Adverse childhood experiences, adult depression, and suicidal ideation in rural Uganda: A cross-sectional, population-based study.

BACKGROUND: Depression is recognized globally as a leading cause of disability. Early-life adverse childhood experiences (ACEs) have been shown to have robust associations with poor mental health during adulthood. These effects may be cumulative, whereby a greater number of ACEs are progressively associated with worse outcomes. This study aimed to estimate the associations between ACEs and adult depression and suicidal ideation in a cross-sectional, population-based study of adults in Uganda. METHODS AND
FINDINGS: Between 2016 and 2018, research assistants visited the homes of 1,626 adult residents of Nyakabare Parish, a rural area in southwestern Uganda. ACEs were assessed using a modified version of the Adverse Childhood Experiences-International Questionnaire, and depression symptom severity and suicidal ideation were assessed using the Hopkins Symptom Checklist for Depression (HSCL-D). We applied a validated algorithm to determine major depressive disorder diagnoses. Overall, 1,458 participants (90%) had experienced at least one ACE, 159 participants (10%) met criteria for major depressive disorder, and 28 participants (1.7%) reported suicidal ideation. We fitted regression models to estimate the associations between cumulative number of ACEs and depression symptom severity (linear regression model) and major depressive disorder and suicidal ideation (Poisson regression models). In multivariable regression models adjusted for age, sex, primary school completion, marital status, self-reported HIV status, and household asset wealth, the cumulative number of ACEs was associated with greater depression symptom severity (b = 0.050; 95% confidence interval [CI], 0.039-0.061, p < 0.001) and increased risk for major depressive disorder (adjusted relative risk [ARR] = 1.190; 95% CI, 1.109-1.276; p < 0.001) and suicidal ideation (ARR = 1.146; 95% CI, 1.001-1.311; p = 0.048). We assessed the robustness of our findings by probing for nonlinearities and conducting analyses stratified by age. The limitations of the study include the reliance on retrospective self-report as well as the focus on ACEs that occurred within the household.
CONCLUSIONS: In this whole-population, cross-sectional study of adults in rural Uganda, the cumulative number of ACEs had statistically significant associations with depression symptom severity, major depressive disorder, and suicidal ideation. These findings highlight the importance of developing and implementing policies and programs that safeguard children, promote mental health, and prevent trajectories toward psychosocial disability.

Introduction

Major depressive disorder, which is characterized by a range of symptoms such as depressed mood, feelings of low self-worth, anhedonia, and decreased energy [1], has been recognized globally as a leading cause of disability [2]. Associated with high rates of morbidity and mortality, depression negatively affects individuals’ social, occupational, and physical functioning [3]. In low- and middle-income countries (LMICs), depression is the leading neuropsychiatric cause of the burden of disease [4], and by 2030, depression is projected to be the leading cause of the global burden of disease [5]. In addition to its adverse effects on quality of life and functioning, depression has consistently been shown to be a strong risk factor for suicide [6].

Adverse childhood experiences (ACEs) include a range of early-life challenges and traumatic events that occur before age 18 and may put an individual at risk for negative outcomes throughout the life course [7]. Such experiences include emotional, physical, and sexual abuse; household dysfunction; and neglect. Research studies from high-income countries (HICs) have shown robust associations between ACEs and adverse mental health outcomes in adulthood, including antisocial behaviors [8], adult life stress [9], smoking [10], problematic substance use [11], depression [12], and suicide [13]. Research has found that ACEs increase children and adolescents’ risk of having low resilience factors [14]. The magnitude of this association may partially depend on individual-level vulnerability. A study from Australia directly testing the diathesis-stress model for depression found that adults who had high predispositional vulnerability and who had experienced more stressful, adverse life experiences were at the highest risk of developing depression [15].

The effects of ACEs on adult health outcomes may be cumulative, whereby individuals who experience more ACEs have greater risk for developing mental ill health as adults [13,16]. The cumulative disadvantage theory posits that early advantage and disadvantage, including that which results from genetic and environmental factors, may compound, resulting in markedly differing trajectories over time [17]. In addition to the cumulative nature of ACEs, some experiences may be more salient than others; one study, for example, found stronger associations between negative mental health outcomes and child maltreatment than with household dysfunction [18]. An umbrella review of 19 meta-analyses found strong associations between childhood sexual abuse and depression in adulthood, as well as other psychosocial and psychiatric outcomes [19].

Childhood adversity is prevalent across LMICs, with evidence demonstrating the exacerbating effects of poverty and family violence on other childhood traumas [20]. Negative socioeconomic conditions in general, including insecure access to food and water, present additional obstacles to healthy development, with documented negative effects on school attendance, health, and well-being [21]. Prior research from other contexts also indicate that neighborhood poverty is positively associated with abuse, child maltreatment [22,23], and household dysfunction including high caretaker stress and depression [24]. Living in unsafe, “stressogenic” environments characterized by economic precarity, a lack of resources, and violence increases children and adolescents’ exposure to stress and trauma [25]. Furthermore, as “poverty begets trauma” [25] and “stress begets stress” [26], individuals who grow up in such environments are at a higher risk of experiencing psychopathology as adults [27,28]. Thus, the pervasive poverty in some LMICs further underscores the importance of studying the relationship between ACEs and adult mental ill health.

While research studies from sub-Saharan Africa indicate a high prevalence of childhood adversity [29-32], research on the associations between ACEs and adult depression and suicidal ideation in these settings is limited. A study from South Africa found that emotional neglect and sexual abuse before age 18 were both associated with suicidality, problematic substance use, and depression 2 years after initial measurement among adolescents and young adults aged 15 to 26 [30]. Depression assessment in this study was based on past-month symptom screening using the Centre for Epidemiologic Studies Depression Scale. Similarly, a prospective study among South African adolescents found a statistically significant, graded relationship between the cumulative number of ACEs and suicide behaviors after adjusting for baseline suicidality [33]. Two studies of children and adolescents in Uganda found that depression was associated with loss of a parent and alcohol consumption [34] and domestic violence [35]. These and other studies have indicated a high prevalence of ACEs and associations with depression and suicidality among children, adolescents, and young adults in sub-Saharan Africa [31,32,36,37]. To address the gap in the literature on ACEs and mental health among adults, we aimed to estimate the associations between ACEs and depression symptom severity, major depressive disorder, and suicidal ideation in a population-based sample of adults in rural Uganda.

Methods

This study is reported following the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guideline (S1 Checklist).

Study setting and population

This cross-sectional study took place in Mbarara District, a rural region of southwestern Uganda, as part of a population-wide longitudinal study [38]. Mbarara District is made up of 16 subcounties, 90 parishes, and 910 villages. In 2014, the population was estimated at 474,144 people [39]. Through an iterative process involving field site visits and informal conversations with local leaders and other prominent village residents, our team selected Nyakabare Parish as the study site. It was smaller than other parishes in the region, which facilitated our team’s ability to capture a whole-population sample; the local leaders were supportive of engaging area residents in a population-based household survey; and the local leaders stated that there was relatively less nongovernmental organization involvement in the area, in terms of service delivery or other development activities. Thus, the present study took place in the 8 villages of Nyakabare Parish, about 20 kilometers outside of Mbarara Town. Parish residents commonly make their livelihoods from subsistence farming and animal husbandry, and income is often supplemented by migratory work; food and water insecurity are common [40-42].

The national Ugandan Violence Against Children Survey estimated that 35% of girls experience sexual violence and 68% of boys experience physical abuse during childhood and adolescence [43]. Additionally, research studies from across Uganda have estimated a prevalence of probable depression among adults between 17% and 29% [44-46]. Despite the high prevalence of depression in Uganda, there are severe resource shortages for mental health care [47]. Only 1% of the overall health budget in Uganda is allocated to mental health [48]. Furthermore, there is currently less than 1 mental health provider for every 100,000 people in the country [49], and there are limited treatment resources for substance use [50,51]. Resource limitations, coupled with poverty and stigma toward mental illness [52], contribute to a wide treatment gap [53].

