Socioeconomic inequalities in expenditures and income committed to the purchase of medicines in Southern Brazil

OBJECTIVE: To describe socioeconomic inequalities regarding the use, expenditures and the income committed to the purchase of medicines. METHODS: A cross-sectional population-based study was carried out with 1,720 adults living in the urban area of Florianópolis, Southern Brazil, in 2009. Cluster sampling was adopted and census tracts were the primary sampling units. Use of medicines and the expenditures incurred in their purchase in the past 30 days were investigated through interviews. Use, expenditures and the income committed concerning medicines were analyzed according to per capita family income, self-reported skin color, age and sex, adjusting for the complex sample. RESULTS: The prevalence of medicine use was 76.5% (95%CI: 73.8; 79.3), higher among women and in older individuals. The mean expenditure on medicine was R$ 46.70, with higher values among women, whites, older individuals and among richer people. While 3.1% of the richest committed more than 15% of their income to purchasing medicine, that fi gure reached 9.6% in the poorest group. The proportion of people that had to buy medicines after an unsuccessful attempt to obtain them in the public health system was higher among the poor (11.0%), women (10.2%) and the elderly (11.1%). A large part of the adults bought medicines contained in the National List of Essential Medicines (19.9%) or in the Municipal List of Essential Medicines (28.6%), with signifi cant differences according to gender, age and income. CONCLUSIONS: There is socioeconomic, age and gender inequality in the income committed to the purchase of medicines, with worse conditions for the poor, older individuals and women. DESCRIPTORS: Income. Health Expenditures. Pharmaceutical Preparations. Health Inequalities. Cross-Sectional Studies.


INTRODUCTION
The Brazilian health system has been legally universal, equitable and integral since the end of the 1980s.The Sistema Único de Saúde (SUS -National Health System), which was created in Brazil in 1990, has incorporated actions to consolidate its principles and guidelines, such as the guarantee of integral therapeutic assistance, including pharmaceutical assistance.
Brazil is one of the few countries in which free access to essential medicines is a right of the citizens, materialized by means of policies and strategies, like Relação Nacional de Medicamentos Essenciais (RENAME -National List of Essential Medicines) and Relação Municipal de Medicamentos Essenciais (REMUME -Municipal List of Essential Medicines). 4 Generic drugs were introduced in Brazil in 1999 to reduce the price of medicines and facilitate the access of the entire population to these inputs.In 2004, the federal government launched Programa Farmácia Popular (Popular Drugstore Program), which aims to reduce the impact of medicines' prices on the family budget based on the co-payment system, in which the public power pays approximately 90% of the medicine's value and the citizen pays directly the rest.a Despite the advancements in the sector's public policies, they do not reach a large part of the population, mainly concerning the rational use of medicines, both in the private market and in the public sector. 7The global investment in health (public and private expenditures) was equivalent to 7.6% of the Brazilian gross domestic product in 2003, 19 showing a lower participation of the public sector in this proportion (45.3%), a value that is close to the one observed in countries with nonuniversal health systems. 19The low public investment in health causes higher private expenditures in this sector, including drug expenditures.
Drug expenditure represents the largest part of the private expenditures on health among the poor.The medicines' weight in the total health expenditures is inversely proportional to income, according to a literature review in 54 low-income countries. 16Similarly, medicines have had the largest share of responsibility for health expenditures in Brazil, reaching approximately 47% of the total expenditures. 3Income committed to drug expenditure, especially when it is not planned, can lead families to even greater health risks, like the reduction in the purchase of food products, and make them become more susceptible to health problems, or, by the economic order, make families become poorer or even indigent.b Brazilian population-based studies have shown that the mean expenditure on medicines in absolute fi gures is higher among older men 12 and increases as income increases. 3The SUS assumes the largest part of the global drug expenditures, mainly among the poor, while the rich present more private expenditures. 5In spite of this, data from Pesquisa Nacional por Amostra de Domicílios 1998 (National Household Sample Survey) presented more income committed to the purchase of medicines among the poor, transposing differences between groups and confi guring it as inequity.The aim of the present study was to describe socioeconomic inequalities concerning the use, expenditures and income committed to the purchase of medicines.The parameters to the fi nal sample were: prevalence of drug expenditures of 50%, sample error of four percentage points, and a 95% confi dence interval.A design effect of 2 was used, 10% were added for possible losses or refusals and 20% to control for confounding factors.This sample was enough to estimate the outcome of any magnitude considering the same sample error.Two-stage sampling was performed.The fi rst stage was formed by the census tracts defi ned by Instituto Brasileiro de Geografi a e Estatística (IBGE -Brazilian Institute of Geography and Statistics) for statistical purposes.The units of the second stage were the households included in the selected census tracts.The study's unit of analysis was the adult individual.All the adult inhabitants of the selected households were eligible for the interview, except for institutionalized individuals or those unable to answer the study's questionnaire (bedridden or with some physical or mental disability).

