Tuberculosis in Salvador , Brazil : costs to health system and families

Correspondence to: Maurício L Barreto Instituto de Saúde Coletiva Universidade Federal da Bahia Rua Padre Feijó, 29, Canela 40110-170 Salvador, BA, Brasil E-mail: mauricio@ufba.br Project funded by the National Health Foundation/Brazilian Ministry of Health and by the Department for International Development of the United Kingdom. Received on 16/5/2003. Reviewed on 22/4/2004. Approved on 1/6//2004.


INTRODUCTION
The report of the Commission on Macroeconomics and Health of the World Health Organization 7 has emphasized that tuberculosis is the most common of the infectious diseases. Approximately onethird of the world's population is infected with Mycobacterium tuberculosis and around 5% to 10% of infected individuals will become ill and infective at some time during their lives. The same report has estimated that, if not controlled, tuberculosis will be responsible for the death of 35 million people over the next 20 years. Brazil ranks tenth in the world in numbers of cases, with around 83,000 cases notified every year, in all age groups but especially among those aged over twenty years old. In the State of Bahia, around 7,800 cases occur every year, which corresponded to an incidence of 60.3/100,000 in-*Details of the methodology utilized for collecting information on the patients can be found at the web site: http://www.isc.ufba.br/ programas/edidn.asp, or by means of contacting the authors, by e-mail: mauricio@ufba.br *Details of the methodology utilized for collecting information on the patients, specifically in Tables 1 to 3, can be found at the web site: http://www.isc.ufba.br/programas/edidn.asp, or by means of contacting the authors, by e-mail: mauricio@ufba.br jective of the questionnaire was to obtain information on the costs that they had had prior to and subsequent to the start of tuberculosis treatment. The new patients were recruited during the months of February and March 2000. For the hospitalized patients, the objective was to verify the costs during the six months preceding the date of hospitalization and the costs during their hospitalization. The hospitalized and multiresistant patients were recruited during the months of July and August 2000. For the multiresistant patients, the objective was to estimate the costs over a one-year period prior to the date of the interview.
For the new outpatients, the questionnaires were applied on three occasions: at the start of their treatment (to register their costs prior to the diagnosis) and three and six months later. This last application coincided with the end of the treatment. For the multiresistant and the hospitalized patients, the questionnaire was applied on a single occasion (the information collected from the patients is shown in Table 1 published on ISC's website.* The data utilized for estimating the costs of the public and private healthcare sectors was collected for the year 1999, from the central administrations of the Secretaria de Saúde do Estado da Bahia (Health Department of the State of Bahia -SESAB) and the Secretaria Municipal de Saúde de Salvador (Municipal Health Department of Salvador -SMS), and from healthcare centers and one private philanthropic entity that were treating tuberculosis in Salvador. For this collection, six healthcare centers from among the twelve centers where tuberculosis control activities were being performed in the municipality of Salvador were selected. This selection was based on the availability of the information needed for calculating the costs. Thus, the centers that had all the information needed for the cost calculations were selected.
The files of the new outpatients and multiresistant patients were also reviewed in order to determine the average number of consultations and tests that these patients did. The files of the hospitalized patients were not reviewed. The costs of the epidemiological surveillance of tuberculosis were not estimated.
For analysis, the procedures of cost accounting were utilized, in which the proportion of the resources that is utilized for each activity is considered in order to habitants in 1999. Of these, approximately 3,000 cases of tuberculosis occur in the city of Salvador every year, with an incidence of 131.2/100,000 inhabitants in 1999. 11 In addition to the direct suffering caused by the disease, tuberculosis has been taking up significant proportions of public budgets in developing countries. It is estimated that, by 2015, investments of the order of US$12 billion will be needed for controlling diseases like Aids, tuberculosis and malaria. 7 The increased costs involved in attending to and controlling tuberculosis are also due to the ever-growing numbers of cases that are resistant to different types of chemotherapy. As well as the high cost to the healthcare system, tuberculosis puts a burden on family budgets, thus worsening poverty conditions. This results not only from the direct costs related to the treatments, but especially from the income losses caused by the disease. 1,2,6,7,9,10,12,13 In Brazil, the economic effects of tuberculosis are not well known. The present study had the objective of estimating the economic costs of tuberculosis for the healthcare system and for the families of patients undergoing treatment for tuberculosis. It also aimed to contribute towards the development of similar studies in other contexts, through laying out in detail the methodology utilized.

