Comprehensive evaluation of the psychosocial parameters of epilepsy: A representative population-based study in Prey Veng (Cambodia)


  • Devender Bhalla,

    1. INSERM UMR 1094, Tropical Neuroepidemiology, Limoges, France
    2. School of Medicine, Institute of Neuroepidemiology and Tropical Neurology, CNRS FR 3503 GEIST, University of Limoges, Limoges, France
    3. Centre Hospitalier Universitaire, Limoges, France
    4. Cambodian Society of Neurology, Phnom Penh, Cambodia
    Search for more papers by this author
  • Kimly Chea,

    1. University of Health Sciences, Phnom Penh, Cambodia
    Search for more papers by this author
  • Hun Chamroeun,

    1. Cambodian Society of Neurology, Phnom Penh, Cambodia
    2. University of Health Sciences, Phnom Penh, Cambodia
    3. Department of Neurology, Calmette Hospital, Phnom Penh, Cambodia
    Search for more papers by this author
  • Chan Vichea,

    1. Department of Neurology, Calmette Hospital, Phnom Penh, Cambodia
    Search for more papers by this author
  • Pierre Huc,

    1. INSERM UMR 1094, Tropical Neuroepidemiology, Limoges, France
    2. School of Medicine, Institute of Neuroepidemiology and Tropical Neurology, CNRS FR 3503 GEIST, University of Limoges, Limoges, France
    Search for more papers by this author
  • Chan Samleng,

    1. Cambodian Society of Neurology, Phnom Penh, Cambodia
    2. University of Health Sciences, Phnom Penh, Cambodia
    3. Department of Neurology, Calmette Hospital, Phnom Penh, Cambodia
    Search for more papers by this author
  • Robert Sebbag,

    1. Department of Access to Medicines, Sanofi, Gentilly, France
    Search for more papers by this author
  • Daniel Gérard,

    1. Department of Access to Medicines, Sanofi, Gentilly, France
    Search for more papers by this author
  • Michel Dumas,

    1. INSERM UMR 1094, Tropical Neuroepidemiology, Limoges, France
    2. School of Medicine, Institute of Neuroepidemiology and Tropical Neurology, CNRS FR 3503 GEIST, University of Limoges, Limoges, France
    Search for more papers by this author
  • Sophal Oum,

    1. University of Health Sciences, Phnom Penh, Cambodia
    Search for more papers by this author
  • Michel Druet-Cabanac,

    1. INSERM UMR 1094, Tropical Neuroepidemiology, Limoges, France
    2. School of Medicine, Institute of Neuroepidemiology and Tropical Neurology, CNRS FR 3503 GEIST, University of Limoges, Limoges, France
    3. Centre Hospitalier Universitaire, Limoges, France
    Search for more papers by this author
  • Pierre-Marie Preux

    Corresponding author
    1. INSERM UMR 1094, Tropical Neuroepidemiology, Limoges, France
    2. School of Medicine, Institute of Neuroepidemiology and Tropical Neurology, CNRS FR 3503 GEIST, University of Limoges, Limoges, France
    3. Centre Hospitalier Universitaire, Limoges, France
    • Address correspondence to Prof. Pierre-Marie Preux, Institut de Neuroépidémiologie et Neurologie Tropicale (EA3174), Faculté de Médecine, 2, rue du Dr Marcland, 87025 Limoges Cedex, France. E-mail:

    Search for more papers by this author



We conducted a population-based study of epilepsy in Prey Veng (Cambodia) to explore self-esteem, fear, discrimination, knowledge–attitude–practice (KAP), social-support, stigma, coping strategies, seizure-provoking factors, and patient-derived factors associated with quality of life (QOL).


The results are based on a cohort of 96 cases and matched controls (n = 192), randomly selected from the same source population. Various questionnaires were developed and validated for internal consistency (by split-half, Spearman-Brown prophecy, Kuder-Richardson 20), content clarity and soundness. Summary, descriptive statistics, classical tests of hypothesis were conducted. Uncorrected chi-square was used. Group comparison was done to determine statistically significant factors, for each domain, by conducting logistic regression; 95% confidence interval (CI) with 5% (two-sided) statistical significance was used.

Key Findings

All questionnaires had high internal consistency. Stress was relevant in 14.0% cases, concealment in 6.2%, denial in 8.3%, negative feelings in public in 3.0%. Mean self-esteem was 7.5, range 0–8, related to seizure frequency. Mean discrimination was least during social interactions. Coping strategies were positive (e.g. look for treatment). Postictal headache, anger, no nearby health facility, etc. were associated with QOL.


