Clinical Trial Methodology
A patient's perspective on multidisciplinary treatment gain for fibromyalgia: An indicator for pre-post treatment effects?†
ClinicalTrials.gov identifier: NCT00268606.
Increasing attention is devoted to the patient's perspective in clinical research and practice. However, the relationship between the patient's view on treatment progress and standardized pre-post changes in health outcomes is not well understood. The objective of this study was to investigate whether the patient's perception of treatment gain converges with pre-post treatment effects of a multidisciplinary treatment as assessed by standardized self-report measures.
During a tailored multidisciplinary treatment for fibromyalgia, validated self-report questionnaires were assessed at baseline and posttreatment on the outcome measures of pain, functional disability, fatigue, anxiety, and negative mood. In addition, the participants were asked to fill in a questionnaire at the end of the treatment assessing the patient's perception of improvement on core outcomes, as well as satisfaction and usefulness of the treatment.
Moderate to relatively high correlations were found between the patient's perception of improvement and pre-post changes on the physical outcomes, in contrast to small or nonsignificant correlations for psychological outcomes. In addition, satisfaction and usefulness were significantly related to pre-post changes on physical outcomes, but no relationship was found with respect to psychological outcomes.
Results suggest that the patient's perception of treatment gain and pre-post changes in outcomes during treatment assess different aspects of the patient's treatment progress, particularly with regard to psychological functioning. Future research on clinical improvements should consider the patient's perception of treatment gain as an independent and clinically relevant outcome, in addition to standardized trial data of pre-post assessments of health outcomes.
The patient's perspective is becoming more important in clinical research and practice. In the view of increasing pharmacologic and nonpharmacologic treatment choices, patients are more intensively involved in decision making regarding treatment, resulting in improved communication and a more profound understanding of the impact of the disease (1, 2). Moreover, patients are regarded more as active partners in self-care treatment, in particular for nonpharmacologic treatment, where the success largely depends on the active collaboration of the patient. Incorporating the patient's view on treatment can also improve the quality of care. For example, it has been found that tailoring treatment more specifically to the patient's needs and characteristics can enhance treatment effects (3–6).
Due to the growing recognition of the importance of the patient's point of view in the care process, the prioritization of outcome variables has shifted. Concepts such as patient satisfaction and the patient's perception of improvement are more frequently applied as an indicator of treatment value, quality, and even outcome (7, 8). However, it is not well understood to what extent the patient's perception of treatment gain is related to changes during treatment assessed with validated, standardized instruments before and after treatment. For example, with regard to self-reported outcome in syndromes with limited pathophysiologic findings such as fibromyalgia (FM), a chronic rheumatologic condition characterized by widespread pain (9), a difference can be found between the patient's view of treatment gain and more standardized assessments such as patient-reported pre-post changes in pain (10). Several explanations for this low to moderate association have been proposed. For example, it could be that statistically significant pre-post changes are not perceived as clinically relevant to the patient (11). The patient's perception of treatment gain might be viewed as a possible approach to judge whether the amount of change during treatment is large enough to be considered meaningful for the patient. Furthermore, issues other than reducing symptoms might be more important from the view of the patient, such as group support, reducing helplessness, improving self-efficacy, or the relationship with the therapist. For example, it has been shown that a positive communication between physicians and patients with rheumatic conditions and chronic low back pain contributed more to the patient's overall satisfaction with treatment than the improvement in physical functioning (12, 13). Another important factor that could explain low to moderate correlations between self-reported pre-post changes and retrospective reports of improvement is a change of perspective of health status, also known as response shift. Comparing self-reported pre- and postassessments relies on the assumption that patients use the same internal standard of well-being. However, during treatment, this standard may shift in the direction of the improved functioning, which might interfere with self-reported pre-post assessments (14, 15). Indeed, studies investigating the influence of response shift during treatment in various conditions reported a significant difference between the actual baseline value and a retrospective estimate of the baseline value, indicating a change in the internal standard during treatment (16–18). Retrospective reports might also be influenced by response shift tendencies in addition to specific factors such as difficulties with recalling the baseline status and the expectancy of gain during treatment (19, 20). Since self-report measures are used more regularly in research and clinical practice and several possible mechanisms of response tendencies might affect these outcomes, it is important to determine the level of correspondence between the patient's perception of treatment gain and patient-reported pre-post changes during treatment.
