Health care utilization and outpatient, out-of-pocket costs for active convulsive epilepsy in rural northeastern South Africa: a cross-sectional Survey

Health care utilization is an important component of both costs and adherence, with the relationship between these factors often bidirectional and complex [10, 11]. Studies from high-income countries show that people with epilepsy utilize health care more often than people without epilepsy due to repeat drug prescriptions [12] and increased emergency room usage [13].

Existing studies in LMICs find a large epilepsy treatment gap - defined as those not on treatment or on inadequate treatment [14]. An analysis of 27 studies from LMICs found an overall treatment gap of 56 % [15]. This suggests that people do not utilize (or receive) the health care they need. Cost of treatment, unavailability of drugs, and inadequate healthcare manpower all contributed significantly to the observed treatment gap [15]. By understanding costs and health care utilization for people with epilepsy, interventions to remove barriers and improve access to care can be developed. In South Africa, where standard AEDs are provided free of charge, understanding drivers of out-of-pocket costs and time spent seeking health care may provide insights on how to reduce the treatment gap.

In the current study, we determined the health care utilization of people with active convulsive epilepsy, estimated the time spent seeking outpatient treatment and calculated direct out-of-pocket costs in a rural South African context. This is the first study of its kind from South Africa.

The Agincourt Health and Socio-Demographic Surveillance System (HDSS) site, located in the Agincourt health sub-district of the Bushbuckridge municipality, comprises 420 km² of semi-arid, rural scrubland, 500 km northeast of Johannesburg, South Africa in Mpumalanga Province adjacent to Mozambique. Established as a research site in 1992, the Agincourt HDSS (http://​www.​agincourt.​co.​za) aims to inform health and development policy through evidence-based research and has generated socio-demographic data since its inception through annual census updates that capture vital statistics, including births, deaths, and in- and out-migrations.

The 2011 mid-year population was roughly 90,000 individuals in 16,000 households and 27 research villages [16]. The majority of the population speaks Xitsonga with one-third of the population of Mozambican origin.

Unemployment remains high in Agincourt with 36 % of the adult population (older than 18 years old) unemployed and currently looking for work [17]. The socio-economic status of the site has improved over the last 20 years, though as of 2011, 20 % of the population was still living in extreme poverty (personal communication, Mark Collinson). Government grants provide much needed financial support, with the majority of Agincourt households receiving at least one (personal communication, Mark Collinson).

Within the Agincourt site, a network of six, government primary health care (PHC) clinics offers free basic, outpatient health services during regular business hours. These services include mother and child care, including immunizations, family planning, testing and treatment for sexually transmitted infections, including HIV, minor trauma and routine care for those with chronic conditions, including epilepsy [18].

The six clinics refer patients to one larger government community health centre within the site. The health centre provides PHC services as well as 24-h acute care and facilities for 48-h patient observation [18]. Nurses run both clinics and health centres, though doctors do visit clinics and health centres.

A public-private partnership community health centre within the site provides HIV, tuberculosis and other chronic disease care and treatment. Clinicians staff this facility.

Referrals from primary health care facilities in the Agincourt research site are made to three government hospitals located 25-55 km from the site. Clinicians at the hospitals primarily diagnose epilepsy and initiate therapy before referring patients to the clinics or health centre for monthly AED collection [19]. People with epilepsy are expected to return to the hospital at 6-month intervals for review of seizure frequency and treatment.

A 2008 cross-sectional survey sought to identify all individuals with active convulsive epilepsy (ACE) living within the Agincourt HDSS [20]. ACE was defined as having ≥2 unprovoked convulsive seizures occurring more than 24 h apart, with at least one seizure occurring in the year preceding the study or currently taking AEDs for epilepsy [20]. After undergoing a clinical history and neurological exam, 311 individuals living within the Agincourt HDSS were diagnosed with ACE by the study neurologist (CRN), yielding an adjusted ACE prevalence of 7.0 per 1000 individuals [20].