The reliability of our questionnaires was high. A positive social environment was noted with many infrequent social and personal prejudices. Not all populations should (by default) be considered as stigmatized or equipped with poor KAP. We addressed themes that have been incompletely evaluated, and our approach could therefore become a model for other projects.

Epilepsy is a major clinical disorder with important social implications (Quintas et al., 2012) resulting in numerous challenges due to stigma, discrimination, low self-esteem, and so on. Although some of these issues have been widely investigated, many remain to be fully addressed. For instance, people with epilepsy (PWE) are known to face stigma, but whether, and to what extent, it negatively influences them (for example in terms of discrimination and the need for coping strategies) when they are making decisions about their lives, is less well defined. Quality of life (QOL) has also been extensively examined but, in most cases, not from the patient's point of view (Schipper et al., 1990). It is essential that patients themselves identify what affects their QOL, thereby minimizing interrater biases. Other factors such as stress, self-esteem (SE), discrimination, social support, and fear related to illness are thought to have some relation to epilepsy, but again they have been examined in only a fragmented manner (Collings, 1990; Eklund & Sivberg, 2003; MacLeod & Austin, 2003; Raty et al., 2007; McCagh et al., 2009). Understanding of these aspects is of utmost importance if we are to understand the individual patient's perceptions of epilepsy and its treatment. Lack of comprehensive information in these areas may delimit any prospective implementation of prevention and control activities. Although incidence and prevalence estimates are lower in Asia than in other low- and middle-income (LAMI) countries, there are plenty of other reasons that make epilepsy more important in Asia than other LAMI regions (Preux & Druet-Cabanac, 2005; Mac et al., 2007; Mbuba et al., 2008). Epilepsy in Cambodia continues to be investigated in a comprehensive manner, and about 86,000 people, on average (prevalence 5.8/1,000), have epilepsy (active plus inactive) (Preux et al., 2011; Bhalla et al., 2012). Plenty of possible opportunities as well as challenges have already been identified in this population (Bhalla et al., 2013). We conducted this population-based study to further develop our understanding of how people in Cambodia perceive epilepsy and its treatment. The objective was to elucidate self-esteem, fear related to epilepsy, discrimination, knowledge–attitude–practice (KAP), stigma, social support, and patient-perceived factors associated with quality of life (QOL) in Prey Veng Province, Cambodia.


Study design, organization, and procedures

This was a population-based door-to-door representative study conducted in Prey Veng province, Cambodia (Fig. 1). A random cluster survey design was used in which each village was considered a cluster. This province was chosen because of its easy accessibility from the National Capital, Phnom Penh, and its sociodemographic features, which are largely representative of Cambodian national data (Table 1). The study population consisted of all adults and children ≥1 year of age (N = 16,510) residing in the households of randomly selected villages (n = 24) within Prey Veng province. All those who had epilepsy (n = 96) were identified using a two-phase door-to-door procedure: screening with a prevalidated Limoges screening questionnaire for epilepsy in tropical countries; and confirmation by a neurologist (Preux et al., 2011). The screening instrument was translated into Khmer and pretested before final application. All subjects were examined with portable electroencephalography (EEG). A detailed history was taken and an examination was conducted; neuroimaging was not done because of operational difficulties. Those with febrile and neonatal seizures were excluded. To facilitate comparison of various parameters, we also recruited a control group. The controls underwent complete examination by the neurologists. Those who were declared free of any neurologic disease were recruited. The controls were randomly selected by using the same population register of the same source population that generated the cases. We recruited two controls for each case. Each control was individually matched for age (±5 years in adults and ±2 years in children <15 years old), gender, and residence (residence in the same village). All controls underwent the same procedures as cases, and had same exclusion and inclusion criteria except the presence of epilepsy.

Table 1. Sociodemographic features of Prey Veng versus Cambodia
ParameterPrey Veng (%)Cambodia (%)
Proportion of those aged 15–64 years56.8 (2008)63.2 (2010)
Percent distribution of illiterate people (>6 years age)23.525.0
Percent distribution of population in lowest wealth quintiles27.920.0
Percent distribution of birth registration (<5 years age)49.754.2
Percent distribution of the de facto household population ill or injured in the previous 30 days82.483.8
Percent distribution of women who attended secondary school16.924.8
Figure 1.

Map and location of Cambodia and Prey Veng province.

At the time of data collection, all cases had a history of treatment with either phenobarbital or valproate in two manners: doorstep cost-free treatment for 1 year (primary treatment period) and subsequent cost-free treatment through primary health care centers (treatment continuation period). The results with regard to treatment aspects are submitted elsewhere; however, treatment data were included while conducting data analyses of all psychosocial parameters of this study. Ethical approval was obtained from the Ministry of Health, Cambodia. Verbal informed consent was obtained from each participant prior to enrolment in the study, as is recommended for surveys in populations where literacy levels are not uniformly high. The questionnaires used are presented in Appendix S1.

