Prevalence and Pattern of Epilepsy Treatment in Different Socioeconomic Classes in Brazil


Address correspondence to Li Min Li, Department of Neurology, Faculty of Medicine, UNICAMP, 3083–970, Campinas, SP, Brazil. E-mail: or


Summary: Purpose: The worldwide prevalence of epilepsy is variable, estimated at 10//1,000 people, and access to treatment is also variable. Many people go untreated, particularly in resource-poor countries.

 Objective: To estimate the prevalence of epilepsy and the proportion of people not receiving adequate treatment in different socioeconomic classes in Brazil, a resource-poor country.

Methods: A door-to-door survey was conducted to assess the prevalence and treatment gap of epilepsy in three areas of two towns in Southeast Brazil with a total population of 96,300 people. A validated screening questionnaire for epilepsy (sensitivity 95.8%, specificity 97.8%) was used. A neurologist further ascertained positive cases. A validated instrument for socioeconomic classification was used.

Results: Lifetime prevalence was 9.2/1,000 people [95% CI 8.4–10.0] and the prevalence of active epilepsy was 5.4/1,000 people. This was higher in the more deprived social classes (7.5/1,000 compared with 1.6/1,000 in the less deprived). Prevalence was also higher in elderly people (8.5/1,000). Thirty-eight percent of patients with active epilepsy had inadequate treatment (19% on no medication); the figures were similar in the different socioeconomic groups.

Conclusion: The prevalence of epilepsy in Brazil is similar to other resource-poor countries, and the treatment gap is high. Epilepsy is more prevalent among less wealthy people and in elderly people. There is an urgent need for education in Brazil to inform people that epilepsy is a treatable, as well as preventable, condition.

Epilepsy is the most common serious chronic non-infective neurological condition in the world (Fernandes and Sander, 1988). Studies from Brazil have reported lifetime prevalence of epilepsy ranging from 11.9/1,000 to 21/1,000 (Fernandes and Sander, 1988; Li and Sander, 2003). Studies carried out in other Latin American countries (for review, see Burneo et al., 2005) revealed similar figures; the prevalence of epilepsy ranged from 12.2 to 19.5/1,000 in Ecuador (Placencia et al., 1992), and was 12.3/1,000 in Bolivia (Nicoletti et al., 2005). Similar prevalence figures (for review see Preux and Druet-Cabanac, 2005) are also observed in other continents in countries with limited resources such as Senegal with a lifetime prevalence of 14.2/1,000 (Ndoye et al., 2005), and China with a lifetime prevalence of 7/1,000 inhabitants (Wang et al., 2003).

The treatment gap, defined as the percentage of people with epilepsy who need treatment but do not receive it (Kale, 2002), is high in low-income countries, with variations among countries (Meinardi et al., 2001; Ndoye et al., 2005). For instance, it was estimated that 53% of patients with active epilepsy in Honduras were not receiving medical treatment (Medina et al., 2005). A survey in Brazil using indirect measurements based on antiepileptic drug distribution estimated that the treatment gap was around 50% (Noronha et al., 2004). In a study in China, it was estimated that 63% of six million people with active epilepsy were not receiving treatment in the week before the survey (Wang et al., 2003).

The present study is the initial phase of the National Demonstration Project on Epilepsy in Brazil (Li and Sander, 2003), part of the Global Campaign Against Epilepsy (Reynolds, 2000), an initiative of the Pan American Health Organization/World Health Organization, the International Bureau for Epilepsy and the International League Against Epilepsy. The objectives of this study are (a) to estimate the prevalence of epilepsy and the proportion of people not receiving adequate treatment for their condition and (b) to ascertain whether socioeconomic class is a determinant of the prevalence and pattern of treatment of epilepsy.


We conducted a door-to-door survey in three areas: Barão Geraldo in the municipality of Campinas, and Jaguaré and Santo Antônio in the municipality of São José do Rio Preto (SJRP), both in the southeastern Brazilian state of São Paulo, to evaluate the prevalence and the treatment gap of epilepsy.

