We aimed to test the acceptability, feasibility, and potential effects of the SAMSON solution on 30-50 patients with hematological cancer through a 2-armed, unblinded, 12-week, pragmatic pilot randomized controlled trial (RCT) [22]. After providing written consent (Multimedia Appendix 2) and completing baseline questionnaires (Multimedia Appendix 3), participants were randomized to either the intervention group (SAMSON solution) or the control group (usual care) at a 1:1 ratio. After 12 weeks, they completed end-of-study surveys (Multimedia Appendix 3).

Using a computer-generated randomization chart, a permuted block randomization of size 4 was used to ensure an even balance of patients in each group throughout the study period. The allocation schedule was generated by a statistician who was blinded to the participants to prevent any predictability when randomizing participants to intervention or control [23].

Patients were recruited from a metropolitan specialized cancer hospital in Melbourne, Australia, between August 2023 and February 2024. Eligible patients were adults (more than 18 years old), diagnosed with hematological cancer, scheduled to commence oral anticancer medicines (OAMs) or commenced the medication for less than 12 months, willing to have OAMs dispensed at the hospital for the duration of the trial, able to communicate in English, and had access to the internet, a smartphone or computer, and telehealth.

Patients were identified by study site nurses, pharmacists, or treating consultants, who were informed about the study and eligible criteria. At their scheduled consultation, they were asked if they would like their details passed on for contact by the research team. If the patient agreed, they were referred to the study RC. Then, the RC contacts the patient for screening and a comprehensive informed consent process, either in person or online.

Demographic was completed at baseline (t₀). Patients’ personal and clinical information was collected from the hospital’s electronic medical record system. The study’s primary outcomes included the patients’ acceptance of SAMSON, measured by the Unified Theory of Acceptance and Use of Technology (UTAUT) [24,25] at 12 weeks (t₁), and study feasibility, measured by predefined rates of recruitment, randomization, retention, intervention adherence, and outcome assessment completion [26,27]. Secondary outcomes included MA, self-reported adherence, toxicity self-management, anxiety, depression and symptoms, and quality of life. Outcomes were assessed at baseline (t₀) and at the end of week 12 (t₁), except MA measured in week 16. All survey data collection (Multimedia Appendix 3) was done through (Research Electronic Data Capture) REDCap (Vanderbilt University) [28].

MA was measured by medication refill adherence (MRA) [30] collected from pharmacy dispensing data. MRA was defined as a percentage calculated from the total days’ supply divided by the number of days of study participation and multiplied by 100. In this study, the patient was considered as optimal adherence if their MRA was≥90%.

Self-reported adherence (Adherence Starts with Knowledge 12 [ASK-12]) was measured [31]; this measure included 12 items in 3 subscales: adherence behavior, health beliefs, and inconvenience or forgetfulness.

Toxicity self-management was measured with the Patient Activation Measure-Short Form [32,33], a 13-item self-report measure assessing the patient’s knowledge, skills, and confidence in the self-management of their disease and related symptoms.

Anxiety, depression, and symptoms were measured by the Patient-Reported Outcomes Measurement Information System (PROMIS) [34] to assess depression, anxiety, pain interference, fatigue, sleep disturbance, and physical function.

Quality of Life was measured by Functional Assessment of Cancer Therapy-General (FACT-G) [35], which is a 27-item self-report scale measuring the quality of life of patients currently undergoing cancer treatment.

Descriptive statistics were used to summarize participant characteristics across study arms, and differences in baseline attributes were assessed using t tests or chi-square tests as appropriate.

SAMSON acceptability was analyzed thematically. The UTAUT aims to examine individual quality dimensions, which means it is a suite of scales rather than one quality measure; therefore, adding up the overall scale of the questionnaire is not suitable. Results of UTAUT surveys were summarized for each of the 5 dimensions with the percentage of patients endorsing Likert scale ratings of 3 (disagree, neither agree nor disagree, and agree). Free text answers to UTAUT questionnaires were narratively summarized to gain further insight into acceptability.

Feasibility was determined based on the traffic light approach. Recruitment feasibility was assessed by the number of patients recruited (consented) divided by the number of patients who were approached to join the study. Randomization feasibility was assessed by the number of patients who were randomized divided by the number of patients who consented. Retention in both arms of the study was assessed by the number of patients who remained at the end of the study divided by the total number of patients who consented to join the study. We also tracked intervention adherence, for example, the percentage of patients who completed tasks on the SAMSON smartphone app and received MI teleconsultations. Compliance data for survey completion was calculated as the percentage of patients who completed surveys at t₀ and t₁t₁ out of the total patients in the study at these time points.

