INTRODUCTION

Globally, Indonesia has a high prevalence of type 2 diabetes (T2DM). Epidemiological projections estimate that the number of people living with T2DM in Indonesia will rise from 19.5 million in 2021 to 28.6 million by 2045 [1]. Two primary goals of managing T2DM are to achieve glycemic control and prevent comorbidities [2,3], given that 72.0% to 83.1% of individuals with T2DM have not met glycemic targets [4]. Accordingly, regular screening for glycemic targets is crucial, as glycemic control is a significant predictor of diabetic complications, mortality, and quality of life [2,3].

Previous studies show that 12.3% to 35.0% of individuals with T2DM report feelings of loneliness [5,6]. In fact, an association between loneliness and glycemic control has been established [7]. Loneliness has been linked to elevated cortisol, interleukin-6, and interleukin-1 levels [8]. Cortisol plays a key mechanistic role in glucose homeostasis by facilitating gluconeogenesis and modulating glycogen metabolism [8,9]; accordingly, increases in cortisol levels correlate with higher plasma glucose and predict future glucose abnormalities [8]. Increases in interleukin-1 and interleukin-6, which are associated with loneliness, represent known risk factors for insulin resistance and T2DM onset [8,10].

The health belief model (HBM) has been used to examine the adoption of health-related applications and individual behavior in seeking online health information [11]. In the effort to develop a diabetes education chatbot, the HBM was used as a foundational framework to improve user engagement in self-management practices [11,12]. The HBM was deliberately integrated into the chatbot functionality. The chatbot addressed perceived susceptibility and severity through tailored risk education and narratives illustrating serious outcomes of inadequate glycemic control. Perceived benefits were addressed by content emphasizing positive outcomes (e.g., that regular blood glucose assessment helps the patient notice patterns early), while perceived barriers were mitigated through customized responses to users’ stated challenges, such as simple exercise and dietary tips [11,12].

The technology acceptance model (TAM) also guided the development of the content, interface, and interaction strategies. Additionally, this study expands on the TAM by incorporating human-centered elements and examining users’ intention to use the chatbot technology, as well as the technology’s ease of use and usefulness [13,14]. Regarding perceived usefulness, communications with the chatbot were evaluated, including tailored dietary suggestions and medication reminders. A conversational interface with prepared response options and information presented in compact, digestible portions reduced cognitive load and facilitated ease of use. To encourage intention to use, the chatbot delivered positive feedback after each interaction. By incorporating these theoretical frameworks from the outset, the chatbot was developed to facilitate behavioral change among individuals with T2DM.

A systematic review and meta-analysis highlighted that chatbot interventions for diabetes self-management have already reached the randomized controlled trial stage. However, the quantitative findings in that body of work offer limited novelty in their current form [15]. Therefore, it is necessary to employ mixed-methods research to capitalize on the strengths of both quantitative and qualitative methodologies within appropriate theoretical frameworks [15]. Moreover, only a few studies have examined chatbots that help reduce loneliness [16], lower hemoglobin A1c (HbA1c) [12], and improve health beliefs [11].

Thus, the objective of this study was to conduct an explanatory mixed-methods evaluation to examine feasibility (intention to use, ease of use, and usefulness) and the preliminary impact of a diabetes-education chatbot simulation on glycemic targets, loneliness, and health beliefs among adults with T2DM, based on the TAM and HBM theoretical approaches.

METHODS

Study Design

A two-stage explanatory sequential mixed-methods approach was implemented (Supplemental Material 1). The first stage, which involved a single group of participants, utilized pre-test and post-test pilot evaluations. In the second stage, focus group discussions and individual interviews were conducted with participants purposively selected to capture likelihood score variation—including those with relatively high and low scores for feasibility and glycemic control—to ensure representation of experiences and to gain insights into which components of the intervention users found more and less useful. Additionally, the qualitative component assessed whether users would be open to adjusting the intervention. It also sought to identify any potential technological difficulties encountered by first-stage participants during the intervention. The rationale for employing an explanatory sequential mixed-methods design was its capacity to provide a comprehensive understanding of the intervention’s impacts. Quantitative data can measure the degree of change, whereas qualitative research offers context, elaboration, and explanatory depth, especially when the quantitative results are unexpected.

