comments 4: The paper cites a large body of literature; however, it is not recent enough and the literature review has a shallow depth in several important aspects such as recent studies in digital service quality, recent quality equity models, post-pandemic digital transformation … etc. =>

Response 4: We added and reviewed approximately 30 recent studies.”

 

Comments 5: The methodology provided clear description of the samples used and clear hypotheses, however, it needs several improvements such as:

1. The sampling method should be detailed and proper justifications for the adopted sample size.

Response 5-1: Our study focuses on users with specific characteristics, namely individuals with prior experience using a digital Oriental medicine platform. Therefore, purposive sampling, which intentionally selects respondents who meet predefined criteria aligned with the research objectives, is more appropriate than convenience sampling. A corresponding explanation has been added to Section 3.1: “A purposive sampling method was employed to recruit respondents with actual usage experience, as such experience was essential for ensuring that their evaluations were relevant and valid for the purposes of this study.”

In addition, the following statement was added to Section 3.4 :

The final sample size of 800 respondents is statistically adequate for PLS-SEM analyses. First, following the “10-times rule” proposed by Hair et al. (2017), the minimum required sample is ten times the largest number of structural paths directed toward a latent construct. In this model, the maximum number of incoming paths is six, resulting in a minimum threshold of 60 observations—well below the 800 responses collected. Second, recent simulation studies recommend between 200 and 500 cases for complex PLS-SEM models with multiple latent constructs, reflective indicators, and moderating effects. The present sample exceeds these recommendations, ensuring stable parameter estimates and high statistical power. Finally, an a priori statistical power analysis (effect size f² = 0.15, α = 0.05, power = 0.95) indicates that approximately 220 responses are sufficient to detect medium effects in SEM models. Therefore, a sample of 800 substantially improves detection power for both direct and moderating effects, reduces sampling error, and enhances the reliability of bootstrapping procedures. Together, these considerations confirm that the sample size is more than sufficient for robust estimation, hypothesis testing, and generalizable interpretation within the PLS-SEM framework.

 

1. Proper explanation of construct operationalization. =>

Response 5-2: The following statement was added to Section 3.3 :

All constructs used in this study were operationalized based on validated scales from prior research. Each construct was represented by three to five reflective indicators measured on a seven-point Likert scale (1 = strongly disagree, 7 = strongly agree). The selection and adaptation of each item followed established conceptual definitions as proposed in previous studies, ensuring both theoretical consistency and contextual relevance for AI-based Oriental medicine digital platforms.

1. The paper does not show if the measurement scale came from Aaker, Yoo & Donthu, or custom design. => Although the origins of the measurement scales are described in the explanations of each individual construct, we accepted the reviewer’s comment and explicitly stated the sources of the measurement scales at the beginning of the relevant section (Section 3.3).

Response 5-3: All measurement items used in this study were adapted from established and validated scales in prior research. The dimensions of brand equity (brand awareness, brand availability, brand image, customer orientation, physical environment quality) were adapted from Aaker [7], Keller [3], and Yoo and Donthu’s consumer-based brand equity scales [15]. Digital service quality (system, information, and service quality) was adapted from Parasuraman, Zeithaml, and Berry’s SERVQUAL/E-SERVQUAL framework [25], and subsequent digital quality studies. Customer satisfaction items were adapted from Oliver’s expectation–disconfirmation model [23] and Cronin & Taylor’s SERVPERF scale [26]. Brand loyalty items followed Oliver [10] and Morgan and Hunt [11]. Environmental uncertainty items were adapted from established scales developed by Lisboa et al. [84].

1. Several important tests should be provided such as Harman’s single-factor test, Marker variable approach … etc. =>

Response 5-4: This issue was not addressed earlier in the manuscript but was instead discussed in Section 4.2.2 (Common Method Bias Test: Assessment of Over- or Underestimation). Specifically, as the data were collected from a single source, Harman’s single-factor test was conducted following the procedure proposed by Podsakoff et al. [103] to examine whether common method bias (CMB) may have inflated or deflated the observed relationships. The results of this test are reported in that section. Additional methods were not included to avoid redundancy.

1. Several important aspects about the control variables should be explained in this section such as the justifications behind selecting them, the number of control variables used, how they are measured, is there any standardization performed on them … etc. =>

Response 5-5: we inserted a statement in Section 3.3.9 indicating that technological uncertainty and market uncertainty were adopted as control variables : This study included two control variables—technological uncertainty and market uncertainty—to account for external environmental factors that may independently influence customer loyalty. The selection of these control variables is theoretically grounded in prior research in strategic management and digital transformation, which demonstrates that environmental dynamism affects customer evaluations and behavioral intentions regardless of service quality or satisfaction [55, 56].

 

1. The detailed justification behind why PLS-SEM is chosen instead of CB-SEM, whether the model is formative or reflective, why including the moderating effect … etc.

Response 5-6: The rationale for using PLS is already stated in Section 3.4 (Research Tool), as follows : The Partial Least Squares (PLS) method underlying SmartPLS is particularly suitable for exploratory and prediction-oriented studies because it: (1) does not require a large sample size, (2) imposes minimal distributional assumptions, and (3) provides robust estimates even when the measurement quality of survey data is limited [86]. In contrast to covariance-based SEM (CB-SEM), which emphasizes model fit and theory confirmation, the PLS approach focuses on maximizing explained variance (R²) and predictive relevance (Q²), making it appropriate for developing or formative constructs such as digital quality.

• In addition, in response to the reviewer’s comment regarding whether the model is formative or reflective and the rationale for including the moderating effect, we inserted the following statement in Section 3.4 (Research Tool): All latent constructs in the study were modeled as reflective rather than formative constructs. This specification is theoretically grounded in the nature of the underlying phenomena: the indicators are manifestations of the construct and are expected to covary. For example, perceptions of digital system quality (stability, responsiveness, and ease of use) are reflective outcomes of an underlying evaluation of digital quality rather than independent dimensions forming the construct. Similarly, satisfaction and loyalty are conceptualized as affective and attitudinal responses, making reflective measurement appropriate.
• The moderating effect of digital quality on the satisfaction–loyalty relationship was included to test whether digital service experiences strengthen or weaken customers' translation of satisfaction into behavioral loyalty. This inclusion is theoretically grounded in digital service and relationship marketing literature, which suggests that high-quality digital interactions enhance trust, perceived value, and relational commitment, thereby amplifying the effect of satisfaction on loyalty. Empirically, moderation enables the model to determine whether digital quality acts merely as an antecedent or also as a boundary condition shaping relational outcomes.

