Measurement properties of the Swedish Brief Pain Coping Inventory-2 in patients seeking primary care physiotherapy for musculoskeletal pain

1 Introduction

Musculoskeletal pain (MSKP) disorders are the main reason people seek primary care physiotherapy. Care of benign MSKP disorders should be based on a biopsychosocial approach where self-management is prioritised over finding a “cure” [1,2,3,4,5]. To deliver biopsychosocial care, healthcare professionals need to be aware of how cognitions, emotions, and behaviours can influence a patient’s MSKP-disorder and health. Incorporating psychological awareness into practice is known as psychologically informed care [6,7]. Many psychological factors, such as beliefs about pain, pain self-efficacy, coping strategies, and psychological flexibility (PF), may be influential in the development and prognosis of MSKP-disorders [8,9,10,11,12]. The assessment of the relative influence, positive or negative, and subsequent treatment of such psychological factors is therefore a central but challenging part of primary care [13,14,15,16]. To enhance psychological awareness, primary care practitioners and researchers can integrate methods from the field of chronic pain management, where cognitive behavioural (CB) models have long been used [16,17,18]. Two well-established, overlapping models are the traditional CB therapy (CBT) model and the PF model. CBT focuses on negative thoughts, attention, avoidance, and MSKP-coping behaviours, while the PF model encompasses pain acceptance, focused awareness, cognitive defusion, values-based action, and self-as-observer [11,17]. PF has been defined as “one’s ability to directly and openly contact experiences in the present moment and persisting or changing behaviour according to what the situation affords and one’s personal goals and values” [19, p. 141]. Patients presenting to primary care may demonstrate psychological inflexibility if they reduce participation in valued activities while rigidly focusing on finding the cause of their pain [20,21]. Patients may also demonstrate maladaptive behaviours such as pain avoidance if they believe pain to be a sign of damage [21]. Primary care practitioners can therefore benefit from time-efficient methods that assess patients’ MSKP-coping strategies and PF. Such methods may also help identify patients who could benefit from psychologically informed interventions, such as acceptance and commitment therapy (ACT) [6,11]. However, as MSKP-coping strategies and PF have not been widely evaluated in primary care populations, their relative significance in the management of MSKP-disorders remains unclear. To facilitate psychologically informed primary care research and practice, reliable and validated patient-reported outcome measures (PROMs) measuring MSKP-coping strategies and PF are needed.

The Brief Pain Coping Inventory (BPCI) has been cited as a suitable PROM for assessing MSKP-coping [22]. The BPCI version 2 (BPCI-2) was developed as an evolution of the BPCI, and both were developed by clinical psychologists as PROMs for use in English-speaking populations attending interdisciplinary care for chronic MSKP [20,23]. The BPCI-2 is a 19-item PROM questionnaire with two subscales measuring traditional pain management strategies (TPMS) and PF. The TPMS subscale has eight items (scored 0–56) and the PF subscale has 11 items (0–77, four reverse scored). Scoring is based on the number of days in the last week a patient has performed a given strategy, with higher subscale scores indicating more adaptive TPMS and PF [24]. The TPMS subscale was created based on CB models and measures self-reported use of exercise, pacing, passive modalities (e.g., ice), relaxation, positive self-statements, and distraction in the preceding week [20,24]. The PF subscale was created based on the PF model and measures self-reported pain acceptance and distress, broad and focused awareness, and values-based engagement in the preceding week [20,24]. As far as we are aware, only three studies have evaluated the measurement properties of the BPCI-2, and all of these were conducted in English-speaking populations of patients attending intensive treatment programmes for chronic MSKP. In these studies, the BPCI-2 demonstrated acceptable structural validity and clinical utility, while the PF subscale demonstrated stronger construct validity than the TPMS subscale [19,20,24]. In summary, the BPCI-2 shows potential as a low-burden PROM for assessing TPMS and PF in primary care. However, prior to its use in Sweden, it is important that the BPCI-2 is subjected to a thorough translation process and its measurement properties subsequently reassessed [25,26].

This study aimed to translate and cross-culturally adapt the BPCI-2 from English to Swedish, to create a Swedish version of the BPCI-2 (BPCI-2:SWE) and to evaluate whether the two BPCI-2:SWE subscales are valid and reliable for use in primary care research and clinical practice.

2 Methods

2.1 Design

This study is reported based on COnsensus-based Standards for the selection of health Measurement INstruments (COSMIN) guidelines [27]. The BPCI-2:SWE was translated and cross-culturally adapted, and its validity and reliability were evaluated in seven stages (I–VII). The methods for stages I-VI were based on guidance from Beaton et al. [28], augmented with the face-validity index (FVI) method and guidance from COSMIN regarding translation and content validity analysis [25,29]. The methods for evaluating construct validity, test–retest reliability, and measurement error in stage VII were based on COSMIN guidance [25]. Details of stages I–VI are given in Figure 1.

Figure 1

The six-stage process for translation of the BPCI-2 from English to Swedish (stages I–IV), content validity, and qualitative content evaluation of the BPCI-2:SWE (stage V) prior to final approval by the expert committee (stage VI). Process based on the methods outlined by Beaton et al. [28] and the COSMIN study design checklist [25]. ^‡The expert committee (n = 11) consisted of the five authors (two associate professors in physiotherapy, one post-doctoral researcher in physiotherapy, one PhD student in physiotherapy, and one certified physiotherapist) as well as the three additional translators, a professor in physiotherapy experienced in PROM development and two patient representatives educated in research methodology. Figure was adapted from Beaton et al. [28].

2.1.1 Participants

Patients eligible for stages V and VII were adult patients (≥18 years) booked for an initial physiotherapy consultation for MSKP at one of the three participating physiotherapy departments within Region Östergötland (RÖ), Sweden. Patients were excluded if they did not demonstrate acceptable Swedish linguistic skills as judged by a native speaking physiotherapist, if impairments made them unable to complete the questionnaires, or if they required urgent medical consultation.

2.1.2 Content validity evaluation (stage V)

Based on COSMIN guidance, the pre-final BPCI-2:SWE’s comprehensibility and relevance was examined using the FVI-method and semi-structured interviews in a cohort of 13 patients [25,29]. The 13 patients were interviewed by one author (JA). The interviews were recorded and transcribed verbatim. Recruitment was ceased once the interviewing author judged that no new insights were evident. The interviews included open and closed questions and enabled calculation of the FVI for clarity and comprehensibility for each item (I-FVI) and subscale (S-FVI/Ave) as described by Yusoff [29]. Minimum FVI scores of 0.8 (range 0.0–1.0) for clarity and comprehensibility for each item (I-FVI) and subscale (S-FVI/Ave) were considered acceptable [29]. The FVI-method was adapted to evaluate each item’s and subscale’s relevance [25,29]. Qualitative content analysis of the interview transcripts was conducted according to Graneheim and Lundman’s [30] method. An inductive analysis of the manifest content was used to explore patterns in the patients’ experiences of the BPCI-2:SWE and create categories and sub-categories regarding comprehensibility and relevance [31]. The transcripts were initially analysed independently by two authors (RT and JA), and two senior authors (KJ and MF) triangulated the results by reviewing the transcripts and analyses.

2.1.3 Expert committee review (stage VI)

The expert committee from stage IV reviewed the content validity results to judge whether cross-cultural adaptation had been achieved.

2.1.4 Validity and reliability evaluation (stage VII)

The BPCI-2:SWE’s validity and reliability were evaluated using a test-retest design in a sample of >100 patients, based on COSMIN guidance [25]. All demographic and PROMs data at baseline and 1-week follow-up were collected via a web-based survey system (Webropol, Sweden). A 1-week follow-up was chosen to avoid overlap in the BPCI-2:SWE scoring and to reduce scope for spontaneous recovery.

2.1.4.1 Construct validity evaluation

As recommended by COSMIN, the BPCI-2:SWE’s construct validity was evaluated via hypothesis testing using the following Swedish validated PROMs [25]:

The psychological inflexibility in pain scale (PIPS): A 12-item PROM with two subscales assessing pain-related avoidance (eight items; 8–56) and cognitive fusion (four items; 4–26). Lower scores on the PIPS indicate greater PF [32]. The PIPS has demonstrated acceptable measurement properties in populations with Whiplash-associated disorder [32] and chronic MSKP [33].

Average pain intensity in the last week: Assessed using an eleven-point numerical rating scale (NRS; 0 = no pain to 10 = worst imaginable pain) [34]. Pain intensity was included as it is a core outcome measure for studies on MSKP [35].

Self-reported general health in the last week: Assessed using the eleven-point EQ-VAS (0 = worst imaginable health to 10 = best imaginable health) [36,37]. The EQ-VAS has shown adequate validity in European populations [38].

Self-reported control over pain: Assessed using one question from the Brief Illness Perception Questionnaire (BIPQ) (0 = absolutely no control to 10 = extreme amount of control) [39]. The BIPQ has shown good construct validity, responsiveness, and reliability [39,40].

Self-reported levels of exercise; general physical activity and sitting time: Assessed using three ordinally scored questions validated by the Swedish national board of health and welfare that ask the patient how much time in the last week they have performed exercise (0 to >120 min), been physically active, e.g., by walking (0 to >300 min) and how many hours they sit during a normal day (sit nearly all day to never sit) [41,42].

3 Results

The translation and expert committee review stages (I–IV) successfully produced a pre-final version of the BPCI-2:SWE that was tested in stage V.

4 Discussion

This study aimed to translate and cross-culturally adapt the BPCI-2 from English to Swedish and evaluate the measurement properties of the BPCI-2:SWE’s subscales in a primary care physiotherapy setting. The BPCI-2:SWE demonstrated excellent content validity, indicating that the translation and cross-cultural adaptation were successful. The PF subscale demonstrated stronger construct validity than the TPMS subscale. Neither of the subscales demonstrated floor or ceiling effects. The TPMS subscale demonstrated good, and the PF subscale demonstrated moderate test–retest reliability. Measurement error constituted a relatively high proportion of both subscale scores at the individual patient level, but the SDC at the group level was low.

To our knowledge, this is the first study to evaluate the BPCI-2 in primary care. Only three previous studies have evaluated the BPCI-2, and all of these recruited English-speaking patients seeking intensive treatment for chronic MSKP [19,20,24]. In comparison to these three studies, the patients in this study were on average older and had shorter pain duration, lower pain intensity, higher PF, and lower use of TPMS, and fewer patients had spinal pain [19,20,24]. When compared with population-based data, the cohorts in this study appear to reflect a typical Swedish primary care population with MSKP [50,51]. No previous study has evaluated the BPCI-2’s content validity, floor or ceiling effects, measurement error, or test–retest reliability. Therefore, only this study’s construct validity results can be directly compared to previous research. The BPCI-2:SWE’s comprehensiveness was not analyzed in this study as the study intended to retain the original 19 items [25]. Factor and internal consistency analyses were not performed as the 19 items combine to form the constructs TPMS and PF, and as the study intended to retain the original 19-item, two-subscale structure [20,25]. In addition, the thorough translation and cross-cultural adaptation were considered sufficient to maintain the two-factor structure previously established [20].

The BPCI-2:SWE was found to be comprehensible and relevant to patients seeking primary care physiotherapy. The excellent FVI-scores, supportive qualitative data, and low levels of missing data indicate that the BPCI-2:SWE, its instructions, and item statements are clear and comprehensible. The qualitative content analysis underlined the BPCI-2:SWE’s relevance as it found that the statements stimulated patients to reflect on how they managed their MSKP. Interestingly, the patients in this study rated the TPMS subscale as more relevant than the PF subscale. This finding conflicts with results from this and previous studies that have consistently found the PF subscale to be more strongly associated with important clinical outcomes [19,20,24]. The PF subscale’s S-FVI/Ave relevance score (0.76) was significantly lowered by item 13’s I-FVI relevance score (0.38). As item 13 was rated as comprehensible (0.92), its low relevance score may be explained by the heterogeneity seen among the stage V cohort. The stage V patients reflect Swedish primary care, where patients range from those with acute MSKP at a single site, where pain may not hinder activity, to those with chronic widespread pain, who may have significant pain-related disability [52]. Despite its low relevance score with patients, item 13 may still help researchers and practitioners identify patients who report pain avoidance. Overall, the construct validity results indicate that patients may be more familiar with TPMS and that they are perhaps less aware of PF’s importance in pain management. The PF subscale could therefore be used as a screening tool to increase both patients’ and practitioners’ awareness of PF [14,15]. Patients with lower scores on the PF subscale may benefit from treatments that target increasing PF, such as ACT [6,11]. Incorporating ACT into primary care physiotherapy has been shown to be feasible and more effective than usual care for treating chronic LBP [16]. In summary, this suggests that PF is relevant for patients with MSKP disorders, but there may be a lack of insight into its significance within primary care rehabilitation.

The TPMS subscale did not demonstrate adequate construct validity on hypothesis testing. This reflects previous evaluations of the BPCI-2, which have shown that TPMS scores correlate weakly with other measures of physical and psychological functioning [19,20,24]. In this study, patients with higher pain intensity reported more restrictive cognitive processes and engaged more actively in TPMS. These results suggest that the TPMS measured by the BPCI-2:SWE may have questionable effectiveness and that some strategies could even be considered maladaptive [21,53]. For example, the National Institute for Health and Care Excellence (NICE) guidelines now discourage the use of transcutaneous electrical nerve stimulation for the treatment of chronic pain and osteoarthritis [2,4]. Patients who frequently use passive TPMS may be those who prioritise pain control over other valued activities. One could argue, based on its consistent lack of construct validity, that the TPMS subscale should be discarded. However, CBT treatments are an effective part of managing chronic pain, and the TPMS subscale could still be used to screen patients for maladaptive coping strategies [6,11,54]. On balance, these results imply that the TPMS subscale requires updating, for example, by reverse scoring the more passive coping strategies. In contrast to the TPMS subscale, the PF subscale demonstrated good construct validity in correlating moderately and in the hypothesised direction with seven out of the eight other PROMs. The correlations between the PF subscale and other PROMs found in this study were broadly in line with results from previous evaluations of the BPCI-2. This suggests that PF is similarly associated with physical and psychological functioning in Swedish primary care populations as in English populations with chronic MSKP [19,20,24]. Beyond evaluations of the BPCI-2, PF has been shown, in six RCTs, to mediate pain interference and disability in patients with chronic pain [11]. Longitudinal primary care studies could therefore use the BPCI-2:SWE to explore PF’s mediating or moderating role in MSKP-outcomes. Overall, these results suggest that greater focus should be placed on PF than TPMS in primary care research and practice.

The BPCI-2:SWE demonstrated adequate reliability at the group level, but results were more uncertain at the individual patient level. Neither of the subscales demonstrated floor or ceiling effects, giving the BPCI-2:SWE scope to be responsive over time. With regard to test–retest reliability, the TPMS subscale’s ICC exceeded the acceptable level of 0.7 [43]. The PF subscale demonstrated moderate test-retest reliability, but the ICC’s 95% CIs ranged from poor to good, making it difficult to draw firm conclusions [46]. The SEM_agreement and SDC_ind for both subscales constituted a relatively high proportion of the total scores at the individual patient level. This measurement error may be acceptable if the BPCI-2:SWE is used as a screening tool in conjunction with a clinical assessment, but further research is required before the BPCI-2:SWE is used as a PROM at the individual patient level. The SDC_group results were more encouraging and suggest that the BPCI-2:SWE has the potential to detect between- and within-group differences [55]. However, there are several limitations in this study that can have impacted the reliability and measurement error results. First, the qualitative findings revealed that patients reported difficulty in scoring based on the number of days a strategy was performed. Future research should therefore assess whether amending the BPCI-2:SWE’s scoring, for example by using an ordinal scale, improves its reliability. Second, in line with COSMIN recommendations and previous MSKP research, the reliability and measurement error analyses were performed for a stable subgroup identified by a GRoC method [25,56]. A stable subgroup was analysed to compensate for the fluctuations in MSKP that are known to lower ICCs [57]. However, this method resulted in a stable subgroup (n = 64) which only marginally exceeded the threshold for adequate statistical analysis [25]. Third, according to the common-sense model of self-regulation, changes in MSKP-coping precede changes in health outcomes [10]. Some of the measurement errors found in this study may therefore have been a real change in TPMS or PF, which occurred prior to patients experiencing a change in their MSKP disorder. In summary, it is hoped that the BPCI-2:SWE can facilitate psychologically informed primary care research and practice. The BPCI-2:SWE demonstrates adequate validity and reliability for use at the group level. For practitioners, the BPCI-2:SWE could be used for screening or to raise awareness of PF or maladaptive TPMS. However, we would not currently recommend the BPCI-2:SWE as a PROM at the individual patient level. Future research should reevaluate the BPCI-2:SWE’s test–retest reliability with a larger stable subgroup. Evaluation of the BPCI-2:SWE’s responsiveness and value as a prognostic tool, as well as PF’s role as a mediator, in longitudinal studies is recommended.

5 Conclusions

The thorough translation process and results of the content validity analysis indicate that the BPCI-2:SWE has been cross-culturally adapted for use within Swedish primary care. The PF subscale demonstrated adequate validity. However, the TPMS subscale did not achieve acceptable construct validity on hypothesis testing, and revision may be warranted. The reliability can be considered acceptable at the group level, but the relatively large individual-level measurement error indicates that caution is recommended before using the BPCI-2:SWE as an outcome measure in clinical practice. Future research should further investigate the BPCI-2:SWE’s reliability, responsiveness, and prognostic utility.