This explanatory mixed-methods study conformed to the Consolidated Standards of Reporting Trials (CONSORT) statement [17] and the Good Reporting of a Mixed Methods Study (GRAMMS) guidelines [18]. The CONSORT diagram of participant recruitment and flow is presented in Supplemental Material 2.

Participants

The quantitative study recruited a convenience sample of 33 participants from 2 diabetes outpatient clinics in Kediri, East Java. This study aimed to recruit 33 participants to evaluate the feasibility and preliminary efficacy of the program, in accordance with prior recommendations for sample size (15–30 participants) in pilot studies and single-group pre–post studies [19,20].

Qualitative data collection involved 18 participants, 9 of whom had high scores on feasibility and glycemic control and 9 of whom had low scores. The selection of participants with the greatest degree of variation was intended to help identify significant patterns that were consistent across various ranges [21] of quantitative scores.

Moreover, the eligibility criteria included Indonesian nationals aged 21–65 years with T2DM who could communicate and write in Indonesian, had internet access, had a smartphone, and had or were willing to install the Telegram application. The study excluded individuals with type 1 diabetes, gestational diabetes, hearing deficits, conditions that posed a risk of harm due to physical limitations or limited physical exercise, or current psychological treatment. To evaluate the risk of harm, individuals were asked whether they were subject to any medical restrictions regarding physical exercise.

“TakonGendhis” Intervention Chatbot

A multidisciplinary team comprising a nurse with certification in diabetes care education, a pharmacist, a psychologist, and a health informatics expert developed an educational chatbot for adults with T2DM. The multidisciplinary team met weekly to develop chat content for “TakonGendhis.” Their initial work defined the scope and objectives of the pilot, including key performance indicators, topics, information regarding improvements, and messages to participants. Supplemental Material 3 illustrates how we built the TakonGendhis chatbot knowledge base based on users’ expressed responses.

Named “TakonGendhis,” which means “Ask Diabetes,” the platform was developed in Bahasa Indonesia with the TAM and HBM as foundational frameworks. This SaaS clinical decision-support web app used patient-generated health data and required no app installation or login. It was compatible with a wide range of hardware, including low-end mobile devices, workstations, and tablets. Chats supported a high-quality multimedia experience, including embedded images and video. Content covered (1) hyperglycemia and hypoglycemia management; (2) exercise, foot care, and diet (short videos and leaflets); (3) glycemic targets; (4) medication-taking behaviors; (5) reminders for medication and glycemic tests; and (6) motivational cards (Table 1).

The chatbot greeted users warmly on launch and asked them to enter their preferred name in the interface. A dashboard monitored enrollment data and participant input, while also flagging areas for improvement using indicators derived from participant responses. The user experience using Telegram on a mobile phone is shown in Supplemental Material 4.

Study Procedures

Pre-intervention and post-intervention surveys were administered face-to-face to eligible participants to collect participant characteristics and clinical data. All data were obtained at baseline and at the end of the study (12 weeks). Participants were instructed to answer questions on loneliness and perceived health beliefs during scheduled clinic visits. They were also required to install the Telegram app and fill out pre-intervention questionnaires. After using the chatbot, participants were scheduled for interviews and completed outcome questionnaires covering glycemic targets, loneliness, and perceived health beliefs. Finally, qualitative data on participants’ concerns and experiences using the TakonGendhis chatbot were collected.

Measurements

Socio-demographic questionnaire

Demographic characteristics included age, body mass index, diabetes duration, gender, marital status, family type, income, and education. The validity and reliability of the demographic questionnaire were documented in a previous investigation [22].

