Erratum to “Non-organic symptom reporting in patients with chronic nonmalignant pain” [Pain 68 (1996) 329-342

Abstract

Patients with chronic non-malignant pain are often suspected of reporting medical symptoms that have non-organic as opposed to purely organic origins. According to the somatization hypothesis, non-organic reporting occurs when affective or other benign physical sensations are misconstrued as symptoms of physical disease. Psychological tests purporting to assess somatization are limited by their self-report format and may be confounded in patients with physical disease or injury. Measures of somatization may also be influenced or biased by underlying differences in depression or anxiety. In order to obtain an unbiased estimate of somatization, therefore, it is necessary to control for the influence of extraneous variables. In the present study, symptom report scales designed to assess somatization, symptom amplification, and disease conviction were administered to a group of 100 patients with chronic non-malignant pain. The strategy was to determine whether any of these tests could account for individual differences in illness behavior. Specifically, the set of dependent measures included: length of disability; frequency of medical visitation; activity level; and level of domestic functioning. The most successful predictor of patient behavior was the Somatization Scale (Derogatis et al. 1974) which correlated positively and significantly with each dependent measure. In order to examine the possibility that scores on this test were biased by differences in organic pathology, three physician pain specialists were asked to rate the morbidity of each item on the scale. A multiple regression analysis was then performed to examine whether differences in symptom morbidity, depression, or anxiety could account for the correlation between symptom ratings and illness behavior. The analysis showed that while depression and anxiety were significantly correlated with measures of illness behavior, the Somatization Scale still accounted for a significant amount of unique variance in three out of five dependent variables. Symptom morbidity was significantly correlated with only one measure of illness behavior (Activity Level). In view of these findings, scores on the Somatization Scale were used to classify 25 patients as Symptom Minimizers and another 25 as Symptom Amplifiers. When compared to Minimizers, Amplifiers were disabled for a significantly greater number of days, reported significantly more impairment in domestic functioning, were significantly less active, visited the doctor significantly more often, and were significantly more distressed. The results suggest that substantial differences in disability and medical visitation may exist among patients who may not differ appreciably in their level of organic pathology. Instead, differences in illness behavior may, to some extent, be mediated by differences in somatization.