 

Comment 6: The results section presents clear results, however, the comparative analysis in the results section is very weak and not sufficient. Actually, it relies in the comparison heavily on really old work. Additionally, it does not provide a comparison against recent work in AI-based healthcare adoption, digital service quality in healthcare, patient satisfaction in AI-based healthcare systems … etc. Moreover, detailed discussion of the contradiction found to mainstream theories should be further justified, elaborated, and discussed in details, for example, the environment uncertainty generally weaken loyalty, the digital quality often strengthen loyalty … etc. Additionally, theoretical interpretation of the findings based on modern approaches (e.g., Digital Experience Quality, Post-pandemic digital healthcare frameworks … etc.) should be provided. 

Response 6: The text was reconstructed with new sentences to fully reflect the theoretical perspective suggested by the reviewer.

 

Comment 7: The results section is mostly descriptive with limited analytical interpretation. Response 7: We have made our best efforts to revise the manuscript.

 

Comment 8: The paper provides several limitations in the conclusion (moderator model limitations, serial mediation model … etc.), however, there are other equally important limitations that should be discussed such as the generalizability of the findings, the limitation of collecting the data at certain point in time, where the some of the findings concerning the relationship between the satisfaction and loyalty might need longitudinal data, the evaluation of AI component such as the explainability, interaction between AI results and patients’ behavior … etc. =>

Response 8: In the Conclusions section, we inserted a comprehensive statement summarizing the limitations pointed out by the reviewer, as follows (Section 6).

Fourth, methodological improvements can be expected. First, future studies need to strengthen external validity by employing samples that include cross-national comparisons or multiple platforms. In addition, the present study is based on cross-sectional data collected at a single point in time. However, in digital healthcare environments, the relationship between satisfaction and loyalty changes over time. Follow-up studies using longitudinal designs or panel data would allow this relationship to be examined more precisely.

Finally, interpretation of the relationship between AI quality and user behavior remains insufficient. Future research should expand the model by incorporating AI-related attributes such as algorithmic transparency, explainability (XAI), perceived fairness, and error tolerance, and should analyze how AI diagnostic results influence user behavior. Recent studies point out that how patients interpret AI diagnostic outcomes, the extent to which they trust them, and how they actually act upon them are closely associated with satisfaction and loyalty.

 

Comment 9: Several important factors are missing in the methodology (e.g., the method of selecting the participants, sampling procedure, randomization method used (if any), if self-selection bias is taken into consideration, if pilot testing is adopted or not … etc.) that makes the reproducibility of the methodology/result very limited.

Response 9: Additional text was inserted in response to the detailed methodological comments.

• In response to the reviewer’s comment, we inserted the following sentence in Section 3.5 (Sample Design and Data Collection):

A non-probability sampling approach was employed, which is common in exploratory studies on digital healthcare platforms. Participation was voluntary and based on user availability; therefore, randomization was not applied. Survey responses were collected from users with diverse demographic and technological backgrounds, and participation was not restricted to technologically proficient users. Anonymity was ensured to mitigate potential self-selection bias.

 

Comment 10: More than 73% of the references are published before 2020, which can be considered a bit old. Meanwhile, around 55% of all references are dated before 2015, which can be considered old (except for the foundational ones). Please try to rely on more recent references.

Response 10: We added more than 20 studies published since 2023.

 

Comment 11: Some references are non-peer reviewed such as [32].

Response 11: We do not consider that all referenced materials must necessarily be limited to peer-reviewed journal articles. Books, industry reports, newspaper articles, and publicly available materials can also provide valuable contextual insights, particularly when they address emerging practices or practical issues. Although such sources are not strictly academic, they offer practically relevant and informative perspectives and were therefore used as supplementary references rather than as primary theoretical foundations.

 

Comment 12: Some references are not related to the context of the manuscript such as [15], [28], [43], [44], [57], [58], [79], [111], [113] … etc.

Response 12: After reviewing the references, inappropriate sources were removed or replaced.

 

Comment 13: Some references are very weakly relevant and can be considered as out of context such as [102].

Response 13: It has been revised as follows: “The non-response bias assessment was revised to more accurately reflect the methodological foundation of the early–late respondent comparison. Specifically, the analysis was rephrased to clarify that the independent-sample t-test was adopted as a commonly used empirical implementation inspired by the logic of Armstrong and Overton (1977), rather than as a statistical procedure directly proposed by the original study. In addition, the interpretation of the results was moderated to emphasize a limited risk of non-response bias rather than asserting its complete absence.”

 

Comment 14: References [79] and [111] are the same reference.

Response 14: It has been revised.

 

Comment 15: Some references are not ranked/indexed such as [4], [5], [30], [32], [40], [49], [32] which can reduce the credibility of the manuscript.

Response 15: All references were reorganized, and the manuscript was revised to fully reflect all of the reviewer’s comments.

 

Comment 16: The iThenticate similarity index shows a really large index (26%), please kindly consider reducing it to the minimum possible.

Response 16: Although most overlaps occurred in technical terms, the manuscript was nevertheless substantially revised

 

Comment 17: “pc web” in Table A2 is repeated twice in many lines in the table, please kindly adjust.

Response 17: It has been revised.

 

Comment 18: “Availi,” in section 4.2.2, I think it should be “Availability”

Response 18: It has been revised.

 

Comment 19: “generally pleasant..” in Table A1, there is one extra period that should be removed.

Response 19: ‘generally’ has been removed.  

 

Comment 20: Some sentences are too long and should be separated into multiple sentences for clear readability such as: “Third, digital quality did not exhibit a significant moderating effect on the satisfaction– loyalty …”, “After excluding invalid responses, 800 valid samples were analyzed ….” … etc.

Response 20: It has been revised.

Reviewer 2 Report

Comments and Suggestions for Authors

• Decide whether the focal object is (a) an AI-based digital platform, (b) Oriental medicine hospitals, or (c) a combined “hybrid service”; then revise the title, abstract, methods, and measurement items to use one coherent terminology throughout.​

• If the context is inherently hybrid, explain explicitly how the digital platform and offline Oriental medicine service are integrated, and what exactly respondents were instructed to evaluate (platform, hospital, or both).

• There is an internal inconsistency in the reported sample size: at one point the paper mentions 800 valid responses, while later sections refer to 526 valid observations in the bootstrapping description. This discrepancy must be resolved and the final N consistently reported across the paper.
• Temper the claim that digital quality is merely a hygiene factor; clearly separate what is empirically shown (indirect effect via satisfaction, small direct effects on loyalty, non-significant moderation) from broader theoretical generalisations.
• Shorten the theoretical background by merging overlapping descriptions of brand equity, service quality, and digital transformation; focus on the constructs actually used in the model and avoid long textbook-style digressions.