TAM questionnaire

The TAM questionnaire was adopted from a previous study and comprised 3 feasibility constructs: perceived usefulness, perceived ease of use, and intention to use [23]. Items were rated on a 7-point scale from 1 (strongly disagree) to 7 (strongly agree). In our study, internal consistency for the 3 constructs —perceived usefulness, perceived ease of use, and intention to use—was strong, with Cronbach alpha values of 0.94, 0.87, and 0.83, respectively.

Glycemic target measures

Individuals with T2DM who met the inclusion criteria were invited to undergo glycemic target measurement at health clinics after an 8-hour fast. Trained phlebotomists performed all venipuncture procedures. Fasting plasma glucose was considered the primary determinant of HbA1c, with increases in fasting plasma glucose correlated with increases in HbA1c [24,25]. For the present investigation, we treated fasting plasma glucose as a continuous variable and defined achievement of a controlled glycemic target as a fasting plasma glucose level of 80–130 mg/dL.

UCLA Loneliness Scale

Hughes et al. [26] developed a 3-item loneliness scale to quantify perceived social connectedness. In the present study, respondents rated how often they felt left out, isolated from others, or lacking companionship on a 3-point Likert scale, from 1 (“hardly ever”) to 3 (“often”). Scores were summed, with a higher total score indicating greater loneliness (range, 3–9). The Cronbach alpha found in this study was 0.92.

Health belief questionnaire

We adopted the health beliefs questionnaire for individuals with diabetes, which comprised 4 dimensions: perceived benefits, perceived severity, perceived susceptibility, and perceived barriers (3 items each). A 5-point Likert scale was used, with higher scores indicating more favorable health beliefs [27]. The Cronbach alpha in this study was 0.72, indicating acceptable internal consistency.

Statistical Analysis

Quantitative analysis

Data were analyzed using SPSS version 26.0 (IBM Corp., Armonk, NY, USA). Descriptive statistics were used to summarize demographic characteristics. For all intervention outcomes, statistical measures, such as mean, standard deviation, and mean difference (with 95% confidence interval), were computed. A paired-samples t-test and the Cohen d were employed to assess differences and effect sizes between pre-test and post-test scores. The p-values of less than 0.05 were considered to indicate statistical significance. Missing data due to loss to follow-up (n=2) were treated as missing, and analyses were conducted on an intention-to-treat basis.

Qualitative analysis

Qualitative data were analyzed using framework analysis, consisting of 5 phases [28]. The transcripts were independently coded by researchers. Codes were then collaboratively reviewed and organized into potential topic indices to map the data and develop key interpretations.

To improve trustworthiness, the qualitative research was integrated with confirmability checks and triangulation [29]. Triangulation helped explain the phenomenon by identifying consistencies and discrepancies across data types. In this study, triangulation employed multiple methods: qualitative (in-depth interviews and focus group discussions) and quantitative (survey responses) data were used to confirm themes. The research team recorded analytic decisions, codebooks, and meeting notes for auditability. Documentation from interviews and focus group discussions was used to increase analytic transparency and reproducibility.

The research team used Microsoft Excel (Microsoft Corp., Redmond, WA, USA) to organize data, track codes, and categorize themes. This program was chosen for its accessibility, flexibility, and ease of use during collaborative meetings, facilitating the comparison of coded segments, visualization of patterns, and establishment of a consensus on theme development [30]. Qualitative data analysis was conducted by a pair of male team members who were trained and experienced in qualitative research methods. A female team member with similar training and experience also participated in discussions and validation of the analysis. The results from the quantitative and qualitative phases were integrated through narrative synthesis [31].

Ethics Statement

The implementation protocol received approval from the Ethics Committee of Universitas Kadiri (reference: 027/24/VII/EC/KEP/UNIK/2024). The study was conducted in accordance with the guidelines outlined in the Declaration of Helsinki. All participants provided written informed consent prior to their involvement in the investigation and could withdraw from the study at any time with confirmation.

