Longitudinal outcome evaluations of Interdisciplinary Multimodal Pain Treatment programmes for patients with chronic primary musculoskeletal pain: A systematic review and meta‐analysis

2.1 Protocol and registration

The study was reported in line with the PRISMA guidelines (Moher et al., 2009, 2015) and the study protocol has been registered in PROSPERO under CRD42018076093.

2.2 Search

We performed our search in Medline and Embase via OVID, and PsycInfo and Cinahl via EBSCOhost from inception to May 2020. The search string was developed by experienced reviewers (SE and JK) and consisted of multiple blocks that were combined with Boolean operators (see File S1). Each block included free-text words as well as specific subject headings. In addition, we searched for grey literature including unpublished studies in the Dart Europe, Open access Theses and Dissertations, NDLTD, ClinicalTrials.gov and WHO ICTRP databases. For each included study, we also performed forward (in Google Scholar) and backward reference searches.

2.3 Eligibility criteria

Randomized controlled trials, as well as case series and cohort studies were included. Study cohorts had to include adult patients with chronic primary musculoskeletal pain for at least 3 months that was primarily perceived in musculoskeletal structures (e.g. bones, joints, muscles or related soft tissues; Treede et al., 2015, 2019). In case of mixed cohorts, at least 75% of the patients had to experience musculoskeletal pain. The criteria for IMPT programmes were based on the definition of Gatchel et al. and had to include (a) a common philosophy treatment in line with the biopsychosocial model of pain; (b) a treatment component where patients actively participated by means of tasks, training and/or exercise; (c) at least three different healthcare professionals from various disciplines that provided the interdisciplinary treatment; (d) a single facility where each patient received treatment (Gatchel et al., 2014). This last criterion excluded care-network settings, but not multicenter trials. Although structured team meetings are considered an important aspect of IMPT programmes (Kaiser et al., 2017), we did not include this as an inclusion criterion, because we expected that not all studies would explicitly report this. Our outcomes were based on the criteria developed by the Initiative on Methods, Measurement, and Pain Assessment in Clinical Trials and included physical functioning, pain interference, depression, anxiety, emotional functioning, anger, self-efficacy, social functioning and pain intensity (see protocol for rationale; Dworkin et al., 2005; Turk et al., 2003). The study had to include at least one outcome that was measured at two-time points: prior to treatment and at least 12 months after the intervention was completed. Studies that focused on patients with post-surgical pain or cancer pain, as well as studies that solely included patients on the basis of specific comorbidity (e.g. depression) were excluded. Articles published in other languages than English, German or Dutch were also excluded.

2.4 Study selection, data extraction and risk of bias

All study selection, data extraction and risk of bias assessments were independently performed by at minimum two different researchers (UK and SE for articles in German, SK, SE and MK for articles in other languages). Researchers used pre-tested forms and compared their input to reach a consensus. In case of disagreement, the study was discussed with other researchers (HW and RS) for a final decision. Study selection was performed in two rounds. In the screening round, abstracts were screened using the Rayyan software package (Ouzzani et al., 2016). Subsequently, full-text studies were assessed on all eligibility criteria.

From the extraction round onwards, we considered patient cohorts—not journal articles—as our primary unit of analysis. In case of multiple articles describing the same cohort, we combined these sources to construct a complete overview of the development over time. The first published article that met our eligibility criteria was used as the primary source and we consulted additional sources, such as protocols or follow-up studies if they contained additional relevant information. If the information sources did not contain all data items of interest, we did not contact the study authors but coded this as ‘not reported’ in our dataset. Our data extraction form included all items from the template for intervention description and replication (TIDieR) checklist to describe the content of the treatment programme in detail (File S2; Hoffmann et al., 2014). Risk of Bias was assessed with the Joanna Briggs Institute Checklist for Case Series, which included 10 criteria (Moola et al., 2017). A response of ‘no’ to any one of the items resulted in a high risk of bias, unless we found a clear indication of a limited impact of that item on the overall study outcome. The risk of bias form, including the scoring instructions, are available in the online multimedia appendix.

