Integrating remote patient monitoring into pulmonary rehabilitation: a feasibility randomised controlled trial

Abstract

Background

Digital remote patient monitoring (DRPM) may enhance pulmonary rehabilitation (PR) by supporting continuous symptom tracking and early intervention, key features of effective disease management. This study examined the feasibility of integrating DRPM within a structured PR programme.

Methods

This was a parallel-group, single-centre, cluster-randomised clinical trial embedded within a quality improvement initiative. Patients were allocated to Standard-PR (16 PR sessions delivered over 6 or 8 weeks) or PR+DRPM. In addition to Standard-PR, the PR+DRPM group remotely monitored and transmitted blood pressure, heart rate, oxygen saturation, body temperature and weight data daily to PR staff during PR and for 12 weeks post-PR. Feasibility outcomes included recruitment, retention, adherence, acceptability, patient safety and impact on clinician workload. Secondary outcomes explored whether DRPM influenced self-management behaviours while preserving established benefits of PR.

Results

78 of 83 eligible patients approached were enrolled (94% recruitment rate; 55% male; mean±sd age 68.4±11 years; mean±sd forced expiratory volume in 1 s 59.1±22.7% predicted). Participant retention was 77% post-PR and 68% at the 12-week follow-up, with similar rates between groups. DRPM adherence was 90% during PR and 89% at follow-up. 100% of participants would recommend DRPM to other patients. A total of 163 compliance calls were made by the vendor. PR staff workload increased by 45 min per participant in the PR+DRPM group addressing DRPM-generated alerts.

Conclusion

Integrating DRPM into PR was feasible based on pre-specified benchmarks, but substantially increased provider workload, underscoring the need for workflow optimisation before initiating a larger trial.

Shareable abstract

Integrating digital remote patient monitoring into pulmonary rehabilitation for COPD is feasible, well accepted and supports self-management, although provider workload increases, warranting further evaluation https://bit.ly/3LKRGFi

Introduction

Pulmonary rehabilitation (PR) is a well-established intervention for managing COPD. Beyond clinically meaningful improvements in symptom control, exercise capacity and health-related quality of life (HRQoL) [1, 2], the benefits of PR extend to enhancing patient self-management through better symptom recognition and help-seeking behaviour [3, 4]. However, many individuals struggle to sustain these behaviours without ongoing interactions with healthcare providers [5, 6]. Across North America, COPD exacerbations remain prevalent and a major contributor to higher morbidity, frequent hospitalisations and a diminished quality of life for those living with the disease [7, 8]. Traditional PR models provide structured support during scheduled sessions but offer limited real-time insights into patients’ clinical status at home. The prevention and mitigation of COPD exacerbations require a model of care that offers sustainable and ongoing support.

Digital health technologies have become more ubiquitous since the COVID-19 pandemic, even extending the reach and benefits of PR beyond traditional clinical settings. Virtual health and telerehabilitation have improved PR accessibility and enabled more people to participate and improve their health [9–13]. Evidence suggests that telerehabilitation is safe, and patients can achieve clinically meaningful outcomes [9, 11–13], although not always matching those achieved with in-person programmes [14].

Digital remote patient monitoring (DRPM) offers the potential to further enhance PR by extending clinical oversight into the home environment for all patients. Daily physiological monitoring can help patients detect early symptom worsening, reinforce self-management behaviours, and facilitate timely and personalised provider feedback [15–17], mechanisms aligned with the goals of PR [3], yet difficult to achieve without home-based data. COPD exacerbations can be predicted through monitoring physiological parameters [15]. DRPM may provide an additional layer of support between PR sessions, offering some patients the timely reinforcement needed to optimise their health and potentially avoid hospitalisations.

Despite the theoretical advantages, evidence on the feasibility and implementation of DRPM within a real-world PR programme (PR+DRPM) remains limited. A key concern is the practicality for patients and providers, specifically regarding patient adherence to the monitoring protocol and provider workload associated with monitoring [18]. Therefore, this study evaluated the feasibility of integrating DRPM into a centre-based PR programme. Specifically, we examined recruitment, retention and adherence to DRPM, programme acceptability, safety and provider workload to inform the design and implementation of a larger randomised controlled trial. Secondary outcomes explored whether DRPM supports self-management behaviours and maintains the established benefits of PR.

