- Email: [email protected]
The development and psychometric evaluation of an instrument to assess spouse responses to pain and well behavior in patients with chronic pain: The Spouse Response Inventory
The Development and Psychometric Evaluation of an Instrument to Assess Spouse Responses to Pain and Well Behavior in Patients With Chronic Pain: The Spouse Response Inventory Lauren Schwartz,* Mark P. Jensen,* and Joan M. Romano† Abstract: Operant behavioral models of chronic pain posit that the pain behaviors and disability of patients with chronic pain can be influenced by social contingencies, such as significant others’ responses to pain and well behaviors. The aim of the present study was to develop and evaluate a comprehensive measure of spouse responses to patient pain and well behaviors, the Spouse Response Inventory (SRI). One hundred four patients with chronic pain and their spouses completed a battery of questionnaires, including the SRI. The final analysis yielded a 39-item inventory divided into 4 scales that assess spouse solicitous and negative responses to patient pain behaviors and spouse encouraging and negative responses to patient well behaviors. Analyses yielded results that were consistent with previous research demonstrating an association between spouse solicitous and negative responses to patient pain behaviors and measures of patient functioning, providing preliminary support for the validity of the SRI scales. The results are discussed in terms of implications for further research and the clinical applicability of the SRI. Perspective: This article presents the psychometric properties of a new measure of spouse responses to patient chronic pain and well behavior. This measure could potentially be helpful to clinicians seeking to assess the extent to which spouse responses may contribute to patient pain and disability and also to researchers who wish to test hypotheses derived from operant theory as applied to chronic pain. © 2005 by the American Pain Society Key words: Chronic pain, marriage, spouse responses.
A
n operant behavioral formulation of chronic pain behavior, first introduced over 3 decades ago,9 has had a significant influence on the understanding and treatment of chronic pain. In this theoretical perspective, pain behaviors (eg, grimacing, limping, or guarding) might come under the control of reinforcing consequences in the environment (eg, social responses of significant others and avoidance of stressful interpersonal or work situations) and be maintained, at least in part, by such consequences. Hence a potential explanation for continued pain behavior in some individuals with chronic pain is the ongoing positive or negative reinforcement of pain behaviors.8 Such operant behavioral principles have been incorporated into many multidisciplinary chronic pain treatment programs5,18 and serve as the theoretical basis for a number of interven-
Received June 23, 2004; Revised November 9, 2004; Accepted December 21, 2004. From the Departments of *Rehabilitation Medicine and †Psychiatry and Behavioral Sciences, University of Washington School of Medicine, Seattle, WA. Address reprint requests to Mark P. Jensen, Department of Rehabilitation Medicine, Box 356490, University of Washington School of Medicine, Seattle, WA 98195-6490. E-mail: [email protected] 1526-5900/$30.00 © 2005 by the American Pain Society doi:10.1016/j.jpain.2004.12.010
tions, such as the behavioral shaping of increased activity through quota-based reactivation. Correlational research supports a primary hypothesis of the operant model: the responses of the social environment can serve as reinforcers of pain behavior and thus play a role in maintaining chronic pain behaviors and dysfunction. For example, observed spouse solicitous responses to patient nonverbal pain behaviors are associated with higher rates of observed pain behaviors in patients reporting greater pain and with greater disability in more depressed patients.25 Spouse solicitousness as reported by patients4,31 and spouses19 has also been associated with higher self-reports of patient pain intensity4,19 and observed patient pain behaviors.31 Kerns and colleagues14 found a similar relationship between reported spouse solicitousness and pain severity for maritally satisfied pain patients and spouses. Flor et al6 found that greater spouse solicitousness was associated with lower activity levels in patients with chronic pain. Findings such as these are consistent with the assumption that solicitous attention to patient pain behaviors from spouses might be reinforcing for many patients. Recent experimental research provides even more support for the hypothesis that environmental responses might influence pain behavior. Flor et al7 assessed pain intensity reports and physiologic responding (electroen-
The Journal of Pain, Vol 6, No 4 (April), 2005: pp 243-252
243
244 cephalography, electrooculography, electromyography, and skin conductance level) in 30 patients with chronic pain and 30 healthy control subjects in response to 4 levels of electric shock (individual pain threshold and 25%, 50%, and 75% of the distance between individual pain threshold and pain tolerance, as assessed during pain threshold and tolerance tests). The study participants were then randomly assigned to be given either (1) positive feedback when subsequent ratings for shock were higher than the average rating previously given and negative feedback when these ratings were lower than the previous averages (up-training condition) or (2) negative feedback for higher ratings and positive feedback for lower ratings (down-training condition). Consistent with operant theory, subjects in the up-training condition provided increasing pain reports over time, and those in the down-training condition provided decreased pain reports over time. Moreover, patients with chronic pain showed a slower extinction rate (ie, maintained their relatively higher reports of pain for a longer period of time) than healthy control subjects when reinforcement was discontinued. The authors concluded that patients with chronic pain appear to be more susceptible to the operant conditioning of increasing pain responses than healthy control subjects. In a similar study, Jolliffe and Nicholas13 randomly assigned half of 46 undergraduates to receive positive verbal reinforcement from an experimenter after each trial of a cold pressor task if their report of pain intensity exceeded that of a previous trial. The other half did not receive this reinforcement. Again, and consistent with an operant model, the mean pain intensity ratings of the subjects in the reinforcement condition were greater than those of the nonreinforced subjects. Operant theory also suggests that pain behaviors might be maintained by a process of negative reinforcement, which is defined as the reduction or removal of an aversive stimulus (not to be confused with punishment, which is the application of an aversive stimulus). For example, it has been hypothesized that pain behaviors can lead to decreased demands on the patient to participate in potentially aversive activities, such as work, chores, exercise, or conflictive interpersonal interactions.8,26,27 However, the relationship between these potential negative reinforcers and patient functioning has not yet been systematically examined in previous research. This is most likely because of the lack of a reliable and valid measure of such possible negative reinforcers. Spouse negative responses (eg, irritation and criticism) to patient pain behaviors have also been examined to determine whether such responses might act as punishers and thus reduce the frequency of pain behaviors. The literature, however, has not yielded consistent findings to support this hypothesis. Some studies have found that negative responses by significant others are associated with a higher level of activity.6 Other studies have found that negative responses were associated with poorer patient physical and psychosocial functioning,27 whereas yet other studies have found no significant relationship between partner negative responses and patient disabil-
Measuring Spouse Responses to Pain 28
14,28
ity, pain intensity, or pain behavior.28 The most consistent finding has been that spouse negative responses are associated with greater patient psychosocial distress and depression,14,22,25,27,28 although this finding has been shown to vary with levels of marital satisfaction.14,28 Overall, these findings suggest that spouse negative responses to pain behaviors might not be associated consistently with a reduced frequency of pain behaviors but do appear to be associated more often with greater patient emotional distress. An important implication of the operant model is that spouses can play a role in encouraging and reinforcing better physical functioning and well behaviors (behaviors incompatible with the sick role, such as maintaining an appropriate activity level) in the person with pain.8 For example, spouse support, defined as emotional support and encouragement during a stationary bicycle exercise task, was shown to be associated with greater exercise persistence in a sample of 50 patients.21 Despite the theoretical potential that spouse responses to patient well behaviors could increase or maintain positive functioning in persons with chronic pain, no assessment instrument has yet been developed to systematically assess spouse responses to patient well behaviors. The most frequently used measures of spouse responses to patient pain behaviors are the Spouse Response subscales from the West Haven–Yale Multidimensional Pain Inventory (WHYMPI).16 The WHYMPI has many strengths. It is theoretically based, is well validated, and measures 3 types of spouse responses to patient pain behavior (solicitous, negative or “aversive,” and distracting responses). Moreover, the solicitous subscale of the WHYMPI is associated positively with observed spouse solicitous behaviors.24 However, the WHYMPI does not, in its current form, differentiate between spouse responses that are potentially positively reinforcing from those that are potentially negatively reinforcing. There might be important differences in how the provision of solicitous responses (potential positive reinforcers) and the removal of aversive responsibilities or conflicts (potential negative reinforcers) might affect patient function. In addition, the characteristics of couples in which these different response patterns might occur could differ in ways that might influence patient and spouse adjustment. Finally, the WHYMPI does not assess spouse responses (either positive or negative) to patient well behaviors. Thus there is a need for a more comprehensive measure to assess these dimensions of spouse responses to patient behavior. The primary goal of the present study was to develop and perform a preliminary evaluation of a comprehensive measure of spouse responses to the behaviors of patients with chronic pain, the Spouse Response Inventory (SRI). Specifically, this questionnaire attempts to assess spouse behaviors likely to reflect positive and negative consequences for patient pain and well behaviors. By measuring a broader range of spouse responses to pain behaviors and well behaviors, the SRI provides a comprehensive description of chronic pain patient-
ORIGINAL REPORT/Schwartz et al Table 1.
245
Reliability Coefficients and Means for the Final SRI Subscales
SUBSCALE Spouse Responses to Pain Behavior Scales Solicitous (combined) Negative Spouse Responses to Well Behavior scales Facilitative Negative
NO. OF ITEMS
INTERNAL CONSISTENCY*
STABILITY COEFFICIENT
MEAN RATING (SD)
19 7
0.93 0.88
0.84 0.83
2.36 (0.85) 0.93 (0.80)
7 6
0.88 0.81
0.73 0.80
2.62 (0.86) 1.57 (0.91)
*Cronbach ␣.
spouse interactions and might, in future studies, allow more comprehensive tests of operant hypotheses. On the basis of operant theory and prior research, we hypothesized that spouse solicitous responses to patient pain behaviors as measured by the SRI (if it is a valid measure of this construct) should be positively related to higher frequencies of patient pain behaviors and higher levels of patient physical disability. Conversely, we hypothesized that, if valid, an SRI measure of spouse encouragement of patient well behaviors should be negatively related to dysfunction. On the basis of previous research, as reviewed above, we did not anticipate that spouse negative responses to patient pain behaviors would necessarily be associated with lower rates of pain behaviors or disability. However, we did hypothesize that if the SRI scale that assesses negative spouse responses to pain behaviors were valid, it should be associated with greater patient depression.
