Interdisciplinary pain treatment programs for highly impaired adolescents with severe chronic pain usually take an active self-management approach [1
]. This means that patients are actively involved in, and personally responsible for, managing their pain by applying active pain coping strategies. Prior to treatment, most patients believe medical interventions will manage their condition [2
], however, during interdisciplinary pain treatment, they learn the skills required to manage their own pain [1
]. Insufficient motivation to self-manage pain is thought to be an important factor contributing to unsatisfactory treatment outcomes in both pediatric and adult pain patients [3
The Pain Stages of Change Questionnaire (PSOCQ), originally developed for adults [7
], is derived from the transtheoretical model [8
], which postulates that individuals experience differing levels of willingness to employ pain self-management strategies in identifiable stages. Four stages of change are assessed by the PSOCQ; Precontemplation, Contemplation, Action and Maintenance [7
]. Briefly, in Precontemplation, patients typically lack personal responsibility for their pain management, looking towards medical interventions for solutions to their pain. Contemplation is characterized by awareness of the need to change pain coping [2
]. The Action stage is identifiable by the recent commencement of self-management strategies, and Maintenance occurs when the patient routinely uses pain self-management strategies in everyday life [2
Guite et al. [2
] adapted the adult version of the PSOCQ for use in adolescents with chronic pain (PSOCQ-A). As parents are usually involved in their child’s pain management, a parent version was also created (PSOCQ-P). Factor analyses combined the Action and Maintenance stages within the PSOCQ-A [2
]. Thus, for adolescents, the questionnaire consists of three, instead of four, stages of change. For parents, four stages of change were confirmed [2
]. Although a German version of the PSOCQ for adults exists [9
], to date, no German versions of the PSOCQ-A and PSOCQ-P are available. Therefore, the overarching aim of the present study was to create validated German versions of the PSOCQ-A and the PSOCQ-P. A previously conducted pilot study revealed low acceptance of both German questionnaires [10
] because participants had difficulty interpreting items, particularly prior to specialized pain treatment in which patients often learn about the concept of active pain self-management for the first time. Therefore, determining the temporal utility of the questionnaire during different stages of treatment was identified as important. To address these goals, two studies were conducted. The aims of Study 1 were to investigate the factor structure of the PSOCQ-A and PSOCQ-P, to create short versions of the questionnaires to increase usability in clinical practice, and to investigate the suitability of the PSOCQ at different stages prior to treatment. The aims of Study 2 were to cross-validate the short versions developed in Study 1 in an independent sample. Sensitivity to change at different time points during and after treatment was investigated, as was criterion validity, by means of associations between readiness to change, pain characteristics and treatment outcomes.
The aim of the present study was to create and validate German versions of the PSOCQ-A and the PSOCQ-P. In Study 1, a short version was created that was cross-validated in Study 2. An acceptable to good model fit was found for both PSOCQ-A and PSOCQ-P. However, internal consistency scores for Precontemplation and Contemplation scales were low for the PSOCQ-A. The results also indicated that it may not be useful to administer the PSOCQ-A before initial presentation to specialized pain care. Readiness to change was associated with pain intensity in the present study, but not with functional impairment, pain coping or with treatment failure three months after discharge. According to both patient- and parent-report, willingness to change appeared to progress over the course of treatment.
Regarding the PSOCQ-A, we confirmed a factor structure consistent with previous studies (Precontemplation, Contemplation, Action/Maintenance) [2
]. For the PSOCQ-P, previous studies have found support for a four-factor solution (Precontemplation, Contemplation, Action, Maintenance) [2
]. We similarly found support for a four-factor solution when including all original questionnaire items. However, these findings were only evident in patients and parents admitted to inpatient IIPT. In the group of adolescents and parents completing the questionnaire prior to adolescent intake at a specialized pain treatment center, model fit was poor.
The problematic factor structure of the PSOCQ-A and PSOCQ-P in individuals completing the questionnaires prior to intake may be due to these patients and parents being unfamiliar with a self-management approach to pain, as noted by intake staff. This may have caused difficulty answering items using the scale of ‘strongly disagree’ to ‘strongly agree’ for items indicative of later stages of change (e.g., “When my pain flares up, I find myself using coping strategies that have worked in the past, such as a relaxation exercise or distractions.”; “I use what I have learned to help keep my pain under control”). Contrastingly, in the original English validation study, patients and their parents completed the PSOCQ-A and the PSOCQ-P prior to initial evaluation in a similar setting to ours, with no difficulties noted by researchers [2
]. Usually patients are referred to specialized centers from other medical subspecialties because previous treatments have failed [2
]. It is not clear how much information has been provided to these patients about pain self-management before presentation to specialized care. Differences in health care systems, or available chronic pain education outside of treatment settings, may be responsible for the discrepant findings.
A German pilot study revealed low acceptance of the PSOCQ questionnaires [10
]. In response, we created short versions of the PSOCQ-A and PSOCQ-P in Study 1 in an attempt to make accurate completion more likely. However, these revised versions lost comparability with the original version. Furthermore, given that the PSOCQ-A resulted in Precontemplation, Contemplation and Action/Maintenance factors, and the PSOCQ-P resulted in Contemplation, Action and Maintenance factors, the parent and child version are no longer directly comparable. Therefore, addressing one issue resulted in other problems in the validity of this tool.
