A qualitative descriptive study design was used to explore the content of SMS text messages exchanged between health care providers and patients or caregivers within the HaH program. This design is appropriate as it contextually facilitates the grounded analysis of real-world SMS text message communications, offering insights into the content, tone, and relational dynamics that shape communication practices in real-world HaH settings [20]. The SRQR (Standards for Reporting Qualitative Research) checklist guided the reporting of this study [21] (Checklist 1).

A convenience sampling approach was adopted, drawing on all available deidentified SMS text message communications exchanged between health care providers, patients, and caregivers enrolled in the HaH program during the COVID-19 pandemic. Eligibility criteria included the following: (1) enrollment in the HaH program during the study period, (2) the availability of WhatsApp message records between the health care team and the patient or caregiver, and (3) SMS text messages written in English. All SMS text messages meeting these criteria and exchanged during the HaH admission period were included in the analysis. In routine care, WhatsApp communication occurred using the patient’s or caregiver’s own device.

This study used retrospective data extracted between August and October 2022 from WhatsApp, a digital messaging platform used by health care providers and patients enrolled in the National University Health System (NUHS) program NUHS@Home. Established in 2020, NUHS@Home delivers hospital-level care at home for patients who would otherwise require inpatient hospitalization, primarily managing acute internal medicine conditions. During the study period, all patients included in this study were admitted as inpatient-level HaH cases and diagnosed with COVID-19, reflecting the patient population admitted at that time rather than a program-specific focus on COVID-19 care. In addition to SMS text messaging, patients received routine clinical monitoring through teleconsultations, remote monitoring, and home visits when clinically indicated. Patients who experienced clinical deterioration could be escalated for further evaluation or hospital admission according to established protocols. In this study, all included patients were managed entirely within the HaH setting and did not require escalation to inpatient hospitalization. As such, the dataset captures SMS text message communications during routine home-based care, and interactions related to acute deterioration may therefore be underrepresented.

WhatsApp was used by the clinical care team as part of routine clinical care, with patients agreeing to its use for service communication because of its high local adoption, support for group communication, and convenience. It was integrated into routine care as an asynchronous communication platform to facilitate both clinical and administrative communications between health care providers and patients or their caregivers. These uses included sharing clinical updates, monitoring symptoms, addressing patient queries, and coordinating logistical matters such as appointments and medication delivery.

The study team comprised an internal medicine clinician (SK), a registered nurse (JSLS), a doctoral-trained researcher (SS), and a public health researcher (SYL) with formal training and experience in qualitative methodology. The diversity of professional backgrounds offers a multidisciplinary lens, further enriching the interpretation and contextualization of communication dynamics within the HaH setting.

The study received ethical approval from the National Health Group Domain-Specific Review Board (reference 2021/00896), which serves as the ethics board of the participating institution. Specific research consent was waived by the ethics board because this study involved a retrospective analysis of deidentified routine care data. To ensure anonymity and confidentiality, data from the participants were deidentified and assigned coded pseudonyms.

Qualitative data were retrospectively extracted from WhatsApp comprising all SMS text messages exchanged between the NUHS@Home team and patients or caregivers between August 2022 and October 2022. Demographic characteristics of patients (eg, age, sex, and ethnicity) were obtained from the electronic medical records to describe the study population. Sample messages included conversations related to episodes of care or administrative coordination. As this was an exploratory retrospective qualitative study, no a priori sample size was specified. All eligible messages exchanged during the study period were included in the analysis. During iterative coding, category development was considered sufficiently comprehensive when no substantially new patterns or insights emerged from further review of the dataset [22].

Rigor was established using the guidelines of credibility, transferability, dependability, and confirmability developed by Lincoln and Guba [23]. Credibility was enhanced through investigator triangulation during data analysis, which involved interdisciplinary team members interpreting the data. This ensured that multiple perspectives were incorporated and findings were grounded in the evidence. Transferability was maintained by providing detailed contextual descriptions of the setting and participant characteristics, allowing readers to assess relevance to other care environments. Dependability was ensured through the comprehensive documentation of an audit trail, comprising raw SMS text message data, coding, and analytic processes, to provide transparency and consistency. Confirmability was enhanced through reflexive journaling and independent coding by 2 researchers, followed by discussion and consensus-building on coding differences. As emphasized by Teh and Lek [24], reflexivity is a vital component of qualitative research that helps to mitigate unconscious researcher bias. The reflexive journal enabled the research team to critically examine their assumptions and decision-making, thereby enhancing the overall trustworthiness of the study [25].

