ORIGINAL ARTICLE

Pain Self-efficacy Mediates the Relationship Between Depressive Symptoms and Pain Severity Jay R. Skidmore, PhD, Alex L. Koenig, MS, Sara J. Dyson, MA, Amy E. Kupper, MS, Melissa J. Garner, MA, and Christopher J. Keller, MA, MS

Objectives: We examined the relationships between depressive symptoms, pain severity, and pain self-efficacy (PSE) in patients with chronic low back pain (CLBP). We hypothesized that change in depressive symptoms would significantly influence change in pain severity, and that PSE indirectly affects this relationship. Materials and Methods: Participants were 109 CLBP patients in a 4week multidisciplinary rehabilitation program for CLBP. They completed measures of PSE, depression, and pain severity at admission and discharge. Structural equation modeling was used to test the significant direct and indirect effects from pretreatment to posttreatment. Results: Change in depressive symptoms significantly predicted change in pain severity in affective (b = 0.358; 95% confidence interval [CI], 0.206-0.480; P = 0.006), sensory (b = 0.384; 95% CI, 0.257-0.523; P = 0.002), and evaluative pain (b = 0.456; 95% CI, 0.285-0.605; P = 0.002). The indirect effects of change in PSE partially accounted for the relationship between change in depressive symptoms and change in sensory (b = 0.105; 95% CI, 0.016-0.241; P = 0.023) and evaluative pain (b = 0.121; 95% CI, 0.010-0.249; P = 0.040). The relationship between change in depressive symptoms and change in affective pain was fully accounted for by the indirect effect of change in PSE (b = 0.203; 95% CI, 0.082-0.337; P = 0.002). Discussion: These findings suggest that pain management and rehabilitation programs for CLBP should specifically target PSE as a key aspect of treatment. Key Words: chronic low back pain, pain self-efficacy, multidisciplinary, depression

(Clin J Pain 2015;31:137–144)

C

hronic pain is a serious and costly health concern in the United States.1–3 The World Health Organization finds that 20% of the world population experiences back pain annually.4 About 60% to 80% of doctor visits are related to a pain and the estimated prevalence of chronic pain ranges from 7% to 30%.5–7 Chronic low back pain (CLBP) is the primary cause of disability and absenteeism in the American workplace,4 for which effective care often requires a comprehensive multidisciplinary approach to treatment.3,8,9 The importance of pain self-efficacy (PSE)—defined as the patient’s confidence in his or her ability to tolerate pain and perform daily activities despite their pain—has been well documented.10–15 Self-efficacy beliefs are often used to predict pain tolerance and aspects of pain experience,11,15,16 and have Received for publication December 20, 2013; revised March 26, 2014; accepted March 1, 2014. From the Clinical Psychology Department, Seattle Pacific University, Seattle, WA. The authors declare no conflict of interest. Reprints: Jay R. Skidmore, PhD, Seattle Pacific University, 3307 Third Ave, West, Seattle, WA 98119 (e-mail: [email protected]). Copyright r 2014 Wolters Kluwer Health, Inc. All rights reserved. DOI: 10.1097/AJP.0000000000000094

Clin J Pain



Volume 31, Number 2, February 2015

been shown to be associated with pain-related variables in individuals with osteoarthritis,17 fibromyalgia,18 rheumatoid arthritis,19 and CLBP.20 Higher levels of PSE are related to positive health outcomes such as lower pain ratings,17 fewer pain behaviors,18,19 and reduced levels of disability.20 Conversely, lower levels of PSE are associated with negative outcomes, including higher levels of anxiety and depressive symptoms, and lower levels of social support.21 Psychological interventions, such as cognitive-behavior therapy (CBT), have proven useful in reducing distress, disability, and psychological morbidity associated with chronic pain.22,23 In addition, multidisciplinary CLBP rehabilitation programs that include aspects of CBT have been shown to significantly improve patients’ quality of life, level of disability, and mood.22–25 Among the important psychological factors, previous research suggests that PSE is an important mechanism in reducing depressive symptoms, disability, and pain severity.8–10 Likewise, depression has been an oft-studied and yet inconsistent predictor of outcomes in multidisciplinary rehabilitation.25,26 The negative influence of pretreatment depression as a predictor of treatment outcomes in more psychologically minded or multidisciplinary CLBP treatment remains unclear. Even fewer studies have investigated PSE and depressive symptoms using multidisciplinary measures for pain (ie, affective, sensory, and evaluative pain). Thus, additional research is needed to further disentangle (1) the use of depressive symptoms as a predictor of CLBP treatment outcomes, and (2) how PSE influences the relationship between depressive symptoms and pain severity. The aim of this study was to examine how change in depressive symptoms relates to change in PSE over the course of a multidisciplinary pain rehabilitation program, and whether or not PSE indirectly affects the relationship between change in depressive symptoms and change in pain severity. We hypothesized that change in depressive symptoms over the course of treatment would be significantly related to change in pain severity (ie, affective, sensory, and evaluative pain) and change in PSE. In addition, we predicted that there would be significant indirect effects (Although previous research has referred to statistical tests of mediation, we will use the language “indirect effects” to refer to mediational hypotheses as articulated in Smith.27) of PSE on the relationship between depressive symptoms and pain severity (ie, affective, sensory, and evaluative pain).

