• David A Fishbain, Daniel Bruns, John M Disorbio, John E Lewis, and Jinrun Gao.
• Department of Psychiatry, Miller School of Medicine at University of Miami, Florida, USA. d.fishbain@miami.edu
• Pain Med. 2010 May 1;11(5):658-69.

ObjectiveIllness uncertainty (IU) theory proposes that patients with chronic illness may have difficulty adjusting to the illness if there is significant diagnostic or prognostic uncertainty. Two dimensions of IU theory are "lack of information about diagnoses or severity of the illness" (LIDSI) and "complexity regarding the health care system" (CRHCS). The primary objective of this study was then to compare the prevalence of IU in community nonpatients, community patients, and rehabilitation patients without pain/chronic pain patients (CPPs)/acute pain patients (APPs) as represented by two items with possible face validity for LIDSI ("doctors puzzled by my problems,"doctors missed something important") and three items with possible CRHCS face validity ("doctors don't believe me,"I need to prove my problem is real,"doctors think my problems are in my head"). The secondary objectives were to determine if the LIDSI items are associated with the CRHCS items and to develop predictor models for the LIDSI items in APPs and CPPs.DesignThe Battery for Health Improvement Research (BHI-R) version was administered to a healthy (pain-free) community sample (N = 1,478), community patient sample (N = 158), rehabilitation patients without pain (N = 110), rehabilitation APPs (N = 326), and rehabilitation CPPs (N = 341). The IU LIDSI and CRHCS items were contained within the BHI-R. These five patient groups were compared for the risk of endorsement of these items. Correlations were developed between the LIDSI and CRHCS items in APPs and CPPs. APPs and CPPs that affirmed IU items were compared with those not affirming the item on a wide range of demographic variables and Behavior Health Inventory (BHI 2) scales. Significant variables (P < or = 0.01) were then utilized as independent variables in predictor models for the LIDSI items.SettingCommunity patients and nonpatients, patients from physical therapy/work hardening/chronic pain/vocational rehabilitation programs, and physicians' offices.ResultsAffirmation for the LIDSI items ranged from 5.04% (community healthy) to 24.9% (CPPs) and for the CRHCS items, from 3.16% (community healthy) to 29.6% (CPPs). CPPs were significantly more likely than community patients to endorse one of the LIDSI items (doctors puzzled by my problem) plus all the CRHCS items. APPs, however, were no more likely than community patients to endorse any LIDSI IU items and two out of the three CRHCS items. LIDSI items were significantly correlated with the CRHCS items in both APPs and CPPs. The following items entered the final logistic regression models for LIDSI in APPs and CPPs: CRHCS items (APPs); items from the Doctor Dissatisfaction scale of the BHI 2 and the scale itself (APPs and CPPS); items related to faulty patient memory (APPs and CPPs); and various other items such as "hard muscles," etc. The models classified 87% (puzzling medical problem) and 91% (doctors missed something) of the APPs correctly. For CPPs, the models classified 79% (puzzling medical problem) and 88% (doctors missed something) of the patients correctly. None of these classifications, however, were better than the base rate.ConclusionLIDSI and CRHCS IU is not unusual in nonpatient and patient groups. However, rehabilitation CPPs are at significantly greater risk than community patients for LIDSI and CRHCS IU. LIDSI IU is associated with CRHCS IU, and LIDSI IU is predicted by a large number of items, the most notable of these being perception of not being believed and dissatisfaction with the physician.

18 keywords...

hide…