Abstract

Human trafficking is increasingly understood as a global public health concern that harms individuals, families, and communities by directly and indirectly causing a multitude of adverse physical and mental health outcomes. 1 Intersecting with a complex range of social determinants of health (eg, income, migration status, social exclusion), 2 human trafficking manifests itself through various forms of commercial sexual exploitation and forced labor. Researchers, policy makers, and survivors have promoted reframing human trafficking, which is traditionally perceived as a law enforcement issue, as a public health issue. 3 The US public health community has broadly responded with high levels of engagement, including delineating research priorities, 4 developing prevention strategies, 5 and advocating for policy changes. 6
Difficult-to-Reach Populations and Use of Public Health Methodologies
Researching human trafficking is challenging because of complex legal definitions, the geographic and demographic diversity of trafficked people, and the inherently hidden nature of the population of interest. The combination of traffickers’ restrictions on trafficked people’s movements, the stigma of self-identification and related criminalization, and a shortage of trained frontline responders creates difficulties in risk identification and monitoring of human trafficking. In 2019, the US National Human Trafficking Hotline identified 11 500 human trafficking incidents nationwide. 7 Although this hotline mechanism is considered one of the best sources of quantitative data available on human trafficking in the United States, it relies strictly on passive data collection (ie, people reporting to the hotline). As such, from a population-level perspective, the number of cases reported will likely represent a small fraction of the total number of trafficked people in the country (estimated to be 400 000 as of 2018). 8 Likewise, passive data from the criminal justice system are likely to underestimate the number of trafficked people. Trafficked people may be reluctant to engage with justice systems in general because of a fear of retaliation from traffickers and a lack of institutional infrastructure to bring cases forward. 9 Instead, this article suggests a shift toward a highly targeted, active surveillance approach. Conducting regular screening and monitoring within locations or organizations where trafficked people are likely to present in order to address other life issues, such as legal or health needs, could improve prospects of successfully identifying trafficked people through interactions with existing ancillary points of contact. 10,11
Many trafficked people have interfaced with health care professionals and visited health care facilities during their experience of being trafficked. 12 -15 Because of health care providers’ interactions with trafficked people and their traditional assessment, recording, and surveillance roles, health care providers are well positioned to help identify trafficked people and provide necessary service referrals and assistance. As a result, more training modules have been developed to educate service providers on human trafficking, 16-18 and more screening tools 19-21 have been created to institutionalize the practice of identifying trafficked people in clinical settings.
Because of the low population prevalence of human trafficking, traditional methods of household or probability sampling would necessitate a large sample size, making this form of data collection and estimation prohibitively complex and expensive. 22 Using methods commonly employed by public health researchers to estimate hidden populations, such as injection drug users and people living with HIV/AIDS, has been suggested as a way to produce more reliable estimates of the size and scope of human trafficking. 23 Studies have generated localized or national estimates of trafficked people that relied on such methods; for example, snowball sampling, 24 respondent-driven sampling, 25 and multiple systems estimation. 26
Sentinel Surveillance Overview
Public health surveillance through the collection, analysis, and interpretation of data is critical to the planning, implementation, and evaluation of public health practice. 27 Adopting a public health approach to anti-trafficking requires systematic surveillance of the problem both locally and globally. 28,29 One alternative surveillance approach that has been previously suggested is human trafficking sentinel surveillance. 30,31 A sentinel surveillance system comprises a network of strategically located nodes (or sentinel sites) where the probability of detecting a condition of interest is higher than average. Sites are selected nonrandomly to sample changes in numbers and characteristics over time. Most commonly applied in the epidemiological surveillance of infectious diseases such as HIV/AIDS 32 -36 and influenza, 37-39 this approach has also been adapted and used for noncommunicable health problems such as domestic violence, 40 occupational injuries, 41 self-harm, 42 and child maltreatment. 43
Sentinel surveillance helps to collect information about trafficked people and their vulnerabilities, the effectiveness of laws and policies, and trends and changes in prevalence and flow that is crucial to anti-trafficking efforts. 44 Sentinel surveillance is often used when population-based case reporting systems are not feasible because of the complexity of necessary data. 45 Because human trafficking is a process that occurs over time and cannot be identified through a traditional confirmatory laboratory or clinical test, more comprehensive information is needed for each case to triangulate negative risk based on a combination of quantitative and qualitative data. A particular advantage of this type of sentinel approach as it relates to human trafficking is the ability to systematically collect detailed information about cases where no one-size-fits-all categorization can be universally evaluated across various locations and contexts. 45 Relying on high-quality information from a limited number of sources where specialized training in case identification takes place is preferable to collecting data from every potential reporting source where such training has not been conducted or prioritized. In limited-resource settings, it may be most effective to concentrate available resources in places with dedicated and experienced personnel who have the capability, opportunity, and motivation to obtain such highly sensitive information.