Methods

Study design and ethical considerations

This single-centre study to evaluate the feasibility of DRPM within PR was embedded within a quality improvement initiative undertaken in partnership with the Alberta Health Services (AHS) Virtual Health programme. The PR staff implemented the quality improvement project as part of routine clinical care using a parallel-group, cluster-randomised clinical trial design. The PR classes were randomised to Standard PR or PR+DRPM using a computer-generated randomisation sequence in Microsoft Excel (Version 2408). Randomisation occurred at the class (cluster) level to minimise selection bias and reduce contamination. Classes were randomised to PR+DRPM or Standard-PR only after classes were fully formed, preventing participants or staff from self-selecting into a condition. Within the PR+DRPM classes, patients were offered the option to receive the DRPM system as part of standard clinical practice. Individuals who declined participation in DRPM continued to receive traditional PR. Given the nature of rehabilitation, the allocation could not be concealed from participants or PR providers once the PR programme began.

The research component was approved by the University of Alberta Ethics Board (Study ID: Pro00066560) and registered with ClinicalTrials.gov (ID: NCT06077994). All participants provided written informed consent to their data being used for research and prior to the completion of any research-related questionnaires. The sole research instrument was the Partners in Health (PIH) self-management scale, a 12-item questionnaire that evaluates patients’ self-management knowledge and behaviours [19, 20]. All other data were collected as part of standard care. The patient consent process emphasised that all decisions regarding participation in the research were voluntary and would not affect access to PR services.

Participant recruitment

Participants were recruited from the Breathe Easy Pulmonary Rehabilitation programme at the G.F. MacDonald Centre for Lung Health in Edmonton, Alberta. As a standard of care [21, 22], a pulmonologist evaluated all prospective PR participants before the programme began. This assessment included an incremental cardiopulmonary exercise test [23] and a 6-min walk test (6MWT) [24]. A brief functional mobility screen was conducted to identify areas where the assessment or exercises might require modification to ensure safe participation, and included: assessment of global spinal mobility, functional shoulder range of motion and the ability to perform a full-bodyweight squat. After clinical assessment, participants were enrolled in a PR class based on personal preference, without knowledge of the class's study group assignments. All patients deemed fit for PR (aged 18 years or older with a diagnosis of a chronic respiratory disease) and enrolled in a class were eligible for this trial. Individuals who were cognitively impaired, nonambulatory or diagnosed with unstable cardiovascular disease were excluded from PR.

Interventions

Standard-PR

All participants attended a PR programme, which was scheduled twice a week for 8 weeks or thrice a week for 6 weeks, for a total of 16 sessions. Each class included 60 min of education and 90 min of supervised aerobic and resistance training following standard PR guidelines and tailored to each patient's ability [21, 22]. Supplementary table S1 outlines educational content delivered across PR sessions. PR attendance was recorded as the number of sessions attended out of 16. PR completion was defined as attending nine sessions, including the essential educational components, and completing the post-PR assessments.

PR+DRPM

In addition to PR, the PR+DRPM group also received a Cloud Dx Connected Health system (Cloud Dx Kitchener, Ontario, Canada; see supplementary figure S1). This system operates via Wi-Fi or a built-in cellular connection and connects to patients’ electronic medical records through the provincial health system. Patients received a password-protected kit at no cost during the fourth class of their PR programme. This timing was intentional; the early weeks of PR include multiple assessments, education sessions and exercise orientation, and therefore delaying the kit orientation helped to limit patient burden at the start of PR. In addition to the physical kit, participants were provided with written instructions on ensuring data quality when taking readings and an action plan to follow if the readings fell outside established thresholds. Within 24 h of receiving the device, the Cloud Dx vendor provided more detailed education over the telephone on how to connect peripheral devices, transmit data and troubleshoot errors. For quality assurance, the vendor assisted participants in taking their initial readings during this education session, verifying the accuracy of data collection.

Using Bluetooth-connected peripheral devices included with the Cloud Dx system, the DRPM protocol required patients to take their vitals (blood pressure, heart rate, oxygen saturation, body temperature and body weight) daily between 07:00 and 11:00. Except for body weight, patients were instructed to take readings after at least 5 minutes of quiet sitting. Daily vital sign measurements continued throughout and up to 12 weeks post-PR, with participants encouraged to review their data for trends to inform self-management.

