Search results

Our search yielded 3105 records (Fig. 1). Following removal of duplicates, we screened the titles and abstracts of a total of 1911 records, of which 112 full-text articles were assessed for eligibility. A total of 22 studies were included in the final analysis.

Fig. 1 Preferred Reporting Items for Systematic Reviews and Meta-Analyses flow diagram. IBSS, International Bibliography of the Social Sciences.

Study characteristics

An overview of study characteristics is shown in Supplementary Tables 1 (cohort studies) and 2 (cross-sectional studies). Included studies were conducted in Canada (n = 8), Sweden (n = 6), Denmark (n = 2), Turkey (n = 2), Australia (n = 1) and the USA (n = 1); two studies collected data from several low- and middle-income countries (n = 2). Of the included studies, 18 used a cohort design and 4 were cross-sectional. Data sources were either registers (n = 15) or clinical records (n = 7).

In line with our inclusion criteria and review focus, all studies reported on mental health service use by young refugee or asylum seekers. Importantly, all included studies looked only at young refugees (i.e. youth who had been granted residency permits) and not asylum seekers (i.e. pending asylum decision). As such, our findings reflect service use in young refugees and not asylum seekers. Studies looked at refugees who were either exclusively under 18 years old (n = 13), up to 25 years old (n = 8) or exclusively between 19 and 25 years old (n = 1). The number of refugees included in the studies ranged from 60 ^{Reference Betancourt, Newnham, Birman, Lee, Ellis and Layne33} to 41 884 ^{Reference Amin, Rahman, Dorner, Björkenstam, Helgesson and Norredam34} . Studies compared service use with the majority population (n = 17) and/or other migrant youth (n = 3), or had no comparator (n = 4). For our synthesis, we chose to present results that have been adjusted for relevant factors (e.g. demographic and socioeconomic factors) whenever possible.

Quality of included studies

Of the 18 cohort studies included, 11 were rated as high quality, ^{Reference Saunders, Gill, Holder, Vigod, Kurdyak and Gandhi15–Reference Gubi, Sjöqvist, Viksten-Assel, Bäärnhielm, Dalman and Hollander17,Reference Amin, Rahman, Dorner, Björkenstam, Helgesson and Norredam34–Reference Taipale, Niederkrotenthaler, Helgesson, Sijbrandij, Berg and Tanskanen41} 4 as moderate quality ^{Reference Betancourt, Newnham, Birman, Lee, Ellis and Layne33,Reference Kamali, Edwards, Anderson, Duku and Georgiades42–Reference Poyraz Fındık, Çeri, Ünver, Perdahlı Fiş, Rodopman Arman and Beşer44} and 3 as low quality. ^{Reference Kane, Ventevogel, Spiegel, Bass, van Ommeren and Tol45–Reference Karadag and Calisgan47} Moderate- and low-quality studies did not adequately address items assessing confounding and follow-up. Two ^{Reference Mazumdar, Chong, Eagar, Fletcher-Lartey, Jalaludin and Smith48,Reference Saunders, Lebenbaum, Lu, Stukel, Urquia and Guttmann49} of the four cross-sectional studies were considered high quality, adequately addressing all aspects of the JBI appraisal checklist. The remaining two studies ^{Reference Toulany, Kurdyak, Stukel, Strauss, Fu and Guan50,Reference Toulany, Vigod, Kurdyak, Stukel, Strauss and Fu51} were assessed as being of moderate quality, as they did not provide sufficient information on whether confounders were identified and addressed. No studies were excluded based on quality, because the quality of existing literature is an important finding in itself. Full details of the quality appraisal can be found in Supplementary Material C.

Q1. Mental health service utilisation

Of the 22 studies that looked at mental health service utilisation, half (n = 11) covered multiple healthcare service types and half (n = 11) reported only one service. Out-patient services were the most frequently reported (n = 11), followed by emergency services (n = 7), in-patient (n = 6), combined service types (n = 5), psychotropic medication prescription or use (n = 4) and primary care (n = 3). Almost half of the studies (n = 10) defined mental health service use as first contact or initiation of medication, with the remaining half looking at rates or amount of use. In addition to overall rates of service use, several studies sought to understand other factors that might influence service use, including the presenting difficulty/diagnosis (n = 10), country or region of origin (n = 5) and parental presence (n = 3).

