A two-tiered search strategy was applied. First, all included studies (12/2019 to 11/2021) of a previous review [17] on COVID-19 health and health-related outcomes, including clinical outcomes such as infections, hospitalisation, mortality but also mental health outcomes or wellbeing and social determinants that affect health, among refugees and other migrant populations worldwide were screened with narrower inclusion criteria (i.e., population, setting, exposure). Second, an update search (12/2021 to 02/2023) in the WHO COVID-19 Research Database (representing 24 bibliographic databases [19]) and The Cochrane Library was conducted, using a search string created along the PECO scheme (see Supplementary Tables S1, S2a,b). Websites of the International Organization for Migration (IOM), the European Centre for Disease Prevention and Control (ECDC) and the European Public Health Association (EUPHA) were searched for further grey literature articles.

We included empirical studies i) investigating refugees (including asylum seekers and IDPs) living in camps or camp-like settings such as reception and collective accommodation centres; and ii) reporting any type of public health and social measures based on the WHO taxonomy [20], such as individual measures, environmental, surveillance, response, or social and physical distancing measures; and iii) reporting any health-related outcomes (physical, mental, child, or maternal health, quality of life, social wellbeing) or social and economic outcomes. German, English and Spanish articles reporting empirical data using any research methodology were included from December 2019 until February 2023. Detailed inclusion and exclusion criteria are listed in the Supplementary Table S3.

Titles, abstracts, and full-texts were screened by two reviewers independently. Conflicts were resolved by discussion or a third reviewer. Subsequently, we applied a third screening step to distinguish between studies reporting intended vs. unintended consequences of public health and social measures. Intended consequences were defined as any impact that reduced SARS-CoV-2 transmission or infection. Only the studies reporting unintended consequences were synthesised. We used the Covidence software for the screening process [21].

JBI checklists appropriate for the study design were used to assess the quality of the included studies [22]. For each study, the scores obtained from two independent ratings were averaged, and studies were grouped based on their scores and classified as high (100%–75% of full score), moderate (74%–50%), and low (<50%) quality studies. Rating discrepancies in case of considerably different scores (i.e., studies rated high by one reviewer and low by another) were resolved by discussion among the team. Questions marked “not applicable” were not included in the overall quality score to avoid artificially downgrading studies.

We extracted bibliographic information (author, year, title), results of the quality assessment, study characteristics (country of study (according to the world bank country classifications by income level [23]), study period, migrant population, sample size and study design/methodology) as well as information on public health and social measures (i.e., description, period of implementation, target of measure (general vs. camp-specific), and context of implementation). To systematically describe unintended consequences, we used the framework of Jabeen [4] to assess their knowability (anticipated/unanticipated), value (positive/negative/neutral), distribution of effects (some/whole population), and temporality (simultaneously/after some time) as well as the relationship of unintended consequences with health, economic, and social outcomes (direct/indirect).

After the extraction we classified the public health and social measures according to the WHO classification, and the unintended consequences according to the CONSEQUENT framework [5]. The framework categorises (unintended) consequences in multiple domains: health, health system, human and fundamental rights, acceptability and adherence, equity and equity-related, social and institutional, economic and resource-related, and ecological, whereas some consequences were not clearly assignable to single domains and therefore count into several domains, so that the percentages below exceed 100%. It also offers a set of potential mechanisms through which they emerge: bio-physiological mechanisms; (re-)action and behaviour change; perception, experience and assessment; available opportunities for (re-)action; environments and environmental exposure; social norms and practices; economic and market mechanisms; the functioning of systems and system components. Based on the frequency of measures identified, we split the public health and social measures into three main categories: a) surveillance and response measures in specific settings, including the protection of specific populations, b) social and physical distancing measures, and c) operational measures involving adaptations or closures of schools, health and social services, businesses, public spaces and administrative offices. We developed mosaic plots using the programming language R 4.3.2 to visualise the linkages between public health and social measures and unintended consequences (see Supplementary Chapter 4) and a diagram to visualise the cascading impacts of public health and social measures inspired by a framework of Carter et al. [24].

As this study is based on published literature, patient and public involvement was not applicable.