Development of the questionnaires

When designing the questionnaires, we initially searched the literature including textbooks, dissertations, scientific publications, and Websites for similar tools. That information was then used to design a preliminary framework for the new questionnaires, along with experts’ opinions. Items were selected based on relevance to the survey's research questions, feedback and input, and suggestions made by experts. The questionnaires were given to epilepsy experts to review and, based on their feedback, necessary changes were made to improve the clarity of items. Experts were asked to comment on the content clarity and soundness of the questionnaire in the context of our research objectives. The questionnaires were developed in English and later translated into Khmer. Local health care field staffs that have expertise in local cultural contexts were asked to comment on the content clarity and soundness of all questionnaires.

The resulting questionnaires covered self-esteem (8 items, 4 positive, and 4 negative), fear related to epilepsy (6 items, related to psychoemotional state; and risk of brain disease, increase in seizure severity/frequency, and premature death); discrimination (7 items, related to school/employment, societal interactions, and family/marital life); presence of stigma (8 items); KAP (11 items, related to different parameters of epilepsy); quality of life (43 items, related to various aspects of epilepsy relevant to both adults and children). Questions (SE, fear associated with epilepsy, discrimination, and stigma) were dichotomous in nature and required a simple yes/no response. The questions were posed and explained verbally to each participant in Khmer, and all were given sufficient time to answer as they deemed appropriate. QOL was patient-derived and all patients were given a list of all possible factors that might affect it (in both adults and children). A verbal explanation was given and they were requested to identify those factors that pertained to their QOL. Proxies were used when respondents were infants or otherwise unable to respond independently. Both primary participant and proxy were present at the interviews, and each response was confirmed by asking primary participants if they agreed with the proxy.


Data entry was carried out with Epi-info (Epi-info 3.5; Centers for Disease Control and Prevention, Atlanta, GA, U.S.A.) and data were analyzed using STATA (Stata Corp., College Station, TX, U.S.A.) version 9.1 for Windows. We assessed the internal reliability and consistency of the questionnaires by adopting suitable models; particularly split-half model, Spearman-Brown prophecy, and Kuder-Richardson (KR) 20 (Gravetter & Forzano, 2009). Regarding questions on SE, fear, and discrimination, subjects scored one for each item with which they agreed, and their total was the sum of their positive responses. Higher totals indicated greater self-esteem, fear, or discrimination experienced by each subject, as the case may be. Social support, stigma, KAP, and QOL were descriptively measured. Various summary and descriptive statistics and classical tests of hypothesis were conducted to determine numbers, means, medians, with respective standard deviation (SD), simple frequencies, t-score, and f- and z-statistics. Uncorrected chi-square was used. Group data were compared to determine statistically significant factors, for each domain, by conducting logistic regression. Group data were compared in several manners: first, between cases and controls, for parameters on which response was available from both these groups; second, for parameters on which response was only available from cases, two groups were made, depending on positive or negative response. For example, for “whether stress causes some of their seizures,” one group consisted of all those cases who responded yes and a second group consisted of those who responded no. Ninety-five percent confidence interval (CI) was used, and the statistical significance (two-sided) was set at 0.05. Required approvals and informed consent were obtained.


Clinical and demographic characteristics of the participants

Our study population comprised 96 cases and 192 matched controls. Males comprised 53.1% of cases and 53.2% of controls (p = 0.9). Median ages of cases and controls were 24.0 years (SD ± 13.6, range 3–70) and 28.0 years (SD ± 14.1, range 3–77), respectively. This difference was statistically significant (p = 0.02). Those aged <20 years represented 65.6% of the cases, and the most frequent age group was 12–20 years (30.2%). Five (2.6%) controls had health problems, namely chronic headache (n = 2) and sleep disturbance, gastric ulcer, and weakness (n = 1 each). All controls were neurologically normal.

Our cases had the following types of epileptic seizures: generalized tonic–clonic (76.0%), myoclonic (3.1%), absence (1.0%), simple plus complex partial (12.5%), olfactory partial (1.0%), mixed (6.2%), and undefined (0.2%). Within this cohort, 44.8% had idiopathic, 37.4% had cryptogenic, and 17.8% had symptomatic epilepsy. The major associated risk factors were the following: family history of epilepsy (odds ratio [OR] 2.6, confidence interval [CI] 1.0–6.6); eventful pregnancy (hemorrhage, hypertension) of the subject's mother (OR 8.6, CI 2.6–28.2); and long and/or difficult birth (OR 8.5, CI 2.1–33.1; Bhalla et al., 2012). Treatment results are submitted elsewhere; in short, the treatment gap was about 66% (primary treatment period) and 75% (treatment continuation period, see above); the number of seizures per month (during the primary treatment period) was 12.6 (before onset of any treatment), dropping to 4.7 after 3 months of treatment, and 1.2 after 12 months. The figure subsequently increased by 183.0% during treatment continuation. Seizure duration decreased by 71.0% during primary treatment (unpublished data).