Content analysis was used to explore the SMS text messages exchanged between health care providers and patients or caregivers within the HaH program. This method was selected for its established applicability in analyzing communication and media-based data, making it well-suited to the study’s focus on clinical SMS text messaging [26]. Content analysis allows researchers to interpret contextual meaning through a systematic coding process, thereby uncovering patterns and themes within communication [27]. Two members of the research team (SS and JSLS), both trained in qualitative research methods, independently reviewed and coded the data. Discrepancies were addressed through collaborative discussions among the research team, ensuring that the findings were accurately represented. Microsoft Word was used to organize and manage the dataset during the coding process. The 3-step inductive approach proposed by Elo and Kyngäs [28] guided the analytical process. In the preparation phase, researchers immersed themselves in the dataset by reading the messages several times to gain familiarity and determine appropriate units of analysis. Message timestamps were used at this stage to establish the chronological sequence of exchanges and to reconstruct conversational threads between health care providers and patients or caregivers. This enabled the identification of message initiation and response patterns within each interaction. In the organizing phase, the data were segmented into meaning units and coded to capture key content. These codes were then grouped into categories and subcategories through a process of abstraction. Finally, in the reporting phase, these categories were refined to reflect the recurring patterns present in the data.

From August 2022 to October 2022, a total of 354 HaH admissions met the study inclusion criteria and were included in the analysis. All included patients were managed within the HaH settings without requiring inpatient hospitalization, and all patients used asynchronous communication via WhatsApp. The mean age of the participants was 57 (SD 17.8) years. Most study participants were female (199/354, 56.2%), and the group was composed predominantly of individuals with Chinese ethnic backgrounds (246/354, 69.5%; Table 1). The primary diagnosis was COVID-19 (N=354), and the mean length of stay in HaH was 8 (SD 4) days. A total of 1218 SMS text messages were examined and coded from the conversations between health care providers and patients or caregivers. The median number of WhatsApp messages exchanged between the care team and patients or caregivers per patient interaction was 8 (IQR 13.5). Among the coded messages, 61.1% (364/596) were provider-initiated and 38.9% (232/596) were patient- or caregiver-initiated, based on the initiation of conversational threads. The median response time for health care providers was 11 (IQR 21.5) minutes during office hours (8 AM-6 PM) and 3 (IQR 6.25) minutes after office hours (6 PM).

Three categories were identified from the analysis of the SMS text messages, reflecting the nature of interactions between the health care providers and patients or caregivers within the HaH program: (1) clinical checks and advice, (2) administrative and transport arrangements, and (3) quality of interpersonal dynamics. These categories were further supported by 13 subcategories (Textbox 1).

This study examined the content of SMS text messages exchanged between health care providers and patients or caregivers within a HaH program in Singapore, offering insights into clinical communication and the relational dynamics enabling remote care delivery. Although the findings were derived from a cohort composed entirely of patients with COVID-19 during August 2022 to October 2022 and comprised low-acuity cases with no escalation to inpatient care, they offer potentially transferable insights into asynchronous communication processes within HaH settings. However, their applicability to other diagnoses should be interpreted cautiously. To our knowledge, this is the first content analysis of asynchronous communications within a HaH service in Singapore.

Communication related to the clinical aspects of care emerged as a key category in text messages exchanged between patients, caregivers, and health care providers within the HaH program. These interactions were centered on clinical consultations, where patients and caregivers sought medical advice related to symptom management, medication clarification, and general health concerns. In parallel, health care providers used these messages to deliver clinical updates, offer guidance, and adjust care plans based on patient conditions. This closely aligns with studies conducted in the United States and China, suggesting that SMS text messaging–based communication between clinicians and patients frequently involved symptom discussions, disease management support, and medication concerns [29-31]. Additionally, frequent SMS text message exchanges related to the monitoring of clinical parameters, such as reminders to report vital sign readings initiated by both health care providers and patients or caregivers, were observed. These were often followed by health care providers’ interpretations of the submitted data, which subsequently guided further clinical decision-making. This communication pattern aligns with findings from a systematic review, which emphasized that remote vital signs monitoring is a foundational element of HaH programs, facilitating the early detection of clinical deterioration and timely intervention when necessary [32]. Collectively, these observations underscore the value of asynchronous communication as a critical enabler of clinical continuity in home-based care. Through timely and responsive digital communications, it bridges the gap between patients and health care providers, allowing for ongoing monitoring, real-time decision support, and patient engagement that aligns with the core principles of the HaH model. As such, future research may benefit from examining how the frequency and quality of digital clinical communication can influence patient outcomes, such as symptom resolution and overall satisfaction with care.

