Mental and substance use disorders and food insecurity among homeless adults participating in the At Home/Chez Soi study.

BACKGROUND: Few studies have examined how food insecurity changes over time when living with severe mental disorders or substance use disorders. This study identifies food insecurity trajectories of homeless adults participating in a trial of a housing intervention and examines whether receiving the intervention and having specific mental and substance disorders predict food insecurity trajectories.
MATERIALS AND METHODS: We studied 520 participants in the Toronto site of the At Home/Chez-Soi project. Food insecurity data were collected at seven times during a follow-up period of up to 5.5 years. Mental and substance use disorders were assessed at baseline. Food insecurity trajectories were identified using group based-trajectory modeling. Multinomial logistic regression was used to examine the effects of the intervention and mental and substance use disorders on food insecurity trajectories.
RESULTS: Four food insecurity trajectories were identified: persistently high food insecurity, increasing food insecurity, decreasing food insecurity, and consistently low food insecurity. Receiving the intervention was not a predictor of membership in any specific food insecurity trajectory group. Individuals with major depressive episode, mood disorder with psychotic features, substance disorder, and co-occurring disorder (defined as having at least one alcohol or other substance use disorder and at least one non-substance related mental disorder] were more likely to remain in the persistently high food insecurity group than the consistently low food insecurity group.
CONCLUSION: A persistently high level of food insecurity is common among individuals with mental illness who have experienced homelessness, and the presence of certain mental health disorders increases this risk. Mental health services combined with access to resources for basic needs, and re-adaptation training are required to enhance the health and well-being of this population.

Introduction

Mental and substance use disorders remain a major public health and social issue among homeless individuals. A recent study conducted in high-income countries found alcohol dependence and drug dependence are among the most common disorders among the homeless population, with a prevalence ranging from 8–58% and 5–54%, respectively [1]. Prevalence estimates for mental disorders, such as psychosis, depression, personality disorder, and post-traumatic stress disorder, are also higher than those reported for the general population of those countries [1,2]. For instance, the prevalence of psychosis among homeless individuals (3–42%) is approximately 3 times higher than the estimate in the general population [1].

Individuals with poor mental health are at a greater risk of poverty and job loss [3,4], social isolation/exclusion and stigmatization [5-7], which can contribute to other personal vulnerabilities [5,6,8-11], such as homelessness and food insecurity. Food insecurity is the lack of or limited access to food or to nutritious diet because of financial constraints, and is estimated to affect more than two thirds of individuals who experience homelessness with a mental disorder. [12-17] While having a severe mental disorder is an underlying contributor to food insecurity [11,18], the relationships are bidirectional; food insecurity is also a risk factor for mental health problems, including depressive symptoms, mood disorder, stress and anxiety, [19-22] suicidality, [19,23,24] substance abuse, [15,25] and poor cognitive performance [26,27].

Prior studies indicate homeless individuals disproportionately experience food insecurity more than the general population [12,13,28,29]. However limited research has quantified the effects of mental disorders on food insecurity over time among people experiencing homelessness [30,31]. Using the At Home/Chez Soi (AH/CS) data, a previous analysis across 5 Canadian cities examined the effect of a Housing First (HF) intervention on food security among homeless people with mental health disorders, and found marginal but inconsistent improvements in food security following provision of housing after 2-year of implementation [31]. In this study, mental health disorders were not the main focus, and food insecurity was modelled at specific points in time, which ignores food security trajectories altogether or assumes that they are homogenous within each intervention group. Recent studies suggest this type of analysis may conceal significant heterogeneity of long-term trajectories within groups [32,33]. Alternatively, by examining the complexity of food insecurity by modelling trajectories, additional insights into these relationships can be revealed. A trajectory-based approach can highlight how trajectories of food insecurity differ across individuals and identify factors that influence individuals to follow a given trajectory.

To build on existing literature, this paper examines the association of mental and substance use disorders with food insecurity trajectories among a sample of homeless adults enrolled in the At Home Chez Soi (AH/CS) Study, a randomized controlled trial of Housing First in Toronto. Specifically, the objectives of the study were to 1) identify trajectories of food insecurity over a period of 5.5 years, 2) test the predictive effect of the housing intervention on the identified patterns of changes in food insecurity, and 3) analyze how mental and substance disorders predict food insecurity trajectory membership. We hypothesize that having a severe mental or substance use disorder will be strong predictors of a trajectory of persistent food insecurity over the follow up period.

Materials and methods

Housing First intervention

This study used data from the Toronto site of the AH/CS study, which was a randomized controlled trial that compared a scattered-site Housing First intervention and support services (HF) to treatment as usual (TAU) [34]. Detailed information on the recruitment and study design has been published elsewhere [35]. In brief, participants were recruited from community agencies, shelters, clinics, and directly from the street between October 2009 and July 2011. The inclusion criteria specified participants to be at least 18 years old, absolutely homelessness or precariously housed, diagnosed with a mental disorder, and not being served by an assertive community treatment program at the time of enrolment.

Prior to randomization, participants were stratified by their level of needs for mental health services. Need level was determined based on an algorithm that included the presence of psychotic disorder or bipolar affective disorder with psychotic symptoms (based on the Mini International Neuropsychiatric Interview (MINI) 6.0), level of community functioning (Multnomah Community Ability Scale), presence of a co-morbid substance use disorder, and history of hospitalization and incarcerations [34,35]. Out of the 575 Toronto participants, 197 participants met criteria to be classified as high level of need and 378 as moderate level of need. Participants were then randomly assigned to the intervention groups according to their level of need, with the high needs group receiving Housing First with assertive community treatment (HF-ACT) and the moderate needs group participants receiving Housing First with intensive case management services (HF-ICM). Alternatively, participants randomized to TAU continue to have access to the same locally available housing or social support services, irrespective of need level. Follow up interviews occurred every 3 months for the first two years after randomization. Once the 2-year follow-up was complete in July 2013 (phase I), the Toronto site received additional funding to follow participants until March 2017 (phase II).

Ethical approval

The Toronto AH/CS study received approval by the Research Ethics Board of St. Michael’s Hospital in Toronto, Canada. All study participants provided written informed consent to participate in the AH/CS study. The AH/CS study is registered with the International Standard Randomized Control Trial Number Register (ISRCTN42520374).

Outcomes

Food security data were collected during in person interviews using the modified version of the US Adult Food Security Survey Module [36]. This instrument has been validated as a measure of food insecurity among individuals who experience homelessness [37-39]. The total score is a sum of 10 items related to food access, which is further dichotomized into food secure (a total score of 0 to 2) and food insecure (a total score of 3–10) [40-42]. Data were collected every 6 months during phase I and every 12 months during phase II for a total of seven time points over the follow-up period. To ensure trajectories could be properly examined, we excluded participants who had food security data for less than two time points (n = 55), resulting in a final sample of 520 participants for the analysis. Comparisons of socio-demographics characteristics between participants and the excluded group were conducted using student t-tests or Fisher’s test, and showed a statistically significant different only for the variable level of need. Out of the 520 study participants, 48.9% had a high level of need compared to 32.9% in the excluded group (p-value: 0.027, see S1 Table). Therefore, we included this variable as an adjustment variable in our models. S2 Table of Means (and standard deviation) of the number of food insecurity assessments of the participants included in this analysis by intervention group and by key mental health disorders is shown in S2 Table.