MATERIALS AND METHODS Participants We used archived data from a multidisciplinary rehabilitation and pain management program in Surrey, England. In this study, individuals were eligible for participation if they (1) were 18 years or older, (2) were able to www.clinicalpain.com |

137

Clin J Pain

Skidmore et al

read, write, and understand English, (3) had persistent low back pain for at least 6 months, (4) had pain associated with a medical condition that was not expected to improve with medical or surgical treatment, and (5) were referred by their general physician. Individuals were ineligible for participation if they were medically unstable or nonambulatory. Others considered ineligible were those who might have a severe personality disorder or psychosis, at a serious risk of suicide, unable to manage their own affairs, or who had substance abuse or addictions severe enough to warrant detoxification. The eligible participants in this study were 113 CLBP patients who completed treatment in a multidisciplinary pain rehabilitation center, of which 109 completed enough data to allow for adequate analysis. All patients included in this study had a primary diagnosis of CLBP and their low back pain had persisted for at least 6 months. Only patients with lower back pain were included in this study; those who had lower back pain in addition to other sites of pain were not included. These patients are similar in most respects to clinical CLBP populations in previous research regarding age, sex ratio, duration of back pain, and number of back surgeries.20,22,25,28 Patients ranged in age from 18 to 72 years old (M = 42.6, SD = 12.0). Sixty-eight percent of the participants were female and 32% were male. The number of prior back surgeries ranged from 0 to 10 (M = 1.1, SD = 1.8). The length of time that patients reported they had experienced pain was from 1 to 50 years (M = 9.6, SD = 10.4). These data were obtained nonintrusively as all patients completed measures for this study (see the Procedure section below) as a standard component of intake evaluation before treatment and then again when discharged after treatment. Rehabilitation for CLBP patients at this pain center involved 4 weeks of residential multidisciplinary treatment (with patients typically going home on the weekends).

Volume 31, Number 2, February 2015

PSE The Pain Self-Efficacy Questionnaire (PSEQ) is a measure of generalized PSE that assesses the confidence of individuals to perform daily activities, despite experiencing pain.28 The PSEQ contains 10 items rated on a 7-point Likert scale from 0 (not at all confident) to 6 (completely confident). Sample items include “I can enjoy things, despite the pain” and “I can still accomplish most of my goals in life, despite the pain.” Studies of the psychometric properties of the PSEQ have demonstrated its reliability and validity.28 The PSEQ has high internal consistency, with a Cronbach a of 0.92.

Pain Severity The short form McGill Pain Questionnaire (SF-MPQ) was developed to assess pain.29 The measure contains 15 items that describe different aspects of pain, and 2 subscales: sensory pain (11 items) and affective pain (4 items). Each item is rated on a 4-point Likert scale from 0 (none) to 3 (severe), and sample items include “Burning,” “Throbbing,” and “Sickening.” The sum of all items is the evaluative pain score, which ranges from 0 (no pain) to 45 (severe pain). The sensory aspect of the pain experience is described in terms of temporal, spatial, thermal, and pressure-like properties. The affective experience of pain involves the emotional aspects of pain, such as fear, tension, and autonomic properties. Finally, the evaluative component of the pain experience consists of a subjective overall intensity combining both the sensory and affective qualities of pain. The SF-MPQ has been widely used in studies related to chronic pain, and has sound psychometric qualities.29 In past research, the Cronbach a has been found to be 0.96 for the sensory scale and 0.97 for the affective scale.30

Depressive Symptoms

Procedure Patients who met the above inclusion criteria were informed that their participation in the rehabilitation program was voluntary before admission and that they had the option of seeking treatment elsewhere or withdrawing from treatment at any time. Informed consent was obtained from all patients, who understood that their deidentified data might be used for the purposes of ongoing research. Multidisciplinary treatment included CBT, relaxation techniques (such as progressive muscle relaxation, hypnosis, and biofeedback), counseling for self-management of medications, as well as physical and occupational therapies. Treatment plans were adjusted to each patient’s specific needs, and care was provided by a team of physicians, psychologists, nurses, physical therapists, and occupational therapists. During the treatment program, patients participated in approximately 6 hours of group and individual treatments per day with clinicians from each discipline. Clinicians met together every day before the program sessions began to coordinate and integrate patients’ specific therapy needs. The goals of the program were to promote self-management of pain symptoms, improve physical fitness, and facilitate return to normal activities.

Measures Participants in this study completed the measures below at a pretreatment evaluation and again at a posttreatment discharge assessment.

