Welcome to the Schizophrenia Resource Centre

Welcome, this website is intended for healthcare professionals in EMEA with an interest in the treatment of schizophrenia. By clicking the link below you are declaring and confirming that you are a healthcare professional

You are here

Long-term clinical course and outcome of schizophrenia in rural Ethiopia: 10-year follow-up of a population-based cohort

Schizophrenia Research, 2-3, 161, pages 414 - 420



Although the few available studies from LMICs report favorable outcome, the course of schizophrenia is more complex than has been indicated so far.


A sample of 361 people with a standardized clinical diagnosis of schizophrenia were recruited from a predominantly rural community in Ethiopia and followed up regularly for an average of 10 years. Psychiatrists used the Longitudinal Interval Follow-up Evaluation chart to carry out assessment of illness course. Duration of time in clinical remission was the primary outcome.


About 61.0% of the patients remained under active follow-up, while 18.1% (n = 65) were deceased. The mean percentage of follow-up time in complete remission was 28.4% (SD = 33.0). Female patients were significantly more likely to have episodic illness course with no inter-episode residual or negative symptoms (χ2 = 6.28, P = 0.012). Nearly 14.0% had continuous psychotic symptoms for over 75% of their follow-up time. Only 18.1% achieved complete remission for over 75% of their follow-up time. Later onset of illness was the only significant predictor of achieving full remission for over 50% of follow-up time in a fully adjusted model. Conventional antipsychotic medications were fairly well tolerated in 80% of the patients and 4.2% (n = 15) experienced tardive dyskinesia.


This population-based study is one of the very few long-term outcome studies of schizophrenia in LMICs. The study demonstrated clearly a differential and more favorable course and outcome for female patients but overall course and outcome of schizophrenia appeared less favorable in this setting than has been reported from other LMICs.

Keywords: Schizophrenia, Course, Outcome, Gender, Follow-up study, Ethiopia.

1. Introduction

Prospective population-based studies of course and outcome of schizophrenia are scarce in low and middle-income countries (LMICs) (Thara and Rajkumar, 1992, Ohaeri, 1993, Thara et al, 1994, Gureje and Cohen, 2011, and Novick et al, 2012). Nevertheless, the clinical and functional course of schizophrenia in LMIC settings is presumed to be favorable ( Sartorius et al., 1978 ).

This discourse of a more favorable outcome in LMICs appears to have first emerged in reports from Africa ( Murphy and Raman, 1971 ). Subsequent data from the World Health Organization (WHO) multi-country studies (the International Pilot Study of Schizophrenia—IPSS—and the Determinants of Severe Mental Disorders—DOSMD) supported the initial reports (Sartorius et al, 1977, Sartorius et al, 1978, and Jablensky et al, 1992). Both the IPSS and DOSMD involved a single country, Nigeria, from Africa. In the IPSS-Nigeria study ( Sartorius et al., 1977 ), 42.6% (n = 60/141) had a diagnosis other than schizophrenia. Participants with a diagnosis of schizophrenia had either paranoid or catatonic schizophrenia. In the DOSMD-Nigeria sample ( Jablensky et al., 1992 ), nearly 20% (n = 26/142) did not receive the core diagnosis of schizophrenia and 76% of the participants were married. The cohort also had one of the highest proportions of participants in receipt of treatment despite a 31% drop out rate. The Nigerian cohort had the best short term outcome in both the IPSS and DOSMD. However, this outcome has to be seen in the context of the atypical nature of the sample.

Since these earlier reports, the notion of better outcome in LMICs has persisted (Kulhara et al, 2009 and Novick et al, 2012). The counter-intuitive nature of the evidence has been a point of contention given the high mortality ( Ran et al., 2007 ), and other poor prognostic factors (Sartorius et al, 1978, Sartorius and Janca, 1996, Mojtabai et al, 2001, Kurihara et al, 2006, and Cohen et al, 2008), including poor access to treatment and long duration of untreated psychosis (DUP) (Shibre et al, 2003 and Farooq et al, 2009), which in turn predisposes to poorer outcomes (Perkins et al, 2005, Kurihara et al, 2006, Ran et al, 2007, Crumlish et al, 2009, Farooq et al, 2009, and Teferra et al, 2011). Other factors such as internalized and public stigma (Shibre et al, 2002a, Shibre et al, 2002b, Van Zelst, 2009, and Assefa et al, 2012), absence of formalized social safety nets leading to over-burdened caregivers ( Shibre et al., 2003 ), and environmental factors are all likely to contribute to poorer outcome in LMICs.

The most recent series of reports on the outcome of schizophrenia come from the Butajira cohort in Ethiopia (Kebede et al, 2005, Alem et al, 2009, and Teferra et al, 2011).

These reports focused on short-term outcomes, which are likely to exaggerate negative outcomes. Therefore, we followed-up patients for additional years to provide evidence on the longer-term outcome of schizophrenia in LMICs. This data extends the evidence on the outcome of schizophrenia in LMICs. First, this is the largest single site study of the outcome of schizophrenia in Africa and, to our knowledge, any other LMIC. Secondly, patients were serially monitored with rich clinical data throughout the course of follow-up. Third, a more complete data was obtained on all subjects at the time of final follow-up. Fourth, participants were recruited from the population and are potentially the first population-based cohort from any LMIC and are likely to be a more true representation of the course of illness. Finally, the duration of follow-up was relatively long, offering a more realistic evidence on the course of schizophrenia.

2. Methods

The details of the case identification methods for the Butajira cohort have been reported previously (Shibre et al, 2002a, Shibre et al, 2002b, Kebede et al, 2003, Kebede et al, 2004, Fekadu et al, 2006, and Alem et al, 2009) and are outlined briefly below.

