In this qualitative study from rural Ethiopia we explored the reasons for non-adherence to antipsychotic medications in persons diagnosed with schizophrenia, from the perspectives of patients, their families, health professionals and research field workers. Several of the reported reasons for non-adherence were in keeping with previous findings from high-income countries [6
], illustrating commonalities across settings. However, the particular contribution of this study was the finding that those factors considered most important to medication adherence were peculiar to this low-income country setting.
Economic factors such as lack of access to adequate nutrition emerged as the most salient factor for non-adherence from the perspective of the respondents. Chronic food insufficiency is still common in the study site [25
] despite recent positive developments in the country, and patients with severe mental illnesses were reported to have more chronic energy deficiency compared with their healthy counterparts [26
]. In fact, a 5-year mortality report from the same cohort of patients with schizophrenia showed severe malnutrition as the second most common cause of death, accounting for 13.2% of the deaths [4
]. In a situation where there is food shortage at the household level, healthy members of the family get the priority.
The inter-relationships between food insufficiency and non-adherence to antipsychotic medication appear to be multi-faceted. Antipsychotic medications are known to increase appetite and craving for food which is generally believed to be mainly through stimulation of the histaminergic neurotransmitter system [27
]. Appetite stimulation by the medication would lead to feeling of hunger (subjectively distressing) or increased food consumption (increased burden on family). While there are some studies from western countries, where there is abundance of high fat and high calorie foods, suggesting increased appetite and the consequent weight gain adversely affecting adherence to antipsychotic medications [28
], there is paucity of data from low income countries where food is scarce. But, a review of the HIV literature from sub-Saharan African countries indicated lack of food or hunger being an important factor affecting adherence. Patients reported increased appetite caused by the medications, and inability to get adequate food forced them to stop the medications [29
]. This same experience was reported by our patients as well as their caregivers.
A previous qualitative exploration of purposively selected patients from the Butajira course and outcome study found that one of the justifications patients gave for ongoing chewing of the leaves of khat, an amphetamine-like stimulant which has anorexigenic properties, was as a means to curb their appetite. Family caregivers who couldn’t provide food would allow patients to chew khat despite being aware of its negative impact on the patient’s illness such as worsening of symptoms [31
]. However, another factor linked inadequate food availability with non-adherence: perceiving the medication to be ‘strong’, requiring the person to take adequate food in order to be able to tolerate it, emerged as a strong factor affecting adherence. The apparent fatigue which could potentially be caused by the typical, low potency antipsychotic medications (e.g. chlorpromazine and thioridazine) that the patients were taking was perceived to be caused by the lack of a balanced diet [32
Another important factor reported to affect medication adherence was the role of the family in the care of patients. In the absence of community-based mental health care, patients relied on family to access care. Patients may be too unwell, lack insight or be disabled by negative and cognitive symptoms of schizophrenia, to be able to remember their appointment dates and attend clinics on their own, particularly as health facilities are usually some distance from the patient’s place of residence. As well as enabling the patient to access care, family members also play an important role in ensuring that their affected relative takes their medication, through reminders and encouragement. A recent randomized controlled study carried out in Peshawar, Pakistan, found that adherence significantly improved whenever patients were supervised by relatives, an approach which the authors called STOPS, Supervised Treatment in Out-Patient for Schizophrenia [33
]. The actions of family members to promote adherence may go beyond mere encouragement. In this study, some family members went so far as to mix the medication with drinks (tea, milk) covertly, while others employed various coercive techniques to make the patients under their care take the medication. Such covert administration of treatment and coercion is commonly practiced on patients living in LAMI countries [34
]. It has been argued that, for severely unwell persons who lack the capacity to make decisions about their treatment, and in the absence of adequate community-based services, such actions may serve to promote, rather than undermine, a patient’s autonomy through restoring their health [35
]. However, patients can be abused in the process, in addition to issues relating to autonomy and consent [36