Data were securely transmitted to a protected repository hosted by the provincial health system through Wi-Fi or cellular connections. The DRPM system generated alerts whenever any physiological reading fell outside the predefined normal range, notifying patients and PR providers via e-mail. The parameters included systolic/diastolic blood pressure readings above 160/105 or below 90/60 mmHg, oxygen saturation <88%, heart rate >100 bpm, temperature >37.5°C or a body weight change of ±3 lbs between days. Patients were instructed to contact a healthcare professional if abnormal readings were accompanied by symptoms such as increased shortness of breath, chest pain or extreme fatigue, and to seek emergency care for severe symptoms (e.g. confusion, drowsiness or chest pain), informing the PR team once safe to do so.

Alerts were reviewed daily by a designated experienced PR provider (physiotherapist or respiratory therapist) and responded to within 24–48 h during regular work hours (Monday to Friday, between 08:00 and 16:00) via telephone or in-person consultation during class. Alerts occurring after hours and on weekends were responded to during the next workday. When an alert occurred, the provider first reviewed the patient's DRPM record; if the initial abnormal reading was followed by a normal reading later that same day, no further action was taken. For alerts that remained abnormal, the provider contacted the patient by phone to review the reading. If the reading appeared questionable, the patient was coached to repeat the measurement while on the phone. When an abnormal value was confirmed, the provider advised the patient on appropriate next steps, which could include initiating their COPD action plan, contacting their primary care physician, or seeking immediate medical attention.

In addition to reactive monitoring, PR providers reviewed patient data weekly to provide proactive care. If a patient missed a reading for >48 h, the Cloud Dx vendor initiated a compliance call and notified PR providers when they were unable to reach the participant. Post-PR, in addition to regular monitoring, PR staff conducted monthly telephone check-ins to ensure patient engagement, assess overall progress and address any concerns not captured through physiological readings.

Outcomes

Feasibility outcomes

The selection of feasibility outcomes was guided by Bowen et al. [25], and included patient recruitment and retention, adherence to the DRPM protocol, intervention acceptability, adverse events and impact on provider workload/resource use. Recruitment success was defined as enrolling ≥80% of eligible participants; retention benchmarks were ≥70% completion of PR and ≥50% participation in the 12-week follow-up. Adherence to DRPM was deemed feasible at ≥80% of expected daily readings. Programme acceptability was considered achieved if ≥80% of participants rated their experience as “somewhat easy/useful” or “very easy/very helpful” on the Technology Utilisation Questionnaire. Safety and operational feasibility were defined as no serious adverse events and <10% equipment loss. Adverse events were recorded and categorised by severity and attribution. Mild events were transient and required no change in medical or rehabilitation management (i.e. skin irritation from the devices, brief discomfort during exercise). Moderate events required a temporary modification or pause in the intervention, but no medical treatment (e.g. mild increase in respiratory discomfort). Severe events were defined as those requiring medical assessment, pharmacological interventions or hospitalisation. Attribution to PR or DRPM was determined by the PR team. Adverse events were tracked throughout the study in patient chart notes, vendor support logs and a Research Electronic Data Capture (REDCap) database [26, 27]. Provider workload was captured objectively through Cloud Dx's internal time-tracking feature. Each time a provider opened a patient file to review an abnormal alert, the platform logged the full session duration, ending only when the provider manually closed the platform after completing patient contact, follow-up and documentation. Workload demand also included a count of vendor-initiated reminder calls. No a priori threshold was set for provider workload.

Secondary outcomes

Exploratory outcomes were assessed during the first week of PR (pre-PR), the last week of PR (post-PR) and the 12-week follow-up, including functional exercise capacity (mean distance of the best two 6MWTs) [24], health status measured using the COPD Assessment Tool (CAT) [28] and chronic disease self-management using the Partners in Health (PIH) scale [18]. Clinically meaningful change was defined a priori using established minimal clinically important differences (MCIDs): 25 m for the 6MWT [29] and 2 points for the CAT [30]. The MCID for the PIH has not been reported [20].

Data analysis

This feasibility study was not powered to detect statistically significant differences in clinical outcomes. A target sample size of 70 participants (35 per group) was chosen based on recommendations for pilot trials, allowing for the assessment of recruitment rates, adherence and acceptability, and to estimate variability for future sample size calculations [31, 32].