138 | www.clinicalpain.com



The Beck Depression Inventory (BDI) is a self-report measure that was originally developed by Beck et al31 for assessing depression in clinical populations. It consists of 21 items, and participants are asked to rate each item on a scale from 0 (not present) to 3 (present) according to how they have been feeling over the past week. Sample items include “I do not feel sad” and “I am so sad or unhappy that I can’t stand it.” The BDI yields a total score ranging from 0 to 63 wherein higher scores indicate greater depressive symptoms. A number of studies support the validity and other psychometric properties of the BDI.31–39 Internal reliability, as measured by the Cronbach a, has been reported at 0.90 to 0.92 in the general population as well as in psychiatric populations.

Data Analysis We managed missing data using multiple imputations. Preliminary analyses assessed the normality of the data, and also included an examination of means and SD for measures in the study. Structural equation modeling was used to test the indirect effects of PSE on the relationship between depression and pain severity. To test the indirect effects of PSE in the hypothesized model (Fig. 1), we used a bootstrap sampling procedure and examined the indirect effects of change in PSE on the relationship between change in depressive symptoms and change in affective, sensory, and evaluative pain.40,41 We used Amos 19.0 to evaluate the indirect effects of change in PSE on the relationship

Copyright

r

2014 Wolters Kluwer Health, Inc. All rights reserved.

Clin J Pain



Volume 31, Number 2, February 2015

Pain Self-efficacy and Depression

Δ Affective Pain

Δ Pain Self-Efficacy Δ Depressive Symptoms Δ Sensory Pain

Δ Evaluative Pain

FIGURE 1. Hypothesized model of relationships between change in depressive symptoms, change in pain self-efficacy, and change in affective pain, sensory pain, and evaluative pain.

between change in depressive symptoms and change in pain severity (Fig. 1).

RESULTS Statistics and Data Analysis Preliminary Analyses Normality and outlier analysis: Data were analyzed using the Statistical Package for the Social Sciences Version 16.0 (SPSS, 2007). Multiple imputation42 was used to manage missing data because Amos does not allow missing data when using bootstrapping. Cases with 25% or more missing data were eliminated (n = 4). Given that structural equation modeling is based on analysis of covariance and that kurtosis affects tests of variance and covariance, an analysis of normality and deletion of outliers is relevant to this study. Thus, after data were combined and entered into SPSS, several steps were taken to assess the normality of the data for the 109 participants using procedures outlined by Byrne43 and Field.44 The following procedures were used to help identify the presence of any extreme outliers. First, the data were inspected visually using frequency histograms, normal

curves, P-P plots, Q-Q plots, and box plots. Second, we examined z-scores for the skewness and kurtosis of variables. We also examined the descriptive data described below to ensure that the values for all variables were within expected parameters (eg, that the range of the data did not exceed that which would be expected for a given measure). We concluded that there was not sufficient evidence to support the deletion of any outliers. Descriptive statistics: Descriptive statistics including the means, SDs, and bivariate correlations for each of the measures used in this study are shown in Tables 1 and 2. Examination of pain predictors: We used linear regression to predict pain outcomes from baseline PSE and depression scores. Results showed that higher PSE at baseline predicted less evaluative pain at discharge (b = 0.368; 95% CI, 5.044 to  1.456; P = 0.001). In addition, higher depression scores at baseline predicted greater evaluative pain at discharge (b = 0.204; 95% CI, 0.028 to 4.078; P = 0.053). Calculation of residualized change scores: To create residualized change scores, we used linear regression to predict discharge scores from the admission scores for the same variable, saving the standardized and unstandardized residual values as new variables. As described by MacKinnon,45 these

TABLE 1. Means (SDs) and Pearson Correlation Values for the BDI, PSEQ, MPQa, MPQs, and MPQe at Admission BDI PSEQ MPQa MPQs MPQe

BDI

PSEQ

MPQa

MPQs

MPQe

22.34 (10.88)

0.51** 21.39 (9.73)

0.36**  0.29** 5.13 (3.28)

0.26**  0.01 0.52** 14.45 (6.94)

0.36** 0.21** 0.37** 0.18** 3.46 (0.93)

**P < 0.01. MSQa indiates McGill Pain Questionnaire - Affective; MPQs, McGill Pain Questionnaire - Sensory; MPQe, McGill Pain Questionnaire - Evaluative.

Copyright

r

2014 Wolters Kluwer Health, Inc. All rights reserved.

www.clinicalpain.com |

139

Clin J Pain

Skidmore et al



Volume 31, Number 2, February 2015

TABLE 2. Means (SDs) and Pearson Correlation Values for the BDI, PSEQ, MPQa, MPQs, and MPQe at Discharge BDI PSEQ MPQa MPQs MPQe

BDI

PSEQ

MPQa

MPQs

MPQe

12.56 (11.14)

0.66** 40.27 (11.84)

0.42**  0.45** 2.98 (2.74)

0.42**  0.38** 0.65** 11.03 (6.22)

0.50** 0.53** 0.41** 0.45** 2.42 (1.10)

**P < 0.01. MSQa indicates McGill Pain Questionnaire - Affective; MPQs, McGill Pain Questionnaire - Sensory; MPQe, McGill Pain Questionnaire - Evaluative.

residual values represent an index of change for each variable because they are equivalent to the variability in the discharge scores after accounting for the portion of their variability that is predicted by the admission scores. Descriptive data for these residualized change scores along with correlations are presented in Table 3.

