The feasibility study was an unblinded trial. Participants were prospectively recruited from EDs and inpatient units at 3 large urban hospital sites in British Columbia from November 2016 to July 2017. As this study was designed to assess the feasibility of the study protocol and refine procedures with a purposive sample gleaned from the feasibility study sites, no power calculation was conducted. Rather, the sample size was estimated from the recruitable number of patients with HF presenting at the ED sites based on administrative data with HF diagnosis and factoring in estimation of eligibility and attrition rates.

The inclusion criteria were as follows: one or more typical clinical HF symptoms; one or more typical clinical HF signs; one or more objective measures of HF, such as brain natriuretic peptide elevation or chest x-ray findings; currently receiving diuretic therapy; and age of 19 years or above. The exclusion criteria were as follows: inability to complete study procedures; no access to a reliable phone for communicating with a nurse; coronary or structural heart intervention during admission that would alter the course of HF convalescence with medical therapies alone; patients not expecting to present back to hospital for further deterioration, such as those wanting to die at home (see Multimedia Appendix 1 for the patient participant eligibility criteria of the feasibility).

Potential participants were first identified via referrals from hospital ED staff and screening hospital ED lists. Throughout recruitment, the study team expanded referral streams to include hospital in-patient wards to increase recruitment numbers. Eligible patients were contacted in the hospital when possible or immediately after discharge. Participants completed the consent process and were enrolled within 7 days from hospital discharge.

This study obtained ethics and associated approvals (no. H16-01076) from the regions where the 3 hospitals operate: University of British Columbia (UBC), Vancouver Coastal Health, and Interior Health Authorities.

The HHM device kits were provided by TELUS Health (technology partner). The kit included a touch screen tablet computer and Bluetooth-connected sensors that included a blood pressure cuff, a pulse oximeter, and a weighing scale. Upon enrollment and after patients returned home, monitoring nurses aimed to contact all participants within 1-3 days by telephone. Nurses trained to provide monitoring during the study period were seconded from cardiac clinics and had extensive experience managing patients with HF. During the first call, the nurse confirmed eligibility and scheduled the delivery of the HHM device kit. Kits were provided free of charge for 60 days and were delivered and set up in participants’ homes by a technician. The protocol sought to have kits delivered within 7 days of discharge from the hospital; however, this varied depending on the individual patient’s scheduling preference.

Participants were taught how to use the device kit by the technicians at the time of delivery and were observed submitting their first set of biometric measurements (ie, blood pressure, oxygen saturation, pulse, and weight) and answering 10-12 yes-or-no questions regarding symptoms (eg, “I feel more short of breath today”) in the technician’s presence. Participants were then asked to continue daily submissions of 1 set of this data per day for the next 60 days. The monitoring nurse sent an enrollment letter to the participant’s primary care provider (PCP) and other health care team members (eg, specialists), if appropriate, to notify them of their patients’ participation in the monitoring program and research study. PCPs or specialists were also asked to review and change the default monitoring limits of the patients (eg, blood pressure, heart rate limits, or oxygen saturation levels) to customize to patients’ needs if required. Dry target weight was also requested if known.

Monitoring data were reviewed by a single monitoring nurse Monday through Friday between 8:00 AM and 4:00 PM (excluding statutory holidays) via a web-based dashboard. The primary monitoring nurse performing the monitoring was a cardiac specialty nurse with over 20 years of experience working in cardiac care settings, such as a heart function clinic.

The platform listed and flagged patient measurements falling outside of predefined default (or customized) values (eg, weight gain of more than 5 lbs in 2 days), changes in symptoms, or missed data submissions. Alerts were flagged as red/severe (eg, systolic blood pressure < 85 mm Hg or >160 mm Hg) or yellow/caution (eg, systolic blood pressure 85-89 mm Hg or 141-160 mm Hg). The monitoring nurse followed up by telephone with the participants on all concerning flags and alerts. The monitoring nurse was authorized to advise the patient on better managing their condition and to provide education on management of the patient’s condition. For medication changes, the monitoring nurse would connect with the most responsible physician (eg, PCP) of the patient to facilitate these changes. In addition, the monitoring nurse would connect with other specialists with whom the patient was attached, if needed. For urgent situations, such as severe shortness of breath, the monitoring nurse was advised to direct the patient to their closest urgent care center or to call 911. Interventions that resulted from consultations between patients and monitoring nurses were documented in monitoring nurses’ notes, and summaries were shared via fax with the participants’ PCP every 2 weeks or as needed. The nurse also provided HF self-management education including advance care planning discussions to participants over the telephone, referring to a binder of HF self-management materials provided to participants upon enrollment into the study. Participants were discharged from the monitoring program after 60 days with a final report summarizing the monitoring data sent to the PCP. Participants were contacted again 30 days after discharge from the HHM program by telephone to complete a follow-up survey and to obtain their feedback on the HHM monitoring program.