Mental and substance use disorders

The following mental health disorders were examined as predictors of food insecurity: depressive episode, panic disorder, mood disorder with psychotic features, posttraumatic stress disorder (PTSD), manic or hypomanic episodes, psychotic disorder, alcohol and substance use disorder. All mental and substance use disorders were identified based on DSM-IV criteria using the MINI 6.0 and were evaluated at the time of screening for study eligibility [34,43].

Co-occurring disorders were defined as a comorbid condition including at least one alcohol or other substance use disorder and at least one non-substance related mental disorder. [44,45]

Covariates

We included the factors used to screen or assign participants to the different AH/CS study groups: level of needs (high vs. moderate level of need), gender (male/female), self-identified ethnic group (white vs. member of non-white/ethnic groups), and intervention group (At home participant vs treatment as usual) [34,35,46]. Based on previous studies on food insecurity, we also adjusted for age, lifetime duration of homelessness (less than 3 years vs. 3 years or more), and education level (middle/high school, completed high school, and university or higher) [42,47].

Statistical analyses

Group-based trajectory modelling was used to identify clusters of individuals who followed a similar pattern of change (trajectory) of food insecurity over time [48]. Using intercept and time as change parameters, the model assumed a logistic distribution of the dichotomous food insecurity variable in order to estimate the latent trajectory groups. We tested the shape of trajectory groups by including higher-order polynomial growth factors (linear, quadratic, and cubic time factor). The optimal number of trajectory groups was determined by using the Bayesian Information Criterion (BIC), while the best fit of trajectory shape was examined using the average posterior probability measure and the weighted odds of correct classification [48,49]. All models were estimated using the command Traj in Stata 15 [48].

Next, we used multinomial logistic regression to examine the effect of HF on trajectory group membership. We fit an unadjusted model including only the intervention group and an adjusted model controlling for gender, age, education level, ethno-racial status, levels of need, and lifetime duration of homelessness.

We also used multinomial logistic regression to separately assess the effect of having a mental or substance use disorder on food insecurity trajectory membership. For each mental health outcome we adjusted for gender, age, education level, ethno-racial group, levels of need, homelessness lifetime duration, and intervention group. Twenty-five individuals (4.8%) were excluded from the multivariable analysis due to missing values for either education or lifetime duration of homelessness. To evaluate the family-wise error rate due to mutilple inferences, we used Bonferroni to compute a corrected overall critical P-value [50]. All analyses were conducted using Stata 15 [51].

Results

Characteristics of study participants are summarized in Table 1. Out of our 520 participants, substance use disorder (37.9%), psychotic disorder (36.3%), and major depressive episode (36.0%) were the most common mental disorders identified. Moreover, 246 of the participants were diagnosed with co-occurring mental and substance use disorders (47.3%).

Table 1

Participants’ characteristics at baseline.

Variable	Total (n = 520)
Female	165 (31.7)
Age (mean(SD))	40.3 (11.8)
Self-identified ethnic group	304 (58.5)
Education level
Middle/high school	241 (46.3)
Completed high school	94 (18.1)
Graduate/postgraduate	170 (32.7)
Missing	15 (2.9)
Level of need
High level	175 (33.7)
Low level	345 (66.3)
Lifetime duration of homelessness
Less than 3 years	225 (43.3)
More than 3 years	270 (51.9)
Missing	25 (4.8)
Intervention
Housing first	282 (54.2)
Treatment as usual	238 (45.8)
Mental and substance disorders
Major depressive episode	187 (36.0)
Manic or Hypomanic episode	57 (11.0)
Post-traumatic stress disorder	126 (24.2)
Panic disorder	74 (14.2)
Mood disorder with psychotic features	107 (20.6)
Psychotic disorder	189 (36.3)
Alcohol disorder	151 (29.0)
Substance disorder	197 (37.9)
Co-occurring disorders	246 (47.3)

Food insecurity trajectories

Based on the Bayesian Information Criterion fit statistics, the model with two quadratic trajectories and two linear trajectories was the best fit model (BIC for the 2-group model = -1635.22; BIC3 = -1579.35, and BIC4 = -1573.34, and BIC5 = 1598.37). The average posterior probability for the 4-group model ranged from 0.75 to 0.81 and the odds of correct classification weighted posterior probability was higher than 5, which also indicate good fit.

The model classified the participants into four food insecurity trajectory groups, persistently high food insecurity, increasing food insecurity (from moderate to high), decreasing food insecurity (from moderate to low), and consistently low food insecurity (Table 2 and Fig 1).

Table 2

Food Insecurity (FI) trajectories.

Parameters		Four-group model of Food insecurity trajectory
		Intercept	linear	Quadratic	Cubic
	Persistent high FI	1.12	0.08
	Increasing FI	0.08	-0.06	0.00
	Decreasing FI	-0.10	0.16	-0.01
	Consistent low food FI	-1.23	-0.06
Membership and Posterior probability of assignment
	Group	Group membership: N (%)	Group APP	OCC weighted
	Persistent high FI	131 (25.2)	0.75	12.74
	Increasing FI	99 (18.9)	0.78	7.70
	Decreasing FI	163 (31.4)	0.77	9.94
	Consistent low food FI	127 (24.4)	0.81	13.39

APP = Average of the maximum posterior probability of assignments.

OCC = Odds of correct classification weighted posterior probability.

Fig 1

Food Insecurity (FI) trajectories over time points.

Housing First and food insecurity trajectory group membership

Table 3 estimates the relative risk ratios (RRRs) of belonging to each food insecurity trajectory group (considering consistently low FI as the reference) for a participant assigned to the HF intervention compared to TAU. Assignment to the Housing First intervention was not statistically associated with any of food insecurity trajectory groups.

Table 3

Effect of Housing First intervention on Food Insecurity (FI) trajectory group membership.

Intervention	Persistent FI		Increasing FI		Decreasing FI
	RRR (95% CI)	P-value	RRR (95% CI)	P-value	RRR (95% CI)	P-value
Unadjusted model
HF vs. TAU	0.86 (0.54 to 1.38)	0.53	0.90 (0.51 to 1.60)	0.72	0.97 (0.6 to 1.52)	0.88
Adjusted model*
HF vs. TAU	0.88 (0.54 to 1.42)	0.59	0.93 (0.52 to 1.70)	0.80	0.97(0.6 to 1.54)	0.90

Consistent low FI is the base outcome

RRR: relative risk ratio

*The model is adjusted for the following variables: Gender, age, education level, self-identified ethnic group, level of need, lifetime homelessness, and intervention group

Mental health disorders and food insecurity trajectory membership

Participants with major depressive episode were more likely to be part of the the persistent food insecurity trajectory group compared to the group of consistent low food insecurity trajectory (RRR = 1.9 [95% CI: 1.1 to 3.2]) (Table 4). Likewise, having a mood disorder with psychotic features was also associated with increased risk of being in the persistent food insecurity group (RRR = 3.4 [95% CI: 1.6 to 6.9]), decreasing FI group (RRR = 4.1 [95% CI: 2.0 to 8.1]), and increasing FI group (RRR = 2.7 [95% CI: 1.2 to 6.2]). Individuals with substance disorder were also more likely to be part of the group that followed the persistent food insecurity trajectory (RRR = 3.0 [95% CI: 1.7 to 5.2]) and the decreasing FI trajectory (RRR = 2.2 [95% CI:1.3 to 3.8]). In contrast, participants with psychotic disorder were less likely to belong to the persistent food insecurity trajectory group (RRR = 0.5 [95% CI: 0.3 to 0.9]) compared to the group of consistent low food insecurity trajectory. Finally, co-occurring disorders were associated with increased likelihood of being classified in the persistent food insecurity trajectory (RRR = 2.78 [95% CI: 1.60 to 4.83]) and in the decreasing food insecurity trajectory (RRR = 2.45 [95% CI: 1.45 to 4.15]) groups, compared to the group of consistent low food insecurity trajectory. Estimates for Mood disorder with psychotic features, Substance disorder, and Co-occurring disorders remained statistically significant after the Bonferroni adjustement for multiple inferences.