Author Response

Comment 1: Decide whether the focal object is (a) an AI-based digital platform, (b) Oriental medicine hospitals, or (c) a combined “hybrid service”; then revise the title, abstract, methods, and measurement items to use one coherent terminology throughout.​

Response 1: We have revised the title as follows:  

Digital Transformation and Sustainable Customer Value in Healthcare: Evidence from an AI-based Diabetes Prognostic Service

 

Comment 2: If the context is inherently hybrid, explain explicitly how the digital platform and offline Oriental medicine service are integrated, and what exactly respondents were instructed to evaluate (platform, hospital, or both).

Response 2: This survey was conducted among patients who had experienced digital platform services provided by Oriental medicine clinics. Accordingly, both Oriental medical services and digital platform services were simultaneously experienced by the patients.

•  
• Comment 3: There is an internal inconsistency in the reported sample size: at one point the paper mentions 800 valid responses, while later sections refer to 526 valid observations in the bootstrapping description. This discrepancy must be resolved and the final N consistently reported across the paper.
• Response 3: 800 is correct, then it has been revised.
•  
• Comment 4: Temper the claim that digital quality is merely a hygiene factor; clearly separate what is empirically shown (indirect effect via satisfaction, small direct effects on loyalty, non-significant moderation) from broader theoretical generalisations.
• Response 4: We tempered the claim that digital quality functions merely as a hygiene factor and replaced it with more academically grounded discussion of the study’s findings. In addition, we newly added more detailed discussion addressing the non-significant moderating effects. In addition, we comprehensively revised the theoretical and managerial implications derived from the empirical results, strengthening theory-driven discussions based on perspectives such as digital experience, the resource-based view, and the dynamic capabilities view.

 

• Comment 5: Shorten the theoretical background by merging overlapping descriptions of brand equity, service quality, and digital transformation; focus on the constructs actually used in the model and avoid long textbook-style digressions.

Response 5: The theoretical background section was revised to adopt a more academic and theory-driven discussion.

Reviewer 3 Report

Comments and Suggestions for Authors

The research model is comprehensive, integrating both traditional brand equity and digital quality constructs. Several claims require stronger empirical grounding, the manuscript currently lacks depth in several foundational areas, and the theoretical framework needs to be updated with recent literature on AI diagnostic systems, digital trust formation, and platform-based healthcare interactions.

1. The introduction presents digital transformation broadly but does not articulate a precise and technically grounded research gap. The connection between AI-based prognostic services and customer loyalty remains conceptual rather than evidence-based. Recent advancements in AI-enabled clinical diagnostics, digital twin models, multimodal disease prediction, and preventive analytics are not discussed, weakening the rationale for your model.
2. The introduction does not differentiate between traditional healthcare service quality and digital platform quality, which is fundamental to your hypotheses. As recommendation, authors may refer to 10.1016/j.talanta.2023.125052 , in which the work on DBAN-based diabetic kidney disease detection can strengthen your justification that AI systems substantially improve predictive accuracy and user trust, reinforcing your argument about the increasing criticality of digital quality. 10.3389/fpubh.2025.1635475 , Their findings on innovation networks can improve the discussion on how digital transformation ecosystems reshape sustainable service value and patient engagement.
3. The theory section should explicitly connect digital quality with AI model performance, robustness, explainability, and multi-device usability, right now the literature review in Theoretical Background and Hypotheses Section summarizes foundational models but lacks synthesis and fails to position your contribution relative to the most recent digital healthcare studies.
4. Several hypotheses appear descriptive rather than derived from observable contradictions or gaps in prior literature. High conceptual overlap between constructs (e.g., digital service quality vs customer satisfaction) suggests a need for stronger justification of discriminant boundaries. To minimize the research gap, you can refer 10.3389/fphys.2023.1233341 , Their outlier detection framework supports the argument that information quality is not only about accuracy but also about resilience to noisy clinical data, which meaningfully improves patient trust and satisfaction. 10.1109/TIM.2025.3562973 , The digital twin diagnostic model demonstrates how system quality—including reliability, latency, and seamless device integration which directly affects user perception and satisfaction. 10.1109/TIM.2025.3563000 , This study on multivariate time-series classification reinforces the methodological alignment for AI-enabled healthcare signals and validates the relevance of using predictive modeling frameworks.
5. Methodology section requires clearer mapping between AI-platform characteristics and questionnaire items. The absence of multi-group analysis or cross-validation of digital quality constructs limits the empirical reliability of your results. Digital quality indicators are not validated against any established AI-healthcare system performance framework (e.g., latency, model confidence, data flow reliability). Authors can refer to 10.47297/wspchrmWSP2040-800503.20251604 , as this paper provides an empirical survey-based methodology that can strengthen your justification for operationalizing digital transformation constructs using multi-item Likert scales.
6. The justification for using PLS-SEM over CB-SEM is generic. No empirical reasons such as non-normality, indicator distribution, or formative constructs are provided.
7. The questionnaire items do not clearly reflect actual interactions with an AI-based prognostic platform. Many items appear designed for traditional hospitals, not digital health ecosystems.
8. In result sections, the demographic distribution skews to older adults but the study does not analyze age segmentation although it affects digital healthcare use.
9. The discussion does not incorporate technical dimensions of AI-driven platforms, such as model reliability, digital workflow integration, or clinical follow-up processes. Limitations do not acknowledge the core methodological gaps: construct overlap, possible common method bias, and lack of real usage data. The conclusion should emphasize the need for future studies to integrate behavioral data, such as platform logs, diagnostic accuracy scores, or user interaction metrics.

Author Response

Comment 1: The introduction presents digital transformation broadly but does not articulate a precise and technically grounded research gap. The connection between AI-based prognostic services and customer loyalty remains conceptual rather than evidence-based. Recent advancements in AI-enabled clinical diagnostics, digital twin models, multimodal disease prediction, and preventive analytics are not discussed, weakening the rationale for your model.

Response 1: We added new sentences explicitly articulating the research gap and strengthened the empirical grounding of AI-based digital quality by referring to the prior studies suggested by the reviewer. In addition, recent technological advancements in preventive analytics—such as AI-enabled clinical diagnostics, digital twin models, and multimodal disease prediction—were newly introduced as follows: “Digital transformation in the era of the Fourth Industrial Revolution has reshaped how healthcare organizations create value, accelerating the integration of artificial intelligence (AI), data analytics, digital twin models, multimodal disease prediction, and digital platforms into clinical and preventive services [1,2].”

