Symptom Trajectories of Youth in Live-In Treatment Facilities Before, During and After the COVID-19 Pandemic

ABSTRACT

COVID-19 and contagion mitigation efforts forced health care workers to very quickly develop and implement new policies aimed at reducing virus transmission. Such safety measures, while helping to reduce COVID-19 may result in decreased therapeutic effectiveness. We examined the impact of COVID-19, as well as several individual differences on the mental and behavioral health trajectories of 176 youths residing in several live-in treatment centers (M age = 15.30 years). Longitudinal data were collected at Intake, Discharge and 6-months post treatment, with timepoints spanning before and after COVID-19 lockdown. We observed an interaction between COVID-19 and prior substance use; youth who reported drug or alcohol use in the 30 days prior to Intake had worse treatment trajectories compared to those without prior use, but only when treatment started before lockdown mandates were implemented. For youth who entered treatment after lockdown had started, there was no difference in treatment outcomes between those with and without prior substance use. We also found an effect of family functioning whereby youth with more adaptive family functioning had better treatment trajectories. Practical implications are discussed.

KEYWORDS:

• Treatment trajectories
• COVID-19
• youth

Practice Implications

• Substance use impacted treatment trajectories; a multi-pronged approach is needed for adolescents with comorbid mental health and substance use issues

• Substance use interacted with lockdown mandates suggesting that the effect of youth substance use may be impacted by mitigation efforts

• Adaptive family functioning improved treatment trajectories and should be a main focus of the treatment journey for youth

The World Health Organization declared COVID-19 a global pandemic in March 2020. Beyond the physical ramifications of COVID-19, there has been a growing number of studies pointing to the influence of social distancing and lockdown mandates on peoples’ mental health (e.g., Li et al., 2021; Xiong et al., 2020). As mandates have been lifted and people have adjusted to post-pandemic life, it is imperative to determine those who may have been disproportionately impacted. Research suggests that certain sub-groups of the population, including adolescents (Chadi et al., 2022; Fancourt et al., 2021), individuals with preexisting mental health issues (Fancourt et al., 2021), and those in live-in treatment facilities (Edwards et al., 2020), may have suffered the most. The current study examined the impact of COVID-19, as well as several individual differences on the mental and behavioral health trajectories (assessed at intake, discharge, and six-months follow-up) of adolescents in several live-in treatment facilities.

COVID-19 and Adolescents

Adolescents may have been disproportionately impacted by the COVID-19 pandemic (Chadi et al., 2022; Fancourt et al., 2021). Lockdown measures resulted in virtual classes, isolation from friends and teachers, increased time spent at home, and cancellation of important life milestones. These changes may amplify risk-factors for adverse mental health outcomes, such as social isolation, loneliness, family conflict, and lack of access to educational and social supports (Patel et al., 2007). Christ and Gray (2022) examined long-term implications of COVID-19 and found that 9-months into the pandemic, even after lockdown mandates had lessened, adolescents reported high levels of loneliness. In fact, over half of the sample indicated that their loneliness was more pronounced at that point in time than it was pre-pandemic. Previous research has found that social isolation in adolescents is associated with symptoms consistent with post-traumatic stress disorder (Sprang & Silman, 2013), as well as increased risk of suicide (Calati et al., 2019). Some research published since the pandemic supports this, finding increased rates of emergency department visits and hospitalizations for suicidality during spring 2020 (e.g., Hill et al., 2021), while others have reported decreased rates (e.g., Kim et al., 2023). Taken as a whole, social distancing and disease containment efforts may be harmful for adolescents.

Adolescents with Pre-Existing Mental Health Problems

There is some evidence to suggest that adolescents with preexisting mental health problems have struggled more than those without (e.g., Cost et al., 2022). For instance, results of surveys from more than 2000 youth in the UK, found that those with prior mental health difficulties reported decreases in their well-being. In March 2020, 32% of the respondents indicated that COVID-19 had made their mental health difficulties “much worse,” and by June 2020 this had increased to 41% (Young Minds, n.d.). In a cross-sectional study examining Canadian teenagers, Cost et al. (2022) found that all adolescents reported deterioration of mental health during the first wave of COVID-19, with rates of deterioration highest for those with a preexisting mental health diagnosis.

