BMJ Open Is Committed to Open Peer Review. As Part of This Commitment We Make the Peer Review History of Every Article We Publish Publicly Available

BMJ Open: first published as 10.1136/bmjopen-2019-034367 on 7 June 2020. Downloaded from

BMJ Open is committed to open peer review. As part of this commitment we make the peer review history of every article we publish publicly available.

When an article is published we post the peer reviewers’ comments and the authors’ responses online. We also post the versions of the paper that were used during peer review. These are the versions that the peer review comments apply to.

The versions of the paper that follow are the versions that were submitted during the peer review process. They are not the versions of record or the final published versions. They should not be cited or distributed as the published version of this manuscript.

BMJ Open is an open access journal and the full, final, typeset and author-corrected version of record of the manuscript is available on our site with no access controls, subscription charges or pay-per-view fees (http://bmjopen.bmj.com).

If you have any questions on BMJ Open’s open peer review process please email [email protected] http://bmjopen.bmj.com/ on September 30, 2021 by guest. Protected copyright. BMJ Open BMJ Open: first published as 10.1136/bmjopen-2019-034367 on 7 June 2020. Downloaded from

Clinical outcomes among individuals with a first episode psychosis attending Butabika National mental referral hospital in Uganda- a longitudinal cohort study. A study protocol

ForJournal: peerBMJ Open review only Manuscript ID bmjopen-2019-034367

Article Type: Protocol

Date Submitted by the 20-Sep-2019 Author:

Complete List of Authors: Akena, Dickens; Makerere University, Psychiatry Semeere, Aggrey; Infectious Diseases Institute Makerere University, Kadama, Philippa ; Makerere University Infectious Diseases Institute Mwesiga, Emanuel ; Makerere University College of Health Sciences Basangwa, David ; Butabika National Referral Hospital Nakku, Juliet; Butabika National Referral and Teaching Mental Hospital, Kampala, Psychiatry Nakasujja, Noeline ; Makerere University College of Health Sciences

Schizophrenia & psychotic disorders < PSYCHIATRY, Public health < Keywords: INFECTIOUS DISEASES, Adult psychiatry < PSYCHIATRY http://bmjopen.bmj.com/

on September 30, 2021 by guest. Protected copyright.

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 1 of 13 BMJ Open

1 2 3

4 BMJ Open: first published as 10.1136/bmjopen-2019-034367 on 7 June 2020. Downloaded from 5 6 7 8 9 I, the Submitting Author has the right to grant and does grant on behalf of all authors of the Work (as defined 10 in the below author licence), an exclusive licence and/or a non-exclusive licence for contributions from authors 11 who are: i) UK Crown employees; ii) where BMJ has agreed a CC-BY licence shall apply, and/or iii) in accordance 12 with the terms applicable for US Federal Government officers or employees acting as part of their official 13 duties; on a worldwide, perpetual, irrevocable, royalty-free basis to BMJ Publishing Group Ltd (“BMJ”) its 14 licensees and where the relevant Journal is co-owned by BMJ to the co-owners of the Journal, to publish the 15 Work in this journal and any other BMJ products and to exploit all rights, as set out in our licence. 16 17 The Submitting Author accepts and understands that any supply made under these terms is made by BMJ to 18 the Submitting Author Forunless you peer are acting as review an employee on behalf only of your employer or a postgraduate 19 student of an affiliated institution which is paying any applicable article publishing charge (“APC”) for Open 20 Access articles. Where the Submitting Author wishes to make the Work available on an Open Access basis (and 21 intends to pay the relevant APC), the terms of reuse of such Open Access shall be governed by a Creative 22 Commons licence – details of these licences and which Creative Commons licence will apply to this Work are set 23 out in our licence referred to above. 24 25 Other than as permitted in any relevant BMJ Author’s Self Archiving Policies, I confirm this Work has not been 26 accepted for publication elsewhere, is not being considered for publication elsewhere and does not duplicate 27 material already published. I confirm all authors consent to publication of this Work and authorise the granting 28 of this licence. 29 30 31 32 33 34 35

36 http://bmjopen.bmj.com/ 37 38 39 40 41 42 43

44 on September 30, 2021 by guest. Protected copyright. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 2 of 13 BMJ Open: first published as 10.1136/bmjopen-2019-034367 on 7 June 2020. Downloaded from 1 2 Clinical outcomes among individuals with a first episode psychosis attending Butabika 3 4 National mental referral hospital in Uganda- a longitudinal cohort study. A study protocal 5 1 2 2 1,4 6 Authors: Dickens Akena , Aggrey Semeere , Philippa Kadama , Emanuel Mwesiga , David 3 3 1 7 Basangwa , Juliet Nakku , Noeline Nakasujja 8 9 Affiliations: 1. Department of Psychiatry, Makerere University College of Health Sciences, 2. 10 Infectious Disease Institute, Makerere University College of Health Sciences, 3. Butabika 11 National Mental referral Hospital, 4. Department of Psychiatry and Mental Health, University of 12 Cape Town. 13 14 Author emails: 15 16 Dickens Akena: [email protected] peer review , [email protected] only 17 18 Aggrey Semeere: [email protected] 19

20 : 21 Philippa Kadama [email protected] 22 23 Emanuel Mwesiga: [email protected] 24 25 David Basangwa: [email protected] 26 27 Juliet Nakku: [email protected] 28 29 Noeline Nakasujja: [email protected] 30 31 Corresponding author: Dickens Akena. P.O.Box 7062, Makerere University. email: 32 http://bmjopen.bmj.com/ 33 [email protected] 34 35 36 37 38 39

40 on September 30, 2021 by guest. Protected copyright. 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 3 of 13 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2019-034367 on 7 June 2020. Downloaded from 1 2 Abstract 3 4 5 Introduction: Psychotic disorders are a major contributor to high morbidity and mortality and 6 significantly contribute to years lived with disability. Predictors of mortality, relapse and barriers 7 to care among patients with first episode psychoses (FEP) have been studied as a means of 8 tailoring interventions to improve patient outcomes. However, little has been done in sub- 9 Saharan Africa to document the incidence and predictors of relapse in patients with FEP. The 10 objectives of this study will be to (a) estimate the cumulative incidence of clinical relapse of 11 psychotic symptoms in patients with FEP, and (b) determine the factors that predict clinical 12 relapse in patients with FEP who showed clinical response to treatment during hospital 13 admission. 14 15 Methods and analysis: We will assemble a cohort of patients with a FEP seen at the Butabika 16 National Mental ReferralFor Hospital peer in Kampala review over a 4-year period.only Participants will be adults 17 ( ≥18 years old), who have received a diagnosis of a psychosis according to the Mini 18 International Neuropsychiatric Instrument (M.I.N.I), with a demonstrable resolution of active 19 symptoms following the use of antipsychotic medications, and deemed clinically stable for a 20 discharge by the health care practitioner. All participants will be required to provide written 21 informed consent. 22 Trained research assistants (RA) will collect Demographic and clinical parameters, age of onset 23 24 of symptoms, diagnostic data using the MINI, physical examination data, symptom severity, 25 level of social and occupational functioning and household income, during the 4 year study 26 period. We will conduct a verbal audit in the event of loss of life. 27 28 Ethics and Dissemination: All participants will provide written informed consent Ethical 29 approvals for the study has been obtained from the Makerere University School of Medicine 30 Research and Ethics Committee and the Uganda National Council for Science and Technology. 31 We will prepare manuscripts for publication in peer reviewed journals

32 http://bmjopen.bmj.com/ 33 34 35 36 37 38 39

40 on September 30, 2021 by guest. Protected copyright. 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 5 of 13 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2019-034367 on 7 June 2020. Downloaded from 1 2 Introduction: Psychotic disorders are a significant contributor of years lived with disability the 3 4 world over, sub-Saharan Africa (SSA) inclusive [1-3]. Psychotic disorders have been shown to 5 predict poor quality of life [4,5], increased health care costs [6,7], and higher mortality mainly 6 due to suicide, accidents, and comorbid infectious diseases [8-10]. Existing literature indicates 7 that psychotic disorders are significantly associated with poor quality of life [4,5], increased 8 health care costs [6,7], and higher mortality due to suicide, accidents, and comorbid infectious 9 diseases [8-10]. Individuals with psychoses are more likely to suffer from comorbid non- 10 psychotic mental illnesses including depression, anxiety and substance misuse disorders [11- 11 13], often referred to as common mental disorders (CMD), as well as other non-communicable 12 diseases such as diabetes mellitus, hypertension, and dyslipidemias [14-16]. 13 14 In high income countries (HIC) a number of longitudinal assessments of patient centered 15 outcomes have been conducted to identify predictors of mortality, relapse and barriers to care 16 among patients withFor first episode peer psychoses review (FEP) [17-19]. Theonly majority of interventions that 17 target improvement in patient outcomes (improvement in medication adherence, stigma 18 reduction and relapse prevention) [20,21] draw their evidence from studies that documented 19 the predictors of outcomes in patients with FEP. A couple of follow-up studies documenting 20 predictors of relapse, and clinical outcomes in patients with FEP have been conducted in SSA 21 [22,23]. Moreover, a number of studies indicate that the presentation of psychoses among 22 individuals of African ancestry is heterogeneous in nature [24-29]. The heterogeneous 23 24 presentation of psychoses in part lead to differences in its incidence, as well as shortages of 25 efficacious treatment options. More work is needed to generate data about the predictors of 26 relapse in SSA— findings from these studies will be critical in designing relapse prevention 27 interventions for individuals with FEP in resource constrained SSA. 28 29 A number of relapse prevention interventions for individuals with a FEP already exist, especially 30 in HIC. However, a number of reasons makes the generalization, and (or) extrapolation of 31 relapse prevention techniques from HIC to SSA inappropriate. First, there is wide variation in

32 the operational definition for first episode psychosis [30] —definitions that could be based on the http://bmjopen.bmj.com/ 33 time of onset of disease or presentation of patients to the health facilities. Secondly, prognosis 34 of FEP and relapse rates have been shown to differ between HIC and low and middle income 35 countries (LMIC) [31,32] —differences in the level of social support have been sighted as some 36 of the reasons [33]. Also, the metabolism of psychotropic medications may vary, in part due to 37 genetic and gender differences [34-36] —variations that are likely to impact on the response 38 (efficacy) to psychotropic medications, and ultimately rates of relapse. Furthermore, variations in 39 the clinical or symptom presentation, conceptualization of psychoses, as well as disease 40 severity in some ethnic groups across multiple populations in the world [25,37-40] may dictate on September 30, 2021 by guest. Protected copyright. 41 the type of treatment that patients receive [41]. Studies that document relapse rates in SSA are 42 43 urgently needed in order for relapse prevention studies to be designed for these specific 44 populations. 45 46 Objectives. The objectives of this study will be to 47 48 (a) Estimate the cumulative incidence (proportion) of a clinical relapse of psychotic symptoms in 49 patients with FEP. Participants will be deemed to have clinically relapsed if they (i) are 50 assessed by a health care worker and deemed to be severely ill requiring admission to the 51 hospital (ii) score ≥20 on the Young Mania Rating Scale (YMRS) [42] or register an increase of 52 25% in Positive and Negative Symptoms of Schizophrenia Scale (PANSS) [43] score from the 53 last measurement 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 6 of 13 BMJ Open: first published as 10.1136/bmjopen-2019-034367 on 7 June 2020. Downloaded from 1 2 (b) Determine the factors that may predict clinical relapse in patients with FEP who showed 3 4 initial clinical response to antipsychotic treatment while under admission in hospital. 5 6 Methods 7 Study setting and design: We are assembling a cohort of patients with a FEP seen at the 8 Butabika National Mental Referral Hospital in Kampala. Patients will be enrolled starting 9 September 2019 and will be followed for 4 years. Participants will have been initially assessed 10 as part of the Neuro-Psychiatric Genetics of African Populations-Psychosis (Neuro-Gap) study 11 [44]. Neuro-GAP is a multi-center study, being conducted in Uganda (at 5 sites: Butabika, Gulu, 12 Naguru, Arua and Mbarara Hospitals), Ethiopia, Kenya and South Africa. While the Neuro-GAP 13 study is cross sectional in nature, we shall follow only participants with a FEP who meet our 14 eligibility criteria (see below). Participants will be enrolled and followed until occurrence of the 15 outcomes (relapse) or death within the 4 year study period. It is estimated that approximately 16 1000 participants willFor be recruited peer for the Neuro-GAPreview study overonly the 4 year period. Individuals 17 with a FEP constitute about 20-30% of study participants recruited in the Neuro-GAP project— 18 these are the individuals from whom we will identify participants for recruitment (criteria below). 19 20 21 Eligibility criteria 22 Inclusion criteria: We shall enroll adults ( ≥18 years old), who have received a diagnosis of a 23 24 psychosis operationally defined as any of: (i) brief psychotic episode; (ii) schizophrenia, or 25 schizophrenia spectrum disorder; or (iii) bipolar affective disorders diagnosed according to the 26 Mini International Neuropsychiatric Instrument (M.I.N.I) version 7.0.2 [45]. Participants will be 27 considered to have a FEP if they have (a) experienced psychotic symptoms for the very first 28 time in their lives, (b) experienced psychotic symptoms before, but are accessing psychiatric 29 care (antipsychotic medications) for the very first time in their lives at the study site, or (c) if 30 already on antipsychotics or used antipsychotic medications for no longer than 6 weeks [30]. 31 Participants need to have demonstrable resolution of active symptoms following the use of

32 antipsychotic medications, and deemed clinically stable for a discharge by the attending health http://bmjopen.bmj.com/ 33 care practitioner. Participants need to be living within a 21 km radius from the hospital. 34 35 Exclusion criteria: Individuals who present with a FEP, but with a substance use disorder as a 36 primary disorder will be excluded. 37 38 Withdrawal criteria: Individuals who commence the use substances in the course of the follow- 39 up and develop a comorbid substance use disorder will be withdrawn from the study.

