Configural frequency analysis as a method of determining patients' preferred decision-making roles in dialysis

BMC Med Inform Decis Mak. 2010 Sep 11;10:47. doi: 10.1186/1472-6947-10-47.


Background: Numerous studies examined factors in promoting a patient preference for active participation in treatment decision making with only modest success. The purpose of this study was to identify types of patients wishing to participate in treatment decisions as well as those wishing to play a completely active or passive role based on a Germany-wide survey of dialysis patients; using a prediction typal analysis method that defines types as configurations of categories belonging to different attributes and takes particularly higher order interactions between variables into account.

Methods: After randomly splitting the original patient sample into two halves, an exploratory prediction configural frequency analysis (CFA) was performed on one-half of the sample (n = 1969) and the identified types were considered as hypotheses for an inferential prediction CFA for the second half (n = 1914). 144 possible prediction types were tested by using five predictor variables and control preferences as criterion. An α-adjustment (0.05) for multiple testing was performed by the Holm procedure.

Results: 21 possible prediction types were identified as hypotheses in the exploratory prediction CFA; four patient types were confirmed in the confirmatory prediction CFA: patients preferring a passive role show low information seeking preference, above average trust in their physician, perceive their physician's participatory decision-making (PDM)-style positive, have a lower educational level, and are 56-75 years old (Type 1; p < 0.001) or > 76 years old (Type 2; p < 0.001). Patients preferring an active role show high information seeking preference, a higher educational level, and are < 55 years old. They have either below average trust, perceive the PDM-style negative (Type 3; p < 0.001) or above average trust and perceive the PDM-style positive (Type 4; p < 0.001).

Conclusions: The method prediction configural frequency analysis was newly introduced to the research field of patient participation and could demonstrate how a particular control preference role is determined by an association of five variables.

Publication types

  • Randomized Controlled Trial

MeSH terms

  • Adult
  • Aged
  • Decision Making*
  • Female
  • Germany
  • Humans
  • Information Seeking Behavior*
  • Male
  • Middle Aged
  • Models, Theoretical
  • Patient Preference / psychology*
  • Physician-Patient Relations*
  • Predictive Value of Tests
  • Renal Dialysis / psychology*