With the growing interest in patient-centered care, patient evaluations of their own experiences have been increasingly used to improve the quality of care. Factors associated with satisfaction are thought to include the structure, process and outcome of care as well as patient sociodemographic, physical and psychological status, and attitudes and expectations concerning medical care. This study considers how factors that can be controlled by the organization (e.g. staff behavior, the environment), as well as factors that cannot be controlled (e.g. patient characteristics) impact satisfaction.
The objective of the study was to identify factors associated with satisfaction among inpatients receiving medical and surgical care for cardiovascular, respiratory, urinary and locomotor system diseases. Although numerous hospitals have developed ongoing programs for the routine assessment of patient perceptions of the quality of care, this study considers the factors that may affect level of satisfaction in a single, prospective study, making it possible for providers to target patients at risk for worse experiences.
Adult patients were eligible for the study when admitted to one to one of the study service units with a stay of more than three days - long enough to experience many aspects of care. The sample size of 550 was devised to allow a statistical detection of a difference in means of at least 6 points on a 100-point satisfaction scale. Two weeks after discharge, 533 patients completed a Patient Judgments Hospital Quality questionnaire that was adapted to a French setting (e.g. elimination of billing section) by using a core set of 48 of the original 106 items, plus an open-ended response question. This instrument covered seven dimensions of satisfaction (admission, nursing and daily care, medical care, information, hospital environment and ancillary staff, overall quality of care and services, recommendations/intentions). Patient sociodemographic, health and stay characteristics, and organization/activity of service were used as independent variables, with patient satisfaction and complaints treated as dependent variables. Analysis inlcuded multivariate ordinal polychotomous and dichotomous logistic stepwise regressions, respectively.
The two strongest predictors of satisfaction for all dimensions were:
- older age and
- better self-perceived health status at admission.
Men tended to be more satisfied than women. Other predictors specific for certain dimensions of satisfaction were: married, Karnofsky index more than 70, critical/serious self-reported condition at admission, emergency admission, choice of hospital by her/himself, stay in a medical service, stay in a private room, length of stay less than one week, stay in a service with a mean length of stay longer than one week.
The most common complaints were about the built environment - living arrangements and amenities such as the room, toilets and bathroom. Satisfaction scores were lowest in this area.
Author identified limitations include:
- A lack of generalizability because general internal medicine services were deliberately excluded to focus on services with higher technical levels of care. However weak differences between services suggest that patient satisfaction factors are similar among services.
- Sampling bias, as some individuals more likely to experience problems with care were excluded (e.g. those discharged to other institutions, those judged incapable of completing the questionnaire).
- Potential random error (unpredictable errors that vary from sample to sample) due to the subjectivity of choosing a unique threshold to define satisfied and nonsatisfied patients
- The nature of the statistcial model chosen, as the Ordinal polychotomous regression model assumes that the odds ratio from level to level of the cumulative categories is the same, which is not always true (threshold effects are likely to occur).