Key metrics were drawn from the Triple Aim Plus framework [7], and both quantitative and qualitative data were collected from patient participants, participants’ PCPs, and the monitoring nurses. The primary outcome was the combination of unscheduled ED revisits in discharged participants within 90 days of discharge or all-cause death. Secondary outcomes included hospital readmission, patient participant health status (eg, quality of life, self-efficacy), end-user experience, and health system cost-effectiveness (including cost reduction and hospital bed capacity).

From October 2016 to June 2017, 519 patients who met the criterion of presenting to the hospital with shortness of breath were referred and screened for further HF eligibility screening: 219 met the study eligibility criteria for HF, and 70 were enrolled. Out of these 70, 47 (67%) participants completed the enrollment survey and thus provided demographic information. The median age of these participants was 75 years (range: 44-93 years), 24/47 (51%) participants identified as male, 36/47 (76%) identified as White (of European descent), and 45/47 (96%) identified English as their language of preference for health care matters.

The top 3 reasons for nonparticipation included patients declining, patients not meeting clinical eligibility criteria, and inability to provide informed consent (see Multimedia Appendix 6 for a summary of the most common reasons for patients not participating). Of the 121 patients who declined to participate, most did not provide a reason (see Multimedia Appendix 7 for a summary of reasons provided by eligible patients who declined to participate).

Of the 70 participants, 47 completed and returned the enrollment survey, 49 completed and returned the 90-day survey, and 35 completed and returned both the enrollment and 90-day survey. The subsamples providing data for each area of analysis are specified in the following section.

Participants were expected to be in the monitoring system for 60 days, submitting data once per day. Actual monitoring adherence (ie, the actual days data were entered) averaged 56 days across our sample, which is a 94% adherence rate based on the 60 days of baseline expectation.

Satisfaction with the monitoring platform as measured by the System Usability Scale [13] resulted in a mean score of 80.0 and a median of 81.4 on a 100-point scale (n=49).

Interviews were also conducted with the 2 monitoring nurses and covered satisfaction, patient–care provider interactions, procedures and usability, and perceived impact. Findings are summarized in Textbox 2. Monitoring nurses emphasized that they perceived positive impacts on participants quality of life and confidence:

PCPs were invited to participate in surveys, but we did not receive any responses from them.

The purpose and scope of this feasibility study were not designed to determine the efficacy of HHM. Nevertheless, we observed important trending of the data showing benefits across all 3 domains of the Triple Aim Plus framework. Most importantly, fewer patients had unscheduled ED revisits, fewer were readmitted to hospital, and the overall length of hospitalization decreased. In terms of quality of life, participants reported improved scores for both HF and general questionnaires. Furthermore, in economic terms, a cost consequence analysis showed that in all health care utilization factors measured, TEC4Home participants demonstrated decreased cost to the health system and decreased out-of-pocket costs. Finally, both participants and providers felt an improved experience in managing the patient’s HF.

The observation that our participants felt better while being monitored was similar to clinical case series and studies demonstrating high participant satisfaction when patients with HF were monitored at home [16]. Critically, our positive results differ from those of previous studies that suggested no benefit [3,4]. We hypothesize that TEC4Home was designed to support a patient in the postdischarge period for 60 days, which may be a reason for our promising results. Our study also measured participants’ quality of life and end-user experience, and both were found to increase. These findings have not been frequently reported in the literature.

This study, conducted in 3 urban sites in British Columbia, enrolled motivated patients. Of the 219 patients fulfilling eligibility criteria, only 70 patients enrolled, resulting in a recruitment rate of 32.0%. It would, therefore, be important to increase identification, recruitment, and enrollment of eligible patients. One challenge experienced was attempting to recruit patients at the height of their exacerbation in the ED. Indeed, other studies have encountered similar challenges in recruiting from the ED with common difficulties, such as time-consuming health record searches, limited research nurse support, and lack of face-to-face communication between patient and researcher [17]. The results we provide in Table 1 helped us to understand some of the factors that led to eligible patients declining participation, such as not feeling well enough to participate and not feeling comfortable using HHM or having someone to help them use it.

