Support for theoretical conceptualizations of the role of close relationships in pain
The operant model of pain (Fordyce 1976) is best known as advancing the notion that significant others including spouses have active roles in the experience of pain. Fordyce (1976) argued that the reinforcement of pain behaviours and the extinction of well behaviours may explain why acute pain behaviours from surgeries or injuries can persist over time. Reinforcement may also explain the persistence of pain behaviours once such behaviours become chronic. Specifically, spouses may reinforce pain behaviours by providing attention or help. Reinforcing or solicitous pain behaviours such as these may lead to increases or at least the maintenance of chronic pain behaviours including verbal expressions of pain, excessive rest, or ambulation problems. As a result, solicitous behaviours may actually encourage pain behaviours, which can lead to greater disability. Spouses may also punish pain behaviours by expressing anger or other forms of negative affect. Finally, spouses may ignore pain behaviours and reinforce well behaviours, both of which would theoretically lead to an extinction of pain behaviours and an increase in activity.
Very little research has tested the operant model in couples facing acute pain. One study examined pain, depression, and spouse social support after a laparoscopic radical prostatectomy as a treatment for prostate cancer (Knoll et al. 2007). Higher levels of pain as reported by patients at two days post-surgery were associated with a drop in their spouses’ provision of instrumental support from 2 days to 2 weeks following the surgery. This was only for pain in other body sites, not at the site of the surgery. The authors suggest that spouses may have interpreted these pains (e.g. headache, sore throat) as attention-seeking behaviours rather than genuine ailments, and so they deliberately reduced their instrumental support as a way to prevent the patients from adopting the sick role.
To date, the strongest evidence for the operant model is found in studies of chronic pain couples, which have been conducted by Romano and colleagues (1991, 1992, 1995, 2000). These studies are of particular value because they are observational studies, and thus, they are a more rigorous test of the model. Couples in which one member had a chronic pain condition were compared to healthy control couples (Romano et al. 1991, 1992, 1995). Participants were videotaped in the laboratory doing a series of routine household activities: sweeping the floor, changing bed sheets, bundling newspapers, and carrying fire logs across the room. They were instructed to perform the tasks together, stressing that the pain patient must be involved at all times. Trained raters coded the videos for six categories of behaviours: Non-verbal pain, verbal pain, solicitous, facilitative, aggressive, and distressed. Non-verbal pain behaviours included facial expressions, such as grimacing, and bodily actions such as limping or groaning; verbal pain behaviours included statements that indicated pain or functional limitations; solicitous behaviours included offering or giving assistance, and statements that expressed concern for the other; facilitative behaviours included compliments, encouragement, and humour; aggressive behaviours involved negative affect directed towards the other (e.g. disapproval, threats, arguments, disagreement); and distressed behaviours included complaints not directed towards the spouse, sadness, and whining. Preliminary analyses revealed that pain patients showed higher rates of overt verbal and non-verbal pain behaviours, and their spouses showed higher rates of solicitous behaviour, compared to control couples (Romano et al. 1991). In addition, pain patients and their spouses both showed significantly lower rates of facilitative behaviours than the control couples, though groups did not differ on the rate of aggression or distressed behaviours (i.e. negative affect). In subsequent analyses, the authors found that solicitous spouse behaviours preceded and followed both verbal and non-verbal pain behaviours more often in the pain couples than in the control couples (Romano et al. 1992). When they examined additional pain adjustment variables, they found that the sequence of spouse solicitousness in response to the patient’s non-verbal pain behaviour was a significant predictor of physical dysfunction, but only in more depressed patients (Romano et al. 1995).
Another study with similar methodology but using additional household chore tasks (folding laundry, cleaning up toys, building a bookcase from bricks and boards, then loading books out of boxes onto the bookcase) confirmed earlier findings: partner solicitous responses to the patient’s pain behaviours were significantly positively associated with the rate of patient non-verbal pain behaviours, while negative partner responses were significantly inversely associated with patient pain non-verbal behaviour rates (Romano et al. 2000). Regression analyses revealed that the rate of partner solicitous behaviours significantly contributed to the prediction of rates of both patient verbal and non-verbal pain behaviours, but partner negative behaviours did not. Taken together, the results of these studies suggest that spouse solicitous behaviours influence patients’ nonverbal pain behaviours. It was noted that the rate of verbal pain behaviours in these studies were low, so the results may have been dampened by restriction of range. Different kinds of interaction tasks may be more appropriate for analysing verbal behaviours, such as a discussion. Another limitation to these studies is that the sequential data were analysed using z scores, which are affected by the number of behaviours that are observed. In other words, couples should not be compared on z scores unless they exhibit exactly the same number of behaviours (Bakeman & Gottman 1997).
Nevertheless, these observational studies are invaluable because they provide convincing evidence that solicitous spouses can maintain pain and disability in chronic pain patients. The results were also confirmed by Paulsen and Altmaier (1995) who found, using similar tasks which only the patient completed, that patient- and spouse-reported spouse solicitousness were consistently associated with higher levels of observed patient pain behaviours. Patients with solicitous spouses engaged in more pain behaviours than patients with non-solicitous spouses, regardless of whether the spouse was present or not. A study by Lousberg and colleagues (1992) also sought to examine this relationship using a different behavioural task. Chronic low back pain patients engaged in a walking-to-tolerance treadmill test during which they walked on a treadmill that became gradually steeper. They were instructed to continue until they had to stop due to pain or fatigue. Each participant performed the test twice: once in the presence of his/her spouse, and once without the spouse. Pain behaviours were measured as pain intensity on a visual analogue scale and the total time spent on the treadmill, while spouse solicitousness was reported by both patients and spouses on a self-report measure. Patients who had solicitous spouses (as rated by the spouse, after a median split) reported increased pain intensity, and they spent less time on the treadmill when their spouse was present, compared to when they were alone. Patients with nonsolicitous spouses showed the opposite effect; they spent more time on the treadmill and reported less pain. The authors interpret this as an increase in pain behaviours in the presence of a solicitous spouse. These results, however, were not found when they used the patients’ ratings of spouse solicitousness: regardless of the patient’s rating of spouse solicitousness, the presence of the spouse during the walking test resulted in more pain and shorter walking time. It is possible that patients and spouses had widely different perceptions of the spouse’s solicitousness with varying degrees of accuracy, which caused these discrepant results.
There are some limitations to how these observational studies can be interpreted. Most importantly, the data are cross-sectional and cannot be interpreted causally. Also, the patients who participated already had chronic pain conditions, so it is not certain that the observed behaviours were related to the progression of an acute condition to a chronic one. Finally, observational studies do not account for how patients interpret spousal behaviours. Nevertheless, the evidence generally supports the operant model’s hypothesis that solicitous spouse behaviours should be detrimental to patients’ pain adjustment.