Study area and population

These areas were chosen because of cooperation between the health centers in each area and the team conducting the demonstration project (DP), and also because a variety of socioeconomic classes were included. Barão Geraldo has a population of 44,000 people, Jaguaré 23,300 people and Santo Antônio 29,000 people.

Clinical and epidemiological definitions

Epilepsy classification

We classified the epilepsy and types of medical treatment according to the following definitions:

  • 1Active epilepsy: any patient who has had recurrent unprovoked seizures with an interval between them of 24 h or more in the previous 24 months.
  • 2Inactive epilepsy: any patient who has had recurrent unprovoked seizures with an interval between them of 24 h or more, but who has been seizure-free for the previous 24 months.
  • 3Adequate epilepsy treatment: any patient with active epilepsy regularly using appropriate antiepileptic drugs (AEDs) as mono- or polytherapy at standard dosage is defined as having adequate epilepsy treatment (Camfield and Camfield, 2000; Betting et al., 2003).
  • 4Monotherapy: any patient who is taking just one AED is defined as taking monotherapy.
  • 5Polytherapy: any patient who is taking more than one AED is defined as taking polytherapy.
  • 6Nontreated epilepsy: any patient with active epilepsy who did not take any AED during the week prior to prevalence day is defined as having nontreated epilepsy.

Assessment of socioeconomic status

The patients were classified into socioeconomic classes, using a validated questionnaire based on household assets (Periscinoto, 1994). This classifies people into seven classes in decreasing order of wealth: A1, A2, B1, B2, C, D, and E. For the purpose of this study, these were grouped into four socioeconomic classes: A (A1 + A2), B (B1 + B2), C, and D (D+E) (see box 1) (Periscinoto, 1994).


Step 1—A door-to-door survey was carried out between September and December 2002. For this survey, conducted in collaboration with INTEP (Instituto Tecnológico de Estudos e Pesquisas, SJRP), a third-party company specializing in public surveys, we used a validated epidemiological questionnaire (with sensitivity 95.8% and specificity 97.8%) for the identification of cases of epilepsy (Borges et al., 2004). If no response was obtained on the first visit to a house, two further visits were made and a letter left before considering this a nonresponse.

Step 2—Those who screened positive in step 1 were invited for a consultation with a neurologist with an interest in epilepsy (LHNM, MAB, ALAN, and LML). At this consultation the diagnosis of epilepsy was made or refuted, based on clinical history. People not diagnosed as having epilepsy were considered false positive and a presumptive diagnosis was made. Treatment status and seizure activity were also established in those confirmed as positive.

Plan of statistical analysis


  • 1Prevalence is defined as the number of people with a history of active or inactive epilepsy per 1,000 people in the population. Prevalence was estimated on the basis of the number of true positive cases divided by total number of people studied expressed as n/1,000 (with a 95% confidence interval estimated using Wilson's method).
  • 2The treatment gap is the number of patients with active epilepsy not on treatment or on inadequate treatment, expressed as a percentage of the total number of people with active epilepsy (Kale, 2002). The treatment gap estimation was based on AED intake in the week prior to the survey as established in step 2.

Prevalence was standardized for age within each social class using a standard population from the same area as the sample population derived from the national population census figures for the year 2000 (IBGE, Censo 2000). This standardisation was carried out to compare different social classes and to eliminate any age bias in the estimations.

The prevalence of epilepsy and the percentage of people with active epilepsy and inactive epilepsy for each social class were then extrapolated to the Brazilian population.

Data sets used for the calculation of prevalence

Prevalence were calculated for two different data sets (A and B):

  • 1Minimum data set. (data set A): This comprises all cases identified by neurologists in step 2.
  • 2Maximum data set. (data set B): This comprises all cases from the minimum data set and an additional estimated number of cases, assumed to have been missed on screening. This was estimated by applying the false negative rate (1 – sensitivity) from the validation of the screening procedure to those who screened positive.


The total population for the study area is approximately 96,300.

Step 1—A total of 54,102 people (50% women) were surveyed in step 1 (Barão Geraldo [Campinas] 21,535, Jaguaré[SJRP] 14,412, Santo Antônio [SJRP] 18,155). The mean age of the population was 38 years (range 0–98 years, standard deviation 25.6). 1,657 people screened positive.

Step 2—Neurologists assessed the diagnosis in 1,657 people, identifying 1,161 as false positive and 496 as true positive (Fig. 1). Therefore, over the three areas, epilepsy was confirmed in 496 (0.9%) people (254 women) with a mean age of 36 years (range 0–90 years, standard deviation 19.1). Two hundred ninety (58.5%) had active epilepsy, 176 (35.5%) had inactive epilepsy and 30 (6.0%) were uncertain when their last seizure occurred.

Figure 1.

Screening survey about epilepsy conducted in Campinas and São José do Rio Preto in 2002 and 2003.

Application of the sensitivity of the screening test (95.83%) to the number who screened positive estimates that 22 people with epilepsy will have been missed by the screening test.


  • •. The minimum lifetime prevalence estimated is 9.2/1,000 people [CI 95% 8.4–10.0] (496/54,102 data set A).
  • •. The maximum lifetime prevalence estimated is 9.6/1,000 people [CI 95% 8.8–10.4] ([496+22]/54,102 data set B).
  • •. The minimum prevalence of active epilepsy is 5.4/1,000 people [CI 95% 4.8–6.0] (290/54,102 data set A).
  • •. The maximum prevalence of active epilepsy is 5.6/1,000 people [CI 95% 5.0–6.3] ([290+12.62]/54,102 data set B).
  • •. In the population of Campinas and São José do Rio Preto the prevalence of both active and inactive epilepsy was highest in those aged 60 years and older (Table 1).Of those with a confirmed diagnosis of epilepsy, 489 (98.6%) completed a questionnaire about socioeconomic classification (173 in Campinas and 316 in SJRP).
Table 1. Epilepsy prevalence by age
Age group (years)Campinas and São José do Rio PretoPercentage Brazilian populationb
n total%Lifetime epilepsyaActive epilepsyInactive epilepsy
nPrevalence (/1,000) [95% CI]nPrevalence (/1,000) [95% CI]nPrevalence (/1,000) [95% CI]n total%
  1. aThe lifetime prevalence includes people with epilepsy who did not know when their last seizure occurred, plus people with active epilepsy and people with inactive epilepsy.

  2. bFrom: IBGE, 2000.

0–4 3,534 6.5 102.8 [1.5–5.2]  82.3 [1.2–4.5]  2  0.6 [0.16–2.1] 16,375,728 9.6
5–9 3,762 7.0 28  7.4 [5.2–10.7]  164.3 [2.6–6.9] 102.7 [1.4–4.9]16,542,327 9.7
10–19 9,72218.0 687.0 [5.5–8.9] 373.8 [2.8–5.2] 293.0 [2.1–4.3]35,287,88220.8
20–5932,28359.7328  10.2 [9.1–11.3]  1885.8 [5.1–6.7]1173.6 [3.0–4.3]87,057,20451.3
60 or more  4801 8.9 62    12.9 [10.1–16.5]    41  8.5 [6.3–11.6]  183.7 [2.4–5.9]14,536,029 8.6
Total54,102100   496 9.2 [8.4–10] 2905.4 [4.8–6.0]1763.3 [2.8–3.8]169,799,170  100  

Prevalence values according to social class are showed in Table 2. To be able to ascertain any true difference according to social class, we carried out a direct standardization of the sample using the total population of the area as the standard. In general terms, no differences were found between the standardized and nonstandardized estimations.

Table 2. Epilepsy in Campinas and São José do Rio Preto according to social classification
Social classesn totalLifetime epilepsyaActive epilepsyInactive epilepsy
  1. aThe lifetime prevalence includes people with epilepsy who did not know when their last seizure occurred, plus people with active epilepsy and people with inactive epilepsy.