RESULTS

Participant Demographics

A total of 33 participants were included, 20 of whom were female and 13 male (Table 2). The largest proportions of participants had an education level of International Standard Classification of Education ≥3 (n=19, 57.6%); lived in a nuclear family (n=13, 39.4%); were married (n=18, 54.5%); and reported internet use ≥3 hr/day (n=23, 69.7%). The mean age was 47.36±6.25 years, mean body mass index was 25.44±3.94 kg/m², and mean diabetes duration was 4.42±1.77 years.

Quantitative Results

Of the 33 eligible participants with T2DM who were offered the chatbot, 31 (93.94%) completed the intervention. Based on intention-to-treat analysis, the impact of the chatbot education intervention was assessed by differences between pre-test and post-test scores (Table 3). Fasting plasma glucose and loneliness scores decreased significantly (t=25.17, effect size=0.95 and t=16.52, effect size=0.90, respectively). All 4 HBM subdomain scores also improved significantly after the intervention (all p<0.001): benefits (5.77 to 9.42; mean difference, −3.65; t=−13.00; effect size=0.85), severity (8.52 to 5.48; mean difference, 3.03; t=15.01; effect size=0.89), susceptibility (10.87 to 6.74; mean difference, 4.13; t=12.50; effect size=0.84), and barriers (10.55 to 7.00, mean difference, 3.55; t=13.56;effect size=0.86).

Moreover, participants rated the chatbot as feasible in TAM terms (usefulness, ease of use, and intention to use). The mean± standard deviation scores were 26.55±1.50 for usefulness, 27.32±0.79 for ease of use, and 34.03±1.74 for intention to use (Table 4).

Qualitative Results

In total, 18 participants from the intervention group completed individual interviews. Four themes were identified: feasibility, beliefs, emotional support, and areas for improvement (Table 5).

Narrative data integration

Integrating the quantitative and qualitative data provided a comprehensive understanding of the chatbot intervention’s effects among individuals with T2DM (Supplemental Material 5). Participants exhibited substantial decreases in fasting plasma glucose and feelings of loneliness, alongside improvements across all HBM subdomains. The chatbot received favorable feasibility ratings, reflecting high perceived usefulness, ease of use, and intention to use (TAM). The qualitative interviews corroborated these findings, with participants emphasizing the chatbot’s feasibility, strengthening of health beliefs, and emotional support, along with recommendations for content improvement. Collectively, these findings illustrate the measurable efficacy and perceived value of the chatbot intervention.

DISCUSSION

Using theoretical approaches derived from the TAM and HBM, this pilot study evaluated the feasibility and impact of a newly developed chatbot application on glycemic targets, loneliness, and perceived health beliefs among adults with T2DM (Supplemental Material 5). In this study, 31 participants completed the pilot intervention, representing a completion rate of 93.94%. This aligns with prior chatbot studies reporting completion rates of 86% [12], 90.22% [32], and 91% [33], supporting the acceptability of the chatbot format.

In prior clinical data, among participants who used a chatbot intervention, the average reduction in HbA1c was 1.04% [12]. A randomized controlled trial also reported reductions in estimated HbA1c in both groups after an app-based chat intervention compared with baseline; however, the between-group difference in change was not statistically significant [34]. Considerable research has shown that chatbots are feasible for supporting diet and physical activity [35]. Similarly, a pilot cohort study reported that an educational chatbot simulation significantly decreased HbA1c [36].

A previous study described chatbot technology as an advanced tool that can effectively address loneliness by integrating emotional, health, and social elements into an interactive, tailored experience [16]. With respect to loneliness, chatbots may help connect users with the information they need and reduce isolation by acting as a “smart technology” companion available around the clock [37]. In addition, higher levels of emotional support have been associated with a lower likelihood of experiencing loneliness [38]. This is consistent with our qualitative finding that a chatbot can provide emotional support for individuals with T2DM. Prior research has also highlighted the potential for chatbot-assisted emotional support in contexts such as depression or low mood, suggesting that chatbots may be particularly valuable for users who struggle with their emotions [39]. This may reflect the potential tendency of individuals with T2DM who feel lonely to seek additional emotional support.