Sampling procedure and data collection

Prior to starting the study, a population census was conducted, including the enumeration of all 1,933 adults across the 758 households in the parish. All adults 18 years and older, and emancipated minors between 16 and 18 years, who reported stable residence in the parish were considered for study participation. Emancipated minors were defined as persons younger than 18 years who are married, are pregnant, live with a biological child in the household, or are responsible for their own livelihood. Exclusion criteria included individuals who could not adequately communicate with the research team due to cognitive impairment; behavioral problems including psychosis, neurological damage, or acute intoxication; and deafness, mutism, or aphasia.

Between 2016 and 2018, a team of research assistants visited the homes of all 1,795 eligible adults to request study participation and obtain written informed consent. After providing consent, participants were interviewed in a private location, often in or near the participant’s home. Interviews were conducted in Runyankore, the local language. Data were collected using the Computer-Assisted Survey Information Collection (CASIC) Builder software program. Questionnaires had built-in logic and skip patterns based on participant responses. All instruments were written in English, translated into Runyankore, and back translated into English in an iterative process to confirm translation fidelity.

Measures

ACEs were assessed using a modified version of the Adverse Childhood Experiences–International Questionnaire (ACE-IQ; S1 Text) [54]. The original ACEs instrument was established in the United States (US) [7]. The ACE-IQ was later developed to increase cultural applicability beyond the US context and capture experiences unique to various international settings [55,56]. To date, the ACE-IQ has been validated for use with Malawian adolescents [57] and for adolescents and adults in Nigeria [58].

We modified the ACE-IQ following an iterative process of focus group discussion with key informants. Some of the original ACE-IQ items were dropped because they were either poorly understood or thought to be less applicable for our population. Additional items on food and water insecurity were added because both food and water insecurity are common in this rural setting [40-42]. The modified ACE-IQ included 16 items about exposures to adverse experiences during the participant’s first 18 years of life. For some items, participants were probed for frequency of the experience. For the purposes of this analysis, however, all items on the instrument were converted into binary variables, with any experience of the event categorized as 1 and no experience of the event categorized as 0.

The 16 items on the ACE-IQ were grouped into 9 types of ACEs: (1) physical abuse; (2) verbal or emotional abuse; (3) attempted or enacted sexual abuse; (4) residence with an adult who used alcohol or drugs; (5) residence with an adult who had mental illness or who attempted suicide; (6) parents separated or divorced; (7) residence with an adult who was sent to prison or jail; (8) observed violence toward mother or grandmother; and (9) food and/or water insecurity. These 9 binary variables were summed to calculate the cumulative number of ACEs (S2 Text).

Depression symptom severity was measured using the Hopkins Symptom Checklist for Depression (HSCL-D). The self-report instrument assesses for symptoms of depression over the past week and has been modified and validated for use among Runyankore-speaking populations [59,60]. The original HSCL-D includes 15 items. Local modifications involved dropping 1 item (“feeling trapped”) and adding 1 item (“don’t care what happens to your health”) [42,60]. For each item, respondents are asked the frequency of the respective symptom (i.e., not at all, not much, much, very much). The total score on the HSCL-D is calculated as the average of the responses. One question asks the participant to describe how often, during the past 7 days, they had thoughts of ending their life. This item was converted into a binary variable. Participants who responded that they thought about ending their life “much” or “very much” were categorized as having significant suicidal ideation [61].

Because depression screening instruments can often yield overestimates of depression prevalence [62,63], we applied a previously developed algorithm to identify major depressive disorder based on the Diagnostic and Statistical Manual of Mental Disorders [64-66]. The 15 items of the HSCL-D were then categorized into the 9 DSM “A” criteria: (A1) depressed mood; (A2) diminished interest or pleasure; (A3) significant weight loss or change in appetite; (A4) insomnia or hypersomnia; (A5) psychomotor agitation; (A6) fatigue or loss of energy; (A7) feeling worthless or guilty; (A8) diminished ability to think or concentrate; and (A9) recurrent thoughts of death. Participants who experienced an HSCL-D item “much” or “very much” in the past week were classified as meeting the respective DSM “A” criteria. We then summed the total number of criteria met, out of a maximum of 9. If a participant reported at least 5 criteria, and met either the A1 criterion (depressed mood) or the A2 criterion (diminished interest or pleasure), s/he was identified as likely meeting diagnostic criteria for a major depressive episode or major depressive disorder.

The demographics questionnaire included questions on age, sex, highest level of educational attainment, marital status, self-reported HIV status (if known), and household asset wealth [67]. Participants were classified into wealth quintile categories (i.e., poorest to richest) based on overall household assets [68,69].

Ethics

This study received ethical approval from the Mbarara University of Science and Technology Research and Ethics Committee and the Partners Human Research Committee. Consistent with national guidelines, we obtained clearance to conduct the study from the Uganda National Council for Science and Technology.

Data analysis

The analysis was not preregistered, but we followed a prespecified analysis plan and tracked any deviations that resulted from peer review (S3 Text). To estimate the bivariate associations between the cumulative number of ACEs and depression symptom severity based on the HSCL-D, we fitted a linear regression model to the data with the cumulative number of ACEs as the sole explanatory variable. We then refitted the model, adjusting for the following covariates: sex, age, marital status, primary education completion, HIV status, and household asset wealth quintile category. To determine the extent to which the associations remained statistically significant across the age range, we first fitted a multivariable linear regression model containing a product term between the cumulative number of ACEs and age, specified as a continuous variable. We then conducted further analyses stratified by age, with participants categorized into 1 of 3 age bins: younger adults (26 years of age and younger), adults (27 to 39 years of age), and older adults (40 years of age and older).

We conducted similar analyses to estimate the unadjusted and adjusted associations between the cumulative number of ACEs and major depressive disorder and suicidal ideation. For these analyses, we fitted Poisson regression models with cluster-correlated robust estimates of variance. When applied to binary dependent variable data, the modified Poisson regression model has been shown to yield estimated incidence rate ratios that can be interpreted straightforwardly as relative risk ratios [70].

In secondary analyses, we grouped the cumulative ACEs score into 4 categories for comparability with prior work [26]: no ACEs or 1 ACE (lowest), 2 or 3 ACEs (low), 4 or 5 ACEs (high), and 6 or more ACEs (highest). These ACEs categories were determined based on the interquartile range (IQR) of participants’ cumulative ACEs scores. Using this specification of the ACEs categories, we fitted the same linear and Poisson regression models as described above. Next, we disaggregated the cumulative ACEs score into each of the 9 types of ACEs, allowing us to estimate the associations between each type of ACE and depression symptom severity, major depressive disorder, and suicidal ideation, without assuming that the different ACEs had equivalent associations.

To probe the robustness of our findings to confounding by unobserved variables, we used methods proposed by Vanderweele and Ding [71]. We calculated the e-value to determine the minimum strength of association on the risk ratio scale that would be required for an unobserved confounder to have with both the exposure (ACEs) and outcome (depression or suicidality), conditional on the measured covariates, in order to explain away the observed associations. Thus, the e-value quantifies the extent to which unobserved confounding might contribute to the findings.

Following general econometric guidance [72,73], all regression models included adjustment for clustering at the village level. In the setting of multiple levels of clustering, consistent confidence intervals (CIs) will be obtained by using cluster-robust variance estimated at the highest level of clustering. In a sensitivity analysis, we refitted the primary regression models adjusting for clustering at the household level. All analyses were conducted using Stata version 16 (College Station, Texas).

Results

Participants

Of 1,795 eligible adult community members (91% response rate), 1,626 were included in the analysis. Slightly over half of the population were women (908 [56%]), and the median age among the 1,602 participants who reported their age was 37 years (IQR 26 to 50), including one 17-year-old emancipated minor. A majority (975 [60%]) of the population had completed at least a primary education. Most adults were either married or cohabiting (993 [61%]), with smaller subsets either separated, divorced, or widowed (288 [18%]) or single/never married (344 [21%]) (Table 1).