The goal of the present study was to examine the relationship between the patient's perception of treatment gain (improvement in core outcomes, satisfaction, and usefulness) and the patient-reported pre-post changes during treatment. During a tailored multidisciplinary treatment, the patient's perception of treatment gain and patient-reported pre-post changes on the core outcome measures of pain, fatigue, functional disability, anxiety, and negative mood assessed with validated self-report measures are assessed. In line with theoretical models and earlier studies on the possible correspondences and differences between these concepts (10–20), we expected no more than moderate correlations between the patient's perception on treatment gain and pre-post changes.
PATIENTS AND METHODS
Participants and procedure.
Patients with FM were referred by rheumatologists from several regional hospitals for participation in a tailored multidisciplinary treatment that was evaluated in a randomized controlled trial. Inclusion criteria consisted of a diagnosis of FM as confirmed by a rheumatologist based on the diagnostic criteria of the American College of Rheumatology (9) and age >18 years. Furthermore, the patients' diagnoses of FM had to be made no more than 5 years prior to the study to ensure a relative early intervention, and only the patients with heightened levels of psychological distress were included (3, 21, 22). Exclusion criteria consisted of severe physical and psychiatric comorbidities that might interfere with the study protocol, FM secondary to another rheumatic condition, illiteracy, and inability to communicate in Dutch. Assessments of the outcome measures were made with valid instruments before treatment and at posttreatment. Furthermore, the participants in the treatment groups were asked to fill out a standardized evaluation questionnaire at the end of the treatment.
Data were collected from 84 consecutive women who participated in the treatment groups. The mean ± SD age of the participants was 41.8 ± 10.3 years (range 21–71 years). The majority of the participants were married or cohabiting (78.6%). With regard to the educational level, 2.4% had a primary level of education, 81.9% had a secondary level of educational, and 15.7% had a tertiary level of education, representing on average 7, 12, and 17 years of formal education, respectively. Between pretreatment and the followup assessment, 5 patients (6%) dropped out. Reasons for dropping out were physical comorbidity that occurred during treatment for 3 patients and lack of sustained motivation to participate in the study for 2 patients. Furthermore, 1 patient who participated in the treatment was excluded due to an incomplete questionnaire assessment, resulting in a sample of 78 patients. When comparing these 6 patients with the total group of 84 patients at pretreatment, no significant differences in demographic variables (e.g., age, marital status, educational level) were found between the groups.
Tailored cognitive–behavioral therapy (CBT) treatment.
The patients received an outpatient treatment tailored to their specific cognitive–behavioral pattern (23). The treatment program was highly structured and consisted of 16 sessions twice a week, plus one booster session 3 months after treatment. The treatments took place in a group setting of 8 participants. Every session started with 2 hours of CBT followed by 2 hours of exercise training. The goal of the treatment was to diminish the cognitive, emotional, behavioral, and social consequences of pain and accompanying symptoms in daily life. The CBT was led by a trained psychotherapist and social worker. On the third, ninth, and fifteenth sessions, the partner of or another significant relation to the patient attended the CBT. The goal of the exercise training was to increase endurance, flexibility, coordination, and relaxation. This training was supervised by experienced physiotherapists and included aerobic exercises (e.g., cycling, gymnastic exercises), anaerobic exercises (e.g., strength and flexibility exercises, functional walking training), hydrotherapy, and relaxation therapy. Throughout the treatment, patients and their spouses received homework assignments each session for the CBT and the exercise training. The final booster session dealt with relapse prevention and further improvement of the attained goals.
Patient's perspective on treatment.
The questionnaire to evaluate the patient's perception of treatment gain was comprised of visual analog scales (score range 0–10, ranging from “not at all” to “very”) to assess the level of satisfaction with the treatment and the overall usefulness of the treatment. Numerous investigators established the validity and reliability of visual analog scales in measuring constructs of primarily patient satisfaction, subjective improvement, and patient-reported quality of life (24–27). In addition, the patient's perception of improvement on the outcome measures pain, fatigue, functional disability, tension, and negative mood were assessed with specific questions on a 4-point Likert scale (where 1 = not, 2 = a little, 3 = fairly, 4 = very). Items consisted of: “Due to the treatment, I have less pain/I feel less fatigued/I experience less disability in daily life (e.g., work, hobbies, social activities)/I feel less tensioned/I feel less depressed.”