Internal consistency of the questionnaires

The estimates obtained from the split-half correlation, Spearman-Brown prophecy, and KR20 were as follows: self-esteem (0.98, 0.99, and 0.98); stigma (0.77, 0.86, and 0.83); discrimination (0.96, 0.98, and 0.96); stress (0.96, 0.98, and 0.97); and fear (0.98, 0.99, and 0.97).


About 14% of cases reported that some seizures occurred as a result of stress (p < 0.00001). For all of them, a reduction in stress also led to a reduction in the number of seizures.


Only 6.2% kept their epilepsy secret from others (p < 0.00001); about 80% had no objection telling their friends about his/her epilepsy; about 3% cases felt that there are negative feelings in public; about 2% were negatively affected by this; about 18% were always thinking about epilepsy (p = 0.00001); and 8.3% had denied having epilepsy at some time (p = 0.00001).

A response was sought on the question of whether epilepsy was more feared than other well-known disorders, namely stroke, cancer, HIV/AIDS. Epilepsy was statistically more feared than a combination of HIV/AIDS, cancer, and stroke (p = 0.00001), and, individually, more feared than cancer (p = 0.07) and stroke (p = 0.01), but not AIDS (p = 0.36).

More cases than controls reported epilepsy as most feared (26.0% vs. 4.1%, p = 0.00001). Epilepsy was least feared by 13.5% and 10.4%, respectively (p = 0.43). More controls (5.2%) reported feeling fearful when seeing someone having a seizure than cases (2.0%) but the difference was not statistically significant (p = 0.21). More controls felt sadness, pity, and helpfulness (84.3%, 97.3%, and 100.0%, respectively) when seeing someone having a seizure than cases (33.3%, 21.8%, and 88.5%, respectively). None reported feeling shock when seeing someone having a seizure. More controls (96.3%) wanted their friends to tell them about their epilepsy than cases (80.0%), p = 0.00001. More controls (7.2%) than cases (3.0%) experienced negative feelings in public, p = 0.004.

Further analyses of cases “always thinking about epilepsy” showed that this tendency was statistically related to all negative domains of self-esteem (p < 0.05), fear of suffocation during seizure (p = 0.002), fear of having more frequent/severe seizures over time (p = 0.06), stressful state (p = 0.003), job loss (p = 0.02), adverse effects of treatment (p = 0.05), noncontinuation of treatment (p = 0.0001), and considering epilepsy as most feared disorder (p = 0.0001). It was not statistically related to seizure frequency (>3 seizures/month; p > 0.05).

Fear associated with epilepsy

The mean degree of fear related to epilepsy was 6.4 (21.0%, standard deviation [SD] 1.6, range 6–12). Comparison between those reported having a fear and those not yielded the following results: fear of suffocation during a seizure (21.1%, p < 0.00001), fear of having a brain tumor (19.2%, p < 0.00001), fear of emotional disturbance (17.3%, p < 0.00001), fear of having more severe or more frequent seizures over time (15.3%, p < 0.00001), fear of decline of intelligence (13.4%, p < 0.00001), and fear of premature death (13.4%, p < 0.00001).

Fear of death was negatively correlated with presence of stress (p = 0.0001), fear of suffocation during seizures (p = 0.0001), fear of brain tumor (p = 0.0001), having more frequent or severe seizures over time (p = 0.001), having generalized seizures (p = 0.09), and having more seizures (>3) per month (p = 0.02). Fear of death was positively correlated with having no seizures (p = 0.09), and considering oneself as worthy as other people (p = 0.02).


The mean self-esteem was 7.5 (SD 1.5, range 0–8). Not being as worthy as others was the major attribute (18.3%), and 11.0% felt themselves complete failures in life because of their epilepsy. Other attributes were insufficient respect (13.3%); and no good qualities, no positive attitude, not having much to be proud of, and feeling useless at times (all 11.0% each). Not being satisfied with oneself was noted in 10.0% of cases. The factors that were negatively correlated with overall self-esteem were the following: considering epilepsy as most feared (p = 0.003), and higher seizure frequency (>3 seizures per month; p = 0.07).