Health beliefs are useful for understanding the mechanisms that increase intention to use a health chatbot [11]. Furthermore, perceived benefits represent key health beliefs that directly influence willingness to use such a tool [11]. Greater perceived benefits and the elimination of barriers have been associated with more preventive actions and a stronger intention to use a health education chatbot [27]. From a practical standpoint, our study supports the HBM as a valuable tool for understanding users’ willingness to use chatbots.

Chatbots demonstrate the capacity to facilitate effective behavior change by providing personalized, interactive, and scalable interventions [15]. They recognize user needs, demonstrate comprehension, and deliver prompt services tailored to user preferences, including goal setting and behavioral monitoring. Chatbots can implement interventions that encourage physical activity and medication adherence, thus promoting health behavior change [11,12]. Moreover, this study utilized the HBM and TAM frameworks, which focus on behavior-change interventions [11,27]. These frameworks help identify key elements of behavior change and guide the development of effective intervention strategies. Conventional diabetes education is typically delivered in person through structured sessions led by healthcare professionals; while helpful, these approaches can be constrained by accessibility, time, and resources. Conversely, the “TakonGendhis” chatbot offers a solution that allows patients to obtain individualized education and behavioral support at their convenience, without ongoing professional engagement.

The integration of HBM and TAM helps clarify how chatbots influence behavior. The chatbot’s perceived usefulness, ease of use, and supportive tone facilitated acceptance (TAM), which in turn heightened users’ perceived susceptibility and benefits and strengthened their ability to act, as articulated by the HBM. Consequently, users became more aware of their risk and more confident in their capacity for action. Within this model, individuals who perceive barriers are more likely to use the chatbot when they believe it provides beneficial health information.

With its participants recruited from 2 diabetes outpatient clinics, this pilot study used a relatively small convenience sample, underscoring the need for future studies with larger, randomly selected samples to increase generalizability. In addition, the study examined only short-term effects and included no follow-up. The chatbot’s technical requirements also limited participation to individuals with smartphones and an active internet connection. Future researchers could implement other health-focused chatbots, assess their feasibility and impact using TAM and HBM concepts, and validate them in randomized controlled trials.

Supplemental Materials

Notes

Data Availability

The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.

Conflict of Interest

The authors have no conflicts of interest associated with the material presented in this paper.

Funding

Kementerian Pendidikan Tinggi, Sains dan Teknologi, Indonesia, for the funding received (109/E5/PG.02.00.PL/2024).

Acknowledgements

The authors thank the Institut Ilmu Kesehatan Bhakti Wiyata Kediri, Taipei Medical University Universitas Dirgantara Marsekal Suryadarma, and Duta Bangsa University Surakarta for their support for this project.

Author Contributions

Conceptualization: Rias YA, Akasyah W, Mulyati TA, Astuti HW, Tsai HT. Data curation: Rias YA, Akasyah W, Mulyati TA, Astuti HW, Noronha H, Sani FN, Tsai HT. Formal analysis: Rias YA. Funding acquisition: Rias YA, Akasyah W, Mulyati TA. Methodology: Rias YA, Tsai HT. Project administration: Rias YA, Tsai HT. Visualization: Rias YA. Writing – original draft: Rias YA, Akasyah W, Mulyati TA, Astuti HW, Noronha H, Sani FN, Tsai HT. Writing – review & editing: Rias YA, Tsai HT.