Baseline participant characteristics and group differences were compared using the t-test. Descriptive statistics were performed for recruitment, retention, adherence, acceptability and provider contact time. Continuous variables were described using the mean±sd, and categorical variables were described using the frequency (n) and percentage (%). Linear mixed-effects models [33, 34] were used to estimate within- and between-group differences in the 6MWT, CAT score and PIH scale. Fixed effects included the treatment arms (Standard PR, PR+DRPM) and time (pre-, post- and 12-week follow-up), and individual patients accounted for random effects. Model assumptions were verified using residual plots, and missing data were addressed through maximum likelihood estimation. Statistical analyses were performed using Stata version 18 (StataCorp LLC, College Station, TX, USA).

Results

Feasibility

83 individuals were approached for participation; five PR classes were randomised to PR+DRPM and four to Standard-PR. Recruitment rates met feasibility benchmarks, with 94% of eligible participants enrolled (n=78; 55% male; mean age 68.4±11 years; mean forced expiratory volume in 1 s 59.1±22.7% predicted). Of these, 41 participants were enrolled in a 6-week programme, and 37 in an 8-week programme. Overall, 77% of participants completed the PR programme, and 68% attended the 12-week follow-up evaluation (figure 1). The dropout rates were similar between the treatment groups (p=0.78). Participant characteristics were comparable at baseline (table 1).

FIGURE 1.

Participant flow diagram. PR: pulmonary rehabilitation; DRPM: digital remote patient monitoring.

TABLE 1.

Baseline characteristics of study patients

Variable	Overall	Intervention	Control	p-value	Study completers	Study dropouts	p-value
Patients, n	78	35	43		53	25
Age years	68.4±11.0	67.1±11.9	69.4±10.5	0.36	67.6±10.7	70±12.5	0.38
Sex, n (%)				0.89			0.92
Male	43 (55)	19 (54)	24 (56)		29 (55)	14(56)
Female	35 (45)	16 (46)	19 (44)		24 (45)	11 (44)
BMI kg·m−2	29.3±6.8	29.3±7.1	29.3±6.7	0.99	29.4±7.24	28.58±5.61	0.64
Smoking history pack-years	33.4±18.2 (n=68)	32.8±20.0 (n=39)	33.8±17.2 (n=29)	0.83	34.6±17.0 (n=44)	31.1±20.1 (n=24)	0.45
FEV1 % predicted	59.1±22.7	55.9±23.2	61.7±22.3	0.26	58.2±22.8	61.2±23.0	0.56
FVC % predicted	83.4±21.3	83.4±22.8	83.1±20.2	0.95	83.4±22.5	83.0±19.1	0.94
FEV1/FVC ratio	55.6±18.0	53.6±19.9	57.2±16.4	0.39	54.2±15.8	58.4±22.1	0.34
LTOT, n (%)	12 (15.4)	6 (17.1)	6 (17.6)	0.66	8 (15)	4 (16)	0.92
Primary diagnosis # , n (%)				0.53			0.31
COPD	49 (62.8)	20 (57.1)	29 (67.4)		33 (62)	16 (64)
Asthma	4 (5.1)	3 (8.6)	1 (2.3)		4 (8)	0 (0)
Interstitial lung disease	9 (11.5)	5 (14.3)	4 (9.3)		5 (9)	4 (16)
Comorbidities ¶ , n (%)
Cerebrovascular	4 (5)	2 (6)	2 (5)	0.8	1 (2)	3 (12)	0.11
Congestive heart failure	9 (12)	4 (13)	5 (12)	0.94	5 (10)	5 (22)	0.17
Diabetes	12 (16.0)	5 (15)	7 (17)	0.78	13 (25)	1 (4)	0.05
Hypertension	45 (58)	19 (56)	26 (61)	0.68	32 (60)	15 (63)	0.86
Liver	7 (9)	3 (9)	4 (10)	0.92	5 (10)	2 (9)	1
Renal	12 (16)	4 (13)	8 (20)	0.42	7 (14)	7 (30)	0.1

Unless otherwise stated, all values are presented as the mean±sd. BMI: body mass index; FEV₁: forced expiratory volume in 1 s; FVC: forced vital capacity; LTOT: long-term oxygen therapy. ^#: missing diagnoses may include α₁-antitrypsin deficiency, bronchiectasis, restrictive musculoskeletal disorders and post-COVID infection; ^¶: these findings are not mutually exclusive.

