Development and first assessment of a questionnaire for health care utilization and costs for cardiac patients

Abstract

Background: The valid and reliable measurement of health service utilization, productivity losses and consequently total disease-related costs is a prerequisite for health services research and for health economic analysis. Although administrative data sources are usually considered to be the most accurate, their use is limited as some components of utilization are not systematically captured and, especially in decentralized health care systems, no single source exists for comprehensive utilization and cost data. The aim of this study was to develop and test a questionnaire for the measurement of disease-related costs for patients after an acute cardiac event (ACE).

Methods: To design the questionnaire, the literature was searched for contributions to the assessment of utilization of health care resources by patient-administered questionnaires. Based on these findings, we developed a retrospective questionnaire appropriate for the measurement of disease-related costs over a period of 3 months in ACE patients. Items were generated by reviewing existing guidelines and by interviewing medical specialists and patients. In this study, the questionnaire was tested on 106 patients, aging 35-65 who were admitted for rehabilitation after ACE. It was compared with prospectively measured data; selected items were compared with administrative data from sickness funds.

Results: The questionnaire was accepted well (response rate = 88%), and respondents completed the questionnaire in an average time of 27 minutes. Concordance between retrospective and prospective data showed an intraclass correlation (ICC) ranging between 0.57 (cost of medical intake) and 0.9 (hospital days) with the other main items (physician visits, days off work, medication) clustering around 0.7. Comparison between self-reported and administrative data for days off work and hospitalized days were possible for n = 48. Respective ICCs ranged between 0.92 and 0.94, although differences in mean levels were observed.

Conclusion: The questionnaire was accepted favorably and correlated well with alternative measurement approaches. This first assessment showed promising characteristics of this questionnaire in different aspects of validity for patients with ACE. However, additional research and more extensive tests in other patient groups would be worthwhile.

Figure 1

Bubble plot of difference (self-report–administrative data) by mean days off work (n = 48). The bold solid line represents the regression line; the thin line is drawn at zero representing exact agreement. The bubble size is proportionate to the number of patients with corresponding values.