REFERENCES

• 1. Sun H, Saeedi P, Karuranga S, Pinkepank M, Ogurtsova K, Duncan BB, et al. IDF Diabetes Atlas: global, regional and country-level diabetes prevalence estimates for 2021 and projections for 2045. Diabetes Res Clin Pract 2022;183: 109119. https://doi.org/10.1016/j.diabres.2021.109119ArticlePubMedPMC
• 2. Davies MJ, Aroda VR, Collins BS, Gabbay RA, Green J, Maruthur NM, et al. Management of hyperglycemia in type 2 diabetes, 2022. A consensus report by the American Diabetes Association (ADA) and the European Association for the Study of Diabetes (EASD). Diabetes Care 2022;45(11):2753-2786. https://doi.org/10.2337/dci22-0034ArticlePubMedPMC
• 3. ElSayed NA, Aleppo G, Aroda VR, Bannuru RR, Brown FM, Bruemmer D, et al. Summary of revisions: standards of care in diabetes-2023. Diabetes Care 2023;46(Suppl 1):S5-S9. https://doi.org/10.2337/dc23-SrevArticlePubMedPMC
• 4. Raccah D, Chou E, Colagiuri S, Gaàl Z, Lavalle F, Mkrtumyan A, et al. A global study of the unmet need for glycemic control and predictor factors among patients with type 2 diabetes mellitus who have achieved optimal fasting plasma glucose control on basal insulin. Diabetes Metab Res Rev 2017;33(3):e2858. https://doi.org/10.1002/dmrr.2858ArticlePubMedPMC
• 5. Hackett RA, Poole L, Hunt E, Panagi L, Steptoe A. Loneliness and biological responses to acute stress in people with Type 2 diabetes. Psychophysiology 2019;56(6):e13341. https://doi.org/10.1111/psyp.13341ArticlePubMedPMC
• 6. Kobos E, Szewczyk A, Kokoszka-Paszkot J, Dziedzic B. Factors associated with loneliness in patients with diabetes mellitus. Nurs Open 2020;8(1):517-524. https://doi.org/10.1002/nop2.655ArticlePubMedPMC
• 7. Huang YC, Cho E, Kuo HJ, García AA. The influences of depression and loneliness on A1C among middle-aged and older adults with diabetes. Psychol Health Med 2023;28(6):1540-1548. https://doi.org/10.1080/13548506.2022.2124287ArticlePubMed
• 8. Tran QA, Nakamura S, Watanabe K, Chei CL, Narimatsu H. The relationship between loneliness and blood glucose: a cross-sectional survey among Japanese. BMC Res Notes 2024;17(1):201. https://doi.org/10.1186/s13104-024-06855-zArticlePubMedPMC
• 9. Habegger KM. Cross talk between insulin and glucagon receptor signaling in the hepatocyte. Diabetes 2022;71(9):1842-1851. https://doi.org/10.2337/dbi22-0002ArticlePubMedPMC
• 10. Matthews T, Rasmussen LJH, Ambler A, Danese A, Eugen-Olsen J, Fancourt D, et al. Social isolation, loneliness, and inflammation: a multi-cohort investigation in early and mid-adulthood. Brain Behav Immun 2024;115: 727-736. https://doi.org/10.1016/j.bbi.2023.11.022ArticlePubMedPMC
• 11. Liu YL, Yan W, Hu B, Li Z, Lai YL. Effects of personalization and source expertise on users’ health beliefs and usage intention toward health chatbots: evidence from an online experiment. Digit Health 2022;8: 20552076221129718. https://doi.org/10.1177/20552076221129718ArticlePubMedPMC
• 12. Nassar CM, Dunlea R, Montero A, Tweedt A, Magee MF. Feasibility and preliminary behavioural and clinical efficacy of a diabetes education chatbot pilot among adults with type 2 diabetes. J Diabetes Sci Technol 2025;19(1):54-62. https://doi.org/10.1177/19322968231178020ArticlePubMedPMC
• 13. Akram S, Buono P, Lanzilotti R. Recruitment chatbot acceptance in a company: a mixed method study on human-centered technology acceptance model. Pers Ubiquit Comput 2024;28: 961-984. https://doi.org/10.1007/s00779-024-01826-4Article