 

Comment 2: The introduction does not differentiate between traditional healthcare service quality and digital platform quality, which is fundamental to your hypotheses. As recommendation, authors may refer to 10.1016/j.talanta.2023.125052 , in which the work on DBAN-based diabetic kidney disease detection can strengthen your justification that AI systems substantially improve predictive accuracy and user trust, reinforcing your argument about the increasing criticality of digital quality. 10.3389/fpubh.2025.1635475 , Their findings on innovation networks can improve the discussion on how digital transformation ecosystems reshape sustainable service value and patient engagement. =>

Response 2: A sentence distinguishing traditional healthcare service quality from digital platform quality was inserted in the Introduction.

• In addition, the two digital service cases suggested by the reviewer were incorporated into the Introduction as follows: In contrast, DBAN-based diagnostic studies empirically demonstrate that AI systems enhance predictive performance and trust, while research on innovation networks clearly explains how digital transformation ecosystems reshape sustainable service value and structures of patient engagement.

Comment 3: The theory section should explicitly connect digital quality with AI model performance, robustness, explainability, and multi-device usability, right now the literature review in Theoretical Background and Hypotheses Section summarizes foundational models but lacks synthesis and fails to position your contribution relative to the most recent digital healthcare studies.

Response 3: The theoretical background section was revised to be more theory-driven.

 

Comment 4: Several hypotheses appear descriptive rather than derived from observable contradictions or gaps in prior literature. High conceptual overlap between constructs (e.g., digital service quality vs customer satisfaction) suggests a need for stronger justification of discriminant boundaries. To minimize the research gap, you can refer 10.3389/fphys.2023.1233341 , Their outlier detection framework supports the argument that information quality is not only about accuracy but also about resilience to noisy clinical data, which meaningfully improves patient trust and satisfaction. 10.1109/TIM.2025.3562973 , The digital twin diagnostic model demonstrates how system quality—including reliability, latency, and seamless device integration which directly affects user perception and satisfaction. 10.1109/TIM.2025.3563000 , This study on multivariate time-series classification reinforces the methodological alignment for AI-enabled healthcare signals and validates the relevance of using predictive modeling frameworks.

Response 4: We sincerely appreciate the reviewer’s insightful and technically rich suggestions. We agree that the cited studies on outlier detection, digital twin–based diagnostics, and multivariate time-series modeling provide important contributions to AI-enabled healthcare research. However, incorporating these methodological frameworks in depth would require a substantial shift in the scope and objectives of the present study, which is primarily focused on examining customer perceptions of digital quality, satisfaction, and loyalty rather than developing or validating diagnostic algorithms. Accordingly, while we acknowledge the relevance of these approaches, fully integrating them would effectively constitute a different line of research. We therefore plan to address these technically oriented issues in a separate, dedicated study that focuses on AI model design and validation in clinical analytics. We believe that maintaining a clear focus in the current manuscript is essential to preserve its theoretical coherence and contribution.

 

Comment 5: Methodology section requires clearer mapping between AI-platform characteristics and questionnaire items. The absence of multi-group analysis or cross-validation of digital quality constructs limits the empirical reliability of your results. Digital quality indicators are not validated against any established AI-healthcare system performance framework (e.g., latency, model confidence, data flow reliability). Authors can refer to 10.47297/wspchrmWSP2040-800503.20251604 , as this paper provides an empirical survey-based methodology that can strengthen your justification for operationalizing digital transformation constructs using multi-item Likert scales.

Response 5: We sincerely appreciate the reviewer’s thoughtful and methodologically insightful comments. We fully acknowledge that a more explicit mapping between AI-platform characteristics and questionnaire items, as well as additional validation procedures such as multi-group analysis or cross-validation, would further enhance the empirical robustness of research on AI-enabled healthcare platforms. However, the primary objective of the present study is to examine users’ perceived digital quality and its effects on satisfaction and loyalty, rather than to evaluate objective system-level performance or validate AI models. Addressing the suggested methodological extensions—such as validation against AI-healthcare system performance frameworks (e.g., latency, model confidence, and data flow reliability)—would require a different research design and additional data sources, which extend beyond the scope of the current study. Accordingly, we plan to address these important methodological issues as part of future research focusing specifically on the integration of system-level AI performance indicators with user perception–based measures. We believe that maintaining a clear focus in the present manuscript allows for a more coherent theoretical and empirical contribution.

 

Comment 6: The justification for using PLS-SEM over CB-SEM is generic. No empirical reasons such as non-normality, indicator distribution, or formative constructs are provided.

Response 6: We appreciate the reviewer’s comment regarding the justification for using PLS-SEM over CB-SEM. The present study is based on multi-item Likert-scale survey data, and preliminary data screening indicated potential deviations from normality for several indicators. Given these distributional characteristics, PLS-SEM—which is more robust to non-normal data—was considered more appropriate than CB-SEM. In addition, the primary objective of this study is prediction-oriented, aiming to explain and predict customer satisfaction and loyalty in an emerging AI-based digital healthcare context rather than to strictly confirm an established theory. Furthermore, considering the complexity of the structural model with multiple latent constructs and paths, as well as the moderate sample size, PLS-SEM was deemed a more suitable analytical approach for the present study.

The relevant sentence was also inserted into the main text, as follows: “3.4. Research Tool

To test the proposed research model, this study employed a variance-based structural equation modeling (SEM) approach using the SmartPLS 4.0 software package for several methodological reasons. First, the data were collected using multi-item Likert-scale measures, and preliminary data screening indicated potential deviations from normality, for which PLS-SEM is more robust than covariance-based SEM (CB-SEM). Second, this study adopts a prediction-oriented approach, aiming to explain and predict customer satisfaction and loyalty in an emerging AI-based digital healthcare context rather than to strictly confirm an established theory. Finally, considering the complexity of the research model with multiple latent constructs and structural paths, as well as the moderate sample size, PLS-SEM was deemed an appropriate analytical technique.

SEM allows researchers to simultaneously examine complex causal relationships among latent constructs and assess both the measurement and structural components of a model [103]. The Partial Least Squares (PLS) method underlying SmartPLS is particularly suitable for exploratory and prediction-oriented studies because it: (1) does not require a large sample size, (2) imposes minimal distributional assumptions, and (3) provides robust estimates even when the measurement quality of survey data is limited [104]. In contrast to covariance-based SEM (CB-SEM), which emphasizes model fit and theory confirmation, the PLS approach focuses on maximizing explained variance (R²) and predictive relevance (Q²), making it appropriate for developing or formative constructs such as digital quality.”

 

 

Comment 7: The questionnaire items do not clearly reflect actual interactions with an AI-based prognostic platform. Many items appear designed for traditional hospitals, not digital health ecosystems.

Response 7: In this study, digital quality was not limited solely to AI. Rather, we aimed to examine patients’ satisfaction and loyalty toward digital healthcare services provided through digital transformation in Oriental medicine institutions.

