This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR mhealth and uhealth, is properly cited. The complete bibliographic information, a link to the original publication on http://mhealth.jmir.org/, as well as this copyright and license information must be included.
Telecoaching approaches can enhance physical activity (PA) in patients with chronic obstructive pulmonary disease (COPD). However, their effectiveness is likely to be influenced by intervention-specific characteristics.
This study aimed to assess the acceptability, actual usage, and feasibility of a complex PA telecoaching intervention from both patient and coach perspectives and link these to the effectiveness of the intervention.
We conducted a mixed-methods study based on the completers of the intervention group (N=159) included in an (effective) 12-week PA telecoaching intervention. This semiautomated telecoaching intervention consisted of a step counter and a smartphone app. Data from a project-tailored questionnaire (quantitative data) were combined with data from patient interviews and a coach focus group (qualitative data) to investigate patient and coach acceptability, actual usage, and feasibility of the intervention. The degree of actual usage of the smartphone and step counter was also derived from app data. Both actual usage and perception of feasibility were linked to objectively measured change in PA.
The intervention was well accepted and perceived as feasible by all coaches present in the focus group as well by patients, with 89.3% (142/159) of patients indicating that they enjoyed taking part. Only a minority of patients (8.2%; 13/159) reported that they found it difficult to use the smartphone. Actual usage of the step counter was excellent, with patients wearing it for a median (25th-75th percentiles) of 6.3 (5.8-6.8) days per week, which did not change over time (
The 12-week semiautomated PA telecoaching intervention was well accepted and feasible for patients with COPD and their coaches. The actual usage of the step counter was excellent, whereas actual usage of the smartphone tasks was lower and decreased over time. Patients who required more contact experienced less PA benefits.
ClinicalTrials.gov NCT02158065; http://clinicaltrials.gov/ct2/show/NCT02158065 (Archived by WebCite at http://www.webcitation.org/73bsaudy9)
Reduction in physical activity (PA) is a major feature of chronic obstructive pulmonary disease (COPD), occurring both as a consequence of disease and driving worse outcomes in the condition [
In this paper, 3 concepts, which are often assessed as part of a process evaluation, have been investigated: (1) acceptability, (2) actual usage, and (3) feasibility of the intervention from both a patient and a coach perspective. In addition, we aimed to investigate their association (ie, actual usage and feasibility) with the effectiveness of the intervention.
First, acceptability is a key concept in the development, evaluation, and the implementation of complex interventions and can have significant impact on the intervention’s effectiveness [
Second, the actual usage of the intervention by patients and health care providers forms an important part of the delivery of PA telecoaching interventions. Actual usage was assessed as the degree to which patients used the components of the interventions as it was designed [
Third, the implementation of this intervention also depends on whether it was considered to be feasible by patients as well by the coaches. Feasibility is defined as “the extent to which a new treatment, or an innovation, can be successfully used or carried out within a given agency or setting” [
Finally, the direct association between both coach feasibility (as assessed by contact time) and actual usage by patients with the effectiveness of the intervention was investigated. The latter insights could lead to improved design and implementation of PA telecoaching interventions in the future as well as optimized selection of patients.
This study was approved by the local ethics committee at each center (Commissie medische ethiek van de universitaire ziekenhuizen KU Leuven [Leuven, S-55919]; Medische ethische toetsingscommissie universitair medisch centrum Groningen [Groningen, Metc 2013.362]; RES Committee London—South East [London and Edinburgh, 13/LO/1660]; Scientific Council of the ‘Sotiria’ General Hospital for Chest Diseases (Athens, 27852/7-10-13); Kantonale Ethikkommission Zürich, and Ethikkommission Nordwest- und Zentralschweiz [Zurich, KEK-ZH-Nr. 2013-0469 and EKNZ2014-192, respectively]).
A convergent mixed-methods research design using quantitative and qualitative data was applied to evaluate the acceptability, actual usage, and feasibility of a PA telecoaching intervention. Both qualitative and quantitative data on the intervention were separately collected and analyzed. Later, these findings were compared for data triangulation, which allowed a more comprehensive understanding of the intervention [
This trial forms part of a 12-week, multicenter randomized controlled trial (1:1 randomization) conducted by the PROactive consortium [
Patients in the intervention group [
Overview of the intervention; 1=sending of “steps data” to smartphone (through Bluetooth); 2=data sent to central database; 3=coach is able to access database; 4=coach is able to manually adjust goals, 5=accessing & closing the different tasks on the smartphone app (automated messages); i.e., (from left to right); morning goal, send activity in the evening, daily feedback (from Monday to Saturday) and weekly feedback (only on Sunday) tasks.
Acceptability was assessed through quantitative data (a project-tailored questionnaire [20 items,
During the final visit of the study (V3), patients were asked to fill in a 20-min self-administered, project-tailored, multiple-choice questionnaire on their experiences with the intervention and the usefulness of its components on a 10-point Likert scale (
Patient interviews were conducted by local PA coaches in each center at V3. Each coach was informed and trained on how to conduct the interview during an investigator’s meeting before the start of the trial. Interviewers from each center were asked to transcribe the answers of the patients to the discussion guide questions and forward them (anonymized) to one researcher (ML) who collected all quotes into 1 Excel file for analysis. In this pooled Excel file, each line represented the verbatim answer of each participant on a question with a number code and a letter representing, respectively, the patient’s ID and the question of the discussion guide.
After completion of the trial, an audiotaped focus group was organized to capture the intervention experience from the perspective of the coaches. Local PA coaches with a diverse background (ie, medical doctor [RAR], physiotherapist [ML, HD], exercise physiologist [ZL], biomedical scientist [MS], and psychologist [AF]; n=6), and 2 experienced physiotherapists who were involved in the development of the intervention (n=2; EGS and Ane Arbillaga-Etxarri (AAE) from the center in Barcelona [IS GLOBAL]) discussed the feasibility, appreciation, possible future adaptations, time investment, and actual usage of the different components of the intervention (
Actual usage of the intervention by patients was assessed objectively through the smartphone app log. A database was derived directly from the smartphone app. This included information about completion of the app tasks and step counter data on a day-by-day basis. Actual usage of the step counter was defined based on the presence of step count data (ie, ≥70 steps for that day). Self-reported actual usage of performing home exercise and the times patients looked at their step counter were assessed subjectively in the project-tailored questionnaire.
Actual usage by the coaches was assessed based on the closure of tasks in the app-linked Web accounts and discussed during the coach focus group.
Coach feasibility was already partly assessed in the main paper of the MrPAPP trial in terms of number of contacts and total amount of contact time between coaches and patients (quantitative data) [
Both actual usage by patients and coach feasibility (ie, contact time) with the intervention were separately linked to the effectiveness of the intervention. This effectiveness was assessed as the change in numbers of steps per day after 12 weeks, measured by the Actigraph GT3x (ACT, Actigraph LLC Pensacola, FL). The latter is a triaxial accelerometer validated for use in patients with COPD [
All statistical analyses were performed with Statistical Analysis Software version 9.4 (SAS Institute, Cary, NC). Continuous variables were expressed as means with SD (normal distribution) or as medians (25th-75th percentiles [P25-P75]; skewed distribution), unless stated otherwise. Categorical variables were expressed as proportions and percentages. The level of significance was set at .05 for all statistical tests. The analyses were based on patients in the intervention group who completed the 12-week intervention (N=159).
Data from the project-tailored questionnaire were scored as categorical variables and reported as frequencies and percentages (ie, number of patients indicating each answer), except for the usefulness ratings of the components, which were expressed as median (P25-P75).
For analysis of the interview data, two researchers (HD and FMR) independently performed thematic analysis on the Excel file containing the verbatim transcriptions of the interview data [
HD and FMR read the data multiple times and descriptively noted down their initial ideas of what is in the data and what is interesting about them.
HD and FMR independently generated an initial list of codes from the data and put the data systematically under certain headings.
Afterwards, they searched for reoccurring themes, which began to emerge from these codes to focus their analysis on a broader level.
HD and FMR refined and defined their themes taken into account the overall message of the analysis. Themes and subthemes were organized and ranked into categories.
HD and FMR came together for group discussion to find an agreement on defining the themes and subthemes, which led to the development of a (final) codebook.
Afterwards, one researcher (ML) applied the final codebook to all verbatim transcripts. After iterative group discussions, data were synthesized and representative example quotes were extracted to illustrate findings and were labeled by a unique participant’s code together with the category of contact time and actual usage score of that participant.
The thematic analysis was conducted inductively (ie, themes emerged from the data, hence without predetermined coding frame) in Excel, without the use of specialized analytic software. Further details on the methodological aspects of the latter analyses have been added to the COnsolidated criteria for REporting Qualitative research (COREQ) checklist (see
During the focus group, 1 PA coach (ML) wrote a consensus summary. A total of 2 PA coaches (HD and MS) independently reviewed the consensus summary based on the audio recording. Additional information that was considered as relevant was independently added by both coaches (HD and MS). Only minor interpretation disagreements occurred between the 2 PA coaches, which were discussed together with a third PA coach (ML). Later, a summary of the focus group was sent for revision to all PA coaches, including those who could not be present at the focus group. A consensus quote on the future implementation of this PA telecoaching intervention was formulated.
Actual usage was compared according to age (<65 vs ≥65 years, Mann-Whitney
In the larger centers (inclusion of at least 20 patients), the contact time with the first 10 patients was compared with the others (Mann-Whitney
Division (into 3 groups) of patients based on total duration and number of contacts between patients and coach. Min=minutes; #=number of contacts; n=number of patients in each group.
We attempted to create 3 equally balanced groups (low, medium, and high) of total contact time (
To analyze the association between (1) the actual usage by patients of different components of the intervention and coach feasibility (ie, contact time) and (2) the effectiveness of the intervention, 2 separate generalized linear model analyses were used in completers with valid PA data (88.1% [140/159] of the completers sample). Change in PA was used as the outcome and contact time and actual usage as the class variables, respectively. Due to their possible influence on the intervention effect, baseline exercise capacity (6MWD), symptom score (mMRC-scale), forced expiratory volume in 1 second (FEV1) % predicted, and the number of acute exacerbations in the previous 12 months were considered as possible (continuous) covariates of the association [
Baseline characteristics of the 159 completers are outlined in
Overall, the PA telecoaching intervention was well received by the patients as 89.3% (142/159) indicated that they “enjoyed taking part in the intervention.” Furthermore, the majority of the patients (59.1%, 94/159) claimed that the intervention coached them “a lot” toward enhancing their PA. Approximately half of the patients (47.2%, 75/159) experienced the proposed weekly increases in step counts as “reasonable,” whereas 37.7% (60/159) and 10.1% (16/159) of the patients experienced these increases as “a little bit too high” and “much too high,” respectively.
Patients rated the usefulness of the step counter (median [P25-P75]; 10 [
When patients were asked to name the most important part of the intervention, 76.1% (121/159) of patients did choose the step counter as the most important part with 93.1% (148/159) of all patients willing to continue using the step counter in the future. In total, 45.9% (73/159) of all patients were willing to continue using the full intervention, with only 8.2% (13/159) of all patients reported to experience working with the smartphone as difficult.
Baseline characteristics of the completers of the trial.
Variables | Intervention completers (n=159) |
Age in years, mean (SD) | 66 (8) |
Gender (male), n (%) | 89 (64) |
BMIa (kg/m2), mean (SD) | 26.9 (5.3) |
FEV1b predicted (%), mean (SD) | 53.9 (19.9) |
6MWDc (m), mean (SD) | 442 (107) |
6MWD predicted (%), mean (SD) | 70.3 (16.5) |
CATd score, mean (SD) | 13 (8) |
QFe (kg), mean (SD) | 31.5 (10.9) |
PAf (steps per day), median (P25-P75)g | 4272 (2783-5768) |
aBMI: body mass index.
bFEV1: forced expiratory volume in 1 second.
c6MWD: 6-minute walking distance; 6MWD was missing in 2 patients.
dCAT: chronic obstructive pulmonary disease (COPD) assessment test.
eQF: quadriceps force; QF was not measured in 2 centers and QF was missing in 27 patients.
fPA: physical activity; valid PA measurements were present in 140 patients.
g25th and 75th percentiles (P25-P75).
Boxplots depicting the usefulness score (0-10 Likert scale) of the different parts of the intervention from the patients’ perspective. “app” between brackets represents messages displayed on the smartphone app.
In total, 91.2% of patients (145/159 of the completers sample) took part in the semistructured interviews at V3. Themes and subthemes that were derived from the verbatim responses of patients to the interview are presented in
Themes of (1) positive experiences and (2) issues or problems emerged from the data.
Positive experiences No technical problems: A large portion of patients stated not to have encountered technical issues with any of the components of the intervention. Working with app: The ease of use with the different components of the intervention was highlighted by patients. Furthermore, patients who had less a priori experience with managing a smartphone device expressed that the learning process of working with this device was smooth.
Issues or problems Help from others: Few patients needed more than a familiarization period before they were able to feel confident about working with the smartphone and its app. Help from both the study team (through phone calls or face-to-face contacts) and from their relatives was considered essential when experiencing problems. Speed of interaction with the app: Some patients felt the speed of the app was slow and perceived the interaction with it as time consuming. Especially, the transfer of step data onto the phone in the evening was delayed for several minutes. App problems: Some patients reported during the interview that working with the app was often hindered (eg, tasks not opening and not possible to send data). Reasons for these app problems were mostly related to issues with the internet connection or Bluetooth problems. Step counter: A small minority of patients expressed their frustration with the step counter that was not always able to detect all steps they performed. Activities such as slow walking, cycling, and arm movements were not measured accurately.
Themes of (1) positive experiences, (2) issues or problems, and (3) outcome emerged from the data.
Positive experiences Step counter: The step counter was judged as the essential part of the intervention by several patients because of its simplicity, feedback, and usefulness. Graphs: Another highly rated aspect of the intervention was the graphical feedback displays that patients received based on the achievement of their goals. According to the patients, it was an interesting and excellent way of motivating them. Nice experience: In general, the intervention was considered as motivating to a large majority of patients across the different centers. Patients claimed it was a fun and interesting experience that helped them toward being more active and feeling better and fitter. Being monitored: One of the most important motivational reasons according to patients to become more active was the feeling of being monitored. Knowing that the coaches were following them up gave them an external motivational cue to be physically active. Family participation: Next to the help from the coaches, patients’ relatives often played an important supportive and stimulating role throughout the intervention. Close relatives of patients (mostly spouses) also bought a step counter to join their wife or husband throughout their coaching.
Issues or problems Goals: One of the most important issues was the increase in the step count goal, which was often too high for patients. This caused some frustration among patients as it was perceived as demotivating to have too high goals and not being able to reach them. Variations: As the intervention was used for a period of 12 weeks, the component of variation in the content of the app was deemed as important according to the patients. Some patients reported that because of the lack of variation, their actual usage of the intervention (in particular with the opening of the messages on the smartphone) lowered. The morning messages with the goal patients needed to achieve were repeated every day of that week and required more variation according to the patients. Barriers: One of the major drawbacks of the intervention according to patients was that it did not take into account several barriers with which they were confronted. When a patient experienced an acute exacerbation, his or her goal was not adjusted immediately. Weather factors were not taken into account within the app. Furthermore, patients regretted that there was no option for them to make the intervention aware that they had other priorities (eg, holidays or days when they needed to watch their grandchildren). Motivational issues: A few patients did not find the app interesting and did not like working with it.
Outcomes New routine: Patients stated that the intervention and the goals resulted in the adoption of new lifestyle routines to be more physically active. They hoped to continue with these more active lifestyles after the intervention finished.
All coaches present at the focus group considered the intervention to be a useful addition to standard care in patients with COPD. The coaches rated the step counter as very useful, mainly attributed to the direct feedback it provided and its ease of use. Technical problems with the smartphone interface intermittently occurred (eg, Bluetooth connection or requests for automatic updates). In addition, coaches reported that a minority of patients felt the smartphone app lacked variation. Considering future long-term use, coaches proposed a more individualized technical training based on individual patient needs (eg, more extensive in patients with difficulties and those needing more contact time). Finally, the coaches regretted that the home exercises did not result in higher step counts and lacked variation, which might explain the low use of the home exercise booklet by patients.
Almost 60% (59.7%, 95/159) of patients wore the step counter for more than 90% of the days they were included in the coaching program, representing a median (P25-P75) of 6.3 (5.8-6.8) days per week with no difference over time within the trial (
In terms of self-reported actual usage, a large majority of the patients (76.7%, 122/159) stated that they looked
Coaches performed 1053 out of the 1161 contacts that appeared on the platform; however, no details on the time of solving the tasks were available.
Feasibility from the perspective of the patients was good as a large proportion of patients reported that the smartphone intervention was not too much of a burden to work with when they were asked how they had experienced the technical aspects of the intervention. Coaches spent significantly more time (
All PA coaches present in the focus group reached consensus that a follow-up of approximately 25 to 30 patients simultaneously for 1 coach would be feasible. It was felt to be beneficial to have 1 coordinating center to discuss day-to-day problems in patient management on a case-by-case approach.
Overview of the different components of the intervention. Definition of actual usage of the different components of the intervention of all completers (n=159 patients) and the minimum and maximum values one can achieve in terms of actual usage were reported when applicable. Actual usage and possible minimum-maximum are expressed as median (P25-P75) days per week for the step counter and the daily tasks on the app. Weekly feedback is expressed as median (P25-P75) percent of weeks in the intervention.
Components of the intervention | Actual usage | |||
|
Definition of actual usage | Median (p25-p75)a | Possible minimum-maximum | |
One-to-one interview with coach discussing motivation, barriers, favorite activities, and strategies to become more active | N/Ab | N/A | N/A | |
Step counter (Fitbug Air; days per week) | A day with ≥70 steps recorded | 6.3 (5.8-6.8) | 0-7 | |
|
||||
|
Send activity data task (days per week) | Patient closes task | 4.1 (2.4-5.6) | 0-7 |
|
Looking to the daily goal task (days per week) | Patient closes task | 4.1 (2.1-5.9) | 0-7 |
|
Looking at the daily feedback task (days per week) | Patient closes task | 2.2 (0.7-4.1) | 0-6 |
|
Looking at the weekly feedback task (% of weeks in the intervention) | Patient closes task | 55 (29-78) | 0-100 |
A booklet containing home exercises | N/A | N/A | N/A | |
Weekly group text messages with activity proposals sent by the coach | N/A | N/A | N/A | |
Contact with the coaches, which was triggered in the case of nonactual usage with wearing the step counter, failure to transmit data, or failure to progress | N/A | N/A | N/A |
a25th and 75th percentiles (P25-P75).
bN/A: not applicable.
Contact time throughout the intervention (only including centers with more than 20 patients). The black bars represent the mean contact time (in min per week) per patient from the first 10 patients that were recruited in each center. White bars represent the mean contact time (in min per week) per patient from the patients that were recruited at a later stage.
Patients in the low (n=49), medium (n=46), and high (n=45) contact time group had a median (P25-P75) total contact time of 25 (10-30), 50 (40-60), and 140 (105-185) min, respectively. Patients who had more contact time with the coaches during the time of the study, had more severe airflow obstruction, tended to have a lower functional exercise capacity (
When groups were divided in 3 according to their overall actual usage score, neither patient characteristics nor effectiveness were different (see
Logistic univariate regression analysis revealed that patients with a low contact time (≤30 min; n=103) after 4 weeks were 3.58 times more likely of achieving the MID improvement of 1000 steps per day (95% CI 1.88-6.82;
Patient baseline characteristics according to the total contact time (only including patients with valid PA measurement; n=140); data are expressed as mean (SD) unless stated otherwise.
Variables | Low contact time (n=49) | Medium contact time (n=46) | High contact time (n=45) | |
Age in years, mean (SD) | 65 (7) | 65 (10) | 68 (6) | .16 |
Gender (male), n (%) | 28 (57) | 34 (74) | 27 (60) | .20 |
BMIa (kg/m2), mean (SD) | 27.8 (5.3) | 26.1 (4.4) | 27.0 (6.4) | .35 |
FEV1b predicted percentage, mean (SD) | 59.5 (22.6) | 54.1 (16.5) | 49.1 (20.5)j | .04 |
6MWDc (m), mean (SD) | 444 (100) | 459 (101) | 411 (113) | .09 |
6MWD predicted percentage, mean (SD) | 71.5 (14.5) | 71.2 (15.0) | 67.4 (19.6) | .29 |
CATd score, median (p25-p75)e | 10 (6-17) | 13 (7-19) | 16 (10-21) | .11 |
QFf (kg), mean (SD) | 33.1 (13.2) | 31.2 (10.0) | 29.2 (10.5) | .33 |
PAg (steps per day), median (p25-p75) | 4542 (3387-5587) | 4377 (3016-6723) | 3186 (2375-5339) | .15 |
Contact time first 4 weeks in minutes, median (p25-p75) | 0 (0-5)h | 10 (5-20)i | 50 (20-85)j | .005 |
aBMI: body mass index.
bFEV1: forced expiratory volume in 1 second.
c6MWD: 6-minute walking distance; 6MWD was missing in 2 patients.
dCAT: chronic obstructive pulmonary disease (COPD) assessment test.
e25th and 75th percentiles (P25-P75).
fQF: quadriceps force; QF was not measured in 2 centers and QF was missing in 27 patients.
gPA: physical activity.
hIndicates statistical significance (
iIndicates statistical significance (
jIndicates statistical significance (
Change in physical activity (PA; mean [SE]) across groups of patients according to total contact time; adjusted for age, baseline functional exercise capacity, baseline forced expiratory volume in 1 second, baseline symptom score and number of acute exacerbations in the previous 12 months.
Patient characteristics according to the total actual usage score (3 groups only including patients with valid physical activity measurement by actigraph, n=140); data are expressed as mean (SD) unless stated otherwise.
Variables | Low actual usage, <47% of usage (n=47) | Medium actual usage, 47% to 75% of usage (n=46) | High actual usage, >75% of usage (n=47) | |
Age in years, mean (SD) | 66 (8) | 66 (9) | 65 (8) | .76 |
Gender (male), n (%) | 31 (66) | 29 (63) | 29 (62) | .91 |
BMIa in kg per m2, mean (SD) | 27.5 (5.3) | 27.6 (6.5) | 26.0 (4.3) | .34 |
FEV1b predicted percentage, mean (SD) | 54.4 (20.3) | 55.2 (19.5) | 53.5 (21.6) | .92 |
6MWDc (m), mean (SD) | 431 (106) | 432 (105) | 454 (107) | .50 |
6MWD predicted percentage, mean (SD) | 69 (17) | 69 (17) | 72 (16) | .61 |
CATd score, median (p25-p75)e | 14 (7-19) | 13 (6-19) | 12 (7-21) | .94 |
QFf (kg), mean (SD) | 32.0 (10.8) | 30.0 (12.9) | 31.1 (9.4) | .73 |
PAg (steps per day) median (p25-p75) | 4369 (2868-5672) | 3850 (2380-6108) | 4540 (2940-6731) | .49 |
aBMI: body mass index.
bFEV1: forced expiratory volume in 1 second.
c6MWD: 6-minute walking distance; 6MWD was missing in 2 patients.
dCAT: chronic obstructive pulmonary disease (COPD) assessment test.
e25th and 75th percentiles (P25-P75).
fQF: quadriceps force; QF was not measured in 2 centers and QF was missing in 27 patients.
gPA: physical activity; valid PA measurements was present in 140 patients.
Change in physical activity (PA; mean [SE] across groups of patients according to overall actual usage score; adjusted for age, baseline functional exercise capacity, baseline forced expiratory volume in 1 second, baseline symptom score and number of acute exacerbations in the previous 12 months.
On the basis of the secondary analysis of the MrPAPP PA telecoaching trial in patients with COPD, this mixed-methods research design study shows that the intervention was feasible and well accepted by both patients and coaches. Given the design of the intervention (ie, patients were contacted when PA was not increasing), patients with high contact time with coaches had less PA improvements, suggesting that the high contact time resulted from either difficulty or reluctance to engage in PA. Furthermore, we observed that the overall level of actual usage with the program components in the entire group did not influence the intervention effect.
The intervention had good acceptability for patients who rated their satisfaction in line with previous PA telecoaching research in a mixed COPD and diabetes type-2 population [
The smartphone app was also well received by patients although to a lesser extent than the step counter. This was associated with a considerably lower actual usage score of patients for the smartphone intervention compared with the step counter. Several factors may explain this relatively lower actual usage. First, a proportion of patients with COPD who owned a smartphone before the study might have caused less fluency with the smartphone (
In literature, mixed results and high heterogeneity are reported on the actual usage with PA coaching Web portals or smartphone apps. During a 4-month, internet-based PA telecoaching program, veterans with COPD logged into the website and uploaded their daily step counts for 5.7 days per month which decreased to 3.0 days per month over a follow-up of 12 months [
Components of the intervention that were not individually tailored (eg, educational activity tips and home exercise booklet) were rated as less useful. This confirms patients’ self-reported actual usage of the home exercise booklet, which was low and is in line with findings from the focus group, in which PA coaches pointed out that the home exercise booklet was not individualized for each specific patient. This highlights the importance of introducing personalized components within PA telecoaching, which has also been suggested in patients with ischemic heart disease who participated in a mobile health cardiac rehabilitation intervention [
In line with the patients, the coaches expressed good acceptability of this PA telecoaching program. On future use of the intervention, coaches reached the following consensus:
1. “The goal of such a PA telecoaching intervention should be that patients are able to use this intervention quasi independently indefinitely. Every 6 months patients could come for a follow-up visit, synchronized with other planned health visits to the outpatient clinic.” Interestingly, our data suggest that 3 months of coaching might be enough for patients to reach a plateau in PA increase (see
2. “As their PA coach it is our task to provide further follow-up by giving them the step counter and occasional phone calls for follow-up.” Such strategies merit further validation, but the statement strengthens the importance of acceptability, actual usage, and feasibility with long-term PA telecoaching programs in this patient population. In addition to the latter perspectives, the coaches highlighted that it is highly important that the preferences and experiences of the patients with the intervention are assessed and taken into account when looking at future implementation. Therefore, future (long-term) PA telecoaching interventions need to ensure whether enough variation within such apps is introduced in addition to those components deemed as the most essential to patients (ie, step counter and contact with the study team). Furthermore, such interventions need to take the occurrence of acute exacerbations into account and involve patients’ relatives as these can play an important role as social support in being physically active [
In terms of coach feasibility, the main paper of the MrPAPP trial revealed that patients were contacted for a median of 50 min throughout the 12 weeks intervention [
Literature about the relationship of both actual usage by the patients and coach feasibility (contact time) of the intervention with the change in PA in telecoaching trials is scarce. In this study, the degree of the overall actual usage score (including wearing the step counter and all the app tasks) was not associated with the effectiveness of the intervention. This is in contrast to a 4-week pilot (telecoaching) study which showed a positive relationship between the degree of actual usage of wearing a smartphone-based activity coach and the benefits from the intervention during the first 2 weeks albeit this association disappearing during the third week [
In contrast to actual usage, the contact time between the coach and patients was associated with the effectiveness of the intervention, that is, a lower effect in those patients in need of more contact time. These patients were the more severe (ie, they have more severe airflow obstruction and tend to have a lower functional exercise capacity) and are more likely to experience exacerbations and therefore, have more chance of triggering coaching-related and/or health-related contacts with their coach. As contact time remained a significant, negative predictor of the change in PA, independent of the patient characteristics, this may point to the inability of some patients to work with the coaching app. This corroborates with the findings of the qualitative part of the study and should not be ignored as a reason for treatment failure. In clinical practice, we would therefore advocate flexible use of these interventions where patients are diverted to other interventions (eg, more supervised exercise programs such as pulmonary rehabilitation) if contact time accumulates. This is important for stratification in future trials.
To the best of our knowledge, this study is the first providing an in-depth analysis of the acceptability, actual usage, and feasibility with a PA telecoaching intervention developed for patients with COPD. Our study is unique as it allows us to investigate these aspects, relating them to physiological characteristics along with the level of response.
The results are based on a combination of quantitative and qualitative research, including information coming from patients as well as from coaches. In addition, the study is performed on the back of a properly powered randomized controlled trial, which was characterized by a comprehensive physiological assessment and objective assessment of PA. Furthermore, this PA telecoaching intervention consists of several behavioral principles (including but not limited to facilitating goal setting, action planning, feedback, and problem solving) which were based on the behavior change technique taxonomy of Michie et al [
First, we only included patients that completed the trial. This could have resulted in a selection bias. Coaches might have spent more time in those patients who subsequently dropped out during their intervention period. However, as only 7% (12/171) of patients discontinued, this is unlikely to have had a large impact on the results. Second, no multiple-comparison post hoc corrections were applied in the quantitative data analysis as these analyses should be regarded as exploratory and in need of independent confirmation. These results help to guide future research; however, they may not be taken as a final judgment and should be interpreted with caution due to the latter limitation. Third, only 1 focus group with a limited number of PA coaches was performed. Therefore, data saturation could not have been reached. Another focus group with participants with a broad background and experience would have been of great value for (1) external validity of findings and (2) to ensure data saturation. Nevertheless, coaches were asked during the focus group whether they had additional comments. In addition, a summary of the focus group was sent to the coaches who could not be present at the focus group for completion of the summary. New themes emerged, which allowed for more data capturing. Fourth, we did not specifically assess capabilities or history of patients with managing the smartphone device or their expectations. In hindsight, this might have provided even more detailed information to predict the therapeutic response to the PA telecoaching intervention. Fifth, for the assessment of acceptability of the intervention, we used a project-tailored questionnaire. In literature, several attempts have been made to measure the quality of mobile health apps; however, no measure from a user perspective has been widely accepted [
In line with general findings of the present behavioral modification program [
This 12-week PA telecoaching intervention was well accepted and feasible for both patients with COPD and their coaches. Actual usage of the step counter was excellent, whereas actual usage of the smartphone tasks was lower and decreased over time. Overall actual usage was not associated with the effect of the intervention. The step counter and direct contact with the coach were perceived as the most useful components of the intervention by the patients. Patients with more need for contact had more severe airflow obstruction, tended to have more severely limited exercise capacity, and experienced less PA benefits. Alternative strategies (including more face-to-face contacts and offering pulmonary rehabilitation programs) might be more effective in these patients.
Home exercise booklet.
Linkcare application platform.
Project-tailored patient satisfaction form.
Patient interview (discussion guide).
Focus group (coaches).
COREQ checklist.
Illustrative quotes interviews.
Actual usage of the different intervention tasks according to gender and age.
Overview of (mean ± SE) steps performed per week by all patients (step counter data).
chronic obstructive pulmonary disease
COnsolidated criteria for REporting Qualitative research
forced expiratory volume in 1 second
Global Initiative for Chronic Obstructive Lung Disease
mobile app rating scale
minimal important difference
multicenter physical activity telecoaching trial
modified Medical Research Council
physical activity
6-minute walking distance
The authors would like to acknowledge Claudia Perez for providing data from the Linkcare app, Maarten Spruit (MS) and Ane Arbillaga-Etxarri (AAE) for their contribution with the data collection. The PROactive project is funded by the Innovative Medicines Initiative Joint Undertaking (IMU JU) #115011. The Leuven study group was supported by the Flemish Research Foundation (grant # G.0871.13). HD was the recipient of a joint ERS/SEPAR Fellowship (LTRF 2015) and is a postdoctoral fellow of the FWO-Flanders. ZL was the recipient of a European Respiratory Society Fellowship, grant number LTRF 2016-6686 and is a postdoctoral fellow of the FWO-Flanders (Fellowship number 12U5618N). FMR is funded by The National Council for Scientific and Technological Development (CNPq), Brazil (249579/2013-8). The Zurich study group was supported by an additional grant of the Lung League Aargau (nonprofit organization) as well as by Swisscom AG who provided 30 sim cards and data usage of up to 1 GB per month. MIP’s contribution to this work was supported by the National Institute for Health Research (NIHR) Respiratory Biomedical Research Unit at the Royal Brompton and Harefield National Health Services (NHS) Foundation Trust and Imperial College, London UK who part fund his salary. The views expressed in this publication are those of the authors and not necessarily those of the NHS, the NIHR or the Department of Health. The sponsors did not have any influence on the design, conduct, and analysis of the study. ISGlobal is a member of the CERCA Programme, Generalitat de Catalunya.
RR, HD, ZL, AF, CDJ, IV, JGA, MIP, Maarten Spruit (MS), RT, NH, and TT contributed to the study protocol and development of the intervention. ML, RAR, HD, ZL, RT, NR, AF, CDJ, SCB, GB, AS, and IV contributed to the data collection. ML, HD, and TT contributed to the data analyses, interpretation of the data, and the writing of the paper. RAR, HD, ZL, RT, NR, AF, CDJ, EGS, FMR, SB, NH, GB, AS, IS, IV, JGA, MP, and TT critically reviewed the paper. TT is the guarantor of the study. All authors had full access to all the data in the study and can take responsibility for the integrity of the data and the accuracy of the data analysis.
None declared.