The majority of the included studies were qualitative studies (n = 22; 62.9%), while ten (28.5%) were quantitative studies–comprising cross-sectional and cohort studies–and three (8.5%) were mixed-method studies (combining quantitative and qualitative research). In terms of migrant populations, 29 (82.9%) investigated refugees and asylum seekers, and three studies each (8.5%) examined IDPs and refugees in detention facilities. Geographically, the studies were conducted in 18 different countries, including five low-income countries (Burkina Faso, Mali, Rwanda, Somalia, Uganda), five low-middle-income countries (Bangladesh, Jordan, Kenya, Lebanon, Palestinian Territories (i.e., Westbank and Gaza)), and eight upper-middle and high-income countries (France, Germany, Greece, Iraq, Italy, Mexico, United Kingdom, and the United States) (see Table 1).

We found public health and social measures related to different categories of the WHO taxonomy, namely, individual measures (such as using personal protective equipment (PPE) or mask wearing) [27–31], surveillance and response measures (i.e., detecting and isolating cases) [12, 28, 30, 32–38], social and physical distancing measures [12, 25, 26, 28–32, 34, 35, 37–56], and less frequently environmental and biological measures [46] (which are not part of this analysis). Social and physical distancing measures were the biggest group and the main focus of our analysis. This category contains various measures including domestic travel restrictions including stay-at-home orders and mobility restrictions, measures for special populations, measures for offices, businesses, institutions and operations, school closures, and the like.

Across all these public health and social measures we identified 113 unintended consequences. 33.6% (n = 38) of unintended consequences related to health consequences including psychosocial health and wellbeing, and physical health and health behaviours. 23.0% (n = 26) referred to the health system, namely, to access to, utilisation of, and quality of health services, and acceptability and adherence to health services. Economic and resource-related unintended consequences accounted for 15.0% (n = 17), while 8.0% (n = 9) were related to human and fundamental rights, specifically autonomy, self-determination, and privacy. Finally, 28.3% (n = 32) of the unintended consequences were related to social and institutional factors, including the legal and political system, social cohesion and wellbeing, safety, security, and crime, conditions of daily living, and education and development. One unintended consequence referred to an ecological consequence in a low-middle income country due to the use of PPE [28]. Table 2 provides a detailed description of each unintended consequence, as noted in the primary studies.

Figure 2 shows the distribution of unintended consequences by public health and social measures. The graph highlights that unintended physical health consequences were primarily reported in the context of individual measures (Figure 2a), particularly hand hygiene and the use of PPE (Figure 2b). Measures related to gatherings, businesses, and services were reported as having a significant impact on mental health, while social and physical distancing measures were primarily reported with respect to health system-related unintended consequences (Figure 2a). Two additional key findings become evident from the illustration. First, international travel restrictions are reported to have a significant impact on social and institutional outcomes, such as increased exposure to crime and delays in or failures of migration projects (Figure 2b). Second, detecting and isolation measures are the primary reported drivers of unintended consequences, particularly regarding human and fundamental rights as well as the acceptability and adherence to health services (Figure 2b).

This section integrates the three main public health and social measure categories (subsections Surveillance and Response Measures in Specific Settings, Social and Physical Distancing Measures, Operational Measures) with their associated unintended consequences. Figure 3 provides an overarching synthesis of the pathways through which public health and social measures give rise to unintended consequences across all three categories, distinguishing direct and indirect pathways (solid and dashed arrows) and highlights underlying mechanisms (blue squares) linking public health and social measures (yellow polygon) to unintended consequences (red shapes). Guided by this integrative framework, the following subsections examine each public health and social measure in detail. More detailed pathway diagrams for each public health and social measure group are provided in the Supplementary Figures S2a–c.

While the pathway analysis primarily focused on unintended consequences with negative effects, we also identified reports of unintended consequences of a different nature. For example, Akhtar et al. reported that a lockdown in the Azraq refugee camp in Jordan was associated with a decrease in symptom severity among residents with post-traumatic stress disorder [40]. Another study described positive unintended outcomes among Rohingya refugees in Bangladesh, where supervision visits implemented for quality control of public health and social measures and mandatory testing of visitors and staff entering healthcare facilities increased trust in authorities and healthcare providers within the community [28]. In addition, Rohingya refugee women living in Cox’s Bazar, Bangladesh, identified new income-generating opportunities through the production of producing face masks, driven by the mandatory mask-wearing requirements as part of individual-level public health and social measures and shortages of available PPE [29]. Finally, at a detention centre in California, several measures–including suspension of visitations, overcrowding of medical wards, and the imposition of mass quarantine in specific units–let to a hunger strike among detainees [41]. While the value of this unintended consequence is debatable, it illustrates that detainees were able to politically organise despite differences in language, culture, and nationality, as well as the threat of penalties.