40 on September 30, 2021 by guest. Protected copyright. 41 Study procedure: 42 43 Identification and consent: Health care workers in the different wards and out-patient 44 departments of Butabika Hospital will be informed about the study. Trained research assistants 45 (RA) shall liaise with the clinicians to identify potential participants for recruitment at the time of 46 admission to the wards. RA’s will then assess participants who are due for discharge (patients 47 with a clinical response to medications) for possible enrollment, and provide them with 48 information about the study. Patients who access care at the out-patient clinic (who may not be 49 admitted) but are eligible will also be approached by the RA for enrollment. The RA’s will invite 50 potential participants to take part in the study, and obtain written informed consent. During the 51 consenting process, the purpose of the study will be described further, the procedures will be 52 explained, and the benefits of taking part in the study will be outlined. Upon demonstrating 53 understanding and being given a chance to ask questions, the potential participant will then 54 provide a witnessed, signed or thumb print consent. 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 7 of 13 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2019-034367 on 7 June 2020. Downloaded from 1 2 We will administer the University of California, San Diego Brief Assessment of Capacity to 3 4 Consent (UBACC) [46] instrument to assess whether the participants have understood the 5 consent process. The UBACC is a 10-item scale comprised of 3 factors that evaluate 6 understanding, appreciation and reasoning. A score of less than 14.5 on three separate 7 occasions indicates that the participant has not understood what the study is all about. Such 8 participants will not be recruited, but will be given a chance to return at a later date for 9 recruitment. Participation in this study is completely voluntary. The patient has the right to 10 withdraw at any time during the study, including at follow-up. We will document the reasons for 11 withdrawal. Each interview is anticipated to last a minimum of 120 minutes (there will be breaks 12 in between the interviews). 13 14 Sample size and power calculation: We used the Leslie-Kish Formula [47] to calculate the 15 sample size for cross sectional studies measuring proportions as the outcome variable. We 16 assume that 25% ofFor recruited peer participants reviewwill relapse within theonly 2-3 year study period. 17 Substituting the figures in the Leslie-Kish Formula (standard deviation of 1.96, and a precision 18 of 5%) yields a sample size of 244 participants. Assuming a 10-20% loss to follow-up of 19 participants, the adjusted sample size will be 292 participants. Our sample size will provide us 20 with enough power (80%) to detect clinically meaningful differences—a score>20 on the YMRS 21 and an increase of 25% of the PANSS score from baseline between the two groups (those who 22 relapse and those who don’t), and identify predictors of a relapse. 23 24 25 Study measurements 26 27 Trained RA will administer the following standardized questionnaires to all participants. 28 29 i. Demographic and clinical parameters: We will (a) document the age, gender, physical 30 address, contact information, marital and employment status, education level, date/month/year 31 of onset of current illness. We will document whether the participant lives within a catchment

32 area of ministry of health supported village health team (VHT) member; we will separately http://bmjopen.bmj.com/ 33 contact the VHT and get their details. We will also request for information from the next of kin for 34 future contact in the event of a loss to follow-up. (b) we will document the age of onset of 35 symptoms, duration of illness before accessing hospital care (acute if it is within 6 months of 36 onset and chronic if it is more than 2 years), whether the participant has received prior treatment 37 for the psychosis (traditional or faith healers), whether or not the patient had a say in the choice 38 of antipsychotics that was prescribed to them. 39 40 ii. We will administer the UBACC [46] to assess participant’s capacity to provide informed on September 30, 2021 by guest. Protected copyright. 41 consent. 42 43 44 iii. The M.I.N.I 7.0.2 [45] psychosis, depression, bipolar affective disorders, substance use 45 disorder, PTSD and generalized anxiety modules will be used to confirm the presence of a 46 psychoses, and other CMD. 47 48 iv. Physical examination for weight and height to calculate the body mass index (a proxy 49 indicator for obesity), a blood pressure measurement to assess for hypertension, and a random 50 blood sugar level (assessed using a glucometer) as a screen for diabetes mellitus. 51 52 v. Symptom severity assessed using the YMRS or PANSS. 53 54 vi. The presence of medication side effects will be assessed using the modified version of the 55 Glasgow Antipsychotic Side Scale (GASS) 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 8 of 13 BMJ Open: first published as 10.1136/bmjopen-2019-034367 on 7 June 2020. Downloaded from 1 2 3 4 vi. The World Health Organization Disability Assessment Schedule Version 2 (WHODAS 2.0) 5 will be used to assess the level of social and occupational functioning of participants. 6 7 vii. Level of house-hold income and health care expenditures will be assessed using the 8 Uganda Bureau of Statistics (UBOS) surveillance guidelines 9 10 viii. Clinical relapse as operationally defined by (a) participant being re-admitted to hospital, (b) 11 participant scoring ≥20 on the YMRS [42] or an increase of 25% in PANSS [43] scores from the 12 last measurement. The YMRS and PANSS will also be used to rate severity of bipolar affective 13 disorders and schizophrenia spectrum disorders respectively, and document a relapse of 14 symptoms. 15 16 ix. We will documentFor mortality peer from any causes review in the participants only using the WHO verbal 17 autopsy scale. 18 19 20 Instrument Baseline Follow-up 21 22 Demographic parameters Yes No 23 UBACC Yes No 24 MINI Yes Yes 25 Physical exam Yes Yes 26 Symptom severity using the YMRS/PANSS Yes Yes 27 GASS Yes Yes 28 WHODAS 2.0 Yes Yes 29 House hold income Yes Yes 30 Clinical relapse No Yes 31 WHO Audit to assess all course mortality No Yes 32 http://bmjopen.bmj.com/ 33 34 35 Participant Follow-up: We shall collect the same data as at baseline from participants at month 36 3 then every six months. In the event that the participant accesses health care and we are not 37 aware, we will review their medical records and abstract information collected during the course 38 of routine care (where they exist). However, all attempts will be made to collect data directly 39 from the participants at all times. Relapse will be defined as any one of: (a) a re-admission of

40 participants to the hospital often based on clinical signs, (b) a score >20 on the YMRS for on September 30, 2021 by guest. Protected copyright. 41 individuals with Bipolar affective disorders at follow-up [42,48], (c) a 25% increase in the total 42 score of the PANSS from baseline in participants with Schizophrenia at follow-up. 43 44 Medical Record Review: There is a possibility that participants will return to access care at the 45 facility and be missed by the RA. They could also be admitted to the facility due to other health 46 complications. We will review participant’ medical charts and abstract information about any 47 admission to the hospital, duration of stay, laboratory parameters, and any other recorded 48 complications. 49 50 Adverse event reporting during participant follow up: We anticipate that this project will 51 have minimal adverse events that are directly related to the study. However, in the event that 52 we observe any adverse events as a result of the use of prescribed medications during clinical 53 54 care or loss of privacy/confidentiality, then we will report it promptly to the relevant regulatory 55 bodies per requirement. Confidentiality could be broken in the event that the RA gets 56 information related to the following: a) participant is suicidal, b) participant threatens to commit a 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 9 of 13 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2019-034367 on 7 June 2020. Downloaded from 1 2 homicide, and c) participants reports a sexual abuse to themselves or other parties. The RA will 3 4 immediately inform the PI about such, and appropriate action will be taken including reporting 5 such cases to the administration of Butabika Hospital. 6 7 Potential risks: There is a potential risk of developing severe psychological distress during the 8 interviews as a result of answering questions that are deemed private by the participant. RA will 9 be trained to identify any of such distress, and the interview will be terminated. Participants may 10 be asked to continue with the interview only if they feel like doing so. Such adversities will be 11 reported to the PI and SOMREC. There is also a risk of having information about participants 12 made available in the public domain. We will guard against this by having all identifying 13 information of the patients locked away in file cabinets and password locked computers. The 14 risk to loss of information is minimal. 15 16 Benefits: There are Forno direct peerbenefits to be review gained by the individuals. only However, the scientific 17 community will be able to get information about the predictors of relapse. 18 19 Community Tracing: In the event that a participant has not appeared for a clinic visit on their 20 scheduled clinic appointment, we will contact them or their appointed person through telephone. 21 If neither the patient nor the contact person can be reached by phone three months from the last 22 date of their scheduled appointment, we will make active attempts to trace the participant at 23 24 their residence by liaising with the VHT based in the same location. In the event that the 25 participant can’t be traced at their place of residence, we will consider them as a potential loss 26 to follow-up. There exists a number of village health team members who provide care to non- 27 mental health clients. We will utilize their knowledge about the village to identify individuals in 28 the cohort. 29 30 31 Data analysis plans:

32 http://bmjopen.bmj.com/ 33 Descriptive analysis: We shall describe the population characteristics by demographics, social 34 and economic strata using simple proportions. Then we shall perform survival analysis using the 35 Aalen-Johansen estimator were death prior to the outcome will be considered a competing risk. 36 For outcome A we shall describe incidence overall at 1 year and 4 years. We shall describe the 37 survival time to relapse for all participants categorizing by psychoses i.e. schizophrenia, 38 schizophrenia spectrum disorders or bipolar affective disorders, considering death as a 39 competing risk. We shall also estimate the incidence rate of relapse for the outcomes above.

40 on September 30, 2021 by guest. Protected copyright. 41 Explanatory analyses: We shall use proportional hazard regression to estimate hazard ratios for 42 43 a clinical relapse (operationally defined above). For this we shall use proportional hazards 44 regression to estimate the hazard ratios for relapse overall. We shall use directed acyclic graphs 45 to define models for evaluation of predictors of relapse. If the power permits, we shall estimate 46 this for each disorder as well. We will document the total length of follow-up period during which 47 a subject is (i) adherent to antipsychotic medication, (ii) in a state of complete remission 48 (defined as having less than the baseline PANSS score at recruitment, and <20 on YMRS) for 49 half the time they are being followed-up , (iii) in partial remission (no reduction in scores from 50 baseline), and (iv) in a psychotic episode (scores on the PANNS and YMRS increase after 51 discharge and never return to baseline). In cases where we have informative censoring, we 52 shall use inverse probability weights from a sample of tracked participants to estimate the 53 outcome. 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 10 of 13 BMJ Open: first published as 10.1136/bmjopen-2019-034367 on 7 June 2020. Downloaded from 1 2 Ethical considerations. We will have a two-layered consent process. (i) Trained RAs will read 3 4 the consent form to the participant, and (ii) then administer the UBACC to ensure that 5 participants have fully understood the rationale for conducting the study and that they are 6 participating well aware of their rights as participants. Only participants who score ≥10 will be 7 enrolled. Ethical approvals for the study has been obtained from the Makerere University School 8 of Medicine Research and Ethics Committee (#REC-REF 2019-033) and the Uganda National 9 Council for Science and Technology (UNCST HS2638). 10 11 A number of ethical considerations are worth pointing out. First, participants may experience 12 severe psychological distress when talking about their illnesses during assessments. We will 13 train our RA to be able to terminate the interviews in the event of severe psychological distress 14 that results from the interview. That said, beyond the distress that will be experienced by 15 participants, this will be a minimal risk study. Secondly, to minimize the risk of loss of privacy 16 and confidentiality, onlyFor the investigatorspeer willreview have access to onlythe study records and test results 17 and the link between personal identifying information and study data. No individual identities will 18 be used in any reports or publications associated with the data from this study. All softcopies of 19 the data will be stored in password locked computers. Hard copies of the questionnaires will be 20 stored in locked file cabinets at the study offices in Butabika National Referral Hospital. 21 22 Data statement: The data set collected from this work will be available for use upon receipt of 23 24 permission from the study team. 25 26 Author contributions: All authors contributed in editing the manuscript and providing 27 constructive feedback. DA Conceptualized, designed and wrote the first draft. SAS contributed 28 to the data analysis component. EM, NN, DB, and JN will train research assistant and provide 29 support supervision during data collection. PK will set up the data bases for analysis. 30 31 Acknowledgement: We acknowledge the editorial work by Prof Dan Stein in the development

32 of the manuscript. http://bmjopen.bmj.com/ 33 34 Funding statement: This study has no funding at the moment. The authors will apply for 35 funding from different agencies to conduct the study. In the interim, the Lead Author will cover 36 the costs of conducting the study out of pocket. 37 38 Patient involvement: No patient involved. This is a research protocol and at the moment, there 39 is no patient involvement in the design of this study protocol. We will include a patient 40 involvement statement at the time of publication, at which point we would have engaged on September 30, 2021 by guest. Protected copyright. 41 patients. 42 43 44 Competing interest: The authors declare no conflict of interest. 45 46 Word count: 3150 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 11 of 13 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2019-034367 on 7 June 2020. Downloaded from 1 2 References 3 4 5 1. Harvey A Whiteford, Louisa Degenhardt, Jürgen Rehm, Amanda J Baxter, Alize J Ferrari, et 6 al. (2013) Global burden of disease attributable to mental and substance use disorders: fi 7 ndings from the Global Burden of Disease Study 2010. The Lancet 8 http://dx.doi.org/10.1016/S0140-6736(13)61611-6 9 10 2. Rössler W, Salize H.J, van Os J, A R-R (2005) Size of burden of schizophrenia and psychotic 11 disorders. Eur Neuropsychopharmacol 15: 399-409. 12 13 3. Assen Jablensky (2000) Epidemiology of schizophrenia: The global burden of disease and 14 disability. European Archives of Psychiatry and Clinical Neuroscience 250: 274-285. 15 16 4. Samuli I. Saarni, ForSatu Viertiö peer , Jonna Perälä review , Seppo Koskinen only , Jouko Lönnqvist , et al. 17 (2010) Quality of life of people with schizophrenia, bipolar disorder and other psychotic 18 disorders. Br J Psychiatry 197: 386-394. doi: 310.1192/bjp.bp.1109.076489. 19 20 5. Erin E Michalak, Lakshmi N Yatham, Raymond W Lam (2005) Quality of life in bipolar 21 disorder: A review of the literature. Health and Quality of Life Outcomes 3: doi:10.1186/1477- 22 7525-1183-1172. 23 24 25 6. Charles Laidia, Prigentd A, Alice Plasd, Marion Leboyera, Guillaume Fond, et al. (2018) 26 Factors associated with direct health carecosts in schizophrenia: Results from theFACE-SZ 27 French dataset. European Neuropsychopharmacology 28: 24–36. 28 29 7. S.Frey (2014) The economic burden of schizophrenia in Germany: A population-based 30 retrospective cohort study using genetic matching. European Psychiatry 29: 479-489. 31 8. Solomon Teferra, Teshome Shibre, Abebaw Fekadu, Girmay Medhin, Asfaw Wakwoya, et al.

