Performance of primary health care workers in detection of mental disorders comorbid with epilepsy in rural Ethiopia

There have been a number of studies indicating the high prevalence of comorbid mental disorders in people with epilepsy [1, 2]. The pooled prevalence of comorbid depression in people with epilepsy is estimated to be 22.9% (95% confidence interval (CI) 18.2–28.4%) and co-morbid anxiety disorders is 20.2% (95% CI 15.3–26.0%), based largely on studies from high-income countries (HIC) [2]. In hospital-based studies from sub-Saharan Africa, prevalence estimates for comorbid depression in people with epilepsy are even higher, e.g. 45.5% [3].

The relationship between epilepsy and mental disorders is complex [4]. There is an increased incidence (Incidence rate ratio (IRR) 1.5–15.7) of depression, anxiety and psychotic disorders before the onset of epilepsy compared to people who do not develop epilepsy [5, 6]. The incidence of these mental disorders is also increased (IRR 2.2–10.9) after the diagnosis of epilepsy compared to people with no epilepsy [5]. The complex causal interrelationship between depression and epilepsy has been most investigated [6]. Functional and structural neuroimaging studies provide evidence of common neuropathology underlying the association between depression and epilepsy [4, 6]. Regardless of the nature of the relationship, the existence of co-morbid mental disorders in people with epilepsy has a detrimental effect on prognosis of epilepsy and quality of life [4] .

Early detection and appropriate management of comorbid mental disorders in people with epilepsy is of paramount importance [4]. The evaluation should include lifetime and current history, family history of depressive, anxiety or attention deficit hyperactivity disorder (ADHD) and psychotic disorder. However, in routine clinical practice there is under-detection and limited treatment of comorbid depression and anxiety in people with epilepsy (PWE) [7]. Failure to recognize and properly manage these comorbid disorders has serious impacts [7]. Those who are diagnosed with mood disorders before the onset of epilepsy have been found to have an increased risk of treatment resistant epilepsy [8] and of developing adverse side effects of anti-epileptic medications [9]. The risk of suicide is also increased three-fold in people with epilepsy compared to the those without epilepsy [10].

In low- and lower middle-income countries (LLMICs), the treatment gap (PWE who are not accessing or unable to access biomedical facilities for diagnosis and treatment and, if accessing biomedical treatment, those not adhering to the prescribed anti-epileptic drug (AED)) for active epilepsy ranges from 25 to 100% [11, 12], even before considering access to care for co-morbid mental health conditions. The treatment gap for mental disorders in the general population in LLMICs is high, ranging from 1.6–15.4% [13]. In recognition of the high treatment gap for both epilepsy and mental disorders globally, the World Health Organization (WHO) recommends integrated management of priority mental, neurological and substance use disorders (MNS) at primary health care level [14, 15]. As part of this effort, the WHO has produced the mental health Gap Action Programme Intervention Guide (mhGAP-IG) comprising evidence-based clinical algorithms for MNS disorders that are suitable for non-specialist workers [14]. Epilepsy is included in mhGAP-IG, together with depression, psychosis, substance use disorders, suicidal behaviour, developmental disorders/child mental health problems, dementia and somatic/anxiety symptoms. Since its launch in 2008, there have been many efforts to adapt, implement and evaluate mhGAP, although there has been limited evaluation of the application of mhGAP-IG to people with co-morbid epilepsy and mental health disorders [16].

In Ethiopia, implementation of mhGAP in primary care settings has been shown to improve identification of people with psychosis and epilepsy [17], but not depression [18]. Qualitative exploration indicated unmet emotional needs in the people receiving mhGAP epilepsy care [19]. However, there have been very few studies to investigate the identification of comorbid mental disorders or the validation of common mental disorders screening instruments in people with epilepsy, especially in sub Saharan Africa [20]. The impact of using a locally validated screening tool on detection of comorbid mental disorders in people with epilepsy was found to improve detection in Zambia [21, 22]. However, further studies are needed to evaluate and determine the detection of comorbidity among people with epilepsy in primary care settings in order to quantify the treatment gap, and to inform the design of interventions to increase detection.