Table 1

Sociodemographic and health characteristics of the sample, by sex.

	Sex
	Female (n = 908, 55.8%)		Male (n = 718, 44.2%)		Total (n = 1,626)
	n		n		n
Age:
Young adults (17–26 years)	224	24.7%	185	25.8%	409	25.2%
Adults (27–39 years)	274	30.2%	214	29.8%	488	30.0%
Older adults (40+ years)	391	43.2%	314	43.7%	705	43.4%
Missing	19	2.09%	5	0.70%	24	1.48%
Education:
Completed Primary School	486	53.5%	489	68.1%	975	60.0%
Did Not Complete Primary School	422	46.5%	229	31.9%	651	40.0%
Married:
Yes	526	57.9%	467	65.0%	933	61.1%
No	382	42.1%	251	35.0%	633	38.9%
Religion:
Protestant	628	69.2%	502	69.9%	1,130	69.5%
Catholic	211	23.2%	174	24.2%	385	23.7%
Born Again Pentecostal	57	6.38%	29	4.04%	86	5.29%
Muslim	10	1.10%	9	1.25%	19	1.17%
Other	2	0.002%	4	0.01%	6	0.004%
HIV Status:
HIV–Positive	108	11.9%	59	8.22%	167	10.3%
HIV–Negative	800	88.1%	659	91.8%	1,459	89.7%
ACE Category:
Lowest (0–1 ACE)	218	24.0%	169	23.5%	387	23.8%
Low (2–3 ACEs)	301	33.2%	199	27.7%	500	30.8%
High (4–5 ACEs)	237	26.1%	212	29.5%	449	27.6%
Highest (6+ ACEs)	152	16.7%	138	19.2%	290	17.8%
Probable Depression (HSCL-D > 1.75):
Yes	239	26.3%	92	12.8%	331	20.4%
No	669	73.7%	626	87.2%	1,295	79.6%
Major Depressive Disorder:
Yes	114	12.6%	45	6.27%	159	9.78%
No	794	87.4%	673	93.7%	1,467	90.2%
Suicidal Ideation:
Yes	22	2.42%	6	0.84%	28	1.72%
No	886	97.6%	712	99.2%	1,598	98.3%

ACEs, adverse childhood experiences; HSCL-D, Hopkins Symptom Checklist for Depression.

aFigures do not add to 100% due to rounding.

Overall, 1,458 participants (90%) had experienced at least 1 ACE before age 18. The median number of ACEs was 3 (IQR, 2 to 5). On average, men reported more ACEs compared with women (3.3 versus 3.28; t = 1.42, p = 0.16), although the difference was not statistically significant. The majority of participants reported physical abuse, verbal or emotional abuse, and residence with an adult who used alcohol or drugs. While the prevalence of most ACEs was comparable across sexes, women were more likely to report experiences of attempted or enacted sexual abuse compared with men, while men were more likely to report verbal or emotional abuse (Table 2).

Table 2

Prevalence of types of ACEs by sex.

	Sex
	Female (n = 908, 55.8%)		Male (n = 718, 44.2%)		Total (n = 1,626)
	n		n		n
Physical Abuse:
Yes	462	50.9%	362	50.4%	824	50.7%
No	446	49.1%	356	49.6%	802	49.3%
Verbal or Emotional Abuse:
Yes	531	58.5%	453	63.1%	984	60.5%
No	377	41.5%	265	36.9%	642	39.5%
Attempted or Enacted Sexual Abuse:
Yes	150	16.5%	75	10.5%	225	13.8%
No	758	83.5%	643	89.6%	1,401	86.2%
Residence with an Adult Who Used Alcohol or Drugs:
Yes	508	56.0%	420	58.5%	928	57.1%
No	400	44.1%	298	41.5%	698	42.9%
Residence with an Adult Who Had Mental Illness or Who Attempted Suicide:
Yes	227	25.0%	181	25.2%	408	25.1%
No	681	75.0%	537	74.8%	1,218	74.9%
Parents Separated or Divorced:
Yes	224	24.7%	183	25.5%	407	25.0%
No	684	75.3%	535	74.5%	1,219	75.0%
Residence with an Adult Who was Sent to Prison or Jail:
Yes	339	37.3%	274	38.2%	613	37.7%
No	569	62.7%	444	61.8%	1,013	62.3%
Observed Violence Toward Mother or Grandmother:
Yes	284	31.3%	234	32.6%	518	31.9%
No	624	68.7%	484	67.4%	1,108	68.1%
Food and/or Water Insecurity:
Yes	252	27.8%	284	39.6%	536	33.0%
No	656	72.3%	434	60.5%	1,090	67.0%

ACEs, adverse childhood experiences.

aFigures do not add to 100% due to rounding

The mean score on the HSCL-D was 1.48 (standard deviation [SD], 0.42). Women had higher depression symptom severity scores compared with men (1.57 versus 1.38; t = −9.36, p < 0.001). Using a score above 1.75 as a cutoff, 331 participants (20%) screened positive for probable depression. However, only 159 participants (10%) met criteria for probable major depressive disorder. Twenty-eight participants (1.7%) provided responses indicative of suicidal ideation.

Analyses

In the linear regression model estimating the association between cumulative ACEs score and depression symptom severity, the cumulative number of ACEs was associated with depression symptom severity (b = 0.046; 95% CI, 0.038 to 0.053; p < 0.001) (S1 Table). After adjusting for covariates, the estimated association remained statistically significant (b = 0.050; 95% CI, 0.039 to 0.061; p < 0.001) (Table 3). Female sex, older age, and being unmarried were also associated with depression symptom severity in the multivariable regression model. Each additional ACE was associated with 0.050/0.42 = 0.119 SD units of increase in depression symptom severity. A 1 SD difference in the cumulative ACEs score was associated with a 2.20 × 0.119 = 0.26 SD difference in depression symptom severity. In sensitivity analyses, the estimates remained qualitatively similar when we clustered the standard errors at the household level (S2 Table).

Table 3

Adjusted linear and Poisson regression models estimating associations between number of ACEs and depression symptom severity, major depressive disorder, and suicidal ideation.

	Depression Symptom Severity		Major Depressive Disorder		Suicidal Ideation
	Adjusted b (95% CI)	p-value	Adjusted RR (95% CI)	p-value	Adjusted RR (95% CI)	p-value
Cumulative No. ACEs	0.050 (0.039–0.061)	<0.001	1.190 (1.109–1.276)	<0.001	1.146 (1.001–1.311)	0.048
Female	0.178 (0.135–0.222)	<0.001	1.858 (1.298–2.660)	0.001	2.620 (1.510–4.544)	0.001
Age (years)	0.003 (0.002–0.004)	0.001	1.002 (0.995–1.010)	0.511	0.989 (0.959–1.019)	0.466
Completed Primary School	−0.087 (−0.174–0.001)	0.051	0.471 (0.293–0.760)	0.002	0.365 (0.210–0.636)	<0.001
Married	−0.068 (−0.113–−0.023)	0.009	0.712 (0.528–0.959)	0.026	1.178 (0.534–2.600)	0.684
HIV–Positive	−0.035 (−0.100–0.030)	0.239	0.809 (0.604–1.085)	0.158	1.071 (0.527–2.176)	0.849
Wealth Quintile Category
Poorest
Second	−0.060 (−0.123–0.004)	0.061	0.783 (0.547–1.120)	0.181	0.311 (0.077–1.254)	0.101
Third	−0.040 (−0.102–0.021)	0.165	0.804 (0.609–1.060)	0.122	0.856 (0.379–1.932)	0.709
Fourth	−0.068 (−0.123–−0.013)	0.022	0.644 (0.461–0.902)	0.010	0.284 (0.066–1.214)	0.089
Richest	−0.035 (−0.124–0.055)	0.390	1.008 (0.651–1.562)	0.970	1.019 (0.414–2.508)	0.967
Constant	1.246 (1.177–1.315)	<0.001	0.065 (0.030–0.139)	<0.001	0.018 (0.004–0.081)	<0.001
Observations	1,602		1,602		1,602
R2 and Pseudo R2	0.157		0.076		0.080

ACEs, adverse childhood experiences; b, beta coefficient; CI, confidence interval; RR, relative risk.