Table 4

Multivariable multinomial logistic regressions Food Insecurity (FI) trajectories and mental disorders adjusted for baseline factors.

	Food Insecurity (FI) Trajectories
Mental and substance use disorder	Persistent high FI		Increasing FI		Decreasing FI
	RRR (95% CI)	P-value	RRR (95% CI)	P-value	RRR (95% CI)	P-value
Major depressive episode	1.87 (1.10 to 3.17)	0.021	1.25 (0.66 to 2.36)	0.491	1.31 (0.79 to 2.19)	0.299
Manic or Hypomanic episode	1.1 (0.51 to 2.36)	0.804	0.55 0.18 to 1.61)	0.274	1.11 (0.54 to 2.29)	0.772
Post-traumatic stress disorder	1.75 (0.97 to 3.14)	0.062	0.9 (0.42 to 1.91)	0.785	1.51 (0.85 to 2.66)	0.156
Panic disorder	1.46 (0.70 to 3.04)	0.316	1.26 (0.53 to 3.02)	0.597	1.46 (0.72 to 2.94)	0.294
Mood disorder with psychotic features	3.35 (1.63 to 6.91)	0.001	2.70 (1.18 to 6.18)	0.018	4.06 (2.02 to 8.15)	0.001
Psychotic disorder	0.54 (0.31 to 0.92)	0.024	0.70 (0.36 to 1.35)	0.290	0.92 (0.56 to 1.53)	0.759
Alcohol disorder	1.32 (0.75 to 2.34)	0.334	1.20 (0.62 to 2.34)	0.590	1.26 (0.73 to 2.18)	0.402
Substance disorder	2.96 (1.69 to 5.21)	0.001	1.36 (0.70 to 2.66)	0.363	2.20 (1.28 to 3.78)	0.004
Co-occurring disorders	2.78 (1.60 to 4.83)	0.001	1.50 (0.78 to 2.86)	0.221	2.45 (1.45 to 4.15)	0.001

RRR: relative risk ratio

Reference group: Consistent low food FI

The model was adjusted for the following variables: Gender, age, education level, self-identified ethnic group, level of need, lifetime homelessness, and intervention group

Bonferroni Corrected overall critical P-value: 0.0014 (See the smile plot S1 Fig)

Boldface indicates statistical significance after Bonferroni Adjustment

Discussion

This study employed a group-based approach to identify trajectories of food insecurity and to examine the predictive effects of a Housing First intervention and mental and substance use disorders on trajectory group membership. We identified four trajectory groups: 1) persistent food insecurity trajectory group, who remained food insecure over the entire study period, 2) consistent low food insecurity trajectory group, who remained low food insecure over the follow-up period, 3) decreasing food insecurity trajectory group, who experienced less food insecurity over time, and 4) increasing food insecurity trajectory group, who experienced more food insecurity over time. These trajectories highlight the diverse experiences of food insecurity among homeless individuals and a need for targeted interventions to reduce food insecurity. For example, individuals within the increasing trajectory group require strategies to stabilize or reverse the increasing food insecurity over time, while the decreasing FI insecurity group requires supports to accelerate the reductions in food insecurity.

Our findings suggest the Housing First intervention does not have a significant association with food insecurity group trajectory membership. This lack of effect is consistent with the findings of a prior analysis by O’Campo et al. [31] which raised concerns about the limited effects of Housing First on food insecurity over two years of follow up. In fact, the existing literature provides little evidence that the Housing First intervention improves other non-housing outcomes, such as quality of life, physical community integration, psychological community integration or mental health. [52]

Homeless individuals with major depressive episode, mood disorder with psychotic features, and substance disorder were more likely to remain persistently food insecure over the study period. We also found co-occurring disorders played a major role in food insecurity trajectories. Having an alcohol- or drug-related disorder along with a non-alcohol/drug disorder was highly associated with membership to the persistent food insecurity trajectory group, compared to the group of the consistent low food insecurity group.

These results are in line with previous studies that exposed the relationship between mental disorders and food insecurity within the homeless population[30,31,53,54]. A study conducted by Parpouchi et al. [30] found that among a sample of homeless individuals in Canada, those having panic disorder, PTSD or major depressive episode, or having two or more mental disorders had higher odds to be food insecure. A previous analysis conducted by O’Campo using At Home/Chez Soi two years of follow-up data found that having more psychiatric symptoms and substance use disorder were associated with lower food security. [31] In a representative sample of Canadian adults, mood or anxiety disorders were associated with severe household food insecurity, while co-occurrence of chronic physical and mental health conditions has a gradient effect on the severity of household food insecurity. [18] Another study in the United States found similar results regarding the association between depressive symptoms and persistent food insecurity among low-income rural families. [11] However, we found no prior studies that analyzed food insecurity trajectory over time.

Surprisingly, we found that having a psychotic disorder was associated with lower risk of following a persistent food insecurity trajectory. We found no previous studies focused on psychotic disorders and persistent food insecurity, which implies a need for more research prior to asserting that individuals with psychotic disorders are less likely to have food insecurity. It is possible that our participants experiencing psychotic disorder tended to report lower food insecurity scores due to the nature of the disorders, which can impair the perception of their every life experiences and temporality [55,56] including food-related pattern and needs.

Three plausible mechanisms may explain the relationship between mental and substance disorders and food insecurity trajectories among individuals with homelessness experience. First, mental disorders impair coping strategies and skills to manage food insufficiency in the face of scarce resources even after being in stable housing [18,57]. While previous studies identify issues related to food storage or access to a kitchen for meal preparation [23], we hypothesize that mental disorders, mainly Mood disorder with psychotic features and substance disorder, limit adjustment to am environment and lifestyle with better food security related supports (i.e., kitchen) and ability to manage resources in a manner that prioritized food security [58]. Second, when mental disorders co-occur with (recent or present) homelessness and structural discrimination, an individual’s ability to gain suitable employment and social capital is impaired, which further limits availability of economic resources to support achieving food security [59-61]. Third, moving to stable housing may eventually increase living costs—transportation, energy bills, household supplies—which result in more competition for scarce resources impacting food security.

The following limitations should be considered when interpreting the present study findings. The group-based trajectories are not fixed pathways that remain invariant over time or across populations, but rather help to identify clusters of people following similar pathways in relation to a given outcome. [48,49] Thus, more research studies are needed to verify whether the trajectories are similar in other homeless populations, such as individuals without severe mental illness. Moreover, participants were asked at each interview about food insecurity over the past month. These recall data may be subject to bias and error. Moreover, personality disorder was not included among the mental disorders screened by the At Home/Chez Soi project, though it is generally highly prevalent among homeless individuals [1]. Therefore, we could not include in the present study.