Methods Subjects Subjects were 104 patients with chronic pain who were being evaluated for possible treatment in the University of Washington Multidisciplinary Pain Program. To be included in the study, subjects had to have experienced pain on a daily or almost daily basis for at least 6 months, be able to read and write English, and be currently married or cohabitating (because 90% of couples were married, all partners will be referred to as spouses). One hundred forty-seven consecutive patients met the inclusion criteria, and 104 (71%) of these agreed to participate in the study. The mean age of study participants was 42.76 years (SD, 11.30) for patients and 43.40 years (SD, 12.23) for spouses. The average length of marriage or cohabitation was 15.00 years (SD, 5.32; range, 1-58 years). Sixty-one percent of the patients were female. Twenty-nine percent of patients were employed part time or full time. Eighty-eight percent of the sample were white, 4% were black, 4% were Hispanic, 2% were Asian, and 2% were Native American. The average duration of patient pain was 6.47 years (SD, 2.54; range, 1-25 years). Thirty-nine percent reported primary complaints of low back pain, leg pain, or both; 26% had mid- or upper-back pain; and the remaining 40% had other musculoskeletal pain problems or headache.
Measures SRI The SRI was developed to assess spouse responses to patient pain and well behaviors. Patients are asked to indicate, on a 5-point Likert scale ranging from “never” to “always,” how often in the previous 2 weeks their spouses had responded in a particular way to the patient’s pain or well behaviors. For the initial version of the SRI, the questionnaire items were selected on the basis of a review of the literature in this area,15,21,29 concepts derived from operant theory, and the clinical judgment of the authors based on a combined 30 years of experience working with patients with chronic pain and their spouses. Part 1 of the SRI assesses spouse responses to patient pain behaviors. Part 1 was initially divided into 3 subscales: (1) Solicitous Responses to Pain Behavior I, which assessed spouse responses that involve providing something likely to be seen as positive or pleasurable in response to pain behaviors (ie, potentially positively reinforcing); (2) Solicitous Responses to Pain Behavior II, which assessed spouse responses that involve removing or decreasing something potentially aversive (ie, potentially negatively reinforcing); and (3) Negative Responses to Pain Behavior, which assessed spouse responses that are likely to be viewed as negative or aversive. Part 2 of the initial version of the SRI was divided into 3 subscales to assess spouse responses to patient well behaviors, such as engaging in a healthy activity (eg, exercise, doing chores, and working): (1) Facilitative Responses to Well Behavior I, which assessed spouse positive responses to the patient engaging in activity or well behaviors; (2) Facilitative Responses to Well Behavior II, which assessed the extent to which the spouse responds to patient well behaviors by removing or decreasing something potentially aversive for the patient; and (3) Negative Responses to Well Behavior, which assessed spouse responses that are likely to be aversive or not encouraging of well behavior. The initial SRI was modified on the basis of preliminary reliability, item, and interscale correlation analyses. The Facilitative Responses to Well Behavior II subscale was deleted because of poor internal consistency (␣ ⫽ .43) and test-retest reliability (stability coefficient, 0.55). The Solicitous Responses to Pain Behavior I and Solicitous Re-
246 Table 2.
Measuring Spouse Responses to Pain
SRI Items and Their Association With Their Parent Scale
ITEM Spouse Solicitous Responses to Pain Behavior scale 1. Got me my pain medication 3. Tried to comfort me by talking to me 4. Did some of my household chores 5. Let me know that he/she was sorry that I was in pain 7. Tried to reassure me 8. Was affectionate with me (kissed or hugged me) 9. Took over some of my job/household responsibilities 11. Tried to keep the tension, stress, or noise level down in the house 12. Paid attention to me by listening to me 13. Talked to someone for me (especially in difficult situation) 15. Gave me a massage 16. Got me something to drink 17. Cooked my favorite meal or took me out to dinner 19. Made sure that others did not bother me 20. Asked what he/she could do to help 21. Got me something to eat 23. Avoided asking me to do something for him or her (such as pick up the kids, put away dishes) 24. Got me a pillow, blanket, or heating pad 25. Called my doctor Spouse Negative Responses to Pain Behavior scale 2. Seemed to criticize me more 6. Did not pay any attention to me 10. Seemed to get irritated with me 14. Ignored me 18. Seemed to get frustrated with me 22. Did not talk to me (gave me the silent treatment) 26. Avoided being physically affectionate with me (did not hug me or kiss me) Spouse Facilitative Responses to Well Behavior scale 1. Seemed happy that I was more active 2. Hugged or kissed me 4. Told me that he/she loves me 6. Said nice things to me 8. Smiled at me 10. Told me that I was doing a good job 12. Told others that I was doing well Spouse Negative Responses to Well Behavior scale 3. Told me that if I didn’t stop now, I’d pay for it later 5. Questioned my ability to do the activity 7. Told me to “slow down” 9. Criticized me for being so active 11. Warned me that I might really hurt myself if I did too much 13. Seemed to worry about me because I was being so active
ASSOCATION (r) WITH PARENT SCALE
LOADING ON SCALE FACTOR
0.60 0.74 0.57 0.72 0.74 0.59 0.58 0.75 0.70 0.71 0.54 0.71 0.62 0.56 0.82 0.68 0.68
0.59 0.71 0.50 0.74 0.74 0.59 0.56 0.77 0.70 0.71 0.54 0.73 0.62 0.56 0.81 0.70 0.67
0.78 0.60
0.78 0.61
0.76 0.67 0.85 0.74 0.79 0.74 0.73
0.73 0.71 0.84 0.78 0.81 0.76 0.72
0.78 0.78 0.70 0.79 0.74 0.86 0.61
0.72 0.80 0.78 0.83 0.79 0.84 0.61
0.74 0.65 0.76 0.74 0.75 0.76
0.78 0.71 0.86 0.75 0.83 0.86
NOTE: The instructions for the SRI are as follows. Before Part 1, they read, “There are two parts to this questionnaire. Please read the directions for each part. Part 1. Instructions: We are interested in how your spouse responds to you when he or she thinks you are in pain. Using the scale below, circle a number for each of the questions to indicate how often your spouse responded to you in that particular way, during the past 2 weeks, when he or she thought that you were in pain.” Following this, “0 ⫽ Never,” “1 ⫽ Rarely,” “2 ⫽ Sometimes,” “3 ⫽ Often,” and “4 ⫽ Always” are listed. Above the specific items, the following sentence stem states, “During the past 2 weeks when my spouse thought that I was in pain, he or she:.” The 26 SRI items for Part 1 (responses to pain behavior items) are then listed, with the response options (0-4) next to each item. After the last pain behavior spouse response item in Part 1 is listed, the sentence “STOP—PLEASE READ THE DIRECTIONS ON THE NEXT PAGE” is printed in bold face at the bottom of the page. The top of the next page (Part 2) then presents the instructions, “We are interested in how your spouse responds to you when you are working or engaging in some other kind of activity, such as exercising or doing household chores. Using the scale below, circle a number for each of the questions to indicate how often your spouse responded to you in that particular way; during the past 2 weeks, when you were being active.” The same 0 to 4 scale used for Part 1 is listed, and the sentence stem listed before the Part 2 items reads, “During the past 2 weeks when my spouse thought that I was being active, he or she:.” The 13 SRI well behavior items are then listed, with the responses options (0-4) next to each item. The SRI scoring protocol is as follows for each subscale: Average Solicitous Responses to Pain Behavior subscale score ⫽ Items (1 ⫹ 3 ⫹ 4 ⫹ 5 ⫹ 7 ⫹ 8 ⫹ 9 ⫹ 11 ⫹ 12 ⫹ 13 ⫹ 15 ⫹ 16 ⫹ 17 ⫹ 19 ⫹ 20 ⫹ 21 ⫹ 23 ⫹ 24 ⫹ 25)/19; Average Negative Responses to Pain Behavior subscale score ⫽ Items (2 ⫹ 6 ⫹ 10 ⫹ 14 ⫹ 18 ⫹ 22 ⫹ 26)/7; Average Facilitative Responses to Well Behaviors subscale score ⫽ Items (1 ⫹ 2 ⫹ 4 ⫹ 6 ⫹ 8 ⫹ 10 ⫹ 12)/7; Average Negative Responses to Well Behaviors subscale score ⫽ Items (3 ⫹ 5 ⫹ 7 ⫹ 9 ⫹ 11 ⫹ 13)/6.
ORIGINAL REPORT/Schwartz et al sponses to Pain Behavior II were combined into a single scale because of their high Pearson correlation coefficient (r) of 0.82 (P ⬍ .001), which suggested a large degree of shared statistical variance. This combined scale was used in all further analyses. The scale items presented in Table 1 make up the final version of the SRI and are scored to create 4 subscales: (1) Solicitous Responses to Pain Behavior; (2) Negative Responses to Pain Behavior; (3) Facilitative Responses to Well Behavior; and (4) Negative Responses to Well Behavior.
Disability The 24-item Roland-Morris Disability Questionnaire (RMDQ)20 was given to patients to assess the effect of pain on physical functioning. The RMDQ was derived from the Sickness Impact Profile Physical Disability subscale.3 The RMDQ has been shown to be a reliable and valid measure of physical disability in patients with low back pain23 and in a heterogeneous sample of patients with chronic pain.12
Pain Intensity Patients were asked to rate their average pain intensity on a 0 to 10 numeric rating scale, with 0 defined as no pain and 10 defined as worst pain imaginable.10,17
Pain Behaviors Patients rated the extent to which they engaged in various pain behaviors during the previous week on the Pain Behavior Checklist (PBC).15 The PBC assesses common pain behaviors, including distorted ambulation, facial-audible expressions, affective distress, and seeking help. Patients rate the frequency of each behavior’s occurrence on a scale of 0 (never) to 6 (very often). The total score was used as a measure of degree of pain behavior displayed.
Depressive Symptoms Self-report of depressive symptoms was assessed with the Beck Depression Inventory (BDI).2 The BDI is a 21item, self-report measure of severity of depressive symptomatology that has been validated for use in patients with chronic pain.29
Procedures Patients completed the battery of questionnaires at the time of their initial evaluation at the University of Washington Medical Center Multidisciplinary Pain Clinic. All subjects were told that participation in this research was voluntary and that data would not be accessible to the clinical evaluation team. The research protocol was approved by the Human Subjects Review Committee of the University of Washington. After the recruitment of the first 30 subjects, the remainder of subjects in the study (74) were asked to complete the SRI (Table 2) for a second time, 2 weeks after the initial administration, to provide test-retest reliability data. These subjects were paid $5 for returning these data by mail, and 50 (67%) complied. There were no significant differences between
247 those subjects who completed the SRI a second time and those who did not complete this measure in terms of duration of pain, self-report of physical disability, or age.