The PR+DRPM group had the Cloud Dx kit for a mean of 119±13 days. One participant initially consented to remote monitoring but declined before data collection began and therefore contributed no DRPM data. Among those who engaged in remote monitoring, adherence to DRPM was 90% [10] during PR and 89% [11] during follow-up, corresponding to an overall adherence rate of 89% [11] across the study period. Figure 2 illustrates the adherence rate for each vital sign at the end of the study. Among the monitored parameters, adherence was lowest for temperature and body weight measurements.

FIGURE 2.

Heat map of participant adherence to vital sign monitoring at the end of the study. The heat map represents the percentage adherence of each patient to the five vital signs: blood pressure (BP), oxygen saturation (S_pO2), heart rate (HR), temperature (Temp) and body weight (BW). Darker shades indicate poor adherence. Each row corresponds to a patient, and each column corresponds to a vital sign. Mean adherence throughout the study was 89% [11]. A total of 23 participants are shown: one PR+DRPM participant was excluded because they did not use the DRPM kit and had no adherence data. PR: pulmonary rehabilitation; DRPM: digital remote patient monitoring.

A total of 496 alerts were generated during the monitoring period. Alerts were related to abnormal blood pressure (n=332) and elevated heart rate (n=109), with fewer alerts associated with low oxygen saturation (n=33), altered body weight (n=20) or high temperature (n=2). Most of the alerts (n=313, 63%) were single abnormal readings that normalised upon repeat testing by the patient. These alerts were reviewed in the patient file by the clinician but were not formally followed up on with the patient. Of the remaining alerts (n=183, 37%) that required a provider phone call and patient review, 17 (9%) led to a recommendation to initiate a COPD action plan, with some (n=8) recommended to seek primary care follow-up. No alert resulted in a recommendation to seek emergency care. Staff burden associated with alert review averaged 45 min per participant across the full monitoring period. The vendor's team completed 163 compliance calls, each lasting 2–3 min.

The acceptability of the DRPM system was assessed using a fit-for-purpose questionnaire, the Technology Utilisation Questionnaire (supplementary material). Participants rated the technology as “somewhat easy/useful” or “very easy/very helpful” (figure 3). Notably, 92% of participants rated their overall experience with the equipment as ≥8 out of 10, and 100% indicated they would recommend DRPM to other patients.

FIGURE 3.

Acceptability of digital remote patient monitoring (DRPM) technology. Results are based on responses from participants in the PR+DRPM group (n=23) to the Technology Utilisation Questionnaire, demonstrating a high level of acceptability among participants. The data highlight participants’ favourable perceptions of the usability, functionality and overall integration of the DRPM technology. PR: pulmonary rehabilitation.

Adverse events and challenges

No adverse events were noted in either group. Challenges encountered during the study related to DRPM were documented in participant chart notes, vendor support logs and research team records. These included:

1. Equipment-related issues: Due to size constraints, one participant required a replacement blood pressure cuff.

2. Connectivity issues: Three participants (9%) experienced intermittent Internet and/or Bluetooth connectivity challenges, which were resolved with vendor support. Since participants could manually enter their vitals to reduce missing data, it was difficult to determine the exact number of people affected by connectivity issues.

3. Equipment return: At study completion, two Cloud Dx kits remained unreturned despite follow-up reminders.

Exploratory outcomes

Attendance in the PR classes was high, particularly in the PR+DRPM group (94% versus 86% in Standard PR, p=0.004). The magnitude of changes in the 6-min walk distance and CAT scores with PR were clinically important [29, 35] and comparable between the groups over time, aligning with previous trials demonstrating improvements post-PR. An increase in PIH score was observed in both groups, indicating improvement in self-management skills, with no significant between-group differences (table 2).

TABLE 2.

