RESUMO
Inpatients' satisfaction with hospital care is often assessed by questionnaire. From a psychometrical standpoint, this method has drawbacks, however. We further investigated which item response format would maximise desirable outcomes regarding characteristics of the sample obtained (response rate and representativeness) and psychometric properties of the instrument (e.g. missing items responses, variance, validity) as an initial study into this question was indecisive. Subjects were 1184 discharged inpatients, of which 728 patients (62%) responded. They filled out a 55-item satisfaction questionnaire, addressing 12 aspects of care, using either a 10-point Evaluation scale ranging from "very poor" to "excellent" (E10) or a 5-point Satisfaction scale ranging from "dissatisfied" to "very satisfied" (S5). Both E10 and S5 showed good psychometric properties, but S5 yielded a better score distribution. Other results also favoured S5, but differences were small. In conclusion, different response formats do not yield widely different results. Thus, in choosing between them, convenience could be a decisive factor.