Sentinel surveillance sites are selected according to numerous criteria, including facility capacity, population representativeness, and geographic dispersal. 46,47 Site selection strikes a balance between optimizing potential for impact and practical considerations such as willingness to participate and financial resource constraints. A sentinel surveillance network for human trafficking would need to comprise sites with a high probability of encountering trafficked people, the necessary infrastructure to support scalable data collection and reporting, and staff members who are sufficiently specialized in victim identification. Within sites, rapid triage tools or decision support systems could assist staff members in their initial screening and identification of trafficked people.
To date, only 2 documented human trafficking studies have explicitly used a sentinel surveillance approach, in which data were collected from deportees using various border crossings as sentinel sites. 44,48 These studies reported that “the lessons learned and applicability of data from sentinel surveillance are numerous, offering insights on hotspot source and destination areas, locality-specific vulnerability factors, and ways to improve the targeting and effectiveness of human trafficking prevention, prosecution, and protection interventions.” 44 Although not novel in its application to human trafficking research, sentinel surveillance is broadly underexplored, particularly through using various types of locations or facilities as data collection sites. Despite being rooted in public health, a sentinel surveillance approach has not been replicated using health care facilities as sentinel sites to understand trends and characteristics of trafficked people in a systematic fashion.
Community Health Centers and Sentinel Surveillance
Federally qualified health centers, which include community health centers (CHCs) that serve medically underserved areas and/or at least 1 of 3 special populations (migratory or seasonal workers, homeless populations, or residents of public housing), provide treatment and services to a disproportionate number of medically underserved populations throughout the United States. The populations served by CHCs often have vulnerabilities that overlap with the most common risk factors for human trafficking, such as migration status, sexual orientation, a history of maltreatment, low educational attainment, socioeconomic status, and high-risk behaviors (eg, substance use). 49,50 More than 90% of CHC patients are lower- to low-income status, and more than one-fifth are uninsured. 51 Seasonal workers in the agricultural industry and homeless people are at increased risk of being trafficked: in 2019, more than 1 million CHC patients were agricultural workers and almost 1.5 million patients were homeless. 51 CHCs coordinate care for patients who experience complex medical, behavioral, and social needs, making them well suited to serve and identify people who have been trafficked. 52 Many CHCs are better equipped than other health care settings to provide resources and “enabling services” (ie, nonclinical services) to trafficked people because of multilingual staff members, case management systems, transportation services, environmental risk reduction, outreach activities, and legal assistance. 53 CHCs typically use a trauma-informed model of care, with high levels of cultural competency that are conducive to ensuring continuity of care for trafficked people. 54
CHCs can play a role in providing services to victims at potentially all stages of the human trafficking cycle. As one physician advocate noted, “It’s not a question of if health centers are treating trafficked victims. We are seeing these patients period.” 55 For example, CHCs can be a first point of contact for a person who is at risk of human trafficking as the person attends scheduled appointments. Trafficked people may come to a CHC for treatment related to physical injuries or mental health problems while they are still being trafficked. Finally, CHCs could play an important role in providing services toward rehabilitation post-trafficking.
Data Collection Infrastructure and Standardization
Primary data on human trafficking are collected by various actors, for various purposes, and with differing mandates for data sharing, and these data streams are rarely interlinked. 56 A lack of standardization in definitions, procedures, and data collection tools prevents any meaningful analysis over time or comparison between areas. 57 Within a network of sentinel sites, a standardized operational definition of human trafficking needs to be adopted to improve the accuracy of surveillance and related responses. A common protocol is necessary to ensure comparability, reliability, and uniformity among sites. Standardization enables systematic training in the collection and reporting of data. It also overcomes barriers to surveillance by improving health care provider education, enabling incentivization for surveillance, reducing complexity of reporting, and enabling systematic and timely feedback. Solutions such as paper-based survey instruments or mobile phone–based questionnaires can be designed to support proactive and consistent screening for signs of human trafficking. Integration with technical systems can facilitate the capture, collation, and analysis of data generated from these screening procedures.
In June 2018, the National Center for Health Statistics released new International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM) codes that support enhanced data collection on human trafficking by allowing health care providers to differentiate trafficked people from people who have experienced different forms of abuse, such as interpersonal violence (Table). 58 These additions included T codes (category for unspecified injuries and external causes) for reporting suspected and confirmed cases of forced labor and sexual exploitation and Z codes (category for social determinants of health) for the observation and examination of trafficked people. 59 At the start of calendar year 2020, human trafficking ICD-10-CM codes were added to the Health Resources and Services Administration (HRSA) CHC Universal Data System (UDS). 60 The codes were added to table 6a of the CHC UDS (see p. 66), which focuses on selected diagnosis and services based on the assumption of significant prevalence of trafficked people among CHC patients or of sentinel interest to HRSA. 61
ICD-10-CM codes for human trafficking a
Abbreviation: ICD-10-CM, International Classification of Diseases, Tenth Revision, Clinical Modification.
