Improving Adherence and Quality of Life of CPAP for Obstructive Sleep Apnea With An Intervention Based On Stratication and Personalization of Care Plans: A Randomized Controlled Trial

Purpose The aim was to determine adherence to CPAP and health-related outcomes in patients with OSA via a comprehensive program based on stratication and individualized care plans, including motivational interview. Methods A multicentre, randomized controlled trial (RCT) design was used in 3 hospitals. The control group followed the usual treatment. Intervention group (PIMA) followed the treatment with an adapted care plan (home visits, telephone, care-center) depending on socio-demographic, clinical and psychological variables, using motivational interviews in all the interventions. The main outcome was adherence, and secondary were quality of life, emotional state, activities, social relationships, perceived competence, and motivation. The clinical trial was registered with www.clinicaltrials.gov (NCT 04691479). 213 patients upper-secondary education) were randomized: 108 were in the intervention group and 105 were in the control group. The result showed a statistically signicant difference in adherence at 90 and 180 days respectively: 129.24 (IC 95% 77.25-181.22) p < 0.0001 and 288.30 (IC95% 187.146–389.47) p < 0.0001. The adherence (hours/day) showed favourable results for the PIMA group compared to the control group at 90 days with a difference of 1.74 hours/day (IC95% 1.18–2.30) p < 0.0001 and at 180 days with a difference of 2.31 (IC95% 1.72–2.91) p < 0.001. The results for secondary outcomes also showed statistically signicant differences in favour of the PIMA group. This study found evidence that a program based on stratication and personalized care plans, through motivational interviewing, improves adherence to CPAP treatment and health-related outcomes than traditional care.


Introduction
In obstructive sleep apnoea/hypopnoea syndrome (OSA), recurring episodes of obstruction are observed during sleep and are associated with daytime sleepiness [1]. The severity of this syndrome is assessed according to the number of events per hour of sleep, using the apnoea-hypopnoea index (AHI). OSA occurs in up to 24% of men [2], 7% of which experience daytime sleepiness, while OSA occurs in only 5% of women 2 . Lack of treatment is related to an increased risk of accidents, loss of quality of life and cardiovascular mortality [3].
Continuous positive airway pressure (CPAP) is the rst-line treatment for moderate-to-severe OSA [4]. Most research considers a patient who uses CPAP a minimum of 4 hours a night during 70% of the week to be a good complier [5].
However, it is to a certain extent obvious, and it is also endorsed by studies [6], that 6 hours would be an excellent metric for adherence to optimize symptoms and obtain a clear reduction in blood pressure in hypertensive patients.
Multidisciplinary approaches based on the study of behavioural changes together with continuous monitoring have shown promising results in improving patient adherence to CPAP intervention [7]. The current de nition of adherence is largely arbitrary and is mainly used by third-party payers to determine CPAP reimbursement but CPAP adherence remains sub-optimal [8].
Patients may bene t from more individualized treatment, which can be done by reducing the risk of OSA consequences on a patient-to-patient basis. New strategies to augment adherence, especially early in the course of CPAP treatment, are needed in the management of OSA [9,10].
Motivational interviewing (MI) is a technique that involves discussing treatment with patients in a way which empowers them to understand their individual therapy needs and drive them to address those. Different studies have demonstrated the usefulness of MI in the CPAP treatment of sleep apnea [11,12]. Therefore, it is very likely that MI, in combination with personalized patient interventions, could further improve adherence.
The aim of this study is to determine whether OSA patient adherence to CPAP can be improved using a comprehensive, multidisciplinary programme, based on a motivational change model and integration of predictive factors of adherence, to obtain a speci c patient pro le and therefore a personalized intervention plan.

Materials And Methods
Design A multi-centre randomized, controlled trial (RCT) design was used, enrolling adult patients with OSA and a prescription for CPAP treatment from the Respiratory Medicine Department at La Princesa University Hospital

Patients
The subjects enrolled in the study were required to have a diagnosis of OSA con rmed by sleep studies with polygraphy and/or polysomnography (criteria of the Spanish Sleep Society) and be naive to CPAP intervention. The study excluded subjects with obesity-related hypoventilation, severe COPD, cognitive disorders and those unable to understand the consent to participate. Prior to enrolment in the study, all patients were informed in detail about the study and signed the consent form to participate.

Interventions
A process was followed using a random number generator in SPSS to systematically assign patients to one group or another, with allocation concealment. Prescribers were responsible for randomization. See Fig. 1.
The control intervention group patients followed the standard of care. The therapy started in the hospital. The nurse performed training in the use of CPAP equipment, mask adjustment, safety, and maintenance. For follow-up, the patient was always referred to the Hospital, with a frequency established by the Spanish Society of Pulmonology and Thoracic Surgery (Days 30, 90 and 180). The follow-up procedure consisted of reviewing the CPAP hour-meter and resolving any, with the necessary corrective actions.
The intervention group 'PIMA' (Spanish acronym: Plan Individualizado de Mejora de la Adherencia; Personalized Adherence Improvement Plan), began the treatment with the educational and training program "MEntA" (Motivational Interview for Adherence) [12]. It consisted of one session with two blocks: educational activity and training activity. Also, the key to all patient interactions was the use of MI. In fact, the nurse used MI not only at the beginning of therapy, but also in every contact that the patient had with the patient throughout the treatment (subsequent follow-up visits, phone calls, etc.).
Subsequently, the nurse performed the strati cation process, which determined the personalized intervention plan using two types of variables: personal and modulation. The variables of the strati cation established the complexity and intensity of adaptations for the patient's treatment needs. Personal variables consider characteristics of the patient such as: age, level of studies, clinical (somnolence and apnoea-hypopnea index), and psychological (motivation and perceived competence). Modulation variables refer to characteristics ranging from the patient's environment (isochrone, which is the time the patient spends from home to the care center) to service options and preferences (for example, use of electronic applications, email, or visit preferences). The results are patients with the "a" pro le (characterized by autonomy and mobility, predisposition to remote-controlled follow-up), patients with the "b" pro le (need for more intensity in follow-up), or patients with the "c" pro le (more di culties to move around and require more intensive treatment). Pro le "d" was applied to patients who were professional drivers, as they require speci c interventions based on their occupation.
Regarding psychological variables for the evaluation of perceived competence, the Perceived Competence Evaluation Questionnaire validated in Adherence to CPAP in OSA (CEPCA) was used [13]. The scores obtained indicated the patient's feelings for self-effectiveness with the CPAP treatment. Motivation was also assessed, using the methodology de ned in the Prochaska and DiClemente transtheoretical model [14], with an open-ended question to the patient which could be classi ed in ve stages: precontemplation (very low motivation), contemplation (low motivation), determination (some motivation), active change (quite motivated) and maintenance (high motivation). Regarding the clinical variables, drowsiness was obtained through the administration of the Epworth Somnolence Test [15], and the apnoea-hypopnea index was taken from the patient's clinical history.
The personal and modulating variables were used to identify the degree of intensity the patient needed (channels and frequency of contact). The psychological and clinical variables obtained in this rst visit were used as "predictive" information for patient adherence (either high, moderate or low). Subsequently, the care plan was initiated based on this information. Low or moderate adherence care plans were more intensive than high adherence relative care plans. In Table 1 we can see how the patient's pro le (a, b, c or d) and the level of adherence determine the care plan. In follow-up visits, psychological and clinical variables were reviewed and the patient's adherence was added (the pro le always remains the same, since their personal characteristics are generally stable). Based on the level of adherence, the care plan was adapted if the patient's situation had changed. For patients with low adherence, telemonitoring was used to measure hours of adherence, AHI and air leaks. Once the follow-up procedure with telemonitoring of 1 month had started, every 10 days, the patient evolution was reviewed. The level of adherence determine the frequencies of interventions, taking into account short term ( rst 90 days of treatment) and after day 90. This con guration allows, on the one hand, the dedication of resources to patients who need more care, and on the other hand, the long-term management of large volumes of patients.
Several outcomes were included in this study. First, adherence to CPAP therapy after 90, 180 and 365 days of treatment was assessed as the primary outcome. The total number of night-time hours during which the CPAP device was used at a therapeutic pressure was recorded by speci c counters within each device and were collected during the study follow-up period. In terms of secondary outcomes, quality of life was assessed using the Visual Analogical Well-being Scale for apnoea [16] and sleepiness was assessed using the Epworth Sleepiness Scale [15].
Motivation was also assessed, as de ned in the transtheoretical model [14] using an open-ended question to the patient, which could be classi ed in the ve stages. Moreover, three health related outcomes (mood, activities, social relationships) were also measured using the following ad hoc question "Taking into account your sleep problems, how would you say you are in terms of mood/activities/social relationships?". The answer alternatives were: Good, Normal/no change, Bad.

Statistical analysis
The sample size was determined based on a previous moderate effect size of .06 because there was no prior literature available with similar characteristics to those proposed in this study. As such, with an alpha value of .05 and a maximum beta error of .20, a total number of 70 subjects were established, which was increased considerably in order to compensate for possible experimental mortality in the study and to reduce the variability of the measures being studied.
Statistical analysis is based on intent-to-treat (ITT) analysis of the participants. A descriptive analysis of the demographic and clinical characteristics of all patients enrolled in the study was performed, as well as a baseline analysis comparing both study groups after randomization. The statistical inference analyses were performed using both parametric (t-student for independent groups) and non-parametric (U-Mann Whitney) models for continuous variables, and crosstab analysis for categorical data. The analyses were performed for the comparison of study groups at two cut-off time points (90 and 180 days) and an ad hoc analysis at one year of follow-up for the most important variables. The Chi-square statistic was used for comparing all categories at one year adherence analysis. The alpha value was 0.05 and all analyses were performed on two-tails. The statistical software used was SPSS V26.

Descriptive statistics
A total of 213 patients were randomized between the two groups (PIMA and Control) (Fig. 2) of whom 108 were in the intervention group and 105 were in the control group. 68.1% were men, 58.7% of the study sample were aged between 51 and 69, and 45.5% had upper-secondary school education. Tests performed to compare control and PIMA groups at the beginning of the RCT showed that groups were equivalent for all baseline variables. Table 2 provides descriptive characteristics of the participants. A total of 9.85% (21 subjects) withdrew from the study. The percentage of withdrawals in the PIMA group was 10.18% (11 subjects) while in the control group it was 9.52% (10 subjects). In the PIMA group, the reasons for this were: 8 subjects for clinical reasons (changes of treatment), 2 lost, 1 abandoned. In the control group, 7 patients withdrew because of voluntary abandonment and 3 patients were lost. None of the withdrawals were due to possible adverse effects.
Inferential statistics results. patients in the control group did not result in increased average sleep per day between these time points. This was also con rmed by non-parametric analysis: p < 0.01. Table 3 for complete results. An ad hoc analysis at one year of follow-up still showed statistically signi cant differences between the PIMA and control groups for the main variables (Tables 3 and 4).80% of patients were compliant at day 365 in the PIMA group compared with 66.3% in the control group. The improvement in overall adherence to CPAP therapy in the PIMA intervention group is clearly demonstrated. At Day 365, the % of non-compliers in the PIMA and control groups were 20% and 33.7% respectively. The decrease in non-adherence can also be observed at earlier time-points during the study, where the % of non-compliers at Day 90 and D180 were 11.8% compared with45.7% and 18.2% compared with 38.9% in the PIMA and control groups respectively. The largest impact of the PIMA intervention on the improvement of adherence to CPAP therapy is observed in the group of subjects with adherence < 2h/day, more so than any higher adherence time category. This is also consistent across all time points.

Discussion
This study has found evidence that a programme based on strati cation labels for patients to obtain personalized care plans, achieves a higher level of e cacy in adherence to the CPAP intervention, as well as an improvement in quality of life, emotional state and motivation, when compared to the standard of care intervention. The data observed in this study shows how an approach that strati es OSA patients based on objective measurements is highly useful in determining speci c plans according to subject pro les and thereby improves adherence with the CPAP intervention.
The results obtained through clinical trials, which have been widely discussed, range from increased number of hours of sleep, quality of life and social relationships, to improvements in cognitive performance and reduced mortality [17]. However, the e cacy of CPAP interventions depends on proper adherence and given its speci c conditions of use, this aspect still requires a signi cant level of research, since it has been observed that, in the past 20 years, the percentage of subjects who do not manage to reach 7 hours of night-time sleep while using this intervention has remained constant at around 34% of subjects [18].
The results of our research show that the strati cation procedure, based on which a personalized adherence improvement plan is put in place, is able to maintain adherence with the therapy for up to 180 days compared to the usual practice for monitoring this type of patient undergoing CPAP intervention. The PIMA programme improves therapeutic adherence by raising patient awareness and adjusting care to consider changes in their lifestyle and technological advances such as telemonitoring [7,11,19]. Previous studies have shown the importance of examination before prescription, due to the in uence of anatomical differences in terms of adherence [20]. It has also been observed that the success of long-term adherence is conditioned, to a large extent, by the patient's behaviour during the rst two weeks [21] together with the immediate social support they receive [22]. Similarly, the use of new technologies, such as telemonitoring enables early intervention with alerts to complications or changes in use [23]. However, the indiscriminate use of telemonitoring does not provide bene ts in terms of adherence, and it seems to be more indicated in situations of early detection or identi cation of barriers to adherence and to certain pro les that present a speci c clinical picture and speci c situation [24].
An important aspect is the cost-bene t impact of an intervention based on personalization of care. PIMA allows sustainability since with strati cation and segmentation, it is identi ed which patients need more attention, without forgetting the long-term treatment of the rest of the patients.
As a patient's pattern of adherence is evident within the rst weeks of CPAP treatment, it is highly likely that patients have already formed perceptions regarding the seriousness of OSA and the bene t of treatment, and thus these cognitions in uence adherence [23]. Education is one way to enhance self-e cacy. MEntA, the educational program used by the PIMA programme is based on the motivational interview, for which there is also extensive evidence on its e cacy in health interventions [19,25]. The factors which the aforementioned studies determine as being predictive of proper adherence are complemented in the PIMA model by the Transtheoretical Model of Change, in its version adapted to sleep apnoea treatment, from the initial version created to predict smoking cessation [26].
Studies have observed that beliefs related to the pathology or e cacy of the treatment, together with the perception of self-e cacy in terms of con dence in starting the treatment and in beliefs regarding proper interaction with the environment, explain an up to 30% difference in the use of the CPAP intervention [27]. A high level of adherence is associated with an internal locus of control [25] (subjects who believe they are in control of their own situation are more likely to take on board the advice provided by the physician) increasing adherence to CPAP [28]. Motivated patients have a higher level of adherence to CPAP [29]. This also applies to subjects with social support, with an improvement in sleep quality and quality of life also being observed in the person sharing a bed with the patient [30]. All of these predictive factors are included in the framework of the PIMA programme and were used to determine the label for each patient and the degree of personalization included to monitor them.
The PIMA intervention in our study was based on a broad, multidisciplinary framework with the aim of personalizing the approach needed by each patient in terms of improving adherence to CPAP. This intervention strives to change the patient's perception of the di culties and drawbacks of the proposed therapy, as well as their possible risks and ability to use it. To this end, these results not only show an improvement in the adherence, but also show statistically signi cant correlations between this adherence and positive correlates such as quality of life and activities. Statistically signi cant relationships were observed between adherence and patients-reportedoutcomes (PROs). These variables are highly related to the e cacy of the CPAP therapy.
The limitations of this study include the study duration, which, although at 1 year is longer than many studies, does not provide insight into longer-term changes and behaviour maintenance.

Conclusions
The PIMA model summarizes and quanti es the most relevant aspects to be considered in caring for CPAP therapy patients. It provides a speci c pro le for each patient aiming to provide personalized care with the help of new information technologies. Ultimately, this methodology improves OSA patients' adherence to CPAP therapy, thereby improving their quality of life and everyday living conditions. Improvement in adherence to CPAP therapy leads to an overall lower burden of disease and decreased costs associated with events arising from poorly-managed OSAS.

Declarations
Funding sources: Not applicable Con icts of Interest: The authors declare that there are no con icts of interest.
Availability of data and material: Data available on request from the authors  Consort diagram of participants randomized to PIMA and Control Groups Quality of Life at 180 days.