32 (2011) Five-year mortality in a cohort of people with schizophrenia in Ethiopia. BMC Psychiatry , http://bmjopen.bmj.com/ 33 11:: 165 doi.org/110.1186/1471-1244X-1111-1165. 34 35 9. Abebaw Fekadu , Girmay Medhin, Derege Kebede, Atalay Alem (2015) Excess mortality in 36 severe mental illness: 10-year population-based cohort study in rural Ethiopia. The British 37 Journal of Psychiatry 206: 289–296. doi: 210.1192/bjp.bp.1114.149112. 38 39 10. Gregory E. Simon, Christine Stewart, Bobbi Jo Yarborough, Frances Lynch, Karen J. 40 Coleman, et al. (2018) Mortality Rates After the First Diagnosis of Psychotic Disorder in on September 30, 2021 by guest. Protected copyright. 41 Adolescents and Young Adults. JAMA Psychiatry doi:10.1001/jamapsychiatry.2017.4437. 42 43 44 11. Chandra Kiran, Suprakash Chaudhury (2016) Prevalence of comorbid anxiety disorders in 45 schizophrenia. Industrial Psychiatry Journal 25: 35-40. doi:10.4103/0972-6748.196045. 46 47 12. Peter F. Buckley, Brian J. Miller, Douglas S. Lehrer, David J. Castle (2009) Psychiatric 48 Comorbidities and Schizophrenia. Schizophrenia Bulletin 35: 383–402, 2009 49 doi:2010.1093/schbul/sbn2135. 50 51 13. Sarah M. Hartz, Carlos N. Pato, Helena Medeiros, Patricia Cavazos-Rehg, Janet L. Sobell, 52 et al. (2014) Comorbidity of Severe Psychotic Disorders With Measures of Substance Use. 53 JAMA Psychiatry 71: 248-240. doi:210.1001/jamapsychiatry.2013.3726 Published online 54 January 1001, 2014. 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 12 of 13 BMJ Open: first published as 10.1136/bmjopen-2019-034367 on 7 June 2020. Downloaded from 1 2 14. Martha Ward, Benjamin Druss (2015) The epidemiology of diabetes in psychotic disorders. 3 4 The Lancet Psychiatry 2: P431-451. 5 6 15. Aniyizhai Annamalai, Urska Kosir, Tek C (2017) Prevalence of obesity and diabetes in 7 patients with schizophrenia. World J Diabetes 8: 390-396 doi: 310.4239/wjd.v4238.i4238.4390. 8 9 16. Vikram Patel, Somnath Chatterji (2015) Integrating Mental Health In Care For 10 Noncommunicable Diseases: An Imperative For Person-Centered Care. HEALTH AFFAIRS 9: 11 1498–1505. 12 13 17. Urs Heilbronner, Myrto Samara, Stefan Leucht, Peter Falkai, Thomas G. Schulze (2016) 14 The Longitudinal Course of Schizophrenia Across the Lifespan: Clinical, Cognitive, and 15 Neurobiological Aspects. Harvard Review of Psychiatry 24: 118-128. 16 doi:110.1097/HRP.0000000000000092.For peer review only 17 18 18. Martin Harrow, James R. Sands, Marshall L. Silverstein, Joseph F. Qoldberg (1997) Course 19 and Outcome for Schizophrenia Versus Other Psychotic Patients: A Longitudinal Study. 20 Schizophrenia Bulletin 23: 287-303. 21 22 19. Erika Jääskeläinen, Marianne Haapea, Nina Rautio, Pauliina Juola, Matti Penttilä, et al. 23 24 (2015) Twenty Years of Schizophrenia Research in the Northern Finland Birth Cohort 1966: A 25 Systematic Review. Schizophrenia Research and Treatment: 1-12. 26 27 20. Xia J, Merinder L.B, Belgamwar M.R (2011) Psychoeducation for schizophrenia (Review). 28 Cochrane Database of Systematic Reviews Art. No.: CD002831. DOI: 29 002810.001002/14651858.CD14002831.pub14651852. 30 31 21. Lene Halling Hastrup, Christian Kronborg, Mette Bertelsen, Pia Jeppesen, Per Jorgensen, et

32 al. (2013) Cost-effectiveness of early intervention in first-episode psychosis: economic http://bmjopen.bmj.com/ 33 evaluation of a randomised controlled trial (the OPUS study). The British Journal of Psychiatry 34 202: 35–41. doi: 10.1192/bjp.bp.1112.112300. 35 36 22. Mahlet Fikreyesus, Matiwos Soboka, Garumma Tolu Feyissa (2016) Psychotic relapse and 37 associated factors among patients attending health services in Southwest Ethiopia: a cross- 38 sectional study. BMC Psychiatry 16. 39 40 23. Atalay Alem, Derege Kebede, Abebaw Fekadu, Teshome Shibre, Daniel Fekadu, et al. on September 30, 2021 by guest. Protected copyright. 41 (2009) Clinical Course and Outcome of Schizophrenia in a Predominantly Treatment-Naive 42 43 Cohort in Rural Ethiopia. Schizophrenia Bulletin 35: 646–654 doi:610.1093/schbul/sbn1029. 44 45 24. McGrath J (2008) Dissecting the Heterogeneity of Schizophrenia Outcomes Schizophrenia 46 Bulletin 34: 247–248, https://doi.org/210.1093/schbul/sbm1133. 47 48 25. Alex Cohen, Ramachandran Padmavati, Maia Hibben, Samuel Oyewusi, Sujit John, et al. 49 (2016) Concepts of madness in diverse settings: a qualitative study from the INTREPID project. 50 BMC Psychiatry 16: 388 DOI 310.1186/s12888-12016-11090-12884. 51 52 26. Hannah E. Jongsma, Charlotte Gayer-Anderson, Antonio Lasalvia, Diego Quattrone, Alice 53 Mulè, et al. (2018) Treated Incidence of Psychotic Disorders in the Multinational EU-GEI Study. 54 JAMA Psychiatry JAMA Psychiatry: 1. 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 13 of 13 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2019-034367 on 7 June 2020. Downloaded from 1 2 27. A. Jablensky, N. Sartorius, G. Ernberg, M. Anker, A. Korten, et al. (1992) Schizophrenia: 3 4 manifestations, incidence and course in different cultures A World Health Organization Ten- 5 Country Study. Psychological Medicine Monograph Supplement 20: 1-97. 6 7 28. Elliot M Goldner, Lorena Hsu, Paul Waraich, Julian M Somers (2002) Prevalence and 8 Incidence Studies of Schizophrenic Disorders: A Systematic Review of the Literature. Can J 9 Psychiatry 47: 833–843. 10 11 29. Sukanta Saha, David Chant, Joy Welham, John McGrath (2005) A Systematic Review of the 12 Prevalence of Schizophrenia. PLoS Med 2: e141 110.1371/journal.pmed.0020141. 13 30. Breitborde NJK, Srihari VH, Woods SW (2009) Review of the operational definition for first- 14 episode psychosiseip. Early intervention in psychiatry 3: 259-265. 15 16 31. Mahlet Fikreyesus,For Matiwos peer Soboka, Garummareview Tolu Feyissa only (2016) Psychotic relapse and 17 associated factors among patients attending health services in Southwest Ethiopia: a cross- 18 sectional study. BMC Psychiatry 16: 354. 19 20 32. N. M. Menzes, T. Arenovich, R. B. Zipursky (2006) A systematic review of longitudinal 21 outcome studies of first-episode psychosis. Psychological Medicine 36: 1349–1362 22 doi:1310.1017/S0033291706007951. 23 24 25 33. Jonathan K. Burns (2012) The Social Determinants of Schizophrenia: An African Journey in 26 Social Epidemiology. Public Health Reviews, : 34 epub ahead of print. 27 28 34. Eva J Brandl, James L Kennedy, Daniel J Müller (2014) Pharmacogenetics of 29 Antipsychotics. CanJPsychiatry 59: 76–88. 30 31 35. Jennie G. Pouget, Shams; TA, Arun K. Tiwari, Daniel J. Müller (2014) Pharmacogenetics

32 and outcome with antipsychotic drugs. Dialogues Clin Neurosci 16: 555-566. http://bmjopen.bmj.com/ 33 34 36. A. Zarina Kraal, Kristen M. Ward, Vicki L. Ellingrod (2017) Sex Differences in Antipsychotic 35 Related Metabolic Functioning in Schizophrenia Spectrum Disorders. Psychopharmacology 36 Bulletin 47: 8–21. 37 38 37. van der Ven E, Bourque F, Joober R, Selten J.P, Malla A.K (2012) Comparing the clinical 39 presentation of first-episode psychosis across different migrant and ethnic minority groups in 40 Montreal, Quebec. Can J Psychiatry 57: 300-308. on September 30, 2021 by guest. Protected copyright. 41 42 43 38. John S. Brekke , Concepcion Barrio (1997) Cross-Ethnic Symptom Differences in 44 Schizophrenia: The Influence of Culture and Minority Status. Schizophrenia Bulletin 23: 305- 45 316,. 46 47 39. Frank Njenga (2007) The concept of mental disorder: an African perspective. World 48 Psychiatry 6: 166-167. 49 50 40. Joanna Teuton, Richard Bentall, Dowrick C (2007) Conceptualizing Psychosis in Uganda 51 The Perspective of Indigenous and Religious Healers. Transcultural Psychiatry 44: 79–114 DOI: 52 110.1177/1363461507074976. 53 54 41. Craig Morgan, Maia Hibben , Oluyomi Esan, Sujit John, Vikram Patel, et al. (2015) 55 Searching for psychosis: INTREPID (1): systems for detecting untreated and first-episode cases 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 14 of 13 BMJ Open: first published as 10.1136/bmjopen-2019-034367 on 7 June 2020. Downloaded from 1 2 of psychosis in diverse settings. Soc Psychiatry Psychiatr Epidemiol 50: 879–893 DOI 3 4 810.1007/s00127-00015-01013-00126. 5 6 42. Young R.C, Biggs J.T, Ziegler V.E, Meyer D.A (1978) A rating scale for mania: reliability, 7 validity and sensitivity. Br J Psychiatry 133: 429-435. 8 9 43. Stanley R. Kay, Abraham Flszbeln, Lewis A. QpJer (1987) The Positive and Negative 10 Syndrome Scale (PANSS) for Schizophrenia. Schizophrenia Bulletin 13: 282-287. 11 12 44. Anne Stevenson, Dickens Akena, Rocky E Stroud, Lukoye Atwoli, Megan M Campbell, et al. 13 (2019) Neuropsychiatric Genetics of African Populations-Psychosis (NeuroGAPPsychosis): a 14 case-control study protocol and GWAS in Ethiopia, Kenya, South Africa and Uganda. BMJ 15 Open: e025469. doi:025410.021136/bmjopen-022018-025469. 16 For peer review only 17 45. Sheehan .D.V, Lecrubier Y, Harnett-Sheehan K (1998) The Mini International 18 Neuropsychiatric Interview (M.I.N.I.): The Development and Validation of a Structured 19 Diagnostic Psychiatric Interview. . J Clin Psychiatry 59: 22-23. 20 21 46. Jeste D.V, Palmer B.W, Appelbaum P.S, Golshan S, Glorioso D, et al. (2007) A new brief 22 instrument for assessing decisional capacity for clinical research. Arch Gen Psychiatry 64: 966- 23 24 974. 25 26 47. Leslie Kish (1965) Survey Sampling,. New York, John Wiley and Sons Inc. 27 28 48. Michael Lukasiewicz, Stephanie Gerard, Adeline Besnard, Bruno Falissard, Elena Perri, et 29 al. (2013) Young Mania Rating Scale: how interpret the numbers? Determination of a severity 30 threshold and of the minimal clinically significant difference in the EMBLEM cohort. International 31 Journal of Methods in Psychiatric Research 22: 46–58.