Cross
The mean income of the heads of households was observed in the 420 urban census tracts according to the 2000 Census and the values ranged from R$ 192.80 to R$ 13,209.50.The sectors were organized in an increasing income order and stratifi ed in deciles.Sixty census tracts were drawn, with systematic selection of six sectors in each income decile.With a minimum of 61 households and a maximum of 810, the coeffi cient of variation of occupied households across sectors was 55%.In order to reduce it, some census tracts were merged and others were divided; thus, 63 census tracts were obtained with a coeffi cient of variation of 32%.Eighteen households were systematically drawn in each of these geographical units.
Data collection was performed by 35 female interviewers.They had completed High School and were fully available to conduct the fi eldwork.The Personal Digital Assistant (PDA) was used to register and store data.Losses were considered adult inhabitants of drawn households that were not found after at least four visits, one on a weekend and another in the night period.
The quality control of the answers was performed on a weekly basis through revision and checking of 15% of the interviews, selected by means of a draw in each sector.Key information contained in the questionnaire was verifi ed by telephone by the fi eld supervisors.Analyzing the agreements, the kappa value varied from 0.6 to 1.0.
The outcomes were the use of medicines, expenditures and the income committed to the purchase of them during 30 days before the interview.The socioeconomic classifi cation was performed according to the interviewees' per capita family income (stratifi ed in quintiles).Information on self-reported skin color was also utilized (white, mixed-color, black, indigenous and yellow; the last two were excluded from this analysis due to the small number of people), as well as gender and age in complete years (20 to 29; 30 to 39; 40 to 49 and 50 to 59).
The question that guided the study was: "In the last 30 days did you take any medicine?".The names, dosages and pharmaceutical presentation of the medicines were asked, as well as the form of acquisition (SUS; commercial drugstore; the person already had it at home or was given by someone).The number of medicines used, the purchase of medicines because it was not possible to obtain them from the SUS, and their presence on RENAME and REMUME were analyzed.
To calculate the drug expenditures, information referring to the last 30 days and self-reported values described in reais (Brazilian currency) were used.The value of the income committed to the purchase of medicines was obtained by the division of the reported expenditure by the per capita family income.Then, the quotient was multiplied by 100 and the value was expressed in the form of percentage.For this calculation, 24 people who had no income were excluded.The variable income commitment was categorized according to cutoff points equal to ≥ 5%, ≥ 10% and ≥ 15%.

RESULTS
The response rate was 85.3% (n = 1,720).The mean age of the interviewees was 38 years (standarddeviation -SD = 11.6) and 55.5% were women.The prevalence of medicines use during 30 days before the interview was 76.5%, 29% higher among women (PR: 1.29; 95% CI: 1.21;1.39)and 7% higher in older adults when compared to younger adults (PR: 1.07; 95%CI: 1.00;1.15) in the adjusted analysis.No statistically signifi cant differences were observed in the use of medicines among the income and skin color strata (Table 1).Approximately one third of the adults who used medicines in the past 30 days used only one, approximately half used two, ¼ reported the use of three and a little less than 10% used four or more medicines.
The mean expenditure was R$ 46.70 (median = R$ 10.00), varying from zero to R$ 920.00.The spent values were 60% higher among the women and increased according to the individuals' income, approximately fi ve times higher among the richer ones (Q5) when compared to the poorer ones (Q1).Higher expenditures were observed among older people (approximately three times higher in the older individuals when compared to the younger ones).White individuals (70% more) and mixed-color individuals (50% more) presented higher expenditures on medicines than black individuals (Table 2).
The income committed to the purchase of medicines was almost three times higher among the poorer individuals (Q1) when compared to the richer ones (Q5).While in the fi rst income quintile people committed, on average, 7.3% of their income to the purchase of medicines, in the richest quintile the proportion was of 2.7% (Table 2).More income was committed among the women (90% more) and among the older individuals (60% more among individuals aged 50 -59 years compared to those aged 20 -29 years) (Table 2).
Figure 1 presents the prevalence of people who spent any value with the purchase of medicines and the proportion of people who committed values equal to or higher than 5%, 10% and 15% of their income to the purchase of medicines, according to the quintiles of per capita family income.Approximately 70% of the richer ones (Q5) committed some of their income to expenditures on medicines, a proportion that is signifi cantly higher than in the other groups (Figure 1A).However, the proportion of people who committed 5% or more of their income to expenditures on medicines was much higher among the poorer individuals (Q1) (Figure 1B).This difference was more signifi cant when higher income commitments were considered.While 6.5% of the richer individuals (Q5) committed 10.0% or more of their income to the purchase of medicines, among the poorer ones, 14.6% presented this situation (Figure 1C).When the focus was on commitment equal to or higher than 15%, 9.6% of the poorer individuals (Q1) were in this condition, compared to 3.1% of the richer individuals (Q5) (Figure 1D).4).