Results Reliability and Internal Consistency of the SRI Scales Table 1 presents the means and SDs, internal consistency (Cronbach ␣), and test-retest stability coefficients for the final version of the SRI scales. Table 2 presents the final SRI items, along with their association (Pearson correlation coefficient) with their parent scale. The internal consistency coefficients are high (all above .80), suggesting that the subscales are measuring relatively unitary dimensions of spouse responding. The test-retest stability coefficients of the subscales ranged from 0.73 to 0.84, indicating acceptable to excellent stability of these scales over a 2-week time interval. The correlations between each item and its parent scale are all strong (r range, 0.56-0.86). Although there were not enough subjects to perform a reliable factor analysis of all of the SRI items, there were an adequate number of subjects to perform reliable factor analyses of the items on each individual scale. The scree test was used to determine the number of factors in each scale and showed that, in general, the SRI items from each scale load primarily onto one main factor, as evidenced by a very large difference between the first and second eigenvalues for all scales. The first 2 eigenvalues associated with the Spouse Solicitous Responses to Pain Behavior, Negative Responses to Pain Behavior, Facilitative Responses to Well Behavior, and Negative Responses to Well Behavior scales were as follows: 8.53 and 1.52; 4.11 and 0.84; 4.20 and 0.94; and 3.81 and 0.62. The loadings of each item with the factor score that represents each spouse response domain (Table 2) are very high and are consistent with the high correlations found between each item and its parent scale.
Scale Intercorrelations Correlational analyses were conducted to evaluate the relationships among the SRI subscales. As can be seen in Table 3. Interscale Correlations for the Final SRI Subscales
SUBSCALE Negative Responses to Pain Behavior Facilitative Responses to Well Behavior Negative Responses to Well Behavior *P ⬍ .05. †P ⬍ .01.
SOLICITUS RESPONSES TO PAIN BEHAVIOR
NEGATIVE RESPONSES TO PAIN BEHAVIOR
FACILITATIVE RESPONSES TO WELL BEHAVIOR
⫺0.48* 0.61†
⫺0.50†
0.38†
⫺0.03
0.18
248
Measuring Spouse Responses to Pain
Regression Analyses Predicting Patient Physical Disability (Roland-Morris Disability Questionnaire)
Table 4.
PREDICTOR AND STEP 1: Pain variables (block) Pain intensity Pain duration 2: Spouse response inventory (block) Solicitous Responses to Pain Behavior Negative Responses to Pain Behavior Facilitative Responses to Well Behavior Negative Responses to Well Behavior
TOTAL R2
R2 CHANGE
F CHANGE
0.05
0.05
2.70
 .19 ⫺.12
0.24
0.20
5.01† .38† .01 ⫺.34† .23*
*P ⬍ .05. †P ⬍ .01.
Table 3, solicitous responses to patient pain behaviors were significantly positively associated with both facilitative and negative responses to well behaviors and inversely correlated with negative responses to patient’s displays of pain. A significant negative association was seen between the Facilitative Responses to Well Behavior and the Negative Responses to Well Behavior scales.
Validity of the SRI: Predictive Validity Separate regression analyses, using a hierarchical regression model, were conducted for the 4 adjustment measures to determine the relationship between the SRI and measures of patient adjustment. The 4 criteria variables in these analyses were as follows: (1) patient physical disability (RMDQ), (2) patient pain behaviors (PBCTotal), (3) patient depression (BDI), and (4) average pain intensity. For each equation, patient pain intensity and pain duration were entered (with the exception of the equation predicting pain intensity, for which only pain duration was entered) in the first step to control for their potentially confounding effects on the relationship between the primary predictor variables of interest (SRI subscales) and the patient functioning variables. In the second step the 4 SRI subscales were entered as a block (Tables 4-7). The regression analysis predicting patient disability scores revealed that the SRI subscales entered as a block made a significant contribution (R2 change ⫽ 0.20, P ⬍ .01), even when controlling for pain-related variables. The Solicitous Responses to Pain Behavior, Facilitative Responses to Well Behavior, and Negative Responses to Well Behavior scales made statistically significant independent contributions to the prediction of patient physical disability, with positive associations found between the Solicitous Responses to Pain Behavior and Negative Responses to Well Behavior scales and patient disability and negative associations between the Facilitative Responses to Well Behavior scale and patient disability. For the regression analysis predicting PBC scores, the SRI subscales entered as a block also made a significant contribution (R2 change ⫽ 0.27, P ⬍ .01). The  values for the Solicitous Responses to Pain Behavior, Negative Responses to Pain Behavior, and Negative Responses to Well Behavior scales were all positive and statistically
significant, indicating that these variables made independent contributions to the prediction of self-reported patient pain behaviors. Higher scores on each of these subscales were associated with more frequent reported pain behaviors. The  value associated with the Facilitative Responses to Well Behavior scale approached statistical significance and was negative ( ⫽ ⫺.22, P ⫽ .051), suggesting a trend toward an inverse relationship between patient perceptions of encouragement of well behavior and reported patient pain behaviors. The SRI subscales as a group accounted for 13% additional variance in depression scores (R2 change ⫽ 0.13, P ⬍ .05). Only the Negative Responses to Pain Behavior scale made a significant independent contribution, with higher levels of negative responding associated with greater depression. As a group, the SRI scales also made a significant contribution to the prediction of pain-rated pain intensity over and above the effects of pain duration. This effect was due primarily to a positive association between patient perceptions of spouse negative responses to patient pain behaviors and pain intensity.
Discussion The present study describes the development and preliminary psychometric evaluation of a self-report inventory designed to assess spouse responses to patient pain and well behaviors. In its final form, the SRI is a 39-item inventory divided into 2 parts, spouse responses to pain behaviors (Part 1) and spouse responses to well behaviors (Part 2), with 2 subscales in each part. The subscales reflect solicitous and negative responses to patient pain behaviors and facilitative and negative responses to patient well behaviors. The test-retest stability coefficients (over a 2-week period) were uniformly high, and factor analyses of the scale items (performed for each scale separately) provided evidence that the SRI items associated with each scale can be interpreted as assessing a single primary domain of patient-perceived spouse responding. Moreover, the item analyses (correlation of each item with its parent scale and the loading of each item on a factor score representing each spouse response domain) indicated strong associations between the individual items
ORIGINAL REPORT/Schwartz et al Table 5.
249
Regression Analyses Predicting Patient Pain Behaviors (Pain Behavior Checklist)
PREDICTOR AND STEP 1: Pain variables (block) Pain intensity Pain duration 2: Spouse response inventory (block) Solicitous Responses to Pain Behavior Negative Responses to Pain Behavior Facilitative Responses to Well Behavior Negative Responses to Well Behavior
TOTAL R2 0.09
R2 CHANGE 0.09
F CHANGE

†
4.85
.23* ⫺.10 0.38
0.27
9.39† .49† .30† ⫺.22 .22*
*P ⬍ .05. †P ⬍ .01.
and the parent scale–factor score for each SRI scale. Overall, these findings provide strong initial support for the reliability and internal consistency of the SRI scales. The associations between the SRI scales and patient reports of their functioning were in directions consistent with what operant theory and previous research (which used other measures of spouse responding) would predict. However, because the SRI scales include responses to well behaviors, some novel findings emerged. For example, the SRI scales that assess patient perceptions of spouse responses that are likely to be viewed as positive by most patients (solicitous responses to pain behaviors and facilitative responses to well behaviors) were positively associated with one another, yet they evidenced different patterns of associations with patient-reported disability. Solicitous responses to pain behaviors were associated with greater disability, whereas facilitative responses to well behaviors were associated with less patient disability. This pattern of associations provides further support for the discriminant validity of these 2 SRI scales; if both scales were tapping into the same domain (perhaps a general perception of patient-perceived positive responding by their spouses), they should have shown a similar association to patient functioning. A positive relationship was seen between discouraging well behavior (Negative Responses to Well Behavior scale scores) and solicitous responding to pain behavior, but a significant positive correlation was also seen between encouraging well behavior (Facilitative Responses to Well Behavior scale) and solicitous responding to pain behavior. Although this might seem contradictory, these findings suggest that patterns of patient-spouse responding might be highly complex, relationship dependent, and context specific. Defining the interrelationships among types of perceived spouse responses, the factors that influence them, and the circumstances under which they are seen requires further investigation. It would have been interesting to determine the extent to which the 2 types of positive responding assessed by the SRI (solicitous responses to pain behaviors and facilitative responses to well behaviors) are associated with marital satisfaction or if marital satisfaction mediates the relationships between perceived spouse responses and patient outcomes. Unfortunately, we did not include a measure of marital satisfaction in this study, and there-
fore the potential contributions of patient perceptions of both positive and negative spouse responses to marital satisfaction must await further research. The results of the analyses using the SRI subscales as predictors of patient functioning were generally consistent with previous research that has examined the association between measures of spouse responses to patient pain behaviors and patient functioning domains and support the construct validity of the SRI subscales. Patient’s perceptions of spouse solicitous responses to pain behaviors, as measured with the SRI, were associated with greater patient-reported physical disability and a higher frequency of pain behaviors. Similarly, SRIassessed encouragement of well behaviors by spouses made an independent contribution to the prediction of lower patient disability, beyond the contribution of the SRI Solicitous scale. In addition, patient perceptions of spouse negative responses to well behaviors were associated with higher patient disability and pain behaviors. Associations between perceived spouse responses to well behaviors and measures of patient functioning have not been demonstrated in the empirical literature to our knowledge and represent understudied, yet theoretically and clinically important, relationships. These preliminary findings support the predictive validity of the SRI scales measuring responses to patient well behaviors and also support the study of the potential effect of spouse responses to patient well behaviors on patient functioning. If further research confirms these findings and the role that spouse responses to well behaviors might play in promoting improved patient functioning, this would support the development of empirically based couples interventions incorporating a focus on spouse responses to both pain and well behaviors to enhance functioning in patients with chronic pain. Patient perceptions of spouse negative responses to patient pain behaviors were predictive of patient reports of more frequent pain behaviors and were not associated with patient physical disability (RMDQ scores). Given the inconsistent findings of prior studies regarding spouse negative responses to pain behaviors, we did not hypothesize that these responses would be associated with pain behaviors or disability. Although such responses are classed as “negative” in terms of their content, they might or might not function to reduce the
250
Measuring Spouse Responses to Pain
Regression Analyses Predicting Patient Depressive Symptoms (Beck Depression Inventory)
Table 6.
PREDICTOR AND STEP 1: Pain variables (block) Pain intensity Pain duration 2: Spouse response inventory (block) Solicitous Responses to Pain Behavior Negative Responses to Pain Behavior Facilitative Responses to Well Behavior Negative Responses to Well Behavior †
TOTAL R2
R2 CHANGE
F CHANGE
0.01
0.01
0.61
 .10 .00
0.14
0.13
2.54* .06 .26* ⫺.14 .16
P ⬍ .01.