  2. The figures shown in brackets for the lifetime, active and inactive epilepsy prevalences are standardized for age within each socioeconomic class. For each category of epilepsy status the standardized prevalence for socioeconomic classes B to D+E was compared with the standardized prevalence for social class A. The results of this comparison were significant (p < 0.05).

A 4,304 184.2 (4.2)2.7–6.6  71.6 (1.7)0.8–3.4  92.1 (1.9)1.1–4.0
B13,2891017.6 (7.3)6.3–9.2 413.1 (2.9)2.3–4.2 443.3 (3.1)2.5–4.4
C22,8582239.8 (10.3)8.6–11.11436.3 (6.6)5.3–7.4 763.3 (3.4)2.7–4.2
D+E13,01315111.6 (10)  9.9–13.6 977.5 (8.7)6.1–9.1 463.5 (2.0)2.7–4.7
Not classified   638 34.71.6–13.7  23.1  11.6

Thirty-eight percent of patients with active epilepsy had inadequate treatment: 12% were on inappropriate medication or suboptimal dose, 19% were not on any medication and in 7% the medication was unknown. None of the patients classified as active epilepsy were considered having inactive epilepsy undergoing medication withdrawal or no-medication. The figures were similar in the different socioeconomic classes (Table 3). There was no clear pattern in the treatment gap over the different age groups studied, although the extremes of life tended to be less adequately treated (Table 4). Although most patients were on a standard dosage of AED treatment, many were at the lower limit of the range, and could potentially have benefited from an increase in dosage. The most common reason for a total lack of treatment was that patients did not want to take it (51%) (Table 5).

Table 3. Treatment of active epilepsy according to social classes in Campinas and São José do Rio Preto
Social classesAdequate treatmentInadequate treatmentActive epilepsy (Total number)Treatment gap (%[95%CI])
MonotherapyPolytherapyInadequate dosage (%[95% CI])Nontreated (%[95%CI])Unknown (%[95%CI])
A  2 21 (14.3 [2.6–51.3])1 (14.3 [2.6–51.3])1 (14.3 [2.6–51.3])  73 (42.9 [16–75]) 
B 21 91 (2.4 [0.4–12.6]) 5 (12.2 [5.3–25.5])5 (12.2 [5.3–25.5]) 4111 (26.8 [16–42])  
C 503222 (15.4 [10.4–22.2])31 (21.7 [15.7–29.1])8 (5.6 [2.9–10.7]) 14361 (42.7 [35–51])  
D+E 461711 (11.3 [6.5–19.2]) 18 (18.6 [12.1–27.4])5 (5.2 [2.2–11.5])  9734 (35.1 [26–45])  
Not classified  2  2
Total 1216035 (12.1 [8.8–16.3])  55 (19.0 [14.9–23.9])19 (6.6 [4.2–10])  290109 (37.6 [32.2–43.3])
Table 4. Treatment gap—distribution by age in Campinas and São José do Rio Preto
Age group (years)Campinas and São José do Rio Preto
Active epilepsy (n)Inadequately treated (n)%95% CI
0–4  8  562.530.6–86.3
5–9 16  743.823.1–66.8
10–19 37 1129.717.5–45.8
20–59188 6534.628.1–41.6
60 or more 41 2151.236.5–65.7
Table 5. Reasons given for being on no treatment by people with active epilepsy
Reasons for not using medicineCampinas and São José do Rio Preto  
n%95% CI
Side effects23.61.0–12.3
Do not know about treatment712.76.3–24.0
Do not want treatment2850.938.1–63.6
Medical orientation112011.6–32.4
Never looked for treatment712.76.3–24.0


This study is the first door-to-door epidemiological survey of epilepsy, the treatment gap and the socioeconomic influence on epilepsy in a general Brazilian population. The prevalence of epilepsy in Brazil is similar to other resource-poor countries (Farmer et al., 1992; Kale, 2002; Placencia et al., 1992; Li and Sander, 2003; Nicoletti et al., 2005; Ndoye et al., 2005), and the treatment gap is high (Meinardi et al., 2001; Kale, 2002; Wang et al., 2003; Noronha et al., 2004; Ndoye et al., 2005). Even in potentially economically active patients aged between 20 and 59 years, over one-third of those with active epilepsy were not on adequate treatment. The data add to the understanding of the treatment gap in epilepsy worldwide. A study in China also showed a high treatment gap (Wang et al., 2003).