32 http://bmjopen.bmj.com/ 33 34 35 36 37 38 39

Clinical outcomes among individuals with a first episode psychosis attending Butabika National mental referral hospital in Uganda- a longitudinal cohort study. A study protocol

ForJournal: peerBMJ Open review only Manuscript ID bmjopen-2019-034367.R1

Article Type: Protocol

Date Submitted by the 18-Feb-2020 Author:

Primary Subject Mental health Heading: http://bmjopen.bmj.com/

Secondary Subject Heading: Patient-centred medicine

Schizophrenia & psychotic disorders < PSYCHIATRY, Adult psychiatry < Keywords: PSYCHIATRY, MENTAL HEALTH

on September 30, 2021 by guest. Protected copyright.

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 1 of 15 BMJ Open

1 2 3

36 http://bmjopen.bmj.com/ 37 38 39 40 41 42 43

44 on September 30, 2021 by guest. Protected copyright. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 2 of 15 BMJ Open: first published as 10.1136/bmjopen-2019-034367 on 7 June 2020. Downloaded from 1 2 Clinical outcomes among individuals with a first episode psychosis attending Butabika 3 4 National mental referral hospital in Uganda- a longitudinal cohort study. A study protocal 5 1 2 2 1, 4 6 Authors: Dickens Akena , Aggrey Semeere , Philippa Kadama , Emanuel Mwesiga , David 3 3 1 7 Basangwa , Juliet Nakku , Noeline Nakasujja 8 9 Affiliations: 1. Department of Psychiatry, Makerere University College of Health Sciences, 2. 10 Infectious Disease Institute, Makerere University College of Health Sciences, 3. Butabika 11 National Mental referral Hospital, 4. Department of Psychiatry and Mental Health, University of 12 Cape Town. 13 14 Author emails: 15 16 Dickens Akena: [email protected] peer review , [email protected] only 17 18 Aggrey Semeere: [email protected] 19

40 on September 30, 2021 by guest. Protected copyright. 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 3 of 15 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2019-034367 on 7 June 2020. Downloaded from 1 2 Abstract 3 4 5 Introduction: Psychotic disorders significantly contribute to high morbidity and mortality. In high 6 income countries, the predictors of mortality, relapse and barriers to care among patients with 7 first episode psychoses (FEP) have been studied as a means of tailoring interventions to 8 improve patient outcomes. However, little has been done to document relapse rates and their 9 predictors in patients with FEP in low resourced high disease burdened sub-Saharan Africa. 10 11 Objective: We shall estimate the rates of relapse of psychotic symptoms and the factors that 12 predict them in patients with FEP over 4years. 13 14 Methods and analysis: We will assemble a cohort of patients with a FEP seen at the Butabika 15 National Mental Referral Hospital in Kampala over 4years. Participants will be adults ( ≥18 years 16 old), who have receivedFor a diagnosis peer of a psychosisreview according only to the Mini International 17 Neuropsychiatric Instrument (M.I.N.I), with a demonstrable resolution of active symptoms 18 following the use of antipsychotic medications, and deemed clinically stable for a discharge by 19 the health care practitioner. All participants will be required to provide written informed consent. 20 Trained research assistants (RA) will collect Demographic and clinical parameters, age of onset 21 of symptoms, diagnostic data using the MINI, physical examination data, symptom severity, 22 level of social and occupational functioning and household income, during the 4 year study 23 24 period. We will conduct a verbal audit in the event of loss of life. We shall perform survival 25 analysis using the Aalen-Johansen estimator, and describe the population characteristics by 26 demographics, social and economic strata using simple proportions. 27 28 Ethics and Dissemination: All participants will provide written informed consent Ethical 29 approvals for the study has been obtained from the Makerere University School of Medicine 30 Research and Ethics Committee and the Uganda National Council for Science and Technology. 31 We will prepare manuscripts for publication in peer reviewed journals

32 http://bmjopen.bmj.com/ 33 34 35 36 37 38 39

40 on September 30, 2021 by guest. Protected copyright. 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 5 of 15 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2019-034367 on 7 June 2020. Downloaded from 1 2 Introduction: Psychotic disorders are a significant contributor of years lived with disability the 3 4 world over, sub-Saharan Africa (SSA) inclusive [1-3]. Psychotic disorders have been shown to 5 predict poor quality of life [4 5], increased health care costs [6 7], and higher mortality mainly 6 due to suicide, accidents, and comorbid infectious diseases [8-10]. Moreover existing literature 7 indicates that psychotic disorders are significantly associated with poor quality of life [4 5], and 8 increased health care costs [6 7]. Individuals with psychoses are more likely to suffer from 9 comorbid non-psychotic mental illnesses including depression, anxiety and substance misuse 10 disorders [11-13], often referred to as common mental disorders (CMD), as well as other non- 11 communicable diseases such as diabetes mellitus, hypertension, and dyslipidemias [14-16]. 12 13 In high income countries (HIC) a number of longitudinal assessments of patient centered 14 outcomes have been conducted to identify predictors of mortality, relapse and barriers to care 15 among patients with first episode psychoses (FEP) [17-19]. The majority of interventions that 16 target improvementFor in patient peer outcomes (improvement review in medication only adherence, stigma 17 reduction and relapse prevention) [20 21] draw their evidence from studies that documented 18 the predictors of outcomes in patients with FEP. A couple of follow-up studies documenting 19 predictors of relapse, and clinical outcomes in patients with FEP have been conducted in SSA 20 [22 23]. Moreover, a number of studies indicate that the presentation of psychoses among 21 individuals of African ancestry is heterogeneous in nature [24-29]. The heterogeneous 22 presentation of psychoses in part lead to differences in its incidence, as well as shortages of 23 24 efficacious treatment options. More work is needed to generate data about the predictors of 25 relapse in SSA— findings from these studies will be critical in designing relapse prevention 26 interventions for individuals with FEP in resource constrained SSA. 27 28 A number of relapse prevention interventions for individuals with a FEP already exist, especially 29 in HIC. However, a number of reasons makes the generalization, and (or) extrapolation of 30 relapse prevention techniques from HIC to SSA inappropriate. First, there is wide variation in 31 the operational definition for first episode psychosis [30] —definitions that could be based on the

32 time of onset of disease or presentation of patients to the health facilities. Secondly, prognosis http://bmjopen.bmj.com/ 33 of FEP and relapse rates have been shown to differ between HIC and low and middle income 34 countries (LMIC) [31 32] —differences in the level of social support have been sighted as some 35 of the reasons [33]. Also, the metabolism of psychotropic medications may vary, in part due to 36 genetic and gender differences [34-36] —variations that are likely to impact on the response 37 (efficacy) to psychotropic medications, and ultimately rates of relapse. Furthermore, variations in 38 the clinical or symptom presentation, conceptualization of psychoses, as well as disease 39 severity in some ethnic groups across multiple populations in the world [25 37-40] may dictate 40 the type of treatment that patients receive [41]. Studies that document relapse rates in SSA are on September 30, 2021 by guest. Protected copyright. 41 urgently needed in order for relapse prevention studies to be designed for these specific 42 43 populations. 44 45 Objectives. 46 47 (a) The primary objective of the study will be to estimate the cumulative rates of relapse 48 (proportion) of a clinical relapse of psychotic symptoms in patients with FEP. The primary 49 outcome will be a clinical relapse defined as (i) re-admission to a hospital after a clinical 50 assessment by a health care worker who deems the patient to be severely ill requiring 51 admission to the hospital (ii) score ≥20 on the Young Mania Rating Scale (YMRS) [42] or 52 register an increase of 25% in Positive and Negative Symptoms of Schizophrenia Scale 53 (PANSS) [43] score from the last measurement 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 6 of 15 BMJ Open: first published as 10.1136/bmjopen-2019-034367 on 7 June 2020. Downloaded from 1 2 (b) The secondary objective is to determine the factors that may predict clinical relapse in 3 4 patients with FEP who showed initial clinical response to antipsychotic treatment while under 5 admission in hospital. Based on previous literature [44-46], we hypothesize that poor adherence 6 to antipsychotic medications (measured using the medication adherence rating scale (MARS) 7 for psychosis [47] will be the main predictor of relapse in the study population. 8 9 Study outcomes: 10 (a) The primary outcome will be a clinical relapse (defined above, clinical acumen or by use of 11 rating scales). 12 13 (b) The secondary outcome will be all cause mortality (assessed using the verbal WHO autopsy 14 scale). 15 16 For peer review only 17 Methods 18 Study setting and design: We are assembling a cohort of patients with a FEP seen at the 19 Butabika National Mental Referral Hospital in Kampala. Butabika National Mental Referral 20 Hospital is a 600 bed hospital located 13km east of Kampala city (the capital city of Uganda with 21 a population of 3.5million) The hospital has 3 acute admission wards, 3 convalescent wards 22 (housing patients with less acute symptoms and ready for discharge), One male and one female 23 24 sick ward (where individuals with physical illnesses are admitted), an alcohol and drug unit, a 25 child and adolescent unit and a private wing. It has a medical out-patient that provides a service 26 to HIV/AIDS patients, a dental clinic and a general out-patients clinic. The out-patients clinics 27 operates week days from 9 am – 5 pm, and at a minimum. Each of the units (in-patient and out- 28 patient) is run by team of Psychiatrists, medical officers, Psychiatric Clinical Officers, Psychiatric 29 Nurses, clinical psychologists and psychiatric social workers. 30 31 Patients will be enrolled starting May 2020 and will be followed for 4 years. Participants will

32 have been initially assessed as part of the Neuro-Psychiatric Genetics of African Populations- http://bmjopen.bmj.com/ 33 Psychosis (Neuro-Gap) study [48]. Neuro-GAP is a multi-center study, being conducted in 34 Uganda (at 5 sites: Butabika, Gulu, Naguru, Arua and Mbarara Hospitals), Ethiopia, Kenya and 35 South Africa. While the Neuro-GAP study is cross sectional in nature, we shall follow only 36 participants with a FEP who meet our eligibility criteria (see below). Participants will be enrolled 37 and followed until occurrence of the outcomes (relapse) or death within the 4 year study period. 38 It is estimated that approximately 1000 participants will be recruited for the Neuro-GAP study 39 over the 4 year period. Individuals with a FEP constitute about 20-30% of study participants 40 recruited in the Neuro-GAP project—these are the individuals from whom we will identify on September 30, 2021 by guest. Protected copyright. 41 participants for recruitment (criteria below). 42 43 44 Eligibility criteria 45 Inclusion criteria: We shall enroll adults ( ≥18 years old), who have received a diagnosis of a 46 psychosis operationally defined as any of: (i) brief psychotic episode; (ii) schizophrenia, or 47 schizophrenia spectrum disorder; or (iii) bipolar affective disorders diagnosed according to the 48 Mini International Neuropsychiatric Instrument (M.I.N.I) version 7.0.2 [49]. Participants will be 49 considered to have a FEP if they have (a) experienced psychotic symptoms for the very first 50 time in their lives, (b) experienced psychotic symptoms before, but are accessing psychiatric 51 care (antipsychotic medications) for the very first time in their lives at the study site, or (c) if 52 already on antipsychotics or used antipsychotic medications for no longer than 6 weeks [30]. 53 Participants need to have demonstrable resolution of active symptoms following the use of 54 antipsychotic medications, and deemed clinically stable for a discharge by the attending health 55 care practitioner. Participants need to be living within a 21 km radius from the hospital. 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 7 of 15 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2019-034367 on 7 June 2020. Downloaded from 1 2 3 4 Exclusion criteria: Individuals who present with a FEP, but with a substance use disorder as a 5 primary disorder will be excluded. 6 7 Study procedure: 8 Identification and consent: Health care workers in the different wards and out-patient 9 departments of Butabika Hospital will be informed about the study. Trained research assistants 10 (RA) shall liaise with the clinicians to identify potential participants for recruitment at the time of 11 admission to the wards. RA’s will then assess participants who are due for discharge (patients 12 with a clinical response to medications) for possible enrollment, and provide them with 13 information about the study. Patients who access care at the out-patient clinic (who may not be 14 admitted) but are eligible will also be approached by the RA for enrollment. The RA’s will invite 15 potential participants to take part in the study, and obtain written informed consent. During the 16 consenting process,For the purpose peer of the study review will be described only further, the procedures will be 17 explained, and the benefits of taking part in the study will be outlined. Upon demonstrating 18 understanding and being given a chance to ask questions, the potential participant will then 19 provide a witnessed, signed or thumb print consent. 20 21 We will administer the University of California, San Diego Brief Assessment of Capacity to 22 Consent (UBACC) [50] instrument to assess whether the participants have understood the 23 24 consent process. The UBACC will be translated into Luganda, the commonly spoken local 25 language at the study site. The UBACC is a 10-item scale comprised of 3 factors that evaluate 26 understanding, appreciation and reasoning. It has been used in the Ugandan setting for the 27 Neuro-GAP project, although it is yet to be culturally adapted for use in these settings. A score 28 of less than 14.5 on three separate occasions indicates that the participant has not understood 29 what the study is all about. Such participants will not be recruited, but will be given a chance to 30 return at a later date for recruitment (within a week). We will record the number of participants 31 who fail the UBACC at baseline and can’t be recruited and those who do so after being invited a