Primary Analysis Tests of direct and indirect effects: We used a bootstrap sampling procedure, as described by Mallinckrodt et al40 to determine the significance of indirect effects. This process involves using the sample as a population reservoir from which a large number of random samples are drawn and continuously replaced so that they have an equal likelihood of being randomly selected on all subsequent drawings.40 In the present study, we specified 1000 bootstrap iterations and followed the suggestions by Mallinckrodt et al40 using 95% bias-corrected confidence intervals and bootstrap estimates of indirect, direct, and total effects. To test the hypothesized indirect effects, we used standardized residualized change scores in place of each variable as indices of change over treatment (Fig. 2). We examined the bivariate a, b, and c0 paths of the hypothesized mediated models, as well as the c (total effect) paths and indirect effects. A relationship is considered to be mediated if the indirect effect is statistically significant and if the direct effect decreases when the mediator is included in the model (ie, when there is a decrease from the total effect, c to c0 , which includes the mediator). The standardized results from the hypothesized model are shown in Table 4. First, model-fit was examined using several indicators. A w2 statistic that is statistically significant (P < 0.05) indicates that the population covariance matrix differs significantly from the sample covariance matrix.43 For this model, w2 (w2 = 40.887, df = 3,

P < 0.001), indicated unsatisfactory fit. However, because the w2 statistic is known to have many problems, other fit indices46 were also examined, which revealed that the fit of the final tested model to the data was poor (GFI = 0.877; CFI = 0.739; RMSEA = 0.301). However, we evaluated modification indices43 to locate parameters that might be freed to covary. Because modification indices are statistically driven, we only made respecifications when substantive rationale supported the parameter change; and we decided to allow a covariance between D2 and D3. The final model is shown in Figure 3. Overall, examination of multiple model-fit indicators revealed that the fit of the final tested model to the data was good (w2 = 5.129, df = 2, P = 0.077; GFI = 0.982; CFI = 0.979; RMSEA = 0.120). All direct effects for the hypothesized model were statistically significant (Table 4). In addition, indirect effects of the hypothesized model were significant for affective pain (b = 0.203; 95% CI, 0.082-0.337; P = 0.002), sensory pain (b = 0.105; 95% CI, 0.016-0.241; P = 0.023), and evaluative pain (b = 0.121; 95% CI, 0.010-0.249; P = 0.040). The total effects, or c paths, of depressive symptoms on (1) sensory pain, and (2) evaluative pain decreased and remained significant when controlling for PSE (c0 paths) indicating partial mediation. However, the total effect, or c path, of depressive symptoms on affective pain decreased and was no longer significant when PSE was also considered. As with the hypothesized model (Table 4), all paths were statistically significant for the final model except the total effect, or c path, of depressive symptoms on affective pain (Table 5). The results of this analysis suggest that the relationships between the change in depressive symptoms scores and the changes in (1) sensory pain and (2) evaluative pain are partially mediated by change in PSE over the course of treatment. In addition, the relationship between change in depressive symptom scores and change

TABLE 3. Pearson Correlation Values for the Standardized Residualized Change Scores for the BDI, PSEQ, MPQa, MPQs, and MPQe BDI PSEQ MPQa MPQs MPQe

BDI

PSEQ

MPQa

MPQs

MPQe

9.93 (10.89)

0.58**  18.88 (12.08)

0.39**  0.43** 2.27 (3.52)

0.41**  0.35** 0.60** 3.54 (6.28)

0.44** 0.39** 0.30** 0.39** 1.05 (1.24)

Means (SDs) of raw difference scores are displayed on the diagonal. Means and SDs of raw difference scores (rather than standardized residualized change scores) are presented because means of standardized residualized change scores are equal to 0 and SDs are equal to 1 (they are deviations from a best fit regression line). **P < 0.01. BDI indicates Beck depression inventory; PSEQ, Pain Self-Efficacy Questionnaire; MSQa, McGill Pain Questionnaire - Affective; MPQs, McGill Pain Questionnaire - Sensory; MPQe, McGill Pain Questionnaire - Evaluative.

140 | www.clinicalpain.com

Copyright

r

2014 Wolters Kluwer Health, Inc. All rights reserved.

Clin J Pain



Volume 31, Number 2, February 2015

Pain Self-efficacy and Depression

D2

Δ Affective Pain

D1 -.35

.17

D3 -.58

Δ Depressive Symptoms

Δ Pain Self-Efficacy -.18 .29

Δ Sensory Pain

.34

-.20 D4

Δ Evaluative Pain

FIGURE 2. Standardized results from the analysis of the hypothesized model.

in affective pain was fully mediated by change in PSE over the course of treatment.