Each model is adjusted for sex, age, primary school completion, marital status, HIV status, and household asset wealth quintile category.

In a multivariable linear regression model containing a product term between the cumulative number of ACEs and age, specified as a continuous variable, the estimated regression coefficient on the product term was statistically significant and negative, suggesting an interaction (b = −0.001; 95% CI, −0.0014 to −0.0006; p < 0.001): The estimated association between ACEs and depression symptom severity weakened with age (S3 Table). Consistent with this regression model, when we fit regression models stratified by age bin, the estimated association between the cumulative number of ACEs and depression symptom severity was largest among younger adults (b = 0.060; 95% CI, 0.047 to 0.073; p < 0.001) and adults (b = 0.064; 95% CI, 0.056 to 0.071; p < 0.001) and weakest among older adults (b = 0.031; 95% CI, 0.017 to 0.046; p = 0.001).

In the multivariable Poisson regression models, an increase in the cumulative number of ACEs was associated with major depressive disorder (adjusted relative risk [ARR] = 1.190; 95% CI, 1.109 to 1.276; p < 0.001). Similarly, the cumulative number of ACEs was associated with suicidal ideation (ARR = 1.146; 95% CI, 1.001 to 1.311; p = 0.048).

The linear regression models estimating the associations between categorical ACEs score and depression symptom severity indicated a graded increase in associations between ACEs category and depression symptom severity, with the fourth category (6 or more ACEs) associated with the highest depression symptom severity score, both before (b = 0.287; 95% CI, 0.227 to 0.346; p < 0.001) (S4 Table) and after adjusting for covariates (b = 0.313; 95% CI, 0.242 to 0.384; p < 0.001) (Table 4). Study participants who reported 6 or more ACEs had 0.313/0.42 = 0.75 greater SD units of depression symptom severity compared with study participants who reported 0 to 1 ACE. The Poisson regression models estimating the association between categorical ACEs score and major depressive disorder also indicated a graded increase in the prevalence of depression across ACE categories. Individuals who had experienced 6 or more ACEs during childhood were over 2 and a half times as likely to meet criteria for major depressive disorder (RR = 2.616; 95% CI, 1.725 to 3.967; p < 0.001; ARR = 2.819; 95% CI, 2.030 to 3.916; p < 0.001) as those who reported 0 to 1 ACE. None of the categorical ACEs scores, however, had a statistically significant association with suicidal ideation (e.g., ≥6 ACEs: RR = 2.669; 95% CI, 0.973 to 7.320; p = 0.057; ARR = 2.340; 95% CI, 0.844 to 6.488; p = 0.102).

Table 4

Adjusted linear and Poisson regression models estimating associations between ACE category and depression symptom severity, major depressive disorder, and suicidal ideation.

	Depression Symptom Severity		Major Depressive Disorder		Suicidal Ideation
	Adjusted b (95% CI)	p-value	Adjusted RR (95% CI)	p-value	Adjusted RR (95% CI)	p-value
ACE Category
Lowest (0–1 ACE)
Low (2–3 ACEs)	0.081 (0.008–0.155)	0.035	1.356 (0.961–1.913)	0.083	1.668 (0.664–4.191)	0.276
High (4–5 ACEs)	0.138 (0.070–0.205)	0.002	1.595 (1.137–2.236)	0.007	1.141 (0.464–2.802)	0.774
Highest (≥6 ACEs)	0.313 (0.242–0.384)	<0.001	2.819 (2.030–3.916)	<0.001	2.340 (0.844–6.488)	0.102
Female	0.178 (0.135–0.222)	<0.001	1.860 (1.287–2.689)	0.001	2.531 (1.434–4.467)	0.001
Age (years)	0.003 (0.001–0.004)	0.001	1.003 (0.995–1.010)	0.492	0.989 (0.957–1.021)	0.491
Completed Primary School	−0.088 (−0.178–0.001)	0.052	0.468 (0.287–0.764)	0.002	0.357 (0.200–0.635)	<0.001
Married	−0.066 (−0.111–−0.020)	0.011	0.715 (0.533–0.960)	0.026	1.199 (0.565–2.541)	0.637
HIV–Positive	−0.034 (−0.101–0.032)	0.260	0.802 (0.606–1.062)	0.124	1.076 (0.544–2.128)	0.833
Wealth Quintile Category
Poorest
Second	−0.062 (−0.125–0.002)	0.055	0.783 (0.546–1.122)	0.182	0.321 (0.079–1.303)	0.112
Third	−0.041 (−0.106–0.023)	0.171	0.792 (0.597–1.051)	0.106	0.891 (0.398–1.993)	0.778
Fourth	−0.072 (−0.128–−0.016)	0.019	0.624 (0.455–0.855)	0.003	0.283 (0.066–1.218)	0.090
Richest	−0.038 (−0.124–0.049)	0.339	0.984 (0.645–1.502)	0.941	1.014 (0.417–2.468)	0.975
Constant	1.300 (1.236–1.365)	<0.001	0.079 (0.035–0.176)	<0.001	0.020 (0.004–0.112)	<0.001
Observations	1,602		1,602		1,602
R2 and Pseudo R2	0.149		0.071		0.081

ACE, adverse childhood experience; b, beta coefficient; CI, confidence interval; RR, relative risk.

Each model is adjusted for sex, age, education, marital status, HIV status, and household asset wealth quintile category.

Linear regression models demonstrated statistically significant associations between each of the 9 types of ACE and depression symptom severity (S5 Table). Th-estimated associations remained statistically significant after adjusting for covariates. In the adjusted models, associations were strongest for attempted or enacted sexual abuse (b = 0.193; 95% CI, 0.127 to 0.259; p < 0.001), observing violence toward the mother or grandmother (b = 0.131; 95% CI, 0.088 to 0.174; p < 0.001), and food and/or water insecurity (b = 0.187; 95% CI, 0.137 to 0.238; p < 0.001). In the adjusted models, every ACE had a statistically significant association with major depressive disorder. In the models estimating associations between each ACE and suicidal ideation, however, the only 2 experiences that had statistically significant associations with suicidal ideation were residence with an adult who was sent to prison or jail (ARR = 2.654; 95% CI, 1.646 to 4.278; p < 0.001) and food and/or water insecurity (ARR = 1.882; 95% CI, 1.155 to 3.065; p = 0.011).

We explored the robustness of our findings to potential confounding from unobserved variables. Using as an example the multivariable Poisson regression estimate for the association between the highest category of ACEs exposure and major depressive disorder, we obtained an e-value of 5.08. Thus, an unobserved confounder would need to have a strength of association, on the risk ratio scale, with the highest category of ACEs exposure and with major depressive disorder of 5.08 each to move our estimated association to include a risk ratio of 1.

Discussion

In this cross-sectional, population-based study of adults in rural Uganda, we demonstrated robust associations between cumulative number of ACEs and depression symptom severity, major depressive disorder, and suicidal ideation. Furthermore, we estimated a graded association, whereby the prevalence of depression was highest among individuals who reported the highest category of ACEs (≥6 experiences). Significant associations were present for all 9 types of ACEs and depression, and between 2 of the ACEs and suicidal ideation.