32 week later. We will ask participants to provide us with their demographic information (age, http://bmjopen.bmj.com/ 33 gender and education level) and examine whether there are significant differences between 34 participants who are able to consent and those who are not. Participation in this study is 35 completely voluntary. The patient has the right to withdraw at any time during the study, 36 including at follow-up. We will document the reasons for withdrawal. Each interview is 37 anticipated to last a minimum of 120 minutes —there will be 2-3 breaks at 40 minute intervals in 38 between the interviews. 39 40 Sample size and power calculation: on September 30, 2021 by guest. Protected copyright. 41 a) For the first objective (proportion of participants who will relapse), we used the Leslie-Kish 42 43 Formula [51] to calculate the sample size for cross sectional studies measuring proportions as 44 the outcome variable. We assume that 25% of recruited participants will relapse within the 2-3 45 year study period. Substituting the figures in the Leslie-Kish Formula (standard deviation of 46 1.96, and a precision of 5%) yields a sample size of 244 participants. Assuming a 10-20% loss 47 to follow-up of participants, the adjusted sample size will be 292 participants. Our sample size 48 will provide us with enough power (80%) to detect clinically meaningful differences—a score>20 49 on the YMRS and an increase of 25% of the PANSS score from baseline between the two 50 groups (those who relapse and those who don’t), and identify predictors of a relapse. 51 52 b) For the second objective of examining the predictors of relapse, using a conservative 53 prevalence estimate for a major predictor, medication non-adherence as 40% in a population of 54 patients with Schizophrenia [44-46], with 211 patients we should able to estimate hazard ratios 55 as high as 2.0 or greater within 2 years with an estimated relapse rate of 25% per year. 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 8 of 15 BMJ Open: first published as 10.1136/bmjopen-2019-034367 on 7 June 2020. Downloaded from 1 2 Thus, our sample size of 292 participants should be adequate to answer both objectives. 3 4 5 6 Study measurements 7 8 Trained RA will administer the following standardized questionnaires to all participants. All study 9 questionnaires will be translated into Luganda, the commonly spoken local language at the 10 study site. 11 12 i. Demographic and clinical parameters: We will (a) document the age, gender, physical 13 address, contact information, marital and employment status, education level, date/month/year 14 of onset of current illness. We will document whether the participant lives within a catchment 15 area of ministry of health supported village health team (VHT) member; we will separately 16 contact the VHT andFor get their peer details. We willreview also request for only information from the next of kin for 17 future contact in the event of a loss to follow-up. (b) we will document the age of onset of 18 symptoms, duration of illness before accessing hospital care (acute if it is within 6 months of 19 onset and chronic if it is more than 2 years), whether the participant has received prior treatment 20 for the psychosis (traditional or faith healers), whether or not the patient had a say in the choice 21 of antipsychotics that was prescribed to them. 22 23 24 ii. We will administer the UBACC [50] to assess participant’s capacity to provide informed 25 consent. 26 27 iii. The M.I.N.I 7.0.2 [49] psychosis, depression, bipolar affective disorders, substance use 28 disorder, PTSD and generalized anxiety modules will be used to confirm the presence of a 29 psychoses, and other CMD. The MINI has been used in multiple Ugandan study settings 30 including the Neuro-GAP project, although it is yet to be validated for use in these settings 31

32 iv. Physical examination for weight and height to calculate the body mass index (a proxy http://bmjopen.bmj.com/ 33 indicator for obesity), a blood pressure measurement to assess for hypertension, and a random 34 blood sugar level (assessed using a glucometer) as a screen for diabetes mellitus. 35 36 v. Symptom severity assessed using the YMRS or PANSS. Both the YMRS and PANSS have 37 been used in Ugandan study settings although it is yet to be validated for use in these settings 38 39 vi. The presence of medication side effects will be assessed using the modified version of the 40 Glasgow Antipsychotic Side Scale (GASS) on September 30, 2021 by guest. Protected copyright. 41 42 43 vi. The World Health Organization Disability Assessment Schedule Version 2 (WHODAS 2.0) 44 will be used to assess the level of social and occupational functioning of participants. There is 45 limited data about the use of the WHODAS in Ugandan populations. 46 47 vii. Level of house-hold income and health care expenditures will be assessed using the 48 Uganda Bureau of Statistics (UBOS) surveillance guidelines 49 50 viii. Clinical relapse as operationally defined by (a) participant being re-admitted to hospital, (b) 51 participant scoring ≥20 on the YMRS [42] or an increase of 25% in PANSS [43] scores from the 52 last measurement. The YMRS and PANSS will also be used to rate severity of bipolar affective 53 disorders and schizophrenia spectrum disorders respectively, and document a relapse of 54 symptoms. 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 9 of 15 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2019-034367 on 7 June 2020. Downloaded from 1 2 ix. Adherence to antipsychotics will be measured using the medication adherence rating scale 3 4 (MARS) for psychosis [47] 5 6 x. We will document mortality from any causes in the participants using the WHO verbal autopsy 7 scale. 8 9 10 Instrument Baseline Follow-up 11 Demographic parameters Yes No 12 UBACC Yes No 13 MINI Yes Yes 14 Physical exam Yes Yes 15 Symptom severity using the YMRS/PANSS Yes Yes 16 GASS For peer review onlyYes Yes 17 18 WHODAS 2.0 Yes Yes 19 House hold income Yes Yes 20 Clinical relapse No Yes 21 Medication adherence using the MARS No Yes 22 WHO Audit to assess all course mortality No Yes 23 24 Pilot data collection: We will conduct a pilot of data collection among 10 participants, and use 25 semi-structured questionnaires to document clarity of study questionnaires, barriers to 26 implementation and ways of circumventing the barriers. We will then make appropriate changes 27 28 to the study protocol and submit an amendment to the IRB before commencement of the study 29 if need be. 30 31 Participant Follow-up: We shall collect the same data as at baseline from participants at month 32 3 then every six months. In the event that the participant accesses health care and we are not http://bmjopen.bmj.com/ 33 aware, we will review their medical records and abstract information collected during the course 34 of routine care (where they exist). However, all attempts will be made to collect data directly 35 from the participants at all times. Relapse will be defined as any one of: (a) a re-admission of 36 participants to the hospital often based on clinical signs, (b) a score >20 on the YMRS for 37 38 individuals with Bipolar affective disorders at follow-up [42 52], (c) a 25% increase in the total 39 score of the PANSS from baseline in participants with Schizophrenia at follow-up.

40 on September 30, 2021 by guest. Protected copyright. 41 Medical Record Review: There is a possibility that participants will return to access care at the 42 facility and be missed by the RA. They could also be admitted to the facility due to other health 43 complications. We will review participant’ medical charts and abstract information about any 44 admission to the hospital, duration of stay, laboratory parameters, and any other recorded 45 complications. 46 47 Adverse event reporting during participant follow up: We anticipate that this project will 48 have minimal adverse events that are directly related to the study. However, in the event that 49 we observe any adverse events as a result of the use of prescribed medications during clinical 50 care or loss of privacy/confidentiality, then we will report it promptly to the relevant regulatory 51 bodies per requirement. Confidentiality could be broken in the event that the RA gets 52 information related to the following: a) participant is suicidal, b) participant threatens to commit a 53 homicide, and c) participants reports a sexual abuse to themselves or other parties. The RA will 54 immediately inform the PI about such, and appropriate action will be taken including reporting 55 such cases to the administration of Butabika Hospital. 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 10 of 15 BMJ Open: first published as 10.1136/bmjopen-2019-034367 on 7 June 2020. Downloaded from 1 2 3 4 Potential risks: There is a potential risk of developing severe psychological distress during the 5 interviews as a result of answering questions that are deemed private by the participant. RA will 6 be trained to identify any of such distress, and the interview will be terminated. Participants may 7 be asked to continue with the interview only if they feel like doing so. Such adversities will be 8 reported to the PI and SOMREC. There is also a risk of having information about participants 9 made available in the public domain. We will guard against this by having all identifying 10 information of the patients locked away in file cabinets and password locked computers. The 11 risk to loss of information is minimal. 12 13 Benefits: There are no direct monetary benefits to be gained by the individuals. However, 14 participants will receive regular assessments for their symptoms every 6 months for four years. 15 The scientific community will be able to get information about the predictors of relapse. 16 For peer review only 17 Community Tracing: In the event that a participant has not appeared for a clinic visit on their 18 scheduled clinic appointment, we will contact them or their appointed person through telephone. 19 If neither the patient nor the contact person can be reached by phone three months from the last 20 date of their scheduled appointment, we will make active attempts to trace the participant at 21 their residence by liaising with the VHT based in the same location. In the event that the 22 participant can’t be traced at their place of residence, we will consider them as a potential loss 23 24 to follow-up. There exists a number of village health team members who provide care to non- 25 mental health clients. We will utilize their knowledge about the village to identify individuals in 26 the cohort. 27 28 Data analysis plans: 29 30 Descriptive analysis: We shall describe the population characteristics by demographics, social 31 and economic strata using simple proportions. Then we shall perform survival analysis using the

32 Aalen-Johansen estimator were death prior to the outcome will be considered a competing risk. http://bmjopen.bmj.com/ 33 For outcome (a) we shall describe relapse rates at 1 year and 4 years. We shall describe the 34 survival time to relapse for all participants categorizing by psychoses i.e. schizophrenia, 35 schizophrenia spectrum disorders or bipolar affective disorders, considering death as a 36 competing risk. For outcome (b), we also estimate the predictors of relapse rates using 37 generalized linear equation estimates 38 39 Explanatory analyses: We shall use proportional hazard regression to estimate hazard ratios for 40 a clinical relapse (operationally defined above). For this we shall use proportional hazards on September 30, 2021 by guest. Protected copyright. 41 regression to estimate the hazard ratios for relapse overall. We shall use directed acyclic graphs 42 43 to define models for evaluation of predictors of relapse. If the power permits, we shall estimate 44 this for each disorder as well. We will document the total length of follow-up period during which 45 a subject is (i) adherent to antipsychotic medication, (ii) in a state of complete remission 46 (defined as having less than the baseline PANSS score at recruitment, and <20 on YMRS) for 47 half the time they are being followed-up , (iii) in partial remission (no reduction in scores from 48 baseline), and (iv) in a psychotic episode (scores on the PANNS and YMRS increase after 49 discharge and never return to baseline). In cases where we have informative censoring, we 50 shall use inverse probability weights from a sample of tracked participants to estimate the 51 outcome. 52 53 Ethical considerations. We will have a two-layered consent process. (i) Trained RAs will read 54 the consent form to the participant, and (ii) then administer the UBACC to ensure that 55 participants have fully understood the rationale for conducting the study and that they are 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 11 of 15 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2019-034367 on 7 June 2020. Downloaded from 1 2 participating well aware of their rights as participants. Only participants who score ≥10 will be 3 4 enrolled. Ethical approvals for the study has been obtained from the Makerere University School 5 of Medicine Research and Ethics Committee (#REC-REF 2019-033) and the Uganda National 6 Council for Science and Technology (UNCST HS2638). 7 8 A number of ethical considerations are worth pointing out. Beyond the distress that may l be 9 experienced by participants, this will be a minimal risk study. To minimize the risk of loss of 10 privacy and confidentiality, only the investigators will have access to the study records and test 11 results and the link between personal identifying information and study data. No individual 12 identities will be used in any reports or publications associated with the data from this study. All 13 softcopies of the data will be stored in password locked computers. Hard copies of the 14 questionnaires will be stored in locked file cabinets at the study offices in Butabika National 15 Referral Hospital. 16 For peer review only 17 Data statement: The data set collected from this work will be available for use upon receipt of 18 permission from the study team. 19 20 Author contributions: All authors contributed in editing the manuscript and providing 21 constructive feedback. DA Conceptualized, designed and wrote the first draft. SAS contributed 22 to the data analysis component. EM, NN, DB, and JN will train research assistant and provide 23 24 support supervision during data collection. PK will set up the data bases for analysis. 25 26 Acknowledgement: We acknowledge the editorial work by Prof Dan Stein in the development 27 of the manuscript. 28 29 Funding statement: This study has no funding at the moment. The authors will apply for 30 funding from different agencies to conduct the study. In the interim, the Lead Author will cover 31 the costs of conducting the study out of pocket.