Patient-derived factors associated with QOL

Headache after seizures (41.6%) was the most frequently reported factor that affected QOL of our case population. Anger (28.1%), no health care facility nearby (27.0%), burden on family due to cost (21.8%), uncertainty of next seizure (13.5%), bed wetting (13.5%), fear of injury (11.4%), and worry if seizures can ever be controlled (11.4%), were other important influences on QOL among cases in Cambodia.


The mean discrimination was 1.0 (21.0%, SD 0.1, on a range of 0–7). The discrimination was related to job (51.3%), or marriage (29.7%), or social interactions (18.9%). The overall difference between those discriminated against (in any form) by others and those not was statistically significant (p = 0.001) and the same result was obtained for discrimination related to job (p < 0.00001), marriage (p < 0.00001), and social interactions (p < 0.00001).

Social support of PWE

A majority of cases (89.5%) reported receiving support from their family, neighbor, or others, with their epilepsy. The overall difference between those with support and those without any support was statistically significant (p < 0.00001).


About 89% of cases felt well informed about epilepsy. About 89% of cases and 96% of controls believed that the general public is well informed about epilepsy (p = 0.009). Knowledge about causes, symptoms, and seizure-provoking factors among cases and controls is presented in Table 2. When asked how individuals with epilepsy (in the general population) feel, the response (from cases and controls, respectively) was that these individuals feel ashamed (51% and 91.6%, respectively, p = 0.00001) or worried (50.0% and 89.5%, respectively, p = 0.00001) or the same as anyone else (26.0% and 3.6%, respectively, p = 0.00001).

Table 2. Knowledge–attitude–practice related to epilepsy in Cambodia
ParameterCases (%)Controls (%)
  1. All figures rounded to single digit.

  2. Causes, causes of epilepsy; Feel, How do you feel when you see someone having a seizure?; Precipitants, endogenous or exogenous precipitants for seizures; Symptoms, symptoms of epilepsy; What to do, What do you do when you see someone having a seizure?

CausesBrain lesion (40), brain infection (37), stress (9), fever (12), genetic (8), tetanus (5), psychological (5), birth injury (1)Brain lesion (68), brain infection (27), fever (19), alcoholism (16), tetanus (10.0), genetic (9), dog bite (6)
SymptomsUnconsciousness (91), up rolling of eyes (85), tongue bite (82), salivation (70), incontinence (40), exhaustion (17), remember nothing after attack (5.5)Unconsciousness (96), tongue bite (95), up rolling of eyes (94), salivation (91), incontinence (29), exhaustion (4)
PrecipitantsExertion (45), full moon (29), hunger (17), fever (17), stress (14), emotional disturbance (10), noise (4), light (1)Exertion (53), full moon (34), fever (24), hunger (10), emotional disturbance (7), light (7), every time (1)
FeelAshamed (54), worried (53), same (28), fearful (4)Ashamed (92), worried (90), fearful (4), same (4), incapable (1)
What to doEnsure ventilation (72), protect head (25), remove dangerous objects (24), take to practitioner (16), pull tongue (11), lay the person down (7), put object between teeth (8)Ensure ventilation (84), protect head (22), remove dangerous objects (15), take to practitioner (15), put object between teeth (6), pull tongue (1)

Strategies to cope with epilepsy

The coping strategies related to epilepsy were the following: look for treatment (71.8%); tell others about their epilepsy (52.0%); and isolate themselves (1.0%).


Many parameters that we assessed (fear, self-esteem, coping strategies) have been incompletely investigated, particularly in LAMI countries. However, their importance is well recognized, and it is crucial to elucidate how epilepsy is perceived in Asia.

Our questionnaires had very high internal reliability and content consistency, indicating that the ways in which we measured parameters using different constructs delivered reliable results. The questionnaires were put through several testing procedures: split-half model, KR20, and Spearman-Brown prophecy prediction. Split-half is an excellent measure when the questionnaire concerned is dichotomous (like ours) and two test administrations are unlikely or undesirable (Cohen & Swerdlik, 2002). High results on split-half indicate that its dichotomous nature was of no disadvantage to our questionnaires. Our split-half results were subsequently validated through KR20, with similarly high estimates. High estimates of our Spearman-Brown prophecy (no <0.83 for any questionnaire) indicate that our questionnaires had appropriate test length and number of questions and that many biases (irrelevant questions, multiscale responses that may artificially increase the reliability of a questionnaire, and no straightforward answer) may therefore have been overcome. Like all questionnaires, ours are also subject to biases related to test-taking skills, response time, ability to guess answers, conduct of the administrator, leniency error, and so on.

Many factors and situations, such as alcohol abuse or stress, may provoke seizures, as happens in 63–92% cases (Frucht et al., 2000). Stress is perceived to have both anticonvulsant (acute stress) and proepileptogenic (chronic stress) effects and has been reported to occur in both prospective and retrospective studies (Haut et al., 2003). Stress in our population was associated with seizures in about 14% of cases, which indicates that stress cannot be a universal precipitant of epileptic seizures, except for some patients, in Cambodia. All stressed individuals reported that reduction of stress reduced seizure frequency, thereby emphasizing the need to identify and counsel them. Focus on the mechanisms through which stress provokes seizures could elucidate epileptogenesis and possibly facilitate antiepilepsy treatment.

Stigma becomes particularly important when it actually affects someone with epilepsy. Theoretically, each person with epilepsy carries some degree of risk of being discriminated against in one way or another, but we noted that discrimination is not a universal phenomenon and only some PWE in Cambodia experience it. In our population, discrimination was greatest in relation to employment (or education among children); others concur (Jacoby et al., 2005a). Discrimination is not just a feature of LAMI countries alone and may vary in its nature or frequency (Jensen & Dam, 1992; Rhodes et al., 2008) across different regions. Discrimination is often believed to stem from negative beliefs such as considering epilepsy as contagious and shameful (Jensen & Dam, 1992; Sidig et al., 2009), or from insufficient knowledge among the public. Other studies in Western populations have reported the presence of attitudes and knowledge gaps that have the potential for encouraging discriminatory behavior (Jacoby et al., 2004). These are not necessarily sufficient explanations, as populations with high degrees of awareness and infrequent contagious beliefs still show high levels of discrimination (Sidig et al., 2009). We hypothesize that this may in fact be a result of genuine concern of others; for PWE, for instance, it may be a genuine concern to minimize the risk of injury while working. This seems more likely in a society, like ours, where social support for epilepsy cases and close social ties are almost a norm. The fact that social discrimination is the least frequent form in our population supports this premise. Further support can be deduced from the fact that nearly all controls reported being helpful when seeing someone having a seizure (Table 2). This underscores the importance of making society conscious of epilepsy (Souza et al., 2002) and promoting closer social ties and positive interpersonal relations.

Many factors that may reflect stigma were infrequent in our population; for instance, there was little tendency for denial and concealment, and isolation was a rare coping strategy. Most of the cases as well as the controls would have no objection to telling a friend he or she had epilepsy. This corroborates our premise that not all populations are likely to be stigmatized and not all populations are negatively influenced by stigma when making epilepsy-related choices. Elsewhere as well, it has been reasoned that stigma has been dealt with “obsessively” and there has been a tendency to overly justify many epilepsy-related parameters through stigma (Muela Ribera et al., 2009). Many studies from elsewhere have contradicted these results (MacLeod & Austin, 2003; Baker et al., 2005a; Hosseini et al., 2010) reporting, for example, concealment, intervention of God, and social withdrawal to cope with discrimination and stigma (Jacoby et al., 2005b; Hosseini et al., 2010).

Nearly one-fourth of our cases had some fear of damage or death due to epilepsy. Fears, such as that of premature death, brain damage, and so on, which we observed in our study, are also observed in other populations (Ward & Bower, 1978; Mittan, 1986; Austin et al., 1995). In our study, there was a statistically significant fear of premature death related to many factors (see relevant results section above) including fear of suffocation during seizures, fear of having more frequent/severe seizures over time, of having more seizures (>3) per month, and of having generalized seizures. Suffocation is an important cause of sudden unexpected death in epilepsy, more so for those having generalized seizures (Nashef & Ryvlin, 2009). This fear may arise due to incorrect body position or difficulties in breathing during the periseizure phase (Nashef & Ryvlin, 2009). Correlation between sudden unexpected death in epilepsy and seizure frequency/severity is also well recognized (Lhatoo & Sander, 2002). There was no correlation between epilepsy-related fears and treatment or its side effects. This is also in contrast to other reports (Reynolds, 1983), and absence of this prejudice may indicate that treatment is likely to be more acceptable in our population. The role of the predominant trade in this province, rice cultivation, must be borne in mind because of the risk of drowning.

We noted that self-esteem was not greatly affected in our population with epilepsy. This is surprising as epilepsy is, more often than not, reported to have a strong negative impact on a person's self-esteem or self-concept (Baker et al., 2005b). Observations similar to ours are, however, reported by other studies, in which no significant differences in self-esteem between epilepsy and nonepilepsy individuals were noted (Lee et al., 2008). The most likely explanation could be the presence of the following three important buffers: social support, seizure controllability (both frequency, and severity), and optimal knowledge (Smith et al., 1991; Amir et al., 1999; Baker et al., 2005b; Gauffin et al., 2010), two of which were well represented in our population.