 

Comment 8: In result sections, the demographic distribution skews to older adults but the study does not analyze age segmentation although it affects digital healthcare use.

Response 8: The primary focus of this study was not on age-based differences but on overall patient satisfaction and loyalty. We acknowledge that examining age-related differences would require a separately designed survey, and therefore plan to address this issue as a topic for future research.

• The discussion does not incorporate technical dimensions of AI-driven platforms, such as model reliability, digital workflow integration, or clinical follow-up processes. Limitations do not acknowledge the core methodological gaps: construct overlap, possible common method bias, and lack of real usage data. The conclusion should emphasize the need for future studies to integrate behavioral data, such as platform logs, diagnostic accuracy scores, or user interaction metrics. => We appreciate the reviewer’s thoughtful comments regarding the Discussion and Conclusions sections. The present study is positioned as a firm-oriented research focusing on the strategic and managerial implications of digital transformation in healthcare service organizations. Accordingly, the Discussion section emphasizes theoretical and managerial implications primarily from an organizational and managerial perspective, rather than from a technical or system-engineering viewpoint.
• Regarding the limitations, we would like to clarify that common method bias (CMB), which was noted by the reviewer, is explicitly addressed in the main text through a dedicated analysis section. In addition, the manuscript discusses other meaningful and contextually relevant limitations that align with the research scope. While we acknowledge that additional limitations could be identified, we intentionally exercised restraint to avoid overburdening the Conclusions section with an excessive number of limitations, which could detract from the clarity and coherence of the study’s contributions.

Nevertheless, in response to the reviewer’s suggestion, we have revised the Conclusions section to explicitly emphasize the need for future studies to integrate behavioral data—such as platform log data, diagnostic accuracy scores, and user interaction metrics—in order to complement perception-based survey data and further strengthen empirical insights into AI-enabled digital healthcare platforms as follows: “Fourth, methodological improvements can be expected. Above all, future studies need to strengthen external validity by employing samples that include cross-national comparisons or multiple platforms. In addition, the present study is based on cross-sectional data collected at a single point in time. However, in digital healthcare environments, the relationship between satisfaction and loyalty changes over time. Follow-up studies using longitudinal designs or panel data would allow this relationship to be examined more precisely.

Finally, interpretation of the relationship between AI quality and user behavior remains insufficient. Future research should expand the model by incorporating AI-related attributes such as algorithmic transparency, explainability (XAI), perceived fairness, and error tolerance, and should analyze how AI diagnostic results influence user behavior. Recent studies point out that how patients interpret AI diagnostic outcomes, the extent to which they trust them, and how they actually act upon them are closely associated with satisfaction and loyalty.”

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

1. The study relies heavily on old / really old references (around 70% of the references are pre-2020 and around 55% of the references are pre-2015), which is point 10 in the previous review report. It is really important to use alternative recent references when possible instead of these old ones.
2. The response to point 4 and the corresponding updates in the manuscript in the previous review report are totally inadequate.
3. The response to comment 6 in the previous review report is totally inadequate, given that this specific point is instrumental in enhancing the manuscript.
4. The response and corresponding updates in the manuscript for point 8 in the previous review report is inadequate, it actually ignored the major portion of the comment (e.g., the evaluation of AI component such as the explainability, interaction between AI results and patients’ behavior … etc.).
5. Still there are references not related to the study such as [5] and [43], or very weakly related (better remove them or use alternative related ones) such as [6], [9], [49], [51], [102] … etc. Actually, this point is indicated in point 12 in the previous review report.
6. References [99] and [134] are the same exact reference. Also, this point is highlighted in point 14 in the previous review report.
7. There are many references in the updated version of the manuscript that are not indexed such as [4], [37], [41], [49], [51] … etc. This point is point 15 in the previous review report.
8. The iThenticate similarity index shows a really large index (26%), please kindly consider reducing it to the minimum possible.

Comments on the Quality of English Language

There are too many sentences in the manuscript that are really long and should be restructured such as "Taken together, the findings suggest ....", "By doing so, this study extends prior literature ....", "In the context of digital healthcare, AI-based ...", "In this study, we examined the strategic implications ..." ... etc. The paper really needs full proofreading.

Author Response

Comment 1: The study relies heavily on old / really old references (around 70% of the references are pre-2020 and around 55% of the references are pre-2015), which is point 10 in the previous review report. It is really important to use alternative recent references when possible instead of these old ones.

Response 1: By newly incorporating recent research trends on “customer experience,” the references were additionally revised and updated as follows:

36. Nguyen, M.; Waller, M.; Pandya, A.; Portnoy, J. (2020), A review of patient and provider satisfaction with telemedicine. Current Allergy and Asthma Reports, 2020, 20, 1-7.
37. Kim, S.M.; Byeon, G.W.; Jeong, S.M.; Kim, Y.M.; Lee, K.U. Satisfaction Survey of Telemedicine Services for the Patients With Dementia in the Vulnerable Area for Medical Service, Journal of Korean Neuropsychiatric Association, 2021, 60, 366-378.
38. Lemon, K.N., Verhoef, P.C. Understanding customer experience throughout the customer journey, Journal of Marketing, 2016, 80, 69-96.
39. Becker, L.; Jaakkola, E.(2020), Customer experience: fundamental premises and implications for research, Journal of the Academy of Marketing Science, 2020, 48, 630-648.
40. O'Connor, G.E.; Myrden, S.; Alkire, L.; Lee, K.; Köcher, S.; Kandampully, J.; Williams, J.D. Digital Health Experience: A Regulatory Focus Perspective, Journal of Interactive Marketing, 2021, 56, 121-136.
41. Verleye, K. The co-creation experience from the customer perspective: its measurement and determinants,” Journal of Service Management, 2015, 26, 321-342.

 

 

Comment 2: The response to point 4 and the corresponding updates in the manuscript in the previous review report are totally inadequate.

Response 2: In response to the reviewer’s comments, and out of respect for the reviewer’s opinion, sentences that incorporate recent research trends on “customer experience” and align them with the authors’ perspective were added to Section 2.1.2 (Service Quality and Digital Transformation) as follows:

(2.1.2. Service Quality and Digital Transformation)

Along with the necessity of telemedicine, which has been actively promoted since COVID-19, technological advancements in the digital healthcare field, which have been accelerating in pace, have led to studies on telemedicine satisfaction [36,37], while at the same time stimulating research that examines the significance of customer experience for customer satisfaction and loyalty [38-41].