Secondary outcomes were analyzed using linear regression. The dependent variable was the outcome at follow-up and the independent variables were arm assignment and the outcome at baseline. Standard checks for normality and homoscedasticity of residuals were conducted for each outcome. All analyses were conducted using StataNow 18 (StataCorp) [36].

A steering committee—which included chief investigators; experts in the fields of digital health, information technology, nursing, pharmacy, psychology, and oncology; and a patients’ representative—was formed to provide support for the study. This study was approved by the HRECs of PMCC (number HREC/95332/PMCC) and Swinburne University of Technology (number 20237273-15836) (Multimedia Appendices 5 and 6). Patients reviewed study details and indicated their consent using e-consent forms. Patients were encouraged to contact the study team if they had any questions or concerns. Patients’ personal and health information in the SAMSON app was encrypted in transit and stored in a secure server at Swinburne. Study team access to patient’s data on the SAMSON web-based dashboard was password protected and limited to the research coordinator (RC) and 4 study nurses and pharmacists. All data analyses were conducted on deidentified data. Patients received a US $31 gift voucher if they completed all surveys in the study.

Among 42 patients who were approached, 33/42 (79%) consented to participate, and 31/33 (94%) completed baseline (t₀) surveys. Of those who completed baseline surveys, 3/31 patients (all from the intervention arm) withdrew from the study (10 %) due to the burden of the disease. All the remaining 28 patients (100%) completed week-12 surveys (12 intervention and 16 control) (Figure 1).

Demographics of randomized patients are provided in Table S2 in Multimedia Appendix 7. There were no significant differences in participants’ demographics between the 2 arms.

The study recruitment rate was 33/42 (79%). Most participants went through the information and consent process via phone, except 4 were recruited on-site. The randomization rate was 31 of 33 (94%). The retention rate was 28 of 33 (85%). The RC conducted phone calls or in-person check-ups at least 2 times during the 12-week study. All recruitment, randomization, and retention rates were higher than the upper threshold.

A total of 13 intervention arm patients (13/15, 87%) used the SAMSON app. The proportions of responses to medication reminders and side effect surveys compared to the scheduled tasks during the 12-week study period were 1061 of 1541 (68.9%) and 128 of 141 (90.8%), respectively. The unmet threshold of responses to medication reminders could be due to the app’s functional issues as reported earlier. In total, 65 MI teleconsultations were delivered from August 2023 to June 2024, of which 35 of 65 (54%) were conducted on time as scheduled, 26 (40%) were later than scheduled, and 4 (6%) were missed. The average length of initial pharmacy consultations was 40 minutes, while nurse follow-up consultations were 40 minutes on average. Reasons for delayed and missed consultation sessions were mostly on the patient side, including not showing up at the appointment (25/30, 83%), not responding to HCP phone calls (4/30, 13%), or international travel (1/30, 3%). Only 2 sessions (7%) were delayed, because the HCP could not match their work schedule. Of the 12 intervention arm patients who were retained until the end of the study, 10 of 12 (83%) received all 5 scheduled teleconsultations, 1 (8%) received 4 teleconsultations and 1 (8%) received 2 teleconsultations.

Baseline surveys were completed by 31/33 (94%) patients. All who stayed until the end of the study (28/28, 100%) completed week-12 surveys. Participants completed all surveys online without any need for support. An alert email was sent by RC to all participants a few days before the survey due date. However, over one-third of participants (10/28, 36%) only completed surveys after being reminded. Details of feasibility results are presented in Table S3 in Multimedia Appendix 10.

In this study, we aimed to test the acceptability, feasibility, and preliminary efficacy of a multicomponent MA solution, SAMSON, to help patients with cancer improve their adherence to OAMs and self-manage their physical and emotional symptoms. Overall, patients and oncology HCPs rated SAMSON as highly acceptable, usable, and useful. These high levels of user satisfaction evidence that the solution meets the various needs of support among patients with cancer to medically adhere to and manage side effects at home [37], as well as the needs of oncology HCPs for a practical and tailored MI training, and a means to regularly monitor and provide ongoing support to patients’ MA [20]. Moreover, qualitative findings suggest that SAMSON has the potential to help patients in MA and symptom self-management, as well as to assist HCPs in monitoring and supporting patients’ adherence. The results of the study help to address the gap of knowledge and the need in oncology practice [17,18] by providing evidence of the high quality and potential effect of a digital multicomponent MA solution.