Our findings have research and public health implications. Since mental health issues and food insecurity are often intertwined chronic conditions, [11,18] it is necessary to account for the unique impact specific mental disorders can have on food insecurity over a long period of time. This is especially relevant when providing housing, as individual mental health needs and trauma should be taken into account but also to ensure access to adequate social and basic needs, including food security program. Homelessness, food insecurity and discrimination are avoidable social problems that affect thousands of people locally and globally; political and social actions such as greater access to social and transitional housing, skill and job support services, a basic income for low income individuals, and interventions to target implicit bias for employers and service providers can help to create healthier and more inclusive societies.

Conclusion

Homeless individuals with mental health disorders can be grouped into food insecurity trajectories, and those with mental health disorders such as major depressive episode, Mood disorder with psychotic features, and substance disorder are more likely to remain persistently food insecure over time. Mental health services combined with access to resources for basic needs, and social and re-adaptation training are required to enhance the health and well-being of this population.

Smile plot of relative risk ratio by food security trajectory groups.

(DOCX)

Click here for additional data file.

Fisher test comparing characteristics between study participants (n = 520) and the excluded group (n = 55).

(DOCX)

Click here for additional data file.

Means (and standard deviation) of the number of food insecurity assessments of the participants included in this analysis by intervention group and by key mental health disorders.

(DOCX)

Click here for additional data file.

Participants’ characteristics at baseline by intervention group.

(DOCX)

Click here for additional data file.

Participants’ characteristics at baseline by food security trajectory group.

(DOCX)

Click here for additional data file.

7 Oct 2019

PONE-D-19-21014

Mental and Substance Use Disorders and Food Insecurity among Homeless Adults Participating in the At Home/Chez Soi Study

PLOS ONE

Dear Dr Lachaud,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

The manuscript has been assessed by two reviewers; their comments are available below.

The reviewers have raised major concerns that need attention in a revision. The reviewers note that a related paper from the same project should be discussed and the relationship between the current work and that publication described in greater detail. The reviewers request additional information on the participants excluded, and raise questions about the statistical analyses undertaken, including the need for correction for multiple comparisons for some of the analyses. The reviewers request further information on the restrictions that apply to data access and note that causal language should be revised as causal conclusions are not supported by the study design.

Could you please carefully revise the manuscript to address the concerns raised by the reviewers?

We would appreciate receiving your revised manuscript by Nov 19 2019 11:59PM. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled 'Manuscript'.

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 look forward to receiving your revised manuscript.

Kind regards,

Iratxe Puebla

Senior Managing Editor, PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at http://www.journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and http://www.journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2.  We note that you have indicated that data from this study are available upon request. PLOS only allows data to be available upon request if there are legal or ethical restrictions on sharing data publicly. For information on unacceptable data access restrictions, please see http://journals.plos.org/plosone/s/data-availability#loc-unacceptable-data-access-restrictions.

In your revised cover letter, please address the following prompts:

a) If there are ethical or legal restrictions on sharing a de-identified data set, please explain them in detail (e.g., data contain potentially identifying or sensitive patient information) and who has imposed them (e.g., an ethics committee). Please also provide contact information for a data access committee, ethics committee, or other institutional body to which data requests may be sent.

b) If there are no restrictions, please upload the minimal anonymized data set necessary to replicate your study findings as either Supporting Information files or to a stable, public repository and provide us with the relevant URLs, DOIs, or accession numbers. Please see http://www.bmj.com/content/340/bmj.c181.long for guidelines on how to de-identify and prepare clinical data for publication. For a list of acceptable repositories, please see http://journals.plos.org/plosone/s/data-availability#loc-recommended-repositories.

We will update your Data Availability statement on your behalf to reflect the information you provide.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Partly

Reviewer #2: Partly

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: No

Reviewer #2: No

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: Although the statistical analyses were conducted appropriately and rigorously to the best of my knowledge, the following major concerns prompted me to question the technical soundness of the manuscript:

The manuscript cites “a recent study” on food insecurity among homeless individuals (lines 84-85), but does not make it clear in the text that this study is an analysis of data from the same parent study (At Home/Chez Soi) as the current manuscript. Because the research questions of the published 2017 paper on food insecurity among At Home/Chez Soi participants appear related to the research questions of the current manuscript, the authors should acknowledge the previously published paper and explicitly state how the research questions and analyses of the current paper are distinct.

The manuscript repeatedly uses the language of impact with reference to substance use disorders and mental disorders, e.g. “this paper examines the impact of mental and substance use disorders on food insecurity trajectories” (lines 93-94). This does not accurately reflect the paper’s research design and analyses, which demonstrate associations between substance use/mental disorders and food insecurity trajectories, but cannot definitively show that substance use/mental disorders impact food insecurity.

The plausible mechanisms suggested (lines 294-306) speak mainly to the general relationship between homelessness and food insecurity (already well-established in prior research) and do not match the study's specific findings on food insecurity trajectories. For example, the authors mention lack of food storage and pressures of survival needs (e.g. finding a place to sleep) as explanatory factors in high rates of food insecurity among homeless people, but there was no difference in food insecurity noted between Housing First and treatment as usual participants in this study. Presumably the Housing First residents would have greater access to food storage, and are not faced with the survival need of finding a place to sleep, so the relevance of this mechanism to the study's findings is not clear.

The statement that “political and social actions are required to create healthier and more inclusive societies” (lines 321-322) is extremely vague; it would be more helpful for the authors to name specific policy or practice implications related to the study findings. The conclusion “Both food related and mental health interventions and services are required to enhance the health and well being of this population” (lines 326-328) is also very vague and does not speak to the study’s findings regarding different food insecurity trajectories and the possibility that people with different vulnerabilities are at varying levels of risk for food insecurity and therefore may require different services or interventions.

Regarding data availability, the authors state that “Data cannot be shared publicly because of ethical restrictions to the data” but do not specify the nature of the restrictions or further elaborate on this.

In addition, I noted the following minor concerns:

Use of the term “alcohol and drug dependence” in line 65 is outdated, as this reflects DSM-IV language.

The origin of the definition of food insecurity provided (lines 74-75) is not clear; many current definitions specify that food insecurity is not just about access to food, but access to nutritious food. It is also not clear if the stated prevalence of food insecurity in homeless individuals with mental disorders is referring to the Canadian context, or globally (line 76).

The meaning of “ethno-racial group (yes or no)” is not clear (line 150).

Gender is not listed as a covariate on p. 7-8, but is named as a control variable in Tables 3 and 4.

The authors discuss dichotomizing the variable of food insecurity (lines 135-136), but then reference three levels of food insecurity (low, moderate, and high) in defining the four food insecurity trajectories (lines 189-191).

Reviewer #2: The authors present results of a study of food insecurity over time in homeless individuals in Toronto participating in a randomized trial of a scattered-site Housing First intervention with support services compared to treatment as usual. The authors identify four subgroups of trajectories of food insecurity (persistently high food insecurity, increasing food insecurity, decreasing food insecurity, and consistently low food insecurity). They found no association between intervention group and trajectory group, but did identify some mental health disorders at higher risk of belonging to certain food insecurity groups. The manuscript will be strengthened if the authors consider the following points.

1. The authors mention that 55 individuals were excluded due to having only 1 food insecurity assessment. This makes sense in the context of the focus on longitudinal patterns in food insecurity. However, I think it will help the reader understand the sample better if the authors include some basic descriptions of those excluded (how many from HF vs TAU, how many were food insecure vs not, how many were high vs moderate level of need).

2. What was the mean (and standard deviation) of the number of food insecurity assessments for the 520 people included in this analysis? Did this differ by intervention group or by key mental health disorders of interest?

3. It would be helpful to show Table 1 also broken down by Intervention group. It also would be helpful to have the distribution of need (moderate vs high) by mental and substance disorders presented in a table or in the text (which might help in interpreting the results in later tables).

3. Are the quadratic terms really needed for the Increasing FI and Decreasing FI groups? The parameters presented in Table 2 are essentially 0. What is the BIC in the 4 group model with all linear trajectories?

4. Under the section starting on line 216, it would be worth reporting the percent randomized to HR in each of the FI trajectory groups to further support the RRR presented in Table 3.

5. In some sense, the trajectory groups might be considered ordered, with Persistent FI the worst, then Increasing FI, then Decreasing FI, and then consistently low FI. Why did the authors use a multinomial logistic model rather than an approach that takes into account this ordering? Taking this into account might better capture who is most at risk for the worse trajectories compared to less worse trajectories.

6. The authors perform 9 different models in Table 4, but there is no adjustment of multiple comparisons.

7. For co-occurring disorders, the authors focus on the results related to persistent FI (lines 239-242), but there was also a finding with increased risk fo decreasing FI.

8. How do the authors interpret the fact that substance abuse and mood disorder with psychotic features are at increased risk of both the persistent FI (a bad thing) and decreasing FI (probably a good indicator)? Same for co-occurring disorders. These conclusions may be different if the authors actually take into account the ordering of the groups as I mention above, because as analyzed, these results are confusing to try to interpret.

Minor edits:

1. line 65: change "vary" to "varies"

2. line 140: there is an extra "."

3. line 150: there is an extra ","

4. Table 1: It would be helpful to know what was captured in Ethno-racial group. There are 304 of these individuals in the sample, so I imagine there are further sub-groups that can be listed here with corresponding frequencies to give the reader a better sense of who is represented in this sample.

5. Table 2: the percentage for Increasing FI should be 19.0 (not 18.9)

6. line 237: what does "and RRR=1.6 [95% CI: 1.0 to 2.7]" refer to? This doesn't seem to match anything in Table 4.

7. line 258 "need to for" should be "need for"

8. In the Abstract, Discussion , and Conclusion, the authors include PTSD in the list of disorders more likely to be in the persistently FI group, but this finding was not significant (though close). The authors also don't mention this in the text of the Results section.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files to be viewed.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (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. Registration is free. 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 figures@plos.org. Please note that Supporting Information files do not need this step.

19 Nov 2019

PONE-D-19-21014

Mental and Substance Use Disorders and Food Insecurity among Homeless Adults Participating in the At Home/Chez Soi Study

Editorial Team

Dear Editor,

I am pleased to re-submit the manuscript entitled “Mental and Substance Use Disorders and Food Insecurity among Homeless Adults Participating in the At Home/Chez Soi Study”. We would like to extend our sincere gratitude to the PlosOne team and the Reviewers for their positive and constructive feedback.

The co-authors and I have carefully reviewed the Reviewers’ comments and provide point-by-point responses to their comments below. In particular, we have addressed the concerns related to the participants excluded, and have also applied the Bonferroni test for multiple comparisons for some of the analyses. We have also changed the statement to the data restrictions and removed all language or expressions that can imply causal relationship.

We include a manuscript with the highlighted changes and a clean version, as well as an appendix file with additional tables. We believe that the revised manuscript is greatly improved and has addressed all Reviewer comments.

Thank you again for your consideration.

Sincerely,

Reviewer #1: Although the statistical analyses were conducted appropriately and rigorously to the best of my knowledge, the following major concerns prompted me to question the technical soundness of the manuscript:

The manuscript cites “a recent study” on food insecurity among homeless individuals (lines 84-85), but does not make it clear in the text that this study is an analysis of data from the same parent study (At Home/Chez Soi) as the current manuscript. Because the research questions of the published 2017 paper on food insecurity among At Home/Chez Soi participants appear related to the research questions of the current manuscript, the authors should acknowledge the previously published paper and explicitly state how the research questions and analyses of the current paper are distinct.

Response: We thank Reviewer #1 for this constructive feedback and appreciate this important point raised. We added more information about this study and explained how is distinct to this manuscript. We state as follows (lines 84-90):

Using the At Home/Chez Soi (AH/CS) data, a recent study across 5 Canadian cities examined the effect of a Housing First (HF) intervention on food security among homeless with mental health disorders, and found marginal but inconsistent improvements in food security following provision of housing after 2-year of implementation(31). In this study, mental health disorders were not the main focus, and food insecurity was modelled at specific points in time, which ignores food security trajectories altogether or assumes that they are homogenous within each intervention group. Recent studies suggest this type of analysis may conceal significant heterogeneity of long-term trajectories within groups (32,33).

In the following paragraph, we explained (line 96-104):

To build on existing literature, this paper examines the association of mental and substance use disorders on food insecurity trajectories among a sample of homeless adults enrolled in the At Home Chez Soi (AH/CS) Study, a randomized controlled trial of Housing First in Toronto. Specifically, the objectives of the study were to 1) identify trajectories of food insecurity over a period of 5.5 years, 2) test the predictive effect of the housing intervention on the identified patterns of changes in food insecurity, and 3) analyze how mental and substance disorders predict food insecurity trajectory membership.

Reviewer: The manuscript repeatedly uses the language of impact with reference to substance use disorders and mental disorders, e.g. “this paper examines the impact of mental and substance use disorders on food insecurity trajectories” (lines 93-94). This does not accurately reflect the paper’s research design and analyses, which demonstrate associations between substance use/mental disorders and food insecurity trajectories, but cannot definitively show that substance use/mental disorders impact food insecurity.

Response: We changed as follows (lines 96):

To build on existing literature, this paper examines the association of mental and substance use disorders on food insecurity trajectories among a sample of homeless adults enrolled in the At Home Chez Soi (AH/CS) Study, a randomized controlled trial of Housing First in Toronto.

Reviewer: The plausible mechanisms suggested (lines 294-306) speak mainly to the general relationship between homelessness and food insecurity (already well-established in prior research) and do not match the study's specific findings on food insecurity trajectories. For example, the authors mention lack of food storage and pressures of survival needs (e.g. finding a place to sleep) as explanatory factors in high rates of food insecurity among homeless people, but there was no difference in food insecurity noted between Housing First and treatment as usual participants in this study. Presumably the Housing First residents would have greater access to food storage, and are not faced with the survival need of finding a place to sleep, so the relevance of this mechanism to the study's findings is not clear.

Response: We have reformulated the entire paragraph in the discussion section (lines 316-327):

First, mental disorders impair coping strategies and skills to manage food insufficiency in the face of scarce resources even after being in stable housing(18,56). While previous studies identify issues related to food storage or access to a kitchen for meal preparation (23), we hypothesize that mental disorders, mainly Mood disorder with psychotic features and substance disorder, limit adjustment to an environment and lifestyle with better food security related supports (i.e., kitchen) and ability to manage resources in a manner that prioritized food security(57). Second, when mental disorders co-occur with (recent or present) homelessness and structural discrimination, an individual’s ability to gain suitable employment and social capital is impaired, which further limits availability of economic resources to support achieving food security(58–60). Third, moving to stable housing may eventually increase living costs—transportation, energy bills, household supplies--which result in more competition for scarce resources impacting food security.

Reviewer: The statement that “political and social actions are required to create healthier and more inclusive societies” (lines 321-322) is extremely vague; it would be more helpful for the authors to name specific policy or practice implications related to the study findings. The conclusion “Both food related and mental health interventions and services are required to enhance the health and wellbeing of this population” (lines 326-328) is also very vague and does not speak to the study’s findings regarding different food insecurity trajectories and the possibility that people with different vulnerabilities are at varying levels of risk for food insecurity and therefore may require different services or interventions.

Response: We reformulated the entire paragraph as follows to provide more specification (lines 336-345):

Since mental health issues and food insecurity are often intertwined chronic conditions,(11,18) it is necessary to account for the unique impact specific mental disorders can have on food insecurity over a long period of time. This is especially relevant when providing housing, as individual mental health needs and trauma should be taken into account but also to ensure access to adequate social and basic needs, including food security program. Homelessness, food insecurity and discrimination are avoidable social problems that affect thousands of people locally and globally; political and social actions such as greater access to social and transitional housing, skill and job support services, a basic income for low income individuals, and interventions to target implicit bias for employers and service providers can help to create healthier and more inclusive societies.

Mental health services combined with access to resources for basic needs, and social and re-adaptation training are required to enhance the health and well-being of this population. (lines 350-352)

Reviewer: Regarding data availability, the authors state that “Data cannot be shared publicly because of ethical restrictions to the data” but do not specify the nature of the restrictions or further elaborate on this.

Response: We changed the statement as follows

Anonymised participant data, the study protocol, informed consent forms, survey forms, and statistical analysis plan from the At Home/Chez Soi Toronto site study will be available to investigators for studies that have received approval from independent research committees or research ethics boards. Data are available from the publication date of this article onwards. Study proposals and data access requests should be sent to Dr Stephen Hwang at Stephen.Hwang@unityhealth.to . All study proposals and data requests will be further reviewed by the At Home/Chez Soi team at the Toronto site. Data sharing agreements between the requestors and At Home/Chez Soi principal investigators need to be completed before accessing the data.

Reviewer: In addition, I noted the following minor concerns:

Use of the term “alcohol and drug dependence” in line 65 is outdated, as this reflects DSM-IV language.

Response: Thank you for the precision. We reformulated it as follows (line 65):

While the prevalence of mental disorders vary across studies, alcohol and drug use are the most common disorders among the homeless population, with prevalence ranging from 8-58% and 5-54%, respectively(1)

We don’t used “alcohol and drug disorders” to avoid too much repetition of the work “disorder” in the same sentence

Reviewer: The origin of the definition of food insecurity provided (lines 74-75) is not clear; many current definitions specify that food insecurity is not just about access to food, but access to nutritious food. It is also not clear if the stated prevalence of food insecurity in homeless individuals with mental disorders is referring to the Canadian context, or globally (line 76).

We have added “or to nutritious diet” to the definition (line 75)

The meaning of “ethno-racial group (yes or no)” is not clear (line 150).

We have added it: (defined as white vs. any ethno-racial or no-white groups) (line 163).

This self-identified ethno-racial was included in Toronto Housing First project not to identify specific (race or ethnicity), but to adapt the services the category.

Goering PN, Streiner DL, Adair C, Aubry T, Barker J, Distasio J, et al. The at Home/Chez Soi trial protocol: A pragmatic, multi-site, randomised controlled trial of a Housing First intervention for homeless individuals with mental illness in five Canadian cities. BMJ Open. 2011;1(2):1–18.

Gender is not listed as a covariate on p. 7-8, but is named as a control variable in Tables 3 and 4.

We have it as a covariate: gender (male/female) line (162)

Reviewer: The authors discuss dichotomizing the variable of food insecurity (lines 135-136), but then reference three levels of food insecurity (low, moderate, and high) in defining the four food insecurity trajectories (lines 189-191).

Thank you. We explained in the methodology how we transformed the original variable presented in the lines 135-136 into the food insecurity trajectory groups (lines 169-177)

We sincerely appreciate the constructive feedback from Reviewer #1!

Reviewer #2: The authors present results of a study of food insecurity over time in homeless individuals in Toronto participating in a randomized trial of a scattered-site Housing First intervention with support services compared to treatment as usual. The authors identify four subgroups of trajectories of food insecurity (persistently high food insecurity, increasing food insecurity, decreasing food insecurity, and consistently low food insecurity). They found no association between intervention group and trajectory group, but did identify some mental health disorders at higher risk of belonging to certain food insecurity groups. The manuscript will be strengthened if the authors consider the following points.

We thank Reviewer #2 for these insightful suggestions and recommendations.

1. Reviewer: The authors mention that 55 individuals were excluded due to having only 1 food insecurity assessment. This makes sense in the context of the focus on longitudinal patterns in food insecurity. However, I think it will help the reader understand the sample better if the authors include some basic descriptions of those excluded (how many from HF vs TAU, how many were food insecure vs not, how many were high vs moderate level of need).

Response: We have added a table A1 in Appendix comparing participant’s socio-demographic characteristics collected at the baseline, including age, gender, Intervention group, level of need, ethno-racial group. Let us mention the exclusion is for those who had less than 2 assessments, including no food data at all.

We also indicated in the text (line 14-5148):

Comparisons of socio-demographics characteristics between participants and the excluded group were conducted using student t-tests or Fisher’s test were conducted and showed statistically significant differences only for the variable level of need (See Table A1 in Appendix), which is included as an adjustment variable in our models.

2. Reviewer: What was the mean (and standard deviation) of the number of food insecurity assessments for the 520 people included in this analysis? Did this differ by intervention group or by key mental health disorders of interest?

Response: We added a Table A2 in Appendix presenting the mean (and standard deviation by intervention groups and key mental health.

We also indicated in the text (lines 148-151):

Table A2 of Means (and standard deviation) of the number of food insecurity assessments of the participants included in this analysis by intervention group and by key mental health disorders is shown in Appendix.

3. Reviewer: It would be helpful to show Table 1 also broken down by Intervention group. It also would be helpful to have the distribution of need (moderate vs high) by mental and substance disorders presented in a table or in the text (which might help in interpreting the results in later tables).

Response: We have added this comparison table in Appendix Table A3. We do not want to present in the text body since this comparison table has been published in previous papers of AH /Chez-Soi Study (so it could be seen a duplication of published results) and the comparison is not the main purpose of this study.

The last one was recently published at the Lancet Psychiatry.