Administrative and transport arrangements were identified as another key category in this study, highlighting the importance of nonclinical operational components in enabling effective care delivery within the HaH model. Numerous SMS text messages reflected the coordination of logistical tasks and administrative and transport-related processes. These interactions included scheduling home visits, organizing the delivery of medications and medical consumables, and facilitating transitions between care settings, such as hospital discharge to home or readmission when clinically indicated. Moreover, administrative communications reflected efforts to clarify documentation requirements, streamline admission and discharge workflows, and address patients’ or caregivers’ queries related to billing. These findings are consistent with studies conducted in the United States and the Netherlands, suggesting that digital messaging platforms are frequently used to manage a wide range of clerical and logistical tasks, such as administrative coordination, logistical planning, and transport arrangements [33,34]. When effectively streamlined, these processes can contribute to a reduction in delays in care delivery and improved patient experiences within remote care settings [35]. Interestingly, qualitative studies conducted in Singapore have highlighted that logistical and operational hurdles are inevitable in the implementation of HaH programs, potentially disrupting care continuity and resulting in delays, increased health care provider workload, and reduced patient satisfaction [36,37]. Therefore, addressing these issues is essential to optimize the effectiveness and scalability of HaH services in Singapore. Furthermore, the establishment of clear and reliable communication pathways between health care teams and patients or caregivers was observed. These interactions included identifying preferred modes of communication, whether via SMS text messaging or telephone calls, and designating a consistent point of contact for the care team. Similarly, this is consistent with a study conducted in the United States, highlighting the importance of aligning communication channels with patient preferences to enhance engagement and satisfaction [38]. Thus, clarifying communication preferences and responsibilities not only fosters continuity of care but also supports timely responsiveness and reduces fragmentation in care delivery. All in all, these findings underscore the integral role of digital communication in supporting the administrative and logistical infrastructure of HaH models. Therefore, stakeholders may consider engaging in co-design processes with end users of HaH programs to develop a tailored digital application aimed at streamlining operational workflows, further alleviating the administrative burden on health care teams while enhancing the overall patient experience and satisfaction.

Another key category identified in this study was the quality of interpersonal dynamics reflected in the digital text message exchanges between the health care providers, patients, and caregivers. These interactions consisted of expressions of empathy and emotional reassurance from health care providers and words of appreciation and gratitude from patients and caregivers in response to the care they received. Notably, although these interactions reflected elements of emotional support, explicit discussions of mental health conditions were not observed. This may reflect the clinically oriented nature of routine HaH communication, the likelihood that sensitive psychosocial concerns were addressed through alternative channels (eg, through synchronous audio or video calls), or a reluctance to disclose mental health issues within an SMS text messaging–based clinical exchange [39,40]. Furthermore, respectful and courteous language was observed across both parties, reinforcing the role of digital communication in fostering trust and mutual respect. These findings are consistent with multisite studies conducted in Australia, Ireland, the United Kingdom, Spain, and Germany, which similarly observed empathetic communication, emotional support, and courteous exchanges within digital platforms [40,41]. Moreover, the presence of digital empathy within asynchronous SMS text messaging–based interactions has been shown to foster rapport, encourage participant engagement, and contribute to positive patient experiences in HaH models [42]. Such relational interactions can potentially help to humanize remote care delivery and are instrumental in cultivating therapeutic alliances, which may positively influence patient engagement, satisfaction, and perceived quality of care. Furthermore, with the contemporary workforce gravitating toward asynchronous communication, professional environments have widely integrated digital tools to facilitate collaboration and optimize team productivity [43]. Future research could explore how these relational dynamics shape the experience of patients, caregivers, and health care providers, which could inform the development of culturally sensitive communication frameworks and targeted training programs for health care providers that will support high-quality relational care in digitally mediated HaH settings.