Pain was assessed using the 6-item pain scale of the Impact of Rheumatic diseases on General health and Lifestyle (IRGL) (28, 29), which assesses physical, psychological, and social health in patients with rheumatic diseases. Patients indicate the severity and frequency of painful episodes and the duration of morning stiffness within the last month (sample item: “During the past month my condition caused pain.”). The Cronbach's alpha was 0.75 in this study. Previous research has shown that the reliability and validity of this questionnaire are highly satisfactory (29).
Fatigue for the previous 2 weeks was assessed with the 8-item fatigue scale of the Checklist Individual Strength (sample item: “I feel fatigued.”) (30, 31). This questionnaire has good discriminative validity and good reliability, with a Cronbach's alpha of 0.89 in this study (31, 32).
Functional disability was assessed with the 7-item mobility scale of the IRGL. The 7 items were rated on a 4-point Likert scale ranging from “always” to “never.” A higher score on this scale indicates a lower level of functional disability (sample item: “The past month I was able to walk the stairs.”). The Cronbach's alpha was 0.82 in this study.
The impact of FM on daily life was assessed with the Fibromyalgia Impact Questionnaire (33–35), which consists of 20 items assessing physical impairment, pain, stiffness, fatigue, and anxiety and depressive symptoms (sample item: “How bad has your pain been?”). Higher scores indicate greater impact of FM on daily functioning. The Cronbach's alpha was 0.85 in this study and earlier studies have shown that the reliability and validity are good (33, 36).
Psychological functioning was assessed with the anxiety and negative mood scales of the IRGL. The 10-item anxiety scale assesses the respondent's anxiety levels in the last month, with a Cronbach's alpha of 0.86 in this study (sample item: “I worry too much about unimportant matters.”), and the 6-item negative mood scale assesses the various negative mood states over the previous week, with a Cronbach's alpha of 0.88 in this study (sample item: “How depressed were you during the past week?”).
To give an indication of the results of the patient's perception of improvement, the percentage of patients reporting no improvement, little improvement, fair improvement, and strong improvement with regard to the core outcomes of physical functioning (pain, fatigue, disability in daily life) and psychological functioning (tension, negative mood) were calculated. In addition, the means and SDs for the levels of satisfaction and usefulness of the treatment were calculated. Furthermore, paired t-tests between pre- and posttreatment were conducted on the outcome measures of pain, fatigue, functional disability, and psychological functioning to get an indication of the change during treatment. Standardized residual gain scores were used to assess changes in physical and psychological functioning from pre- to posttreatment (37). These scores were calculated by regressing the outcome variables at the posttreatment assessments on the pretreatment scores of the outcome measures. In order to study the relationship between the pre-post changes and the patient's perception of improvement, Pearson's correlation coefficients were calculated using the entire range of response categories of improvement on every outcome reported by the patients and the pre-post changes on the core outcomes (Table 1). Furthermore, Pearson's correlation coefficients were calculated between the pre-post changes and the patient's scores on usefulness and satisfaction (Table 2). Finally, Pearson's correlation coefficients were calculated between the patient's perception of improvement and their scores on usefulness and satisfaction (Table 3). After checking for possible outliers and normal distribution, the Pearson's correlation coefficients were calculated. All analyses were conducted with the entire sample of 78 patients, with the exception of satisfaction and usefulness. The question about the level of satisfaction was added to the evaluation questionnaire a few weeks after the start of the study, resulting in available data for 51 patients, and results on the level of usefulness were available for 72 patients due to missing data.
Table 1. Pearson's correlation coefficients between the patient's perception of improvement and the pre-post change scores on the outcome measures
Table 2. Pearson's correlation coefficients between the levels of satisfaction and usefulness with the treatment from the patient's perspective and pre-post change scores on the outcome criteria
Table 3. Pearson's correlation coefficients between the levels of usefulness and satisfaction with the treatment from the patient's perspective and the patient's perception of improvement on the outcome criteria
Change during treatment.
The results of the patient's perception of treatment gain showed a mean ± SD satisfaction score of 8.1 ± 1.4 and a mean ± SD score of 8.4 ± 1.4 on usefulness of the treatment. Furthermore, approximately 75% of the participants reported a subjective improvement on physical functioning (pain: 27% no improvement, 36% little improvement, 28% fair improvement, 9% strong improvement; fatigue: 27% no improvement, 50% little improvement, 18% fair improvement, 5% strong improvement; functional disability: 11% no improvement, 35% little improvement, 45% fair improvement, 9% strong improvement). The majority of the participants reported an improvement on psychological functioning (tension: 10% no improvement, 31% little improvement, 44% fair improvement, 15% strong improvement; negative mood: 5% no improvement, 36% little improvement, 50% fair improvement, 9% strong improvement). With regard to the pre-post measures, paired t-tests showed a significant improvement at posttreatment on all of the outcome measures, specifically pain (mean ± SD change −2.8 ± 4.3; t = 5.68, P < 0.001), fatigue (mean ± SD change −10.6 ± 10.7; t = 8.70, P < 0.001), functional disability (mean ± SD change 3.3 ± 3.6; t = −8.10, P < 0.001), negative mood (mean ± SD change −2.7 ± 3.8; t = 6.23, P < 0.001), and anxiety (mean ± SD change −3.3 ± 4.8; t = 6.07, P < 0.001).
