Analysing performance of primary health care using the ACG Case-Mix System

Sammanfattning: The overall aim of this thesis was to analyse the importance of co-morbidity level for different aspects of primary health care (PHC) using a measure of co-morbidity. Paper I examines whether there were differences in referral rates from different Lithuanian PHC models after adjustment for co-morbidity level. Paper II assesses to what extent the co-morbidity level could explain concurrent costs in Swedish PHC. Paper III evaluates the importance of age, gender and co-morbidity level on an individual’s choice of listing with either a public or a private PHC practice. Paper IV studies the importance of other factors such as gender, age, geographical distance and socioeconomic status (SES) for secondary health care use when taking co-morbidity level into account. In Paper I patients listed with rural state-owned PHC practices had significantly lower referral rates to secondary health care (SHC) than those in the other three models when adjusted for co-morbidity level. A patient’s increased co-morbidity level correlated with a higher physician- to self-referral ratio. In Paper II gender, age and listing with specific PHC explained only 15% of the variance in individual costs for PHC. When co-morbidity level measured with the ACG Case-Mix System was added the explanation of costs increased to 60–63%. Paper III found that older individuals, women and those with a higher co-morbidity level were listed with the public practice more often than with the private one. Higher co-morbidity level and older age were also associated with higher odds of re-listing to the public practice and remaining listed there. In Paper IV, with co-morbidity level taken into account, other factors influenced higher use of SHC. Men in the studied population were more likely to visit emergency wards. Male gender of the patient was associated with higher SHC costs and number of hospitalization days. Odds of incurring SHC costs decreased with higher age in the population. However, higher age of patients was associated with higher SHC costs and hospitalization days. The SHC costs and hospitalization days of patients living further from hospital were significantly lower. Both higher education and lower income were independently related to higher odds of incurring SHC costs in the population. Contrary to this, patients with high income had lower rates of emergency ward visits, SHC costs and hospitalization days. Co-morbidity level and patient’s place of residence are important factors that explain referral patterns. Patients referred by physicians are more likely to have a higher co-morbidity level, which indicates a positive role of gatekeeping. The co-morbidity level as measured by the ACG Case-Mix System explains patients’ costs in PHC to a high degree and can be a useful tool for creating a needs-based reimbursement system. Measuring co-morbidity level can help us understand more about the choice of PHC provider in elderly and patients with one or more chronic diseases. This could be important for physicians when treating patients and for health care managers when planning and organizing PHC. In addition to co-morbidity level the other factors as gender, age, distance to SHC and socioeconomic status (SES) influence use of SHC. Individuals with different SES, gender or living further from SHC may have difficulties and needs that PHC should better recognize.