The aim of the current study was to evaluate the performance of primary health care (PHC) workers in detecting comorbid mental disorders, by comparing them to a screening scale and comparing their detection to a standardised reference diagnosis of comorbid mental disorders in people with epilepsy in rural Ethiopia. It also tries to clarify the misdiagnosis of PHC in relation to the different sociodemographic factors.

In this cross-sectional survey, the performance of primary health care (PHC) workers in diagnosing comorbid mental disorders against a standardised measure and a screening instrument (SRQ-20) for common mental disorders was examined. The sensitivity of the PHC workers’ diagnoses was low, although they had high specificity in relation to the standardised reference diagnosis. The psychometric properties of SRQ-20 indicate an optimal cut-off score of 9 and above, with moderate sensitivity and specificity but low positive predictive value. When the two diagnostic methods (SRQ-20 screening augmented by PHC workers diagnosis) were combined, the sensitivity was markedly improved but the positive predictive value remained low. Misdiagnosis of comorbidity by PHC worker was significantly associated with increasing age.

The low sensitivity of the PHC workers in detection of depression in our study sample is consistent with other studies carried out in Ethiopia and other parts of the world [18, 37‐39]. In this current study, less than half (45.5%) of the people with epilepsy and comorbid mental disorders were detected by PHC. Very little research attention has been paid to investigating the detection of comorbid mental disorders in people with epilepsy in routine clinical practice in low-income country settings [22]. Under detection and management of comorbid common mental disorders has consistently been found to be a problem in high income countries (HIC), despite the high prevalence of co-morbidity in people with epilepsy [7, 40]. One of the reasons identified for this under-detection is the soloed approach to care for people with epilepsy and mental health conditions, both in terms of inadequate training of neurologists in the psychiatric aspects of mental disorders and poor communication between the neurologist and psychiatrist [40]. In addition, failure of the training programme on psychiatric aspects of commonly occurring neurologic disorders for psychiatry residents, lack of interest in neurologic literature and the absence of psychiatrist in the team of neurologists were seen in a study from the USA [40].

This issue in low income countries like Ethiopia is different from the HIC where more frontline management of epilepsy is expected to be carried out in primary care, with little access to either neurologists or psychiatrists [24]. Health professionals working at the primary health care level are also expected to detect and manage five priority mental disorders based on their training through mhGAP [24]. It was shown previously that the PHC workers in rural primary health care in Ethiopia were more likely to detect people who presented with psychological than somatic symptoms, even though somatic symptoms are the more common presenting symptoms, and tended to detect those who had more severe forms of depression [18]. There were also multiple system and individual level barriers for under detection of depression by PHC workers. Some of these barriers included poor training of the PHC workers on mental disorders, non-biomedical explanatory models of mental disorders by patients and their family members, and somatic symptoms being the common presenting complaints of the patients [18]. In people with epilepsy, overlaps between symptoms of depression and the side effects of older classes of antiepileptic medications (e.g. phenobarbitone, phenytoin) makes early detection of depression challenging [41].

The sensitivity of SRQ-20 in screening for depression was not at a satisfactory level in people with epilepsy but the scale performed relatively well in the general population of Ethiopia and in a similar setting of Eritrea [30, 42, 43]. There is minimal evidence on the validation of SRQ-20 in special populations like people with epilepsy which has made it difficult to compare our findings to previous work. Systematic reviews of validation studies of screening instruments for common mental disorders in LMICs in the general population have shown that the instruments with best performance are those which have been locally developed from scratch for specific population in a specific setting [44]. The SRQ-20 is unusual because it was developed with LMIC contexts in mind, drawing on symptoms from scales used in both HIC and LMIC settings [28]. In the previous studies of SRQ-20 in Ethiopia, the local adaptation and attempts to make the instrument culturally sensitive are likely to have contributed to the observed validity [42]. However, in people with epilepsy, the high number of somatic items on the SRQ-20 may overlap with impacts of inadequately controlled epilepsy (e.g. on sleep) and the effects of antiepileptic drug side effects [45].