Summary of regression models for the secondary outcomes

Outcomes	Coefficient	95% CI	p-value
6MWD
Main effects
PR programme#	−8.26	(−52.21–35.69)	0.71
Post-PR – Pre-PR¶	26.67	(9.52–43.62)	<0.01
12-week FU – Pre-PR¶	16.17	(−1.94–34.28)	0.08
Interaction
Post-PR – Pre-PR × programme#	8.04	(−17.58–33.66)	0.54
12-week FU – Pre-PR × programme#	10.11	(−19.85–40.07)	0.51
Standard-PR group at baseline	371.67	(341.75–401.60)
CAT score
Main effects
PR programme#	−1.00	(−4.19–2.19)	0.54
Post-PR – Pre-PR¶	−1.57	(−3.35–0.21)	0.09
12-week FU – Pre-PR¶	−2.65	(−4.54– −0.76)	0.01
Interaction
Post-PR – Pre-PR × programme#	−0.41	(−3.08–2.26)	0.76
12-week FU – Pre-PR × programme#	0.39	(−2.56–3.34)	0.80
Standard-PR group at baseline	20.66	(18.46–22.86)
PIH scale
Main effects
PR programme#	−3.38	(−8.50–1.73)	0.19
Post-PR – Pre-PR¶	9.54	(5.60–13.50)	<0.01
12-week FU – Pre-PR¶	7.85	(3.73–11.97)	<0.01
Interaction
Post-PR – Pre-PR × programme#	2.65	(−3.03–8.33)	0.36
12-week FU – Pre-PR × programme#	2.38	(−3.58–8.35)	0.43
Standard-PR group at baseline	75.91	(72.35–79.47)

Bolded p-values indicate significant findings. Greater 6MWD, higher PIH scores and lower CAT scores indicate better outcomes. 6MWD: 6-min walk distance; PR: pulmonary rehabilitation; FU: follow-up; CAT: COPD Assessment Tool; PIH: Partners in Health. ^#: the Standard-PR programme is the reference category for the main effect of the programme and the interaction terms; ^¶: Pre-PR is the reference category for the main effects of time.

Discussion

The high recruitment, retention and adherence and strong patient acceptability rates demonstrate the feasibility of integrating DRPM into a real-world PR programme. Adherence rates were high (89%%) throughout the study period, indicating that participants were willing and able to integrate remote monitoring into their daily routines throughout PR and for 12 weeks following PR completion. Additionally, most participants reported positive experiences with the technology, with 100% stating they would recommend it to others and 92% expressing high satisfaction with the equipment. These findings indicate that DRPM can be successfully implemented within a centre-based PR programme and that participants perceive value in ongoing remote monitoring as part of their disease management strategy. The main challenge noted in integrating DRPM into PR was the increased provider workload, which raises concerns about its practicality and necessitates workflow adjustments prior to a full-scale efficacy trial.

Despite prior evidence indicating that older people encounter challenges with digital literacy [18, 36, 37], this study demonstrated high adherence to DRPM, including during the 3-month follow-up period with limited provider contact. The structured onboarding process and proactive technical support provided may have facilitated engagement, underscoring the importance of these elements for future work. A scoping review by Wilson et al. [38] identified key facilitators of e-health adoption among older adults, including improved self-efficacy in using the technology, addressing privacy concerns and integrating e-health programmes within broader health services. The integration of the Cloud Dx system within patients’ electronic health records may have supported patient acceptance, although its role in adherence is likely more complex. Integration with the electronic health record also enabled real-time access to data for the entire PR team, allowing feedback to be embedded directly within routine clinical interactions. This may have contributed to greater adherence by promoting accountability and relevance to patients, particularly when data were discussed by trusted healthcare providers during PR. From a social cognitive theory perspective, this integration may have positively influenced key behavioural determinants such as self-efficacy, perceived outcome expectancies and social reinforcement [39]. Future studies should explore how digital health tools can be better paired with behavioural supports delivered within interdisciplinary teams to enhance motivation and long-term engagement with self-monitoring behaviours.

A key consideration for interpreting feasibility is the operational burden associated with DRPM. Although patient adherence and acceptance were high, provider workload increased substantially. Monitoring alerts, responding to patient-generated data and engaging patients added ∼45 min per participant, not accounting for vendor-initiated reminder calls. While manageable in a small feasibility study, this increased workload may not be sustainable in a routine PR programme. Moreover, most alerts represented transient deviations in vital signs that resolved, with only a few requiring providers’ follow-up. A small proportion of alerts (9%) prompted formal interventions, but none resulted in emergency room visits or unplanned hospitalisations. Although exacerbation history was not captured at baseline, the participants’ clinical profile suggests moderate exacerbation risk. The low number of actionable alerts likely reflects both participant stability during the PR programme and the need for more individualised alert thresholds to reduce nonactionable notifications in future work. Future implementation should explore refining DRPM alert thresholds, establishing alert triage and redistributing monitoring tasks across a broader care team as strategies to reduce provider burden.

In addition to provider workload, device management emerged as a practical challenge during implementation. At the time of study completion, two DRPM kits remained unreturned, despite clear discussions about participant agreement at the start of the study and multiple follow-up reminders. While this represents a small proportion of the total devices distributed, it raises important logistical and cost considerations for future studies and scale-up. Clear return protocols, patient agreements or alternative strategies such as refundable deposits, automated reminders and the use of lower-cost disposable technologies may be necessary to ensure sustainability in large-scale implementations of remote monitoring interventions.

Given that DRPM might increase treatment burden for patients, its impact on established PR outcomes was assessed. Clinically significant improvements were observed in the 6MWT, with CAT scores showing gains approaching the MCID. The absence of between-group differences in functional exercise capacity and health status indicates that adding DRPM does not compromise established PR outcomes. Notably, PR completion rates were higher in the PR+DRPM group, suggesting that incorporating digital health technology into PR may enhance adherence to PR [40]. Additionally, patient experience with PR was comparable in both groups, indicating that DRPM did not diminish programme satisfaction. These findings suggest that short-term clinical outcomes are similar across groups, and that digital health interventions may help sustain engagement in PR.

This study has several strengths. First, it was conducted within a real-world PR programme, enhancing ecological validity and demonstrating the practicality of implementing DRPM alongside established PR workflows. Second, while previous studies have highlighted clinician burden and data overload in remote monitoring programmes [18, 41], none have directly quantified their impact within PR as in this study. Lastly, a key strength of the study is its retrospective alignment with the recently published PANACEA framework [42], which outlines best practice for assessing home-based monitoring in chronic lung disease. Although the framework was not available at the time of study design, our approach aligns with many of its key domains, including access, patient experiences and cost (supplementary table S2), and highlights several areas for future work.

Several limitations should be noted. First, the study sample may not fully represent the broader population with chronic lung disease, particularly those with lower digital literacy or varying levels of self-management readiness, which could influence adherence rates. Second, provider workload totals captured through the DRPM device included only time spent responding to abnormal alerts; vendor-initiated call durations and other indirect tasks were not recorded, likely underestimating the overall implementation burden. Third, the absence of qualitative data from participants, clinicians and vendors limits our understanding of how individuals engaged with the DRPM system and the contextual factors influencing usability. Future mixed-methods research incorporating interviews or focus groups will be essential to better understand these experiential factors and inform broader implementation. Finally, as is the nature of feasibility studies, the small sample size, single-centre design and short follow-up duration limit the ability to draw conclusions about clinical efficacy.

Future work should explore the use of personalised thresholds to reduce nonactionable notifications and provider workload. Such an approach would involve adjusting alerts to patients’ baseline physiological values and patterns, rather than relying on standardised thresholds, thereby reducing false positives and directing provider attention to clinically meaningful deviations. A future trial should also evaluate behavioural responses to PR+DRPM, such as time to initiate action plans and changes in physical activity, as well as downstream outcomes, including hospitalisations avoided. Ultimately, a large trial powered to reduce COPD-related hospitalisations is needed, which would include an economic evaluation examining the increased costs of PR+DRPM, balanced against healthcare utilisation avoided.

Conclusion

In summary, integrating DRPM into PR met pre-specified feasibility thresholds for recruitment, retention, adherence, acceptability and safety. However, the study also revealed challenges related to increases in provider workload as a result of DRPM alerts. These findings suggest that while the model is promising, further refinement is required to ensure clinical and operational sustainability. Future research should explore strategies to optimise alert management and assess long-term clinical and economic impact before broader implementation.

Supplementary material

Please note: supplementary material is not edited by the Editorial Office, and is uploaded as it has been supplied by the author.

Figure S1

01305-2025.SUPPLEMENT

Supplementary material

01305-2025.SUPPLEMENT2

Table S1 and S2

01305-2025.SUPPLEMENT3

Data availability

Patient clinical data cannot be shared publicly due to privacy and ethical considerations. De-identified patient data underlying this study are available from the corresponding author upon reasonable request and subject to institutional ethics approval.