Stergiopoulos, V., Mejia-Lancheros, C., Nisenbaum, R., Wang, R., Lachaud, J, O'Campo, P., and Hwang, S.W. (2019). The long-term effects of rent supplements and mental health support services on housing and health outcomes of homeless adults with mental illness: outcomes of the extended At Home/Chez Soi randomized controlled trial, The Lancet Psychiatry., https://doi.org/10.1016/S2215-0366(19)30371-2

3. Reviewer: Are the quadratic terms really needed for the Increasing FI and Decreasing FI groups? The parameters presented in Table 2 are essentially 0. What is the BIC in the 4 group model with all linear trajectories?

Response: We have tested all the possibilities, including 4-group model with all linear trajectories. The BIC for 4 group model with all linear trajectories is -1632.59. In addition, the Average of the maximum posterior probability of assignment, other good-fitness criteria, is better for the model selected.

4. Reviewer: Under the section starting on line 216, it would be worth reporting the percent randomized to HR in each of the FI trajectory groups to further support the RRR presented in Table 3.

Response: We have added Table A4 in Appendix which shows Participants’ Characteristics at baseline by Food Security Trajectory Group including the variable intervention and all key mental health to see the percent randomized.

5. Reviewer: In some sense, the trajectory groups might be considered ordered, with Persistent FI the worst, then Increasing FI, then Decreasing FI, and then consistently low FI. Why did the authors use a multinomial logistic model rather than an approach that takes into account this ordering? Taking this into account might better capture who is most at risk for the worse trajectories compared to less worse trajectories.

Response: As you mentioned “in some sense”, we also have difficult to accept a strict order when looking at the Trajectory group graph mainly for groups 2 and 3. Hence, that is why we believe that imposing that order might bias the results and the interpretation.

6. Reviewer: The authors perform 9 different models in Table 4, but there is no adjustment of multiple comparisons.

Response: We want to thank Reviewer#2 for this insight comment. We computed the Bonferroni Adjustment for multiple testing as suggested. We have added in the methodology:

To evaluate the family-wise error rate due to multiple inferences, we use Bonferroni to compute a Corrected overall critical P-value (49). (lines 187-188)

The value overall critical p-value is presented at the foot of the Table 4, and the smile plot Fig A1 in Appendix. (lines 267-268)

We also explained in the section of Results (lines 257-259):

Estimates for Mood disorder with psychotic features, Substance disorder, and Co-occurring disorders remained statistically significant after the Bonferroni adjustement for multiple inferences.

7. Reviewer: For co-occurring disorders, the authors focus on the results related to persistent FI (lines 239-242), but there was also a finding with increased risk for decreasing FI.

Response: We have added it the analysis. (lines 256-257)

8. Reviewer: How do the authors interpret the fact that substance abuse and mood disorder with psychotic features are at increased risk of both the persistent FI (a bad thing) and decreasing FI (probably a good indicator)? Same for co-occurring disorders. These conclusions may be different if the authors actually take into account the ordering of the groups as I mention above, because as analyzed, these results are confusing to try to interpret.

Response: The reference category for the analysis is Consistent low food FI, as we added as a footnote to the Table 4 (line 265). Such as, it is a better status compared to the other food insecurity trajectory groups, including Decreasing FI group. Hence, it seems that the estimates for substance abuse and mood disorder with psychotic (and also co-occurring disorders) go to the expected direction, higher than 1, compared to the Consistent low food FI group. Yet again, the fact the estimates for Increasing FI are all not statistically significant expresses our difficulty to impose a strict order between groups 2 and 3.

Reviewer: Minor edits:

1. line 65: change "vary" to "varies"

We have changed it

2. line 140: there is an extra "."

We have removed it

3. line 150: there is an extra ","

We have removed it

4. Table 1: It would be helpful to know what was captured in Ethno-racial group. There are 304 of these individuals in the sample, so I imagine there are further sub-groups that can be listed here with corresponding frequencies to give the reader a better sense of who is represented in this sample.

Response: We have added the definition Ethno-racial (defined as: white vs. any ethno-racial or no-white groups). (line 163)

This self-identified ethno-racial was included in Toronto Housing First project not to identify specific (race or ethnicity), but to adapt the services the category.

Goering PN, Streiner DL, Adair C, Aubry T, Barker J, Distasio J, et al. The at Home/Chez Soi trial protocol: A pragmatic, multi-site, randomised controlled trial of a Housing First intervention for homeless individuals with mental illness in five Canadian cities. BMJ Open. 2011;1(2):1–18.

5. Table 2: the percentage for Increasing FI should be 19.0 (not 18.9)

Thank you. We have corrected it.

6. line 237: what does "and RRR=1.6 [95% CI: 1.0 to 2.7]" refer to? This doesn't seem to match anything in Table 4.

Thanks. We have removed it

7. line 258 "need to for" should be "need for"

We have corrected it.

8. In the Abstract, Discussion, and Conclusion, the authors include PTSD in the list of disorders more likely to be in the persistently FI group, but this finding was not significant (though close). The authors also don't mention this in the text of the Results section.

We removed PTSD in this list of disorders in the Abstract, Discussion, and Conclusion.

We thank Reviewer #2 again for all of these extremely helpful comments!

4 Feb 2020

PONE-D-19-21014R1

Mental and Substance Use Disorders and Food Insecurity among Homeless Adults Participating in the At Home/Chez Soi Study

PLOS ONE

Dear Dr Lachaud,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

The manuscript has been evaluated by two reviewers, and their comments are available below.

The reviewers have raised a number of minor concerns that need attention. They request further clarification on the terminology used, and suggest some grammar corrections.

Could you please revise the manuscript to carefully address the concerns raised?

We would appreciate receiving your revised manuscript by Mar 19 2020 11:59PM. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.

To enhance the reproducibility of your results, we recommend that if applicable 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/plosone/s/submission-guidelines#loc-laboratory-protocols

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled 'Manuscript'.

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 look forward to receiving your revised manuscript.

Kind regards,

Carmen Melatti, PhD

Associate Editor

PLOS ONE

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: (No Response)

Reviewer #2: (No Response)

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: No

Reviewer #2: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: The authors have addressed most of my concerns noted in the previous review. I have noted a few additional or remaining concerns:

In lines 75-77, it is still not clear if the prevalence statistic “ and is estimated to affect more than two thirds of individuals who experience homelessness with a mental disorder” is in reference to Canada, or elsewhere.

The paper is inconsistent in its terminology with regard to if “mental disorders” includes substance use disorders. The first paragraph of the introduction (starting with line 64) implies that mental disorders does include substance use disorders, but elsewhere throughout the paper (including in the title), the authors use the term “mental and substance use disorders” – suggesting that mental disorders is not inclusive of substance use disorders. The language should be edited to ensure that it is consistent throughout the paper.

In line 146, the authors note a statistically significant difference between study participants and participants excluded from the study dataset. The authors should describe what direction this difference is in, e.g. which group shows a greater level of need?

Line 97 should say “with food insecurity trajectories” instead of “on food insecurity trajectories” – this is a minor change but important in terms of not implying causality.

There are grammatical errors throughout the paper, including the word “people” missing after “homeless” in line 85. The sentences in lines 145-151 are confusing to read, due to grammatical errors (e.g. the first sentence uses the phrase “were conducted” twice). The paper should be thoroughly edited to correct such errors.