Less than half of the participants diagnosed by the PHC workers as having comorbid mental disorders scored above the optimum cut off point of SRQ-20. Specific items of the SRQ-20 were not differentially reported by the participants diagnosed to have comorbid depression by the PHC workers. This could be due to the small number of participants diagnosed to have comorbid depression by the PHC workers. Some of the somatic and depressive symptom items of SRQ-20 were also highly prevalent in both population (with or without depression). It has been previously shown in the same setting that people with depression spontaneously reported symptoms which are partially represented in the typical diagnostic criteria [46]. The concept of depression was not well recognized as an illness in this community unless it is usually associated with disruption of function [46]. Since the PHC workers working in rural settings are usually accustomed to the cultural ways of expressing distress, this could be inconsistent with the symptoms of depression represented in SRQ-20 and reduce the sensitivity.

This study has also demonstrated that use of the two methods (SRQ-20 screening augmented by the clinical evaluation) of detection comorbidity of depression is promising. This augmentation is seen by the high specificity of PHC workers which has compensated for the low specificity of SRQ-20. As people with epilepsy are a high risk population, the routine use of depression screening is highly recommended in HIC [7]. It is also recommended that the ideal screening tool should be able to detect depression even when patients are presenting with somatic complaints [47]. PHC workers knowing the common terms patients use for emotional symptoms and their relevance will help in identification of mental disorders [47]. However, in our study, the screening questionnaire was administered by research staff and was not used as part of clinical decision-making. Implementation of screening questionnaires for depression in routine PHC settings has yielded mixed results in high-income countries [48], indicating the need for future studies to evaluate the effectiveness of introduction of routine screening in people with epilepsy in LLMICs.

The inclusion and evaluation of three different diagnostic methods simultaneously in routine clinical care was one of the strengths of this study. It is also one of the few studies carried out on detection of comorbid mental disorders in people with epilepsy in sub-Saharan Africa. The results of this study indicate that attention needs to be focused on improving the detection of co-morbid mental health problems integrated within the scale-up of primary care-based mental health care in Ethiopia and similar settings. This study was a pragmatic study which has followed mhGAP-IG criterion for evaluation of people with epilepsy. The use of psychiatric nurses for the standardised clinical assessment rather than psychiatrists means that it may not be considered a gold standard. However, in the Ethiopian setting, psychiatric nurses take on additional clinical roles compared to their counterparts in high-income countries, including diagnostic assessment, and would have had experience and competence in this area. Even though OPCRIT+ allows local clinicians to probe and explore responses in a flexible way, it has not been validated in this specific context. The other limitation of this study was the absence of people with non-generalized seizures. This is likely to be because of the limited clinical experience of PHC workers in management of mental or neurological disorders before the implementation of mhGAP. Even after the mhGAP training the diagnosis of focal seizures and non-epileptic seizures needs experience and an EEG. Neither the primary health care facility nor the PHC workers are equipped with this skill or resource. The low detection of anxiety disorders in this study could also be due to the presentation of depression in Ethiopian culture, which tends to be a combination of anxiety, somatic and depressive symptoms rather than typical DSM criteria, alongside the non-biomedical causal attributions of depressive/anxiety symptoms in this society [27]. It is possible that psychiatry nurses tended to prioritise a diagnosis of depression over anxiety disorders as this reflected their experince of presentation in clinical settings.

Further, studies are needed on interventions to increase detection. The development, cultural adaptation and evaluation of the impact of symptom screening tools in routine settings is a promising avenue for future research. Training programmes for PHC workers in LMICs, such as mhGAP, may benefit from a more horizontally integrated diagnostic algorithms which facilitate detection of co-morbidity. Contextualisation using local expression of mental distress may also increase the impact on detection.

In conclusion the detection of co-morbid mental disorders in people with epilepsy by PHC workers was low. The use of screening instruments augmented by the clinical skill of the PHC workers may possibly improve the detection of mental disorders but needs further evaluation.