Combined service types

Six studies reported mental health service use as a variable that combined several types of services and compared refugee service use with that of the majority population. Studies looked at refugees who were exclusively under 18 years old, ^{Reference Barghadouch, Kristiansen, Jervelund, Hjern, Montgomery and Norredam36,Reference Berg, Ljunggren and Hjern37,Reference Kamali, Edwards, Anderson, Duku and Georgiades42} those aged up to 25 years ^{Reference Gubi, Sjöqvist, Viksten-Assel, Bäärnhielm, Dalman and Hollander17,Reference Amin, Rahman, Dorner, Björkenstam, Helgesson and Norredam34} and those exclusively between 19 and 25 years. ^{Reference Björkenstam, Helgesson, Norredam, Sijbrandij, de Montgomery and Mittendorfer-Rutz38} Overall, these studies showed significantly lower service use in refugees in comparison with their Swedish, Danish and Canadian counterparts. For example, Kamali et al ^{Reference Kamali, Edwards, Anderson, Duku and Georgiades42} found that young refugees were 36% less likely than their Ontario-born peers to have had any mental health service contact (adjusted odds ratio (aOR) 0.64, 95% CI 0.58–0.71). Similarly, Amin et al ^{Reference Amin, Rahman, Dorner, Björkenstam, Helgesson and Norredam34} found that young refugees had a 25% lower likelihood of being treated for common mental health difficulties in comparison with their Swedish counterparts (adjusted hazard ratio (aHR) 0.75, 95% CI 0.73–0.77). Only one study reported overall combined mental health service use in unaccompanied and accompanied minors separately, finding a lower likelihood of service use for both in comparison with Swedish-born individuals (unaccompanied minors aHR 0.65, 95% CI 0.58–0.72; accompanied minors aHR 0.75, 95% CI 0.72–0.78). ^{Reference Björkenstam, Helgesson, Norredam, Sijbrandij, de Montgomery and Mittendorfer-Rutz38}

The three studies that investigated country or region of origin reported that refugees from most regions of origin exhibited decreased use of mental health services in comparison with those belonging to the majority population. ^{Reference Amin, Rahman, Dorner, Björkenstam, Helgesson and Norredam34,Reference Barghadouch, Kristiansen, Jervelund, Hjern, Montgomery and Norredam36,Reference Berg, Ljunggren and Hjern37} For example, Berg et al ^{Reference Berg, Ljunggren and Hjern37} reported that young refugees from low- and middle-income countries (aHR 0.34, 95% CI 0.28–0.42; aHR 0.51, 95% CI 0.46–0.56, respectively) had significantly lower service use rates, whereas those from high-income countries did not differ from the majority population (aHR 0.92, 95% CI 0.70–1.21). Two studies also reported particularly low rates for young people from sub-Saharan Africa, ^{Reference Barghadouch, Kristiansen, Jervelund, Hjern, Montgomery and Norredam36} the Horn of Africa ^{Reference Amin, Rahman, Dorner, Björkenstam, Helgesson and Norredam34} and Syria. ^{Reference Amin, Rahman, Dorner, Björkenstam, Helgesson and Norredam34} One exception to this trend was reported by Amin et al, ^{Reference Amin, Rahman, Dorner, Björkenstam, Helgesson and Norredam34} in which young refugees from Iran demonstrated a slightly higher likelihood of being treated for common mental health disorders than majority peers (aHR 1.15, CI 1.05–1.26).

Only one study looked at the potential impact of age of arrival on mental health service use. Björkenstam et al ^{Reference Björkenstam, Helgesson, Norredam, Sijbrandij, de Montgomery and Mittendorfer-Rutz38} reported that those that had arrived in Sweden over the age of 16 years had a 40% less likelihood of service use compared with the majority population (aHR 0.60, 95% CI 0.53–0.67), whereas those that arrived before age 7 years had only a 20% reduced likelihood (aHR 0.81, 95% CI 0.77–0.85).