DISCUSSION
Women and older individuals use more, spend more and commit more income to drug expenditures compared to men and to younger people.Individuals who self-reported their skin color as black presented less expenditures on medicines than white individuals.An important socioeconomic inequality was observed in drug expenditures and in the income committed to the purchase of medicines, i.e., the poorer the individual, the higher the income committed, but the richer the individual, the higher the expenditures on medicines.
The prevalence of medicines use was higher than the ones observed in other national studies with adults, whose prevalences vary from 49.7% 2 to 69.9%, 17 and the same occurs in international studies (from 36.8% 13 to 67.4%). 11The differences in prevalences may derive from differences in the analyzed age groups and in the recall period.If, on the one hand, a longer recall time can increase the exposure to the use of medicines, on the other hand it can increase the memory bias, mainly in the population that does not use medicines in a continuous way. 6Furthermore, differences in the pharmaceutical policies adopted in different countries may infl uence the behavior of medicines use and may partly explain the distinct prevalences when we compare the national and international panoramas.
Higher utilization of medicines among women and older individuals corroborate the findings of the literature, 7,9,11 which also point to higher concern about health and higher utilization of the health services in the female sex. 10 It is expected that individuals are affected by a higher number of chronic diseases as age advances and, consequently, use more the health services.The combination of these factors directly influences the increase in the use of medicines. 8dditionally, the medicalization of health adopted by a large part of professionals and institutions, besides health actions and programs that are specifi c to women, like adolescents, pregnant women and puerperal women, contribute to explain this scenario. 14The expenditures mean was higher among women and older individuals, accompanying the greater use of these inputs in these groups, and white individuals spent more than black individuals, similarly as found in the United States, 20 but the association lost signifi cance after adjustment.
Drug expenditures were higher among the richer individuals, which suggests higher availability of fi nancial resources and also the possibility that the SUS has a more comprehensive coverage among the poorer individuals, potentially minimizing the absolute expenditures of this group.e The lower expenditures among the poorer individuals may be explained by the expansion of health programs in the SUS, like Farmácia Básica (Basic Drugstore), Estratégia da Saúde da Família (Family Health Strategy) and Farmácia Popular (Popular Drugstore), as well as the changes in prices of goods and services and the income of poor families, which may reduce or even eliminate the health expenditures to allocate them to other goods and services.c Besides, the higher absolute expenditures on the acquisition of these products among the greatest consumers, i.e., women and older individuals, is coherent, confi rming the fi ndings of the national literature. 1 The inequality in the income committed to the purchase of medicines was worse to the poorer individuals.When the committed income was analyzed in at least 10% and 15% of the total per capita income, the difference in the committed income was 2.2 times and 3.1 times higher in the poorer individuals when compared to the richer ones.This was also observed in the literature when the general health expenditures were considered, with higher income commitment among women, the poorer individuals and the older ones. 18e higher commitment of health expenditures can become catastrophic if the health expenditure  represents at least 40% of the family income after the payment of their subsistence expenditures. 20In a study conducted among 59 countries investigated in relation to catastrophic health expenditures, Brazil presented the second highest prevalence of families in this situation (10.3%), behind Vietnam (10.5%). 21Of the health expenditures, those directed at the purchase of medicines are among the most signifi cant. 3e results of the present study and the representation of medicines in the catastrophic expenditures show the importance of the State's protection against these expenses due to the implications deriving from this process, like, for example, the non-utilization or the underutilization of medicines, interfering in the possibility of adequate treatment, cure, aggravation or health complications.The State's role refers to the attempt to reduce the fi nancial obstacle, minimized by the availability or the reduction in medicines' prices.Also, there is the role of the prescriber, who should consider the user's capacity to pay and the availability of medicines in the public sector at the moment of the prescription. 13In addition, macroeconomic and social measures are fundamental, so as to impact the market of consumption and access to goods and services, like better income distribution, the access to public health services and general life conditions.
The studied sample presents external validity for the adult population residing in the urban area of Florianópolis.The uniform distribution of the losses in the family income deciles contributed to this condition of sample inference.The distribution by sex and age group was similar to the one found in the IBGE estimates for the adult population of the municipality for the year of 2009.
The potential memory bias involved in the self-reported data, like the use of medicines and the expenditures on the purchase of the products in the 30 days before the interview can be considered a limitation of the study.
To minimize it, the interviewees were asked to provide prescriptions, packages and/or the description leafl et of medicines utilized in the past 30 days.
Considering the high impact of the private drug expenditure on people's budget and the potential of minimization of this phenomenon through the SUS, the better utilization and organization of RENAME and REMUME might represent important mechanisms to reduce private expenditures regarding the public policies, as approximately 50% of the people do not use the medicines they need because they do not have access to them through the SUS.f Pharmaceutical policies that promote the rational use of the medicines should be implemented, improving health assistance and reducing drug expenditures.