Exclusive reliance upon family support is also problematic when such support is not consistently present, as was reported to be the case by participants in this study. Caregiver burden, both economic and emotional, may contribute to the waning support over time, although a previous report from this study showed that accessing effective treatment was one of the most important factors in alleviating burden [37
]. Without the support of families, many persons with schizophrenia and other severe mental disorders in this setting become vagrants or destitute, contributing to their greatly reduced life expectancy [4
The other important factor reported to affect adherence was lack of knowledge about the nature of the illness, the medications and the expected outcome from the treatment. Considering apparent absence of symptoms as cure led to discontinuation of treatment. Or on the contrary, they expected too much from the treatment, and when ‘cure’ didn’t happen, they abandoned the medications and sought help from alternative health care such as holy water treatment, ‘tsebel
’, or consulted traditional/ indigenous healers. In a rural Ethiopian setting where the burden of acute, curable infectious disease is high, patients and caregivers may find it difficult to shift to a chronic disease model of care and accept the need to continue taking medication [38
Stigma was another factor reported to affect adherence. Having a severe mental illness such as schizophrenia is highly stigmatizing in the area, with low prospects for work or marriage [20
]. It is not only the patients, but also their relatives who are stigmatized [39
]. This issue was particularly relevant for patients who were no longer suffering from any symptoms. After showing some improvement, patients were liable to decline the offer of continued care and stop their medication. They were not even willing to talk with the mental health field workers or psychiatric nurses whenever they visited them in their villages.
In this study, one of the reported factors affecting adherence to antipsychotic medications was lack of insight. As reported time and again by family members, lack of insight led to refusal to take medications. Lack of insight is one of the hallmarks of psychosis [40
]. There are several explanations for lack of insight in schizophrenia including cognitive deficits, denial of symptoms or a lack of adaptive personal narrative understanding of illness [41
]. Although we didn’t specifically look into reasons for lack of insight in the study participants, given the diverse nature of their clinical status, the aforementioned reasons could apply to them as well. There is a complex relationship between insight and medication adherence. Several studies have shown that it is probably one of the most important factors affecting adherence and thereby resulting in poorer outcome [12
]. On the contrary, there are reports which showed that patients might continue to take medication, although they do not believe they have mental illness, because of relief of symptoms from taking medications [24
Medication side effects were frequently cited as an important factor leading to non-adherence in this setting. The only antipsychotic medications available to the patients in the study were a very limited selection of first generation antipsychotic medications (FGAs). These FGAs have the potential to cause distressing side effects in a context where medications to counter side effects is erratically available. Several studies on medication adherence identify side effects as a very important factor affecting adherence [14
]. Unfortunately, the hope that second generation medications would have greater tolerability and, therefore, improved adherence has not been borne out in practice [47
The behavior of some of the health care providers was also mentioned as a factor affecting adherence. Several studies have shown that the therapeutic alliance with patients plays an important role in adherence to prescribed treatments [49
Substance abuse was also reported to be a common reason for non-adherence. Patients who chewed khat and drank alcohol had a reportedly low level of adherence to antipsychotic medications. In keeping with these findings, a previous qualitative study nested in the Butajira cohort study also reported that patients did not like the idea of taking khat and medication together because of fears that harm would result from combining two powerful chemicals [31
]. Co-morbid substance abuse is known to adversely affect adherence to treatment in patients with schizophrenia in high-income countries, and a similar picture appears to pertain in a rural African setting [8
The main limitation in this study is the broad definition of non-adherence i.e. subjective report by patients of not taking their medication, family members’ report or a report by health professional based on a missed clinic appointment. Not all participants, patients and families mentioned non-adherence as a problem for them personally, which could be related to social desirability bias or fear of retribution because the interviewer was a doctor working in the psychiatry clinic. Without an objective measure of adherence we couldn't evaluate the impact of social desirability. However, we did know their adherence history as documented in the monthly follow-ups by the Butajira study nurses. So, we had contemporaneous, independent indicators of adherence and were not relying solely on self-report.