Between September 2010 and January 2011, trained fieldworkers attempted to visit each individual diagnosed with ACE and/or his/her caregiver. If consent was given, the respondent was asked questions regarding health care utilization over the previous 12-months. Questions on out-of-pocket costs associated with utilization of outpatient care for the treatment of ACE were asked regarding the last visit to the biomedical health care facility. The questionnaire was forward- and back- translated into the local language and piloted with local community members. The English version of the questionnaire is available as an Additional file (Additional file 1).

The annual cost of outpatient, out-of-pocket care was calculated by multiplying the average cost per visit by the average number of visits reported in the last 12-months.

The time traveling to a traditional healer was not collected in this study and only traditional healer fees are reported as results.

Two hundred thirty-nine (96 %) individuals reported receiving non-epilepsy related health care when they last needed it, which for 154 individuals (62 %) was in the previous 30-days. For 231 (91 %) individuals, it was in the last year. Four individuals (1 %) reported being denied care and one participant cited inadequate levels of biomedical caregiver’s skill as reason for not receiving the necessary health care. Six individuals did not report the reason for not receiving care when they last needed it.

One hundred eighty-four (74 %) had ever sought general care from both a biomedical facility (hospital, health centre, or clinic) and a traditional healer, whilst 63 individuals (25 %) only sought general care from a biomedical facility and 1 person sought general care only from a traditional healer.

Within the last year, 132 people with epilepsy (53 %) report consulting for epilepsy at a clinic, 162 (65 %) at a hospital and 34 (14 %) with traditional healers. One hundred sixty-eight individuals (67 %) reported previously seeking epilepsy care from both a biomedical facility and a traditional healer, whilst 61 individuals (24 %) have sought care from only biomedical facilities. Eleven individuals (4 %) reported only seeking care for epilepsy from traditional healers, whilst 10 individuals sought no care- neither from a biomedical facility nor a traditional healer. People with epilepsy sought general care at the government clinic more frequently than epilepsy-related care (94 % versus 76 %), whilst utilizing hospital care at similar levels for both epilepsy and general care (91 % versus 86 %) (Table 2). The average number of hospital visits in the previous 12-months was 4.06 (standard deviation (SD): 4.75) and average number of clinic visits was 6.5 (SD: 5.75).

The median cost per clinic visit was found to be $1.74 (IQR: $0-$4.80) (mean: 2.74; SD 3.74), whilst the median cost per outpatient hospital visit was found to be $9.08 (IQR: $6.11-$12.91) (mean: 9.73; SD 5.34) – a statistically significant difference (p < 0.001). The costs for epilepsy care at clinic and hopsital can be found in Table 3. The average annual cost for out-of-pocket, outpatient biomedical care at a clinic or hospital was found to be $58.41.

The median travel and waiting time at the hospital was twice that of the clinic (120 min/60 min and 240 min/120 min, respectively), whilst the time spent with a health professional was the same (10 min) (Table 4). The majority of respondents walked to the clinic but took a public taxi to the hospital. During their last visit to the hospital, 30 % of respondents saw a physician and 70 % saw a nurse. During their last visit to the clinic, 11 % saw a physician and 89 % saw a nurse.

During the last visit to a traditional healer 160 individuals (87 %) had to pay for services, with payment in all cases being cash. The median cost for a traditional healer’s services was $52.36 ($34.90-$87.26). One individual reported only having to pay the traditional healer if s/he was cured of epilepsy.

We found disclosed lifetime traditional healer usage to be high in people with epilepsy for both epilepsy- and non-epilepsy-related care (73 % and 74 %, respectively), although the frequency of visits for both general and epilepsy-related care was substantially less than visits to both biomedical facilities within the 12 months preceding this study. A national South African survey found similar patterns, though lower utilization levels, and suggests that poorer, rural, black populations utilize traditional healers more than urban, non-black populations [27].

Utilization of traditional healers within the previous 12 months was low when compared to those who reported ever seeking traditional healer care for epilepsy. It may be that traditional healer care was sought during the initial onset of seizures and care discontinued as the seizures continued and traditional treatment proved ineffective. Still, nearly 15 % of people with epilepsy reported at least one visit to the traditional healer for epilepsy care within the previous 12-months.