2.1. Study participants

The recruitment of participants took place between March 1998 and May 2001 and followed a two-stage method. In the first stage, 68,378 adults aged 15–49 years were interviewed through a house-to-house survey using the Composite International Diagnostic Interview (CIDI 2.1) (Sartorius and Janca, 1996 and Ustün et al, 1997), augmented by the Butajira key informant method (Shibre et al, 2002a and Shibre et al, 2002b). For those who screened positive in the first stage, diagnostic assessment was conducted using Schedules for Clinical Assessment in Neuropsychiatry (SCAN 2.1) (Sartorius and Janca, 1996 and Ustün et al, 1997). Following the SCAN interview, 321 cases (at initial enrollment) and 40 incident cases (through leakage survey) with confirmed schizophrenia (American Psychiatric Association, 1994 and World Health Organization, 1994), were recruited into the cohort. Thus 361 participants were followed up for a mean follow-up period of 10 years (range: 1–156 months), ( Fig. 1 .). All patients had access to basic psychiatric treatment, mainly first generation antipsychotics (haloperidol, chlorpromazine, thioridazine and fluphenazine decanoate) and tricyclic antidepressants (amitriptyline and imipramine), which are also commonly used in other psychiatric clinics in the country. Doses of antipsychotic medications, in chlorpromazine equivalent doses ( Taylor et al., 2009 ), were between 25 mg and 300 mg/day. Tolerability of antipsychotic medications was assessed clinically as part of the final review.


Fig. 1 Schizophrenia cohort identification and follow-up.

Clinical follow-ups were carried out monthly at the Butajira psychiatric clinic and two outreach clinics established by the research project. At each visit, any complaints, pertinent findings from a mental state examination, overall clinical impression and treatments were recorded. All patients with established diagnoses of schizophrenia who had at least one follow-up assessment after enrollment were included in the analysis.

2.2. Instruments

The following instruments were completed for all patients annually; Basic Information on the Study Subjects and informants, Scales for Assessment of Negative Symptoms (SANS) (Andreasen, 1982 and Andreasen, 1984), Scales for Assessment of Positive Symptoms (SAPS) ( Andreasen, 1984 ), Life Chart Schedule (LCS) ( Sartorius and Janca, 1996 ), Mania Rating Scale (MRS) ( Young et al., 1978 ) and the Hamilton Rating Scale for Depression (HRSD) ( Hamilton, 1960 ). The following forms were completed as indicated: refusers and deceased patients form, Broad Rating Scale (BRS) ( Sartorius and Janca, 1996 ) and physician reviewed verbal autopsy questionnaires (Soleman et al, 2006 and Abbas et al, 2011). The mean carer-burden score and carer-perceived stigma score at baseline were computed from the Family Interview Schedule (FIS) ( Sartorius and Janca, 1996 ). A comprehensive physical examination, assessment for involuntary muscle movements and neurological soft signs were carried out at enrollment ( Shibre et al., 2002b ).

Course and outcome data was collected using the Longitudinal Interval Follow-up Evaluation (LIFE) chart ( Keller et al., 1987 ). LIFE chart is a semi-structured questionnaire suitable for summarizing longitudinal data. Four psychiatrists who were trained for four days administered the LIFE chart. Training was conducted by a trainer (AF) who was trained by the developers of the LIFE chart methodology in the USA. The Ethiopia training included formal presentations by the trainer, ratings of recorded training video interviews, and pilot interviews in selected community samples around Butajira. The LIFE chart has been in use for over four decades and validated for use in different settings and for gathering longitudinal clinical data in psychiatric illnesses (Warshaw et al, 2001 and Judd et al, 2005). Although it was initially designed to assess the longitudinal course of affective symptoms, utility of the LIFE chart for rating the course of any psychiatric disorder is accepted ( Keller et al., 1987 ) and has been used in LMIC ( Shibre et al., 2014 ). The key rating tool in the LIFE chart is the Psychiatric Status Rating (PSR). The PSR is a six-point symptom severity rating, ranging from one to six. A rating of one indicates the absence of symptoms while a rating of six is consistent with severe symptoms. The PSR ratings allow calculation of mean scores or categorization of these symptom severity ratings into three threshold categories. Thus, a score of one and two were indicative of remission; a score of three and four were indicative of partial remission or a sub-threshold state, and a score of five and six were indicative of the occurrence of a full syndromal episode as defined in the DSM-IV. For alcohol and khat abuse or dependence, a three-point scale, corresponding to the three threshold categories, remission, partial remission or sub-threshold state and episode, was used. The PSR ratings were completed using all available information: reports by patients and family, the monthly clinical records, annual symptomatic and functional ratings, reports from psychiatric nurses, and reports from the project outreach workers who had monthly contacts with the patients and their families. When patients were unable to attend the psychiatric clinic for the final assessment, raters visited the patients' house with their permission. Two consultant psychiatrists trained in the use of the LIFE chart (AF and TS) supervised the ratings. Data on reliability of the PSR scores, medication tolerability and extrapyramidal side effects (EPS) of medication were all scored using items in the LIFE-chart. Level of functional impairment was measured using Global Assessment of Functioning (GAF).

2.3. Outcome measures

The primary outcome was the duration of follow-up in a state of remission, expressed in terms of percentage of the follow-up time [(months in remission / total months of follow up) × 100]. The PSR scores were used to summarize data. All clinical states described (remission, partial remission and episode) followed definitions of the DSM-IV course-specifiers ( APA, 1994 ). Remission was equated to score of one or two. The period in remission was then dichotomized at 50%. Participants who achieved remission for over 50% of their follow-up time were considered to have a “good clinical outcome”. Earlier reports from the same cohort also used a similar dichotomization (Alem et al, 2009 and Teferra et al, 2011).