While previous research from sub-Saharan Africa has shown associations between ACEs and depression among adolescents and young adults [30-37], our findings demonstrate that the associations are consistent across all age ranges within a general population of adults. Moreover, our study extends prior work by assessing major depressive disorder and suicidal ideation in addition to using a standardized depression screening instrument. The estimated associations between ACEs and major depressive disorder and suicidal ideation are reflective of research showing strong associations between high levels of adversity and major depressive disorder among older adults [74], as well as research indicating associations between ACEs and suicidal ideation and suicide attempts across the lifespan [13,75]. Additionally, the finding that participants who reported food and water insecurity during childhood were at higher risk for suicidal ideation is consistent with a previously published multicountry study showing an increased odds of suicide attempt among adolescents reporting severe food insecurity [76]. Finally, the results suggest that our findings are unlikely to be completely explained by confounding from unobserved variables.

Limitations

These findings must be considered in the context of existing limitations. The reliance on self-report of ACEs during adulthood presents challenges, as retrospective self-report can be subject to recall bias. As Baldwin and colleagues (2019) identified in their systematic review and meta-analysis, retrospective and prospective accounts of childhood adversity can be inconsistent [77]. A study from the South African Birth to Twenty cohort found low concordance between adolescent and young adults’ retrospective reports of ACEs and their caregivers’ prospective reports [78]. If people with depression are more likely to recall ACEs during childhood, then this measurement error could bias our estimated associations away from the null. A second limitation is that, due to the cross-sectional nature of this study, we are unable to infer causal relations between these interrelated factors. However, our estimates were robust to different specifications, and the e-value analysis suggests that an unmeasured confounder would need to be very strongly associated with both ACEs and depression in order to fully explain away the observed associations.

A third limitation, common to nearly all studies using the ACEs questionnaire, is the lack of detail regarding each experience [79]. For example, one question elicits whether a family member or household-dwelling adult went to prison or jail while the participant was a child or adolescent. However, the instrument does not probe for details about the relationship between the study participant and this adult figure, including the quality of their relationship or the importance of this adult figure’s role in their life. Another consideration is the effect of age on reactions to stressful situations. The current instrument does not ask participants how old they were when the experience occurred.

Fourth, our survey instrument may have yielded underestimates of some of the ACEs. Some of the participants lived with their parents or grandparents in intergenerational households, and such living arrangements could have limited our ability to accurately collect sensitive data. However, research assistants ensured that interviews were conducted in a private location out of earshot of other members of the household. Somewhat related to this limitation, physical discipline is normative in East Africa [80,81]. Responses to the physical abuse questions may potentially underrepresent the extent of the experience if participants responded negatively to those ACE-IQ questions, i.e., because they perhaps believed such behaviors to be standard practice. Although these limitations could have caused us to underestimate the prevalence of certain ACEs, they would only have biased our estimates of the associations between ACEs and the mental health outcomes if the factors leading participants to underreport ACEs were also associated with the mental health outcomes.

Lastly, while this study employed a modified ACE-IQ adapted for the local context, the questionnaire focused primarily on experiences encountered within the household and/or perpetrated by a household-dwelling adult. However, research has demonstrated pervasive violence toward children in Uganda enacted outside the household and perpetrated by other individuals, including school staff, peers, neighbors, and strangers [43,82].

Implications for research, clinical Practice, and public policy

To address limitations in our research, future work may use prospective study designs such as that exemplified by Cluver and colleagues in South Africa [33]. Additionally, studies may involve mixed methods, incorporating qualitative interviews that probe for additional detail on ACEs. By developing a better understanding of how developmental stage interacts with ACEs to influence later outcomes, we can better adapt intervention programs for children and their caregivers. Finally, future studies should assess abuse and other traumatic events experienced both within and outside the household. These findings can thereby be used to inform policies and practice to address violence perpetrated at multiple levels within the community.

To address individual-level vulnerability and reduce the “stressogenic” environments that put individuals at higher risk of childhood adversity and adult depression [25], programs and policies are needed to provide support for children, adolescents, and their families. Preventive programs in schools may focus on supporting natural protective factors among children. Furthermore, socioeconomic interventions may provide additional support to parents during stressful economic times. A systematic review found that 35% of socioeconomic interventions, including housing, conditional cash transfer, and income supplementation, reduced children and adolescents’ exposure to ACEs [83]. Research from LMICs has similarly demonstrated positive effects of cash handouts on health, nutrition, school attendance, and cognitive development [84]. While individual-, family-, and neighborhood-level programs are important to addressing abuse, household dysfunction, and neglect, the high level of childhood adversity seen among this population needs to be recognized as a wider public health policy issue.

While it is vital that future policies and programs address high rates of ACEs in this region, many ACEs reflect larger structural and systematic barriers in Uganda, including poverty and economic insecurity. Thus, while targeting specific ACEs to prevent child and adult psychopathology [85], there remains a need to address depression, suicidal ideation, and other associated negative mental health outcomes. Due to the cross-sectional nature of this study, we were unable to determine causality between ACEs and poor mental health during adulthood. For example, adults may react differently to children who have mental health problems during childhood and/or adolescence, including both internalizing and externalizing disorders, thereby predisposing these children and adolescents to certain types of ACEs. Furthermore, early mental health problems may modify individuals’ perceptions of the world and their experiences. As we are not able to rule out reverse causality as a potential explanation for the observed associations, the present findings may have implications for mental health care among both children and adults. Considering high rates of childhood adversity, the high prevalence of depression, and pervasive barriers to accessing behavioral and mental health care across the country, it is critical that mental health be given more recognition and attention within the public health and health systems agenda in Uganda [47,49,86]. Improvements in behavioral and mental health care may in turn provide support for child caregivers and thereby reduce future ACEs [87].

Conclusions

Given the potential relationship between ACEs and adult depression, interventions are needed to prevent childhood adversity and respond to health and social systems in this context. By addressing multilevel factors contributing to these experiences, programs may reduce ACEs among children and adolescents. Based on our interpretation of the present findings, such intervention may improve trajectories toward poor mental health during adulthood. Given high rates of depression and challenges associated with reducing certain ACEs in this context, programs must also be developed to address barriers to accessing mental health and psychosocial support services.

STROBE checklist.

(DOCX)

Click here for additional data file.

Modified Adverse Childhood Experiences–International Questionnaire (ACE-IQ).

(DOCX)

Click here for additional data file.

Calculating the cumulative ACEs score from the modified version of the ACE-IQ–Binary Version.

(DOCX)

Click here for additional data file.

Methods.

(DOCX)

Click here for additional data file.

Unadjusted linear and Poisson regression models estimating associations between cumulative number of ACEs and depression symptom severity, major depressive disorder, and suicidal ideation.

(DOCX)

Click here for additional data file.

Adjusted linear and Poisson regression models estimating associations between number of ACEs and depression symptom severity, major depressive disorder, and suicidal ideation, using standard errors clustered at the household level.

(DOCX)

Click here for additional data file.

Linear regression model with product term between cumulative number of ACEs and age (specified as a continuous variable), and linear regression models estimating associations between cumulative number of ACEs and depression symptom severity, stratified by age category.

(DOCX)

Click here for additional data file.

Unadjusted linear and Poisson regression models estimating associations between ACEs category and depression symptom severity, major depressive disorder, and suicidal ideation.

(DOCX)

Click here for additional data file.

Linear and Poisson regression models estimating associations between each type of ACE and depression symptom severity, major depressive disorder, and suicidal ideation.

(DOCX)

Click here for additional data file.

17 Jun 2020

Dear Dr Satinsky,

Thank you for submitting your manuscript entitled "Associations between adverse childhood experiences and adult depression symptom severity, major depressive disorder, and suicidal ideation in rural Uganda: A cross-sectional, population-based study" for consideration by PLOS Medicine.

Your manuscript has now been evaluated by the PLOS Medicine editorial staff [as well as by an academic editor with relevant expertise] and I am writing to let you know that we would like to send your submission out for external peer review.

However, before we can send your manuscript to reviewers, we need you to complete your submission by providing the metadata that is required for full assessment. To this end, please login to Editorial Manager where you will find the paper in the 'Submissions Needing Revisions' folder on your homepage. Please click 'Revise Submission' from the Action Links and complete all additional questions in the submission questionnaire.

