Table I.
Multilevel univariable analysis of factors associated with being a “difficult patient,” accounting for clustering at the physician level
| Variable | Coefficient (SE) | P value | |
|---|---|---|---|
| Patient demographics | |||
| Age | 0.04 (0.01) | <.001∗ | |
| Sex (reference group: males) | Female | 0.18 (0.34) | .600 |
| Race (reference: Chinese) | Malay | −0.34 (0.55) | .527 |
| Indian | −0.37 (0.60) | .537 | |
| Caucasian | −2.09 (2.25) | .355 | |
| Others | −1.51 (0.76) | .047∗ | |
| Marital status (reference: single) | Married | −1.26 (0.43) | .003∗ |
| Dating | 0.41 (0.65) | .523 | |
| Divorced/separated | 0.45 (0.85) | .595 | |
| Widowed | −0.88 (1.51) | .557 | |
| Education level | −0.14 (0.19) | .447 | |
| Visit type (reference group: follow-up visit) | First visit | 0.30 (0.46) | .517 |
| Paying rate (reference group: subsidized) | Self-paying | 0.42 (0.63) | .506 |
| All the following variables were adjusted for patient age, sex, and race | |||
| Disease characteristics | |||
| Disease type (reference: eczema) | Psoriasis | −0.88 (0.42) | .035∗ |
| Disease duration (y) | 0.01 (0.02) | .416 | |
| Objective disease severity | 0.01 (0.005) | .004 | |
| Patient cognitive and emotional constructs | |||
| Quality of life impairment | 0.57 (0.11) | <.001∗ | |
| Disease cyclicity | 0.75 (0.23) | .001∗ | |
| Illness coherence | −0.73 (.24) | .002∗ | |
| Personal control | −0.90 (0.26) | .001∗ | |
| Treatment control | −0.88 (0.29) | .003∗ | |
| Resilience | −0.069 (0.05) | .126 | |
| Anxiety | 0.13 (0.03) | <.001∗ | |
| Consultation factors | |||
| Patient’s self-consciousness | 0.07 (0.15) | .636 | |
| Patient’s need to emphasize symptoms | 0.43 (0.16) | .007∗ | |
| Patient’s reporting of physician trust | −0.09 (0.03) | .003∗ | |
| Patient’s reporting of physician empathy | −0.07 (0.06) | .248 | |
| Physician’s perception of patient's degree of symptom reporting | 3.62 (0.34) | <.001∗ | |
| Physician’s management plan | Escalated or switched treatment | 2.15 (0.36) | <.001∗ |
| Reframe mindset by comparing with other patients | 3.25 (0.43) | <.001∗ | |
| Set expectations re. chronicity | 2.46 (0.37) | <.001∗ | |
| Explored nonmedical factors contributing to symptoms | 1.75 (0.39) | <.001∗ | |
| Taught coping strategies for symptoms | 1.90 (0.46) | <.001∗ | |
| Spent longer time than usual listening to the patient | 5.37 (0.39) | <.001∗ | |
| Wrote a memo for the purpose of excuses | 7.42 (1.00) | <.001∗ | |
| Felt pressure to wrap up the consult | 9.04 (0.95) | <.001∗ | |
| None of the above used | −4.24 (0.45) | <.001∗ | |
P < .05. Disease cyclicity, illness coherence, as well as personal and treatment control were assessed using the revised illness perception questionnaire; resilience was assessed using the Brief Resilience Scale; and anxiety was assessed using the Generalised Anxiety Disorder-7 scale. Self-consciousness and the need to emphasize symptoms were single questions scored on the Likert scale. The objective disease severity was taken as a product of body surface area and investigator/physician global assessment score. Perceived patient exaggeration of symptoms was reported by physicians as underplaying, overplaying, or reported symptoms accurately. The components of the physician management plan were scored as done or not done. Additional variables tested and found to be not significant include the following: patient’s personality using the 10-Item Personality Index, physician demographics (age, sex, race, rank, years of experience) and physician burnout.3