*P ⬍ .05.
frequency of the behaviors they follow (ie, act as punishers in operant behavioral terms). It is possible that such responses might in fact serve to increase or maintain pain behavior if attention of any type (either solicitous or negative) is reinforcing in certain situations or couples. This possibility was suggested by Papas et al,20 who found that a subgroup of patients who were classified as having spouses who were “negatively attentive” on the basis of their responses to the Multidimensional Pain Inventory also reported the highest scores on measures of pain, interference, and depression and the lowest scores on activity level (except social activity) and life control. The authors offered as one explanation of their findings the possibility that both positive and negative responses might have become salient attentional reinforcers of pain behaviors, particularly in the absence of other sources of attention, such as occupational and social contacts. Further research is needed using validated measures of spouse responding to test this hypothesis in a larger sample of couples and to determine whether marital satisfaction might play a role in determining the valence that such responses have for couples. Patient perceptions of spouse negative responses to patient pain behaviors were associated with greater depression in the patient. This finding is consistent with previous research on individuals with chronic pain using other measures of spouse negative responding14,20 and supports the validity of the Spouse Negative Responses to Patient Pain Behavior scale. The association between patient-perceived negative spouse responses and patient depression in this study is also consistent with the general marital and depression literature, which finds a strong association between negative behaviors in one partner and poorer functioning in the other partner.1,30 Initially, we were interested in evaluating whether there are differences in patient functioning associated with spouse solicitous responses to pain behaviors categorized as potentially positively reinforcing versus negatively reinforcing. However, there was such a strong degree of statistical association between the subscales measuring these 2 response categories that it seemed most appropriate to consolidate these scales. Moreover, the subscale designed to assess negative reinforcement
of well behaviors had poor psychometric qualities and consequently was eliminated from the inventory. Given the failure to create single scales that reflect patient perceptions of spouse responses that are potentially negatively reinforcing (to patient pain or well behaviors), the current findings suggest the possibility that although operant theory supports the concept of negative reinforcement as being distinct from positive reinforcement, it might be difficult for patients to distinguish between these 2 types of spouse responses. Alternatively, it is possible that patients are able to distinguish between these responses and accurately rate them but that spouses who provide frequent positive reinforcement might also be those who provide frequent negative reinforcement in response to patient pain behaviors. In this case an empiric distinction between these 2 types of spouse responses might be very difficult to achieve. Unfortunately, then, despite the potential for positive and negative reinforcement to have different effects on patient functioning, at this time, the SRI cannot be used to test for these potential differential effects. The current findings are limited by the study’s correlational design and reliance on patient self-report data. Self-reports might be influenced by a number of factors, such as social desirability and patient affect. Thus the results should be considered preliminary and warrant replications with multiple modes of measurement. A second limitation of the study is that the SRI items were rationally rather than empirically derived and thus form a restricted sample of potential responses to pain and well behaviors. It is quite possible that other types of spouse responses to both pain and well behaviors might occur in samples of couples in which one person has chronic pain. In addition, as indicated above, we did not assess marital satisfaction in this study. Because significant associations between measures of marital satisfaction and patient perceptions of spouse responses to pain and well behaviors likely exist and because previous research has shown that the association between spouse responses to patient behaviors and patient functioning might vary as a function of marital satisfaction,14 an examination of these associations would have strengthened the study.
ORIGINAL REPORT/Schwartz et al Table 7.
251
Regression Analyses Predicting Average Patient Pain Intensity (0-10 Numeric Rating
Scale) PREDICTOR AND STEP 2: Pain duration 3: Spouse response inventory (block) Solicitous Responses to Pain Behavior Negative Responses to Pain Behavior Facilitative Responses to Well Behavior Negative Responses to Well Behavior
TOTAL R2 0.07 0.13
R2 CHANGE 0.07 0.06
F CHANGE †
7.59 2.80*
 ⫺.27† .08 .24* .22 ⫺.01
*P ⬍ .05. †P ⬍ .01.
A potential limitation of the current version of the SRI is its limited focus on patient activity level as representing the domain of well behavior. Verbal behaviors that communicate hope for the future, perceived ability to cope with pain, or general positive self-statements might also be considered well behaviors. A number of theoretic models, particularly cognitive-behavioral theory,29 as well as a model of motivation for pain self-management,11 hypothesize a causal link between patient verbal responses (both what they say to others and what they tell themselves) and patient functioning. A potentially useful avenue for future research would be to examine the association between spouse responses to these verbal behaviors and measures of current and future patient functioning. Another limitation is that this study relied solely on self-report for both the SRI and the validity criterion measures. The reliance on self-report for all measures can work to artificially increase the association between variables because of shared method variance. In light of its limitations, the current study should be viewed as representing an initial examination of a new measure of spouse responses to patient behaviors; more work is needed to examine the reliability and validity of the measure in larger samples of diverse patients with chronic pain and their spouses-partners. Specifically, further research is needed to evaluate the relationship between the SRI and observed partner responses to patient pain and well behaviors, as well as other questionnaire measures of spouse responses to patient pain and well behaviors (ie, the WHYMPI) and measures of patient functioning. Such re-
search would be strengthened if criterion measures (of patient functioning) were assessed through direct observation rather than self-report. The sensitivity of the SRI to changes in treatment is also unknown. Further research is also needed to cross-validate the results on other samples of individuals with chronic pain. In summary, the findings from this study suggest that the SRI provides a relatively simple and reliable assessment of perceived responses of significant others to pain and well behaviors. The results also provide preliminary support for the construct validity of the SRI scales in that they demonstrate associations that are consistent with those predicted by theory and found using measures of similar constructs. The SRI might be useful in future research evaluating relationships between spouse responses and patient functioning. The assessment of spouse responses to well behaviors might also have important implications for the treatment of couples with chronic pain, given the possibility that encouragement of well behaviors might be associated with lower levels of disability and might contribute to maintaining treatment gains. Moreover, discouraging spouse solicitous responses to pain behaviors and spouse negative responses to both pain and well behaviors might also be important for adaptive patient functioning. Examining family responses that encourage or promote patient functioning is an important area for continued research, particularly in the current health care environment emphasizing decreasing health care use and improving health care outcomes.
References
4. Block AR, Kremer EF, Gaylor M: Behavioral treatment of chronic pain: the spouse as a discriminant cue for pain behavior. Pain 9:243-252, 1980
1. Beach SRH, Sandler EE, O’Leary KD: Depression in Marriage: A Model for Etiology and Treatment. New York, NY, Guilford, 1990 2. Beck AT, Ward CH, Mendelson M, Mock J, Erlbaugh J: An inventory for measuring depression. Arch Gen Psychiatry 4:461-571, 1961 3. Bergner M, Bobbit RA, Carter WB, Gilson BS: The sickness impact profile: development and final revision of a health status measure. Med Care 19:787-805, 1981
5. Cinciripini PM, Floreen A: An outcome evaluation of a behavioral program for chronic pain. J Behav Med 5:375389, 1982 6. Flor H, Kerns RD, Turk DC: The role of spouse reinforcement, perceived pain, and activity levels of chronic pain patients. J Psychosom Res 31:251-259, 1987. 7. Flor H, Knost B, Birbaumer N: The role of operant condi-
252
Measuring Spouse Responses to Pain
tion in chronic pain: an experimental investigation. Pain 95:111-118, 2002
ship between spouse solicitousness and pain behavior: searching for more evidence. Pain 51:75-79, 1992
8. Fordyce WE: Behavioral Methods for Chronic Pain and Illness. St Louis, Mo, C.V. Mosby, 1976
20. Papas RK, Robinson ME, Riley JL 3rd: Perceived spouse responsiveness to chronic pain: three empirical subgroups. J Pain 2:262-269, 2001.