Of the three studies looking at the influence of time since immigration, two found that care use increased over time of residence in Sweden. ^{Reference Berg, Ljunggren and Hjern37,Reference Björkenstam, Helgesson, Norredam, Sijbrandij, de Montgomery and Mittendorfer-Rutz38} Berg et al ^{Reference Berg, Ljunggren and Hjern37} reported that this was particularly evident for those who had received residency on the grounds of asylum (aHR <5 years: 0.59, 95% CI 0.48–0.72; aHR 10+ years: 0.74, 95% CI 0.64–0.86) rather than family reunification (aHR <5 years: 0.25, 95% CI 0.17–0.36; aHR 10+ years: 0.37, 95% CI 0.28–0.50). The remaining study found that use was highest in the first 2 years in Sweden (aOR 2.30, 95% CI 2.09–2.54), and that this was largely driven by unaccompanied refugee minors (aOR 3.39, 95% CI 2.96–3.85), although rates were still elevated in accompanied refugee minors (aOR 1.51, 95% CI 1.31–1.79). ^{Reference Gubi, Sjöqvist, Viksten-Assel, Bäärnhielm, Dalman and Hollander17} After 6 years, use for both became significantly lower than that for Swedish youth and, at 12 years, was no longer significant. ^{Reference Gubi, Sjöqvist, Viksten-Assel, Bäärnhielm, Dalman and Hollander17} Notably, this study adjusted only for gender and parental income, whereas the other two also adjusted for age.

All six studies looking at overall service use also explored use according to presenting difficulty. Young refugees generally had lower service use rates for most disorders when compared with majority youth, with this being particularly evident for neurodevelopmental disorders and difficulties related to substance use. ^{Reference Gubi, Sjöqvist, Viksten-Assel, Bäärnhielm, Dalman and Hollander17,Reference Betancourt, Newnham, Birman, Lee, Ellis and Layne33,Reference Björkenstam, Helgesson, Norredam, Sijbrandij, de Montgomery and Mittendorfer-Rutz38} For example, de Montgomery et al ^{Reference de Montgomery, Petersen and Jervelund16} found a significantly lower likelihood of any contact for neurodevelopmental disorders and drug-related contact for both refugee boys (neurodevelopmental aOR 0.03–0.15; drug-related aORs 0.29–0.7) and girls (neurodevelopmental aOR 0.11–0.16; drug-related aOR 0.23–0.40).

In contrast, three studies found significantly higher use for trauma-related difficulties in comparison with the majority population. Björkenstam et al ^{Reference Björkenstam, Helgesson, Norredam, Sijbrandij, de Montgomery and Mittendorfer-Rutz38} found that this changed depending on time since immigration. They reported that refugees with <6 years of residency had a threefold likelihood of care use for PTSD (aHR 3.31, 95% CI 2.40–4.56), whereas those who had lived in Sweden for 10 years or more were only at 1.5-fold likelihood (aHR 1.52, 95% CI 1.25–1.85). ^{Reference Björkenstam, Helgesson, Norredam, Sijbrandij, de Montgomery and Mittendorfer-Rutz38} Two studies also found significantly higher use in refugees for schizophrenia and other non-affective psychotic disorders. ^{Reference de Montgomery, Petersen and Jervelund16,Reference Björkenstam, Helgesson, Norredam, Sijbrandij, de Montgomery and Mittendorfer-Rutz38} For example, Björkenstam et al ^{Reference Björkenstam, Helgesson, Norredam, Sijbrandij, de Montgomery and Mittendorfer-Rutz38} reported that refugee immigrants had 1.5-fold likelihood of use in comparison with their Swedish peers (aHR 1.49, 95% CI 1.29–1.71).