The “cognitive experience” distinguished within customer experience is measured, at the stage of operational definition, through experiences of knowledge acquisition, information acquisition, understanding of disease, and acquisition of medical information. In addition, the construction of “social experience” is composed of operational definitions that capture experiences such as a sense of intimacy with the platform, connectedness, consideration shown by platform members, and friendliness. Finally, “emotional experience” is operationally defined through components that involve experiencing reduced anxiety, a sense of relief, and calmness.

Among these three components of customer experience, cognitive experience corresponds to “information quality,” which is emphasized by the researchers of this study as sufficient, diverse, up-to-date, accurate, and useful information, while “social experience” corresponds to “relationship quality,” which is one of the components of traditional service quality.

Meanwhile, the “emotional experience” discussed within customer experience constitutes another antecedent of customer satisfaction that is distinct from traditional service quality, but it is also linked to “service quality,” which emphasizes whether the content provided by the platform proposed in this study is personalized.

Perspectives that emphasize customer experience constructs [39-41] regard them as mediating elements that connect traditional service quality constructs with customer satisfaction and loyalty. Considering that the “social experience” addressed in customer experience corresponds to traditional relationship quality, it appears that fully separating the digital quality construct from traditional service quality constructs requires further empirical examination. In the same context, this study seeks to examine the structural antecedent–consequence relationships among service quality, satisfaction, and loyalty by extending digital quality arising from digital transformation as one of the service quality constructs.

 

In addition, new references were added as follows:

36. Nguyen, M.; Waller, M.; Pandya, A.; Portnoy, J. (2020), A review of patient and provider satisfaction with telemedicine. Current Allergy and Asthma Reports, 2020, 20, 1-7.
37. Kim, S.M.; Byeon, G.W.; Jeong, S.M.; Kim, Y.M.; Lee, K.U. Satisfaction Survey of Telemedicine Services for the Patients With Dementia in the Vulnerable Area for Medical Service, Journal of Korean Neuropsychiatric Association, 2021, 60, 366-378.
38. Lemon, K.N., Verhoef, P.C. Understanding customer experience throughout the customer journey, Journal of Marketing, 2016, 80, 69-96.
39. Becker, L.; Jaakkola, E.(2020), Customer experience: fundamental premises and implications for research, Journal of the Academy of Marketing Science, 2020, 48, 630-648.
40. O'Connor, G.E.; Myrden, S.; Alkire, L.; Lee, K.; Köcher, S.; Kandampully, J.; Williams, J.D. Digital Health Experience: A Regulatory Focus Perspective, Journal of Interactive Marketing, 2021, 56, 121-136.
41. Verleye, K. The co-creation experience from the customer perspective: its measurement and determinants,” Journal of Service Management, 2015, 26, 321-342.

 

Comment 3: The response to comment 6 in the previous review report is totally inadequate, given that this specific point is instrumental in enhancing the manuscript.

Response 3: In accordance with the reviewer’s perspective, sentences emphasizing aspects that are consistent with or divergent from customer experience research were inserted into the conclusion section as follows:

Despite the research results and beneficial implications, the limitations of this study and future research directions are as follows. First, the empirical findings identified in this study are not substantially different from a series of findings reported in recent customer experience research that has attracted increasing attention [39–41]. Just as customers’ cognitive experiences have a positive effect on service satisfaction, digital quality resulting from digital transformation—particularly information quality—stimulates customer satisfaction and leads to customer loyalty. Among the elements of system quality and service quality provided by digital quality, some elements lead to satisfaction and loyalty through customers’ emotional experiences. Positive evaluations by customers derived from traditional relationship quality lead to satisfaction and loyalty through customers’ social experiences. Nevertheless, this study did not examine customers’ cognitive, social, and emotional experiences as mediating elements that connect the structural relationships among service quality, customer satisfaction, and loyalty. Therefore, it is necessary to more closely examine the structural relationships among these elements through extended research.

 

 

Comment 4: The response and corresponding updates in the manuscript for point 8 in the previous review report is inadequate, it actually ignored the major portion of the comment (e.g., the evaluation of AI component such as the explainability, interaction between AI results and patients’ behavior … etc.).

Response 4: Please see the Response 2 this time.

 

Comment 5: Still there are references not related to the study such as [5] and [43], or very weakly related (better remove them or use alternative related ones) such as [6], [9], [49], [51], [102] … etc. Actually, this point is indicated in point 12 in the previous review report.

Response 5: After reviewing the references, inappropriate sources were removed or replaced. However, I would like to express my sincere apologies for the fact that there were more inappropriate sources. Accordingly, we have once again conducted a careful review and sought to remove the inappropriate sources and replace them with alternative ones. Based on the results of this review, we would like to provide the following response.

1) In the case of Reference 5, it is a reference that another reviewer specifically suggested for use, and the authors reviewed it, summarized its key content, and newly incorporated it into the present manuscript.

5. Chen, X.; Chen, C.; Tian, X.; He, L.; Zuo, E.; Liu, P.; Xue, Y.; Yang, J.; Chen, C.; Lv, X. DBAN: An improved dual branch attention network combined with serum Raman spectroscopy for diagnosis of diabetic kidney disease. Talanta. 2024, 266, 125052.

2) In the case of Reference 43, the authors read the article and extracted and organized its content, and the reference is linked to the content presented in the main text.

43. Na, D.; Park, S. Fusion chain: A decentralized lightweight blockchain for IoT security and privacy. Electronics, 2021, 10, 391.

3) In the case of Reference 6, it is a reference that another reviewer specifically suggested for use, and the authors reviewed it, summarized its key content, and newly incorporated it into the present manuscript.

6. Hu, F.; Yang, H.; Qiu, L.; Wang, X.; Ren, Z.; Wei, S.; Zhou, H.; Chen, Y.; Hu, H. Innovation networks in the advanced medical equipment industry: supporting regional digital health systems from a local–national perspective. Frontiers in Public Health. 2025, 13, 1635475.

4) In the case of Reference 9, it is a classic reference that is linked to the following sentence.

Improvements in infrastructural quality tend to reach a diminishing return or ceiling effect, which is theoretically consistent with Herzberg’s motivation–hygiene theory [9]

9. Herzberg, F. The Motivation to Work; John Wiley & Sons: New York, NY, USA, 1959.
   5) In the case of Reference 49, it is one of the prior studies used by the authors to define the components of digital quality.

Digital quality is conceptualized as a multi-faceted framework encompassing the performance of the system, the reliability and usefulness of digital information, and the overall caliber of service delivered. The measurement indicators were adapted from prior research by Lee [49], Kwak, and Lee [88], Hwang and Lee [51], and Kim [89].

49. 49. Lee, J.W. A study on quality evaluation of digital libraries. Journal of the Korean Society for Library and Information Science, 2004, 38, 143-172.