Impact of COVID-19 on Adolescents in Live-In Treatment Facilities

More is becoming known about the mental health consequences of COVID-19 for adolescents in the general population, however, less is known about clinical samples. Perhaps not surprisingly, the pandemic has highlighted gaps in mental health services. Palinkas et al. (2021) for instance, conducted interviews with State Mental Health Authorities representing 21 different states. Many respondents discussed a decreased capacity to deliver services, resulting from staff being sick with COVID-19, refusals to work due to fear of infecting family members, and staff burnout. In many states, the pandemic exacerbated a shortage of trained providers, a trend that many mental health care service providers had already observed for years (Palinkas et al., 2021).

In terms of health services, the pandemic forced health care providers to very quickly develop and implement new policies aimed at reducing virus transmission. Included in the mandates were social distancing protocols, cancellation of group activities, and establishment of cohort units (Government of Canada, 2021). Bojdani et al. (2020) suggested that such safety measures, while helping to reduce COVID-19 transmission, are likely to result in decreased therapeutic effectiveness. For example, many facilities were forced to temporarily end face-to-face contact with parents, an important source of support for adolescents in live-in treatment. Adolescents in treatment at the time of the first lockdown may therefore have experienced significant upheaval to their treatment regimens compared to those who started later, when care providers may have had more time to adjust their practices.

Individual Differences and COVID-19

Gender

While the nature and timing of the pandemic may have impacted inpatient treatment services, symptom trajectories may be further impacted by individual differences. Gender, for instance, has long been known to influence the presentation of symptoms. Research emerging since COVID-19 suggests that adolescent girls have been disproportionately negatively impacted compared to boys. For instance, throughout the pandemic, girls were found to feel more isolated, stressed, and unsafe (J. Wang et al., 2021). Additionally, more so than boys, girls indicated that the pandemic negatively impacted their day-to-day life, as well as their physical and mental health (Halldorsdottir et al., 2021). Consistent with pre-pandemic gender differences, rates of depression have been found to be higher for adolescent girls, compared to boys (e.g., Halldorsdottir et al., 2021). There is also some evidence that externalizing problems may have been exacerbated during lockdown, however, gender patterns are mixed. For instance, Craig et al. (2023) found that gender diverse youth reported higher levels of ADHD compared to males and females, while both gender diverse youth and females reported higher levels of oppositional defiant disorder. There were no gender differences in conduct disorder. It is therefore important to consider gender in the context of COVID-19 when examining treatment trajectories in order to inform effective treatment strategies.

Family Functioning

Another variable known to impact symptom presentation and treatment, is family functioning. Adaptive family functioning is related to positive youth outcomes, including positive mental health, well-being, life satisfaction, and self-esteem (Hakvoort et al., 2010). Conversely, adolescents who grow up in families characterized by maladaptive family functioning are more likely to suffer from depression and anxiety, as well as behavioral problems such as antisocial behavior, delinquency, and substance use (Hakvoort et al., 2010; Sijtsema et al., 2013). In terms of treatment trajectories, research widely supports that more adaptive family functioning, as well as parental involvement in the treatment process is related to treatment success (e.g., Ciao et al., 2015; Creighton & Mills, 2016).