We presented these findings to our research study patient advisory committee and collectively developed the following ways to improve recruitment of eligible patients: increase engagement with the hospital staff members to raise awareness of TEC4Home to improve referral rates of potential patients, and identify optimal times to approach patients during patients’ hospital stays (eg, when they begin to feel better) or being careful to avoid recruiting while in the ED, especially when patients are being admitted to hospital.

The feasibility study also helped us to devise the following refinements to recruitment procedures: refining recruiter materials, including brochures, videos on how to use the HHM tool kit, and other materials to demystify study procedures and pique the participants’ interest; reviewing and clarifying the eligibility criteria with recruiters so they can conduct the screening and enrolment of patients more effectively; and establishing an ongoing support-and-feedback loop for site recruiters to promote consistent understanding and support for patient engagement.

Although a before–after analysis using administrative data is appropriate for a feasibility study, a more definitive trial design with a true control comparison is required. One site was unable to provide administrative data. Further, the discrepancy in responses and return rates between administrative data and survey data will be addressed in the RCT with an emphasis on provincial databases being the definitive source for our primary outcome.

Our results suggests that participants are generally able to perform self-monitoring routinely over 2 months postdischarge from the hospital. Our participants submitted data on nearly 95% of days, which compares favorably to similar studies. A recent telemonitoring feasibility study with patients with type 2 diabetes mellitus found lower adherence rates compared to our trial [18]. We further found the success rate of presurvey completion to be 67% (47/70), postsurvey completion to be 70% (49/70), and pre- and postsurvey completion by the same participants to be 50% (35/70). We identified ways, such as using scheduled reminders, to increase survey returns in our RCT.

Comments from patients in the survey revealed that those who persisted in using the HHM found the experience very helpful to support their self-management. Technology usability was another area that needed improvement—how to make the equipment more user friendly. Survey comments from monitoring clinicians were also highly useful to improving the types of data we needed to include in the dashboard, the workflow of monitoring nurses when contacting participants and their PCP, and how often the monitoring report summary should be sent to PCPs. We were unable to obtain feedback from participants’ PCPs in this feasibility study and needed to understand the reasons for it. We conducted a focus group with the PCPs to explore this issue prior to conducting the full RCT.

This study allowed us to examine our data collection procedures and scope. Overall, our evaluation framework was able to guide the collection of metrics to assess the outcomes that we would like to measure: patient outcome, end-user experiences, and health system cost-effectiveness. The data we collected helped us determine the primary outcome differences before and after monitoring, a basis for hospital utilization and cost comparisons, and end-user experiences based on the validated scales that we selected. This study also guided us in improving the scope of data we would be collecting, such as additional provincial databases covering clinical baseline measures to learn more about our participants at enrollment (eg, severity of illness, comorbidities), more health care utilization indicators (eg, specialized medical services), health care system and patient costs (eg, prescription drugs dispensed), and vital statistics that include accurate and more detailed mortality data. Additionally, the lack of data from a control comparison is a true limitation of any feasibility study design and will be addressed in the upcoming RCT. We also lengthened the period of data collection from just 90 days before and after enrollment to include 3 additional 3-month periods, up to 1 year before and after enrollment to provide insight into long-term effects. A shift from the VR-12 to the EuroQol-5D (EQ-5D) quality of life assessment tool will allow for the calculation of quality-adjusted life-year analysis. All of these refinements based on the feasibility results were implemented in the second phase of TEC4Home, a large-scale RCT designed to further examine home health monitoring across 22 British Columbia hospitals from urban, regional, and rural communities.

This feasibility study better prepared us for a planned multicenter RCT by helping us understand how best to engage patients in eligibility assessment, recruitment, and retention and how to refine our evaluation framework and metrics collection. Furthermore, analysis of the data we collected provided encouragement that HHM can be beneficial for patients with HF post discharge. Findings from this feasibility study provided practical lessons that allowed us to conduct the multicenter RCT as well as identify early positive signals of the benefits of HHM for patients with HF.