2.4. Data analysis

EpiData version 3.1 was used for double data entry and the Statistical Package for Social Sciences (SPSS version 15) was used for data analysis. Data were analyzed in person-months to determine the proportion of time each study participant spent in treatment and in various clinical states. Episode, as per the PSR definition, was considered only when the patient fulfilled episode criteria according to the requirements of the DSM-IV. Sociodemographic and clinical factors known to be associated with remission were included in a logistic regression model. Negative symptoms, disorganized symptoms and psychotic symptoms scores were summarized from the SANS and SAPS data.

2.5. Ethical considerations

The Research and Publication Office of Addis Ababa University approved the study.

Informed consent was obtained from all study subjects. Free medication, were made available for all study participants throughout the follow-up period.

3. Results

3.1. Sample characteristics

A total of 361 patients were included. Three participants were excluded from this analysis ( Fig. 1 ). At baseline, 75% had illness duration of more than 2 years, 82.0% were males, 54.8% had never been married and 70.2% had severe functional impairment ( Table 1 ). At the time of the final assessment, 60.9% of the patients were under active follow-up ( Table 1 ).

Table 1 Baseline demographic characteristics and follow-up status of participants. Butajira cohort. 2012.

  Characteristics Number (%)
Gender Male 296 (82.7)
Female 62 (17.3)
Age at enrollment < 20 28 (7.9)
20–29 136 (38.5)
30–39 129 (36.5)
40–49 60 (17.5)
Marital status Single 193 (54.8)
Married 97 (27.6)
Formerly married 63 (17.6)
Education Non-literate 189 (53.8)
Literate 162 (46.2)
Employment status Unemployed 214 (59.8)
Employed 144 (40.2)
Level of functional impairment a None 11 (3.1)
Mild 17 (4.7)
Moderate 79 (22.0)
Severe 251 (70.1)
Follow-up status b Active 218 (60.9)
Deceased 65 (18.1)
Out migrated 20 (5.6)
Refuser 42 (11.7)
Homeless/vagrant 13 (3.6)
Clinical status during the follow-up period Mean (SD) percent time in complete remission 28.4 (sd = 33.0)
Mean (SD) percent time in partial remission 30.9 (sd = 35.4)
Mean (SD) percent time in psychotic episode 37.1 (sd = 31.2)

a As determined by Global Assessment of Functioning (GAF).

b Data missing in some category.

3.2. Gender and follow-up profile

The proportion of female participants who remained under active follow-up was higher than for males but the difference did not reach statistical significance (72.6% versus 60.1%, χ2 = 3.38, p = 0.066). Female participants were less likely to be deceased (12.9% vs. 14.4%), lost to follow-up as a result of address change (3.1% vs. 6.1%) or to refuse follow-up (8.1% vs.12.5%). Thirteen patients (11 males and 2 females) had experienced a period of homelessness during follow-up ( Fig. 2 .). Altogether, 65 participants died, giving a cumulative mortality rate of 18.2%. Most of the deceased were men (n = 57; 87.7%) and suicide accounted for 13.8% (9/65) of deaths. Only men had committed suicide.


Fig. 2 Gender specific follow-up status of patients with schizophrenia. Butajira cohort. 2012.

3.3. Gender and follow-up PSR

There was a progressive decline in mean PSR scores over the follow-up period. Among females, the mean PSR score declined from 3.8 (SD = 1.5) in year 1 to 2.5 (SD = 1.6) at year 10. Females had a lower mean PSR score at all annual assessments compared to males ( Fig. 3 ) indicating a more favorable course of illness.


Fig. 3 Gender and annual PSR scores of patients with schizophrenia. Butajira, Ethiopia. 2012. PSR scores range from 1 to 6. Higher PSR mean score represents poorer outcome. No statistically significant difference was observed at between the two groups during the follow-up years.

3.4. Pattern of illness course

Over one third of the patients, 37.8% (n = 135), had episodic course while 19.0% (n = 68) had continuous psychotic illness ( Table 2 ). Only 11.9% were continuously in full remission following a single episode. Males were more likely to have an episodic course pattern with inter-episode residual or negative symptoms (χ2 = 9.16, P = 0.002). The number of relapses per person ranged from one to 10, with a mean number of relapses of 1.55 (SD = 1.78). Overall there were 225 psychotic relapses during the follow-up period, with 152 relapses occurring from a state of full remission and 144 relapses from a state of partial remission or sub-threshold state.

Table 2 Course patterns according to the DSM-IV course descriptors stratified by gender (male = 289, female = 62). Butajira cohort. 2012.

Course pattern Overall Male a Female Chi-square P-value
N (%) N % N %
Episodic with inter-episode residual/negative symptoms 135 (37.8) 120 41.5 13 21.0 9.16 0.002
Episodic with no inter-episode residual symptoms 65 (18.2) 45 15.6 18 29.0 6.28 0.012
Continues psychotic symptoms 68 (19.0) 54 18.7 14 22.6 0.50 0.481
Single episode followed by partial remission 38 (10.6) 31 10.7 5 8.1 0.39 0.531
Single episode followed by full remission 42 (11.8) 33 11.4 9 14.5 0.46 0.495
Unspecified/uncertain course pattern 6 (1.7) 4 1.4 3 1.4 1.03 0.310

a Some data on course pattern missing for males.

3.5. Percentage of follow-up time in different clinical states

Over a third, 35.7%, never experienced complete remission, with 29.6% experiencing continuous symptoms—13.7% in episode and 15.9% in sub-threshold (15.9%) states for over 75% of their follow-up time. Only 18.1% achieved complete remission for over 75% of their follow-up time ( Table 3 ). The percentage of patients who had complete remission for more than 75% of the follow-up time showed a continuous increment with the duration of follow-up ( Table 4 ).

Table 3 Percentage of time patients spent in different clinical and treatment status. Butajira cohort. 2012.

  Percentage of total follow-up time (n = 358)
0% 1–5.99% 6–15.99% 16–45.99% 46–75.99% 76–100%
Clinical status
Complete remission (PSR 1,2) 128 (35.7) 25 (7.0) 36 (10.0) 59 (16.4) 46 (22.8) 65 (18.1)
Partial remission (PSR 3,4) 52 (14.5) 30 (8.4) 39 (10.9) 106 (29.5) 75 (20.9) 57 (15.9)
Psychotic episode PSR (5,6) 32 (8.9) 100 (27.9) 62 (17.3) 71 (19.8) 45 (12.5) 49 (13.7)
Percent time on any antipsychotic medication 9 (2.5) 47 (13.1) 60 (16.7) 106 (29.5) 66 (18.4) 71 (19.8)
Antipsychotic treatment status
Percentage time participants taking antipsychotic medication expressed in chlorpromazine equivalent dose
 < 100 mg/d 81 (36.5) 17 (7.7) 22 (9.9) 42 (18.9) 30 (13.5) 30 (13.5)
 100 mg–299.99 mg/d 43 (34.4) 8 (6.4) 13 (10.4) 16 (12.8) 16 (12.8) 29 (23.2)
 > 300 mg/d 0 0 1 (100) 0 0 0

Table 4 Comparative course characteristics of patients with schizophrenia at different times of follow-up. Butajira cohort. 2012.

Course characteristics categories Percentage of patients in course categories Improvement through time
Butajira cohort mean follow-up 3.4 years (%) Butajira cohort mean follow-up 5 years (%) Butajira cohort mean follow-up 10 years (%)
Continuous illness with psychotic episodes or residual symptoms 30.8 30.3 19.0 ↑↑
Psychotic for < 5% of the follow-up time 36.8 36.2 36.8??
Psychotic for > 75% of the follow-up time 1.3 30.3 13.7
Complete remission for > 75% of the follow-up time 5.7 21.2 18.1
On antipsychotic medication for > 75% of the follow-up time 12.9 6.9 36.7 ↑↑
On antipsychotic medication for > 50 + of the follow-up 22.0 (n = 70) 39.4 (n = 121) 36.0 (n = 129)
Never used antipsychotic medication during the follow-up 9.1 4.0 3.34

– None, ↑ some improvement, ↑↑ remarkable improvement.

Over 97.0% of the patients received antipsychotic treatment at least for a month during the follow-up period. Over one third (37.0%) of the participants were prescribed antipsychotic medications for over 75.0% of their follow-up time. Complete remission for over 75.0% of the follow-up time occurred in 13.5% (n = 30) of the participants who received less than 100 mg chlorpromazine equivalent doses and in 23.2% (n = 29) of those who received 100 mg–300 mg chlorpromazine equivalent doses. Antipsychotic medications were fairly well tolerated, with 50.0% (n = 178) showing no EPS, with 20.1% (n = 72) showing mild EPS, 11.2% (n = 40) moderate EPS and 4.7 (n = 17) severe EPS. A further 4.2% (n = 15) had tardive dyskinesia.

3.6. Predictors of clinical remission

In the univariate model, low baseline SANS scores (OR = 0.94, CI = 0.91, 0.97) and low disorganized symptom scores (OR = 0.87, CI = 0.80, 0.95) were associated with achieving full remission for over 50% of the follow-up time. In the fully adjusted model, illness onset after the age of 20 was the only factor that predicted achieving full remission (Table 5 and Table 6).

Table 5 Sociodemographic and clinical factors associated with remission among the Butajira cohort. 2012.

Characteristics   N No (%) in remission for 50% + of the follow-up time P-value Crude OR (95% CI)
Sex Male 296 83 (28.0)   Ref.
Female 62 23 (37.1) 0.156 0.66 (0.37, 1.17)
Age < 20 27 13 (48.2)   Ref.
20–29 136 42 (30.9)   0.41 (.021, 1.11)
30–39 129 37 (28.7)   0.43 (0.19, 1.01)
40 + 60 14 (23.3) 0.023 0.33 (0.13, 0.86)
Marital status (n =) Single 192 53 (27.6)   Ref.
Married 97 36 (37.1)   1.55 (0.92–2.60)
Other 62 16 (25.8) 0.184 0.91 (0.48, 1.75)
Education (n =) Non-literate 189 51 (27.0)   Ref.
Literate 161 55 (34.2) 0.145 1.40 (0.89, 2.22)
Residence (n =) Urban 87 25 (28.7)   Ref.
Rural 271 81 (29.9) 0.838 1.06 (0.62,1.80)
Family and other social support None 158 51 (32.3)   Ref.
Present 147 44 (29.9) 0.658 0.90 (0.55, 1.46)
Religion (n =) Orthodox Christians 80 24 (30.0)   Ref.
Moslem 257 73 (28.4)   0.93 (0.53, 1.60)
Protestant 19 9 (47.4) 0.218 2.1 (0.76, 5.82)
On antipsychotic 50% of follow-up time or more 129 41 (31.8)   1.18 (0.74, 1.88)
  < 50% of the follow-up time 229 65 (28.4) 0.499 Ref.
Sub-type of schizophrenia Paranoid 131 47 (35.9)   Ref.
Disorganized 14 2 (14.3)   0.30 (0.06, 1.39)
Catatonic 27 6 (22.2)   0.51 (0.19, 1.35)
Undifferentiated 132 34 (25.8)   0.62 (0.37, 1.05)
Residual 44 14 (31.8)   0.83 (0.40, 1.73)
Other 10 3 (30.0) 0.326 0.77 (0.19, 3.10)
Speed of onset (n =) Acute within 3 months 180 58 (32.2)   Ref.
Insidious 3–12 months 51 14 (27.5)   0.80 (0.40, 1.59)
Insidious > 12 months 70 12 (17.1) 0.058 0.44 (0.22, 0.87)
Duration of illness (n =), y 2 or less years 90 26 (28.9)   Ref.
More than 2 231 75 (32.5) 0.535 1.18 (0.69, 2.02)
Age at onset (age at first clear recognition of psychotic symptoms) 19 or less 92 36 (39.1)   Ref.
20–45 250 67 (26.8) 0.028 0.57 (0.34, 0.94)
Experiencing EPS Absent 178 58 (32.6)   Ref.
Present 144 39 (27.1) 0.285 0.77 (0.47, 1.25)

Table 6 Association of clinical and social factors with being in remission for 50% or more of the follow-up time. Butajira cohort. 2012.

Characteristics In remission for < 50% of the follow-up period In remission for 50%+ of the follow-up period P-value OR (95% CI)
N, mean, sd N, mean, sd
Negative symptoms 246, 13.2, 7.2 106,10.0, 7.5 < 0.001 0.94 (0.91,0.97)
Disorganized symptoms 247, 3.1, 3.0 106, 2.0, 2.5 0.001 0.87 (0.80,0.95)
Psychotic symptoms 347, 3.9, 2.9 105, 3.6, 3.0 0.333 0.96 (0.89,1.04)
Mean carer burden score at baseline 234, 1.4, 1.0 106, 1.3, 0.9 0.218 0.86 (0.69,1.09)
Mean carer perceived stigma score at baseline 234, 0.65, 0.6 106, 0.62, 0.6 0.711 0.92 (0.61,1.40)

4. Discussion

We report on the long-term course and outcome of schizophrenia in a predominantly rural community from a LMIC setting. Generally, women have a more favorable course of illness compared with men and this is consistent with other reports (Sartorius et al, 1977, Kebede et al, 2005, Alem et al, 2009, and Novick et al, 2012). The majority of the patients in this cohort had at least two relapses during the 10-year follow-up period, with close to 57% of the participants experiencing either episodic course or continuous psychotic symptoms. This is a very high percentage when compared to earlier reports from LMICs ( Harrison et al., 2001 ). The percentage of participants experiencing a single psychotic episode followed by complete remission (11.8%) was also low compared to the WHO study in LMICs and in high-income country sites ( Jablenisky et al., 1992 ).

The percentage of patients experiencing continuous illness in this cohort was also higher than that reported in the WHO study for the LMIC sites ( Jablenisky et al., 1992 ). Our findings are also less favorable than a more recent multi country study that looked into regional differences in the treatment response and short-term outcome of schizophrenia ( Novick et al., 2012 ).

Our results with respect to the clinical course of illness coupled with a high mortality rate ( Teferra et al., 2011 ) indicate that the outcome of schizophrenia in this rural setting is not favorable. Although this poorer outcome might be attributed to the prevalence cohort, the incident cases and those with DUP under two years (n = 90) did not have a much better outcome. Although the number of deaths due to suicide in our sample is lower than has been reported elsewhere ( Wiersma et al., 1998 ), it is higher than that reported in a 10-year follow-up in a LMIC ( Thara et al., 1994 ). Unlike earlier reports from the Butajira cohort (Alem et al, 2009 and Teferra et al, 2011), being on treatment for a longer period of time did not predict remission. In the long-term, negative symptoms, which are predictive of poor prognosis and poor outcome (Kirkpatrick et al, 2001 and Tek et al, 2001) predominate. In line with this, having a low negative symptom score at baseline predicted good outcome and this persisted in the adjusted model. Our finding of late age onset of illness predicting remission is in line with the accepted pattern ( Crumlish et al, 2009 ).

There is an ongoing debate about the course and outcome of schizophrenia. Nevertheless, earlier reports from our cohort (Kebede et al, 2004, Kebede et al, 2005, Teferra et al, 2011, and Shibre et al, 2014) and our current findings are in line with the emerging trend (Patel et al, 2006, Ran et al, 2007, Burns, 2008, and Cohen et al, 2008), which suggests that the outcome of schizophrenia in LMIC might actually be less favorable.

In terms of comparing our findings with studies from Africa, the main studies we could compare our results with were the findings of the IPSS and DOSMD from Nigeria (Sartorius et al, 1977 and Jablensky et al, 1992). The outcome in our sample is demonstrably inferior. However, these studies from Africa were atypical in several respects. (1) 20%–40% of the cases did not have core schizophrenia; (2) Those with core schizophrenia have characteristics that support good prognosis—high proportion married at baseline, and with paranoid and catatonic schizophrenia, predominating; (3) A large proportion were in receipt of treatment; (4) A large proportion of cases dropped out follow-up and there was no clear accounting of their status. Given these considerations, comparison with these studies is difficult. Therefore, our study extends the evidence on the outcome of schizophrenia in Africa. The relevant characteristics that may also contribute to the utility of our study are: first, our sample was much larger than that of the Nigeria sample; second, our sample was recruited through population-based survey thus enhancing representativeness; thirdly, our cohort was predominantly treatment naive at recruitment; fourthly, the duration of follow-up is long and indicative of the more longitudinal course of schizophrenia. Compared to the outcomes we reported after two- and five-years of follow-up (Kebede et al, 2005 and Teferra et al, 2011), the proportion of patients in continuous illness or psychotic episode was lower at 10 years while the proportion of patients achieving remission for over 75% of the follow-up time had increased. Treatment coverage also increased over the years; over 96.0% received antipsychotic medication at least for a month during the follow-up period. This figure is higher than that reported in a longitudinal study in patients with schizophrenia in Europe ( Nielson et al., 2010 ). The concept of tolerability might need a different approach and more careful analysis. In an earlier medication non-adherence study from the same population, factors such as poverty and family support were the most important reasons for non-compliance compared to tolerability indicating that there are more salient factors that might need further study ( Teferra et al., 2013 ). Overall, medications in our cohort were well tolerated, with a relatively low rate of tardive dyskinesia (4.2%). This rate is lower than the 17.0%–30.0% generally reported ( Gershanik and Go´mez Are´valo, 2011 ), even exceeding 30.0% in some studies 30.0% ( Koshino et al., 1992 ). The difference does not appear to be due to the dose of medications in our sample. Doses of medications given to our patients were comparable to studies mentioned above (Koshino et al, 1992, Nielsen et al, 2010, and Gershanik and Go´mez Are´valo, 2011). In most LMICs, first generation antipsychotics remain the mainstay of treatment. Moreover, second generation antipsychotics are expensive, unavailable and are not without serious side effects (Addington et al, 1995 and Bressan et al, 2004).

4.1. Limitations

Over 65.0% of the cohort had illness duration of over two years at enrollment ( Kebede et al., 2003 ). Nonetheless the sub-group of patients with incident schizophrenia or short DUP did not show any better outcome.

4.2. Conclusion

This study fills an important knowledge gap about the long-term course and outcome of schizophrenia in LMICs. Given the large treatment naïve and population-based nature of the cohort, the methodological rigor around ascertainment and diagnosis, as well as the long follow-up period with serial measurement-based assessments, the findings are worthy of note. Given the conflicting evidence regarding clinical course of schizophrenia in LMICs, a definitive population-based incidence study is required. First generation antipsychotic medications are still the only options available in most rural low-income settings. This study provides evidence of relatively low side effects in a real life setting, supporting plans to make these medications available widely as part of integration of mental health services in primary health care.

Role of funding source

The study is funded by the Stanley Medical Research Institute (SMRI) grant. SMRI had no role in the design, analysis, interpretation or reporting of the findings.


TS, AA, DK, AF, GM, and CH contributed to the design of the study, data collection, data analysis and write-up. LJ and GK contributed to the design of the study, and the write up. ST contributed to data collection, analysis and write-up.

Conflict of interest



The Stanley Medical Research Institute (SMRI) is acknowledged for the financial support. All study participants and involved staffs are acknowledged for their support and effort.


  • Abbas et al., 2011 S.M. Abbas, A.Y. Alam, A. Majid. To determine the probable causes of death in an urban slum community of Pakistan among adults 18 years and above by verbal autopsy. J. Pak. Med. Assoc.. 2011;61(3):235-238
  • Addington et al., 1995 D.E. Addington, J.A. Toews, J.M. Addington. Risperidone and tardive dyskinesia: a case report. J. Clin. Psychiatry. 1995;56(10):484-485
  • Alem et al., 2009 A. Alem, D. Kebede, A. Fekadu, et al. Clinical course and outcome of schizophrenia in a predominantly treatment-naive cohort in rural Ethiopia. Schizophr. Bull.. 2009;35(3):646-654 Crossref
  • American Psychiatric Association, 1994 American Psychiatric Association. Diagnostic and statistical manual of mental disorders. IV ed. (American Psychiatric Association, Washington, DC, 1994)
  • Andreasen, 1982 N.C. Andreasen. Negative versus positive schizophrenia: definition and validation. Arch. Gen. Psychiatry. 1982;39:789-794 Crossref
  • Andreasen, 1984 N.C. Andreasen. Scale for the assessment of negative symptoms/scale for the assessment of positive symptoms [manual] University of Iowa Press. (University of Iowa Press, Iowa City. Iowa, 1984)
  • Assefa et al., 2012 D. Assefa, T. Shibre, L. Asher, A. Fekadu. Internalized stigma among patients with schizophrenia in Ethiopia: a cross-sectional facility-based study. BMC Psychiatry. 2012;12:239 Crossref
  • Bressan et al., 2004 R.A. Bressan, H.M. Jones, L.S. Pilowsky. Atypical antipsychotic drugs and tardive dyskinesia: relevance of D2 receptor affinity. J. Psychopharmacol.. 2004;18(1):124-127 Crossref
  • Burns, 2008 J. Burns. Dispelling a myth: developing world poverty, inequality, violence and social fragmentation are not good for outcome in schizophrenia. Afr. J. Psychiatry. 2008;12:200-205
  • Cohen et al., 2008 A. Cohen, V. Patel, R. Thara, O. Gureje. Questioning an axiom: better prognosis for schizophrenia in the developing world?. Schizophr. Bull.. 2008;34(2):229-244
  • Crumlish et al., 2009 N. Crumlish, P. Whitty, M. Clarke, et al. Beyond the critical period: longitudinal study of 8-year outcome in first-episode non-affective psychosis. Br. J. Psychiatry. 2009;194(1):18-24 Crossref
  • Farooq et al., 2009 S. Farooq, M. Large, O. Nielssen, W. Waheed. The relationship between the duration of untreated psychosis and outcome in low-and-middle income countries: a systematic review and meta-analysis. Schizophr. Res.. 2009;109(1–3):15-23 Crossref
  • Fekadu et al., 2006 A. Fekadu, D. Kebede, A. Alem, et al. Clinical outcome in bipolar disorder in a community-based follow-up study in Butajira-Ethiopia. Acta Psychiatr. Scand.. 2006;114:426-434 Crossref
  • Gershanik and Go´mez Are´valo, 2011 O.S. Gershanik, G.J. Go´mez Are´valo. Typical and atypical neuroleptics. Handb. Clin. Neurol.. 2011;100:579-599 Crossref
  • Gureje and Cohen, 2011 O. Gureje, A. Cohen. Differential outcome of schizophrenia: where we are and where we would like to be. Br. J. Psychiatry. 2011;199(3):173-175 Crossref
  • Hamilton, 1960 M. Hamilton. A rating scale for depression. J. Neurol. Neurosurg. Psychiatry. 1960;23:56-62 Crossref
  • Harrison et al., 2001 G. Harrison, K. Hopper, T. Craig, et al. Recovery from psychotic illness: a 15 and 25-year international follow-up study. Br. J. Psychiatry. 2001;178:506-517 Crossref
  • Jablensky et al., 1992 A. Jablensky, N. Sartorius, G. Ernberg, et al. Schizophrenia: manifestations, incidence and course in different cultures. A World Health Organization ten-country study. Psychol. Med. Monogr. Suppl.. 1992;20:1-97 Crossref
  • Judd et al., 2005 L.L. Judd, H.S. Akiskal, P.J. Schettler, et al. Psychosocial disability in the course of bipolar I and II disorders: a prospective, comparative, longitudinal study. Arch. Gen. Psychiatry. 2005;62(12):1322-1330 Crossref
  • Kebede et al., 2003 D. Kebede, A. Alem, T. Shibre, et al. Onset and clinical course of schizophrenia in Butajira-Ethiopia—a community-based study. Soc. Psychiatry Psychiatr. Epidemiol.. 2003;38(11):625-631 Crossref
  • Kebede et al., 2004 D. Kebede, A. Alem, T. Shibre, et al. The sociodemographic correlates of schizophrenia in Butajira, rural Ethiopia. Schizophr. Res.. 2004;69(2–3):133-141 Crossref
  • Kebede et al., 2005 D. Kebede, A. Alem, T. Shibre, et al. Short-term symptomatic and functional outcomes of schizophrenia in Butajira, Ethiopia. Schizophr. Res.. 2005;78(2–3):171-185 Crossref
  • Keller et al., 1987 M.B. Keller, P.W. Lavori, B. Friedman, et al. The longitudinal interval follow-up evaluation. A comprehensive method for assessing outcome in prospective longitudinal studies. Arch. Gen. Psychiatry. 1987;44(6):540-548 Crossref
  • Kirkpatrick et al., 2001 B. Kirkpatrick, R.W. Buchanan, D.E. Ross, et al. A separate disease within the syndrome of schizophrenia. Arch. Gen. Psychiatry. 2001;58(2):165-171 Crossref
  • Koshino et al., 1992 Y. Koshino, S. Madokoro, T. Ito, et al. A survey of tardive dyskinesia in psychiatric inpatients in Japan. Clin. Neuropharmacol.. 1992;15(1):34-43 Crossref
  • Kulhara et al., 2009 P. Kulhara, R. Shah, S. Grover. Is the course and outcome of schizophrenia better in the ‘developing’ world?. Asian J Psychiatr. 2009;2(2):55-62 Crossref
  • Kurihara et al., 2006 T. Kurihara, M. Kato, H. Kashima, et al. Excess mortality of schizophrenia in the developing country of Bali. Schizophr. Res.. 2006;83(1):103-105 Crossref
  • Mojtabai et al., 2001 R. Mojtabai, V.K. Varma, S. Malhotra, et al. Mortality and long-term course in schizophrenia with a poor 2-year course: a study in a developing country. Br. J. Psychiatry. 2001;178(1):71-75 Crossref
  • Murphy and Raman, 1971 H.B. Murphy, A.C. Raman. The chronicity of schizophrenia in indigenous tropical people. Results of a twelve-year follow-up survey in Mauritius. Br. J. Psychiatry. 1971;118(546):489-497 Crossref
  • Nielsen et al., 2010 J. Nielsen, P. le Quach, C. Emborg, et al. 10-year trends in the treatment and outcomes of patients with first-episode schizophrenia. Acta Psychiatr. Scand.. 2010;122(5):356-366 Crossref
  • Novick et al., 2012 D. Novick, J.M. Haro, J. Hong, et al. Regional differences in treatment response and three year course of schizophrenia across the world. J. Psychiatr. Res.. 2012;46(7):856-864 Crossref
  • Ohaeri, 1993 J.U. Ohaeri. Long-term outcome of treated schizophrenia in a Nigerian cohort. Retrospective analysis of 7-year follow-ups. J. Nerv. Ment. Dis.. 1993;181(8):514-516
  • Patel et al., 2006 V. Patel, A. Cohen, R. Thara, O. Gurege. Is the outcome of schizophrenia really better in developing countries?. Rev. Psiquiatr.. 2006;28(2):149-152 Crossref
  • Perkins et al., 2005 D.O. Perkins, H. Gu, K. Boteva, J.A. Lieberman. Relationship between duration of untreated psychosis and outcome in first-episode schizophrenia: a critical review and meta-analysis. Am. J. Psychiatry. 2005;162(10):1785-1804 Crossref
  • Ran et al., 2007 M.S. Ran, E.Y. Chen, Y. Conwell, et al. Mortality in people with schizophrenia in rural China: 10-year cohort study. Br. J. Psychiatry. 2007;190:237-242 Crossref
  • Sartorius and Janca, 1996 N. Sartorius, A. Janca. Psychiatric assessment instruments developed by the World Health Organization. Soc. Psychiatry Psychiatr. Epidemiol.. 1996;31(2):55-69 Crossref
  • Sartorius et al., 1977 N. Sartorius, A. Jablensky, R. Shapiro. Two-year follow-up of the patients included in the WHO International Pilot Study of Schizophrenia. Psychol. Med.. 1977;7:529-541 Crossref
  • Sartorius et al., 1978 N. Sartorius, A. Jablensky, R. Shapiro. Cross-cultural differences in the short-term prognosis of schizophrenic psychoses. Schizophr. Bull.. 1978;4(1):102-113 Crossref
  • Shibre et al., 2002a T. Shibre, D. Kebede, A. Alem, et al. An evaluation of two screening methods to identify cases with schizophrenia and affective disorders in a community survey in rural Ethiopia. Int. J. Soc. Psychiatry. 2002;48(3):200-208 Crossref
  • Shibre et al., 2002b T. Shibre, D. Kebede, A. Alem, S. Kebreab, et al. Neurological soft signs (NSS) in 200 treatment-naive cases with schizophrenia: a community-based study in a rural setting. Nord. J. Psychiatry. 2002;56(6):425-431 Crossref
  • Shibre et al., 2003 T. Shibre, D. Kebede, A. Alem, et al. Schizophrenia: illness impact on family members in a traditional society–rural Ethiopia. Soc. Psychiatry Psychiatr. Epidemiol.. 2003;38(1):27-34 Crossref
  • Shibre et al., 2014 T. Shibre, C. Hanlon, G. Medhin, et al. Suicide and suicide attempts in people with severe mental disorders in Butajira, Ethiopia: 10 year follow-up of a population-based cohort. BMC Psychiatry. 2014;14(150):1-12
  • Soleman et al., 2006 N. Soleman, D. Chandramohan, K. Shibuya. Verbal autopsy: current practices and challenges. Bull. World Health Organ.. 2006;84(3):239-245 Crossref
  • Taylor et al., 2009 D. Taylor, C. Panton, S. Kapur. The Moudsley prescribing guideline. Informa Healthcare 10th ed. (Informa Healthcare is the publisher, London, UK, 2009)
  • Teferra et al., 2011 S. Teferra, T. Shibre, A. Fekadu, G. Medhin, et al. Five-year mortality in a cohort of people with schizophrenia in Ethiopia. BMC Psychiatry. 2011;11:165 10.1186/1471-244X-11-165 Crossref
  • Teferra et al., 2013 S. Teferra, C. Hanlon, T. Beyero, et al. Perspectives on reasons for non-adherence to medication in persons with schizophrenia in Ethiopia: a qualitative study of patients, caregivers and health workers. MBC Psychiatry. 2013;13(168):1-9
  • Tek et al., 2001 C. Tek, B. Kirkpatrick, R.W. Buchanan. A five-year follow up study of deficit and non-deficit schizophrenia. Schizophr. Res.. 2001;49(3):253-260 Crossref
  • Thara and Rajkumar, 1992 R. Thara, S. Rajkumar. Gender differences in schizophrenia. Results of a follow-up study from India. Schizophr. Res.. 1992;7(1):65-70 Crossref
  • Thara et al., 1994 R. Thara, M. Henrietta, A. Joseph, et al. Ten-year course of schizophrenia-the Madras longitudinal study. Acta Psychiatr. Scand.. 1994;90(5):329-336 Crossref
  • Ustün et al., 1997 B. Ustün, W. Compton, D. Mager, T. Babor, O. Baiyewu, S. Chatterji, et al. WHO study on the reliability and validity of alcohol and drug use disorder instruments: overview of methods and results. Drug Alcohol Depend.. 1997;47(3):161-169
  • Van Zelst, 2009 C. Van Zelst. Stigmatization as an environmental risk in schizophrenia: a user perspective. Schizophr. Bull.. 2009;35(2):293-296 Crossref
  • Warshaw et al., 2001 M.G. Warshaw, I. Dyck, J. Allsworth, et al. Maintaining reliability in a long-term psychiatric study: an ongoing inter-rater reliability monitoring program using the longitudinal interval follow-up evaluation. J. Psychiatr. Res.. 2001;35(5):297-305 Crossref
  • Wiersma et al., 1998 D. Wiersma, F.J. Nienhuis, C.J. Slooff, R. Giel. Natural course of schizophrenic disorders: a 15-year follow up of a Dutch incidence cohort. Schizophr. Bull.. 1998;24(1):75-85 Crossref
  • World Health Organization, 1994 World Health Organization. International Statistical Classification of Diseases and Related Health Problems (ICD). (WHO, Geneva, 1994)
  • Young et al., 1978 R.C. Young, J.T. Biggs, V.E. Ziegler, D.A. Meyer. A rating scale for mania: reliability, validity and sensitivity. Br. J. Psychiatry. 1978;133:429-435 Crossref


a Addis Ababa University, Ethiopia

b University of Toronto, Ontario Shores Centre for Mental Health Sciences, Canada

c Addis Ababa University, Aklilu-Lemma Institute of Pathobiology, Ethiopia

d Addis Ababa University, School of Public Health, Addis Ababa, Ethiopia

e WHO Regional Office for Africa, Brazzaville, Congo

f Umeå University, Division of Psychiatry, Sweden

g King's College London, Institute of Psychiatry, Health Services and Population Research Department, Centre for Global Mental Health, London, UK

h King's College London, Institute of Psychiatry, Department of Psychological Medicine, London, UK

lowast Corresponding author at: Addis Ababa University, P.O. Box 9086, Addis Ababa, Ethiopia.