 

 

6) In the case of Reference 51, it is one of the prior studies used by the authors to define the components of digital quality.

 

51. 51. Hwang, J.Y.; Lee, E.B. A review of studies on the service quality evaluation of digital libraries: on the basis of evaluation models and measures methodologies. Journal of Korean Library and Information Science Society. 2009, 40, 243-265.

 

7) In the case of Reference 102, it is also a prior study that was used in the study by Cho and Lim (2011) to conceptualize brand loyalty in the context of Oriental medicine services. The authors of the present study likewise used these two prior studies as references, as the present research also pertains to the field of Oriental medicine services.

84. Cho, S.; Lim, Y. Effects of hospital brand image on patient satisfaction and trust. Korean J. Health Serv. Manag. 2011, 5, 1–16.
    
    

Accordingly, brand loyalty is defined as a combination of favorable attitudes, advocacy behavior, and repeat usage intention toward an Oriental medicine service provider. Following the work of Sirgy and Samli [102] and Cho and Lim [84], this construct was measured using four items: (1) I have a very favorable (good) attitude towards this Platform; (2) I once told others that this Platform is good; (3) I have may other Platform to choose from, but I tend to use this Platform; and (4) I will continue to use this Platform (see Appendix A. Table A3).

 

102. Sirgy, M.J.; Samli, A.C. A path analytic model of store loyalty involving self-concept, store image, geographic loyalty, and socioeconomic status. J. Acad. Mark. Sci. 1985, 13, 265–291.
     
     

Comment 6: References [99] and [134] are the same exact reference. Also, this point is highlighted in point 14 in the previous review report.

Response 6: I am really sorry. We revised the references after you pointed out the duplicate citation in your previous comment, but we did not recognize that the reference you identified this time was also duplicated. As you suggested, we have corrected the reference format to appropriately consolidate the identical references as follows:

[p.30] Second, a serial mediation model can serve as a parsimonious alternative framework. In the present context, a sequential chain such as Digital Quality → (Higher-Order) Perceived Quality → Satisfaction → Trust/Attachment → Loyalty may offer a more theoretically appropriate structure. This implies that the mechanism is multi-stage mediation rather than moderation. Empirically, it is expected that (a) digital quality functions as an antecedent of satisfaction, and (b) trust and attachment (affective factors) operate as pure mediators in the satisfaction–loyalty linkage [99].

 

Comment 7: There are many references in the updated version of the manuscript that are not indexed such as [4], [37], [41], [49], [51] … etc. This point is point 15 in the previous review report.

Response 7: All of the references you pointed out are listed in both the in-text citations and the reference list as follows:

4. Park, H.-S. Determinants of patients satisfaction and intent to revisit oriental medical hospitals. Journal of the Korea Academia-Industrial Cooperation Society 2015, 16, 2726-2736. https://doi.org/10.5762/kais.2015.16.4.2726/
5. Kim, S.M.; Byeon, G.W.; Jeong, S.M.; Kim, Y.M.; Lee, K.U. Satisfaction Survey of Telemedicine Services for the Patients With Dementia in the Vulnerable Area for Medical Service, Journal of Korean Neuropsychiatric Association, 2021, 60, 366-378.
6. Yang, O.S.; Kim, C.G. The impact of corporate digital transformation on corporate performance: focusing on the mediating effects of functional, symbolic, experiential, emotional, and social values. Korean Management Consulting Review, 2023, 10, 113-133.
   49. Lee, J.W. A study on quality evaluation of digital libraries. Journal of the Korean Society for Library and Information Science, 2004, 38, 143-172.
7. 51. Hwang, J.Y.; Lee, E.B. A review of studies on the service quality evaluation of digital libraries: on the basis of evaluation models and measures methodologies. Journal of Korean Library and Information Science Society. 2009, 40, 243-265.

 

 

Comment 8: The iThenticate similarity index shows a really large index (26%), please kindly consider reducing it to the minimum possible.

Response 8: In the first round of revisions, we substantially reduced the level of duplication. When checked directly using iThenticate, the results—similar to those from the first revision—indicate that a considerable portion of the remaining similarity originates from the reference list. We consider this to be an issue that is difficult to fully resolve through revision alone. Nevertheless, in respect of the reviewer’s comments, we made our best efforts in the second round of revisions and further reduced the overall similarity index to 26%. A closer examination of the flagged similarities shows that they do not stem from substantive or critical sections of the manuscript, but rather from repeated keywords, administrative or formulaic expressions, and the reference list.

[Evidence from iThenticate]

 

Comment 9: There are too many sentences in the manuscript that are really long and should be restructured such as "Taken together, the findings suggest ....", "By doing so, this study extends prior literature ....", "In the context of digital healthcare, AI-based ...", "In this study, we examined the strategic implications ..." ... etc. The paper really needs full proofreading.

Response 9: Thank you for this helpful comment. We acknowledge that several sentences in the manuscript were overly long and could hinder readability. In response, we have carefully reviewed the entire manuscript and systematically restructured long and complex sentences—including those beginning with phrases such as “Taken together, the findings suggest…,” “By doing so, this study extends prior literature…,” “In the context of digital healthcare, AI-based…,” and “In this study, we examined the strategic implications…”—to improve clarity and conciseness. In addition, a full proofreading of the manuscript has been conducted to enhance overall readability, sentence flow, and grammatical accuracy.

(Please see attached file to confirm which sentences are revised on this comment):