During the pandemic, families with young children and adolescents faced immense stress and hardships related to school and daycare closures, financial and employment uncertainties, as well as fears related to COVID-19 (Coller & Webber, 2020; Gassman-Pines et al., 2020). Gadermann et al. (2021) found that parents with children under the age of 18 reported significantly worse mental health than those parents with older children, as well as increased fear of domestic violence and abuse, increased alcohol consumption, and suicidal thoughts. The negative impacts on parents may have a cascading effect on children. For instance, youth who experience parental violence and abuse are at increased risk of subsequent mental health problems (Cicchetti & Toth, 2016; Moretti & Craig, 2013) and parents’ heavy episodic drinking has been linked with increased likelihood of problematic drinking among adolescents (Homel & Warren, 2019). As we move forward from the pandemic, it is important to consider that consequences may be long-lasting, as family dynamics and relationships have been re-structured and re-organized (Prime et al., 2020). Given the impact of COVID-19 on families, it is important to examine the interplay between family functioning and the pandemic when examining treatment trajectories.

Substance Use

Another variable we will consider when examining treatment trajectories is substance use. Adolescent substance use and misuse is associated with several mental health concerns (Hovens et al., 1994) and is common in adolescent psychiatric patients (Mangerud et al., 2014). Substance use is also known to impact treatment outcomes and risk of remission. In a review of the literature by Williams et al. (2000), one of the best predictors of unsuccessful treatment outcomes was higher levels of pre-treatment substance use and in a 3-year longitudinal study of adolescents with psychiatric disorders, smoking and trying illicit drugs were associated with persisting psychiatric disorders (Gardvik et al., 2021).

Findings related to substance use and the pandemic are mixed. There is some research to support a decrease in substance use during the early phases of school closures and lockdowns (See Layman et al., 2022). Dumas et al. (2020) found an overall decrease in adolescent substance use during the pandemic, however, of those adolescents who continued to use, the frequency of alcohol and cannabis use increased. Others have found that while use of certain substances decreased (e.g., alcohol), use of others increased (e.g., nicotine; Pelham et al., 2021). Whether an adolescent continues to use substances during lockdowns may be dependent on a variety of factors. For instance, in a large sample of Canadian adolescents, Craig et al. (2023) found that those adolescents who did engage in alcohol and marijuana use were less concerned about the effects of COVID-19 on the population’s health. The authors suggest that adolescents may be using substances as a recreational activity, regardless of COVID-19 safety protocols. Given these inconsistent findings, more needs to be known about the interplay of COVID-19 and substance use. In this study, we explored this relationship by examining the interaction between substance use and the timing of pandemic mandates.

Current Study

The number of studies examining the impact of COVID-19 on adolescent mental health is growing, however, have been criticized for largely relying on cross-sectional approaches (Bouter et al., 2022; Chadi et al., 2022). Accurate assessment of the impact of the pandemic and contagion mitigation efforts requires longitudinal data that includes pre and post pandemic onset timepoints. Of the longitudinal studies that have been completed, most use community samples of adolescents (e.g., Liu et al., 2021). The present study will address these gaps by examining longitudinal indicators of mental health (assessed at intake, discharge and 6-months follow-up) in a clinical sample of adolescents with timepoints that include pre and post pandemic onset.

Method

Participants

Data were collected between January 2018 and January 2022 from numerous live-in service providers in the U.S. as part of a large on-going research initiative developed to understand and optimize treatment outcomes. Service providers vary across locations, participant profiles, and therapeutic approaches, however, all offer live-in services for adolescents dealing with mental health, externalizing, or relationship issues. The final dataset used for the present study included 176 adolescents from 22 agencies.

Procedure

We examined symptom change trajectories across intake, discharge, and 6-months follow-up. To examine the impact of the COVID-19 lockdown, participants were classified as having intake and discharge timepoints before or after April 1, 2020, which was the date by which all U.S. States had instituted lockdown measures. This resulted in three COVID-19 groups: adolescents who started and ended treatment before lockdown (PRECOVID_ONSET), those who started and ended treatment after lockdown mandates had been implemented (POSTCOVID_ONSET), and those who started treatment before the lockdown, but were discharged after, meaning the lockdown was mandated while they were in treatment (DURINGCOVID_ONSET). All participating agencies enter participant data into a central data warehouse that is protected and governed by a dedicated research center at the University of New Hampshire. The protocol was approved at the Research Ethics Board at the University of New Hampshire. All adolescents in the sample consented to contribute data.