Please re-submit your manuscript within two working days, i.e. by .

Login to Editorial Manager here: https://www.editorialmanager.com/pmedicine

Once your full submission is complete, your paper will undergo a series of checks in preparation for peer review. Once your manuscript has passed all checks it will be sent out for review.

Feel free to email us at plosmedicine@plos.org if you have any queries relating to your submission.

Kind regards,

Caitlin Moyer, Ph.D.,

Associate Editor

PLOS Medicine

20 Nov 2020

Dear Dr. Satinsky,

Thank you very much for submitting your manuscript "Associations between adverse childhood experiences and adult depression symptom severity, major depressive disorder, and suicidal ideation in rural Uganda: A cross-sectional, population-based study" (PMEDICINE-D-20-02698R1) for consideration at PLOS Medicine.

I apologize for the delay in returning a decision on your manuscript. Your paper was evaluated and discussed among all the editors here. It was also sent to three independent reviewers, including a statistical reviewer. The reviews are appended at the bottom of this email and any accompanying reviewer attachments can be seen via the link below:

[LINK]

In light of these reviews, I am afraid that we will not be able to accept the manuscript for publication in the journal in its current form, but we would like to consider a revised version that addresses the reviewers' and editors' comments. Obviously we cannot make any decision about publication until we have seen the revised manuscript and your response, and we plan to seek re-review by one or more of the reviewers.

In revising the manuscript for further consideration, your revisions should address the specific points made by each reviewer and the editors. Please also check the guidelines for revised papers at http://journals.plos.org/plosmedicine/s/revising-your-manuscript for any that apply to your paper. In your rebuttal letter you should indicate your response to the reviewers' and editors' comments, the changes you have made in the manuscript, and include either an excerpt of the revised text or the location (eg: page and line number) where each change can be found. Please submit a clean version of the paper as the main article file; a version with changes marked should be uploaded as a marked up manuscript.

In addition, we request that you upload any figures associated with your paper as individual TIF or EPS files with 300dpi resolution at resubmission; please read our figure guidelines for more information on our requirements: http://journals.plos.org/plosmedicine/s/figures. While revising your submission, please upload your figure files to the PACE digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at PLOSMedicine@plos.org.

We expect to receive your revised manuscript by Dec 11 2020 11:59PM. Please email us (plosmedicine@plos.org) if you have any questions or concerns.

***Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.***

We ask every co-author listed on the manuscript to fill in a contributing author statement, making sure to declare all competing interests. If any of the co-authors have not filled in the statement, we will remind them to do so when the paper is revised. If all statements are not completed in a timely fashion this could hold up the re-review process. If new competing interests are declared later in the revision process, this may also hold up the submission. Should there be a problem getting one of your co-authors to fill in a statement we will be in contact. YOU MUST NOT ADD OR REMOVE AUTHORS UNLESS YOU HAVE ALERTED THE EDITOR HANDLING THE MANUSCRIPT TO THE CHANGE AND THEY SPECIFICALLY HAVE AGREED TO IT. You can see our competing interests policy here: http://journals.plos.org/plosmedicine/s/competing-interests.

Please use the following link to submit the revised manuscript:

https://www.editorialmanager.com/pmedicine/

Your article can be found in the "Submissions Needing Revision" folder.

To enhance the reproducibility of your results, we recommend that you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see http://journals.plos.org/plosmedicine/s/submission-guidelines#loc-methods.

Please ensure that the paper adheres to the PLOS Data Availability Policy (see http://journals.plos.org/plosmedicine/s/data-availability), which requires that all data underlying the study's findings be provided in a repository or as Supporting Information. For data residing with a third party, authors are required to provide instructions with contact information for obtaining the data. PLOS journals do not allow statements supported by "data not shown" or "unpublished results." For such statements, authors must provide supporting data or cite public sources that include it.

We look forward to receiving your revised manuscript.

Sincerely,

Caitlin Moyer, Ph.D.

Associate Editor

PLOS Medicine

plosmedicine.org

-----------------------------------------------------------

Requests from the editors:

1.Data availability statement: Thank you for your willingness to make your data and code available. At this time, please provide information on where the anonymized data may be accessed.

2. Protocol: Did your study have a prospective protocol or analysis plan? Please state this (either way) early in the Methods section.

a) If a prospective analysis plan (from your funding proposal, IRB or other ethics committee submission, study protocol, or other planning document written before analyzing the data) was used in designing the study, please include the relevant prospectively written document with your revised manuscript as a Supporting Information file to be published alongside your study, and cite it in the Methods section. A legend for this file should be included at the end of your manuscript.

b) If no such document exists, please make sure that the Methods section transparently describes when analyses were planned, and when/why any data-driven changes to analyses took place.

c) In either case, changes in the analysis-- including those made in response to peer review comments-- should be identified as such in the Methods section of the paper, with rationale.

3. Abstract: Methods and Findings: Early on in this section, please mention the key features of the study design (e.g. population and setting, number of participants, years during which the study took place, and main outcome measures). Please provide some summary demographic information pertaining to the study participants, including how outcome measures were obtained (ACEs numbers, depression symptom severity/MDD diagnosis, suicide ideation).

4. Abstract: Methods and Findings: For the reported results on number of ACEs and depressive symptoms, risk of major depressive disorder, and suicidal ideation, please provide the p values in addition to providing confidence intervals (please also define the abbreviation CI at first use).

5. Abstract: Methods and Findings: Please mention the important variables that are adjusted for in the analyses.

6. Abstract: Methods and Findings: In the last sentence of the Abstract Methods and Findings section, please describe the main limitation(s) of the study's methodology.

7. Abstract: Conclusion: In the first sentence, we suggest you address the results of the study prior to mentioning the implications; the phrase "In this study, we observed ..." may be useful.

8. Author Summary: At this stage, we ask that you include a short, non-technical Author Summary of your research to make findings accessible to a wide audience that includes both scientists and non-scientists. The Author Summary should immediately follow the Abstract in your revised manuscript. This text is subject to editorial change and should be distinct from the scientific abstract. Please see our author guidelines for more information: https://journals.plos.org/plosmedicine/s/revising-your-manuscript#loc-author-summary

9. Throughout the text: Please use square brackets for in-text citations, like this [1].

10. Methods: Line 85: Please remove the trademark symbol.

11: Methods: Line 90: Please provide the modified version of the ACE-IQ and reference it along with your description as a supporting information file.

12. Methods: Line 119-123: As mentioned by a reviewer, please do provide some further description of the algorithm used to identify major depressive disorder, even though the references are provided.

13: Results: Line 199-202: Please provide the p values in addition to the 95% CIs for the unadjusted and adjusted associations between ACEs and depression severity scores.

14: Results: Line 219-221: Please also provide p values in addition to 95% CIs for unadjusted and adjusted associations between major depressive disorder and ACEs number.

15. Results: Line 221-224: Please provide p values in addition to 95% CIs for unadjusted and adjusted associations between suicidal ideation and cumulative ACE score.

16. Results Lines 225-239: for the linear regression models with categorical ACE score, the comparisons between major depressive disorder between ACE categories, and the association between ACE category and suicidal ideation, please provide p values for adjusted and unadjusted results. Please present the unadjusted results (in a table) and reference the location in the text here.

17: Results: Lines 244-255: Please include p values together with 95% CIs for the associations between each of the types of ACE and depression symptoms, and please include a table of the unadjusted analyses.

18. Results: Lines 252-255: Please present the results with 95% CI and p value for the association between having a parent/adult in prison and suicidal ideation, as this is called out in the text.

19. Discussion: Please slightly reorganize the Discussion as follows: a short, clear summary of the article's findings; what the study adds to existing research and where and why the results may differ from previous research; strengths and limitations of the study; implications and next steps for research, clinical practice, and/or public policy; one-paragraph conclusion.