DISCUSSION The goal of the present study was to examine the relationship between change in depressive symptoms and change in pain severity, and to evaluate whether change in PSE indirectly affects this relationship. These findings offer support for both of our hypotheses. Consistent with our first hypothesis, change in depressive symptoms predicted change in pain severity. Specifically, higher levels of depressive symptoms significantly predicted higher pain severity (ie, affective, sensory, and evaluative). Thus,

participants who had fewer depressive symptoms rated their pain as less severe over the course of treatment. Our second hypothesis, that change in PSE would indirectly affect the relationship between change in depressive symptoms and change in pain severity, was also supported. Specifically, the relationship between change in depressive symptoms and change in sensory and evaluative pain was partially accounted for by change in PSE. Thus, change in depressive symptoms was related to change in sensory and evaluative pain partially because of the change that occurred in PSE beliefs. Participants who had fewer depressive symptoms over the course of treatment rated their sensory and evaluative pain as less severe, at least in part because of their increased confidence to tolerate pain and perform daily activities despite their pain. In addition,

TABLE 4. Hypothesized Model—Bootstrap Results to Test Significance of Direct and Indirect Effects (Standardized Values)

95% CI Path/Effect a BDI-PSEQ b PSEQ-MPQa b PSEQ-MPQs b PSEQ-MPQe c (Total effect) BDI-MPQa c (Total effect) BDI-MPQs c (Total effect) BDI-MPQe c0 BDI-MPQa c0 BDI-MPQs c0 BDI-MPQe a b (indirect effect) BDI-MPQa a b (indirect effect) BDI-MPQs a b (indirect effect) BDI-MPQe

b

SE

Lower

Upper

P

 0.577  0.347  0.177  0.201 0.358 0.384 0.456 0.167 0.286 0.337 0.203 0.105 0.121

0.062 0.102 0.090 0.103 0.103 0.085 0.080 0.093 0.092 0.101 0.068 0.055 0.060

 0.687  0.540  0.378  0.411 0.206 0.257 0.285  0.064 0.100 0.143 0.082 0.016 0.010

0.428 0.145 0.021 0.012 0.480 0.523 0.605 0.308 0.456 0.530 0.357 0.241 0.249

0.003 0.002 0.028 0.037 0.006 0.002 0.002 0.199 0.008 0.002 0.002 0.023 0.032

BDI indicates Beck depression inventory; CI, confidence interval; PSEQ, Pain Self-Efficacy Questionnaire; MSQa, McGill Pain Questionnaire - Affective; MPQs, McGill Pain Questionnaire - Sensory; MPQe, McGill Pain Questionnaire - Evaluative.

Copyright

r

2014 Wolters Kluwer Health, Inc. All rights reserved.

www.clinicalpain.com |

141

Clin J Pain

Skidmore et al



Volume 31, Number 2, February 2015

D2

Δ Affective Pain

D1

.40

-.35

.17

D3 -.58 Δ Depressive Symptoms

Δ Pain Self-Efficacy -.18 .29 Δ Sensory Pain .34

-.20 D4

Δ Evaluative Pain

FIGURE 3. Standardized results from the analysis of the final model.

the relationship between change in depressive symptoms and change in affective pain was explained by change in PSE. For instance, participants who had fewer depressive symptoms over the course of treatment rated their affective pain as lower at intake than at discharge because their PSE increased. These findings suggest that increased confidence in the ability to tolerate pain and perform daily activities despite pain severity is related to decreases in depressive symptoms over treatment and, thus, lessens emotional difficulty and tension related to the pain experience. These findings highlight the importance of targeting self-efficacy beliefs as an intervention strategy for reducing depressive symptoms and pain severity in CLBP patients. Previous research has shown that PSE may have more contribution to how or why someone becomes depressed

than pain intensity and disability alone.47,48 However, our findings suggest that pain severity, in terms of the emotional experience of pain, can actually decrease as a result of self-efficacy beliefs. In addition, the overall intensity and sensory experience of pain may, in part, decrease as a result of self-efficacy beliefs. Consistent with these findings, previous research has also found that changes in depressive symptoms are highly related to changes in pain intensity and disability, suggesting that depressive symptoms are more important to target in therapy than disability or pain intensity alone.25 Furthermore, several studies have shown that depression in chronic pain patients differs from typical clinical depression in other populations.15,49 Specifically, depressed pain patients tend to focus their negative cognitions on their

TABLE 5. Final Model—Bootstrap Results to Test Significance of Direct and Indirect Effects (Standardized Values)