9. Fordyce WE, Fowler R, Lehmann J, DeLateur B: Some implications of learning in problems of chronic pain. J Chronic Dis 21:179-190, 1968 10. Jensen MP, Karoly P, Braver S: The measurement of clinical pain intensity: a comparison of six methods. Pain 27:117126, 1986 11. Jensen MP, Nielson WR, Kerns RD: Toward the development of a motivational model of pain self-management. J Pain 4:477-492, 2003 12. Jensen MP, Strom S, Turner JA, Romano JM: Validity of the Sickness Impact Profile Roland Scale as a measure of dysfunction in chronic pain patients. Pain 50:157-162, 1992 13. Jolliffe CD, Nicholas MK: Verbally reinforcing pain reports: an experimental test of the operant model of chronic pain. Pain 107:167-175, 2004 14. Kerns RD, Haythornthwaite J, Southwick S, Giller EL: The role of marital interaction in chronic pain and depressive symptom severity. J Psychosom Res 34:401-408, 1990. 15. Kerns RD, Haythornthwaite J, Rosenberg R, Southwick S, Giller EL, Jacob MC. The Pain Behavior Checklist (PBCL): factor structure and psychometric properties. J Behav Med 14: 155-167, 1991 16. Kerns RD, Turk DC, Rudy TE: The West Haven-Yale Multidisciplinary Pain Inventory (WHYMPI). Pain 23:356-435, 1985 17. Kremer E, Atkinson JH, Ignelzi RJ: Measurement of pain: patient preference does not confound measurement. Pain 10:241-248, 1981 18. Loeser JD, Egan KJ: Managing the Chronic Pain Patient: Theory and Practice at the University of Washington Multidisciplinary Pain Center. New York, NY, Lippincott-Raven Publishers, 1989 19. Lousberg R, Schmidt AJM, Groenman NH: The relation-
21. Patrick L, D’Eon J: Social support and functional status in chronic pain patients. Can J Rehab 9:195-201, 1996 22. Revenson TA, Schiaffino KM, Majerovitz SD, Gibofsky A: Social support as a double-edged sword: the relation of positive and problematic support to depression among rheumatoid arthritis patients. Soc Sci Med 33:807-813, 1991 23. Roland M, Morris R: Development of a reliable and sensitive measure of disability in low back pain. Spine 8:141144, 1983 24. Romano JM, Turner JA, Friedman LS, Bulcroft RA, Jensen MP, Hops H: Observational assessment of chronic pain patientspouse behavioral interactions. Behav Ther 22:549-567, 1991 25. Romano JM, Turner JA, Jensen MP, Friedman LS, Bulcroft RA, Hops H, Wright SF: Chronic pain patient-spouse behavioral interaction predict disability. Pain 63:353-360, 1995 26. Roy R: Chronic pain and marital difficulties. Health Social Workers 10:199-207, 1995 27. Schwartz L, Slater MA, Birchler GR: The role of pain behaviors in the modulation of marital conflict in chronic pain couples. Pain 65:227-233, 1996 28. Turk DC, Kerns RD, Rosenberg R: Effects of marital interaction on chronic pain and disability: examining the downside of social support. Rehab Psychol 37:259-274, 1992 29. Turk DC, Meichenbaum D, Genest M: Pain and Behavioral Medicine: A Cognitive-behavioral Perspective. New York, NY, Guilford Press, 1983 30. Weissman MM: Advances in psychiatric epidemiology: rates and risks for major depression. Am J Public Health 77:445-451, 1987. 31. Williamson D, Robinson ME, Melamed B: Pain behavior, spouse responsiveness, and marital satisfaction in patients with rheumatoid arthritis. Behav Modif 21:97-118, 1997
A
n operant behavioral formulation of chronic pain behavior, first introduced over 3 decades ago,9 has had a significant influence on the understanding and treatment of chronic pain. In this theoretical perspective, pain behaviors (eg, grimacing, limping, or guarding) might come under the control of reinforcing consequences in the environment (eg, social responses of significant others and avoidance of stressful interpersonal or work situations) and be maintained, at least in part, by such consequences. Hence a potential explanation for continued pain behavior in some individuals with chronic pain is the ongoing positive or negative reinforcement of pain behaviors.8 Such operant behavioral principles have been incorporated into many multidisciplinary chronic pain treatment programs5,18 and serve as the theoretical basis for a number of interven-
Received June 23, 2004; Revised November 9, 2004; Accepted December 21, 2004. From the Departments of *Rehabilitation Medicine and †Psychiatry and Behavioral Sciences, University of Washington School of Medicine, Seattle, WA. Address reprint requests to Mark P. Jensen, Department of Rehabilitation Medicine, Box 356490, University of Washington School of Medicine, Seattle, WA 98195-6490. E-mail: [email protected] 1526-5900/$30.00 © 2005 by the American Pain Society doi:10.1016/j.jpain.2004.12.010
tions, such as the behavioral shaping of increased activity through quota-based reactivation. Correlational research supports a primary hypothesis of the operant model: the responses of the social environment can serve as reinforcers of pain behavior and thus play a role in maintaining chronic pain behaviors and dysfunction. For example, observed spouse solicitous responses to patient nonverbal pain behaviors are associated with higher rates of observed pain behaviors in patients reporting greater pain and with greater disability in more depressed patients.25 Spouse solicitousness as reported by patients4,31 and spouses19 has also been associated with higher self-reports of patient pain intensity4,19 and observed patient pain behaviors.31 Kerns and colleagues14 found a similar relationship between reported spouse solicitousness and pain severity for maritally satisfied pain patients and spouses. Flor et al6 found that greater spouse solicitousness was associated with lower activity levels in patients with chronic pain. Findings such as these are consistent with the assumption that solicitous attention to patient pain behaviors from spouses might be reinforcing for many patients. Recent experimental research provides even more support for the hypothesis that environmental responses might influence pain behavior. Flor et al7 assessed pain intensity reports and physiologic responding (electroen-
The Journal of Pain, Vol 6, No 4 (April), 2005: pp 243-252
243
244 cephalography, electrooculography, electromyography, and skin conductance level) in 30 patients with chronic pain and 30 healthy control subjects in response to 4 levels of electric shock (individual pain threshold and 25%, 50%, and 75% of the distance between individual pain threshold and pain tolerance, as assessed during pain threshold and tolerance tests). The study participants were then randomly assigned to be given either (1) positive feedback when subsequent ratings for shock were higher than the average rating previously given and negative feedback when these ratings were lower than the previous averages (up-training condition) or (2) negative feedback for higher ratings and positive feedback for lower ratings (down-training condition). Consistent with operant theory, subjects in the up-training condition provided increasing pain reports over time, and those in the down-training condition provided decreased pain reports over time. Moreover, patients with chronic pain showed a slower extinction rate (ie, maintained their relatively higher reports of pain for a longer period of time) than healthy control subjects when reinforcement was discontinued. The authors concluded that patients with chronic pain appear to be more susceptible to the operant conditioning of increasing pain responses than healthy control subjects. In a similar study, Jolliffe and Nicholas13 randomly assigned half of 46 undergraduates to receive positive verbal reinforcement from an experimenter after each trial of a cold pressor task if their report of pain intensity exceeded that of a previous trial. The other half did not receive this reinforcement. Again, and consistent with an operant model, the mean pain intensity ratings of the subjects in the reinforcement condition were greater than those of the nonreinforced subjects. Operant theory also suggests that pain behaviors might be maintained by a process of negative reinforcement, which is defined as the reduction or removal of an aversive stimulus (not to be confused with punishment, which is the application of an aversive stimulus). For example, it has been hypothesized that pain behaviors can lead to decreased demands on the patient to participate in potentially aversive activities, such as work, chores, exercise, or conflictive interpersonal interactions.8,26,27 However, the relationship between these potential negative reinforcers and patient functioning has not yet been systematically examined in previous research. This is most likely because of the lack of a reliable and valid measure of such possible negative reinforcers. Spouse negative responses (eg, irritation and criticism) to patient pain behaviors have also been examined to determine whether such responses might act as punishers and thus reduce the frequency of pain behaviors. The literature, however, has not yielded consistent findings to support this hypothesis. Some studies have found that negative responses by significant others are associated with a higher level of activity.6 Other studies have found that negative responses were associated with poorer patient physical and psychosocial functioning,27 whereas yet other studies have found no significant relationship between partner negative responses and patient disabil-
Measuring Spouse Responses to Pain 28
14,28
ity, pain intensity, or pain behavior.28 The most consistent finding has been that spouse negative responses are associated with greater patient psychosocial distress and depression,14,22,25,27,28 although this finding has been shown to vary with levels of marital satisfaction.14,28 Overall, these findings suggest that spouse negative responses to pain behaviors might not be associated consistently with a reduced frequency of pain behaviors but do appear to be associated more often with greater patient emotional distress. An important implication of the operant model is that spouses can play a role in encouraging and reinforcing better physical functioning and well behaviors (behaviors incompatible with the sick role, such as maintaining an appropriate activity level) in the person with pain.8 For example, spouse support, defined as emotional support and encouragement during a stationary bicycle exercise task, was shown to be associated with greater exercise persistence in a sample of 50 patients.21 Despite the theoretical potential that spouse responses to patient well behaviors could increase or maintain positive functioning in persons with chronic pain, no assessment instrument has yet been developed to systematically assess spouse responses to patient well behaviors. The most frequently used measures of spouse responses to patient pain behaviors are the Spouse Response subscales from the West Haven–Yale Multidimensional Pain Inventory (WHYMPI).16 The WHYMPI has many strengths. It is theoretically based, is well validated, and measures 3 types of spouse responses to patient pain behavior (solicitous, negative or “aversive,” and distracting responses). Moreover, the solicitous subscale of the WHYMPI is associated positively with observed spouse solicitous behaviors.24 However, the WHYMPI does not, in its current form, differentiate between spouse responses that are potentially positively reinforcing from those that are potentially negatively reinforcing. There might be important differences in how the provision of solicitous responses (potential positive reinforcers) and the removal of aversive responsibilities or conflicts (potential negative reinforcers) might affect patient function. In addition, the characteristics of couples in which these different response patterns might occur could differ in ways that might influence patient and spouse adjustment. Finally, the WHYMPI does not assess spouse responses (either positive or negative) to patient well behaviors. Thus there is a need for a more comprehensive measure to assess these dimensions of spouse responses to patient behavior. The primary goal of the present study was to develop and perform a preliminary evaluation of a comprehensive measure of spouse responses to the behaviors of patients with chronic pain, the Spouse Response Inventory (SRI). Specifically, this questionnaire attempts to assess spouse behaviors likely to reflect positive and negative consequences for patient pain and well behaviors. By measuring a broader range of spouse responses to pain behaviors and well behaviors, the SRI provides a comprehensive description of chronic pain patient-
ORIGINAL REPORT/Schwartz et al Table 1.
245
Reliability Coefficients and Means for the Final SRI Subscales
SUBSCALE Spouse Responses to Pain Behavior Scales Solicitous (combined) Negative Spouse Responses to Well Behavior scales Facilitative Negative
NO. OF ITEMS
INTERNAL CONSISTENCY*
STABILITY COEFFICIENT
MEAN RATING (SD)
19 7
0.93 0.88
0.84 0.83
2.36 (0.85) 0.93 (0.80)
7 6
0.88 0.81
0.73 0.80
2.62 (0.86) 1.57 (0.91)
*Cronbach ␣.
spouse interactions and might, in future studies, allow more comprehensive tests of operant hypotheses. On the basis of operant theory and prior research, we hypothesized that spouse solicitous responses to patient pain behaviors as measured by the SRI (if it is a valid measure of this construct) should be positively related to higher frequencies of patient pain behaviors and higher levels of patient physical disability. Conversely, we hypothesized that, if valid, an SRI measure of spouse encouragement of patient well behaviors should be negatively related to dysfunction. On the basis of previous research, as reviewed above, we did not anticipate that spouse negative responses to patient pain behaviors would necessarily be associated with lower rates of pain behaviors or disability. However, we did hypothesize that if the SRI scale that assesses negative spouse responses to pain behaviors were valid, it should be associated with greater patient depression.
Methods Subjects Subjects were 104 patients with chronic pain who were being evaluated for possible treatment in the University of Washington Multidisciplinary Pain Program. To be included in the study, subjects had to have experienced pain on a daily or almost daily basis for at least 6 months, be able to read and write English, and be currently married or cohabitating (because 90% of couples were married, all partners will be referred to as spouses). One hundred forty-seven consecutive patients met the inclusion criteria, and 104 (71%) of these agreed to participate in the study. The mean age of study participants was 42.76 years (SD, 11.30) for patients and 43.40 years (SD, 12.23) for spouses. The average length of marriage or cohabitation was 15.00 years (SD, 5.32; range, 1-58 years). Sixty-one percent of the patients were female. Twenty-nine percent of patients were employed part time or full time. Eighty-eight percent of the sample were white, 4% were black, 4% were Hispanic, 2% were Asian, and 2% were Native American. The average duration of patient pain was 6.47 years (SD, 2.54; range, 1-25 years). Thirty-nine percent reported primary complaints of low back pain, leg pain, or both; 26% had mid- or upper-back pain; and the remaining 40% had other musculoskeletal pain problems or headache.