32 http://bmjopen.bmj.com/ 33 Patient involvement: During the pilot phase of the project, we will engage with patients and 34 document any barriers to implementation of the study procedures. 35 36 Competing interest: The authors declare no conflict of interest. 37 38 Word count: 4225 39

32 http://bmjopen.bmj.com/ 33 8. Solomon Teferra, Teshome Shibre, Abebaw Fekadu, et al. Five-year mortality in a cohort of 34 35 people with schizophrenia in Ethiopia. BMC Psychiatry , 2011;11::165 36 doi.org/10.1186/471-244X-11-165 37 38 9. Abebaw Fekadu , Girmay Medhin, Derege Kebede, et al. Excess mortality in severe mental 39 illness: 10-year population-based cohort study in rural Ethiopia. The British Journal of 40 Psychiatry 2015;206:289–96. doi: 10.1192/bjp.bp.114.149112 on September 30, 2021 by guest. Protected copyright. 41 42 43 10. Gregory E. Simon, Christine Stewart, Bobbi Jo Yarborough, et al. Mortality Rates After the 44 First Diagnosis of Psychotic Disorder in Adolescents and Young Adults. JAMA 45 Psychiatry 2018;doi:10.1001/jamapsychiatry.2017.4437 46 47 11. Chandra Kiran, Suprakash Chaudhury. Prevalence of comorbid anxiety disorders in 48 schizophrenia. Industrial Psychiatry Journal 2016;25(1):35-40. doi:10.4103/0972- 49 50 6748.196045. 51 52 12. Peter F. Buckley, Brian J. Miller, Douglas S. Lehrer, et al. Psychiatric Comorbidities and 53 Schizophrenia. Schizophrenia Bulletin 2009;35(2):383–402, 2009 54 doi:10.1093/schbul/sbn135 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 13 of 15 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2019-034367 on 7 June 2020. Downloaded from 1 2 3 13. Sarah M. Hartz, Carlos N. Pato, Helena Medeiros, et al. Comorbidity of Severe Psychotic 4 Disorders With Measures of Substance Use. JAMA Psychiatry 2014;71(3):248-0. 5 doi:10.1001/jamapsychiatry.2013.3726 Published online January 1, 2014 6 7 14. Martha Ward, Benjamin Druss. The epidemiology of diabetes in psychotic disorders. The 8 Lancet Psychiatry 2015;2(5): P431-51 9 10 11 15. Aniyizhai Annamalai, Urska Kosir, Tek C. Prevalence of obesity and diabetes in patients 12 with schizophrenia. World J Diabetes 2017;8(8):390-96 doi: 10.4239/wjd.v8.i8.390 13 14 16. Vikram Patel, Somnath Chatterji. Integrating Mental Health In Care For Noncommunicable 15 Diseases: An Imperative For Person-Centered Care. HEALTH AFFAIRS 2015;9:1498– 16 505 For peer review only 17 18 19 17. Urs Heilbronner, Myrto Samara, Stefan Leucht, et al. The Longitudinal Course of 20 Schizophrenia Across the Lifespan: Clinical, Cognitive, and Neurobiological Aspects. 21 Harvard Review of Psychiatry 2016;24(2):118-28. doi:10.1097/HRP.0000000000000092. 22 23 18. Martin Harrow, James R. Sands, Marshall L. Silverstein, et al. Course and Outcome for 24 Schizophrenia Versus Other Psychotic Patients: A Longitudinal Study. Schizophrenia 25 26 Bulletin 1997;23(2):287-303 27 28 19. Erika Jääskeläinen, Marianne Haapea, Nina Rautio, et al. Twenty Years of Schizophrenia 29 Research in the Northern Finland Birth Cohort 1966: A Systematic Review. 30 Schizophrenia Research and Treatment 2015:1-12 31 32 20. Xia J, Merinder L.B, Belgamwar M.R. Psychoeducation for schizophrenia (Review). http://bmjopen.bmj.com/ 33 34 Cochrane Database of Systematic Reviews 2011(6):Art. No.: CD002831. DOI: 35 10.1002/14651858.CD002831.pub2 36 37 21. Lene Halling Hastrup, Christian Kronborg, Mette Bertelsen, et al. Cost-effectiveness of early 38 intervention in first-episode psychosis: economic evaluation of a randomised controlled 39 trial (the OPUS study). The British Journal of Psychiatry 2013;202:35–41. doi: 40 10.1192/bjp.bp.112.112300 on September 30, 2021 by guest. Protected copyright. 41 42 43 22. Mahlet Fikreyesus, Matiwos Soboka, Garumma Tolu Feyissa. Psychotic relapse and 44 associated factors among patients attending health services in Southwest Ethiopia: a 45 cross-sectional study. BMC Psychiatry 2016;16(354 DOI 10.1186/s12888-016-1076-2) 46 47 23. Atalay Alem, Derege Kebede, Abebaw Fekadu, et al. Clinical Course and Outcome of 48 49 Schizophrenia in a Predominantly Treatment-Naive Cohort in Rural Ethiopia. 50 Schizophrenia Bulletin 2009;35(3):646–54 doi:10.1093/schbul/sbn029 51 52 24. McGrath J. Dissecting the Heterogeneity of Schizophrenia Outcomes Schizophrenia Bulletin 53 2008;34(2):247–48, https://doi.org/10.1093/schbul/sbm133 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 14 of 15 BMJ Open: first published as 10.1136/bmjopen-2019-034367 on 7 June 2020. Downloaded from 1 2 3 25. Alex Cohen, Ramachandran Padmavati, Maia Hibben, et al. Concepts of madness in diverse 4 settings: a qualitative study from the INTREPID project. BMC Psychiatry 2016;16:388 5 DOI 10.1186/s12888-016-1090-4 6 7 26. Hannah E. Jongsma, Charlotte Gayer-Anderson, Antonio Lasalvia, et al. Treated Incidence of 8 Psychotic Disorders in the Multinational EU-GEI Study. JAMA Psychiatry 2018;JAMA 9 Psychiatry(75):1 10 11 12 27. A. Jablensky, N. Sartorius, G. Ernberg, et al. Schizophrenia: manifestations, incidence and 13 course in different cultures A World Health Organization Ten-Country Study. 14 Psychological Medicine Monograph Supplement 1992;20:1-97 15 16 28. Elliot M Goldner,For Lorena peerHsu, Paul Waraich, review et al. Prevalence only and Incidence Studies of 17 Schizophrenic Disorders: A Systematic Review of the Literature. Can J Psychiatry 18 19 2002;47:833–43 20 21 29. Sukanta Saha, David Chant, Joy Welham, et al. A Systematic Review of the Prevalence of 22 Schizophrenia. PLoS Med 2005;2(5):e141 10.1371/journal.pmed.0020141 23 24 30. Breitborde NJK, Srihari VH, Woods SW. Review of the operational definition for first- 25 26 episode psychosiseip. Early intervention in psychiatry 2009;3(4):259-65 doi: 27 10.1111/j.1751-7893.2009.00148.x[published Online First: Epub Date]|. 28 29 31. Mahlet Fikreyesus, Matiwos Soboka, Garumma Tolu Feyissa. Psychotic relapse and 30 associated factors among patients attending health services in Southwest Ethiopia: a 31 cross-sectional study. BMC Psychiatry 2016;16:354

32 http://bmjopen.bmj.com/ 33 34 32. N. M. Menzes, T. Arenovich, R. B. Zipursky. A systematic review of longitudinal outcome 35 studies of first-episode psychosis. Psychological Medicine 2006;36:1349–62 36 doi:10.017/S0033291706007951 37 38 33. Jonathan K. Burns. The Social Determinants of Schizophrenia: An African Journey in Social 39 Epidemiology. Public Health Reviews, 2012:34 epub ahead of print.

40 on September 30, 2021 by guest. Protected copyright. 41 42 34. Eva J Brandl, James L Kennedy, Daniel J Müller. Pharmacogenetics of Antipsychotics. 43 CanJPsychiatry 2014;59(2):76–88 44 45 35. Jennie G. Pouget, Shams; TA, Arun K. Tiwari, et al. Pharmacogenetics and outcome with 46 antipsychotic drugs. Dialogues Clin Neurosci 2014;16:555-66 47 48 49 36. A. Zarina Kraal, Kristen M. Ward, Vicki L. Ellingrod. Sex Differences in Antipsychotic 50 Related Metabolic Functioning in Schizophrenia Spectrum Disorders. 51 Psychopharmacology Bulletin 2017;47(2):8–21. 52 53 37. van der Ven E, Bourque F, Joober R, et al. Comparing the clinical presentation of first- 54 episode psychosis across different migrant and ethnic minority groups in Montreal, 55 Quebec. Can J Psychiatry 2012;57(5):300-8 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 15 of 15 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2019-034367 on 7 June 2020. Downloaded from 1 2 3 38. John S. Brekke , Concepcion Barrio. Cross-Ethnic Symptom Differences in Schizophrenia: 4 The Influence of Culture and Minority Status. Schizophrenia Bulletin 1997;23(2):305-16, 5 6 39. Frank Njenga. The concept of mental disorder: an African perspective. World Psychiatry 7 2007;6(3):166-67 8 9 40. Joanna Teuton, Richard Bentall, Dowrick C. Conceptualizing Psychosis in Uganda The 10 11 Perspective of Indigenous and Religious Healers. Transcultural psychiatry 12 2007;44(1):79–114 DOI: 10.1177/1363461507074976 13 14 41. Craig Morgan, Maia Hibben , Oluyomi Esan, et al. Searching for psychosis: INTREPID (1): 15 systems for detecting untreated and first-episode cases of psychosis in diverse settings. 16 Soc PsychiatryFor Psychiatr peer Epidemiol review 2015;50:879–93 DOIonly 10.1007/s00127-015-1013-6 17 18 19 42. Young R.C, Biggs J.T, Ziegler V.E, et al. A rating scale for mania: reliability, validity and 20 sensitivity. Br J Psychiatry 1978;133:429-35 21 22 43. Stanley R. Kay, Abraham Flszbeln, Lewis A. QpJer. The Positive and Negative Syndrome 23 Scale (PANSS) for Schizophrenia. Schizophrenia Bulletin 1987;13(2):282-87 24 25 26 44. Patricia L. Gilbert, M. Jackuelyn Harris, Lou Ann McAdams, et al. Neuroleptic Withdrawal 27 in Schizophrenic Patients. A Review of the Literature. Arch Gen Psychiatry 1995;52:173- 28 88 29 30 45. Delbert Robinson, Margaret G. Woerner, Jose Ma. J. Alvir, et al. Predictors of Relapse 31 Following Response From a First Episode of Schizophrenia or Schizoaffective Disorder. 32 Arch Gen Psychiatry 1999;56(3):241-47. doi:10.1001/archpsyc.56.3.241 http://bmjopen.bmj.com/ 33 34 35 46. Gunnar Morken, Jan H Widen, Rolf W Grawe. Non-adherence to antipsychotic medication, 36 relapse and rehospitalisation in recent-onset schizophrenia. BMC Psychiatry , 37 2008;3(32):doi:10.1186/471-244X-8-32 38 39 47. Thompson K, Kulkarni J, Sergejew A.A. Reliability and validity of a new Medication 40 Adherence Rating Scale (MARS) for the psychoses. Schizophr Res 2000;5(42):3 on September 30, 2021 by guest. Protected copyright. 41 42 43 48. Anne Stevenson, Dickens Akena, Rocky E Stroud, et al. Neuropsychiatric Genetics of 44 African Populations-Psychosis (NeuroGAPPsychosis): a case-control study protocol and 45 GWAS in Ethiopia, Kenya, South Africa and Uganda. BMJ open 2019:e025469. 46 doi:10.1136/bmjopen-2018-69 47 48 49 49. Sheehan .D.V, Lecrubier Y, Harnett-Sheehan K. The Mini International Neuropsychiatric 50 Interview (M.I.N.I.): The Development and Validation of a Structured Diagnostic 51 Psychiatric Interview. . J Clin Psychiatry 1998;59:22-23 52 53 50. Jeste D.V, Palmer B.W, Appelbaum P.S, et al. A new brief instrument for assessing 54 decisional capacity for clinical research. Arch Gen Psychiatry 2007;64(8):966-74. 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 16 of 15 BMJ Open: first published as 10.1136/bmjopen-2019-034367 on 7 June 2020. Downloaded from 1 2 3 51. Leslie Kish. Survey Sampling,. New York, 1965;John Wiley and Sons Inc. 4 5 52. Michael Lukasiewicz, Stephanie Gerard, Adeline Besnard, et al. Young Mania Rating Scale: 6 how interpret the numbers? Determination of a severity threshold and of the minimal 7 clinically significant difference in the EMBLEM cohort. International Journal of Methods 8 in Psychiatric Research 2013;22(1):46–58 9 10 11 12 13 14 15 16 For peer review only 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31

32 http://bmjopen.bmj.com/ 33 34 35 36 37 38 39

Clinical outcomes among individuals with a first episode psychosis attending Butabika National mental referral hospital in Uganda- a longitudinal cohort study. A study protocal for a longitudinal cohort study.