Extrapolation of our findings to the Brazilian population (170 million) suggests that about one million people are affected by epilepsy of whom over 380,000 will not be on adequate treatment. Although the prevalence of the treatment gap showed no trend over the different socioeconomic classes, the ratio of active to inactive epilepsy was different, with the prevalence of active epilepsy higher than that of inactive in the less wealthy groups. In this study, socioeconomic status appeared to be associated with active epilepsy, as we found a higher prevalence in poorer population strata. This finding is in part similar to a study carried out in Iceland (Hesdorffer et al., 2005), which showed that socioeconomic status is a risk factor for epilepsy in adults, but not in children. A UK study found the incidence of epilepsy to be higher in more socioeconomically deprived populations (Heaney et al., 2002). The higher prevalence of active epilepsy in more socioeconomically deprived groups found in the current study cannot be explained by the treatment gap, which shows no similar trend. One possible explanation could be different causes of epilepsy; poor people may be exposed to causes that lead to more severe epilepsy. Further studies are required to assess this issue.

The reasons for the treatment gap may be multifactorial (Farmer et al., 1992), ranging from logistic aspects of health care delivery to ignorance of existence of medical treatment. In this study, the treatment gap was defined as no treatment or inadequate treatment. The main reason given by people with active epilepsy not on treatment was no desire for treatment. A further one-quarter either never sought medical treatment or did not know about the existence of medical treatment for the condition. It would be interesting to understand how much of the lack of desire for treatment is related to ignorance of effective treatment.

Campinas and São José do Rio Preto are located in one of the wealthier regions of Brazil, where good public and private health care systems are available. Medication is free, which is not the case in most parts of the country. Therefore, our findings are likely to represent the best scenario in the spectrum of epilepsy management in Brazil. The fact that the treatment gap was fairly constant over the different socioeconomic groups could suggest that the current health services do not discriminate against the poorer groups but other reasons may exist.

A commitment from the government at all levels (federal, state, and local) is required to make epilepsy a priority in the health agenda (Li et al., 2005). The results of this study suggest that the aim should be to improve the overall health management of people with epilepsy, and to provide a continuing supply of first line AEDs. They also suggest that a campaign to inform the population that epilepsy is a treatable condition should be implemented. These measures should reduce significantly the percentage of people with epilepsy on inadequate treatment. The measures might also be expected to reduce the prejudice and stigma in epilepsy and to improve the quality of life of people with epilepsy and their families.


Acknowledgments:  We are indebted to the staff of the health units of Barão Geraldo, Jaguaré, and Santo Antônio, in particular Maria Inez Brambilla, Patrícia Castro, Luciana Machado, Regina Célia Nogueira Gomes, Vera Lúcia B. Reis Prietto, Elias Ferreira do Carmo, Maria Marlete Bastos do Nascimento, Maria Auxiliadora Rodrigues de Oliveira, for their support and help in carrying out this study.

We thank Antônio Carlos Carvalho from INTEP who helped coordinate the screening survey of this study.

This study was funded by WHO and PAHO and in part by FAPESP 02//11871-8 and UNICAMP.

Box 1. Determination of Social Classes

Index of economic classification related to power of consumption


Social classes in Brazil are calculated largely from the presence in the household of a number of items (such as television, washing machine and freezer), or one or more cars, and the employment of a servant (up to 29 points). A further five points are based on the education of the head of the household.

Table 6. Classification of the social classes in Brazil
ClassesPointsPercentage in Brazil (%)
A130–34 1.4
A225–29 3.3
B121–24 7.7
D 6–1042.7
E 0–521.1