20. Discussion: In the conclusion paragraph, please temper the statements to avoid inferring the direction of the relationship (ACE leads to depression) and policy/health systems recommendations based on this- while this is an interpretation of the findings, given the limitations mentioned and the cross sectional nature of the study, causal implications should be avoided (lines 381-383).

21. Table 1 and Table 2: Please remove the % sign from the columns headed “%” and please define all abbreviations such as ACE and HSCLD, in the legends.

22. Table 3 and Table 4: Please provide the actual p values associated with the adjusted results (rather than noting p<0.05, etc). In the legend, please provide the abbreviations for ACE, and HSCLD-15, and note which variables were adjusted for in these analyses. Please provide the unadjusted results (and reference in the text around Line 201, 220 and 223)- if desired this can be presented as a supporting information file. Please also clarify whether the suicidal ideation column is a comparison with total number of ACE, or cumulative ACE score (as mentioned in the text).

23. S1 Appendix and S2 Appendix: Please note the actual p values, rather than * for p<0.05, for example. Please note in the legend of S1 Appendix what variables were adjusted for. For both, please also provide the unadjusted results.

24. Checklist: Please ensure that the study is reported according to the STROBE guideline, and include the completed STROBE checklist as Supporting Information. When completing the checklist, please use section and paragraph numbers, rather than page numbers. Please add the following statement, or similar, to the Methods: "This study is reported as per the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guideline (S1 Checklist)."

Comments from the reviewers:

Reviewer #1: See attachment

Michael Dewey

Reviewer #2: This is a very good paper, thorough and well constructed. It follows closely the established ACE methodology conducted by the CDC in the US, and replicated in other high-income settings. Its findings are perhaps not totally surprising, but they are nonetheless very important. We have always assumed that ACE's work in similar ways in LMIC, but this is a very robust investigation of what is actually an empirical question, and the findings are very clear. In fact, they are so close to the findings of the original CDC study (which had a very similar method - retrospective reporting of a population cohort) that this is worth remarking on. ACE impact people in Sub-Saharan Africa just like they do in America. This does matter and should be published.

I would have been interested to see an analysis by gender, but I don't think this should be required at all - just a thought (and gender is in the models). I should add that I'm not so familiar with the adult mental health literature in Eastern Africa, so I have taken as accurate that this kind of study has not been conducted before in the region.

I would recommend this paper for publication, and commend the authors. I'd be interested to know what the objections of previous reviewers were. I would also be very interested for future papers to see the impacts of ACEs on other health, economic and social outcomes. Lucie Cluver.

Reviewer #3: Overall, the methodology and writing is solid. To make an impact, it needs to go further: it needs to really contextualize this piece of research in the larger field, it needs to make a case for what new information it is bringing to the field, and it needs to dive deeper into a rich discussion of what those findings mean for practice. As written, the discussion is not sufficiently developed.

* The literature review seems to neglect a lot of work that has been done on ACEs and psychological health in Africa. For example, Cluver et al in South Africa; Manyema et al with young adults in South Africa, and Kidman et al in Malawi have all explicitly studied ACEs and suicidality, depression or a similar outcomes. These are not perfect parallels to the research question or age group here, but would provide additional context.

* The introduction states that this line of research has been done in HIC, but not LMIC. However, there is no nuanced discussion of how these contexts might matter to the relationship under study.

* There is also little on how ACEs get embodied. A sentence or two on how ACEs translate into later depression would be helpful.

* Why was this particular area (Nyakabare Parish) chosen for this study? Why a census instead of a random sample of a larger area?

* The paper states that the ACE-IQ was modified, but it isn't clear how it was modified. Please provide details. How was it coded? It mentions a point for each item, but typically the coding for the ACE-IQ is a bit more complicated and based at least on a point per type of adversity (which can cover multiple items). Table 2 seems to list these, indicating the range should be 0-9. Is this correct? Later you refer to 16 - but that should be questions not ACEs. And why was food insecurity included? That decision is fine, but a justification is needed. I think some of the other questions are a little different too. This will be confusing when researchers try to compare ACE prevalence across populations/papers without adequate descriptions.

* It would be helpful to have a bit more on the algorithm for the depression diagnosis in the paper, especially since it is the key outcome.

* Analyses adjust for clustering at the village level, but what about the household level? If you only interviewed one per household, this should be in the methods. However, based on the numbers presented, I don't think this is the case.

* Do you have the statistical power to examine individual ACEs and suicide ideation?

* What value do you get out of the individual ACE models? Given the similarity in coefficients, are they important alone, or do they serve as a proxy for overall ACE exposure? For these analyses to be included, it would be essential to motivate them the introduction. Why isn't a cumulative score a better indicator; why would individual ACEs matter and which ones would you expect to see generate a large association? In the absence of such, this comes off as rather exploratory.

* The paragraph starting with line 275 seems to fit better in an introduction. I am not clear how it sheds further light on the new findings. I am also not clear where poverty comes into the discussion of ACEs and depression.

* The last sentence of that paragraph implies that the study found little access to support or resources; this isn't reported in the paper.

* The paragraph starting with line 289 also seems out of place. This is background material, and better suited for an introduction. It is not again related to the findings. The focus on "stressogenic" environments and poverty in particular seems to be introduced for the first time here, and is background - it isn't something we learn from or sheds light on the findings.

* The discussion on implications seems like it could have been written before this study was conducted. It doesn't answer new questions about what should be done using your new findings. The authors start to bring in some of this - such as the fact that ACEs were incredibly high compared to other contexts - but don't drill down. They also start to acknowledge the limited mental health resources available in Uganda, but don't really go into how the new information on ACEs relates to this. The authors make the argument that the high rates of depression warrant more investment in mental health, but this isn't the core of the paper/findings - in fact, these rates were given as background. What other approaches could be useful in this context, given the new info? If there are Ugandan authors, perhaps they can help.

Any attachments provided with reviews can be seen via the following link:

[LINK]

Submitted filename: satinsky.pdf

Click here for additional data file.

14 Dec 2020

Submitted filename: acedepression_response-to-reviewers_plosmed_20201211.docx

Click here for additional data file.

16 Apr 2021

Dear Dr. Satinsky,

Thank you very much for re-submitting your manuscript "Associations between adverse childhood experiences and adult depression symptom severity, major depressive disorder, and suicidal ideation in rural Uganda: cross-sectional, population-based study" (PMEDICINE-D-20-02698R2) for review by PLOS Medicine.

I have discussed the paper with my colleagues and the academic editor and it was also seen again by one of the original reviewers. I am pleased to say that provided the remaining editorial and production issues are dealt with we are planning to accept the paper for publication in the journal.

The remaining issues that need to be addressed are listed at the end of this email. Any accompanying reviewer attachments can be seen via the link below. Please take these into account before resubmitting your manuscript:

[LINK]

***Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.***

In revising the manuscript for further consideration here, please ensure you address the specific points made by each reviewer and the editors. In your rebuttal letter you should indicate your response to the reviewers' and editors' comments and the changes you have made in the manuscript. Please submit a clean version of the paper as the main article file. A version with changes marked must also be uploaded as a marked up manuscript file.

Please also check the guidelines for revised papers at http://journals.plos.org/plosmedicine/s/revising-your-manuscript for any that apply to your paper. If you haven't already, we ask that you provide a short, non-technical Author Summary of your research to make findings accessible to a wide audience that includes both scientists and non-scientists. The Author Summary should immediately follow the Abstract in your revised manuscript. This text is subject to editorial change and should be distinct from the scientific abstract.

We expect to receive your revised manuscript within 1 week. Please email us (plosmedicine@plos.org) if you have any questions or concerns.

We ask every co-author listed on the manuscript to fill in a contributing author statement. If any of the co-authors have not filled in the statement, we will remind them to do so when the paper is revised. If all statements are not completed in a timely fashion this could hold up the re-review process. Should there be a problem getting one of your co-authors to fill in a statement we will be in contact. YOU MUST NOT ADD OR REMOVE AUTHORS UNLESS YOU HAVE ALERTED THE EDITOR HANDLING THE MANUSCRIPT TO THE CHANGE AND THEY SPECIFICALLY HAVE AGREED TO IT.

