Patient satisfaction revisited: a multilevel approach

Area of Science:

• Healthcare Management
• Health Services Research
• Patient Experience

Background:

• Patient satisfaction surveys are crucial for healthcare benchmarking.
• Current Dutch practices report satisfaction univariately, ignoring case mix factors.
• This study addresses the need for case mix adjustment in patient satisfaction data.

Purpose of the Study:

• To determine if patient satisfaction differences stem from hospital, department, or patient characteristics.
• To identify suitable case mix variables for correcting patient satisfaction scores in benchmarking.

Main Methods:

• Cross-sectional surveys of patients from Dutch academic and general hospitals in 2005.
• Multilevel analysis using intra-class correlation coefficients (ICCs) to assess variance at hospital and department levels.
• Investigation of patient characteristics (age, gender, education, health status, mother language) and hospital-level factors (size, type, population density, response rate).

Main Results:

• Intra-class correlation coefficients (ICCs) for hospital and department levels ranged from 0% to 4%, indicating minor variance attributed to these levels.
• Patient characteristics, particularly age, health status, and education, significantly influenced satisfaction scores.
• Hospital-level factors and other patient characteristics showed less impact on satisfaction variance.

Conclusions:

• A substantial portion of patient satisfaction variance resides at the individual patient level.
• Hospital and department levels account for only a minor part of the variance in patient satisfaction.
• Age, health status, and education are recommended as key case mix variables for correcting patient satisfaction outcomes in the Netherlands.