We assessed QOL directly from a patient's perspective. This approach is used only infrequently, and the use of different questionnaires (Gilliam et al., 1997; Martin et al., 2005) limits possible comparisons. QOL is essentially subjective, and patients, not physicians, should determine what affects it. Self-administered questionnaires emphasize patients’ own feelings regarding their QOL, regardless of how they correspond to reality, thus limiting any interrater bias. Our results showed that the effect of epilepsy on QOL is determined by a limited number of factors only, particularly postictal headache (PIH) and anger due to seizures. PIH is frequently (15–51% cases) present but is often neglected (Kasteleijn-Nolst Trenite et al., 2010) and briefly investigated (D'Alessandro et al., 1987; Schon & Blau, 1987). It was similarly frequent (41.6% cases) in our population. Its high frequency among epilepsy subjects indicates that both may share common pathophysiologic mechanisms (Kasteleijn-Nolst Trenite et al., 2010). To date, neither the International Headache Society nor the International League Against Epilepsy mention that headache/migraine may be the sole ictal epileptic manifestation. Many characteristics of our study (young age of population, generalized seizures, long duration of epilepsy) are particularly found to be predisposing factors for PIH among epilepsy subjects (Ekstein & Schachter, 2010). Anger is briefly reported as a QOL-related factor (Arunkumar et al., 2000). A 2-year follow-up study among people with drug-resistant epilepsy showed that resective surgery led to a decline in anger, which in turn led to improved QOL (Meldolesi et al., 2007). Anger can be an outcome of many factors such as difficulty in finding a treatment, which most of our patients look for (see relevant results). Patients experience difficulties in switching from a “sick” to a “normal” role and are easily disappointed if expectations of positive life changes are not rapidly met. Other QOL factors such as cost and lack of a nearby health care facility are to be expected in populations with infrastructural deficiencies (Barennes et al., 2011). Other factors such as uncertainty of subsequent seizure (Gilliam et al., 1997; Martin et al., 2005); bed wetting; fear of injury; and worry if seizures can ever be controlled affected approximately 15% of cases individually (Gilliam et al., 1997; Malhi & Singhi, 2005; Martin et al., 2005). These factors have already been shown to have a negative correlation with QOL among patients with epilepsy. Elsewhere, it has been clearly demonstrated that the QOL of patients with epilepsy is not limited to seizure frequency or seizure counts alone (Jacoby & Baker, 2000), indicating the importance of factors such as those noted in our study.

Although our population was rural, both cases and controls had a high level of familiarity and favorable KAP toward epilepsy (Table 2). This differs from results in other populations that widely report that neither epilepsy patients (Long et al., 2000) nor the general public (Kim et al., 2003) is knowledgeable about epilepsy. Our results also refute the assumption that all rural populations are by default likely to have poorer (or derogatory) KAP scenarios (Bishop & Boag, 2006). Lack of knowledge may promote a feeling of incapacity and may reinforce the misconception that epilepsy has no treatment (Fernandes et al., 2005; Pandian et al., 2006). This is not likely to be the case in our population because of the presence of positive coping strategies and a relatively small treatment gap, as noted earlier in this population (Bhalla et al., 2012). Shame and worry were, however, noted among both cases and controls. Despite the presence of shame, concealment of epilepsy was infrequent in our population. This again is in contrast to many other populations (Dalrymple & Appleby, 2000; Hosseini et al., 2010).

We did not intend to carry out any age-specific or seizure-specific analyses because we analyzed the study population as “one population group,” that is, a mix of people of all ages and seizure types. This allowed our results to be closer to those in a real average general population. Besides this, a statistically significant difference in the age of cases and controls might have influenced our results to some but unknown extent. The use of our inventories can sometimes be time-consuming, even though their use is essential to derive a complete picture of what PWE actually feel and believe in any given population. With practice, the interview time can be reduced. It is often believed that illiteracy may affect a patient's ability to effectively understand and respond to questions, but patients in low-income countries, although illiterate, have a long history of epilepsy, which in most cases spans a number of years. It is therefore unlikely that patients, though illiterate, would be unaware of what is happening to them and to what extent.

Although the content of our QOL questionnaire was not directly developed by taking input from the patients, this does not diminish its value. As shown elsewhere in this article, our QOL questionnaire included 43 issues that are possible barriers to a good QOL among PWE. Issues of particular relevance to children were also included, for instance the bad taste of tablets. Therefore, patients had many issues from which to choose those correlating with their own life situation. Furthermore, our QOL questionnaire extracted issues seldom extracted by physician-derived QOL inventories. Prominent examples include anger, postictal headache, bed wetting, and the bad taste of tablets.