• 14. Davis FD. Perceived usefulness, perceived ease of use, and user acceptance of information technology. Manag Inf Syst Q 1989;13(3):319-340. https://doi.org/10.2307/249008Article
• 15. Wu Y, Zhang J, Ge P, Duan T, Zhou J, Wu Y, et al. Application of chatbots to help patients self-manage diabetes: systematic review and meta-analysis. J Med Internet Res 2024;26: e60380. https://doi.org/10.2196/60380ArticlePubMedPMC
• 16. Rodríguez-Martínez A, Amezcua-Aguilar T, Cortés-Moreno J, Jiménez-Delgado JJ. Qualitative analysis of conversational chatbots to alleviate loneliness in older adults as a strategy for emotional health. Healthcare (Basel) 2023;12(1):62. https://doi.org/10.3390/healthcare12010062ArticlePubMedPMC
• 17. Hopewell S, Chan AW, Collins GS, Hróbjartsson A, Moher D, Schulz KF, et al. CONSORT 2025 statement: updated guideline for reporting randomized trials. JAMA 2025;333(22):1998-2005. https://doi.org/10.1001/jama.2025.4347ArticlePubMed
• 18. O’Cathain A, Murphy E, Nicholl J. The quality of mixed methods studies in health services research. J Health Serv Res Policy 2008;13(2):92-98. https://doi.org/10.1258/jhsrp.2007.007074Article
• 19. Kunselman AR. A brief overview of pilot studies and their sample size justification. Fertil Steril 2024;121(6):899-901. https://doi.org/10.1016/j.fertnstert.2024.01.040ArticlePubMedPMC
• 20. Li Y, Chung TY, Lu W, Li M, Ho YW, He M, et al. Chatbot-based mindfulness-based stress reduction program for university students with depressive symptoms: intervention development and pilot evaluation. J Am Psychiatr Nurses Assoc 2025;31(4):398-411. https://doi.org/10.1177/10783903241302092ArticlePubMed
• 21. Palinkas LA, Horwitz SM, Green CA, Wisdom JP, Duan N, Hoagwood K. Purposeful sampling for qualitative data collection and analysis in mixed method implementation research. Adm Policy Ment Health 2015;42(5):533-544. https://doi.org/10.1007/s10488-013-0528-yArticlePubMedPMC
• 22. Rias YA, Thato R, Tsai HT. Physical activity and its synergistic interaction with low platelet-lymphocyte ratio levels increasing the sleep quality in type 2 diabetes mellitus. Sci Diabetes Self Manag Care 2023;49(4):303-313. https://doi.org/10.1177/26350106231173675ArticlePubMed
• 23. Ayanwale MA, Ndlovu M. Investigating factors of students’ behavioral intentions to adopt chatbot technologies in higher education: perspective from expanded diffusion theory of innovation. Comput Hum Behav Rep 2024;14: 100396. https://doi.org/10.1016/j.chbr.2024.100396Article
• 24. Palella E, Cimino R, Pullano SA, Fiorillo AS, Gulletta E, Brunetti A, et al. Laboratory parameters of hemostasis, adhesion molecules, and inflammation in type 2 diabetes mellitus: correlation with glycemic control. Int J Environ Res Public Health 2020;17(1):300. https://doi.org/10.3390/ijerph17010300ArticlePubMedPMC
• 25. Abdullah A, Alkandari A, Longenecker JC, Devarajan S, Alkhatib A, Al-Wotayan R, et al. Glycemic control in Kuwaiti diabetes patients treated with glucose-lowering medication. Prim Care Diabetes 2020;14(4):311-316. https://doi.org/10.1016/j.pcd.2019.12.001ArticlePubMed
• 26. Hughes ME, Waite LJ, Hawkley LC, Cacioppo JT. A short scale for measuring loneliness in large surveys: results from two population-based studies. Res Aging 2004;26(6):655-672. https://doi.org/10.1177/0164027504268574ArticlePubMedPMC
• 27. Patil K, Kulkarni M. Can we trust Health and Wellness Chatbot going mobile? Empirical research using TAM and HBM. In: 2022 IEEE Region 10 Symposium (TENSYMP); 2022 Jul 1-3; Mumbai; p. 1-6. https://doi.org/10.1109/TENSYMP54529.2022.9864368
