Health Outcomes and Costs of Referral for Elective Surgery: New Zealand 2003-2006

Abstract

Background: Elective surgery is a major output of the health sector. For instance, in 2010/11 approximately 145,000 patients received elective surgery in New Zealand. This study uses data from the ‘Pathways to care and elective surgery study,’ a cohort study performed between 2003 and 2006 of patients referred for elective surgery in New Zealand. A total of 1,603 individuals were included in the Pathways study, of whom 472 received elective surgery during the study period. Information on health-related quality of life (HRQoL) was measured at regular intervals, using two instruments – EQ-5D and SF-36. Information was also collected on patient costs, health sector costs, and societal costs as a whole. This information could provide guidance for improving the allocation of resources in the elective surgery sector. Objectives: 1. To estimate the health-related quality of life (HRQoL) outcomes of a sample of individuals referred for elective surgery using the EuroQol 5-Dimensions (EQ-5D) and Short-Form 6-Dimensions (SF-6D) instruments; and to examine the strength of association between the change in HRQoL pre-surgery and in the period up to 6 months following surgery, with the length of wait for surgery, and patient characteristics. 2. To measure the level of agreement between the EQ-5D and SF-6D HRQoL instruments and compare the performance of each instrument against the review criteria of practicality, reliability, and responsiveness. 3. To estimate the personal, government, and societal costs of individuals referred for elective surgery; and to estimate the strength of association between these costs, HRQoL, the length of wait for surgery, and patient characteristics. Methods: The EQ-5D was derived using Devlin’s algorithm and the SF-6D was derived using Brazier’s algorithm. Multivariate regression models were used to assess the determinants of HRQoL and log-transformed economic costs pre-surgery and in the period up to 6 months post-surgery (including the pre-surgery period). The agreement (correlation between instruments’ index scores), practicality (completion rate), test-retest reliability (reproducibility of repeated measures), and responsiveness (change following elective surgery) was assessed for the EQ-5D and SF-6D. Results: The median waiting time for surgery was 6.4 months. Regression analysis of repeated measures showed that, at the population level, HRQoL was stable over the pre-surgery period and significantly improved following surgery. Patients with more urgent conditions at the time of referral (as assessed by the referring GP) reported significantly worse HRQoL outcomes and higher costs over the referral period. The EQ-5D and SF-6D showed poor agreement at the time of referral. The test-retest reliability of the SF-6D (ICC: 0.640) was higher than the EQ-5D (ICC: 0.507), however the EQ-5D was more responsive to change in HRQoL following surgery. The mean societal cost of waiting for surgery was NZD 1,842 per patient, and increased to NZD 11,113 per patient in the period from referral to 6 months post-surgery. Higher HRQoL (better health) was associated with lower societal costs pre-surgery and in the period up to 6 months post-surgery. Conclusions: A first conclusion to be drawn from the analysis is that there are wide variations in the costs and benefits associated with referral for elective surgery across referral specialties. As a consequence, economics-based evaluations of competing elective surgeries may have important implications for the allocation of resources in the elective surgery sector. A second conclusion concerns the measures of health-related quality of life. The EQ-5D and SF-6D instruments did not provide interchangeable index scores. Moreover, there were clinically important differences in the performance of each instrument across repeated interviews in the measurement of HRQoL outcomes. Researchers undertaking and interpreting quality of life research need to give careful consideration to these differences and what they might mean for policy recommendations.