Out-patient services

Eleven studies reported on out-patient mental healthcare use. Nine of these compared service use with the majority population, ^{Reference de Montgomery, Petersen and Jervelund16,Reference Betancourt, Newnham, Birman, Lee, Ellis and Layne33,Reference Axelsson, Bäärnhielm, Dalman and Hollander35,Reference Barghadouch, Kristiansen, Jervelund, Hjern, Montgomery and Norredam36,Reference Poyraz Fındık, Çeri, Ünver, Perdahlı Fiş, Rodopman Arman and Beşer44,Reference Mazumdar, Chong, Eagar, Fletcher-Lartey, Jalaludin and Smith48,Reference Saunders, Lebenbaum, Lu, Stukel, Urquia and Guttmann49,Reference Toulany, Vigod, Kurdyak, Stukel, Strauss and Fu51} two with non-refugee migrants ^{Reference Axelsson, Bäärnhielm, Dalman and Hollander35,Reference Mohamud, Toulany, Grandi, Altaf, Fu and Strauss43} and one conducted a within-group comparison. ^{Reference Toulany, Kurdyak, Stukel, Strauss, Fu and Guan50} Studies included refugees who were exclusively under 18 years old ^{Reference Betancourt, Newnham, Birman, Lee, Ellis and Layne33,Reference Axelsson, Bäärnhielm, Dalman and Hollander35,Reference Barghadouch, Kristiansen, Jervelund, Hjern, Montgomery and Norredam36,Reference Saunders, Gandhi, Wanigaratne, Lu, Stukel and Glazier40,Reference Mohamud, Toulany, Grandi, Altaf, Fu and Strauss43,Reference Poyraz Fındık, Çeri, Ünver, Perdahlı Fiş, Rodopman Arman and Beşer44,Reference Toulany, Kurdyak, Stukel, Strauss, Fu and Guan50,Reference Toulany, Vigod, Kurdyak, Stukel, Strauss and Fu51} and those aged up to 25 years old. ^{Reference de Montgomery, Petersen and Jervelund16,Reference Mazumdar, Chong, Eagar, Fletcher-Lartey, Jalaludin and Smith48,Reference Saunders, Lebenbaum, Lu, Stukel, Urquia and Guttmann49} Nine studies found significantly lower use in refugee youth. For example, both Barghadouch et al ^{Reference Barghadouch, Kristiansen, Jervelund, Hjern, Montgomery and Norredam36} and de Montgomery et al ^{Reference de Montgomery, Petersen and Jervelund16} reported decreased out-patient service use for refugee boys (adjusted rate ratio (aRR) 0.47, 95% CI 0.44–0.51; aOR 0.39, 95% CI 0.34–0.43) and girls (aRR 0.42, 95% CI 0.39–0.46; aOR 0.34, 95% CI 0.31–0.39). This appeared to be consistent across all regions of origin, but particularly for sub-Saharan African youth. ^{Reference Barghadouch, Kristiansen, Jervelund, Hjern, Montgomery and Norredam36} Betancourt et al found no significant difference in clinic-based out-patient therapy but reported that, in comparison with US youth, refugees were significantly more likely to receive in-home counselling (11.8 v. 3.6%, P < 0.05). ^{Reference Betancourt, Newnham, Birman, Lee, Ellis and Layne33}

In contrast, Axelsson et al ^{Reference Axelsson, Bäärnhielm, Dalman and Hollander35} looked specifically at service use by unaccompanied refugee minors and reported significantly higher likelihood of out-patient service use in unaccompanied refugee minors (aHR 1.10, 95% CI 1.01–1.18) in comparison with both the majority population and accompanied migrant minors. They also looked at whether there was a difference in mean time from immigration to out-patient care for unaccompanied refugee minors versus accompanied migrant minors, finding that unaccompanied refugees had significantly lower mean times (1.37 v. 3.10 years, P < 0.001). ^{Reference Axelsson, Bäärnhielm, Dalman and Hollander35}