Materials

Youth Outcome Questionnaire (Y-OQ-SR; Ridge et al., 2009)

The Y-OQ-SR is a 64-item self-report questionnaire assessing psychosocial, relational, and behavioral distress and is commonly used to measure treatment progress and outcomes. The “total” distress score is a constellation of intrapersonal, somatic, interpersonal, social, and behavioral dysfunction as well as critical health items (e.g., suicidal ideation). Higher scores indicate greater distress, with total scores ≥ 47 considered clinically problematic. The Y-OQ-SR has good internal consistency, test-retest reliability, and concurrent validity (e.g., Ridge et al., 2009), as well as excellent sensitivity to change (McClendon et al., 2011). In the present study we will use the total score to assess treatment trajectories.

McMaster Family Assessment Device (FAD; Epstein et al., 1983)

Family functioning (e.g., feelings of acceptance and cohesion within the family) was rated with the 12-item general family functioning sub-scale from the McMaster Family Assessment Device, with higher scores indicating more maladaptive levels of family functioning. The tool has been widely used in research and clinical settings and demonstrates good to excellent reliability (see Hamilton & Carr, 2016) and adequate convergent and divergent validity (Miller et al., 1985).

Substance Use

Participants were asked to report the frequency of alcohol or drug use over the last 30 days on a 7-point Likert type scale with the following responses: 1 (Not at all), 2 (Less than once a month), 3 (Once a month), 4 (A couple of times a month), 5 (Once a week), 6 (A couple of times a week), and 7 (Daily). Due to a small number of participants within each of the 7 possible responses participants were dichotomized into two groups: Yes Drug Use over the last 30 days and No drug use over the last 30 days.

Gender and Age

Age at intake was calculated by subtracting date of birth from date at intake. Gender was operationalized by two means of assessment. At intake, service providers documented youth’s gender as Male, Female, or Unknown. Participants were also given the opportunity to self-identify their gender with the following question: “Which of the following choice best describes your gender identity?” with choices including “male,” “female,” “gender fluid,” “not sure,” and “I identify as (please specify).” All participants were coded as “male,” “female” or “gender diverse.” In the case of discrepancy between staff and self-reports, self-identification was given priority.

Statistical Analyses

To examine trajectories of change in self-reported psychosocial distress (e.g., Y-OQ-SR total) across treatment, we fit a mixed effects model using R software (R Core Team, 2022). Prior to modeling, the general linear and non-linear trajectories for the Y-OQ-SR total were examined to analyze time trends both within and between subjects. A non-linear trendline was observed, which is more suitably modeled by a quadratic relationship. Likelihood ratio tests were conducted to compare linear growth curve models to quadratic growth curve models, and in all cases the quadratic fit was superior. We observed sufficient variability between the intercepts (mean Y-OQ-SR total scores at intake) and the slopes (change in scores within a subject across intake, discharge and 6-months follow-up) to fit a random slope model. Fixed parameters included the intercept, the linear slope, and the non-linear slope of the model, which was included using raw polynomials due to the presence of a quadratic relationship. In addition to the fixed parameters, we also modeled five covariates including COVID-19 group, gender, drug use, age, and family functioning. All categorical groups were dummy coded prior to analyses and references categories were made for comparisons. Reference categories included: COVID-19 group (PRECOVID_ONSET), gender (Female), and drug-use (Yes prior drug use). For instance, results related to COVID-19 group will be expressed as comparisons to the PRECOVID_ONSET group. To determine whether the effect of COVID-19 may have differed based on the sub-groups of the other variables, we examined estimated marginal means and pairwise comparisons. To determine the Random effects we included participant to account for intra-individual differences in the intercept and slope across the three time-points. All models were fit using restricted maximum likelihood (REML) and all p-values were derived by approximating the degrees of freedom using the Satterthwaite method (Satterthwaite, 1946).

Results

Descriptive Statistics

The initial data set consisted of 2003 adolescents at intake, 1157 adolescents at discharge, and 262 adolescents at six-month follow-up. When we matched across timepoints on all relevant measures our final sample size was 176, with a mean age of 15.30 years, (SD = 1.48). At intake, there were no significant differences in YOQ-SR scores, gender, or age between participants in our matched sample and those who left treatment early. On average, participants spent 355.88 days in treatment. There were 87 adolescents who identified as female, 77 as male, and 12 as gender diverse (e.g., transgender, non-binary). There were 71 participants in the PRECOVID_ONSET group, 83 in the DURINGCOVID_ONSET group and 22 in the POSTCOVID_ONSET group. In the “Yes Drug Use over the last 30 days” there were 21 participants and in the “No drug use over the last 30 days” group there were 94 (missing, n = 61). Y-OQ-SR total scores across intake, discharge, and 6-months follow-up can be seen in Table 1.

Table 1. Y-OQ-SR Total Scores Across Intake, discharge, and 6-months follow-up.

	Y-OQ-SR total
	n (%)	Intake	Discharge	6-months follow up
COVID-19 groups
PRECOVID_ONSET	71 (40%)	59.4	28.4	36.7
DURINGCOVID_ONSET	83 (47%)	69.6	30.3	43.2
POSTCOVID_ONSET	22 (13%)	63.1	41.8	44.0
Gender
Female	87 (49%)	67.2	29.0	40.4
Male	77 (44%)	59.1	31.4	39.6
Gender diverse	12 (7%)	81.8	42.0	49.4
Drug Use
Yes prior drug use	21 (12%)	91.3	50.2	58.5
No prior drug use	94 (53%)	64.1	30.0	37.9
All participants	176 (100%)	64.7	30.9	40.7

Y-OQ-SR Total Treatment Trajectories

Results from the fixed effects parameters for the Y-OQ-SR total show a significant negative slope (−57.01; 95% CI = [−66.26, −47.76]; p < .001), meaning that Y-OQ-SR total scores sharply decreased (e.g., improved) from intake. The quadratic slope term was also significant, (21.67; 95% CI = [17.38, 25.96]; p < .001) meaning that following the initial decrease, each subsequent timepoint showed a slight increase (e.g., worsening) in the slope, supporting a non-linear trend.

We then examined the fixed effects for the categorical covariates, including COVID-19 group, gender, and drug-use. We observed significant effects for drug use. Holding all other variables constant, compared to those who did not use drugs in the last 30 days, adolescents who reported drug or alcohol use had significantly higher (e.g., more maladaptive) Y-OQ-SR total scores (−22.57; 95% CI = [−34.71, −10.42]; p < .001). As a follow-up, we examined the estimated marginal means and pairwise contrasts between drug use and COVID-19 groups. The difference in Y-OQ-SR means was significant for the PRECOVID_ONSET group (32.99; SE = 10.86; p = .003) and the DURINGCOVID_ONSET group (28.46; SE = 8.63; p < .001), whereby youth with prior drug use had significantly higher Y-OQ-SR total scores compared to youth without prior drug use. In the POSTCOVID_ONSET group there was no difference in Y-OQ-SR total scores between the two drug use groups. The main effects of COVID-19 group and gender were not significant. Figure 1 shows the longitudinal treatment trajectories for all categorical variables (note that we have included the non-significant effects of COVID-19 group and gender for visualization purposes only).

Figure 1. Y-OQ-SR total treatment trajectories by COVID-19 group (non-significant effect), gender (non-significant effect) and drug/alcohol use (significant effect) across intake, discharge, and 6-months follow-up.

Following this, we examined the fixed effects of the continuous variables (e.g., FAD and age). The effect of FAD on treatment trajectories was significant (11.34; 95% CI = [3.81, 18.87]; p < .001), showing a positive linear trend between FAD scores and Y-OQ-SR total scores controlling for the effects of all other fixed parameters. A one unit increase in FAD scores corresponds to a 11.34 unit increase in Y-OQ-SR total scores. The effect of age on treatment trajectories was not significant.

Lastly, we examined the random components of the model, which describe variability associated with the intercept and the slope parameters. The variability around the intercept estimate was (sd = 21.76 95% CI = [15.68, 25.99]), which accounted for 47% of the total variance. This term represents the variability around the average Y-OQ-SR total score at the first timepoint (e.g., intake) holding all other covariates constant. The variability associated with the slope estimate was (sd = 13.09; 95% CI = [8.63, 16.93]), which accounted for 17% of the total variance. This term represents the variability around the average linear change in YOQ scores. The variability associated with residual terms was (sd = 19.18; 95% CI = [16.87, 21.86]), this accounted for 36% of the total variance. This term represents the leftover or unexplained variability after accounting for the random intercept and random slope estimates.

Discussion

The present study examined longitudinal treatment trajectories in a clinical sample of adolescents to determine the impact of the COVID-19 lockdown, as well as several other individual differences. In terms of the COVID-19 lockdown, we found no significant differences in treatment trajectories between the three onset groups, meaning that trajectories did not differ based on the timing of treatment in relation to lockdown mandates. Based on past research highlighting the disruptive nature of lockdowns on live-in treatment services (e.g., Palinkas et al., 2021), this result was somewhat unexpected. However, results may suggest that even in the face of unprecedented stress and hardship, live-in facilities were able to adjust and maintain similar treatment pathways. Several studies documented significant increases in the use of telehealth services from off-site clinicians. Youth who have grown up emersed in social media may be prime targets for this type of approach (Burns et al., 2016). Indeed, research stemming from before COVID-19 supports that clinicians are able build therapeutic alliances and rapport with youth and their families (Goldstein & Glueck, 2016). In the current study we do not know whether the sites included in our analyses made use of telehealth services, therefore, this should be considered speculative.

In terms of the other variables included in the model, we observed a significant effect of drug use, whereby adolescents who reported drug or alcohol use in the 30 days prior to starting treatment had significantly worse Y-OQ-SR treatment trajectories compared to those who did not use drugs or alcohol. Adolescents who use substances frequently report using multiple substances (Crummy et al., 2020), which is known to lead to poorer treatment outcomes (L. Wang et al., 2017). Moreover, adolescents who use substances also tend to suffer from higher levels of mood and disruptive disorders compared to adolescents who do not use (Kandel et al., 1999). Greenbaum et al. (1991) found that in a sample of adolescents with severe emotional disturbances, the odds of also having a substance use disorder was 2.37 times greater for those adolescents entering live-in treatment compared to other types of treatment. This is important, as differential combinations of substance use and psychiatric illness can require different treatment approaches (Connor et al., 2013; Mefodeva et al., 2022). In the present study, we did not differentiate sites by participant profiles or therapeutic approaches. It is possible that treatment programs dedicated primarily to mental health or relationship issues do not have the capacity to adequately deal with the complexity of comorbid substance use. In the future, it would be of interest to stratify programs more specifically (e.g., by participant profiles and therapeutic profiles) to gain a more nuanced understanding of the relationship between substance use, mental health and treatment success. Overall, results highlight the importance of treatment approaches that are able to address comorbidity among substance use and mental health.

It is important to note that the effect of drug and alcohol use was dependent on the timing of the COVID-19 lockdown. Specifically, adolescents who reported prior drug or alcohol use before treatment had higher Y-OQ-SR total scores compared to those without prior use, but only for those adolescents whose treatment started before lockdown mandates were implemented (e.g., in the PRECOVID_ONSET and DURINGCOVID_ONSET groups). For adolescents who entered treatment after lockdown had started, there was no difference in treatment outcomes between those with and without prior alcohol or drug use. A potential explanation for this could be related to overall frequency of use. For instance, the questionnaire used to assess substance use ranged from “not at all” to “daily.” Due to missing data, we dichotomized substance use into only two groups (yes prior use and no prior use). Had we been able to examine this variable with more precision, we may have found that while youth continued to use substances after lockdown mandates, the overall frequency of use was substantially lower in the POSTCOVID_ONSET group compared to the PRECOVID_ONSET and DURINGCOVID_ONSET groups. This would be in alignment with some previous research, finding that substance use decreased in the early phases of school closures and lockdowns (Layman et al., 2022). As a result, adolescents who started treatment after lockdown may have had limited access and opportunity for substance use thus removing any treatment differences between use and no use groups. Another possibility may stem from the impact of COVID-19 and contagion mitigation efforts on treatment services. For instance, while largely considered negative, there is some research suggesting a limited number of positive outcomes resulting from the nature of the new mandates. Pagano et al. (2021) conducted semi-structured interviews with directors of live-in substance abuse treatment programs during lockdown. Positive effects included the opportunity to improve aspects of operations during a slow-down in services and government recognition of substance abuse treatment facilities as essential. It is possible that other live-in treatment facilities experienced similar positive effects, which may have mitigated the impacts of prior drug use in the POSTCOVID_ONSET group.

The final factor that impacted treatment trajectories was family functioning, in that those with more adaptive family functioning at intake had better treatment trajectories. During adolescence peer relationships become very influential, however, relationships with parents are crucial for youth’s well-being. Several studies have found that youth with more supportive parents have fewer internalizing and externalizing problems (Luyckx et al., 2014; Lyell et al., 2020) as well as better self-esteem and social well-being (Demaray et al., 2005; Lyell et al., 2020). Similarly, poor parental supervision, parental conflict, and parent aggression are among some of the strongest predictors of later criminal and delinquent behaviors (Shader, 2003) as well as substance abuse (Rodríguez-Ruiz et al., 2023). Our results highlight that more adaptive family functioning at the beginning of treatment may act as a protective factor supporting treatment efforts and should continue to be a focus of the treatment journey for youth.

The present study is notable in that results are longitudinal, examining treatment trajectories from before and after lockdown mandates were implemented. It also adds to our understanding of COVID-19’s impact on youth, a group who has been thought to be especially vulnerable. The limitations of this study raise several interesting questions for further research. First, we collapsed findings across several different types of live-in treatment facilities with different patient profiles and treatment approaches. In the future it would be of interest to know whether some types of programs fared better (or worse) than others. Given the impact of prior drug use and maladaptive family functioning, programs that specifically address substance use and include parents in the treatment process may be more beneficial. Second, we examined the impact of the COVID-19 lockdown on treatment trajectories, however, we do not know how each program adapted to the lockdown and how the changes impacted their standard of care. Future research including this level of detail would allow us to more precisely understand why some programs may have struggled more or less than others. Lastly, we were unable to differentiate drug use by exact frequency and type of drug. This is an important area of future research as certain combinations of substances and psychiatric illness can require different treatment approaches (Mefodeva et al., 2022).

Conclusion

Although the impact of COVID-19 was devastating, results from the present study suggest that as a standalone factor, lockdown mandates did not influence treatment trajectories of adolescents in live-in treatment. This is promising given results from anecdotal reports highlighting the stressful nature of COVID-19 for frontline workers (e.g., Palinkas et al., 2022). Moving forward, time and resources should be focused on pandemic response plans in an effort to decrease the immense toll it takes on frontline staff, while remaining diligent to treatment efforts. Prior drug use and family functioning both emerged as important variables impacting treatment trajectories. Live-in treatment approaches should be multi-pronged to better meet the needs of adolescents with comorbid drug use and parents should be seen as active participants.

Disclosure statement

No potential conflict of interest was reported by the author(s).