It is confusing to use the generic term “ethno-racial group” in Table 1. If this refers to people who do not identify as white, I would rephrase it here as something like “member of non-white racial or ethnic group.” In their response to reviewers, the authors write “This self-identified ethno-racial was included in Toronto Housing First project not to identify specific (race or ethnicity), but to adapt the services the category” – I don’t understand what this means. I would also recommend saying “non-white” instead of “no-white” in line 163.

In line 284, the authors should reiterate that the “prior study by O’Campo et al.” also used the At Home/Chez Soi Trial data. It is not that surprising that two analyses of data from the same study found similar findings about the effects of Housing First on food insecurity. Similarly, line 299 should be rewritten to convey that “another study conducted by O’Campo” is not an independent study, but another analysis of the At Home data.

Reviewer #2: The authors have addressed the majority of my earlier concerns. In their response to the reviews (and within the manuscript), authors refer to an Appendix with tables and a figure, but I do not see the Appendix with the submission.

There are also a couple of typographical errors:

line 44: "ressources"

line 117: "symtpoms"

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files to be viewed.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (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. Registration is free. 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 figures@plos.org. Please note that Supporting Information files do not need this step.

5 Feb 2020

Comments to the Author

Reviewer #1: The authors have addressed most of my concerns noted in the previous review. I have noted a few additional or remaining concerns:

Comment: In lines 75-77, it is still not clear if the prevalence statistic “and is estimated to affect more than two thirds of individuals who experience homelessness with a mental disorder” is in reference to Canada, or elsewhere.

Answer: Thank you for asking for this precision. We reformulated it as follows:

A recent study conducted in high-income countries found alcohol dependence and drug dependence are among the most common disorders among the homeless population, with a prevalence ranging from 8-58% and 5-54%, respectively(1). Prevalence estimates for mental disorders, such as psychosis, depression, personality disorder, and post-traumatic stress disorder, are also higher than those reported for the general population of those countries(1,2). For instance, the prevalence of psychosis among homeless individuals (3-42%) is approximately 3 times higher than the estimate in the general population(1).

Comment: The paper is inconsistent in its terminology with regard to if “mental disorders” includes substance use disorders. The first paragraph of the introduction (starting with line 64) implies that mental disorders does include substance use disorders, but elsewhere throughout the paper (including in the title), the authors use the term “mental and substance use disorders” – suggesting that mental disorders is not inclusive of substance use disorders. The language should be edited to ensure that it is consistent throughout the paper.

Answer: The text has been edited to avoid this confusion.

Mental and substance use disorders remain a major public health and social issue among homeless individuals. A recent study conducted in high-income countries found alcohol dependence and drug dependence are among the most common disorders among the homeless population, with a prevalence ranging from 8-58% and 5-54%, respectively(1). Prevalence estimates for other mental disorders, such as psychosis, depression, personality disorder, and post-traumatic stress disorder, are also higher than those reported for the general population of those countries(1,2). For instance, the prevalence of psychosis among homeless individuals (3-42%) is approximately 3 times higher than the estimate in the general population(1).

Comment: In line 146, the authors note a statistically significant difference between study participants and participants excluded from the study dataset. The authors should describe what direction this difference is in, e.g. which group shows a greater level of need?

Answer: Corrected as follows:

Comparisons of socio-demographics characteristics between participants and the excluded group were conducted using student t-tests or Fisher’s test, and showed a statistically significant different only for the variable level of need. Out of the 520 study participants, 48.9% had a high level of need compared to 32.9% in the excluded group (p-value: 0.027, see Table A1 in Appendix). Therefore, we included this variable as an adjustment variable in our models.

Comment: Line 97 should say “with food insecurity trajectories” instead of “on food insecurity trajectories” – this is a minor change but important in terms of not implying causality.

Answer: Corrected

Comment: There are grammatical errors throughout the paper, including the word “people” missing after “homeless” in line 85. The sentences in lines 145-151 are confusing to read, due to grammatical errors (e.g. the first sentence uses the phrase “were conducted” twice). The paper should be thoroughly edited to correct such errors.

It is confusing to use the generic term “ethno-racial group” in Table 1. If this refers to people who do not identify as white, I would rephrase it here as something like “member of non-white racial or ethnic group.” In their response to reviewers, the authors write “This self-identified ethno-racial was included in Toronto Housing First project not to identify specific (race or ethnicity), but to adapt the services the category” – I don’t understand what this means. I would also recommend saying “non-white” instead of “no-white” in line 163.

Answer: line 85 corrected.

lines 145-151: Thank you. We reviewed it and removed the words “were conducted”, which were typed twice. We also reviewed the entire manuscript to be sure such typos errors it doesn’t repeat.

Thank you for the suggestion for ethno-racial group. We rephrased as follows:

“Self-identified ethnic group (white vs. member of non-white/ethnic groups)”

We also changed it in the table 1 and the footnotes of the tables 3 and 4.

Comment: In line 284, the authors should reiterate that the “prior study by O’Campo et al.” also used the At Home/Chez Soi Trial data. It is not that surprising that two analyses of data from the same study found similar findings about the effects of Housing First on food insecurity. Similarly, line 299 should be rewritten to convey that “another study conducted by O’Campo” is not an independent study, but another analysis of the At Home data.

Answer: line 284 corrected. “prior study by..” was replaced by “prior analysis by…”

Line 289 “another study conducted by O’Campo” was replaced by “A previous analysis conducted by O’Campo using At Home/Chez Soi two years of follow-up data found that”

Thank for your insightful comments!

Reviewer #2: The authors have addressed the majority of my earlier concerns. In their response to the reviews (and within the manuscript), authors refer to an Appendix with tables and a figure, but I do not see the Appendix with the submission.

Answer: added it directly at the end of the manuscript

Comment: There are also a couple of typographical errors:

line 44: "ressources"

Answer: Corrected

Comment: line 117: "symtpoms"

Answer: Corrected

Thank for !

7 Apr 2020

Mental and Substance Use Disorders and Food Insecurity among Homeless Adults Participating in the At Home/Chez Soi Study

PONE-D-19-21014R2

Dear Dr. Lachaud,

We are pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it complies with all outstanding technical requirements.

Within one week, you will receive an e-mail containing information on the amendments required prior to publication. When all required modifications have been addressed, you will receive a formal acceptance letter and your manuscript will proceed to our production department and be scheduled for publication.

Shortly after the formal acceptance letter is sent, an invoice for payment will follow. To ensure an efficient production and billing process, please log into Editorial Manager at https://www.editorialmanager.com/pone/, click the "Update My Information" link at the top of the page, and update your user information. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, you must inform our press team as soon as possible and no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

With kind regards,

Markos Tesfaye, M.D., Ph.D

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Thank you for addressing all of the latest comments from the reviewers.

I see a couple of typos in the document that you might wish to correct.

1. Line 168: "non-white racial or ethnic group" probably needs to be removed.

2. (Appendix) Tables A1 and A4 : you might wish to replace "ethno-racial group as in table A3.

Reviewers' comments:

9 Apr 2020

PONE-D-19-21014R2

Mental and Substance Use Disorders and Food Insecurity among Homeless Adults Participating in the At Home/Chez Soi Study

Dear Dr. Lachaud:

I am pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please notify them about your upcoming paper at this point, to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

For any other questions or concerns, please email plosone@plos.org.

Thank you for submitting your work to PLOS ONE.

With kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Prof. Markos Tesfaye

Academic Editor

PLOS ONE