Characteristics of trafficked adults and children with severe mental illness: a historical cohort study

Transcription

1 Characteristics of trafficked adults and children with severe mental illness: a historical cohort study Siân Oram, Mizanur Khondoker, Melanie Abas, Matthew Broadbent, Louise M Howard Lancet Psychiatry 2015; 2: Published Online October 18, S (15) See Comment page 1048 Health Service and Population Research (S Oram PhD, M Abas MRCPsych, Prof L M Howard MRCPsych), Biomedical Research Centre (M Broadbent MA), King s College London, London, UK; and Department of Applied Health Research, University College London, London, UK (M Khondoker PhD) Correspondence to: Dr Siân Oram, Section for Women s Mental Health, King s College London, PO31 David Goldberg Centre, London SE5 8AF, UK sian.oram@kcl.ac.uk Summary Background Evidence regarding the mental health needs of trafficked people is limited; however, prevalence of depression and post-traumatic stress disorder is high among trafficked people who are in contact with shelter services. We aimed to investigate the sociodemographic and clinical characteristics of trafficked people with severe mental illness. Methods We did a historical cohort study of trafficked people in contact with secondary mental health services in South London, UK, between Jan 1, 2006, and July 31, We searched and retrieved comprehensive clinical electronic health records for over patients from the Case Register Interactive Search database to identify trafficked patients. A matched cohort of non-trafficked adults was generated by simple computer-generated random selection of potential controls for each case within the parameters of matching criteria. We extracted data on sociodemographic and clinical characteristics and abuse history, and used multiple imputation to deal with missing data. We fitted logistic regression models to compare trafficked and non-trafficked patients. Findings We identified 133 trafficked patients, including 37 children. 78 (81%) of 96 adults and 25 (68%) of 37 children were female. 19 (51%) of 37 children were trafficked for sexual exploitation. Among both adults and children, the most commonly recorded diagnoses were post-traumatic stress disorder, severe stress, or adjustment disorder (27 adults [28%] and ten children [27%]) and affective disorders (33 adults [34%] and ten children [27%]). Records documented childhood physical or sexual abuse among trafficked adults (41 [43%]) and children (28 [76%]), and adulthood abuse among trafficked adults (58 [60%]). Trafficked adults were more likely to be compulsorily admitted as a psychiatric inpatient than non-trafficked adults (adjusted odds ratio 7 61, 95% CI ; p=0 002) and had longer admissions (1 48, ; p=0 045). No association was found between trafficking status and either adverse pathway into care (adjusted odds ratio 0 91, 95% CI ; p=0 82) or substance misuse problems (0 55, ; p=0 12). Interpretation Severe mental illness in trafficked people is associated with longer admissions and high levels of abuse before and after trafficking. Evidence is needed on the effectiveness of interventions to promote recovery for this vulnerable group. Funding Department of Health Policy Research Programme. Introduction Human trafficking is the recruitment and movement of people most often through the use of deception, threat, coercion, or the abuse of vulnerability for the purposes of exploitation. 1 Although evidence regarding the mental health needs of trafficked people is limited, findings suggest a high prevalence of depression, anxiety, and posttraumatic stress disorder (PTSD) among trafficked men and women who are in contact with shelter services. 2 5 To our knowledge, no previous research has been done with a clinical sample of trafficked people. Accordingly, little is known about the sociodemographic and clinical characteristics of trafficked people who have severe mental illness. Therefore, we aimed to investigate the sociodemographic and clinical characteristics of trafficked people with severe mental illness. As a secondary objective, we compared clinical characteristics (adverse pathways to care and length of admission) of trafficked people with severe mental illness with matched non-trafficked patients. Methods Study design We did a historical cohort study of trafficked and matched non-trafficked patients who were in contact with secondary mental health services between Jan 1, 2006, and July 31, The South London and Maudsley National Health Service (NHS) Foundation Trust (SLaM) Biomedical Research Centre (BRC) Case Register Interactive Search (CRIS) system provided data for this study. Details about the development, structure, and function of the BRC CRIS have been reported by Stewart and colleagues. 6 SLaM has almost complete monopoly on the provision of secondary mental health services for four London boroughs: Croydon, Lambeth, Lewisham, and Southwark. 6 The four boroughs have an aggregate population of about 1 2 million people and do not differ substantially from London as a whole with regards to sex, age, education, or socioeconomic status. In 2006, the SLaM established the Patient Journey System, an integrated electronic clinical record used across all SLaM Vol 2 December 2015

2 Research in context Evidence before this study In 2012, Oram and colleagues published a systematic review of 16 studies that reported health outcomes and experiences of violence among trafficked people. The investigators searched MEDLINE, PubMed, PsycINFO, Embase, and Web of Science databases from the date of inception to Aug 31, 2011, using a combination of medical subject headings and text words relating to human trafficking (eg, human trafficking and trafficked people ) and to health (eg, health, mental disorder, and violence ). No language restrictions were used. Oram and colleagues found a high prevalence of symptoms of depression, anxiety, and post-traumatic stress disorder among women who had been trafficked for sexual exploitation and who were in contact with shelter services. On Feb 25, 2015, we searched MEDLINE, Embase, and PsycINFO for reports published since 2011, using search terms including human trafficking, trafficked people, trafficked women, trafficked men, trafficked children and health, mental disorder, mental illness, depression, anxiety, PTSD, and suicide, with no language limits set. Since 2012, a case file review of trafficked men; an analysis of risk factors for mental disorders among female survivors in Moldova; and a survey of men, women, and children in contact with post-trafficking services in southeast Asia have reported similar findings. To our knowledge, no studies have reported on a clinical sample of trafficked people or on trafficked people with severe mental illness. Added value of this study Using an innovative data resource, we provide, to our knowledge, the first evidence of the sociodemographic and clinical characteristics of trafficked people with severe mental illness. Implications of all the available evidence Mental health services care for trafficked people with a range of diagnoses, including psychoses. In addition to experiencing several severe traumas while being trafficked, many of these people had also experienced abuse before trafficking and ongoing violence and threats of harm since escaping the trafficking situation. Findings suggest that severe mental illness in trafficked people is associated with longer psychiatric admissions and high levels of abuse before and after trafficking. Mental health professionals need to be aware of indicators of trafficking and of how to respond appropriately to suspicions or disclosures of this form of abuse. services that provides a comprehensive record of all clinical information recorded during patients contacts with the SLaM. The CRIS database allows for searching and retrieval of anonymised full patient records from the Patient Journey System for over individuals. 6 We identified exposed and non-exposed patients from the CRIS database. Exposed participants were patients at SLaM whose care team had recorded concerns that they might have been trafficked, and non-exposed participants were patients at SLaM whose notes did not suggest that they had experienced human trafficking. We matched exposed and non-exposed participants for sex, age (±2 years), primary diagnosis, type of initial care (inpatient or non-inpatient), and year of most recent service contact. Trafficking was defined in accordance with the UN Protocol to Prevent, Suppress and Punish Trafficking in Persons, Especially Women and Children (ie, the recruitment or movement of people, by means such as force, fraud, coercion, deception, and abuse of vulnerability, for the purposes of exploitation), 1 and included international and internal trafficking. To establish the cohort of trafficked patients, we used trafficking search terms to search the free-text clinical case notes and correspondence of all patients in contact with SLaM services between Jan 1, 2006, and July 31, Patients whose records included one or more trafficking search term were assessed for eligibility by reviewing patients clinical notes and correspondence between the health-care team and other professionals involved in the patients care (eg, general practitioners, social services, solicitors, and voluntary sector services). One researcher assessed eligibility (ie, documented concerns that the patient might have been trafficked as per the UN definition of human trafficking), and a second researcher (SO) independently assessed the first ten records identified by the initial search and an additional random 10% of records. A cohort of non-trafficked patients was generated from the Patient Journey System using simple computergenerated random selection from all potential controls for each case within the parameters of the matching criteria. We aimed for a case:control ratio of 1:4. Ethics approval for the research use of CRIS-derived anonymised databases was granted by an independent research ethics committee (Oxfordshire C, reference 08/H0606/71). An oversight committee reviews all applications to use the CRIS database, and gave approval for this study (reference 11/025). Procedures Unless otherwise specified, measures were recorded in clinical notes at entry into care. Terms used to search free-text clinical notes are listed in the appendix. Patients were categorised as having been trafficked if their free-text clinical notes including, but not limited to, clinical notes made at entry into care suggested that their care team believed they had been or had probably been trafficked; for example, because they were informed by the patient or a third party of experiences compatible with the definition of human trafficking, or that the patient was involved in criminal proceedings against their trafficker, was claiming See Online for appendix Vol 2 December

3 asylum in relation to their experiences while trafficked, or was receiving social services or voluntary sector support as a victim of trafficking. The sociodemographic characteristic of sex is routinely recorded. Age at first contact was calculated by subtracting the date of birth from the date of first contact with SLaM services, both of which are routinely recorded. Information on marital and living alone status was extracted from structured fields (eg, dates, integers, and pick lists) and free-text notes. Patients were categorised as having experienced childhood abuse if free-text notes suggested physical or sexual abuse at age 17 years or younger, and as having experienced adulthood abuse if notes suggested physical or sexual abuse at age 18 years or older. Data on past contact with secondary mental health services and psychiatric inpatient admission were extracted from free-text clinical notes. Secondary mental health services provide specialist care for people with severe mental health problems, based either in hospitals or in the community. Primary ICD-10 7 diagnosis is routinely recorded in the Patient Journey System. For patients whose primary diagnosis changed over time, primary diagnosis at most recent contact was extracted. If no diagnosis was recorded, clinical records were independently reviewed by two consultant psychiatrists (LMH and MA) and a diagnosis assigned. Patients were categorised as having substance misuse problems if one or more of the following were recorded at entry into services: (1) primary or secondary diagnosis of a substance misuse disorder; (2) Health of the Nation Outcome Scales (a clinical outcome measure routinely used and recorded by English mental health services) substance misuse subscale scored at 2 or higher; 8 (3) substance misuse was indicated on standard risk assessment; and (4) current or historical drug use or alcohol misuse was indicated in the clinical free-text notes. Patients were categorised as having deliberately self-harmed if their clinical free-text notes suggested a deliberate self-harm event. An adverse pathway into care was defined as referral to SLaM via either the emergency department or the police service. The referral source is routinely recorded in clinical notes. Inpatient admission and discharge dates are routinely recorded and were used to identify inpatient admissions and total duration of inpatient admission. Dates of compulsory inpatient admission (ie, detained under section 2 or section 3 of the Mental Health Act 1983) are also routinely recorded. Total duration of SLaM care was calculated by subtracting the date of first referral from the date of final discharge, excluding any periods between referrals (upper date limit of Jan 24, 2013). Statistical analysis All analyses were done in Stata version 11. We calculated descriptive statistics proportions for categorical variables and means and SDs for quantitative variables to describe the sociodemographic and clinical charac teristics of the sample. We fitted generalised linear mixed-effects models specifically, random intercept logistic regression models with trafficking status (coded 1=trafficked, 0=non-trafficked) as the outcome variable to compare characteristics of trafficked and matched non-trafficked patients. We included a random intercept model for the matching identifier representing clusters of matched individuals in the logistic models to take account of possible correlation (ie, nonindependence) of matched individuals. History of contact with secondary mental health services, history of psychiatric inpatient admission, substance misuse problems, childhood abuse, adulthood abuse, and total duration of SLaM services were investigated as potential confounders and entered simultaneously in the logistic regression models. We investigated patterns of missing data and used logistic regression to identify predictors of missing data. Missing data on covariates were assumed to be missing at random and the multiple imputation model included all variables used in the presented analysis and those that predicted missing data. We used multiple imputation via chained equations, 9 using the user contributed Stata command ice for imputing missing data. We calculated the percentage of missing cases in our dataset and included that many imputed datasets. 10 We used the xtlogit command to fit random intercept logistic regression models to the multiply imputed data. xtlogit approximates the integral of the likelihood over random effects using the adaptive Gauss-Hermite quadrature method, which is regarded as superior to the alternative approximation methods (eg, the Laplace approximation). We combined estimates and standard errors of the parameter estimates from the imputed datasets according to Rubin s rule 11 and implemented this using the Stata command mim. We did complete case analyses as sensitivity analyses. Role of the funding source The funder of the study had no role in study design or conduct, data collection, data management, data analysis, data interpretation, or writing of the report. The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication. Results Searches of the CRIS database identified 691 patients whose records included one or more trafficking search terms. 558 of the 691 records were excluded (eg, because they related to traffic accidents, drug trafficking, or to patients interest in the issue of human trafficking). The final sample included 133 trafficked patients, 37 of whom were aged younger than 18 years at first contact. For 30 patients, no diagnosis was recorded and so a diagnosis was assigned by LMH and MA. Initial inter-rater Vol 2 December 2015

4 agreement was high (0 97) and in all cases consensus was reached on the diagnosis. Trafficking status, adverse pathways into care, duration of contact with SLaM services, inpatient admission at first contact with SLaM services, and year of most recent contact with SLaM services were associated with missing data. As shown in table 1, 78 (81%) of the 96 trafficked adults were female. 46 (48%) of the sample was aged between 18 and 25 years; the mean age was 26 7 years (SD 6 8, range 18 49). Patients came from 33 countries; the most commonly recorded countries of origin were Nigeria (17 adults [18%]), China (nine [9%]), and Uganda (seven [7%]). 56 (58%) of the trafficked adults had been trafficked for sexual exploitation, with ten (10%) trafficked for domestic servitude, and eight (8%) trafficked for other forms of exploitation (table 1). No details were available regarding type of exploitation for 22 (23%) of the trafficked adult sample. Records that did not include detailed accounts of patients experiences of trafficking typically included other relevant information that suggested the patient was a victim of trafficking; for example, that the patient was involved in criminal proceedings against their trafficker, was claiming asylum in relation to their experiences while trafficked, or was receiving social services or voluntary sector support as a victim of trafficking. Physical abuse during childhood was reported in 29 (30%) adults and sexual abuse during childhood in 23 (24%), and physical abuse during adulthood occurred in 40 (42%) and sexual abuse during adulthood in 43 (45%) of adults. The prevalence of physical and sexual abuse after trafficking was also high (13 [14%] and ten [10%] adults, respectively). Clinical notes suggested that 11 adults (11%) experienced post-trafficking abuse perpetrated by an intimate partner. Table 2 describes the sociodemographic characteristics and trafficking experiences of the trafficked children in contact with SLaM services. 25 (68%) of 37 trafficked children were female. 20 (54%) were aged between 16 and 17 years; the mean age was 14 9 years (SD 2 5, range 8 17). Patients came from 17 countries; the most commonly recorded countries of origin were Nigeria (seven children [19%]) and Afghanistan (five [14%]). Patients notes suggested that 12 (32%) of the children had been trafficked for sexual exploitation and 12 (32%) for domestic servitude or other forms of exploitation. No information regarding type of exploitation was recorded for 13 (35%) children (table 2). Records that did not contain detailed accounts of children s experiences of trafficking typically included other relevant information that suggested the patient was a victim of trafficking. 22 (59%) children reported physical abuse and 19 (51%) reported sexual abuse; 28 (76%) reported either physical or sexual abuse. Table 3 shows the clinical characteristics of the trafficked adults in contact with SLaM services. The most common diagnoses were PTSD, severe stress, or adjustment disorders, diagnosed in 27 adults (28%), and affective disorders in 33 (34%), followed by schizophrenia and related disorders in 14 (15%), which was diagnosed in seven (39%) of 18 males and seven (9%) of 78 females. 21 (22%) of 96 adults had previous contact with secondary mental health services: eight before trafficking and 13 since leaving the trafficking situation. 13 patients Sex Female 78 (81%) Male 18 (19%) Region of origin Europe 24 (25%) Africa 46 (48%) Asia 16 (17%) Other 8 (8%) Unknown 2 (2%) Age (years) (48%) (38%) (15%) Relationship status Single 86 (90%) Partnered 7 (7%) Unknown 3 (3%) Living arrangements Alone 24 (25%) With partner only 6 (6%) With children only 8 (8%) With others 38 (40%) Unknown 20 (21%) Children under 18 years Yes 51 (53%) No 33 (34%) Unknown 12 (13%) Type of exploitation Sexual 56 (58%) Domestic servitude 10 (10%) Other 8 (8%) Unknown 22 (23%) Experiences of child abuse Physical or sexual 41 (43%) Physical 29 (30%) Sexual 23 (24%) None recorded 55 (57%) Experiences of adulthood abuse Physical or sexual 58 (60%) Physical 40 (42%) Sexual 43 (45%) None recorded 38 (40%) Trafficked adults (n=96) Data are number (%). Data provided by the South London and Maudsley National Health Service Foundation Trust Biomedical Research Centre Case Register Interactive Search. Some totals do not add to 100 because of rounding. Table 1: Sociodemographic and clinical characteristics of trafficked adults in contact with secondary mental health services in South London, Vol 2 December

5 (14%) had previously been admitted as psychiatric inpatients: seven before trafficking and six since leaving the trafficking situation. 31 (32%) of trafficked adults had adverse pathways into care (table 3). 54 (56%) of trafficked adults were referred after contact with health services, including primary care (32 adults [33%]) and emergency departments (28 [29%]). Maternity services were the third most frequent source of health-service referrals (eight [10%] of 78 women) after primary care and emergency departments. While under the care of SLaM services, 34 (35%) of the adult sample were admitted as psychiatric inpatients, and 20 (21%) had one or more compulsory admissions (table 3). Current or historic substance misuse problems were documented in 33 (34%) of individuals, and 21 (22%) had one or more deliberate self-harm events while in contact with SLaM services. 77 (80%) of individuals were prescribed psychotropic drugs while under the care of SLaM services, including antidepressants in 54 (56%), antipsychotics in 44 (46%), and hypnotics in 41 (43%). Table 4 describes the clinical characteristics of the sample of trafficked children; to preserve anonymity, Sex Female 25 (68%) Male 12 (32%) Region of origin Africa 20 (54%) Asia 11 (30%) Other 5 (14%) Unknown 1 (3%) Age (years) (16%) (30%) (54%) Living arrangements Alone 7 (19%) With foster parents 13 (35%) With others 13 (35%) Unknown 4 (11%) Type of exploitation Sexual 12 (32%) Domestic servitude or other labour 12 (32%) Unknown 13 (35%) Experiences of child abuse Physical or sexual 28 (76%) Physical 22 (59%) Sexual 19 (51%) None recorded 9 (24%) Trafficked children (n=37) Data are number (%). Data provided by the South London and Maudsley National Health Service Foundation Trust Biomedical Research Centre Case Register Interactive Search. Some totals do not add to 100 because of rounding. Table 2: Sociodemographic characteristics of trafficked children in contact with secondary mental health services in South London, results are not reported where cell counts were fewer than five. As with the adult sample, the most common diagnoses were PTSD, severe stress, or adjustment disorders, (ten children [27%]) and affective disorders Trafficked adults (n=96) ICD-10 primary diagnosis Schizophrenia and related disorders 14 (15%) Affective disorders 33 (34%) PTSD, severe stress, or adjustment disorder 27 (28%) Other* 10 (10%) No disorder 9 (9%) Unknown 3 (3%) Psychotropic medication while a patient at the SLaM Any 77 (80%) Anticonvulsants or mood stabilisers 7 (7%) Antidepressants 54 (56%) Antipsychotics 44 (46%) Benzodiazepines 28 (29%) Hypnotics 41 (43%) History of substance misuse problems Yes 33 (34%) No 31 (32%) Unknown 32 (33%) History of contact with secondary mental health services Yes 21 (22%) No 61 (64%) Unknown 14 (15%) History of psychiatric inpatient admission Yes 13 (14%) No 70 (73%) Unknown 13 (14%) Adverse pathway into SLaM care Yes 31 (32%) No 62 (65%) Unknown 3 (3%) SLaM psychiatric inpatient admission at first contact Yes 15 (16%) No 81 (84%) SLaM inpatient admission Yes 34 (35%) No 62 (65%) Compulsory psychiatric admission while a patient at the SLaM Yes 20 (21%) No 76 (79%) Deliberate self-harm while a patient at the SLaM Yes 21 (22%) No 75 (78%) Data are number (%). Data provided by the SLaM Biomedical Research Centre Case Register Interactive Search. Some totals do not add to 100 because of rounding. PTSD=post-traumatic stress disorder. SLaM=South London and Maudsley National Health Service Foundation Trust. *Including substance misuse disorder. Table 3: Clinical characteristics and illness severity of trafficked adults in contact with secondary mental health services in South London, Vol 2 December 2015

6 Trafficked children (n=37) ICD-10 primary diagnosis Affective disorders 10 (27%) PTSD, severe stress, or adjustment disorder 10 (27%) Other 11 (30%) No disorder 6 (16%) Psychotropic medication while a patient at the SLaM Any 15 (41%) Antidepressants 13 (35%) Antipsychotics 8 (22%) Benzodiazepines 11 (30%) History of contact with secondary mental health services Yes 8 (22%) No 24 (65%) Unknown 5 (14%) Adverse pathway into SLaM care Yes 7 (19%) No 30 (81%) SLaM inpatient admission Yes 8 (22%) No 29 (78%) Deliberate self-harm while a patient at the SLaM Yes 10 (27%) No 27 (73%) Data are number (%). Data provided by the SLaM Biomedical Research Centre Case Register Interactive Search. Some totals do not add to 100 because of rounding. PTSD=post-traumatic stress disorder. SLaM=South London and Maudsley National Health Service Foundation Trust. Table 4: Clinical characteristics and illness severity of trafficked children in contact with secondary mental health services in South London, (ten children [27%]). Eight (22%) children had previous contact with secondary mental health services, and seven (19%) had adverse pathways into care. While under the care of SLaM services, eight (22%) were admitted as psychiatric inpatients. 84 patients who were aged 18 years or older at first contact with SLaM services had a recorded ICD-10 psychiatric disorder and were compared with a randomly selected matched sample of 287 non-trafficked SLaM service users (a ratio of 1:3 4). This ratio is below the target ratio of 1:4 because in some cases the CRIS database did not include as many as four non-exposed matched participants. The proportion of missing data ranged from 0% to 37 4% (with a mean of 6 6%), and 215 (58%) of 371 patients had missing data in one or more variables; thus, we imputed 58 datasets. Findings from the complete case analyses were consistent with those from the analysis of multiply imputed data, with the exception of the association between trafficking status and substance and abuse problems; in complete case analysis, trafficked people showed weak evidence of reduced odds of substance misuse problems (odds ratio [OR] 0 56, 95% CI ; p=0 070) compared with Adverse pathway into SLaM care Compulsory psychiatric admission History of substance misuse problems Duration of SLaM inpatient treatment admissions Unadjusted Adjusted* OR (95% CI) p value OR (95% CI) p value 0 79 ( ) ( ) ( ) < ( ) ( ) ( ) ( ) ( ) Data provided by the SLaM Biomedical Research Centre Case Register Interactive Search. Based on imputed data. Patients were matched for sex, age (±2 years), diagnosis, inpatient admission at first contact, and year of most recent contact. OR=odds ratio. SLaM=South London and Maudsley National Health Service Foundation Trust. *Adjusted for previous contact with secondary mental health services, previous admission as a psychiatric inpatient, substance misuse problems, childhood abuse, adulthood abuse, and duration of contact with SLaM services. Table 5: Logistic regression analyses comparing the clinical outcomes of trafficked and matched non-trafficked adults in contact with secondary mental health services in South London, analysis of the multiply imputed data, which showed no association (0 75, ; p=0 37). The proportion of missing data with regards to substance misuse problems was high at 37 4%, and thus we judged that the complete case analysis was likely to be biased. 12 Table 5 reports comparative analyses based on a random intercept logistic regression model from this matched sample. Trafficked patients were more likely to be compulsorily admitted as a psychiatric inpatient than non-trafficked patients (OR 5 47, 95% CI ; p<0 0001). The association remained after adjusting for potential confounders (OR 7 61, 95% CI ; p=0 002). Trafficked patients also had a longer duration of inpatient stay compared with non-trafficked patients (OR 1 61, 95% CI ; p=0 001), which remained in adjusted analyses (1 48, ; p=0 045). No association was found between trafficking status and either adverse pathway into care (adjusted OR 0 91, 95% CI ; p=0 82) or substance misuse problems (0 55, ; p=0 12). Discussion Research into the mental health needs of trafficked people is limited. 2 Using an innovative data resource, we provide, to our knowledge, the first evidence regarding the sociodemographic and clinical characteristics of trafficked people with severe mental illness. Most of our sample were female and had been trafficked for the purposes of sexual exploitation. This finding is consistent with the national profile of identified cases of human trafficking during the study period. 13 PTSD and affective disorders were the most commonly recorded diagnoses, although schizophrenia and related disorders were recorded for 15% of the adult sample. A fifth of adults and children in our study had a history of contact with secondary mental health services, including for several patients a history of inpatient admission that preceded the trafficking experience. To our knowledge, no previous study has reported on mental Vol 2 December

7 disorders experienced before trafficking. Although mental disorders are likely to be induced or exacerbated by traumatic experiences while an individual is being trafficked, 14 poor mental health might contribute to vulnerability to trafficking, including through social marginalisation and economic insecurity. In addition to documenting a high prevalence of childhood and adulthood abuse among trafficked people with severe mental illness, we found evidence of a continued vulnerability to abuse after escaping exploitation including domestic violence and sexual assault. Ongoing interpersonal abuse seems to contribute to poor mental health 15 and should be taken into consideration during therapeutic interventions and during risk assessment and planning. Trafficked and matched non-trafficked patients pathways into care seemed to be broadly similar, with one exception maternity services emerged as a potentially important route into mental health care for female survivors of human trafficking. Once under the care of the mental health service, trafficked people were more likely to be compulsorily admitted as psychiatric inpatients and had a longer duration of inpatient admission. The reasons for trafficked patients longer duration of inpatient stay are not clear, but might be because of their complex social needs. Addressing social and welfare needs (including meeting basic needs for food, clothing, and appropriate housing; supporting the regularisation of immigration status or return to the country of origin and participation in criminal proceedings against their traffickers; and providing opportunities for education, employment, vocational training, and social integration) seems to be important in predicting mental health adverse outcomes in trafficked people. 16,17 In this study, we used an innovative data resource that provides a unique opportunity to identify patients who are usually difficult to recruit into clinical studies. Psychiatric case registers that include complete electronic health records offer exciting opportunities for future research into the mental health needs of trafficked people. They do, however, have important limitations, including in particular that much information of interest is not recorded in a standardised way. 18 We included patients whose care team had recorded concerns that they might have been trafficked. Clinical records varied in the level of detail regarding patients experiences of human trafficking. For example, in situations in which health-care professionals became aware that their patient had been trafficked through correspondence with another professional (eg, law enforcement, immigration, social services, voluntary sector, or other health professional), less detail was recorded about patients experiences of trafficking. Although establishing the type of exploitation experienced was not possible for a fifth of adults and a third of children in our sample, all returned records were reviewed against the UN definition of human trafficking and against the study protocol, with an independent review of a sample of the records by a second researcher. Records that did not include detailed accounts of patients experiences of trafficking typically included other relevant information that suggested the patient was a victim of trafficking; for example, that the patient was involved in criminal proceedings against their trafficker, was claiming asylum in relation to their experiences while trafficked, or was receiving social services or voluntary sector support as a victim of trafficking. Nonetheless, patients inaccurately referred to as having experienced human trafficking by the professionals involved in their care might have been subsequently misclassified by the research team. Also, a much larger number of trafficked patients were probably not included because the professionals involved in their care were unaware that they had experienced trafficking or had not documented their concerns appropriately. Mental health professionals detection and recording of abuse is low; for example, previous research suggests that mental health services detect between 10% and 30% of domestic violence experienced by their patients. 19 The generalisability of the findings beyond the study setting is unclear further research in other settings is needed. Not all trafficked people will need support from secondary mental health services, or indeed be able to access care. Nonetheless, this study provides urgently needed evidence on the needs of trafficked people with severe mental illness. So far, mental health research has been predominantly done with trafficked people who are receiving shelter or outreach support from civil society organisations and has not included people with psychotic disorders or those admitted for psychiatric care. 14,20,21 The most common mental disorders in this sample affective disorders, PTSD, stress, and adjustment disorders have also been described in previous research with trafficked people. 14 In addition to experiencing several severe traumas while being trafficked, many patients had also experienced violence, abuse, and threats of harm both before and since escaping trafficking. Many will also have complex social and legal needs, are likely to have lost contact with their family and be far from home, and have been disadvantaged in access to education, social activities, and physical health care. Service providers need to be aware of broad strategies to support people with complex trauma. 22,23 Evidence-based inter ventions for PTSD and depression include narrative exposure therapy, cognitive behavioural therapy for depression, trauma-focused cognitive behavioural therapy, and further drug treatment However, the effective ness of these interventions in promoting the recovery of trafficking survivors is unknown. All survivors are likely to need support from agencies with experience helping survivors of trafficking, legal advice regarding decisions about regularising immigration status or returning to their home country, and assistance in participating in criminal proceedings against their traffickers. Mental health services care for trafficked people with a range of diagnoses, including psychoses. Mental health Vol 2 December 2015

8 professionals need to be aware of indicators of possible trafficking and how to respond appropriately to suspicions or disclosures of this form of abuse. Treatment should follow clinical guidelines and take account of abuse suffered before, during, and after trafficking, drawing on models of good practice used with domestic violence, sexual violence, and torture. Evidence is now urgently needed on the effectiveness of mental health interventions for trafficked people. Contributors SO, MA, MB, and LMH conceived and designed the study and acquired data. SO, MK, and LMH analysed and interpreted the data. SO and LMH wrote the first draft of the manuscript. MK, MA, and MB critically revised the manuscript. All authors have given their approval for the publication of this manuscript and agree to be accountable for all aspects of the work in ensuring that the questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. Declaration of interests We declare no competing interests. Acknowledgments SO, MA, and LMH are supported by the Department of Health Policy Research Programme (115/0006). LMH is also supported by a National Institute for Health Research (NIHR) Research Professorship (NIHR-RP-R ) and by the NIHR SLaM BRC-Mental Health. This report is independent research commissioned and funded by the Department of Health Policy Research Programme (Optimising Identification, Referral and Care of Trafficked People within the NHS 115/0006). The views expressed in this publication are those of the authors and not necessarily those of the Department of Health. The study was supported by the CRIS system, funded and developed by the NIHR Mental Health BRC at SLaM and King s College London and a joint infrastructure grant from Guy s and St Thomas Charity and the Maudsley Charity. References 1 UN. Protocol to prevent, suppress and punish trafficking in persons, especially women and children, supplementing the United Nations convention against transnational organized crime, G.A. Res. 55/25(2000) load=true (accessed Oct 1, 2015). 2 Oram S, Stöckl H, Busza J, Howard LM, Zimmerman C. Prevalence and risk of violence and the physical, mental and sexual health problems associated with human trafficking: systematic review. PLos Med 2012; 9: e Turner-Moss E, Zimmerman C, Howard L, Oram S. Labour exploitation and health: a case series of men and women seeking post-trafficking services. J Immigr Minor Health 2014; 16: Kiss L, Pocock N, Naisanguansri V, et al. Health of men, women, and children in post-trafficking services in Cambodia, Thailand, and Vietnam: an observational cross-sectional study. Lancet Glob Health 2015; 3: e Abas M, Ostrovschi NV, Prince M, Gorceag VI, Trigub C, Oram S. Risk factors for mental disorders in women survivors of human trafficking: a historical cohort study. BMC Psychiatry 2013; 13: Stewart R, Soremekin M, Perera G, et al. The South London and Maudsley NHS Foundation Trust Biomedical Research Centre (SLaM BRC) Case Register: development and descriptive data. BMC Psychiatry 2009; 9: WHO. The ICD-10 classification of mental and behavioural disorders: clinical descriptions and diagnostic guidelines. Geneva: World Health Organization, Royal College of Psychiatrists. Health of the Nation Outcome Scales (HoNOS). London: Royal College of Psychiatrists, Royston P. Multiple imputation of missing values. Stata J 2004; 4: White IR, Royston P, Wood AM. Multiple imputation using chained equations: issues and guidance for practice. Stat Med 2011; 30: Rubin DB. Multiple imputation for nonresponse in surveys. New York: John Wiley and Sons, Bennett DA. How can I deal with missing data in my study? Aust N Z J Public Health 2001; 25: Serious Organised Crime Agency (SOCA). Statistical data: national referral mechanism data, gov.uk/publications/national-referral-mechanism-statistics/502- national-referral-mechanism-statistics-end-of-year-summary-2014 (accessed May 5, 2015). 14 Abas M, Ostrovschi NV, Prince M, Gorceag VI, Trigub C, Oram S. Risk factors for mental disorders in women survivors of human trafficking: a historical cohort study. BMC Psychiatry 2013; 13: Trevillion K, Oram S, Feder G, Howard LM. Experiences of domestic violence and mental disorders: a systematic review and meta-analysis. PLoS One 2012; 7: e Patel N, Kellezi B, Williams A. Psychological, social and welfare interventions for psychological health and well-being of torture survivors (review). Cochrane Database Syst Rev 2014; 11: CD Ross C, Dimitrova S, Howard LM, Dewey M, Zimmerman C, Oram S. Human trafficking and health: a cross-sectional survey of NHS professionals contact with victims of human trafficking. BMJ Open 2015; 5: e Perera G, Soremekun M, Breen G, Stewart R. The psychiatric case register: noble past, challenging present, but exciting future. Br J Psychiatry 2009; 195: Howard L, Trevillion K, Khalifeh H, Woodall A, Agnew-Davies R, Feder G. Domestic violence and severe psychiatric disorders: prevalence and interventions. Psychol Med 2010; 40: Hossain M, Zimmerman C, Abas M, Light M, Watts C. The relationship of trauma to mental disorders among trafficked and sexually exploited girls and women. Am J Public Health 2010; 100: Tsutsumi A, Izutsu T, Poudyal AK, Kato S, Marui E. Mental health of female survivors of human trafficking in Nepal. Soc Sci Med 2008; 66: Herman JL. Trauma and Recovery. New York: Basic Books, NICE. Clinical Guideline 26. Post-traumatic stress disorder (PTSD): the management of PTSD in adults and children in primary and secondary care. London: National Institute for Health and Clinical Excellence, Anderson I, Ferrier I, Baldwin R, et al. Evidence-based guidelines for treating depressive disorders with antidepressants: a revision of the 2000 British Association for Psychopharmacology guidelines. J Psychopharmacol 2008; 22: Baldwin DS, Anderson IM, Nutt DJ, et al. Evidence-based guidelines for the pharmacological treatment of anxiety disorders: recommendations from the British Association for Psychopharmacology. J Psychopharmacol 2005; 19: Vol 2 December