Our results extend the existing literature on public health and social measures in two important ways. First, while studies in the general population, specific settings or populations in vulnerable situations often focus on single measures or single outcome domains [6–9, 58] – most commonly mental health or access to services–we show that public health and social measures in camps and camp-like settings lead to interconnected and cascading unintended consequences [35, 53] across multiple domains, including health, economic stability, human rights, and social and institutional factors. Second, unlike previous reviews [58] that examined specific public health and social measures or specific population groups, our study considers all public health and social measures and explicitly traces the pathways and mechanisms through which these measures result in unintended consequences in highly constrained living environments such as camps and camp-like settings.

By applying the CONSEQUENT framework [5], we were able to move beyond a descriptive listing of unintended consequences and systematically identify the mechanisms and pathways through which public health and social measures affect residents of camps and camp-like settings. As mentioned previously, we observed that trust played a central role in shaping pathways and mechanisms, highlighting its crucial role as a foundation of health systems [59], particularly during health emergencies. The framework identifies five “root mechanisms” that should be considered when planning and implementing public health and social measures. These mechanisms call for counteracting actions: taking context into account, ensuring stakeholder buy-in and participation, acting on reliable evidence, addressing root causes, and strategically allocating scarce resources [5].

Using Jabeen’s framework [4] to classify unintended consequences proved challenging, as relevant information in primary studies were often unavailable, and in most cases, the unintended consequences were not clearly described, appearing instead as incidental findings within the result sections. Applying the CONSEQUENT framework as an analytical lens to identify unintended consequences and underlying mechanisms provided valuable insights into the ways public health and social measures impacted residents of camps and camp-like settings. The framework enabled a clear distinction between direct and indirect effects of measures and helped classify the mechanisms linking public health and social measures to unintended consequences. We found that the most impactful mechanisms were those that restricted refugees’ agency and self-efficacy. Understanding and considering these mechanisms and pathways–including restricted agency and self-efficacy of camp residents, mistrust, and contextual constraints–can help explain how public health and social measures lead to unintended consequences in camps and camp-like settings. This knowledge can guide the development and planning of countermeasures, ensuring that negative unintended consequences will be avoided from the outset or addressed promptly if they emerge. Public health and social measures are complex interventions implemented within complex systems and therefore require appropriate, non-linear solutions and careful planning with the engagement of all relevant stakeholders. Simple, one-size-fits-all approaches are insufficient, particularly when multiple forms of marginalisation intersect.

We encountered several challenges in applying existing frameworks to our study. For example, public health and social measures were often not reported in detail, and national or regional lockdown measures, in particular, were rarely specified, which may affect the comparability and generalisability across studies. While some studies described public health and social measures–such as the detention of infected individuals or measures of spatial separation in crowded accommodations–their unintended consequences were often not analysed or supported with data. In most cases, unintended consequences were only reported “incidentally” in the discussion sections of primary studies. Only thirteen out of 35 included studies explicitly aimed to analyse unintended consequences [29, 30, 34, 35, 38, 41–43, 47, 49, 52, 53, 57], and none applied a clear concept or an a priori definition of unintended consequences. Future research should prioritise transparent reporting and systematic evaluation of public health and social measures to strengthen the evidence base for decision-making and to better anticipate unintended consequences, fostering more effective and equitable public health responses in refugee reception and accommodation centres.

The heterogeneity of studies conducted in camps and camp-like settings for refugees, asylum seekers, and IDPs across both high-income and low-income countries limited their direct comparability. Nevertheless, despite these differences, the settings–regardless of the country of study–share common contextual factors such as overcrowding and limited access to social and material resources, which shape living conditions. In addition, camps are often located in remote areas, restricting access to essential services such as healthcare and education. Accordingly, we assume that the outcomes and mechanisms identified in this study are likely to be similar across collective (refugee) accommodations in different countries, as well as in other institutionalised settings that limit the residents’ agency.