Two of the above studies also reported specifically on private out-patient service use in comparison with Danish youth, with both reporting significantly lower use in both refugee boys and girls. ^{Reference de Montgomery, Petersen and Jervelund16,Reference Barghadouch, Kristiansen, Jervelund, Hjern, Montgomery and Norredam36} For example, Barghadouch et al ^{Reference Barghadouch, Kristiansen, Jervelund, Hjern, Montgomery and Norredam36} found aRRs of 0.25 (95% CI 0.22–0.28) and 0.16 (95% CI 0.14–0.18) in refugee boys and girls, respectively. ^{Reference Barghadouch, Kristiansen, Jervelund, Hjern, Montgomery and Norredam36} Rates were low across all regions of origin, but particularly for sub-Saharan African children. De Montgomery et al ^{Reference de Montgomery, Petersen and Jervelund16} also reported that the numbers of consultations with private practitioners were also lower in young refugees.

In addition, one study looking specifically at refugee mental health service use in the context of the COVID-19 pandemic reported a 20% increase (aRR 1.20) in mental health service visits for refugees that were new to care during the first year of the pandemic. ^{Reference Toulany, Kurdyak, Stukel, Strauss, Fu and Guan50} This increase was not seen in those who had pre-existing and ongoing contact with services. ^{Reference Toulany, Kurdyak, Stukel, Strauss, Fu and Guan50}

Three studies reported on presenting difficulties in out-patient services. Both studies conducted in Turkey reported neurodevelopmental disorders as the most common presenting difficulty. ^{Reference Poyraz Fındık, Çeri, Ünver, Perdahlı Fiş, Rodopman Arman and Beşer44,Reference Karadag and Calisgan47} For Karadag and Calisgan, ^{Reference Karadag and Calisgan47} 34% of refugee youth presenting to the service was diagnosed with an intellectual disability. For Poyraz Fındık et al, ^{Reference Poyraz Fındık, Çeri, Ünver, Perdahlı Fiş, Rodopman Arman and Beşer44} attention-deficit hyperactivity disorder (ADHD) was the most common presenting difficulty for both refugees and non-refugees. However, they reported that depression and PTSD were more common in refugee children, with all PTSD diagnoses in the service being in the refugee group. Notably, they also reported that the non-refugee group was seven times more likely to receive a diagnosis than the refugee group (1.2 v. 8.8% did not receive a diagnosis). However, both studies were rated poorly in our quality appraisal, and therefore their findings should be interpreted with caution.

The remaining study, ^{Reference de Montgomery, Petersen and Jervelund16} conducted in Denmark, reported significantly fewer and shorter out-patient contacts for affective, neurotic and stress-related disorders in refugee girls compared with non-refugee girls, as well as for schizophrenic disorders in refugee boys compared with non-refugee boys.

In-patient services

Six studies reported on in-patient service use in comparison with the majority population. Studies looked at refugees who were exclusively under 18 years old ^{Reference Betancourt, Newnham, Birman, Lee, Ellis and Layne33,Reference Axelsson, Bäärnhielm, Dalman and Hollander35,Reference Barghadouch, Kristiansen, Jervelund, Hjern, Montgomery and Norredam36,Reference Saunders, Gandhi, Wanigaratne, Lu, Stukel and Glazier40} and those aged up to 25 years. ^{Reference Saunders, Gill, Holder, Vigod, Kurdyak and Gandhi15,Reference de Montgomery, Petersen and Jervelund16} Findings appeared to differ depending on where studies were conducted. All three studies conducted in North America reported no significant difference in service use when comparing young refugees with majority population youth, ^{Reference Saunders, Gill, Holder, Vigod, Kurdyak and Gandhi15,Reference Betancourt, Newnham, Birman, Lee, Ellis and Layne33,Reference Saunders, Gandhi, Wanigaratne, Lu, Stukel and Glazier40} with Saunders et al ^{Reference Saunders, Gill, Holder, Vigod, Kurdyak and Gandhi15} also reporting comparable rates across all places of origin for refugees.