Other psychometric issues for the PSOCQ-A were also apparent in our study, with low (≤0.60) internal consistency found in both the original and shortened versions of the Precontemplation subscale. The Precontemplation subscale of the shortened PSOCQ-A version, in particular, had only three items and very low internal consistency, and therefore should be used with caution. Previous studies have consistently found the lowest internal consistency scores for the Precontemplation subscale [2
]. Additionally, the value of the Precontemplation subscale in the pediatric chronic pain population is questionable, as only 2.7% of adolescent pain patients in this stage were at the start of in-patient treatment in the present study. This distribution is in sharp contrast to previous studies with adolescents, where one third to one half of the participants were categorized as being in the Precontemplation stage prior to intake [2
]. Some important methodological differences should be noted here, with our sample being collected during their admission to in-patient pain treatment, whereas all four pediatric studies were collected prior to intake (with three out of four through mail-out questionnaires). However, in adult studies assessing the stage of change at a similar stage of treatment to our sample, Jensen et al. [24
] and Strong et al. [25
] found that only a small number of adult patients were categorized in Precontemplation, and that most patients were in the Contemplation stage. They concluded that only patients who are motivated and ready to change complete the intake procedures and are admitted for specialized treatment [24
]. This indicates that the Precontemplation subscale in the PSOCQ-A may have some value prior to treatment, however, once admission to specialist treatment has been confirmed, this subscale no longer appears to have much clinical utility.
Despite difficulties with factor structure and low internal consistencies, change in stages across treatment, as measured by the PSOCQ-A and the PSOCQ-P, was in the expected directions and reflected treatment progress. Logan et al. [3
] and Sherry et al. [26
] also demonstrated significant changes in readiness to change during IIPT using the PSOCQ-A and PSOCQ-P. In combination, evidence suggests that interdisciplinary pediatric pain treatment increases readiness to change [3
], with the PSOCQ able to monitor this component of treatment progress.
The assessment of criterion validity in the PSOCQ-A, in line with findings by Guite et al. [2
], identified significant associations between pain intensity and readiness to change stage, in which higher pain intensity was found in adolescents in the Precontemplation stage. Other studies have additionally found significant associations between readiness to change stage and other pain characteristics, such as pain-related disability or passive pain coping [2
], that were not replicated in the present study. Important methodological differences should be noted here. While other studies assessed correlations between subscale scores and outcome scores, here we assessed for statistical differences between groups defined by their stage category. As the aim of the PSOCQ is to identify the stage of change at a given time, the observation of between-group differences is an important component of validity. Our study suggests limited concurrent criterion validity of the PSOCQ-A.
Similar assessments of parental stages of change through the PSOCQ-P identified associations with child coping, but not pain-related outcomes. This is consistent with other research identifying that adolescent psychological components were inversely associated with parental stage of change, including fear of pain [3
], pain catastrophizing [18
] and anxiety [18
], although these were not assessed in this study. Parental stage of change was associated with adolescent pain in one study. In this case, the Precontemplation stage alone was associated with adolescent higher pain-related disability [2
]. Generally, our results indicated some level of consistency regarding the concurrent criterion validity of the PSOCQ-A and PSOCP-P, however this was not a strong finding when assessing individual stages of change with outcomes.
Limited predictive criterion validity was found in our study. Consistent with cross-sectional associations with pain intensity, we found that in adolescents, transferring to a higher stage of change during treatment was associated with a reduction in their pain intensity post-treatment, however, did not find that readiness to change was a significant predictor of disability or treatment failure, nor did we find that stage of change at the end of treatment was associated with post-treatment outcomes. Furthermore, pretreatment scores could not be used for prediction in this study due to low variance, because most adolescent patients were in the same (Contemplation) stage. A study with similar methodology found somewhat consistent results. Logan et al. reported that increases in readiness to change over treatment were associated with decreases in functional disability and maladaptive psychological outcomes [3
]. The authors reported very few meaningful associations with treatment outcomes were found for individual stages of change assessed at pre-treatment. In combination, these findings suggest the utility of the tool may be in tracking mindset shift as identified by a change in stage of change, rather than identifying a particular stage.
4.1. Clinical Implications
Given that the PSOCQ-A and PSOCQ-P were developed and validated in the United States, it was important for us to ensure psychometric properties were maintained in the German translation, particularly as findings from the German pilot study indicated that the population had difficulty interpreting items. Currently, given our findings, we do not recommend using the PSOCQ-A and PSOCQ-P as a measure of the stage of change. This is for a few reasons: validity was not achieved in a pre-clinical sample; associations between stages of change and pain-related outcomes tended to be seen with regard to changes across stages, rather than regarding specific stages themselves; and treatment failure was not predicted.
4.2. Strengths and Limitations
The strengths of this study are the large sample in Study 1, the application of the survey in different settings (outpatient vs. inpatient), cross-sectional and longitudinal validation, and the focus on improving the usability of the PSOCQ in clinical practice. However, a focus on clinical utility resulting in the shortening of the questionnaires limits the comparability of the German versions with other studies. Furthermore, the German versions have only been tested in a single pain-treatment center.