METHODS
To study the cost of tuberculosis for patients and their families, the sample size was defined by convenience, to include 146 patients who had received a diagnosis of tuberculosis for the first time, 33 hospitalized patients and 39 patients with multiresistant tuberculosis. The new cases were recruited from among the recently diagnosed cases of tuberculosis at ten healthcare centers in Salvador, State of Bahia. These ten centers attend to around 90% of the tuberculosis cases in the city. The hospitalized patients were recruited from among the patients in the reference hospital for tuberculosis treatment in the city, and the multiresistant patients were recruited from the outpatient service of this same hospital.
The information was collected using a standardized questionnaire that was administered by nurses, nursing auxiliaries (or technicians) and social workers who had been trained for this purpose. For the new patients beginning outpatient treatment, the ob-determine the unit cost of this activity. 4 For each segment studied, specific indicators were utilized.*

Cost to the families
Of the 146 new cases that were initially included in the sample, four were lost. Table 1 presents the total costs to the families (US$1,145,686), for each cost item and patient group. The average cost to the family varied according to the type of treatment needed. For the family of a patient who was diagnosed with tuberculosis for the first time and was treated as a new outpatient (six months of treatment) the cost was US$266. For the family of a patient with multiresistant tuberculosis (one year of treatment), the estimated cost was US$333 and, for the family of a hospitalized patient (average of 18 days), the cost was US$96.
The average cost to a family for a hospitalized patient was calculated only for the number of days of hospitalization. When this patient left hospital and proceeded with outpatient treatment, it was assumed that the family would have an increase in its costs, which could be considered to be close to the estimate for outpatient treatment (US$266).

Cost to public services
The outlays on attending to and controlling tuberculosis within the public sector, for each health facility, are presented in Table 2. This shows the cost estimates for new cases according to the healthcare center. The two most relevant cost items for healthcare units were the costs of the professionals exclusively within the program and the cost of medications. The variations in the cost of the treatment that were observed in the six centers studied result from the fact that some centers have professionals who are not available in others (for example, nutritionists), and from the variation in the productivity of the centers. Some centers function as referral centers for large areas, thus attracting greater numbers of patients. Consequently, these centers present lower average costs than other centers at which smaller numbers of patients are treated. In this respect, as can be seen in Table 2, the second healthcare center attended to a   Table 3 shows that the average estimated cost for the treatment of one new case of tuberculosis in Salvador was US$103. For chemoprophylactic treatment, it was US$ 73, and for retreatment, it was US$115. Table 3 also presents the estimated costs of applying a dose of BCG vaccine (US$1.40) and searching for defaulters (US$12) and the average cost of hospitalization (US$1,468).
The average cost of hospitalization was estimated only for the public reference hospital for hospitalization of patients with tuberculosis. The patients with a diagnosis of multiresistant tuberculosis were also treated at this hospital. The estimated average cost for treating one of the multiresistant patients, for a oneyear course of treatment, was US$2,744. Around 95% of this amount was due to the expenditure on medications alone, paid for by the public healthcare system. Table 3 also shows that, out of the total public ex-penditure in Salvador for tuberculosis control, approximately 65% was incurred on hospitalization, 32% on outpatient treatment (of new and multiresistant cases) and just 3% on prevention.

Cost to the private sector
The cost of tuberculosis to the private sector is shown in Table 4. The available costs are always greater in the private sector than in the public sector, for every activity. This is mainly due to the fact that the salaries paid to healthcare professionals are higher in the private sector than in the public sector. Table 5 shows the consolidation of costs related to tuberculosis for the public and private sectors and for families. The greatest proportion of the costs is borne by the public sector (62%), followed by the families (33%) and the philanthropic entity (5%).