By contrast, two studies conducted in Denmark demonstrated decreased use of in-patient care among refugees. ^{Reference de Montgomery, Petersen and Jervelund16,Reference Barghadouch, Kristiansen, Jervelund, Hjern, Montgomery and Norredam36} Both reported decreased likelihood of in-patient care for both refugee boys (aOR 0.59, 95% CI 0.50–0.70; aOR 0.59, 95% CI 0.43–0.85) and girls (aOR 0.46, 95% CI 0.39–0.54; aOR 0.50, 95% 0.47–0.85, respectively). Barghadouch et al ^{Reference Barghadouch, Kristiansen, Jervelund, Hjern, Montgomery and Norredam36} also looked at the impact of place of origin and reported that, although rates for almost all regions were significantly lower, those for boys from the former Yugoslavia were particularly low (aRR 0.20, 95% CI 0.08–0.49). They also reported increased rates for children from sub-Saharan Africa, but notably these were not significant.

The remaining study investigated service use in unaccompanied refugee minors, and reported a higher likelihood of using in-patient care (aHR 1.30, 95% CI 1.10–1.45) in comparison with both the majority population and accompanied migrant minors. ^{Reference Axelsson, Bäärnhielm, Dalman and Hollander35} Looking at mean time from immigration to in-patient care, they found that unaccompanied refugees had significantly lower mean times in comparison with accompanied migrant minors (1.58 v. 3.86 years, P < 0.001). ^{Reference Axelsson, Bäärnhielm, Dalman and Hollander35}

Only one study looked at the length of hospitalisation by presenting difficulty. ^{Reference de Montgomery, Petersen and Jervelund16} In terms of duration of hospitalisation, it found that refugee boys had significantly longer hospital stays for affective disorders and schizophrenic disorders when compared with their Danish peers. In contrast, refugee girls had significantly shorter hospitalisations for affective disorders than their Danish peers.

Emergency services

Seven studies reported on refugee mental health contact with emergency departments. Studies included refugees who were exclusively under 18 years old ^{Reference Betancourt, Newnham, Birman, Lee, Ellis and Layne33,Reference Barghadouch, Kristiansen, Jervelund, Hjern, Montgomery and Norredam36,Reference Saunders, Gandhi, Wanigaratne, Lu, Stukel and Glazier40} and those aged up to 25 years old. ^{Reference Saunders, Gill, Holder, Vigod, Kurdyak and Gandhi15,Reference de Montgomery, Petersen and Jervelund16,Reference Gill, Saunders, Gandhi, Gonzalez, Kurdyak and Vigod39,Reference Saunders, Lebenbaum, Lu, Stukel, Urquia and Guttmann49} Findings differed depending on where studies were conducted. All four studies conducted in Ontario, Canada reported greater risks or rates of emergency room contact. ^{Reference Saunders, Gill, Holder, Vigod, Kurdyak and Gandhi15,Reference Gill, Saunders, Gandhi, Gonzalez, Kurdyak and Vigod39,Reference Saunders, Gandhi, Wanigaratne, Lu, Stukel and Glazier40,Reference Saunders, Lebenbaum, Lu, Stukel, Urquia and Guttmann49} For example, Saunders et al ^{Reference Saunders, Gill, Holder, Vigod, Kurdyak and Gandhi15} found significantly higher rates of first mental health contact in emergency departments among refugees in comparison with the majority population (aRR 1.17, 95% CI 1.13–1.21).

In contrast, the two studies conducted in Denmark found decreased rates among both refugee boys and girls. ^{Reference de Montgomery, Petersen and Jervelund16,Reference Barghadouch, Kristiansen, Jervelund, Hjern, Montgomery and Norredam36} For example, de Montgomery et al ^{Reference de Montgomery, Petersen and Jervelund16} reported decreased likelihood of first-time emergency room contact in refugee boys (aOR 0.65, 95% CI 0.55–0.76) and girls (aOR 0.49, 95% CI 0.42–0.57), but noted that this was significantly higher in refugee boys presenting with schizophrenia (aOR 1.61, 95% CI 1.22–2.21). Barghadouch et al ^{Reference Barghadouch, Kristiansen, Jervelund, Hjern, Montgomery and Norredam36} also considered place of origin, and reported significantly lower rates in youth from the former Yugoslavia (risk ratio girls 0.37, 95% CI 0.28–0.49; risk ratio boys 0.27, 95% CI 0.17–0.41) and girls from sub-Saharan Africa (risk ratio 0.45, 95% CI 0.30–0.68), but significantly higher rates for girls from the Middle East and North Africa (risk ratio 1.41, 95% CI 1.20–1.65). The remaining study, conducted in the USA, found no significant differences in emergency room contact. ^{Reference Betancourt, Newnham, Birman, Lee, Ellis and Layne33}

Two studies looked at emergency room use by presenting difficulty. Saunders et al ^{Reference Saunders, Lebenbaum, Lu, Stukel, Urquia and Guttmann49} looked at first-contact emergency department visits and reported a higher proportion of visits in refugees versus Ontario-born youth for substance-related disorders (74.5 v. 66.8%), residual self-harm (69.8 v. 50.8%), acute stress (62.1 v. 47.5%) and anxiety (57.3 v. 50.4%). Rates were similar only for mood/affective disorder (38.8 v. 38%) and psychotic disorders (43.1 v. 40.7%). De Montgomery et al ^{Reference de Montgomery, Petersen and Jervelund16} looked at the number of emergency room visits and found significantly greater levels for affective disorders in both refugee girls and boys in comparison with their Danish counterparts. They also reported that refugee girls had signifintly more emergency room visits for neurotic and stress-related disorders.

Primary care

Three studies reported refugee primary care mental health contact. All three included only those refugees who were exclusively under the age of 18 years. ^{Reference Betancourt, Newnham, Birman, Lee, Ellis and Layne33,Reference Kane, Ventevogel, Spiegel, Bass, van Ommeren and Tol45,Reference Fine, Kane, Spiegel, Tol and Ventevogel46} Only one of these studies compared refugees with the majority population, and found no significant difference between refugees and US youth; however, it did report that refugee immigrants were more likely than non-refugees to use primary care (23.3 v. 5.8%, P < 0.05). ^{Reference Betancourt, Newnham, Birman, Lee, Ellis and Layne33} The remaining two studies utilised the same data-set to look at service use within refugee primary healthcare facilities. ^{Reference Kane, Ventevogel, Spiegel, Bass, van Ommeren and Tol45,Reference Fine, Kane, Spiegel, Tol and Ventevogel46} They found that service use rates per 1000 per month were lower among children <5 years compared with refugees aged 5 years and above, ^{Reference Kane, Ventevogel, Spiegel, Bass, van Ommeren and Tol45,Reference Fine, Kane, Spiegel, Tol and Ventevogel46} and were higher among boys (weighted mean rate 1.13, s.d. 1.75) than girls (weighted mean rate 0.80, s.d. 1.20). ^{Reference Kane, Ventevogel, Spiegel, Bass, van Ommeren and Tol45} Notably, both studies were rated as poor during quality appraisal, which should be kept in mind when considering their findings.

Two studies reported mean rates of service use in refugee primary healthcare facilities by presenting difficulty/diagnosis. Because Fine et al ^{Reference Fine, Kane, Spiegel, Tol and Ventevogel46} is an update of Kane et al, ^{Reference Kane, Ventevogel, Spiegel, Bass, van Ommeren and Tol45} only the results of the former ^{Reference Fine, Kane, Spiegel, Tol and Ventevogel46} will be presented. They found that, for both boys and girls under 5 years, the highest rates of visits during the study period were for ‘epilepsy/seizure’, with rates per 1000 per month ranging from 0.33 to 0.59. ^{Reference Fine, Kane, Spiegel, Tol and Ventevogel46} This was followed by intellectual disability, with rates per 1000 per month ranging from 0.02 to 0.11. ^{Reference Fine, Kane, Spiegel, Tol and Ventevogel46} Service use rates for all other presenting difficulties/diagnoses were negligible. ^{Reference Fine, Kane, Spiegel, Tol and Ventevogel46}

Medication

Four studies looked at psychotropic medication use in young refugees, those who were exclusively under 18 years old ^{Reference Axelsson, Bäärnhielm, Dalman and Hollander35,Reference Karadag and Calisgan47} and those aged up to 25 years old. ^{Reference de Montgomery, Petersen and Jervelund16,Reference Taipale, Niederkrotenthaler, Helgesson, Sijbrandij, Berg and Tanskanen41} Three studies found that refugees had lower medication prescription and use across gender and most presenting difficulties. ^{Reference de Montgomery, Petersen and Jervelund16} For example, Taipale et al ^{Reference Taipale, Niederkrotenthaler, Helgesson, Sijbrandij, Berg and Tanskanen41} found young refugees had 57% reduced odds of initiating antidepressants in comparison with their Swedish peers (aOR 0.43, 95% CI 0.39–0.58), and that this was consistent across youth seen for depression, anxiety and PTSD (aOR 0.42–0.49). This appeared to vary slightly according to country of origin, with refugee youth from Somalia (odds ratio 0.70) and Iraq (odds ratio 0.79) being less likely to initiate antidepressant use in comparison with those from the former Yugoslavia. Notably, refugees also more frequently had a lag time >7 days between prescription and dispensing of antidepressants (16.8 v. 8.6%). Taipale et al ^{Reference Taipale, Niederkrotenthaler, Helgesson, Sijbrandij, Berg and Tanskanen41} also reported that having fewer than 5 years of residency in Sweden was associated with decreased initiation of antidepressants (odds ratio 0.76, 95% CI 0.63–0.92), and older age was associated with increased initiation (odds ratio 1.07, 95% CI 1.04–1.10).

The remaining study found that unaccompanied refugee minors (aHR 1.10, 95% CI 1.01–1.19) were more likely to be prescribed medication in comparison with their Swedish peers. ^{Reference Axelsson, Bäärnhielm, Dalman and Hollander35} Unaccompanied refugee minors also had significantly lower mean time to first prescribed medication in comparison with accompanied migrant minors (1.58 v. 4.47 years, P < 0.001). ^{Reference Axelsson, Bäärnhielm, Dalman and Hollander35}

Three studies looked at whether this differed by type of medication. Taipale at al ^{Reference Taipale, Niederkrotenthaler, Helgesson, Sijbrandij, Berg and Tanskanen41} reported that, alongside decreased antidepressant initiation, young refugees were also less likely to be started on other psychopharmacotherapy (51.5 v. 61.8%), particularly anxiolytics (52.3 v. 60%). They also found that hypnotics were more commonly prescribed (34.8 v. 32.8%). ^{Reference Taipale, Niederkrotenthaler, Helgesson, Sijbrandij, Berg and Tanskanen41} Karadag et al ^{Reference Karadag and Calisgan47} reported that an equal proportion of refugee children was prescribed selective serotonin reuptake inhibitors (SSRIs) and antipsychotics (16 and 17%, respectively). Axelsson et al ^{Reference Axelsson, Bäärnhielm, Dalman and Hollander35} reported that, in comparison with their Swedish peers, unaccompanied refugee minors had a higher likelihood of being prescribed all psychotropic medication (aOR 1.27–3.25), except for ADHD medication (aOR 0.06, 95% CI 0.04–0.10). This was particularly pronounced in regard to sedatives, with unaccompanied minors having over three times the likelihood of prescription (aOR 3.25, 95% CI 3.01–3.51). ^{Reference Axelsson, Bäärnhielm, Dalman and Hollander35}

Q2. Access to mental health services

Only one included study looked at factors associated with access to mental health services, and highlighted language ability as a potential barrier to service use. ^{Reference Mohamud, Toulany, Grandi, Altaf, Fu and Strauss43} Mohamud et al ^{Reference Mohamud, Toulany, Grandi, Altaf, Fu and Strauss43} found that refugees were significantly less likely to use virtual mental health visits in comparison with economic immigrants, regardless of their Canadian language ability. However, refugees with non-Canadian language ability had a 9% lower likelihood of virtual care utilisation compared with refugees with Canadian language ability.