The Human Immunodeficiency Virus (HIV) is a global public health issue with more than 36 million affected. The highest prevalence is found in Sub-Saharan Africa where nearly 5 % of adults live with HIV, being 69 % of all people living with HIV in the world [1]. Although HIV mortality was huge in the 1980s, the development in the 1990s of highly active antiretroviral therapy (ART) was determinant to reduce rates of death and mother-to-child transmission. However, only in developed countries, with general access to ART, HIV behaves as a chronic condition. By 2012, more than half of the HIV world population is in developing countries with inequalities in the access to ART and with a huge rate of mortality [2].
Hence, there are two completely different realities in health policy and planning when looking at HIV. First, in developed countries with mostly universal access to ART, HIV is a chronic condition. There, the challenge in health systems is to become high-performing chronic care systems [3] and efficient given the increasing evolution of health expenditures [4]. That is feasible by looking at the way in which health services are provided [5] to chronic patients and to whether patients are suffering more than one chronic condition at the same time [6] in order to adapt their demand to their need. Here, it remains unknown whether there is equity in the use of health services by patients of different socioeconomic background even when access is guaranteed by the health system. Second, in low and middle-income countries where the challenge is to increase the access to ART [7]. Worldwide, inequity in access is a crucial challenge dealing with HIV. However, most studies look at inequity from the second perspective, focusing on increasing international rates of access to ART where it is needed [8], its role on prevention [9], on mortality [10], its impact in different groups of individuals as sex workers [11] or injecting drug users [12], or early diagnosis in infants [13]. We contribute to the literature by developing a different analysis looking at the first mentioned reality: HIV health policy and planning in developed countries where access to ART is guaranteed.
We look at the Basque Country, a region in Spain, where access to public healthcare services is universal and free at the point of use, where ART is free and patients only pay a copayment rate in other pharmacological treatments [14]. We utilize individual data from the entire Basque population (2.26 million inhabitants) on diagnoses, socioeconomic information and standardized health expenditures. We test whether patients of different socioeconomic background equally use different types of health services.
Data and methods
We utilize the database prepared by the population stratification program (PREST) of the Basque Country including the practical totality of its population: every individual covered on 31 August 2011 by the public health insurance in the Basque Country and who was covered for at least 6 months in the previous year, regardless of whether they made any contact with or use of the Basque Health Service. The analysis refers to one year, from September 1st, 2010 to August 31st, 2011. There are 2,262,698 individuals, being 50.90 % female. As for the age distribution, 15 % are children (younger than 18) and 20 % are over 65, being the average individual 43.69 years old. It is therefore important to remark that it is not a random sample but the real population in terms of health policy and planning.
Our dataset combines three types of information. First, diagnoses information is based on hospital discharges, emergency department, primary care medical records, and prescriptions. They all are coded according to the ICD-9-CM [15] (diagnoses) and ATC [16] (pharmaceuticals). Second, out of utilization we obtain individual standardized health expenditures. Third, we utilize socioeconomic information. For the sake of our analysis, we only look at the HIV diagnosis in order to differentiate among HIV patients and non-HIV reported individuals. With respect to health expenditures, the cost of the public health services provision is based on use. However, there are no market prices within the Basque Health Service and costs are estimated through standardization of total health expenditures per type of service. We take into account the number of visits to primary care, specialist care, Accident & Emergency, rehabilitation sessions, outpatient care, laboratory tests, radiological examinations, and various outpatient procedures such as dialysis, radiotherapy and chemotherapy. Cost of hospitalization and outpatient surgery is assigned through the cost-weights of the corresponding diagnosis-related groups (DRGs). Finally, the cost of ART, provided in public hospitals in the Basque Country, has been calculated as the average cost of all ART provided in 2012. Total number of patients with ART was 5002 and total expenditure in ART treatment was of 38501376€, for an average cost per treatment of 7697€. Finally, the cost of pharmaceutical prescriptions (excluding ART) recorded in electronic health records is based on market prices. Information on socioeconomic status is derived out of the deprivation index (DI), an ordinal variable elaborated for Spain in 2008 [17] categorizing into five socioeconomic groups (SEG) by quintiles. The DI allows for the estimation of socioeconomic and environmental inequities among inhabitants by censal code. It takes into account five dimensions including the percentages of residents who are manual workers, unemployed, temporary employees, or have an inadequate level of educational attainment, overall and also specifically among young people.
We use OLS regressions to identify whether there are inequities in the use of of health services provision of any type for HIV patients in a population in which access to ART and other treatments is free and granted. Following the risk adjustment literature, our dependent variable is health expenditures for HIV patients on age groups, which, as mentioned above, is directly related to the utilization of health services (standardized health expenditures by use). With respect to our independent variables, because we do not observe all variables that might affect health expenditures we use fixed effect by socioeconomic groups (SEG). We do the same for the different types of health expenditures (provision) and for total health expenditures. In that further estimation approach, we take into account the individual age and the number of chronic conditions suffered as the risk adjustment literature states that greater need of health provision derives in greater expected health expenditure [18]. The aim is to control for healthcare needs, taking the individual number of comorbidities as a proxy. In order to construct those variables, we utilize a list of 52 health conditions defined by the research team based on the related literature [19–21]. We capture the role of multimorbidities by defining three categories: 1 to 3 comorbidities, 4 to 6, and 7 or more. The specification for our estimation model is given by:
$$ Health\kern0.5em Expenditure{s}_i={\displaystyle \sum_j{\alpha}_j ag{e}_{ij}+{\displaystyle \sum_k{\beta}_kSE{G}_{ki}+{\displaystyle \sum_l{\delta}_l NMor{b}_{li}+{\varepsilon}_i}}} $$