page	before	after
18	Taken together, these results confirm that the measurement model demonstrates high internal reliability and adequate convergent validity, providing a robust foundation for subsequent structural model analysis.	Overall, these results confirm that the measurement model demonstrates high internal reliability and adequate convergent validity, providing a robust foundation for subsequent structural model analysis.
18	Taken together, all constructs in the proposed model demonstrated adequate discriminant validity, confirming that each latent variable captured a unique and theoretically coherent dimension within the model framework.	In summary, all constructs in the proposed model demonstrated adequate discriminant validity, confirming that each latent variable captured a unique and theoretically coherent dimension within the model framework.
27	Taken together, the results highlight the limitations of a unidirectional interpretation of uncertainty as a purely detrimental factor. In the context of AI-based digital healthcare services, uncertainty can paradoxically function as a catalyst that deepens user–platform relationships and fosters customer loyalty. This finding provides an important extension to existing theories by demonstrating that uncertainty may play a fundamentally different role in data-intensive, predictive service environments.	To sum up, the results highlight the limitations of a unidirectional interpretation of uncertainty as a purely detrimental factor. In the context of AI-based digital healthcare services, uncertainty can paradoxically function as a catalyst that deepens user–platform relationships and fosters customer loyalty. This finding provides an important extension to existing theories by demonstrating that uncertainty may play a fundamentally different role in data-intensive, predictive service environments.
29	Taken together, the findings suggest a coherent theoretical mechanism in which environmental uncertainty enhances the perceived value of strategic resources, amplifies users’ experience of dynamic capabilities, strengthens platform dependence, and ultimately fosters customer loyalty. By doing so, this study extends prior literature that has predominantly treated uncertainty as a negative external condition, demonstrating that in data- and AI-intensive industries such as digital healthcare, uncertainty can instead serve as a structural condition that reinforces relational stability and sustainable competitive advantage.	In summary, the findings suggest a coherent theoretical mechanism in which environmental uncertainty enhances the perceived value of strategic resources, amplifies users’ experience of dynamic capabilities, strengthens platform dependence, and ultimately fosters customer loyalty. By doing so, this study extends prior literature that has predominantly treated uncertainty as a negative external condition, demonstrating that in data- and AI-intensive industries such as digital healthcare, uncertainty can instead serve as a structural condition that reinforces relational stability and sustainable competitive advantage.
6	In this study, system quality denotes the degree to which a system is readily accessible, safeguarded by adequate security measures, and designed for effortless use.	This study adopts the definition of system quality as the degree to which a system is readily accessible, safeguarded by adequate security measures, and designed for effortless use.
8	In this study, digital quality reflects the comprehensive performance of digital services in terms of accessibility, protection of information, user-centered design, and timely responsiveness [40,71].	Digital quality reflects the comprehensive performance of digital services in terms of accessibility, protection of information, user-centered design, and timely responsiveness [40,71].
12	In this study, digital quality is conceptualized as a multi-faceted framework encompassing the performance of the system, the reliability and usefulness of digital information, and the overall caliber of service delivered.	Digital quality is conceptualized as a multi-faceted framework encompassing the performance of the system, the reliability and usefulness of digital information, and the overall caliber of service delivered.
10	All measurement items used in this study were adapted from established and validated scales in prior research. The dimensions of brand equity (brand awareness, brand availability, brand image, customer orientation, physical environment quality) were adapted from Aaker [7], Keller [3], and Yoo and Donthu’s consumer-based brand equity scales [15].	All measurement items used were adapted from established and validated scales in prior research. The dimensions of brand equity (brand awareness, brand availability, brand image, customer orientation, physical environment quality) were adapted from Aaker [7], Keller [3], and Yoo and Donthu’s consumer-based brand equity scales [15].
14	In this study, customer satisfaction is treated as a dual construct: one component reflects judgments tied to specific service encounters, while the other captures the overall satisfaction that emerges from repeated and ongoing interactions with the service [94].	Customer satisfaction is treated as a dual construct: one component reflects judgments tied to specific service encounters, while the other captures the overall satisfaction that emerges from repeated and ongoing interactions with the service [94].
14	Accordingly, brand loyalty in this study is defined as a combination of favorable attitudes, advocacy behavior, and repeat usage intention toward an Oriental medicine service provider.	Accordingly, brand loyalty is defined as a combination of favorable attitudes, advocacy behavior, and repeat usage intention toward an Oriental medicine service provider.
14	In this study, the measurement items for customer satisfaction were adapted from Oliver [23], Ragunathan and Irwin [95], and Lee and Ra [96].	The measurement items for customer satisfaction were adapted from Oliver [23], Ragunathan and Irwin [95], and Lee and Ra [96].
14	In this study, uncertainty is conceptualized as a two-dimensional construct comprising technological uncertainty and market uncertainty.	Uncertainty is conceptualized as a two-dimensional construct comprising technological uncertainty and market uncertainty.
15	In this study, a total of 800 valid responses were analyzed, exceeding the minimum threshold recommended by Hair et al. [105] for PLS-SEM (10 times the maximum number of paths directed toward any latent construct).	A total of 800 valid responses were analyzed, exceeding the minimum threshold recommended by Hair et al. [105] for PLS-SEM (10 times the maximum number of paths directed toward any latent construct).
17	In this study, the diagnostic analysis results showed that the maximum VIF was 3.766, the minimum was 1.002, and the mean was 2.260.	The diagnostic analysis results showed that the maximum VIF was 3.766, the minimum was 1.002, and the mean was 2.260.
18	In this study, factor analysis was performed on 11 latent variables and 58 observed indicators (see Appendix B, Table A5). A factor represents a set of variables exhibiting high intercorrelations, effectively reducing the dimensionality of the data into theoretically meaningful components.	Factor analysis was performed on 11 latent variables and 58 observed indicators (see Appendix B, Table A5). A factor represents a set of variables exhibiting high intercorrelations, effectively reducing the dimensionality of the data into theoretically meaningful components.
26	In contrast, the digital quality in this study—comprising information, system, and service quality—represents a lower-order functional quality that is predominantly utilitarian, focusing on attributes such as accuracy, stability, and responsiveness.	In contrast, the digital quality—comprising information, system, and service quality—represents a lower-order functional quality that is predominantly utilitarian, focusing on attributes such as accuracy, stability, and responsiveness.
29	In this study, we examined the strategic implications of brand equity—encompassing both the traditional components of perceived quality and the digital quality dimensions emphasized in digital transformation, namely system quality, information quality, and service quality—for customer satisfaction and customer loyalty, using cases from the digital healthcare industry.	This study examined the strategic implications of brand equity—encompassing both the traditional components of perceived quality and the digital quality dimensions emphasized in digital transformation, namely system quality, information quality, and service quality—for customer satisfaction and customer loyalty, using cases from the digital healthcare industry.
29	Third, the positive effect of environmental uncertainty on platform dependence and customer loyalty observed in this study can be more rigorously theorized through the lenses of the Resource-Based View (RBV) and the Dynamic Capabilities View (DCV).	Third, the positive effect of environmental uncertainty on platform dependence and customer loyalty observed can be more rigorously theorized through the lenses of the Resource-Based View (RBV) and the Dynamic Capabilities View (DCV).
30	By doing so, this study extends prior literature that has predominantly treated uncertainty as a negative external condition, demonstrating that in data- and AI-intensive industries such as digital healthcare, uncertainty can instead serve as a structural condition that reinforces relational stability and sustainable competitive advantage.	This study extends prior literature that has predominantly treated uncertainty as a negative external condition by showing that, in data- and AI-intensive industries such as digital healthcare, uncertainty may function as a structural condition that reinforces relational stability and sustainable competitive advantage.

 

 

 

Author Response File: Author Response.docx

Round 3

Reviewer 1 Report

Comments and Suggestions for Authors

All the comments highlighted in the previous review report have been adequately addressed in authors' response and the updated manuscript.