Timeliness of communication and professional tone were identified as another key category in the SMS text message exchanges between health care providers, patients, and caregivers in this study. Several interactions revealed delayed responses to nonclinical inquiries, particularly those sent after standard office hours. This finding aligns with studies conducted in China and the United States, which also reported frequent response delays after office hours [29,44]. Notably, timely communication has been shown to significantly influence patient satisfaction, likely due to its role in reinforcing feelings of safety, emotional reassurance, and trust, which are essential qualities valued in HaH models of care [44,45]. In addition to timeliness, the tone of communication provided further insights into the relational dynamics between health care providers, patients, and caregivers. Our findings highlighted that a formal tone was often used in messages on clinical care or medical guidance, reflecting a focus on clarity and professionalism. In contrast, informal tones were more frequently observed in administrative or logistical messages, indicating a conversational style. These patterns are congruent with studies conducted in the United States, which highlighted that clinically significant messages were communicated in a more formal tone, while operational matters were addressed in a neutral or informal tone [46,47]. Together, these findings suggest that tone modulation according to message purpose may play an important role in balancing clinical professionalism with relational rapport in HaH care. Future research could explore how tone preferences vary across cultural and demographic groups and examine how tone influences perceived empathy, patient engagement, and satisfaction within digital health care communication.

Our findings suggest several recommendations to strengthen digital communication practices within HaH care. First, 2-way asynchronous communication among patients, caregivers, and providers appears to be an important enabler of HaH care, helping patients and caregivers feel connected and supported, while allowing the care team to convey information efficiently. Second, the use of an existing platform that patients and caregivers are highly familiar with is critical to optimize their responsiveness and experience. Third, guidelines for asynchronous communications for HaH teams will be helpful to optimize clinical effectiveness and patient engagement. This may include types of content suitable for asynchronous communication (eg, instructive or informative rather than discussive), language and tone (eg, professional yet friendly), target response times, and after-hour response expectations. Dedicated training may be necessary for health care providers on the effective use of asynchronous communication tools as their use expands. Understanding the nature of these communications also has implications on the development of future AI-assisted communication tools that can be trained to answer such queries.

Future research directions can explore differences in communication among older adults and how SMS text messaging–based relational elements, such as empathy and mutual respect, shape the experiences of patients and caregivers. Intervention-focused studies could build on these findings using theory-informed approaches. Further work could also examine how tone preferences vary across cultural or demographic groups. Our findings also suggest a need for digital platforms that are better tailored to the operational realities of HaH care to streamline administrative tasks and reduce the burden on health care teams. Future research could also explore how generative AI technologies, such as large language models and conversational agents, could augment SMS text messaging–based communication within HaH programs by supporting routine queries and summarizing message threads. Nevertheless, given the central role of empathy and relational dynamics identified in this study, careful consideration will be critical to ensure that such innovations support the human dimensions of care. These offer valuable opportunities to inform more patient-centered, culturally attuned, and operationally effective communication strategies that support high-quality HaH care.

This study has several limitations. First, the dataset primarily comprised English-language communications, which may not fully capture the experiences of non-English-speaking participants. This limits the inclusivity and cultural representativeness of the findings, particularly in a multilingual context like Singapore. Additionally, the sample did not adequately reflect the broader demographic trends of an aging population, who are key users of the HaH services. Hence, future studies may benefit from purposive sampling strategies that include a more diverse linguistic and age profile, such as older adults, to provide a more comprehensive understanding of digital communication in HaH care. Second, the retrospective nature of the SMS text message data limits insights into the broader context of communication exchanges, as asynchronous messages may not have fully captured the full scope of interactions, particularly those that occurred following or preceding a verbal or face-to-face encounter. As such, future studies may consider triangulating message data with complementary methods, such as semistructured interviews or ethnographic observations, to better contextualize communication dynamics. Third, due to the time period of sampling, all patients were admitted for COVID-19. However, although remote monitoring, home visits, and virtual consultations are also used in other HaH diagnostic groups, communication patterns may differ by condition severity, symptom profile, and care trajectory. Therefore, transferability to a broader HaH population should be interpreted with caution. Finally, although qualitative content analysis was used to derive insights, the potential for researcher bias remains inherent due to the interpretative process. To mitigate this, systematic coding procedures and regular team discussions were used within the research team.

This study offered valuable insights into the nuances of digital communication between health care providers, patients, and caregivers within a HaH program in Singapore, in the context of the COVID-19 pandemic. The findings underscore the critical interplay between operational efficiency, the quality of interpersonal relationships, and clinical effectiveness in shaping the overall success of digitally mediated care within HaH settings. These interconnected elements not only facilitate seamless care coordination but also foster patient trust, engagement, and overall satisfaction. These insights underscore the need for communication frameworks and training programs that are clinically robust, operationally efficient, and attuned to the relational dynamics essential for high-quality home-based care. As HaH models continue to expand, future research should integrate these communication principles into health care provider training, digital platform development, and implementation strategies to promote more person-centered, culturally responsive, and sustainable care delivery.