Regarding SAMSON’s feasibility, we noted a high level of engagement with the SAMSON solution in terms of the overall tasks completed by patients on the SAMSON smartphone app and the MI teleconsultations completed by both HCPs and patients. The high acceptability and feasibility of SAMSON is a result of several factors. First, co-design and rigorous design framework were applied to develop SAMSON [22]. By involving end users and stakeholders throughout all design stages of the intervention, co-designing helped to improve the solution’s acceptability, desirability, and usability [38,39]. The use of design frameworks, for example, design science research methodology in this study, could enhance the artifact’s design, which is crucial in digital intervention development, and improve its acceptance, usage, and efficacy [40]. Furthermore, the design framework usage can improve the rigor and translatability of research, which is currently limited or poorly reported in the development of available MA interventions in cancer [18]. Second, after development, individual components of SAMSON were successfully tested by end users on their acceptability, usability, and usefulness [19,20]. To the best of our knowledge, SAMSON is the first comprehensive digital MA solution in cancer that is co-designed, theory-based, evidence-based, and rigorously developed and tested.

Although overall acceptance and feasibility were relatively high, some HCPs and patients in the trial reported several technical issues with the SAMSON app and desired better visualization, more functionality, and further instructions and training. This feedback will be used to further improve the solution in the future. A couple of study HCPs were concerned about the time required for MI consultations. Time constraints have been reported as one of the barriers to MI implementation in clinical practice [41]. It is noted that in this study, SAMSON was delivered as an add-on service on top of the hospital’s usual care, which required extra time and effort from busy HCPs. In addition, delivering MI consultations to promote adherence was a newly acquired skill by the study HCPs, thus, they required more time to master this approach. A holistic approach, including the hospital’s mechanisms to provide continuous MI monitoring and training, as well as support (both in terms of skills and resources) to facilitate and maintain HCPs’ confidence and motivation in using this skill set can be a solution to tackle MI implementation barriers [41,42].

The overall recruitment, randomization, retention, and data collection compliance rates were higher than the preset upper thresholds. A high attrition rate is generally one of the major concerns in digital health RCTs compared to RCTs testing more “traditional” interventions [43]. However, the high recruitment rate and the low rate of loss to follow-up in this trial indicate that the SAMSON solution and online pragmatic RCT design are feasible in busy oncology clinical settings, as long as appropriate methodological strategies are applied. Specifically, in this study, clinical nurses and pharmacists from the Hematology Department were recruited and funded to support study recruitment and deliver teleconsultations. The RC dedicated additional time to building rapport with participants and consistently assisted them through all processes of the trial (consenting, baseline assessment, randomization, app installation, technical training and support, and outcome assessments). In addition, different channels of communication, for example, emails, phone calls, and SMSs were used to follow-up and motivate participants throughout the study. The importance of check-ups and follow-ups postrandomization in trials of digital interventions, where participants are required to have certain levels of digital literacy to perform tasks of the intervention [44], has been emphasized in several studies [45]. Future digital trials may benefit from these recruitment and retention strategies.

In this trial, a combination of techniques was used to measure MA, including self-report measures and prescription refill reports. Although triangulation of measurements could improve the accuracy of adherence [46], its interpretation needs to be cautious. The MRA using pharmacy computer records, which is objective [47], does not guarantee that all dispensed medications were consumed by patients. The self-reported adherence survey (ASK-12) is simple, but often subjective and more about barriers to adherence than actual adherence status [31]. Future studies should consider using high-accuracy methods, such as the Medication Event Monitoring System [5], to measure MA.

This study has some limitations. Like many digital health studies, our sample was predominantly Caucasian and highly educated [48,49]. Consequently, the findings may not be representative of patients who are non-Caucasian or have lower socioeconomic status, who might face higher barriers to adherence and could potentially benefit more from the SAMSON solution than their Caucasian or higher socioeconomic status peers [50,51]. Participants were recruited from the Hematology department at PMCC, one of Australia’s leading oncology hospitals, and were prescribed only 1 oral anticancer regimen. Therefore, the results may not be generalizable to those who use multiple anticancer medications or receive care in low-resource oncology settings. A more targeted recruitment strategy focusing on underserved cancer patient populations with other types of cancer in various levels of oncology care institutions is warranted.

Before undertaking this pilot trial, both components of the SAMSON solution were co-designed and developed based on evidence and theory, and then individually tested on target users. The results of this study confirmed that SAMSON is acceptable, usable, and useful for both HCPs and patients with cancer. Both the SAMSON solution and the pragmatic RCT design are feasible in real-life oncology settings. These findings are very encouraging, given the numerous challenges in applying RCT as an evaluation design for digital health interventions [52]. Our next steps will involve refining the SAMSON solution based on participants’ feedback from this study and conducting a full RCT to evaluate its clinical and economic effectiveness.