Study design

This study used qualitative description to explore patients’ experience in using SETAF to manage their AF condition. Qualitative description was preferred as it presents data from the participant’s point of view and avoids interpretation through the lens of the researcher [15]. This approach is known to be useful for collecting data to develop and refine an intervention [16]. The 21 item Standards for Reporting Qualitative Research (SRQR) was used to guide the reporting of this study (S2 File) [17].

Setting and sample

SETAF is an application integrated within the Philips Motiva system to complement future integrated care services for AF patients in Singapore through patient education, monitoring and self-management. Through SETAF, patients received motivational messages, health related surveys and self-management triage assessments. In addition, participants had access to educational videos and patient information sheets. The decision algorithm for the self-management triage and activity calendar for the SETAF application can be found as S1 Fig. The paired blood pressure and heart rate monitor automatically uploaded all measurements into the Motiva cloud database and enabled both patients and their care team to view the history and trends. The complete features of SETAF are described in Table 1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Components and features of SETAF.

https://doi.org/10.1371/journal.pone.0262033.t001

At the beginning of the study, each participant was provided a tablet device using Android software and blood pressure monitor from the Philips Motiva system to access the SETAF application. A brief tutorial was conducted by Philips to familiarize the patient with blood pressure measurement and the functions of the device and Motiva tool, which was delivered directly to the patient’s homes. The patients were allowed 6 weeks to explore and use SETAF on their own, based on instructions provided on how to use the device. After the end of the 6-week period, a time was arranged to collect back the tablet and device from the user, at which point an in-depth face to face interview was conducted to understand the user’s experience based on a semi-structured questionnaire.

Participants were recruited between February-March 2018 via purposive sampling from the outpatient clinic of a heart center of a hospital in Singapore for this study. English speaking patients with stable AF condition aged 21 years and above were eligible for the study. Stable AF was defined as no cardiovascular hospitalization and having no change to the patient’s medication or therapy in the past 3 months. Participants were approached at the outpatient clinic and asked if they would like to take part in the study, with the study explained to them. Consent was taken and a contact number was obtained from the patient to arrange for a suitable time for the blood pressure monitor and tablet to be delivered to the patient.

Data collection

After using SETAF for 6 weeks, a single session, face-to-face, individual, semi-structured in-depth interview was carried out in the participant’s homes by a trained research staff (AC), who was part of the research team. The research staff (AC) followed a flexible interview guide which encouraged participants to talk at length about their experiences and acceptability of the device and its features. All interviews were audio-recorded with the participant’s consent and each interview lasted on average 30 minutes (min. 15 and max. 45 minutes). In addition, the patients were asked to complete a short 4-item survey after the semi-structured interview to measure frequency of use, usefulness of device, desire to continue using the device and willingness to pay for the device. The full interview guide and 4-item survey are included as S1 and S3 Files.

Ethical considerations

The Institutional Review Board approved all aspects of the study (National Healthcare Group Domain Specific Review Board reference: 2017/00068). The study conforms with the principles outlined in the Declaration of Helsinki [18]. Written informed consent was obtained for all study participants.

Data analysis

All interviews were transcribed verbatim by a research staff. Analysis was done inductively by the authors (JN & BW), following a thematic analysis approach as described by Braun and Clarke [19]. This approach was preferred to allow identification of themes and subthemes linked to the dataset. The authors familiarized themselves with the data by listening to the audio recordings repeatedly and reading the interview transcripts (data immersion). Following that, interview transcripts were reviewed line by line to generate the initial codes across the entire data corpus (code generation). Constant comparative analysis of interview transcripts and codes helped to refine the emergent codes which were then grouped into categories to identify themes and subthemes (searching for themes). Discussions between the co-researchers (JN, BW & LTW) took place frequently to compare interpretations and further define and refine themes and subthemes until a consensus was reached. Nvivo 12 (QSR International. Burlington, MA.) software was used for qualitative analyses.

The results from the short 4-item survey were analyzed descriptively. Percentages (%) were used to summarize frequency of use, usefulness of device and desire to continue using the device. Willingness to pay for the device was presented as a mean value together with the minimum and maximum range.