ForJournal: peerBMJ Open review only Manuscript ID bmjopen-2019-034367.R2

Article Type: Protocol

Date Submitted by the 04-May-2020 Author:

Primary Subject Mental health Heading: http://bmjopen.bmj.com/

Secondary Subject Heading: Patient-centred medicine

Schizophrenia & psychotic disorders < PSYCHIATRY, Adult psychiatry < Keywords: PSYCHIATRY, MENTAL HEALTH

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 1 of 15 BMJ Open

1 2 3

36 http://bmjopen.bmj.com/ 37 38 39 40 41 42 43

44 on September 30, 2021 by guest. Protected copyright. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 2 of 15 BMJ Open: first published as 10.1136/bmjopen-2019-034367 on 7 June 2020. Downloaded from 1 2 Clinical outcomes among individuals with a first episode psychosis attending Butabika 3 4 National mental referral hospital in Uganda- a longitudinal cohort study. A study protocal 5 for a longitudinal cohort study. 6 1 2 2 1, 4 7 Authors: Dickens Akena , Aggrey Semeere , Philippa Kadama , Emanuel Mwesiga , David 3 3 1 8 Basangwa , Juliet Nakku , Noeline Nakasujja 9 10 Affiliations: 1. Department of Psychiatry, Makerere University College of Health Sciences, 2. 11 Infectious Disease Institute, Makerere University College of Health Sciences, 3. Butabika 12 National Mental referral Hospital, 4. Department of Psychiatry and Mental Health, University of 13 Cape Town. 14 15 Author emails: 16 For peer review only 17 Dickens Akena: [email protected] , [email protected] 18 19 Aggrey Semeere: [email protected] 20 21 : 22 Philippa Kadama [email protected] 23 24 Emanuel Mwesiga: [email protected] 25 26 David Basangwa: [email protected] 27 28 Juliet Nakku: [email protected] 29 30 Noeline Nakasujja: [email protected] 31

32 http://bmjopen.bmj.com/ 33 Corresponding author: Dickens Akena. P.O.Box 7062, Makerere University. email: 34 [email protected] 35 36 37 38 39

40 on September 30, 2021 by guest. Protected copyright. 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 3 of 15 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2019-034367 on 7 June 2020. Downloaded from 1 2 Abstract 3 4 5 Introduction: Psychotic disorders significantly contribute to high morbidity and mortality. In high 6 income countries, the predictors of mortality, relapse and barriers to care among patients with 7 first episode psychoses (FEP) have been studied as a means of tailoring interventions to 8 improve patient outcomes. However, little has been done to document relapse rates and their 9 predictors in patients with FEP in low resourced, high disease burdened sub-Saharan Africa. 10 11 Objective: We shall estimate the rates of relapse of psychotic symptoms and the factors that 12 predict them in patients with FEP over 4 years. 13 14 Methods and analysis: We will assemble a cohort of patients with a FEP seen at the Butabika 15 National Mental Referral Hospital in Kampala over a 4 year period. Participants will be adults 16 ( ≥18 years old), whoFor have received peer a diagnosis review of a psychosis only according to the Mini 17 International Neuropsychiatric Instrument (M.I.N.I), with a demonstrable resolution of active 18 symptoms following the use of antipsychotic medications, and deemed clinically stable for a 19 discharge by the health care practitioner. All participants will be required to provide written 20 informed consent. Trained research assistants (RA) will collect Demographic and clinical 21 parameters, age of onset of symptoms, diagnostic data using the MINI, physical examination 22 data, symptom severity, level of social and occupational functioning and household income, 23 24 during the 4 year study period. We will conduct a verbal audit in the event of loss of life. We 25 shall perform survival analysis using the Aalen-Johansen estimator, and describe the population 26 characteristics by demographics, social and economic strata using simple proportions. 27 28 Ethics and Dissemination: All participants will provide written informed consent. Ethical 29 approvals for the study has been obtained from the Makerere University School of Medicine 30 Research and Ethics Committee and the Uganda National Council for Science and Technology. 31 Findings will be published in peer reviewed journals.

32 http://bmjopen.bmj.com/ 33 34 35 36 37 38 39

40 on September 30, 2021 by guest. Protected copyright. 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 5 of 15 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2019-034367 on 7 June 2020. Downloaded from 1 2 Introduction: Psychotic disorders are a significant contributor of years lived with disability the 3 4 world over, sub-Saharan Africa (SSA) inclusive [1-3]. Psychotic disorders have been shown to 5 predict poor quality of life [4 5], increased health care costs [6 7], and higher mortality mainly 6 due to suicide, accidents, and comorbid infectious diseases [8-10]. Moreover existing literature 7 indicates that psychotic disorders are significantly associated with poor quality of life [4 5], and 8 increased health care costs [6 7]. Individuals with psychoses are more likely to suffer from non- 9 psychotic mental illnesses comorbidity including depression, anxiety and substance misuse 10 disorders [11-13], often referred to as common mental disorders (CMD), as well as other non- 11 communicable diseases such as diabetes mellitus, hypertension, and dyslipidemias [14-16]. 12 13 In high income countries (HIC) a number of longitudinal assessments of patient centered 14 outcomes have been conducted to identify predictors of mortality, relapse and barriers to care 15 among patients with first episode psychoses (FEP) [17-19]. The majority of interventions that 16 target improvementFor in patient peer outcomes (improvement review in medication only adherence, stigma 17 reduction and relapse prevention) [20 21] draw their evidence from studies that documented 18 the predictors of outcomes in patients with FEP. A couple of follow-up studies documenting 19 predictors of relapse, and clinical outcomes in patients with FEP have been conducted in SSA 20 [22 23]. Moreover, a number of studies indicate that the presentation of psychoses among 21 individuals of African ancestry is heterogeneous in nature [24-29]. The heterogeneous 22 presentation of psychoses in part lead to differences in its incidence, as well as shortages of 23 24 efficacious treatment options. More work is needed to generate data about the predictors of 25 relapse in SSA— findings from these studies will be critical in designing relapse prevention 26 interventions for individuals with FEP in resource constrained SSA. 27 28 Relapse prevention interventions for individuals with a FEP already exist, especially in HIC. 29 However, a number of reasons makes the generalization, and (or) extrapolation of relapse 30 prevention techniques from HIC to SSA inappropriate. First, there is wide variation in the 31 operational definition for first episode psychosis [30] —definitions that could be based on the

32 time of onset of disease or presentation of patients to the health facilities. Secondly, prognosis http://bmjopen.bmj.com/ 33 of FEP and relapse rates have been shown to differ between HIC and low and middle income 34 countries (LMIC) [31 32] —differences in the level of social support have been sighted as some 35 of the reasons [33]. Also, the metabolism of psychotropic medications may vary, in part due to 36 genetic and gender differences [34-36] —variations that are likely to impact on the response 37 (efficacy) to psychotropic medications, and ultimately rates of relapse. Furthermore, variations in 38 the clinical or symptom presentation, conceptualization of psychoses, as well as disease 39 severity in some ethnic groups across multiple populations in the world [25 37-40] may dictate 40 the type of treatment that patients receive [41]. Studies that document relapse rates in SSA are on September 30, 2021 by guest. Protected copyright. 41 urgently needed in order for relapse prevention studies to be designed for these specific 42 43 populations. 44 45 Objectives. 46 47 (a) The primary objective of the study will be to estimate the cumulative rates of relapse 48 (proportion) of a clinical relapse of psychotic symptoms in patients with FEP. The primary 49 outcome will be a clinical relapse defined as (i) re-admission to a hospital after a clinical 50 assessment by a health care worker who deems the patient to be severely ill requiring 51 admission to the hospital (ii) score ≥20 on the Young Mania Rating Scale (YMRS) [42] or 52 register an increase of 25% in Positive and Negative Symptoms of Schizophrenia Scale 53 (PANSS) [43] score from the last measurement 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 6 of 15 BMJ Open: first published as 10.1136/bmjopen-2019-034367 on 7 June 2020. Downloaded from 1 2 (b) The secondary objective will be to determine the factors that may predict clinical relapse in 3 4 patients with FEP who showed initial clinical response to antipsychotic treatment while under 5 admission in hospital. Based on previous literature [44-46], we hypothesize that poor adherence 6 to antipsychotic medications (measured using the medication adherence rating scale (MARS) 7 for psychosis [47] will be the main predictor of relapse in the study population. 8 9 Study outcomes: 10 (a) The primary outcome will be a clinical relapse (defined above, clinical acumen or by use of 11 rating scales). 12 13 (b) The secondary outcome will be all cause mortality (assessed using the verbal WHO autopsy 14 scale). 15 16 Methods For peer review only 17 Study setting and design: We are assembling a cohort of patients with a FEP seen at the 18 Butabika National Mental Referral Hospital in Kampala. Butabika National Mental Referral 19 Hospital is a 600 bed hospital located 13km east of Kampala city (the capital city of Uganda with 20 a population of 3.5million) The hospital has 3 acute admission wards, 3 convalescent wards 21 (housing patients with less acute symptoms and ready for discharge), one male and one female 22 sick ward (where individuals with physical illnesses are admitted), an alcohol and drug unit, a 23 24 child and adolescent unit and a private wing. It has a medical out-patient that provides a service 25 to HIV/AIDS patients, a dental clinic and a general out-patients clinic. The out-patient clinics 26 operates week days from 9 am – 5 pm, and at a minimum. Each of the units (in-patient and out- 27 patient) is run by team of psychiatrists, medical officers, psychiatric clinical officers, psychiatric 28 nurses, clinical psychologists and psychiatric social workers. 29 30 Patients will be enrolled starting May 2020 and will be followed for 4 years. Participants will 31 have been initially assessed as part of the Neuro-Psychiatric Genetics of African Populations-

32 Psychosis (Neuro-Gap) study [48]. Neuro-GAP is a multi-center study, being conducted in http://bmjopen.bmj.com/ 33 Uganda (at 5 sites: Butabika, Gulu, Naguru, Arua and Mbarara Hospitals), Ethiopia, Kenya and 34 South Africa. While the Neuro-GAP study is cross sectional in nature, we shall follow only 35 participants with a FEP who meet our eligibility criteria (see below). Participants will be enrolled 36 and followed until occurrence of the outcomes (relapse) or death within the 4 year study period. 37 It is estimated that approximately 1000 participants will be recruited for the Neuro-GAP study 38 over the 4 year period. Individuals with a FEP constitute about 20-30% of study participants 39 recruited in the Neuro-GAP project—these are the individuals from whom we will identify 40 participants for recruitment (criteria below). on September 30, 2021 by guest. Protected copyright. 41 42 43 Eligibility criteria 44 Inclusion criteria: We shall enroll adults ( ≥18 years old), who have received a diagnosis of a 45 psychosis operationally defined as any of: (i) brief psychotic episode; (ii) schizophrenia, or 46 schizophrenia spectrum disorder; or (iii) bipolar affective disorders diagnosed according to the 47 Mini International Neuropsychiatric Instrument (M.I.N.I) version 7.0.2 [49]. Participants will be 48 considered to have a FEP if they have (a) experienced psychotic symptoms for the very first 49 time in their lives, (b) experienced psychotic symptoms before, but are accessing psychiatric 50 care (antipsychotic medications) for the very first time in their lives at the study site, or (c) if 51 already on antipsychotics or used antipsychotic medications for no longer than 6 weeks [30]. 52 Participants need to have demonstrable resolution of active symptoms following the use of 53 antipsychotic medications, and deemed clinically stable for a discharge by the attending health 54 care practitioner. Participants need to be living within a 21 km radius from the hospital. 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 7 of 15 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2019-034367 on 7 June 2020. Downloaded from 1 2 Exclusion criteria: Individuals who present with a FEP, but with a substance use disorder as a 3 4 primary disorder will be excluded. 5 6 Study procedure: 7 Identification and consent: Health care workers in the different wards and out-patient 8 departments of Butabika Hospital will be informed about the study. Trained research assistants 9 (RA) shall liaise with the clinicians to identify potential participants for recruitment at the time of 10 admission to the wards. RA’s will then assess participants who are due for discharge (patients 11 with a clinical response to medications) for possible enrollment, and provide them with 12 information about the study. Patients who access care at the out-patient clinic (who may not be 13 admitted) but are eligible will also be approached by the RA for enrollment. The RA’s will invite 14 potential participants to take part in the study, and obtain written informed consent. During the 15 consenting process, the purpose of the study will be described further, the procedures will be 16 explained, and the benefitsFor of peer taking part inreview the study will be onlyoutlined. Upon demonstrating 17 understanding and being given a chance to ask questions, the potential participant will then 18 provide a witnessed, signed or thumb print consent. 19 20 We will administer the University of California, San Diego Brief Assessment of Capacity to 21 Consent (UBACC) [50] instrument to assess whether the participants have understood the 22 consent process. The UBACC will be translated into Luganda, the commonly spoken local 23 24 language at the study site. The UBACC is a 10-item scale comprised of 3 factors that evaluate 25 understanding, appreciation and reasoning. It has been used in the Ugandan setting for the 26 Neuro-GAP project, although it is yet to be culturally adapted for use in these settings. A score 27 of less than 14.5 on three separate occasions indicates that the participant has not understood 28 what the study is all about. Such participants will not be recruited, but will be given a chance to 29 return at a later date for recruitment (within a week). We will record the number of participants 30 who fail the UBACC at baseline and can’t be recruited and those who do so after being invited a 31 week later. We will ask participants to provide us with their demographic information (age,

32 gender and education level) and examine whether there are significant differences between http://bmjopen.bmj.com/ 33 participants who are able to consent and those who are not. Participation in this study is 34 completely voluntary. The patient has the right to withdraw at any time during the study, 35 including at follow-up. We will document the reasons for withdrawal. Each interview is 36 anticipated to last a minimum of 120 minutes —there will be 2-3 breaks at 40 minute intervals in 37 between the interviews. 38 39 Sample size and power calculation: 40 a) For the first objective (proportion of participants who will relapse), we used the Leslie-Kish on September 30, 2021 by guest. Protected copyright. 41 Formula [51] to calculate the sample size for cross sectional studies measuring proportions as 42 43 the outcome variable. We assume that 25% of recruited participants will relapse within the 2-3 44 year study period. Substituting the figures in the Leslie-Kish Formula (standard deviation of 45 1.96, and a precision of 5%) yields a sample size of 244 participants. Assuming a 10-20% loss 46 to follow-up of participants, the adjusted sample size will be 292 participants. Our sample size 47 will provide us with enough power (80%) to detect clinically meaningful differences—a score>20 48 on the YMRS and an increase of 25% of the PANSS score from baseline between the two 49 groups (those who relapse and those who don’t), and identify predictors of a relapse. 50 51 b) For the second objective of examining the predictors of relapse, using a conservative 52 prevalence estimate for a major predictor, medication non-adherence as 40% in a population of 53 patients with Schizophrenia [44-46], with 211 patients we should able to estimate hazard ratios 54 as high as 2.0 or greater within 2 years with an estimated relapse rate of 25% per year. 55 Thus, our sample size of 292 participants should be adequate to answer both objectives. 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 8 of 15 BMJ Open: first published as 10.1136/bmjopen-2019-034367 on 7 June 2020. Downloaded from 1 2 Study measurements 3 4 5 Trained RA will administer the following standardized questionnaires to all participants. All study 6 questionnaires will be translated into Luganda, the commonly spoken local language at the 7 study site. 8 9 i. Demographic and clinical parameters: We will (a) document the age, gender, physical 10 address, contact information, marital and employment status, education level, date/month/year 11 of onset of current illness. We will document whether the participant lives within a catchment 12 area of ministry of health supported village health team (VHT) member; we will separately 13 contact the VHT and get their details. We will also request for information from the next of kin for 14 future contact in the event of a loss to follow-up. (b) we will document the age of onset of 15 symptoms, duration of illness before accessing hospital care (acute if it is within 6 months of 16 onset and chronic ifFor it is more peer than 2 years), review whether the participant only has received prior treatment 17 for the psychosis (traditional or faith healers), whether or not the patient had a say in the choice 18 of antipsychotics that was prescribed to them. 19 20 ii. We will administer the UBACC [50] to assess participant’s capacity to provide informed 21 consent. 22 23 24 iii. The M.I.N.I 7.0.2 [49] psychosis, depression, bipolar affective disorders, substance use 25 disorder, PTSD and generalized anxiety modules will be used to confirm the presence of a 26 psychoses, and other CMD. The MINI has been used in multiple Ugandan study settings 27 including the Neuro-GAP project, although it is yet to be validated for use in these settings 28 29 iv. Physical examination for weight and height to calculate the body mass index (a proxy 30 indicator for obesity), a blood pressure measurement to assess for hypertension, and a random 31 blood sugar level (assessed using a glucometer) as a screen for diabetes mellitus.