95% CI Path/Effect a BDI-PSEQ b PSEQ-MPQa b PSEQ-MPQs b PSEQ-MPQe c (total effect) BDI-MPQa c (Total effect) BDI-MPQs c (Total effect) BDI-MPQe c0 BDI-MPQa c0 BDI-MPQs c0 BDI-MPQe a b (indirect effect) BDI-MPQa a b (indirect effect) BDI-MPQs a b (indirect effect) BDI-MPQe Covariance D2’-D3

b

SE

Lower

Upper

P

 0.577  0.347  0.177  0.201 0.358 0.384 0.456 0.167 0.286 0.337 0.203 0.105 0.121 0.402

0.062 0.102 0.090 0.103 0.103 0.085 0.080 0.093 0.092 0.101 0.068 0.055 0.060 0.116

 0.687  0.540  0.378  0.411 0.206 0.257 0.285  0.064 0.100 0.143 0.082 0.016 0.010 0.234

0.428 0.145 0.021 0.012 0.480 0.523 0.605 0.308 0.456 0.530 0.357 0.241 0.249 0.688

0.003 0.002 0.028 0.037 0.006 0.002 0.002 0.199 0.008 0.002 0.002 0.023 0.032 0.000

BDI indicates Beck depression inventory; CI, confidence interval; PSEQ, Pain Self-Efficacy Questionnaire; MSQa, McGill Pain Questionnaire - Affective; MPQs, McGill Pain Questionnaire - Sensory; MPQe, McGill Pain Questionnaire - Evaluative.

142 | www.clinicalpain.com

Copyright

r

2014 Wolters Kluwer Health, Inc. All rights reserved.

Clin J Pain



Volume 31, Number 2, February 2015

bodies or health status.50 Our results add to these findings by suggesting that one specific reason that depressive symptoms may decrease in treatment programs for CLBP may be because PSE beliefs were improved. The mechanism of change for self-efficacy beliefs was not examined directly in this study, but in keeping with Bandura et al12,51 we would suggest it was the combination of physical therapy integrated with psychological treatment that promoted the improvement in our patients. More specifically, as patients were encouraged to gradually increase their physical exercise (in terms of both time spent and amount of weight lifted), their own self-observations of increasing physical functioning and ability to use cognitive-behavioral paincoping strategies proved they could do more even if they had some continuing or recurrent pain symptoms. Hence, improvements in PSE were most likely a function of patients observing their own behavioral change during physical activities.

Limitations Although the present study offers valuable findings, it is important to recognize several potential limitations. First, the present study lacked an untreated control group, which would have added important comparison information to our study. Second, we were not able to determine whether the robustness of our findings remained after discharge. Follow-up assessments after discharge (eg, 3, 6, and/or 12 mo) would provide information regarding the effectiveness of self-efficacy interventions. In addition, the findings are based upon self-report measures that are susceptible to a variety of threats to validity. Furthermore, replication of this study in more culturally diverse (racial, ethnic, socioeconomic, and geographical) groups is needed for wider generalizability. It remains unclear as to whether CBT alone targeted PSE or if it was a combination of modalities. It is important to consider that the present treatment program incorporated relaxation techniques (with hypnosis and/or biofeedback) and a wide range of physical and occupational therapies, in addition to CBT. Although increasing PSE should be an integral component to any pain management program, there may be multiple methods to achieving such goals. Despite these limitations, this study has several strengths that make it a unique and important contribution to the pain management research literature.

Implications and Future Research The present study offers important implications for future research. One logical extension of the present research would be to investigate and implement possible strategies that aid in increasing self-efficacy in multidisciplinary pain management programs, to reduce both depressive symptoms and pain severity. In addition, the incorporation of a control group and follow-up measures, as previously mentioned, in future studies would add to the robustness of our present findings and allow us to determine whether the increase in self-efficacy beliefs is maintained. Furthermore, investigation into how each treatment modality and combinations of treatment modalities influences PSE will be necessary for refining pain management programs. Bandura51 suggested that self-efficacy beliefs could be improved through skill mastery, sharing vicarious experiences, verbal persuasion, and providing information about the individual’s physiological and affective states. Copyright

r

Pain Self-efficacy and Depression

Multidisciplinary pain management programs, such as the one in the present study, typically incorporate many physical therapy components and a broad array of cognitivebehavioral techniques in these rehabilitation programs. Yet other factors may also influence pain intensity and disability. The effectiveness of multidisciplinary programs can be influenced by follow-up interviews,24 supervisor experience,52 intervention components, patient characteristics,53 and readiness to change.54 Thus, although self-efficacy may be one target of intervention, there are still many other aspects of the multidisciplinary treatment programs that deserve further investigation.