Patient's perception of treatment gain and pre-post change.
With regard to the relationship between the patient's perception of improvement and the pre-post change, Pearson's correlation coefficients revealed a significant correlation between the patient's perception of improvement on pain with the pre-post change score on pain (r = 0.69, P < 0.01), between the patient's perception of improvement on fatigue and the pre-post change score on fatigue (r = 0.53, P < 0.01), and also between the patient's perception of improvement on functional disability with the pre-post change scores on functional disability (r = −0.41, P < 0.01) (Table 1). With regard to the impact of FM on daily life, significant correlations were also found between the patient's perception of improvement on all of the outcomes. A small but significant correlation was found between the patient's perception of improvement on negative mood with the pre-post change scores on this outcome measure (r = 0.27, P < 0.05), whereas the correlation between the patient's perception of improvement on tension with the pre-post change score on anxiety was not significant (r = 0.17, P > 0.05).
In order to investigate the relationship between the level of satisfaction and usefulness with treatment and pre-post changes, Pearson's correlation coefficients were calculated (Table 2). The level of satisfaction was significantly correlated with the pre-post change scores on the impact of FM on daily life and the physical outcome measures of pain and fatigue (r = 0.30, P < 0.01). A trend was found for functional disability (r = −0.25, P = 0.07). However, nonsignificant correlations were found on negative mood and anxiety. With regard to the level of usefulness, significant correlations were found with the impact of FM on daily life and the physical outcome measures of pain, fatigue, and functional disability. Nonsignificant correlations were found for negative mood and anxiety.
Patient's perception of improvement and satisfaction and usefulness.
Finally, to study the relationship between the patient's perception of improvement on pain, fatigue, functional disability, negative mood, and anxiety, and the levels of satisfaction and usefulness, Pearson's correlation coefficients were calculated (Table 3). Significant correlations were found between satisfaction and usefulness and the patient's perception of improvement on all of the outcomes (satisfaction: r = 0.31–0.46; usefulness: r = 0.36–0.47), with the exception of the relationship between satisfaction and pain, for which a trend was found (r = 0.27, P = 0.06).
The main goal of this study was to investigate the relationship between the patient's perspective of treatment gain (patient's perception of improvement, satisfaction, and usefulness) and the pre-post change scores during treatment, assessed with the validated self-report outcome measures. The results showed moderate to relatively high significant correlations between the patient's perception of improvement and the pre-post change scores with regard to the physical outcomes (pain, fatigue, and functional disability) and for the impact of FM on daily life, whereas small significant or even nonsignificant correlations were found for the psychological outcomes (negative mood and anxiety). Similar results were found between the levels of satisfaction and usefulness and the pre-post change scores, i.e., significant correlations were found with regard to physical outcomes and nonsignificant correlations were found with regard to psychological outcomes. In conclusion, the patient's perception on treatment gain and patient-reported pre-post changes in outcome criteria should be considered as relatively independent constructs when investigating treatment efficacy, especially when examining psychological functioning.