Measures SRI The SRI was developed to assess spouse responses to patient pain and well behaviors. Patients are asked to indicate, on a 5-point Likert scale ranging from “never” to “always,” how often in the previous 2 weeks their spouses had responded in a particular way to the patient’s pain or well behaviors. For the initial version of the SRI, the questionnaire items were selected on the basis of a review of the literature in this area,15,21,29 concepts derived from operant theory, and the clinical judgment of the authors based on a combined 30 years of experience working with patients with chronic pain and their spouses. Part 1 of the SRI assesses spouse responses to patient pain behaviors. Part 1 was initially divided into 3 subscales: (1) Solicitous Responses to Pain Behavior I, which assessed spouse responses that involve providing something likely to be seen as positive or pleasurable in response to pain behaviors (ie, potentially positively reinforcing); (2) Solicitous Responses to Pain Behavior II, which assessed spouse responses that involve removing or decreasing something potentially aversive (ie, potentially negatively reinforcing); and (3) Negative Responses to Pain Behavior, which assessed spouse responses that are likely to be viewed as negative or aversive. Part 2 of the initial version of the SRI was divided into 3 subscales to assess spouse responses to patient well behaviors, such as engaging in a healthy activity (eg, exercise, doing chores, and working): (1) Facilitative Responses to Well Behavior I, which assessed spouse positive responses to the patient engaging in activity or well behaviors; (2) Facilitative Responses to Well Behavior II, which assessed the extent to which the spouse responds to patient well behaviors by removing or decreasing something potentially aversive for the patient; and (3) Negative Responses to Well Behavior, which assessed spouse responses that are likely to be aversive or not encouraging of well behavior. The initial SRI was modified on the basis of preliminary reliability, item, and interscale correlation analyses. The Facilitative Responses to Well Behavior II subscale was deleted because of poor internal consistency (␣ ⫽ .43) and test-retest reliability (stability coefficient, 0.55). The Solicitous Responses to Pain Behavior I and Solicitous Re-
246 Table 2.
Measuring Spouse Responses to Pain
SRI Items and Their Association With Their Parent Scale
ITEM Spouse Solicitous Responses to Pain Behavior scale 1. Got me my pain medication 3. Tried to comfort me by talking to me 4. Did some of my household chores 5. Let me know that he/she was sorry that I was in pain 7. Tried to reassure me 8. Was affectionate with me (kissed or hugged me) 9. Took over some of my job/household responsibilities 11. Tried to keep the tension, stress, or noise level down in the house 12. Paid attention to me by listening to me 13. Talked to someone for me (especially in difficult situation) 15. Gave me a massage 16. Got me something to drink 17. Cooked my favorite meal or took me out to dinner 19. Made sure that others did not bother me 20. Asked what he/she could do to help 21. Got me something to eat 23. Avoided asking me to do something for him or her (such as pick up the kids, put away dishes) 24. Got me a pillow, blanket, or heating pad 25. Called my doctor Spouse Negative Responses to Pain Behavior scale 2. Seemed to criticize me more 6. Did not pay any attention to me 10. Seemed to get irritated with me 14. Ignored me 18. Seemed to get frustrated with me 22. Did not talk to me (gave me the silent treatment) 26. Avoided being physically affectionate with me (did not hug me or kiss me) Spouse Facilitative Responses to Well Behavior scale 1. Seemed happy that I was more active 2. Hugged or kissed me 4. Told me that he/she loves me 6. Said nice things to me 8. Smiled at me 10. Told me that I was doing a good job 12. Told others that I was doing well Spouse Negative Responses to Well Behavior scale 3. Told me that if I didn’t stop now, I’d pay for it later 5. Questioned my ability to do the activity 7. Told me to “slow down” 9. Criticized me for being so active 11. Warned me that I might really hurt myself if I did too much 13. Seemed to worry about me because I was being so active
ASSOCATION (r) WITH PARENT SCALE
LOADING ON SCALE FACTOR
0.60 0.74 0.57 0.72 0.74 0.59 0.58 0.75 0.70 0.71 0.54 0.71 0.62 0.56 0.82 0.68 0.68
0.59 0.71 0.50 0.74 0.74 0.59 0.56 0.77 0.70 0.71 0.54 0.73 0.62 0.56 0.81 0.70 0.67
0.78 0.60
0.78 0.61
0.76 0.67 0.85 0.74 0.79 0.74 0.73
0.73 0.71 0.84 0.78 0.81 0.76 0.72
0.78 0.78 0.70 0.79 0.74 0.86 0.61
0.72 0.80 0.78 0.83 0.79 0.84 0.61
0.74 0.65 0.76 0.74 0.75 0.76
0.78 0.71 0.86 0.75 0.83 0.86
NOTE: The instructions for the SRI are as follows. Before Part 1, they read, “There are two parts to this questionnaire. Please read the directions for each part. Part 1. Instructions: We are interested in how your spouse responds to you when he or she thinks you are in pain. Using the scale below, circle a number for each of the questions to indicate how often your spouse responded to you in that particular way, during the past 2 weeks, when he or she thought that you were in pain.” Following this, “0 ⫽ Never,” “1 ⫽ Rarely,” “2 ⫽ Sometimes,” “3 ⫽ Often,” and “4 ⫽ Always” are listed. Above the specific items, the following sentence stem states, “During the past 2 weeks when my spouse thought that I was in pain, he or she:.” The 26 SRI items for Part 1 (responses to pain behavior items) are then listed, with the response options (0-4) next to each item. After the last pain behavior spouse response item in Part 1 is listed, the sentence “STOP—PLEASE READ THE DIRECTIONS ON THE NEXT PAGE” is printed in bold face at the bottom of the page. The top of the next page (Part 2) then presents the instructions, “We are interested in how your spouse responds to you when you are working or engaging in some other kind of activity, such as exercising or doing household chores. Using the scale below, circle a number for each of the questions to indicate how often your spouse responded to you in that particular way; during the past 2 weeks, when you were being active.” The same 0 to 4 scale used for Part 1 is listed, and the sentence stem listed before the Part 2 items reads, “During the past 2 weeks when my spouse thought that I was being active, he or she:.” The 13 SRI well behavior items are then listed, with the responses options (0-4) next to each item. The SRI scoring protocol is as follows for each subscale: Average Solicitous Responses to Pain Behavior subscale score ⫽ Items (1 ⫹ 3 ⫹ 4 ⫹ 5 ⫹ 7 ⫹ 8 ⫹ 9 ⫹ 11 ⫹ 12 ⫹ 13 ⫹ 15 ⫹ 16 ⫹ 17 ⫹ 19 ⫹ 20 ⫹ 21 ⫹ 23 ⫹ 24 ⫹ 25)/19; Average Negative Responses to Pain Behavior subscale score ⫽ Items (2 ⫹ 6 ⫹ 10 ⫹ 14 ⫹ 18 ⫹ 22 ⫹ 26)/7; Average Facilitative Responses to Well Behaviors subscale score ⫽ Items (1 ⫹ 2 ⫹ 4 ⫹ 6 ⫹ 8 ⫹ 10 ⫹ 12)/7; Average Negative Responses to Well Behaviors subscale score ⫽ Items (3 ⫹ 5 ⫹ 7 ⫹ 9 ⫹ 11 ⫹ 13)/6.
ORIGINAL REPORT/Schwartz et al sponses to Pain Behavior II were combined into a single scale because of their high Pearson correlation coefficient (r) of 0.82 (P ⬍ .001), which suggested a large degree of shared statistical variance. This combined scale was used in all further analyses. The scale items presented in Table 1 make up the final version of the SRI and are scored to create 4 subscales: (1) Solicitous Responses to Pain Behavior; (2) Negative Responses to Pain Behavior; (3) Facilitative Responses to Well Behavior; and (4) Negative Responses to Well Behavior.
Disability The 24-item Roland-Morris Disability Questionnaire (RMDQ)20 was given to patients to assess the effect of pain on physical functioning. The RMDQ was derived from the Sickness Impact Profile Physical Disability subscale.3 The RMDQ has been shown to be a reliable and valid measure of physical disability in patients with low back pain23 and in a heterogeneous sample of patients with chronic pain.12
Pain Intensity Patients were asked to rate their average pain intensity on a 0 to 10 numeric rating scale, with 0 defined as no pain and 10 defined as worst pain imaginable.10,17
Pain Behaviors Patients rated the extent to which they engaged in various pain behaviors during the previous week on the Pain Behavior Checklist (PBC).15 The PBC assesses common pain behaviors, including distorted ambulation, facial-audible expressions, affective distress, and seeking help. Patients rate the frequency of each behavior’s occurrence on a scale of 0 (never) to 6 (very often). The total score was used as a measure of degree of pain behavior displayed.
Depressive Symptoms Self-report of depressive symptoms was assessed with the Beck Depression Inventory (BDI).2 The BDI is a 21item, self-report measure of severity of depressive symptomatology that has been validated for use in patients with chronic pain.29
Procedures Patients completed the battery of questionnaires at the time of their initial evaluation at the University of Washington Medical Center Multidisciplinary Pain Clinic. All subjects were told that participation in this research was voluntary and that data would not be accessible to the clinical evaluation team. The research protocol was approved by the Human Subjects Review Committee of the University of Washington. After the recruitment of the first 30 subjects, the remainder of subjects in the study (74) were asked to complete the SRI (Table 2) for a second time, 2 weeks after the initial administration, to provide test-retest reliability data. These subjects were paid $5 for returning these data by mail, and 50 (67%) complied. There were no significant differences between
247 those subjects who completed the SRI a second time and those who did not complete this measure in terms of duration of pain, self-report of physical disability, or age.
Results Reliability and Internal Consistency of the SRI Scales Table 1 presents the means and SDs, internal consistency (Cronbach ␣), and test-retest stability coefficients for the final version of the SRI scales. Table 2 presents the final SRI items, along with their association (Pearson correlation coefficient) with their parent scale. The internal consistency coefficients are high (all above .80), suggesting that the subscales are measuring relatively unitary dimensions of spouse responding. The test-retest stability coefficients of the subscales ranged from 0.73 to 0.84, indicating acceptable to excellent stability of these scales over a 2-week time interval. The correlations between each item and its parent scale are all strong (r range, 0.56-0.86). Although there were not enough subjects to perform a reliable factor analysis of all of the SRI items, there were an adequate number of subjects to perform reliable factor analyses of the items on each individual scale. The scree test was used to determine the number of factors in each scale and showed that, in general, the SRI items from each scale load primarily onto one main factor, as evidenced by a very large difference between the first and second eigenvalues for all scales. The first 2 eigenvalues associated with the Spouse Solicitous Responses to Pain Behavior, Negative Responses to Pain Behavior, Facilitative Responses to Well Behavior, and Negative Responses to Well Behavior scales were as follows: 8.53 and 1.52; 4.11 and 0.84; 4.20 and 0.94; and 3.81 and 0.62. The loadings of each item with the factor score that represents each spouse response domain (Table 2) are very high and are consistent with the high correlations found between each item and its parent scale.
Scale Intercorrelations Correlational analyses were conducted to evaluate the relationships among the SRI subscales. As can be seen in Table 3. Interscale Correlations for the Final SRI Subscales
SUBSCALE Negative Responses to Pain Behavior Facilitative Responses to Well Behavior Negative Responses to Well Behavior *P ⬍ .05. †P ⬍ .01.
SOLICITUS RESPONSES TO PAIN BEHAVIOR
NEGATIVE RESPONSES TO PAIN BEHAVIOR
FACILITATIVE RESPONSES TO WELL BEHAVIOR
⫺0.48* 0.61†
⫺0.50†
0.38†
⫺0.03
0.18
248
Measuring Spouse Responses to Pain
Regression Analyses Predicting Patient Physical Disability (Roland-Morris Disability Questionnaire)
Table 4.
PREDICTOR AND STEP 1: Pain variables (block) Pain intensity Pain duration 2: Spouse response inventory (block) Solicitous Responses to Pain Behavior Negative Responses to Pain Behavior Facilitative Responses to Well Behavior Negative Responses to Well Behavior
TOTAL R2
R2 CHANGE
F CHANGE
0.05
0.05
2.70
 .19 ⫺.12
0.24
0.20
5.01† .38† .01 ⫺.34† .23*
*P ⬍ .05. †P ⬍ .01.
Table 3, solicitous responses to patient pain behaviors were significantly positively associated with both facilitative and negative responses to well behaviors and inversely correlated with negative responses to patient’s displays of pain. A significant negative association was seen between the Facilitative Responses to Well Behavior and the Negative Responses to Well Behavior scales.
Validity of the SRI: Predictive Validity Separate regression analyses, using a hierarchical regression model, were conducted for the 4 adjustment measures to determine the relationship between the SRI and measures of patient adjustment. The 4 criteria variables in these analyses were as follows: (1) patient physical disability (RMDQ), (2) patient pain behaviors (PBCTotal), (3) patient depression (BDI), and (4) average pain intensity. For each equation, patient pain intensity and pain duration were entered (with the exception of the equation predicting pain intensity, for which only pain duration was entered) in the first step to control for their potentially confounding effects on the relationship between the primary predictor variables of interest (SRI subscales) and the patient functioning variables. In the second step the 4 SRI subscales were entered as a block (Tables 4-7). The regression analysis predicting patient disability scores revealed that the SRI subscales entered as a block made a significant contribution (R2 change ⫽ 0.20, P ⬍ .01), even when controlling for pain-related variables. The Solicitous Responses to Pain Behavior, Facilitative Responses to Well Behavior, and Negative Responses to Well Behavior scales made statistically significant independent contributions to the prediction of patient physical disability, with positive associations found between the Solicitous Responses to Pain Behavior and Negative Responses to Well Behavior scales and patient disability and negative associations between the Facilitative Responses to Well Behavior scale and patient disability. For the regression analysis predicting PBC scores, the SRI subscales entered as a block also made a significant contribution (R2 change ⫽ 0.27, P ⬍ .01). The  values for the Solicitous Responses to Pain Behavior, Negative Responses to Pain Behavior, and Negative Responses to Well Behavior scales were all positive and statistically
significant, indicating that these variables made independent contributions to the prediction of self-reported patient pain behaviors. Higher scores on each of these subscales were associated with more frequent reported pain behaviors. The  value associated with the Facilitative Responses to Well Behavior scale approached statistical significance and was negative ( ⫽ ⫺.22, P ⫽ .051), suggesting a trend toward an inverse relationship between patient perceptions of encouragement of well behavior and reported patient pain behaviors. The SRI subscales as a group accounted for 13% additional variance in depression scores (R2 change ⫽ 0.13, P ⬍ .05). Only the Negative Responses to Pain Behavior scale made a significant independent contribution, with higher levels of negative responding associated with greater depression. As a group, the SRI scales also made a significant contribution to the prediction of pain-rated pain intensity over and above the effects of pain duration. This effect was due primarily to a positive association between patient perceptions of spouse negative responses to patient pain behaviors and pain intensity.
Discussion The present study describes the development and preliminary psychometric evaluation of a self-report inventory designed to assess spouse responses to patient pain and well behaviors. In its final form, the SRI is a 39-item inventory divided into 2 parts, spouse responses to pain behaviors (Part 1) and spouse responses to well behaviors (Part 2), with 2 subscales in each part. The subscales reflect solicitous and negative responses to patient pain behaviors and facilitative and negative responses to patient well behaviors. The test-retest stability coefficients (over a 2-week period) were uniformly high, and factor analyses of the scale items (performed for each scale separately) provided evidence that the SRI items associated with each scale can be interpreted as assessing a single primary domain of patient-perceived spouse responding. Moreover, the item analyses (correlation of each item with its parent scale and the loading of each item on a factor score representing each spouse response domain) indicated strong associations between the individual items
ORIGINAL REPORT/Schwartz et al Table 5.
249
Regression Analyses Predicting Patient Pain Behaviors (Pain Behavior Checklist)
PREDICTOR AND STEP 1: Pain variables (block) Pain intensity Pain duration 2: Spouse response inventory (block) Solicitous Responses to Pain Behavior Negative Responses to Pain Behavior Facilitative Responses to Well Behavior Negative Responses to Well Behavior
TOTAL R2 0.09
R2 CHANGE 0.09
F CHANGE

†
4.85
.23* ⫺.10 0.38
0.27
9.39† .49† .30† ⫺.22 .22*
*P ⬍ .05. †P ⬍ .01.
and the parent scale–factor score for each SRI scale. Overall, these findings provide strong initial support for the reliability and internal consistency of the SRI scales. The associations between the SRI scales and patient reports of their functioning were in directions consistent with what operant theory and previous research (which used other measures of spouse responding) would predict. However, because the SRI scales include responses to well behaviors, some novel findings emerged. For example, the SRI scales that assess patient perceptions of spouse responses that are likely to be viewed as positive by most patients (solicitous responses to pain behaviors and facilitative responses to well behaviors) were positively associated with one another, yet they evidenced different patterns of associations with patient-reported disability. Solicitous responses to pain behaviors were associated with greater disability, whereas facilitative responses to well behaviors were associated with less patient disability. This pattern of associations provides further support for the discriminant validity of these 2 SRI scales; if both scales were tapping into the same domain (perhaps a general perception of patient-perceived positive responding by their spouses), they should have shown a similar association to patient functioning. A positive relationship was seen between discouraging well behavior (Negative Responses to Well Behavior scale scores) and solicitous responding to pain behavior, but a significant positive correlation was also seen between encouraging well behavior (Facilitative Responses to Well Behavior scale) and solicitous responding to pain behavior. Although this might seem contradictory, these findings suggest that patterns of patient-spouse responding might be highly complex, relationship dependent, and context specific. Defining the interrelationships among types of perceived spouse responses, the factors that influence them, and the circumstances under which they are seen requires further investigation. It would have been interesting to determine the extent to which the 2 types of positive responding assessed by the SRI (solicitous responses to pain behaviors and facilitative responses to well behaviors) are associated with marital satisfaction or if marital satisfaction mediates the relationships between perceived spouse responses and patient outcomes. Unfortunately, we did not include a measure of marital satisfaction in this study, and there-
fore the potential contributions of patient perceptions of both positive and negative spouse responses to marital satisfaction must await further research. The results of the analyses using the SRI subscales as predictors of patient functioning were generally consistent with previous research that has examined the association between measures of spouse responses to patient pain behaviors and patient functioning domains and support the construct validity of the SRI subscales. Patient’s perceptions of spouse solicitous responses to pain behaviors, as measured with the SRI, were associated with greater patient-reported physical disability and a higher frequency of pain behaviors. Similarly, SRIassessed encouragement of well behaviors by spouses made an independent contribution to the prediction of lower patient disability, beyond the contribution of the SRI Solicitous scale. In addition, patient perceptions of spouse negative responses to well behaviors were associated with higher patient disability and pain behaviors. Associations between perceived spouse responses to well behaviors and measures of patient functioning have not been demonstrated in the empirical literature to our knowledge and represent understudied, yet theoretically and clinically important, relationships. These preliminary findings support the predictive validity of the SRI scales measuring responses to patient well behaviors and also support the study of the potential effect of spouse responses to patient well behaviors on patient functioning. If further research confirms these findings and the role that spouse responses to well behaviors might play in promoting improved patient functioning, this would support the development of empirically based couples interventions incorporating a focus on spouse responses to both pain and well behaviors to enhance functioning in patients with chronic pain. Patient perceptions of spouse negative responses to patient pain behaviors were predictive of patient reports of more frequent pain behaviors and were not associated with patient physical disability (RMDQ scores). Given the inconsistent findings of prior studies regarding spouse negative responses to pain behaviors, we did not hypothesize that these responses would be associated with pain behaviors or disability. Although such responses are classed as “negative” in terms of their content, they might or might not function to reduce the
250
Measuring Spouse Responses to Pain
Regression Analyses Predicting Patient Depressive Symptoms (Beck Depression Inventory)
Table 6.
PREDICTOR AND STEP 1: Pain variables (block) Pain intensity Pain duration 2: Spouse response inventory (block) Solicitous Responses to Pain Behavior Negative Responses to Pain Behavior Facilitative Responses to Well Behavior Negative Responses to Well Behavior †
TOTAL R2
R2 CHANGE
F CHANGE
0.01
0.01
0.61
 .10 .00
0.14
0.13
2.54* .06 .26* ⫺.14 .16
P ⬍ .01.