• 28. Gale NK, Heath G, Cameron E, Rashid S, Redwood S. Using the framework method for the analysis of qualitative data in multi-disciplinary health research. BMC Med Res Methodol 2013;13: 117. https://doi.org/10.1186/1471-2288-13-117ArticlePubMedPMC
• 29. Ahmed SK. The pillars of trustworthiness in qualitative research. J Med Surg Public Health 2024;2: 100051. https://doi.org/10.1016/j.glmedi.2024.100051Article
• 30. Ose SO. Using Excel and Word to structure qualitative data. J Appl Soc Sci 2016;10(2):147-162. https://doi.org/10.1177/1936724416664948Article
• 31. Fetters MD, Curry LA, Creswell JW. Achieving integration in mixed methods designs-principles and practices. Health Serv Res 2013;48(6 Pt 2):2134-2156. https://doi.org/10.1111/1475-6773.12117ArticlePubMedPMC
• 32. Munir M, Miraj F, Raza H, Hussain OA, Khan AA, Siddiqi DA, et al. Feasibility of an artificially intelligent vaccines chatbot in Pakistan: a mixed methods evaluation. Eur J Public Health 2021;31(Suppl_3):ckab164.507. https://doi.org/10.1093/eurpub/ckab164.507Article
• 33. Otero-Agra M, Jorge-Soto C, Cosido-Cobos ÓJ, Blanco-Prieto J, Alfaya-Fernández C, García-Ordóñez E, et al. Can a voice assistant help bystanders save lives? A feasibility pilot study chatbot in beta version to assist OHCA bystanders. Am J Emerg Med 2022;61: 169-174. https://doi.org/10.1016/j.ajem.2022.09.013ArticlePubMed
• 34. Gong E, Baptista S, Russell A, Scuffham P, Riddell M, Speight J, et al. My diabetes coach, a mobile app-based interactive conversational agent to support type 2 diabetes self-management: randomized effectiveness-implementation trial. J Med Internet Res 2020;22(11):e20322. https://doi.org/10.2196/20322ArticlePubMedPMC
• 35. Mash R, Schouw D, Fischer AE. Evaluating the implementation of the GREAT4Diabetes WhatsApp chatbot to educate people with type 2 diabetes during the COVID-19 pandemic: convergent mixed methods study. JMIR Diabetes 2022;7(2):e37882. https://doi.org/10.2196/37882ArticlePubMedPMC
• 36. Magee MF, Nassar CM, Dunlea R, James B, Montero AR, Sack P, et al. 330-OR: engagement, satisfaction, and preliminary impact of a personalized diabetes chatbot in a diverse cohort of adults with type 2 diabetes. Diabetes 2022;71(Suppl_1):330-OR. https://doi.org/10.2337/db22-330-ORArticle
• 37. Stefano Valtolina S, Hu L. Charlie: a chatbot to improve the elderly quality of life and to make them more active to fight their sense of loneliness. In: Proceedings of the 14th Biannual Conference of the Italian SIGCHI Chapter (CHItaly ‘21); 2021 Jul 11–13; Bolzano; Article No. 19. https://doi.org/10.1145/3464385.3464726
• 38. Bilecen B, Diekmann I, Faist T. Loneliness among Chinese international and local students in Germany: the role of student status, gender, and emotional support. Eur J High Educ 2024;14(3):470-488. https://doi.org/10.1080/21568235.2023.2215992Article
• 39. Chin H, Song H, Baek G, Shin M, Jung C, Cha M, et al. The potential of chatbots for emotional support and promoting mental well-being in different cultures: mixed methods study. J Med Internet Res 2023;25: e51712. https://doi.org/10.2196/51712ArticlePubMedPMC

Figure & Data

References

Citations

Citations to this article as recorded by  

• Peningkatan Keterampilan Wawancara Kontekstual Melalui Interaksi Siswa Dan Google Gemini Di SMP Negeri 1 Tlanakan
  Habiburrahman Habiburrahman, Rofi’ati Nur Diana Islam, Nadia Fitri, Halilatur Rohemah
  Indonesian Journal of Innovation Multidisipliner Research.2026; 4(2): 3651.     CrossRef