DISCUSSION
The results from the present study can be discussed in relation to three principal findings:

-The families bore 33% of the total costs of tuberculosis in the city of Salvador, in 1999
The main criticisms directed towards studies in which the sample is defined by convenience are re-  lated to possible bias in the data and the consequent limitations on the inferences that can be drawn. Such bias may, for example, be connected with the fact that the patients in the sample might come from socioeconomic strata that differ from those that would characterize the population affected by tuberculosis. However, around 90% of the patients with tuberculosis who were registered in Salvador were being attended at the ten healthcare centers from which the sample of new outpatients was extracted. Also, the patients selected for the study were chosen randomly from the two shifts in which attendance was provided. These two facts are evidence against any possible influence of such bias.
Among the patients in the sample studied, around 21% originated from the health district of the railway suburb of Salvador, and around 80% of them had low income levels (up to two minimum salaries) and low educational level (up to four years of schooling): data not tabulated. In a study on the spatial distribution of mortality due to tuberculosis in Salvador, Mota et al 8 showed that the greatest risk of death due to tuberculosis was also observed in this health district, where the population is characterized by the lowest income and education levels in the city of Salvador.
The data from the national home sampling survey (Pesquisa Nacional por Amostra de Domicílios -PNAD) in 1999 5 showed that 40% of the urban population of northeastern Brazil was in the family income band of up to two minimum monthly salaries, corresponding to an average annual family income of the order of US$994. Considering that a large proportion of tuberculosis cases come from the population within this income band, and that the family expenditure on a new case of tuberculosis that is treated as an outpatient (estimated as US$266), it can be seen that one tuberculosis case committed around 27% of the average family income. This means that more than one-quarter of the income of poor families in the city of Salvador was committed in the event that one family member was affected by tuberculosis. The loss of income due to temporary incapacitation for work was what contributed most to the cost incurred by the families. Among the patients studied, around 80% depended on their health conditions in order to maintain their work and daily earnings, since their occupations were as masons, wall painters, and domestic employees, among others, relying on precarious contracts and without social security protection. For the multiresistant patients, the item of loss of income related to periods prior to the diagnosis. The majority of the multiresistant patients had not done any productive activity for a long time, nor did they have any prospects of performing any such activity.
The findings were similar to others found in the international literature. Similar percentages of loss of income due to temporary incapacitation for work (53.5% of the costs for the families of new cases and 67.5% for the families of hospitalized patients) were found in India. 10 In Thailand, it was found that 20% of family income was committed because of tuberculosis: 15% on the cost of medications and consultations and 5% on loss of income. 6 Memory bias, which is inherent to all studies that seek recollection of information from the past through questionnaires or interviews, must also be considered as a limitation on the results found.

-The public sector bore 62% of the total costs of tuberculosis in Salvador and accomplished few activities for prevention of this disease
The difficulty in accessing information on the expenses of all of the twelve units that treated tuberculosis in Salvador, in 1999, meant that these costs were only estimated for six healthcare centers for which the information was available. In addition to this, the expenses relating to epidemiological surveillance and educational campaigns were not calculated in the cost analysis. For the hospital cost, the decision to base the estimate on a single hospital was due to the fact that, as well as being the reference hospital for the treatment of tuberculosis, around 92% of all hospitalizations due to tuberculosis in the municipality were taking place at this hospital.
Despite the restrictions on the estimates of the public costs described here, it is possible that they may be close to the real values. The methodology utilized very closely reflects the real average costs for the public sector relating to the treatment of tuberculosis.
The present findings indicate that, out of the total public expenditure on tuberculosis, around 65% was directed towards hospitalization. Data from the Department of Information and Information Technology of the Brazilian national health system (Datasus) have revealed that, in 1999, around 12% of the patients hospitalized with a diagnosis of tuberculosis first obtained this diagnosis in the hospital, thus beginning their treatment.