*P ⬍ .05.
frequency of the behaviors they follow (ie, act as punishers in operant behavioral terms). It is possible that such responses might in fact serve to increase or maintain pain behavior if attention of any type (either solicitous or negative) is reinforcing in certain situations or couples. This possibility was suggested by Papas et al,20 who found that a subgroup of patients who were classified as having spouses who were “negatively attentive” on the basis of their responses to the Multidimensional Pain Inventory also reported the highest scores on measures of pain, interference, and depression and the lowest scores on activity level (except social activity) and life control. The authors offered as one explanation of their findings the possibility that both positive and negative responses might have become salient attentional reinforcers of pain behaviors, particularly in the absence of other sources of attention, such as occupational and social contacts. Further research is needed using validated measures of spouse responding to test this hypothesis in a larger sample of couples and to determine whether marital satisfaction might play a role in determining the valence that such responses have for couples. Patient perceptions of spouse negative responses to patient pain behaviors were associated with greater depression in the patient. This finding is consistent with previous research on individuals with chronic pain using other measures of spouse negative responding14,20 and supports the validity of the Spouse Negative Responses to Patient Pain Behavior scale. The association between patient-perceived negative spouse responses and patient depression in this study is also consistent with the general marital and depression literature, which finds a strong association between negative behaviors in one partner and poorer functioning in the other partner.1,30 Initially, we were interested in evaluating whether there are differences in patient functioning associated with spouse solicitous responses to pain behaviors categorized as potentially positively reinforcing versus negatively reinforcing. However, there was such a strong degree of statistical association between the subscales measuring these 2 response categories that it seemed most appropriate to consolidate these scales. Moreover, the subscale designed to assess negative reinforcement
of well behaviors had poor psychometric qualities and consequently was eliminated from the inventory. Given the failure to create single scales that reflect patient perceptions of spouse responses that are potentially negatively reinforcing (to patient pain or well behaviors), the current findings suggest the possibility that although operant theory supports the concept of negative reinforcement as being distinct from positive reinforcement, it might be difficult for patients to distinguish between these 2 types of spouse responses. Alternatively, it is possible that patients are able to distinguish between these responses and accurately rate them but that spouses who provide frequent positive reinforcement might also be those who provide frequent negative reinforcement in response to patient pain behaviors. In this case an empiric distinction between these 2 types of spouse responses might be very difficult to achieve. Unfortunately, then, despite the potential for positive and negative reinforcement to have different effects on patient functioning, at this time, the SRI cannot be used to test for these potential differential effects. The current findings are limited by the study’s correlational design and reliance on patient self-report data. Self-reports might be influenced by a number of factors, such as social desirability and patient affect. Thus the results should be considered preliminary and warrant replications with multiple modes of measurement. A second limitation of the study is that the SRI items were rationally rather than empirically derived and thus form a restricted sample of potential responses to pain and well behaviors. It is quite possible that other types of spouse responses to both pain and well behaviors might occur in samples of couples in which one person has chronic pain. In addition, as indicated above, we did not assess marital satisfaction in this study. Because significant associations between measures of marital satisfaction and patient perceptions of spouse responses to pain and well behaviors likely exist and because previous research has shown that the association between spouse responses to patient behaviors and patient functioning might vary as a function of marital satisfaction,14 an examination of these associations would have strengthened the study.
ORIGINAL REPORT/Schwartz et al Table 7.
251
Regression Analyses Predicting Average Patient Pain Intensity (0-10 Numeric Rating
Scale) PREDICTOR AND STEP 2: Pain duration 3: Spouse response inventory (block) Solicitous Responses to Pain Behavior Negative Responses to Pain Behavior Facilitative Responses to Well Behavior Negative Responses to Well Behavior
TOTAL R2 0.07 0.13
R2 CHANGE 0.07 0.06
F CHANGE †
7.59 2.80*
 ⫺.27† .08 .24* .22 ⫺.01
*P ⬍ .05. †P ⬍ .01.
A potential limitation of the current version of the SRI is its limited focus on patient activity level as representing the domain of well behavior. Verbal behaviors that communicate hope for the future, perceived ability to cope with pain, or general positive self-statements might also be considered well behaviors. A number of theoretic models, particularly cognitive-behavioral theory,29 as well as a model of motivation for pain self-management,11 hypothesize a causal link between patient verbal responses (both what they say to others and what they tell themselves) and patient functioning. A potentially useful avenue for future research would be to examine the association between spouse responses to these verbal behaviors and measures of current and future patient functioning. Another limitation is that this study relied solely on self-report for both the SRI and the validity criterion measures. The reliance on self-report for all measures can work to artificially increase the association between variables because of shared method variance. In light of its limitations, the current study should be viewed as representing an initial examination of a new measure of spouse responses to patient behaviors; more work is needed to examine the reliability and validity of the measure in larger samples of diverse patients with chronic pain and their spouses-partners. Specifically, further research is needed to evaluate the relationship between the SRI and observed partner responses to patient pain and well behaviors, as well as other questionnaire measures of spouse responses to patient pain and well behaviors (ie, the WHYMPI) and measures of patient functioning. Such re-
search would be strengthened if criterion measures (of patient functioning) were assessed through direct observation rather than self-report. The sensitivity of the SRI to changes in treatment is also unknown. Further research is also needed to cross-validate the results on other samples of individuals with chronic pain. In summary, the findings from this study suggest that the SRI provides a relatively simple and reliable assessment of perceived responses of significant others to pain and well behaviors. The results also provide preliminary support for the construct validity of the SRI scales in that they demonstrate associations that are consistent with those predicted by theory and found using measures of similar constructs. The SRI might be useful in future research evaluating relationships between spouse responses and patient functioning. The assessment of spouse responses to well behaviors might also have important implications for the treatment of couples with chronic pain, given the possibility that encouragement of well behaviors might be associated with lower levels of disability and might contribute to maintaining treatment gains. Moreover, discouraging spouse solicitous responses to pain behaviors and spouse negative responses to both pain and well behaviors might also be important for adaptive patient functioning. Examining family responses that encourage or promote patient functioning is an important area for continued research, particularly in the current health care environment emphasizing decreasing health care use and improving health care outcomes.
References
4. Block AR, Kremer EF, Gaylor M: Behavioral treatment of chronic pain: the spouse as a discriminant cue for pain behavior. Pain 9:243-252, 1980
1. Beach SRH, Sandler EE, O’Leary KD: Depression in Marriage: A Model for Etiology and Treatment. New York, NY, Guilford, 1990 2. Beck AT, Ward CH, Mendelson M, Mock J, Erlbaugh J: An inventory for measuring depression. Arch Gen Psychiatry 4:461-571, 1961 3. Bergner M, Bobbit RA, Carter WB, Gilson BS: The sickness impact profile: development and final revision of a health status measure. Med Care 19:787-805, 1981
5. Cinciripini PM, Floreen A: An outcome evaluation of a behavioral program for chronic pain. J Behav Med 5:375389, 1982 6. Flor H, Kerns RD, Turk DC: The role of spouse reinforcement, perceived pain, and activity levels of chronic pain patients. J Psychosom Res 31:251-259, 1987. 7. Flor H, Knost B, Birbaumer N: The role of operant condi-
252
Measuring Spouse Responses to Pain
tion in chronic pain: an experimental investigation. Pain 95:111-118, 2002
ship between spouse solicitousness and pain behavior: searching for more evidence. Pain 51:75-79, 1992
8. Fordyce WE: Behavioral Methods for Chronic Pain and Illness. St Louis, Mo, C.V. Mosby, 1976
20. Papas RK, Robinson ME, Riley JL 3rd: Perceived spouse responsiveness to chronic pain: three empirical subgroups. J Pain 2:262-269, 2001.
9. Fordyce WE, Fowler R, Lehmann J, DeLateur B: Some implications of learning in problems of chronic pain. J Chronic Dis 21:179-190, 1968 10. Jensen MP, Karoly P, Braver S: The measurement of clinical pain intensity: a comparison of six methods. Pain 27:117126, 1986 11. Jensen MP, Nielson WR, Kerns RD: Toward the development of a motivational model of pain self-management. J Pain 4:477-492, 2003 12. Jensen MP, Strom S, Turner JA, Romano JM: Validity of the Sickness Impact Profile Roland Scale as a measure of dysfunction in chronic pain patients. Pain 50:157-162, 1992 13. Jolliffe CD, Nicholas MK: Verbally reinforcing pain reports: an experimental test of the operant model of chronic pain. Pain 107:167-175, 2004 14. Kerns RD, Haythornthwaite J, Southwick S, Giller EL: The role of marital interaction in chronic pain and depressive symptom severity. J Psychosom Res 34:401-408, 1990. 15. Kerns RD, Haythornthwaite J, Rosenberg R, Southwick S, Giller EL, Jacob MC. The Pain Behavior Checklist (PBCL): factor structure and psychometric properties. J Behav Med 14: 155-167, 1991 16. Kerns RD, Turk DC, Rudy TE: The West Haven-Yale Multidisciplinary Pain Inventory (WHYMPI). Pain 23:356-435, 1985 17. Kremer E, Atkinson JH, Ignelzi RJ: Measurement of pain: patient preference does not confound measurement. Pain 10:241-248, 1981 18. Loeser JD, Egan KJ: Managing the Chronic Pain Patient: Theory and Practice at the University of Washington Multidisciplinary Pain Center. New York, NY, Lippincott-Raven Publishers, 1989 19. Lousberg R, Schmidt AJM, Groenman NH: The relation-
21. Patrick L, D’Eon J: Social support and functional status in chronic pain patients. Can J Rehab 9:195-201, 1996 22. Revenson TA, Schiaffino KM, Majerovitz SD, Gibofsky A: Social support as a double-edged sword: the relation of positive and problematic support to depression among rheumatoid arthritis patients. Soc Sci Med 33:807-813, 1991 23. Roland M, Morris R: Development of a reliable and sensitive measure of disability in low back pain. Spine 8:141144, 1983 24. Romano JM, Turner JA, Friedman LS, Bulcroft RA, Jensen MP, Hops H: Observational assessment of chronic pain patientspouse behavioral interactions. Behav Ther 22:549-567, 1991 25. Romano JM, Turner JA, Jensen MP, Friedman LS, Bulcroft RA, Hops H, Wright SF: Chronic pain patient-spouse behavioral interaction predict disability. Pain 63:353-360, 1995 26. Roy R: Chronic pain and marital difficulties. Health Social Workers 10:199-207, 1995 27. Schwartz L, Slater MA, Birchler GR: The role of pain behaviors in the modulation of marital conflict in chronic pain couples. Pain 65:227-233, 1996 28. Turk DC, Kerns RD, Rosenberg R: Effects of marital interaction on chronic pain and disability: examining the downside of social support. Rehab Psychol 37:259-274, 1992 29. Turk DC, Meichenbaum D, Genest M: Pain and Behavioral Medicine: A Cognitive-behavioral Perspective. New York, NY, Guilford Press, 1983 30. Weissman MM: Advances in psychiatric epidemiology: rates and risks for major depression. Am J Public Health 77:445-451, 1987. 31. Williamson D, Robinson ME, Melamed B: Pain behavior, spouse responsiveness, and marital satisfaction in patients with rheumatoid arthritis. Behav Modif 21:97-118, 1997