Predictors of patient satisfaction.

Correlates of patient satisfaction at varying points in time were assessed using a survey with 2-week and 3-month follow-up in a general medicine walk-in clinic, in USA. Five hundred adults presenting with a physical symptom, seen by one of 38 participating clinicians were surveyed and the following measurements were taken into account: patient symptom characteristics, symptom-related expectations, functional status (Medical Outcomes Study Short-Form Health Survey [SF-6]), mental disorders (PRIME-MD), symptom resolution, unmet expectations, satisfaction (RAND 9-item survey), visit costs and health utilization. Physician perception of difficulty (Difficult Doctor Patient Relationship Questionnaire), and Physician Belief Scale. Immediately after the visit, 260 (52%) patients were fully satisfied with their care, increasing to 59% at 2 weeks and 63% by 3 months. Patients older than 65 and those with better functional status were more likely to be satisfied. At all time points, the presence of unmet expectations markedly decreased satisfaction: immediately post-visit (OR: 0.14, 95% CI: 0.07-0.30), 2-week (OR: 0.07, 95% CI: 0.04-0.13) and 3-month (OR: 0.05, 95% CI: 0.03-0.09). Other independent variables predicting immediate after visit satisfaction included receiving an explanation of the likely cause as well as expected duration of the presenting symptom. At 2 weeks and 3 months, experiencing symptomatic improvement increased satisfaction while additional visits (actual or anticipated) for the same symptom decreased satisfaction. A lack of unmet expectations was a powerful predictor of satisfaction at all time-points. Immediately post-visit, other predictors of satisfaction reflected aspects of patient doctor communication (receiving an explanation of the symptom cause, likely duration, lack of unmet expectations), while 2-week and 3-month satisfaction reflected aspects of symptom outcome (symptom resolution, need for repeat visits, functional status). Patient satisfaction surveys need to carefully consider the sampling time frame as well as adjust for pertinent patient characteristics.