CONCLUSIONS In summary, our findings suggest that the indirect effects of PSE partially accounted for the relationship between change in depressive symptoms and change in sensory and evaluative pain, and fully explained the relationship between depressive symptoms and change in affective pain. PSE is an important factor in predicting and managing pain severity and depressive symptoms, and thus should be considered when designing and implementing rehabilitation programs for chronic pain. Increasing selfefficacy beliefs over the course of rehabilitation seems to be a central mechanism in the cognitive-behavioral treatment of chronic pain and in reducing pain-related depression. REFERENCES 1. Melzack R. Evolution of the neuromatrix theory of pain. The Prithvi Raj lecture, Presented at the Third World Congress of World Institute of Pain, Barcelona 2004. Pain Pract. 2005; 5:85–94. 2. Turk DC, Gatchel RJ. Psychological Approaches to Pain Management: A Practitioner’s Handbook. New York: Guilford Press; 2002. 3. Turk DC, Meichenbaum D, Genest M. Pain and Behavioral Medicine: A Cognitive-Behavioral Perspective. New York: Guilford Press; 1983. 4. Ehrlich GE. Back pain. J Rheumatol. 2003;67:26–31. 5. Cosser C. Hypnosis in the treatment of chronic pain: an ecosystem approach. Aus J Clin Exp Hypn. 2002;301:156–169. 6. Frischenschlager O, Pulcher I. Psychological management of pain. Disabil Rehabil. 2002;24:416–432. 7. Loeser JD. Low back pain. In: Loeser JD, Butler SH, Chapman CR, Turk DC, eds. Bonica’s Management of Pain. 3rd ed. Philadelphia: Lippincott Williams & Wilkins; 2001: 1508–1509. 8. Meredith P, Strong J, Feeney JA. Adult attachment, anxiety, and pain self-efficacy as predictors of pain intensity and disability. Pain. 2006;123:146–154. 9. Turk DC, Okifuji A. Psychological factors in chronic pain: evolution and revolution. J Consult Clin Psychol. 2002;70: 678–690. 10. Arnstein P. The mediation of disability by self efficacy in different samples of chronic pain patients. Disabil Rehabil. 2000;22:794–801. 11. Asghari A, Nicholas MK. Pain self-efficacy beliefs and pain behaviour. a prospective study. Pain. 2001;94:85–100. 12. Bandura A, O’Leary A, Taylor C, et al. Perceived self-efficacy and pain control: opioid and nonopioid mechanisms. J Pers Soc Psychol. 1987;53:563–571. 13. Morley S, Eccleston C, Williams A. Systematic review and meta-analysis of randomized controlled trials of cognitive behavior therapy and behavior therapy for chronic pain in adults, excluding headache. Pain. 1999;80:1–13. 14. Pincus T, Santos R, Morely S. Depressed cognitions in chronic pain patients are focus on health: evidence from a sentence completion task. Pain. 2007;130:84–92.

2014 Wolters Kluwer Health, Inc. All rights reserved.

www.clinicalpain.com |

143

Clin J Pain

Skidmore et al

15. Rudy T, Lieber S, Boston J, et al. Psychosocial predictors of physical performance in disabled individuals with chronic pain. Clin J Pain. 2003;19:18–30. 16. Turk DC. Clinical effectiveness and cost-effectiveness of treatments for patients with chronic pain. Clin J Pain. 2002;18: 355–365. 17. Keefe FJ, Lefebvre JC, Maixner W, et al. Self-efficacy for arthritis pain: relationship to perception of thermal laboratory pain. Arthritis Care Res. 1997;10:177–184. 18. Buckelew SP, Parker JC, Keefe FJ, et al. Self-efficacy and pain behavior among subjects with fibromyalgia. Pain. 1994;59: 377–384. 19. Buescher K, Johnston J, Parker J, et al. Relationship of selfefficacy to pain behavior. J Rheumatol. 1991;18:968–972. 20. Woby S, Watson P, Roach N, et al. Coping strategy use: does it predict adjustment to chronic back pain after controlling for catastrophic thinking and self-efficacy for pain control? J Rehabil Med. 2005;37:100–107. 21. Geyh S, Nick E, Stirnimann D, et al. Biopsychosocial outcomes in individuals with and without spinal cord injury: a Swiss comparative study. Spinal Cord. 2012;50:614–622. 22. Hoffman BM, Papas RK, Chatkoff DK, et al. Meta-analysis of psychological interventions for chronic low back pain. Health Psychol. 2007;26:1–9. 23. Morley S, Williams AC, Black S. A confirmatory factor analysis of the Beck Depression Inventory in chronic pain. Pain. 2002;99:298–298. 24. Dysvik E, Kvaløy JT, Natvig GK. The effectiveness of an improved multidisciplinary pain management programme: A 6and 12-month follow-up study. J Adv Nurs. 2012;68:1061–1072. 25. Glombiewski J, Hartwich-Tersek J, Rief W. Depression in chronic back pain patients: prediction of pain intensity and pain disability in cognitive-behavioral treatment. Psychosomatics. 2010;51:130–136. 26. van der Hulst M, Vollenbroek MM, Ijzerman MJ. A systematic review of sociodemographic, physical and psychological predictors of multidisciplncary rehabilitation: or, “back school” treatment outcome in patients with chronic low back pain. Spine. 2005;30:813–825. 27. Smith ER. On mediation. J Pers Soc Psychol. 2012;102:1–3. 28. Nicholas MK. The pain self-efficacy questionnaire: taking pain into account. Eur J Pain. 2007;11:153–163. 29. Melzack R. The short-form McGill Pain Questionnaire. Pain. 1987;30:191–197. 30. Green CR, Hart-Johnson T. The association between race and neighborhood socioeconomic status in younger black and white adults with chronic pain. J Pain. 2012;13:176–186. 31. Beck AT, Ward CH, Mendelson M, et al. An inventory for measuring depression. Arch Gen Psychiatry. 1961;4:561–571. 32. Arnau RC, Meagher MW, Norris MP, et al. Psychometric evaluation of the Beck Depression Inventory II with primary care medical patients. Health Psychol. 2001;20:112–119. 33. Beck AT, Steer RA, Brown GK. Manual for the Beck Depression Inventory-II. San Antonio, TX: Psychological Corporation; 1996. 34. Buckley TC, Parker JD, Heggie BS. A psychometric evaluation of the BDI-II in treatment-seeking substance abusers. J Subst Abuse Treat. 2001;20:197–204.