aData source: American Hospital Association. 58
The addition of 20e Human Trafficking to table 1 of the HRSA CHC UDS was both a timely and welcome addition, because all CHCs must use the UDS as their foundation within electronic health records (EHRs), indicating a drive to capture human trafficking data systematically across all HRSA-funded CHCs. One study that attempted to examine human trafficking through an EHR data extraction before the implementation of these human trafficking–specific ICD-10-CM codes found this approach to be “next to impossible.” 62 Using these codes to track diagnosis and related symptoms can further epidemiological understanding of the incidence of trafficking in the United States. 63 It can also prove useful for reimbursement mechanisms that can account for otherwise nonbillable anti-trafficking measures, such as identification, intervention, and referral of trafficked people. 63 Moreover, establishing baseline data on the systemwide number of patients encountered who are suspected or confirmed trafficking victims is necessary to design robust research studies with sufficient statistical power and sampling frames. 62 Baseline data are crucial in describing characteristics, inferring causality, and developing effective responses to address human trafficking and root causes of risk.
Practical Obstacles to Implementation
For human trafficking cases to be reported in the UDS, they would need to be diagnosed as part of a CHC visit by a clinical provider of health care services. An enabling service provider such as a case manager, health educator, or legal services provider cannot make a diagnosis in the UDS because the interaction does not qualify as a patient visit. Diagnoses in nontraditional settings, such as self-referrals that do not result in a service provider visit, would not be counted as human trafficking in the UDS. The UDS also has an “own field” clause related to diagnosis of human trafficking, in which diagnoses are only applicable to field expertise. 61 Human trafficking must be identified during a qualified visit by a categorically qualified clinical provider of health care services, inevitably diminishing reporting and related response.
Medical–legal partnerships (MLPs) have been promoted in CHCs, and legal services can be paid for as an enabling service of CHCs. 64 As exemplified by the MLP model, attorneys can be part of the health care team and have expertise in human trafficking response services (eg, T visas, which afford trafficked people temporary nonimmigrant status), but attorney referrals do not count as a qualified visit, nor are attorneys considered as service providers who are qualified to identify human trafficking, even if they have the relevant expertise. An attorney performing enabling or ancillary services who identifies human trafficking or performs legal actions linked to the CHC (eg, applying for a T visa) would not directly be counted as human trafficking in the UDS. Not all services or diagnoses by service providers are counted as visits in the UDS and services or diagnoses by “nonproviders” (ie, people not contributing to health care or medical services, or people who are not qualified clinical providers of health care services) are not counted as visits in the UDS.
Another obstacle is that service providers may be hesitant to use the ICD-10-CM codes and document related information about human trafficking in records because of concerns about confidentiality and harm that may affect the patient. Sensitive information, including potentially stigmatizing information contained in codes or descriptions, may be seen by people outside the health care team, such as coders, health insurance workers, or even traffickers themselves. These obstacles collectively create limitations on implementation because not all people who have expertise in human trafficking and identify a trafficked person are able to report it in the UDS, and health care provider hesitancy to use human trafficking codes may lead to an underestimation of human trafficking cases.
Opportunities
Although surveillance has traditionally been undertaken by physicians, other health care staff members and social service providers can be included as sources of surveillance-related information. Using these other sources has been suggested as a way to overcome deficiencies in reporting and adequately capture data by widening the net of who is able to take part in surveillance data collection. 65 Another general barrier to identification even with properly motivated service providers is the power dynamic that makes trafficked people reluctant to disclose any sensitive information. Screening tools that are designed with trafficked people in mind are important to overcome this consideration, and digital technology may provide a confidential, anonymous, and accessible way for trafficked people to self-report exploitation. 66 Such tools could assist in the preliminary screening of potential trafficked people; for example, with trained volunteers in community networks triaging cases that are then referred to case management in the CHC when severe risk factors are identified. This method of triaging cases could help improve identification closer to where exploitation actually occurs, such as workplaces that are far removed from traditional health care facilities.
Data on human trafficking can be captured to a greater extent by expanding who is able to record and report cases, by overcoming training barriers through the use of digital technology, and by moving toward hybrid systems such as telehealth visits. Further recommendations on operational aspects can be found in the Centers for Disease Control and Prevention’s Staged Development Tool for Sentinel Surveillance. 67 This tool helps to assess current capacity in public health activities and to develop a roadmap for achieving a high level of functioning. By providing an overview of system characteristics such as strategic direction, engagement, and impact, improvements made across these dimensions can help sentinel sites progress from basic to advanced. 67
Conclusion
Sentinel surveillance has multiple theoretical and practical characteristics that make it well suited for application in systematic data collection and understanding trends and patterns of human trafficking. Previous studies using this approach have shown promise, but application of this approach has been limited. Trafficked people regularly interface with the health care system in some fashion, and a growing impetus for increased education and training for service providers has developed in the past few years. The combination of patient populations served in CHCs, the enabling services CHCs provide, and recent developments in data collection infrastructure outlined here makes CHCs ideal facility types for the scalable operationalizing of sentinel surveillance of human trafficking in the United States.
Footnotes
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
