JMIR Publications

ABSTRACT

Background:

Combination therapies delivered remotely via the Internet or mobile devices are increasingly being used to improve and promote the self-care of chronic conditions. However, little is known about the long-term effects of these interventions. The aim of this study is to evaluate the effectiveness of a multimodal intervention program using a mobile device that measures associated variables such as catastrophizing, pain acceptance and quality of life in people with chronic pain in the outpatient setting. A total of 209 patients with chronic musculoskeletal pain were randomly assigned to one of the two study arms. The intervention group received a standard, interactive psychosocial therapy-type program of activities through a smartphone for 6 weeks. The control group only had access to the “Find out more” section of the application, which contains audiovisual material for pain management based on a self-help approach. Three months after treatment, significant differences were observed in the intervention group for the outcome variable of catastrophizing (PCS) (-0.59 vs 0.2; P = .0056) and the PCS subscale of helplessness (-0.65 vs 0.01; P = .07); rumination ( 1.23 vs -0.59; P = .04) and magnification (0.1 vs 0.86; P = .02). Pain management improved immediately after the CPAQ (0.38 vs 0.05; P = .0007) and EQ-5D (0.43 vs -0.01; P = .0016) treatments. The results suggest that multidimensional treatments reduce catastrophizing and improve quality of life and psychological flexibility immediately after treatment and that the effects on catastrophizing are maintained for at least three months in chronic pain patients.

Objective:

The aim of this study is to evaluate the effectiveness of a multimodal intervention program using an interactive smartphone or mobile device application to assess pain perception by means of associated variables such as catastrophizing, pain acceptance and quality of life in people with chronic pain. The study is based on the hypothesis that subjects assigned to the intervention group will exhibit less catastrophizing and emotional distress, more acceptance of pain and improve their ability to perform activities of daily living according to self-care values, as well as experience better quality of life and suffer less symptoms, pain intensity, anxiety and depression. These outcomes will be measured after completion of the intervention period and at three months follow-up and compared to a control group in line with the recommended outcomes in chronic pain research

Methods:

A randomized controlled clinical trial was performed using parallel treatment groups.Block randomization was used to ensure similar numbers of subjects in each group and intervention phase Mobile application-based multimodal treatment To develop the contents of the multimodal treatment for the mobile device application (app), a systematized literature review was carried out in a first phase following the Scottish Intercollegiate Guideline Network (SIGN) guidelines. As part of a broader research program aiming to improve pain management, we first located articles that describe the basic concepts of multiprofessional treatments for chronic pain, provide recommendations and include the published evidence for the selected topics of interest. The purpose was to identify the characteristics of pharmacological and non-pharmacological interventions for people affected by chronic pain in different settings by means of a bibliographic search and reading of the literature, a synthesis of the results and an assessment of the evidence. The second part of our study consisted in the prioritization of self-care recommendations for chronic pain selected from the reviewed literature using a participatory approach. Specifically, we designed a multimodal intervention protocol combining physical exercise, psychoeducational therapy, health assets and pharmacological treatment, which was delivered using a mobile device. The effectiveness of the intervention was subsequently evaluated by means of a clinical trial.38 The interventions delivered in the mobile app include ACT and mindfulness exercises to promote greater pain acceptance, reduce the aversive component associated with pain and help patients to dispassionately recognize and observe both the pain and related thoughts and emotions. Another group of activities aims to raise awareness of the individual’s own values through a series of activities to recover a meaningful life project. 23,24,39,40 The exercises section provides tools and resources for patients to improve their physical, mental and emotional well-being. The proposed activities in this section include empowerment, stretching, relaxation, walking or low-intensity exercises to help patients acquire good habits and learn about alternatives to improve their day to day life.41 The activities in the pharmacological section aim to help patients better understand medications that improve the intensity of pain. For each medication, the most common side effects and characteristics related to pain relief are described, as well as which drugs are best suited to the patient’s current health state, as well as the risks of taking more than the recommended dose of medication and how to identify warning signs. 42,43,44 The activities in the health assets section are designed to improve patients’ self-esteem and health by having them identify the individual, physical, institutional, associational, economic and cultural assets and resources available in their community to help them better cope with situations of vulnerability and stress.45,46 Development of the mobile application Our research team developed an app called NO + Dolor (NO + Pain) which contains an Android or iOS user interface. The application includes links to several multimedia resources (mainly audios and videos) and was designed based on game dynamics (gamification) to improve users’ concentration, attention and motivation. Indications on how to use the application and correctly perform the measures are provided in the instructions section of the app. As regards the technical characteristics of the app, the IT part required the design of a relational database implemented with MySQL Community Server 5.6. For the server side, a Java 1.8 communications API was implemented with RESTful architecture between the patients’ mobile apps and the database using a client-server pattern. For the client side, hybrid mobile applications were implemented for the Android and iOS operating systems. Apache Cordova 9.0 and jQuery Mobile 1.4.5 frameworks were integrated into the application and HTML5 and CSS3 technologies were used for the presentation layer (https://apps.apple.com/es/app/no-más-dolor-edcno/id1458096026). Participants The study population comprised residents in the municipalities belonging to the Cordoba South Health District of the Andalusian Health Service, Spain, who were registered in the health service’s user database (BDU). Patients were recruited from each of the district’s 11 primary care centers by two collaborators, a nurse and a physician with experience in the follow-up of chronic pain patients. The sample was drawn from the database using the Diraya electronic medical records. The criterion for inclusion was being attended by primary care physicians and nurses. The database search was carried out on 30 June 2019 and identified 297 patients diagnosed with chronic musculoskeletal pain, of which 205 were women and 92 were men. Accepting an alpha error of 0.05 and a beta error of 0.2 (statistical power of 80%) in a two-tailed test and estimating 15% loss to follow-up, n = 296 is required to detect a statistically significant difference between two means at 3 points in our outcome variable with an estimated standard deviation of 12. All participants were under pharmacological treatment for chronic musculoskeletal pain (analgesics, anti-inflammatory drugs, antidepressants, anticonvulsants or opioids) as previously indicated by their primary care physician and asked to sign an informed consent form. The inclusion and exclusion criteria for the subjects participating in this study are presented in Table 1. Ethical Aspects The study was approved by the Cordoba Research Ethics Committee of the Andalusian Public Health System. Informed consent was approved by the Cordoba Research Ethics Committee and completed by the participants. The study is registered at ClinicalTrial.gov (NCT04509154). Randomization and Blinding Block randomization with a block size of 4 was used. The only stratification criterion was the reference health center of the patient in question. An automated recruitment form hosted on the REDCap platform of the Maimonides Biomedical Research Institute of Cordoba (IMIBIC) was used to randomize the patients by simply clicking a button. The data were transferred and recorded in an electronic notebook using the Data Entry Manager (CDR) system. The statistician, the principal investigator and the co-investigators of the study who evaluated the results were not involved in patient recruitment and were blinded to group assignment. The twenty-two recruiters from the 11 primary care health centers of the Cordoba South Health District in 2019 who recruited the patients in 2019 were also responsible for randomizing the patients (i.e., clicking the button on the automated recruitment form) and were not blinded. Treatment Procedures All patients received a written invitation from their primary care physician or nurse to participate in the study. Two 8-hour face-to-face sessions were held at their reference health center led by one nurse and one primary care physician with experience in the follow-up of chronic pain patients. At the group meeting, the patients who agreed to participate voluntarily in the study were informed that they would receive instructions by email on how to download the mobile application with the contents of the treatment. The patients were also informed that if they were selected to participate in the intervention group, the treatment would last from 6 to 8 weeks. The control group would only have access to the “Find out more” section of the application, which contains audiovisual materials for pain management from a self-help approach, such as information on the origin of chronic pain and advice for pain treatment and relaxation. The intervention group received the treatment via their smartphones for a period of six weeks after completing the two face-to-face sessions. Pain intensity was measured daily on an 11-point numerical rating scale (NRS) when the participants accessed the application. To assess the treatments, self-reported questionnaires were sent to the participants by email and collected at three time points: at admission to the program (T1), at week six of the intervention (T2) and three months after the intervention (T3). The participants completed all three questionnaires at home and returned them by email. Smartphone-based Intervention The intervention consisted of the implementation of a program of standard, interactive psychosocial therapy activities. The pain management application enables automatic monitoring, skills training, social support, education, goal setting, and the achievement of four components: psychological wellness, exercise, pharmacological treatment and health assets. Each week, the participants received three activities for each of the above components via the NO + dolor app until completing the six weeks of treatment. All the activities are designed to be done weekly except for the walking challenge, which is done daily. The first time a participant completes an activity, they are awarded a star. Participants can do the proposed activity as many times as they like but no more rewards are given until the following week. The more activities they complete, the more stars they are awarded and the higher the percentage of goals reached by the patients each week (Appendix1). The application also has a “Consultation” section with a contact form where the participants can send any questions or comments. The form is then sent by email to the researchers so that they can respond to the inquiries. Assessment Measures The Spanish adaptation 47,48,49 of the Pain Catastrophizing Scale (PCS) 50 was used to measure the main outcome variable of the study. 51(51) The total score on the PCS is calculated by summing the responses to the 13 items and ranges from 0 to 52. The PCS subscales comprise three dimensions: a) rumination is scored from 0 to 16 (difficulty inhibiting repetitive pain-related thoughts and inability to seek solutions); b) magnification is scored from 0 to 12 (tendency to exaggerate distressing situations and negative aspects of pain and perceiving oneself as unable to control pain); and c) hopelessness/helplessness is scored from 0 to 24 (inability to cope effectively with pain. Higher scores indicate higher levels of catastrophizing. A score of 30 or more has been considered a cut-off point for clinically significant levels of catastrophizing. The Spanish version of the PCS has been shown to have adequate internal consistency (Cronbach’s α = .79), convergent validity and classificatory value, test-retest reliability (intraclass correlation coefficient = .84) and sensitivity to change in size (effect size >2). The Spanish adaptation 52 of the Chronic Pain Acceptance Questionnaire (CPAQ) was used 22 to measure engagement in life activities despite pain; the willingness to experience pain without trying to control, change or avoid it; the ability to recognize the chronicity of pain and the need to avoid or control pain. The CPAQ is a 20-item self-report questionnaire that rates pain acceptance on a 7-point Likert-type scale ranging from 0 (never true) to 6 (always true). The maximum possible score on the CPAQ is 120, with higher scores indicating higher pain acceptance. Initial studies on the acceptance and adaptation of the CPAQ have shown adequate internal consistency and the expected correlations with measures of physical functioning and psychological distress. Subsequent studies have evaluated the content and dimensions of the questionnaire and identified two factors: activity engagement (Cronbach’s α = 0.82) and pain willingness (Cronbach’s α = 0.78). 22 The Spanish adaptation of the EQ-5D 53,54 was used to measure health-related quality of life (HRQoL). This version can be used both in relatively healthy individuals (general population) and in groups of patients with different conditions. Individuals assess their own health state first by level of severity in different dimensions (descriptive system) and then on a more general visual analogue scale (VAS) of 0 to 100 (“worst imaginable health state” and “best imaginable health state,” respectively). A third component of the EQ-5D is the social values index obtained for each health state generated by the instrument, which describes respondents’ health state according to five dimensions: mobility, self-care, usual activities, pain/discomfort and anxiety/depression. Regarding the instrument’s psychometric properties, the test-retest reliability ranges from 0.86 to 0.90 and numerous studies have demonstrated its validity and sensitivity to change. 55,56 We included the question on subjective global improvement rated by the VAS from 0 to 100: “We would like you to indicate on this scale how good or bad your health state is today”. Pain intensity was measured using an 11-point NRS ranging from 0 (No pain) to 10 (Pain as bad as you can imagine). The format of this rating is established in the latest IMMPACT recommendations on core outcome measures for chronic pain clinical trials. 36 Data analysis A descriptive analysis was performed for the quantitative variables with mean (M) and standard deviation (SD) and for the qualitative variables with recounts (n) and proportions (%). Goodness-of-fit to a normal distribution was determined using the Shapiro–Wilk test and homogeneity of variance was assessed using Levene’s test. The quantitative variables of the treatment and control groups were compared using the Mann–Whitney U test and Pearson’s Chi-square was used to compare the qualitative variables. Additionally, the variations between the different phases were assessed by means of the percentage change rescaled with log base 2. The association between the quantitative variables was determined using bivariate (Pearson’s linear correlation coefficient or Spearman’s rho) and partial correlations controlling for the variables age and sex. A linear mixed effects model Laird & Ware, (1982) 57 was subsequently used to assess changes over time for the repeated measurements of the pain questionnaire scores at three time points between the control group and treatment group. Linear mixed effects models account for variability between subjects and variability between repeated measurements in the same subject simultaneously. To obtain different trajectories for each group (experimental vs. control) over time, we included the intercept and slope effect as random effects; and time, group and the interaction term (group × time) as fixed effects. The variance-covariance structure was fixed to an unstructured matrix and the random effects and error terms were assumed to have a normal distribution. Furthermore, Cohen’s d was calculated based on the results of the linear mixed model (Cohen, 1988). The R-project nlme package (version 3.5.0) was used to estimate all the regression models. The established level of statistical significance was P < .05.

Results:

A total of 297 subjects (205 females and 92 males) were initially invited to participate in the study. Of these, one was excluded for not meeting the inclusion criteria, 78 because they did not attend the meeting and 20 who declined to participate. A total of 198 subjects (151 women and 47 men) were randomized to one of the two study arms (Fig. 1). After randomization, five subjects were excluded for declining to participate, dropping out before the intervention or because they did not know how to use the technologies. In the intervention group, 15 patients did not complete the intervention. The analysis was performed with patients who had completed all three questionnaires at baseline, upon completion of the intervention and three months after the intervention. Table 2 shows the demographic data and baseline characteristics of the sample by groups. The outcome variables showed measures for PCS (P = .20), CPAQ (P = .07) and EQ-5D (P = .26) at the beginning of the intervention (baseline) in which the participants of both groups did not differ in pain catastrophizing, quality of life or pain acceptance. No differences were found in the sociodemographic variables or use of medications (Table 2). Primary Outcome: Catastrophizing The descriptive results for differences between the control and intervention groups regarding measures of catastrophizing at baseline, upon completion of the treatment and during follow-up are presented in Table 3. Between-group variations over time for the primary outcome variable are presented in Table 4. Between-Group Effects Immediately following the treatment (T2), statistically significant improvements were observed in the intervention group for catastrophizing (20.86 vs 31.82; P = .001) and the catastrophizing subscales of helplessness (8.91 vs 14.22; P = .001), rumination (4.05 vs 6.32; P = .004) and magnification (8.29 vs 11.29; P = .007). However, at 3 months of follow-up, the mean remained lower (25.78 vs 31.41; P = .09), although the differences were slightly below our established level of statistical significance. Within-Group Effects Positive effects were observed at the different treatment times according to the percentage change rescaled by log base 2. Specifically, positive effects were found for in the intervention group (T2-T1) in catastrophizing between the baseline and post-treatment phases (P = 3.93E+11) and in the subscales of helplessness (-0.72 vs 0.1; P = .002) and rumination (-1.59 vs -0.53; P = .0003), although no significant changes were found for magnification (0.2 vs 0.77; P = .137). Significant results were also found for catastrophizing in the three-month follow-up with respect to the baseline (-0.59 vs 0.2; P = .0056) and the subscales of helplessness (-0.65 vs 0.01; P = .07), rumination ( 1.23 vs -0.59; P = .04) and magnification (.1 vs .86; P = .02), all of which improved three months after completing the treatment. We also measured changes in the pain questionnaire scores between the control and the treatment group at the three time points applying the linear mixed-effects model. Statistical differences were only found between the two groups for changes in the PCS score over time. In addition, the interaction effect between time and the intervention group is 6.47 (P = .001), thus indicating a significant decrease in PCS scores over time in the intervention group compared to the control group ( Table 5). Secondary Outcomes Table 6 shows the results for the variables acceptance (CPAQ), quality of life (EQ-5D) and overall health state (EQ-VAS), while Table 7 shows variations over time for the secondary outcomes. Between-Group Effects As regards pain acceptance (CPAQ), no significant differences were found between the two groups after treatment (68.23 vs 63.82; P = .143) or at 3 months following the intervention (64.48 vs 65.67; P = .468). In terms of quality of life (EQ-5D), the intervention group showed significant improvement at the end of the treatment (0.55 vs 0.41; P = .008), although these differences were not maintained after the 3-month follow-up (0.43 vs 0.39; P = .299). The assessment of overall health state (EQ-VAS) registered in the daily records showed significant improvements in the intervention group compared to the control group (52.68 vs 38.68; P = .02) at the end of the treatment and that these differences were maintained over time (51.05 vs 39.96; P = .02). Within-Group Effects Regarding variations between the different phases, a positive effect was observed immediately following the intervention (T2-T1) in both acceptance (0.38 vs 0.05; P = .0007) and quality of life (0.43 vs -0.01; P =. 0016), but the positive effect on overall satisfaction with health was not maintained (0.25 vs -0.27; P = .132). As shown in Table 6, no significant differences were found during follow up for CPAQ (T3-T2: -0.22 vs 0.08, P = .143; T3-T1: -0.1 vs 0.12, P = .468), for EQ-5D (T3 T2: 0.09 vs 0.04, P = .172; T3-T1: 0.08 vs 0.12, P = .479) or for overall health state (T3-T2: -0.05 vs -0.08, P = .489; T3 T2 0.18 vs 0.18, P = .344). Table 8 shows the Fisher–Freeman–Halton descriptive analysis of the proportion of participants with clinically significant improvement immediately after treatment according to the EQ-5D subscale. In the intervention group, clinical improvement in pain intensity ranged from 37% improvement in moderate pain to 22.7% improvement in severe pain (P = .04). The exact test results showed that a significantly higher proportion of participants who received the multimodal treatment improved in the mobility subscale (P = .039) and activities of daily living such as going to work, leisure time and family activities (P = .045) immediately after receiving the treatment. These improvements were not significant three months after the intervention.

Conclusions:

The results of our study suggest that multidimensional treatments for adults with chronic pain improve catastrophizing, quality of life and psychological flexibility immediately after treatment and that the effects are maintained for the primary outcome of catastrophizing for at least 3 months following treatment. The study has also shown that non-pharmacological treatments that include physical and psychoeducational therapy to promote active participation work well in combination with pharmacological strategies and that such interventions improve self-reliance in chronic pain patients and help them to cope constructively with pain. The NO + dolor application we have developed uses gamification to teach patients distraction methods and divert their attention away from pain, as well as mindfulness techniques to improve pain acceptance. It also provides patients a well-paced program of exercises and information on the proper use of medications to avoid side effects and helps them to identify health assets to engage in pleasurable activities or find the resources they need. Moreover, the app-based mobile interventions we have presented here are flexible, self-directed and promote self-care in patients with chronic pain and can be used to complement face-to-face pain treatments. Preventive interventions for people with chronic pain designed from a salutogenic approach such as ours are essential to promote well-being and prevent further decline in health throughout life. Clinical Trial: The study is registered at ClinicalTrial.gov (NCT04509154).