Figure 1 .
Figure 1.Per capita family income committed to drug expenditures according to the population's per capita income quintiles: (A) Income commitment 0%; (B) Income commitment 5%; (C) Income commitment 10% and (D) Income commitment 15%.Florianópolis, SC, 2009. 15 -sectional population-based study with 2,016 adults aged 20-59 years from the urban area of Florianópolis (Southern Brazil), carried out between August 2009 and January 2010.The study was called EpiFloripa.Florianópolis is the capital city of the State of Santa Catarina.It is a medium-sized municipality, with population estimated at 408,163 inhabitants in 2009.c In the most recent national census (year 2000), Florianópolis presented per capita income of R$ 701.42 (equivalent to US$ 354.25),Gini index of 0.57 and municipal Human Development Index (MHDI) of 0.875, the highest one among the Brazilian capital cities. d

Table 1 .
Sample according to sociodemographic variables, prevalence of medicine use and crude and adjusted prevalence ratios.Florianópolis, Southern Brazil, 2009.The statistical analyses were performed using the Stata svy commands to correct the design effect of complex samples and incorporate the sample weights.It is available in the program Stata version 9.0.
aThe variable with the lowest n was skin color (n=1678) b Adjusted by socioeconomic and demographic variables c US$1 is equivalent to R$1.75The study was approved by the Research Ethics Committee of Universidade Federal de Santa Catarina (Process no.351/2008).

Table 2 .
Mean expenditure and proportion of income spent with medicines in the 30 days before the interview, according to income quintiles, age, gender and skin color.Florianópolis, Southern Brazil, 2009.

Table 3
presents differences related to the purchase of medicines in the SUS according to gender, age, skin color and per capita family income.Approximately one out of every fi ve people obtained at least one medicine in the SUS, a signifi cantly higher value among the poorer individuals (PR: 6.20 95%CI 3.50;10.97),women (PR: 1.56 95%CI 1.18;2.06),older individuals (PR: 1.76 95%CI 1.29;2.42)and the blacks (PR: 1.44

Table 3 .
Proportion of people who obtained medicines through the Sistema Único de Saúde (National Health System), who purchased at least one medicine not obtained through the SUS, who presented expenditures on medicines from Relação Nacional de Medicamentos Essenciais (National List of Essential Medicines)and from Relação Municipal de Medicamentos Essenciais(Municipal List of Essential Medicines),according to socioeconomic and demographic variables.Florianópolis, Southern Brazil, 2009.
a US$1 is equivalent to R$1.75 SUS: Sistema Único de Saúde (National Health System) RENAME: Relação Nacional de Medicamentos Essenciais (National List of Essential Medicines) REMUME: Relação Municipal de Medicamentos Essenciais (Municipal List of Essential Medicines)

Table 4 .
Crude and adjusted prevalence ratios of people who obtained medicines through the Sistema Único de Saúde (National Health System), who purchased at least one medicine not obtained through the SUS, who presented expenditures on medicines from Relação Nacional de Medicamentos Essenciais (National List of Essential Medicines)and from Relação Municipal de Medicamentos Essenciais (Municipal List of Essential Medicines),according to socioeconomic and demographic variables.Florianópolis, Southern Brazil, 2009.