2.5 Data analysis

The data extraction form included sample size (per measurement moment), age, sex, pain duration, nationality, method of recruitment, patient eligibility criteria, exclusion criteria, study design, type of outcome measures, and outcomes for all available time points on measurement instruments of interest. If treatment intensity was expressed in days, we assumed 6 h of treatment per day. Because IMPT programmes are generally considered as a treatment of last resort, we specifically paid attention to obtaining information on attrition (Jeffery et al., 2011). We obtained pre-, post- and final-follow-up sample sizes to calculate attrition rates for post-treatment and follow-up. When a cohort presented data for two or more outcome measures within one domain, we selected the most commonly used instrument.

2.5.1 Descriptive analysis

To investigate the heterogeneity between the included IMPT programmes, study, patient and intervention characteristics were summarized in tables. Intervention descriptions were extracted and each separate component was then classified into one of 10 possible categories. These components indicate the various means by which each IMPT programme aims to optimize daily life functioning and wellbeing. Education referred to modalities that were primarily concerned with the transfer of information from healthcare providers or experts to patients. All modalities regarding physical training, such as stretching, hydrotherapy and walking were categorized as exercise. Graded activity was only coded if the modality explicitly used the term graded activity or if the activities gradually and time-contingent increased after a baseline measurement. Modalities that described (cognitive) behavioural approaches, including problem-solving training, exposure in vivo, rational emotive therapy or ACT were classified as (cognitive) behavioural treatment. Breathing techniques, autogenic training, mindfulness, and applied relaxation techniques were classified as relaxation. Although self-management, defined as ‘the intrinsically controlled ability of an active, responsible, informed and autonomous individual to live with the medical, role and emotional consequences of his chronic condition(s) in partnership with his social network and the healthcare provider(s)' (Van De Velde et al., 2019), is likely to be influenced various treatment components, some IMPT programmes included specific treatment sessions where coping with pain, setting realistic life goals and problem-solving skills were discussed. These types of sessions were classified as generic self-management skill training. Pharmacological treatment was only coded when medication was provided in response to chronic pain. Medication withdrawal procedures were coded as ‘other’. Workplace visits, and ergonomic advice at the workplace were coded as workplace advice. The category body awareness included physical awareness and psychomotor exercises that aimed to improve the recognition of bodily signals. The last category—team meetings—was only coded when the patient actively participated in the team meetings. The categories were inductively developed by first extracting and then clustering the modalities of the first search into global categories (by SE and SK). In the final dataset, these 10 categories covered more than 90% of the treatment modalities. All remaining modalities were coded as ‘other’. The description of each of the modalities and the classification were registered. A similar process was performed for healthcare providers. The following professions were coded as ‘physician’: occupational physician, rehabilitation physician, general practitioner and not otherwise specified physician. Other physician specialists (e.g. psychiatrist, orthopaedic surgeon, anesthesiologist) who were mainly involved in consulting instead of a coordinating role were coded as ‘other’. Disciplines such as clinical psychologists, general psychologists and behavioural therapists were classified as ‘psychologist’. Physical therapists and physiotherapists were classified as ‘physical therapist’. Social workers and social counsellors were classified as ‘social worker’. Occupational therapists and nurses were classified accordingly.

To assess to what extent treatment programmes aligned their programme with individual patient characteristics and preferences (i.e. tailoring), we classified each programme into low, medium or high tailoring. We defined low tailoring as any form of personalized goal-setting, because this would allow patients to relate treatment content and progress to their personal situation. All studies received at minimum a ‘low’ tailoring classification because we assumed that all interdisciplinary programmes require some form of collaborative goal-setting at the start of treatment. We classified programmes as medium tailoring, when they selected or optionally provided specific treatment components based on patient-specific needs or preferences. High tailoring involved a fully personalized treatment programme, with varying duration and treatment activities and modules, based on each patient's clinical assessment.

2.5.2 Main data analysis

In addition to pain intensity, we included seven key outcome measures as outcomes in this analysis, divided over three domains: physical health, mental health and social health. For physical health, we included physical functioning and pain interference. We extracted of the outcomes depression, anxiety, anger, and self-efficacy beliefs within the mental health domain. For social health we only included social functioning. All outcomes were defined in the study protocol. For each of these outcomes that were present within a cohort, we used the available data to calculate effect sizes for pre-post, post-follow-up and pre-follow-up contrasts. To calculate effect sizes, we used the method of Becker et al's standardized mean change (1988), with the modifications that were suggested by Morris (2000). The model assumes that the outcomes are normally distributed at both time points, with separate means but equal variances. Furthermore, the model corrects for a pre-post within-group correlation. Because we did not have access to the original data of the included cohorts, we imputed this value (Lipsey & Wilson, 2001). For all studies, we imputed the median correlation (r = .59) of a meta-analysis that investigated the range of within-group correlation values in active treatment groups (Balk et al., 2012). This value is comparable to other studies that have imputed within-group meta-analyses (Clond, 2016; Roberts et al., 2017). In addition, a sensitivity analysis for the within-subject correlation is available in the multimedia appendix for all r-values between 0 and 1. Sample sizes lower than n = 10 were not included in the analysis, because this could lead to inaccurate estimates of the standardized mean gain (Morris, 2000). In some interventions, the main treatment programme was followed by follow-up treatment activities to enhance maintenance. In these situations, we considered end of treatment as the moment that the main treatment programme (ie. that covered the core of the treatment procedures) ended. Hence, follow-up meetings, booster sessions or reinforcement sessions were not considered as main treatment and could continue after post-treatment assessments. All assessments within 1 month after end of treatment were considered as a ‘post’ measure. We used the last available time point for the follow-up contrast. We calculated standard deviations from standard errors by multiplying them with the square root of the corresponding sample size (Higgins et al., 2019). If medians and range were provided, we used the formula of Hozo et al. (2005) to estimate the mean value and corresponding SD. For studies that presented change scores, we calculated final value mean scores and imputed the baseline standard deviation. If the latter was not available, the study was not included in the meta-analysis. For medians and interquartile ranges (IQRs), we estimated means and SDs using the assumption that the IQR width is 1.35 SD (Higgins et al., 2019). In case of missing measures of variability at follow-up, we imputed the baseline value or otherwise used the mean SD of the remaining trials that reported on that outcome. If data of the cohort was presented for different subgroups, we calculated one composite mean and SD. For data that was only presented in figures (e.g. boxplots), we measured the central tendency and measure of dispersion if the figure was of sufficient quality.

Subsequently, we summarized the effect sizes per outcome, by describing the direction of effect for each of the included cohorts over time. We a priori decided not to perform any pooling, because this was not in line with our study aims and we expected substantial heterogeneity among the included studies. To facilitate interpretation of the effect sizes, we re-expressed the median pre-post effect size on the most commonly used measurement instrument, using the weighted standard deviation of all available post-intervention scores of that instrument. To assess the statistical heterogeneity of the study outcomes we also calculated the I2 and the Q test for each outcome domain at every time point. A statistically significant Q test rejects the hypothesis that all effect sizes are equal (Huedo-Medina et al., 2006). In addition, the I2 index provides an indication of the proportion of variability in observed effects that is either due to between-study variability or due to within-study variability (ie. sampling error; Borenstein et al., 2017). This analysis was performed with the R metaphor package in RStudio (R Core Team, 2013; RStudio Team, 2020; Viechtbauer, 2010).

2.5.3 Exploratory data analysis

To further explore the included cohorts, we developed an online multimedia appendix that contains interactive forest plots for each outcome, time series and study characteristic tables. The appendix can be accessed via https://stefanelbers.shinyapps.io/deployment/. The time series show the development over time of a measurement on a standardized scale (expressed as a percentage of the maximum score of each particular measurement instrument), as well as the raw scores and standard deviations. To standardize the scores, we obtained the distance of each mean and the unfavourable end of the scale, divided by the total distance of the scale and multiplied by 100. These plots also allow for the comparison of specific cohorts over time with respect to a particular outcome. To accommodate future updates of this systematic review, the appendix also contains contact information to encourage readers to pinpoint any inaccuracies or to suggest cohorts that have not yet been included in the current review. Finally, all data extraction files are listed in this appendix, including any comments that have been made regarding handling specific difficulties with that specific cohort (e.g. dealing with change scores, or imputing missing SDs). File S3 contains the R code that has been used for all analyses in this review as well as the deployment of the appendix. These files can also be accessed in a Github repository via: https://github.com/stefanelbers/impt.meta_analysis/tree/master/deployment.

留言 (0)

沒有登入
gif