Please ensure that the paper adheres to the PLOS Data Availability Policy (see http://journals.plos.org/plosmedicine/s/data-availability), which requires that all data underlying the study's findings be provided in a repository or as Supporting Information. For data residing with a third party, authors are required to provide instructions with contact information for obtaining the data. PLOS journals do not allow statements supported by "data not shown" or "unpublished results." For such statements, authors must provide supporting data or cite public sources that include it.

To enhance the reproducibility of your results, we recommend that you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. Additionally, PLOS ONE offers an option to publish peer-reviewed clinical study protocols. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols

Please review your reference list to ensure that it is complete and correct. If you have cited papers that have been retracted, please include the rationale for doing so in the manuscript text, or remove these references and replace them with relevant current references. Any changes to the reference list should be mentioned in the rebuttal letter that accompanies your revised manuscript.

Please note, when your manuscript is accepted, an uncorrected proof of your manuscript will be published online ahead of the final version, unless you've already opted out via the online submission form. If, for any reason, you do not want an earlier version of your manuscript published online or are unsure if you have already indicated as such, please let the journal staff know immediately at plosmedicine@plos.org.

If you have any questions in the meantime, please contact me or the journal staff on plosmedicine@plos.org.

We look forward to receiving the revised manuscript by Apr 23 2021 11:59PM.

Sincerely,

Caitlin Moyer, Ph.D.

Associate Editor

PLOS Medicine

plosmedicine.org

------------------------------------------------------------

Requests from Editors:

1.From the academic editor: Methods: Line 133-134: Is it relevant to say that the study is part of a population based cohort? Please clarify to avoid confusion.

2. From the academic editor: Methods: Line 200-201: Although the Hopkins Symptom Checklist has been validated in Uganda, please confirm whether it has been validated in this Ugandan language or cultural group (or in a rural setting).

3. We suggest revising the title to: “Adverse childhood experiences, adult depression and suicidal ideation in rural Uganda: A cross-sectional, population-based study”

4. Please revise the short title to: “Adverse childhood experiences and adult depression in Uganda”

5. Data availability statement: Please update the statement at this time, with your plan for making the anonymized data and code available ( with links for access to anonymized data and Stata code in a repository or similar).

6. Abstract: Methods and Findings: Line 10: Should “elicited” be “assessed” or similar? (“...ACES were elicited using a modified…”)

7. Abstract: Line 15: Please remove “significant” before suicidal ideation.

8. Introduction: Line 122 (and throughout text): For in-text references, where a range or list of references are noted within brackets, please do not include spaces [31,32].

9. Introduction: Line 122: Please qualify “...no study has yet…” with “to our knowledge” or similar.

10. Methods: Line 159: Please provide the number of adults represented by the 758 households.

11. Methods: Line 168: Please provide the number of eligible adults.

12. Methods: Line 267-268: Please clarify the rationale for not adjusting at the household level, in line with the response to reviewer 3, comment 7. For example, in the response to reviewers it is mentioned that the estimates in Table 3 are consistent between clustering at village vs household level, but you also note that confidence intervals become less precise adjusting for household clustering. Please consider incorporating this into the supporting information as an additional analysis.

13. Results: Line 299: Please remove “significant” before suicidal ideation.

14. Discussion: Thank you for including the sections describing the strengths and limitations of the work, as well as implications for future research, practice, and policy. If possible, please consider expanding on the paragraph between lines 380-386, where you describe your findings in the context of the existing literature.

15. Discussion: Line 377: Please consider changing the wording of “risk of developing depression” as the study is cross-sectional it may be more conservative to avoid implying any temporal nature of the association.

16. Discussion: Line 384-386: Please revise this sentence to make the meaning more clear: “Finally, we demonstrate that our findings are such that only strong confounding by unobserved variables could completely explain them.” This could be modified to “...results suggest that our findings are unlikely to be completely explained by unobserved variables, fo example…” or similar.

17. Discussion: Line 472: Please consider tempering “prevent” to “improve” here, to avoid any causal implications.

18. References: Please use the "Vancouver" style for reference formatting, and see our website for other reference guidelines https://journals.plos.org/plosmedicine/s/submission-guidelines#loc-references

(For example, in Ref 4, PLoS Medicine should be PLoS Med).

19. Supporting Information: Thank you for including the revised supporting information documents. Please submit a finalized version for each item (rather than the tracked changes versions). It would be helpful for each item to be included in an independent file and labeled as such (S1_Analysis Plan, S1_Checklist, S1_Table, etc.)

20. S3 Appendix (Analysis plan): It would be helpful to include in this Prespecified Analysis Plan document a brief mention of the specific outcomes and analyses planned, to go along with your statement that all the described analyses were specified at the outset of the study. Thank you for thoroughly documenting the changes made to the analyses during the course of peer review.

Comments from Reviewers:

Reviewer #1: The authors have addressed my points.

I may not have been sufficiently clear about the quintiles issue. My point was that this is potentially ambiguous as there are only four quintiles. It would have been possible to use them in the regression but that is not what the authors did. I agree this is a bit picky. I suggested quintile categories or even the everyday English word fifths but if the authors insist I would not want to man the barricades over it. I realise I am fighting a losing battle here anyway and will soon have to accept people talking about dividing a ample into two medians.

Michael Dewey

Any attachments provided with reviews can be seen via the following link:

[LINK]

23 Apr 2021

Submitted filename: acesdepression_response-to-reviewers-2_plosmed_20210423.docx

Click here for additional data file.

29 Apr 2021

Dear Dr Satinsky,

On behalf of my colleagues and the Academic Editor, Charlotte Hanlon, I am pleased to inform you that we have agreed to publish your manuscript "Adverse childhood experiences, adult depression, and suicidal ideation in rural Uganda: A cross-sectional, population-based study" (PMEDICINE-D-20-02698R3) in PLOS Medicine.

Before your manuscript can be formally accepted you will need to complete some formatting changes, which you will receive in a follow up email. Please be aware that it may take several days for you to receive this email; during this time no action is required by you. Once you have received these formatting requests, please note that your manuscript will not be scheduled for publication until you have made the required changes.

Please also address the following two editorial requests:

- Data availability statement: Thank you for providing the GitHub link. At this time, it seems that only the ReadMe file is posted. Please also upload the data files. Also, please update the Data Availability Statement within the manuscript submission system.

- Reference List: Please double-check the formatting of the References. For example, some information appears to be missing from Reference 14, Reference 34, Reference 35. Please update Reference 78.

In the meantime, please log into Editorial Manager at http://www.editorialmanager.com/pmedicine/, click the "Update My Information" link at the top of the page, and update your user information to ensure an efficient production process.

PRESS

We frequently collaborate with press offices. If your institution or institutions have a press office, please notify them about your upcoming paper at this point, to enable them to help maximise its impact. If the press office is planning to promote your findings, we would be grateful if they could coordinate with medicinepress@plos.org. If you have not yet opted out of the early version process, we ask that you notify us immediately of any press plans so that we may do so on your behalf.

We also ask that you take this opportunity to read our Embargo Policy regarding the discussion, promotion and media coverage of work that is yet to be published by PLOS. As your manuscript is not yet published, it is bound by the conditions of our Embargo Policy. Please be aware that this policy is in place both to ensure that any press coverage of your article is fully substantiated and to provide a direct link between such coverage and the published work. For full details of our Embargo Policy, please visit http://www.plos.org/about/media-inquiries/embargo-policy/.

To enhance the reproducibility of your results, we recommend that you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. Additionally, PLOS ONE offers an option to publish peer-reviewed clinical study protocols. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols

Thank you again for submitting to PLOS Medicine. We look forward to publishing your paper.

Sincerely,

Caitlin Moyer, Ph.D.

Associate Editor

PLOS Medicine