Application usage data were presented as percentage of participants who watched the educational videos and read the motivational messages. The total number of vital signs monitoring logged for each patient was divided by the duration of study (42 days) to calculate average number of vital signs monitoring done in a day for each patient. The completion rate of the self-management triage assessments was categorized into less than or equal to 25%, between 26 to 50%, between 51 to 75% and between 76% to 100%.

The findings from both the qualitative and quantitative data were then organized and presented according to the TAM.

Rigour

Several measures were taken to adhere to the criteria of trustworthiness (credibility, transferability, dependability and confirmability) [20]. The interviewer established rapport with the participants and prolonged engagement during the interview to ensure credibility. Transcripts were cross-checked by the authors who did not take part in the transcription process. Two authors were actively involved in the coding process to allow for data triangulation and an audit trail was maintained during the data analysis. In addition, the collection of multiple sources of data allowed for data triangulation between qualitative and quantitative findings. Descriptions of the participant characteristics, context and verbatim quotes were provided in this paper to allow readers to exercise judgement on the transferability of findings.

Demographic profile

A total of 138 participants were approached to participate in this study. Of these, 37 were enrolled in the study while the remaining declined participation citing reasons such as time constraints, finding it too troublesome, not comfortable with using technological gadgets and not being able to speak or read English. Of the 37 participants, 19 were interviewed. 4 of the participants who were not interviewed declined to complete the 4-item survey, however, their application usage data were still included in the analysis. The participant characteristics are summarized in Table 2. Majority of them were male and of Chinese ethnicity. At the time of the study, they had been diagnosed with AF for 1–25 years.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Participant characteristics.

https://doi.org/10.1371/journal.pone.0262033.t002

The technology acceptance model

Fig 1 illustrates the qualitative and quantitative findings according to the basis of TAM.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Technology acceptance model for SETAF.

https://doi.org/10.1371/journal.pone.0262033.g001

Perceived usefulness

Perceived ease of use

Behavioral intention to use

Of the 33 participants who completed the survey, 94% (31/33) reported using SETAF almost every day (6–7 days a week), while the remaining 6% (2/33) used it frequently (4–5 days a week). Of participants, 21% (7/33) strongly agreed, 73% (24/32) agreed and 6% (2/33) were neutral that SETAF was useful in helping them manage their AF condition. When asked if they would continue using SETAF after the study, 21% (7/33) strongly agreed, 49% (16/33) agreed, 24% (8/33) were neutral and 6% (2/33) disagreed. The participants were willing to pay an average of $112.08 SGD (range $10–500) to purchase SETAF to continue using it.

Actual use behavior

Seven educational videos of varying topics were included in the application. The length of the videos ranged from approximately 2 to 25 minutes. The patient interactions with the videos are summarized in Fig 2. The most watched videos were titled ‘Atrial Fibrillation Explained’ (91.9%) and ‘Clot Formation and Stroke Risks’ (78.4%) while the least watched video was ‘Cook Right and Eat Smart for Healthy Ageing’ (56.8%).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Patient interactions with educational videos.

https://doi.org/10.1371/journal.pone.0262033.g002

Motivational messages were generated from the message bank and sent to the participants at regular intervals of approximately 2 messages per week during the study. The patient interactions with messages from the application are summarized in Fig 3. The percentage of patients who read the messages decreased over time from 94.6% for the first motivational message to 67.6% for the last motivational message.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Patient interactions with motivational messages.

https://doi.org/10.1371/journal.pone.0262033.g003

Table 3 summarizes the participants’ usage of SETAF’s blood pressure monitor and self-management triage assessment. On average, most of the participants used SETAF to measure their blood pressure at least once a day (75.7%). The remaining 24.3% of the participants used SETAF to measure blood pressure less than once a day during the study. The self-management triage assessment was pushed out to the participants every day during the study. However, the completion rates varied greatly between participants. 35.1% of the participants completed less than 25% of the self-management triage assessments and 45.9% of the participants completed more than 75% of the assessments.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Use of blood pressure monitor and self-management triage assessment.

https://doi.org/10.1371/journal.pone.0262033.t003

Principal findings

To our knowledge, this is the first qualitative study in Singapore to explore the acceptability and usability of an mHealth intervention for AF patients. Acceptance of SETAF relied on its potential to fill the gaps in current AF care while being easy and convenient for patients to use. In particular, they perceived the patient education and self-monitoring functions as useful for managing AF. They were also reassured by the link the app provided to their healthcare providers. This perceived usefulness is reflected in their high usage of SETAF, with all of the participants reporting using it at least 4–5 days a week Moreover, more than half of the participants would continue to use SETAF and were willing to pay for it. Interestingly, the SETAF application usage data showed that participants were more likely to interact with AF-related content, while interaction with the motivational messages declined over time. Our study also identified suggestions which could improve the acceptance of SETAF, including: better personalization of content, the need for psychosocial support, reducing language barriers and accessibility of SETAF. These findings support the acceptance of mHealth technology amongst the majority of our AF patients and their insights can inform the design of future mHealth interventions for them. For those patients who were unwilling to use SETAF in the future, an improved version of SETAF could potentially change their opinion.

Perceived usefulness: Improved patient education

Education is an essential component of care as patients’ knowledge is a known determinant of therapy uptake and adherence [21–23]. However, previous studies found that AF patients had poor knowledge of their condition and therapy [21–24]. Clinicians were cited as their main source of information and they supplemented this with independent internet searches, which had limitations such as potentially leading to inaccurate and misleading information [23]. Similarly, our study provided evidence that reliance on healthcare professionals and the internet for patient education continued to be suboptimal for our patients. Many participants in the present study felt they had limited opportunities to learn clinical relevant knowledge, which was imparted within the constraints of a clinic consultation which resulted in inadequate knowledge at baseline. However, they reported increased knowledge of AF after using the SETAF application, which in turn led to improved perceptions of the necessity of their treatment. Thus, mHealth has the potential to serve as a supplementary educational tool for delivering curated and validated content to increase health literacy.

Perceived ease of use: Language barriers

On the other hand, reading ability also plays a vital role in health literacy [11]. mHealth interventions should not only ensure scientifically validated content but also be tailored to the literacy levels of their users to maximize its benefits [11, 25]. This is further emphasized by the findings in our subtheme to reduce language barriers. Participants recalled encountering difficult terms which they had to search for. One participant even reported not attempting the surveys that were part of the education component of SETAF due to limited comprehension. AF patients in Singapore tend to be older with lower levels of formal education [13]. Given the multiethnic groups in Singapore, the lack of language options was also considered to be a shortcoming. These considerations should be taken into account when developing the educational content of mHealth interventions.

Perceived usefulness: Improved self-management

During the development phase of SETAF, vital signs monitoring and access to advice from healthcare professionals were identified as desirable features for the mHealth tool [13]. Hence, a paired blood pressure monitor was included in the intervention to allow patients to measure their heart rate and blood pressure. These measurements were then automatically uploaded into the cloud database and made accessible to both patients and their healthcare team. From the SETAF usage data, we found that most patients used the blood pressure monitoring function at least once a day. The study participants felt that this was a better way to keep track of their vital signs compared to traditional paper and pen methods and likely improved adherence to self-monitoring.

SETAF included a self-management triage assessment algorithm which would provide patients with actionable feedback based on the severity of their symptoms such as palpitations, chest pain and shortness of breath. The programmed decision algorithm may direct patients to call the hotline for medical advice for low-risk patients or to seek help immediately by calling an ambulance for emergent cases. We found only modest usage of the self-management triage assessment, and less than half of the participants completed 75% or more of the triage assessments. However, during the interview, most of the patients were aware of this feature and the hotline available to contact for medical advice. When prompted, the participants did not use the hotline as they felt well but found it to be useful and thought it could potentially reduce unnecessary hospital visits during symptomatic episodes. In addition, participants felt reassured that their care team would be notified of abnormal readings from the blood pressure and heart rate monitor. This sense of reassurance from mHealth monitoring was also observed in a qualitative study exploring patients’ experiences of using an mHealth application for management of Chronic Obstructive Pulmonary Disease [26]. Patients with AF are frequent users of the emergency department due to the sporadic and unpredictable nature of the disease [10]. Furthermore, traditional ambulatory care settings provide only a snapshot of the patient’s AF condition and the reliance on patient-reported symptoms makes it impossible to detect periods of silent AF which augment stroke risk and cardiomyopathy from untreated tachycardia [12]. Aljuaid et al. demonstrated the feasibility of smartphone electrocardiogram (ECG) monitoring to lower emergency room and clinic visits for post-ablation AF patients [27]. Similarly, the Mobile Atrial Fibrillation App (mAFA) trial which includes cardiac rhythm monitoring and two-way communication with doctors found reduced risk of adverse clinical events and rehospitalization [28–31]. Such technological advancements of mHealth show promise to transform the delivery and quality of AF care.

Perceived usefulness: Tailoring and customization features

In recent years, text message based mHealth interventions have gained popularity as a potential way to modify health behaviors. Pooled evidence consistently shows that text messaging interventions have a positive effect in health promotion [28, 30]. Interestingly, while many participants in the present study found the daily reminders from SETAF to be useful and helped to reinforce lifestyle changes, some thought that the messages became repetitive and lost effectiveness over time. In addition, the application usage data showed a decreasing trend in percentage of patients who read the motivational messages over time. These findings support the notion that targeted, tailored and personalized messages are more engaging and increases intervention efficacy [28]. Targeted messages cater the content to a population sub-group based on shared group characteristics, tailored messages are unique to an individual’s assessed characteristics and personalized messages included the participant’s name. In addition, messaging frequency was also found to moderate the intervention effectiveness [32]. Interventions that allowed users to customize the frequency and timing of the messages presumably led to the messages being delivered optimally to individual’s needs. The use of tailoring strategies is thought to enhance message processing and acceptance by the users and consequently improve the intervention efficacy [23, 33]. The optimal strategies for this intervention are unclear and will need further study.

Perceived usefulness; addressing psychosocial needs

Another key finding was the need for psychosocial support. Patient anxiety and depression is a well-known issue among AF patients and has recently been identified as an education need [34]. In the present study, participants discerned lack of knowledge as a contributing factor to their anxiety and suggested that the addition of emotional education and peer support could help to reduce it. This is congruent with current evidence that education has a small but positive effect on anxiety and depression for AF patients [8]. In addition, the European Society of Cardiology guideline has underscored the role of psychosocial management in integrated AF care [11]. Current mHealth interventions for AF patients are largely focused on patient education, medication adherence and self-monitoring [13]. To our knowledge, no mHealth interventions have addressed the psychosocial needs of AF patients. Future mHealth interventions for AF management should consider the addition of emotional education and other features of psychosocial support for a more holistic approach.

Strengths and limitations of study

A strength of our study was the conduct of in-depth interviews based on actual trial use of an mHealth tool. Other studies exploring the acceptance and use of mHealth tools so far are based on expressed intention to use and not actual usage. In addition, the qualitative data was supplemented by quantitative data from the survey and application usage data. This allowed for data triangulation between the different data sets and gave us a deeper insight into the participant’s usage behavior. Our sample is also ethnically diverse and includes patients with a wide range of years living with AF. This reflects the different needs and preferences of newly diagnosed and chronic AF patients. Of note, the mHealth tool used in this study was designed based on the findings of a previous study by Cher et al. [13]. Tailoring SETAF to the needs and preferences of patients in the local context which were identified in the previous study may have helped to achieve the perceived usefulness, ease of use and acceptance by the patients in this study.

However, the findings of this study should be considered in the context of a few limitations. Firstly, due to the limitations of the mHealth tool, which was only available in English, we were not able to capture the perspectives of other non-English speaking AF patients who may have specific preferences and mobile health technology use behaviors. Secondly, improvements in AF knowledge and self-management were self-reported during the interviews and no objective measurements were taken to confirm the participants claims. Nonetheless, the purpose of this study was not to assess the effectiveness of the mHealth tool but to explore its acceptability and usefulness for AF patients.