144 | www.clinicalpain.com



Volume 31, Number 2, February 2015

35. Carmody DP. Psychometric characteristics of the Beck Depression Inventory-II with college students of diverse ethnicity. Int J Psychiatry Clin Pract. 2005;9:22–28. 36. Krefetz DG, Steer RA, Gulab NA, et al. Convergent validity of the Beck Depression Inventory-II with the Reynolds Adolescent Depression Scale in psychiatric inpatients. J Pers Asses. 2002;78:451–460. 37. Sprinkle SD, Lurie D, Insko SL, et al. Criterion validity, severity cut scores, and test-retest reliability of the Beck Depression Inventory-II in a university counseling center sample. J Couns Psychol. 2002;49:381–385. 38. Steer RA, Geetha K, Ranieri WF, et al. Use of the Beck Depression Inventory with adolescent psychiatric outpatients. J Psychopathol Behav Assess. 1998;20:127–137. 39. Storch EA, Roberti JW, Roth DA. Factor structure, concurrent validity, and internal consistency of the Beck Depression Inventory-second edition in a sample of college students. Depress Anxiety. 2004;19:187–189. 40. Mallinckrodt B, Abraham WT, Wei M, et al. Advances in testing the statistical significance of mediation effects. J Couns Psychol. 2006;53:372–378. 41. Shrout PW, Bolger N. Mediation in experimental and nonexperimental studies: new procedures and recommendations. Psychol Methods. 2002;7:422–445. 42. Enders CK. Applied Missing Data Analysis. New York: Guilford Press; 2010. 43. Byrne MB. Structural Equation Modeling With AMOS: Basic Concepts, Applications, and Programming. New York: Taylor and Francis Group LLC; 2010. 44. Field AP. Discovering Statistics Using SPSS: (and Sex, Drugs and Rock ’n’ Roll). London: Sage Publications; 2009. 45. MacKinnon DP. Introduction to Statistical Mediation Analysis. New York: Taylor & Francis Group; 2008. 46. Kline RB. Principles and Practice of Structural Equation Modeling. 2nd ed. New York: Guilford Press; 2005. 47. Arnstein P, Caudill M, Mandle CL, et al. Self efficacy as a mediator of the relationship between pain intensity, disability and depression in chronic pain patients. Pain. 1999;80:483–491. 48. Wright GE, Parker JC, Smarr K, et al. Risk factors for depression in rheumatoid arthritis. Arthritis Care Res. 1996;9:264–272. 49. Nicholas M, Coulston C, Asghari A, et al. Depressive symptoms in patients with chronic pain. Med J Aus. 2009; 19:66–70. 50. Pincus T, Williams A. Models and measurements of depression in chronic pain. J Psychosom Res. 1999;47:211–219. 51. Bandura A. Self Efficacy: the Exercise of Control. New York: W.H. Freeman and Company; 1997. 52. Williams ACDC. Cognitive-behavioural pain management: lessons learned. In: McQuai H, Kalso E, More RA, eds. Systematic Reviews in Pain Research: Methodology Refined. Seattle: IASP Press; 2008:275–284. 53. Dijkstra A, Vlaeyen JWS, Rijnen H, et al. Readiness to adopt the self-management approach to cope with chronic pain in fibromyalgic patients. Pain. 2001;90:37–45. 54. Kerns RD, Rosenberg R, Jamison RN, et al. Readiness to adopt a self-management approach to chronic pain: the Pain Stages of Change Questionnaire (PSOCQ). Pain. 1997; 72(1–2):227–234.

Copyright

r

2014 Wolters Kluwer Health, Inc. All rights reserved.