Several explanations could be offered for this moderate association between the patient's perception of treatment gain and the pre-post change assessed with validated self-report measures. First, other outcomes that emphasize the impact of FM on daily functioning or other factors such as the patient-doctor relationship or trust in treatment could be more important from the patient's point of view (7). This might be particularly relevant for psychological outcomes. Second, a change in how people perceive their health (response shift) might account for this finding because the internal standard of well-being might have shifted due to the improved functioning during treatment, interfering with the postassessment and making the comparison with retrospectively reported improvement difficult. In addition, patients received a cognitive–behavioral intervention aimed at changing cognitions and behaviors, which might have directly influenced their internal standard of well-being. Our findings suggest that a response shift might be especially important when assessing psychological symptoms. Future research is needed to investigate the magnitude and significance of response mechanisms in both pharmacologic and nonpharmacologic interventions. Third, the patient's perception of treatment gain was assessed with a single question for every outcome, making the results more mood dependent and susceptible to subjective interpretations (38). For example, a significant correlation was found between the pre-post change in anxiety and the patient's perception of improvement in depressive symptoms and fatigue, but not the patient's perception of improvement in tension, suggesting that the single item on tension might not cover the key properties of the anxiety concept. Furthermore, the patient's perception of improvement is retrospectively reported, which could also be influenced by factors such as a recall bias or difficulties recalling the baseline status (19). Fourth, it is overall well known that concepts such as pain and functional disability show some conceptual overlap. For example, in this study, significant correlations were found across all of the dimensions, with small to moderate intercorrelations between indicators of physical and psychological functioning (0.20 ≤ r ≤ 0.52). Although in the development of health status instruments the relative independence of these factors is shown through factor analyses, these results suggest that the concepts of pain, fatigue, functional disability, negative mood, and anxiety might be interpreted by the patients as a broader concept of overall symptoms or improvement. Furthermore, there could also be a discrepancy between the pre-post change score on validated instruments and the extent to which these changes are noticeable and meaningful for the patient. Finally, the discrepancy was especially apparent with regard to the psychological outcomes of distress and it could be argued that patients with conditions characterized by chronic physical symptoms are more reflective about their physical functioning than their psychological functioning.
The results of our study suggest that pre-post changes and the patient's perception of treatment gain encompass different constructs and are both relevant when investigating treatment efficacy. Pre-post measures are usually considered the gold standard in clinical research because they are assessed with validated, reliable measures that can be easily compared across studies. However, little is known about the magnitude and significance of response shifts, which could influence these assessments and the clinical relevance of these measures. Patient satisfaction and retrospective reports of improvement correspond closely to the patient's point of view, and are therefore clinically relevant. Furthermore, the patient's perception of treatment gain might also offer an important indicator for treatment adherence because studies have indicated that satisfied patients are more likely to be compliant with the treatment and maintain newly acquired behavior (24). The disadvantages of these measures are the possible influences of recall bias, social desirability, and response shift mechanisms. Therefore, including both the patient's perception of treatment gain and the self-reported pre-post outcomes seems to be particularly important with regard to the clarification of mechanisms influencing response tendencies, assessment of psychological functioning, and the feasibility of newly developed treatment approaches. In addition, there is a need for concepts that incorporate the patient's perception on treatment gain more clearly, such as the patient-perceived satisfactory improvement (39), which includes satisfactory and meaningful improvements for the individual. Overall, more research is needed to study these concepts and the clinical relevance from the perspective of the patient in more detail.
Some limitations of the present study can be mentioned. In our study, we found preliminary support that the patient's perception of treatment gain and the pre-post changes encompass different constructs using self-report questionnaires. However, although concepts of patient satisfaction and the perception of treatment gain are increasingly applied, more research is needed to investigate the level of correspondence with other methods such as in-depth interviews and behavioral measures or clinical judgments (40, 41). Furthermore, the relatively low level of correspondence could also be due to methodologic issues of the questionnaires. For example, the validity and reliability of the evaluation questionnaire were not extensively studied and the questions on the patient's perception of improvement, such as the tension/anxiety concept, might be improved. Finally, the level of correspondence between the patient's perception of treatment gain and the standardized, self-reported outcome criteria should be investigated in other medical conditions and with other treatments. The repeatedly reported low to moderate correlations between the patient's perception of treatment gain and validated outcome measures in FM could be different in medical conditions with other pathology, such as rheumatoid arthritis.
In conclusion, the present study demonstrates that the patient's perception of treatment gain is only moderately related to pre-post changes on the outcome measures, suggesting that these are relatively distinctive concepts and should not be used interchangeably. Future research should aim at incorporating the patient's perspective in addition to validated measures for the assessment of outcome criteria.
All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Ms van Koulil had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study conception and design. Van Koulil, Kraaimaat, van Lankveld, van Riel, Evers.
Acquisition of data. Van Koulil, Kraaimaat, van Lankveld, van Riel, Evers.
Analysis and interpretation of data. Van Koulil, Kraaimaat, van Lankveld, van Riel, Evers.
We would like to thank M. Effting, H. Cats, T. van Helmond, A. Vedder, H. van Hoorn, M. Limborgh, A. J. L. de Jong, J. F. Haverman, and K. J. Korff for their help in collecting the research data.