Our research concurs with findings that suggest the costs associated with traditional healers are substantially higher than costs of seeking care at biomedical facilities [28]. The traditional healers’ costs we present only include traditional healers’ fees and do not include time and money spent traveling to and waiting to be seen by a traditional healer, which would likely increase the total costs. Our results likely represent the lower bound of costs associated with traditional healer costs. The higher cost per visit may be financially prohibitive and result in the lower frequency of utilization. Furthermore, the aggregate cost of traditional healers, taking into account the lower frequency but higher per visit cost, may be similar to the cost of biomedical facilities, which are utilized more often, but cost, per visit, substantially less.

Understanding determinants of traditional healer utilization in this context is warranted as the cultural understanding of epilepsy has been shown to be an important determinant in care-seeking behavior [29]. Identifying ways of working with traditional healers to reduce the epilepsy treatment gap, however, may be more productive than attempting to change individual’s behavior.

Seeking care at the hospital took significantly longer and cost significantly more- more than twice the cost of seeking care at the clinic. As costs (including lost time) associated with seeking epilepsy treatment have commonly been found to be associated with the epilepsy treatment gap [15], increasing the utilization of care at clinics by strengthening epilepsy care delivery will likely reduce costs and potentially improve the treatment gap. Furthermore, reducing transportation cost could result in an overall reduction in out-of-pocket costs given that transportation costs have been highlighted as a significant cost in previous work in the area [28].

The overall annual out-of-pocket costs for out-patient epilepsy care at biomedical facilities was found to be $58.41, on average, yet this only represents a portion of the total cost of epilepsy care in rural South Africa.

Patients’ AED costs are likely to be lower in South Africa than other LMICs due to the free availability of AEDs at health facilities. No patient reported paying for AEDs in this study. In contrast, two other African studies found AEDs to contribute 10 % [8] and 22 % [9] of the total direct epilepsy treatment-related costs. One study from rural Kenya found that people with epilepsy who had to pay for AEDs were less likely to seek care [30], suggesting that reducing or removing the cost of AEDs might improve accessibility and consequently adherence [31]. Future work exploring adherence patterns in South Africa, where AEDS are free, is warranted and could be a natural experiment to establish whether free availability of AEDs to patients truly does improve adherence.

While this study presents previously unavailable information on utilization patterns and out-of-pocket costs for people with ACE in rural South Africa, there are a number of potential limitations.

We only examined people with ACE in this study. We recognize that this cohort does not represent all epilepsies and limits the interpretation of the results of this study to people with ACE only. However, convulsive epilepsies are often associated with the largest morbidity, stigma and mortality [32] and potentially pose a greater burden to individuals and families compared to non-convulsive epilepsies.

This study did not measure indirect costs associated with ACE, whilst a number of studies have found that this cost can be significant- ranging from 12 % to 85 % of the total annual costs due to epilepsy [7]. As such, this study likely underestimates the total cost of ACE. This study also only explored out-of-pocket, outpatient costs associated with ACE and did not include costs associated with hospital admissions or epilepsy-related surgeries (in-patient costs). Studies from Italy have found in-patient costs to be significant, especially in new-onset epilepsy [33, 34]. This study only examined ACE, and outpatient treatment patterns and costs for other types of epilepsy may vary; however, further studies are needed to verify this.

We chose an out-of-pocket costing approach for this study. In countries, such as South Africa, with high levels of unemployment and substantial inequality, using the GDP per capita as a proxy for productivity may not represent the actual productivity loss due to epilepsy and, some argue, fails to account for the value of human life [35, 36]. As such, a wider, societal costing approach was not adopted in this study. Further development of methodologies to allow for the accurate calculation of productivity costs in areas of high unemployment where formal sector employment is scarce is needed.

All costs and utilization analyzed in this study were self-reported. Recall bias or intentional bias in reporting may result in misreporting of costs and utilization associated with seeking epilepsy care. Every effort, through careful probing by well-trained fieldworkers, was made to ensure the collection of accurate data.

Finally, this study only explored out-of-pocket costs of epilepsy to the patient. Given the free availability of AEDs in South Africa and the subsidized direct, medical health care costs within the public healthcare system, epilepsy-related treatment costs to the health care system are potentially greater than in other African countries.