Some studies have reported a direct impact of a high seizure frequency or of a particular seizure type on the psychosocial health (stigma and QOL) of PWE (Baker et al., 1997). Conversely, other studies have found no such effect (Smith et al., 1991). In our study as well, “always thinking about epilepsy” (a factor for internalized stigma) was statistically unrelated to seizure frequency (p > 0.05). Many factors apart from seizure frequency or seizure type have been implicated in stigma. They include factors related to the presence of a family history of stigmatizing disorder, personal income, fear, employment, treatment, and neurobiologic factors (Broussard et al., 2012). Some of these factors have had a more profound stigma-promoting effect in our population as well (see relevant results above).

In contrast to epilepsy-related factors (e.g., seizure frequency or type), QOL in our population was related to those determinants that mirror the “phenomenon of epilepsy.” Such factors included seizure uncertainty, postictal headache, and anger. Thus it is likely that not just stigma but also QOL is determined by deeper neurobiologic mechanisms that underlie epilepsy. Such a neurobiologic correlation was earlier proposed for stigma alone (Hermann & Jacoby, 2009).

There was a strong correlation between seizure frequency (and/or seizure type) with self-esteem as well as epilepsy-related fear (of premature death). Such a correlation was reported earlier for self-esteem (Gauffin et al., 2010) alone. These correlations may indicate that self-esteem and fear among PWE can likely be negotiated through optimal seizure control, especially for those with generalized seizures.


A summary of all results is given in Table 3. Complete management of epilepsy includes amelioration of both visible and not so visible challenges. By integrating identification and management of stress, PIH, and anger, into epilepsy care, we are likely to not only reduce the physical burden but also to improve the QOL of patients with epilepsy. A pathogenic overlap among stress, PIH, and epilepsy should be studied further to better understand epileptogenesis. Excellent motivation exists for agencies to establish a treatment program, as many patients look for it. With treatment, reducing psychosocial burden (e.g., self-esteem and fear) is also likely. Many of our results provide insight and solutions not only for Cambodia but also for other populations that face psychosocial defects.

Table 3. Summary results of various psychosocial parameters related to epilepsy in Cambodia
ParameterSummary results
  1. AIDS, acquired immunodeficiency syndrome; GE, generalized epilepsy; KAP, knowledge, attitude, and practice; QOL, quality of life.

StressStress reduction can help 14% of cases
Focus on stress mechanisms to understand seizurogenesis
StigmaNo tendency to deny or conceal, or any negative feelings in public
Epilepsy more feared than AIDS, cancer, and stroke
Epilepsy most feared for one-fourth of cases
Eighteen percent of cases always think about epilepsy (related to fear, stress, low self-esteem)
FearFear of suffocation and death in about 44% of cases
Correlated to GE, more seizures per month, low self-esteem
Self-esteemNot notably low
Correlated to considering epilepsy as most feared, more seizures per month
QOLPIH, anger, no health care facility, cost burden, uncertainty of next seizure, bed wetting, fear of injury, worry if seizures can ever be controlled
DiscriminationNot greatly present
Most related to job (getting a job, unfair treatment at work, job loss), and least for social interactions (shunned, avoided, insulted)
Social supportExtensive
Most important social buffer
KAPCases and controls are well informed
Most cases believe the public is well informed as well
Challenges related to shame, and worry
One-fourth of cases feel same (no shame, worry, dependency, incapacity, fear, depression)
Coping strategiesSeventy-two percent look for treatment
Fifty-two percent would tell others about their epilepsy status
One percent would isolate themselves due to epilepsy


Many epilepsy-related parameters that have been incompletely evaluated were addressed in this study; and the study is one of few conducted in an Asian setting. This can become a model for other projects. Our questionnaires had very high reliability in a Cambodian cultural context, and future studies may examine their generalizability to other cultures. Cambodia exhibited a positive social environment where many social and personal prejudices were infrequently present. Not all populations should by default be considered to be stigmatized or equipped with poor knowledge, attitude, or practice. Care should include appropriate identification and management of stress, anger, and PIH, at least. Treatment programs should be set up, as many patients look for treatment. Data are generated from limited resources so it becomes very important for agencies to take timely and suitable steps.


We acknowledge the support of Mr. William FRANCIS (France) for copyediting the final draft of the manuscript. This project received funding from the Department of Access to Medicines, Sanofi, France.


None of the authors has any conflict of interest to disclose. We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.