32 http://bmjopen.bmj.com/ 33 v. Symptom severity assessed using the YMRS or PANSS. Both the YMRS and PANSS have 34 been used in Ugandan study settings although it is yet to be validated for use in these settings 35 36 vi. The presence of medication side effects will be assessed using the modified version of the 37 Glasgow Antipsychotic Side Scale (GASS) 38 39 vi. The World Health Organization Disability Assessment Schedule Version 2 (WHODAS 2.0) 40 will be used to assess the level of social and occupational functioning of participants. There is on September 30, 2021 by guest. Protected copyright. 41 limited data about the use of the WHODAS in Ugandan populations. 42 43 44 vii. Level of house-hold income and health care expenditures will be assessed using the 45 Uganda Bureau of Statistics (UBOS) surveillance guidelines 46 47 viii. Clinical relapse as operationally defined by (a) participant being re-admitted to hospital, (b) 48 participant scoring ≥20 on the YMRS [42] or an increase of 25% in PANSS [43] scores from the 49 last measurement. The YMRS and PANSS will also be used to rate severity of bipolar affective 50 disorders and schizophrenia spectrum disorders respectively, and document a relapse of 51 symptoms. 52 53 ix. Adherence to antipsychotics will be measured using the medication adherence rating scale 54 (MARS) for psychosis [47] 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 9 of 15 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2019-034367 on 7 June 2020. Downloaded from 1 2 x. We will document mortality from any causes in the participants using the WHO verbal autopsy 3 4 scale. 5 6 7 Instrument Baseline Follow-up 8 Demographic parameters Yes No 9 UBACC Yes No 10 MINI Yes Yes 11 Physical exam Yes Yes 12 Symptom severity using the YMRS/PANSS Yes Yes 13 GASS Yes Yes 14 WHODAS 2.0 Yes Yes 15 House hold income Yes Yes 16 For peer review only 17 Clinical relapse No Yes 18 Medication adherence using the MARS No Yes 19 WHO Audit to assess all course mortality No Yes 20 21 Pilot data collection: We will conduct a pilot of data collection among 10 participants, and use 22 semi-structured questionnaires to document clarity of study questionnaires, barriers to 23 implementation and ways of circumventing the barriers. We will then make appropriate changes 24 25 to the study protocol and submit an amendment to the IRB before commencement of the study 26 if need be. 27 28 Participant Follow-up: We shall collect the same data as at baseline from participants at month 29 3 then every six months. In the event that the participant accesses health care and we are not 30 aware, we will review their medical records and abstract information collected during the course 31 of routine care (where they exist). However, all attempts will be made to collect data directly 32 from the participants at all times. Relapse will be defined as any one of: (a) a re-admission of http://bmjopen.bmj.com/ 33 participants to the hospital often based on clinical signs, (b) a score >20 on the YMRS for 34 individuals with Bipolar affective disorders at follow-up [42 52], (c) a 25% increase in the total 35 score of the PANSS from baseline in participants with Schizophrenia at follow-up. 36 37 38 Medical Record Review: There is a possibility that participants will return to access care at the 39 facility and be missed by the RA. They could also be admitted to the facility due to other health complications. We will review participant’ medical charts and abstract information about any 40 on September 30, 2021 by guest. Protected copyright. 41 admission to the hospital, duration of stay, laboratory parameters, and any other recorded 42 complications. 43 44 Adverse event reporting during participant follow up: We anticipate that this project will 45 have minimal adverse events that are directly related to the study. However, in the event that 46 we observe any adverse events as a result of the use of prescribed medications during clinical 47 care or loss of privacy/confidentiality, then we will report it promptly to the relevant regulatory 48 bodies per requirement. Confidentiality could be broken in the event that the RA gets 49 information related to the following: a) participant is suicidal, b) participant threatens to commit a 50 homicide, and c) participants reports a sexual abuse to themselves or other parties. The RA will 51 immediately inform the PI about such, and appropriate action will be taken including reporting 52 such cases to the administration of Butabika Hospital. 53 54 Potential risks: There is a potential risk of developing severe psychological distress during the 55 interviews as a result of answering questions that are deemed private by the participant. RA will 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 10 of 15 BMJ Open: first published as 10.1136/bmjopen-2019-034367 on 7 June 2020. Downloaded from 1 2 be trained to identify any of such distress, and the interview will be terminated. Participants may 3 4 be asked to continue with the interview only if they feel like doing so. Such adversities will be 5 reported to the PI and SOMREC. There is also a risk of having information about participants 6 made available in the public domain. We will guard against this by having all identifying 7 information of the patients locked away in file cabinets and password locked computers. The 8 risk to loss of information is minimal. 9 10 Benefits: There are no direct monetary benefits to be gained by the individuals. However, 11 participants will receive regular assessments for their symptoms every 6 months for four years. 12 The scientific community will be able to get information about the predictors of relapse. 13 14 Community Tracing: In the event that a participant has not appeared for a clinic visit on their 15 scheduled clinic appointment, we will contact them or their appointed person through telephone. 16 If neither the patientFor nor the contactpeer person review can be reached byonly phone three months from the last 17 date of their scheduled appointment, we will make active attempts to trace the participant at 18 their residence by liaising with the VHT based in the same location. In the event that the 19 participant can’t be traced at their place of residence, we will consider them as a potential loss 20 to follow-up. There exists a number of village health team members who provide care to non- 21 mental health clients. We will utilize their knowledge about the village to identify individuals in 22 the cohort. 23 24 25 Data analysis plans: 26 27 Descriptive analysis: We shall describe the population characteristics by demographics, social 28 and economic strata using simple proportions. Then we shall perform survival analysis using the 29 Aalen-Johansen estimator were death prior to the outcome will be considered a competing risk. 30 For outcome (a) we shall describe relapse rates at 1 year and 4 years. We shall describe the 31 survival time to relapse for all participants categorizing by psychoses i.e. schizophrenia,

32 schizophrenia spectrum disorders or bipolar affective disorders, considering death as a http://bmjopen.bmj.com/ 33 competing risk. For outcome (b), we also estimate the predictors of relapse rates using 34 generalized linear equation estimates 35 36 Explanatory analyses: We shall use proportional hazard regression to estimate hazard ratios for 37 a clinical relapse (operationally defined above). For this we shall use proportional hazards 38 regression to estimate the hazard ratios for relapse overall. We shall use directed acyclic graphs 39 to define models for evaluation of predictors of relapse. If the power permits, we shall estimate 40 this for each disorder as well. We will document the total length of follow-up period during which on September 30, 2021 by guest. Protected copyright. 41 a subject is (i) adherent to antipsychotic medication, (ii) in a state of complete remission 42 43 (defined as having less than the baseline PANSS score at recruitment, and <20 on YMRS) for 44 half the time they are being followed-up , (iii) in partial remission (no reduction in scores from 45 baseline), and (iv) in a psychotic episode (scores on the PANNS and YMRS increase after 46 discharge and never return to baseline). In cases where we have informative censoring, we 47 shall use inverse probability weights from a sample of tracked participants to estimate the 48 outcome. 49 50 Public and patient involvement: During the pilot phase of the project, we will engage with 51 patients and document any barriers to implementation of the study procedures. 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 11 of 15 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2019-034367 on 7 June 2020. Downloaded from 1 2 Ethics and dissemination plans. 3 4 5 Ethical considerations: We will have a two-layered consent process. (i) Trained RAs will read 6 the consent form to the participant, and (ii) then administer the UBACC to ensure that 7 participants have fully understood the rationale for conducting the study and that they are 8 participating well aware of their rights as participants. Only participants who score ≥10 will be 9 enrolled. Ethical approvals for the study has been obtained from the Makerere University School 10 of Medicine Research and Ethics Committee (#REC-REF 2019-033) and the Uganda National 11 Council for Science and Technology (UNCST HS2638). 12 13 A number of ethical considerations are worth pointing out. Beyond the distress that may l be 14 experienced by participants, this will be a minimal risk study. To minimize the risk of loss of 15 privacy and confidentiality, only the investigators will have access to the study records and test 16 results and the link Forbetween personalpeer identifying review information onlyand study data. No individual 17 identities will be used in any reports or publications associated with the data from this study. All 18 softcopies of the data will be stored in password locked computers. Hard copies of the 19 questionnaires will be stored in locked file cabinets at the study offices in Butabika National 20 Referral Hospital. 21 22 Dissemination plans: A manuscript will be prepared from these findings and submitted to peer 23 24 reviewed journal for publication. Findings will also be presented at local and international 25 conferences. We will hold a dissemination workshop and provide results to the patients who 26 have participated in the pilot phase as well as other stakeholders to whom these findings are 27 important in shaping policy and practice. 28 29 Data statement: The data set collected from this work will be available for use upon receipt of 30 permission from the lead author (request can be made through an email). 31

32 Author contributions: All authors contributed in editing the manuscript and providing http://bmjopen.bmj.com/ 33 constructive feedback. DA Conceptualized, designed and wrote the first draft. SAS contributed 34 to the data analysis component. EM, NN, DB, and JN will train research assistant and provide 35 support supervision during data collection. PK will set up the data bases for analysis. 36 37 Acknowledgement: We acknowledge the editorial work by Prof Dan Stein in the development 38 of the manuscript. 39 40 Funding statement: This study has no funding at the moment. The authors will apply for on September 30, 2021 by guest. Protected copyright. 41 funding from different agencies to conduct the study. In the interim, the Lead Author will cover 42 43 the costs of conducting the study out of pocket. 44 45 46 Competing interest: The authors declare no conflict of interest. 47 48 Word count: 4225 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 12 of 15 BMJ Open: first published as 10.1136/bmjopen-2019-034367 on 7 June 2020. Downloaded from 1 2 References 3 4 5 1. Harvey A Whiteford, Louisa Degenhardt, Jürgen Rehm, et al. Global burden of disease 6 attributable to mental and substance use disorders: fi ndings from the Global Burden of 7 Disease Study 2010. The Lancet 2013;http://dx.doi.org/10.1016/S0140-6736(13)61611- 8 9 2. Rössler W, Salize H.J, van Os J, et al. Size of burden of schizophrenia and psychotic 10 disorders. Eur Neuropsychopharmacol 2005;15(4):399-409. 11 12 13 3. Assen Jablensky. Epidemiology of schizophrenia: The global burden of disease and disability. 14 European Archives of Psychiatry and Clinical Neuroscience 2000;250(6):274-85 15 16 4. Samuli I. Saarni, ForSatu Viertiö peer , Jonna Perälä review , et al. Quality only of life of people with 17 schizophrenia, bipolar disorder and other psychotic disorders. Br J Psychiatry 18 2010;197(5):386-94. doi: 10.1192/bjp.bp.109.076489 19 20 21 5. Erin E Michalak, Lakshmi N Yatham, Raymond W Lam. Quality of life in bipolar disorder: A 22 review of the literature. Health and quality of life outcomes 2005;3(72):doi:10.1186/477- 23 7525-3-72 24 25 6. Charles Laidia, Prigentd A, Alice Plasd, et al. Factors associated with direct health carecosts in 26 27 schizophrenia: Results from theFACE-SZ French dataset. European 28 Neuropsychopharmacology 2018;28:24–36 29 30 7. S.Frey. The economic burden of schizophrenia in Germany: A population-based retrospective 31 cohort study using genetic matching. European Psychiatry 2014;29(8):479-89

32 http://bmjopen.bmj.com/ 33 34 35 36 37 38 39

40 on September 30, 2021 by guest. Protected copyright. 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml