Service Use as a Predictor of Change in Mental Health Problems Among Children: A Prospective Cohort Study
Abstract
Objective:
Psychosocial interventions for children’s mental health problems typically differ in several characteristics, such as therapist training, content, motivation for treatment, and extent of comorbid conditions among patients, depending on whether the interventions take place in clinical research studies or in real-life settings. Accordingly, the effects found in research studies may not be generalizable to typical service provision. The authors sought to examine the potential associations between receiving usual care and later psychiatric symptoms, impairment, and potential improvements in social skills.
Methods:
Participants (N=996) drawn from the 2003–2004 birth cohorts in Trondheim, Norway, included children who received usual care and those who did not receive any services (as a control group). The children were assessed with biennial clinical interviews from ages 4 to 14 years. Random intercept, cross-lagged panel models were combined with propensity scoring to adjust for measured time-varying and all unmeasured time-invariant confounders.
Results:
Usual care was not associated with alterations in social skills or impairment due to mental health problems. Similarly, usual care provided to 7- to 12-year-olds did not predict changes in the number of symptoms of psychiatric disorders. However, usual care received at ages 0–4 and 5–6 predicted a slight increase in the number of psychiatric symptoms 2 years later. No significant associations between usual care and improved outcomes were detected.
Conclusions:
These observational findings reveal the need to implement existing evidence-based approaches in usual care and to develop evidence-based approaches to the complex cases often seen in specialty and community care systems.
HIGHLIGHTS
•
The authors used random intercept, cross-lagged analysis and propensity scoring to examine the outcomes of usual care for mental health problems among children ages 4–14 years.
•
Receiving usual care was not associated with any significant improvement in social skills, mental health–related impairment, or psychiatric symptoms, compared with receipt of no care services.
•
The effect of usual care provided in community settings on mental health problems among children may be small or even absent, a possibility that should be investigated further.
Hundreds of controlled studies have examined the effects of therapeutic interventions for psychiatric disorders among children and adolescents and have found a moderate impact of these interventions on these disorders (1). Such research studies are typically conducted in resource-rich settings characterized by rigorous training of therapists, monitoring of fidelity to treatment protocol, and inclusion criteria limiting comorbid conditions. However, usual care takes place in a variety of settings, involves a variety of assessments, and is provided by professionals with varying educational backgrounds and degrees of training. In addition, presence of comorbid conditions is the norm, and the therapy may or may not be manualized and is typically not closely monitored (2). For policy makers, health care planners, service providers, parents, and patients alike, the question is whether such usual care reduces the prevalence of mental health problems and associated impairments. To answer this question, one cannot rely on generalizing from controlled studies of specific therapies for a narrow set of disorders. Ethical and legislative regulations prohibit assigning help-seeking children to a no-treatment or a waitlist condition. Instead, studies with observational designs that limit the impact of confounding must be performed. To address the paucity of such studies, the aim of this study was to use novel analytical techniques in order to investigate whether usual care for mental health problems is significantly associated with beneficial mental health outcomes among young children.
Few longitudinal studies have examined how children receiving usual care for mental health problems fare compared with children who do not receive such care. We identified three prospective cohort studies with community samples that fit this description. The major obstacle to drawing causal conclusions from observational studies is that such studies typically do not control for factors influencing both service use and mental health. In the first study, Zwaanswijk and colleagues (3) found that service use or any control variables (i.e., gender, age, parents’ education, and contact with general practitioners) were not significantly associated with any changes in mental health problems. Applying a covariate approach and adjusting analyses for initial symptom level and previous symptom change, Angold and colleagues (4) reported that usual care in specialty child and adolescent mental health services (CAMHS) had a positive effect on psychiatric symptoms. Because a range of factors beyond severity of symptoms lead to initiation of service use (5, 6), these findings may be attributable to unmeasured confounders, as the authors acknowledge. Jörg et al. (7) applied propensity score matching in which a quasi-experimental design was constructed by creating a control group as similar as possible to the treatment group in terms of the measured characteristics. The authors found that the mental health of children who used services improved less than the mental health of those who did not use services. Importantly, although propensity score matching may be advantageous to a traditional covariate approach, the results of such matching approach still are based on measured confounders and unmeasured confounders cannot be ruled out.
Within-person approaches (8) can be used to adjust for one type of unmeasured confounding, that is, factors that do not vary over time (e.g., stable impacts of genes or personality or stable attitudes toward the health care system). However, results from within-person approaches can still be influenced by time-varying confounders. We combined within-person analysis with propensity scoring to adjust for measured time-varying confounders and all unmeasured time-invariant confounders.
Beyond a reduction in symptoms, a common aim of service provision is to alleviate impairment due to mental health problems. Moreover, for children and adolescents to be functional—in terms of interactions with peers and family and participation in school—social skills are beneficial, and improved social skills can also be considered a desired outcome of interventions.
Following up with a community sample of children ages 4–14 years biennially, we examined whether receiving usual care in early and middle childhood predicts changes in symptoms of psychiatric disorders, associated impairments, and social skills, after adjustment for all unmeasured time-invariant and measured time-varying confounders. We hypothesized that receiving usual care would be associated with greater improvements in these conditions, compared with not receiving usual care.
Methods
Study Design
This study utilized data from the Trondheim Early Secure Study (9), a prospective longitudinal cohort study focusing on children’s psychosocial development, following up two birth cohorts with biennial assessments. Data were collected in 2-year increments at six time points (T) (i.e., T1, 2007–2009; T2, 2009–2011; T3, 2011–2013; T4, 2013–2015; T5, 2015–2017, and T6, 2017–2019).
Participants
Children and parents from the 2003 and 2004 birth cohorts in the city of Trondheim, Norway, were invited to participate in the study during routine health checkups for 4-year-olds at local well-child clinics. Of the 3,013 families with sufficient Norwegian language skills (N=176 were excluded because of a lack of language skills), 2,475 consented to participate. Participants were stratified into four groups on the basis of scores on the Strengths and Difficulties Questionnaire (SDQ), version 4–16 years, completed by parents (cutoff scores: 0–4, 5–8, 9–11, and 12–14, with higher scores indicating more severe mental health problems) (10). Children with higher SDQ scores were oversampled to increase the number of participants in the sample with mental health problems (drawing probabilities, indicating likelihood of being drawn for further participation, by stratum: 0.37, 0.48, 0.70, and 0.89, respectively), and 1,250 were invited to participate in the first examination (T1) at 4 years of age. We obtained diagnostic and service use information from 80% (N=996 of 1,250) of consenting participants (mean±SD age=4.5±0.3 years) at T1. Follow-ups were performed after 2 years (T2: mean age=6.7±0.2, N=752), 4 years (T3: mean age=8.8±0.2, N=670), 6 years (T4: mean age=10.5±0.2, N=678), 8 years (T5: mean age=12.5±0.7, N=638), and 10 years (T6: mean age=14.4±0.2, N=628). Lower social skills at T1 and T4 predicted attrition at T2 and T5, respectively (T2: OR=1.02, 95% CI=1.00–1.03; T5: OR=1.05, 95% CI=1.02–1.07), and lower socioeconomic status at T4 predicted attrition at T5 (OR=1.38, 95% CI=1.02–1.88). A higher propensity for service use at T2 predicted reduced attrition at T3 (OR=0.54, 95% CI=0.31–0.96). According to Cox-Snell R2 values, the predictors at T1 accounted for 2% of the attrition between T1 and T2, and T3 and T4 predictors accounted for 2% and 4%, respectively, indicating that it was unlikely that selective attrition had influenced results. The Regional Committee for Medical and Health Research Ethics, Mid-Norway, approved the study.
Norwegian Health Care System
In Norway, health services for those <18 years are provided for free. Services are divided into CAMHS and community services (e.g., social services, well-child clinics, general practitioners, child protection services, community psychologists, and school health nurses). Information about the number of children utilizing these various community services was not available. Data for CAMHS reveal that in 2018, the health region that covers the city of Trondheim served 8,330 children (5.4% of the population) ages 0–18, of whom 52.2% were boys and 44.8% were <12 years old (11).
Measures
Service use.
Parents were interviewed at T1–T6 with the Child and Adolescent Service Assessment (CASA) (12). The CASA identifies help provided for psychiatric problems. Service use is defined as efforts to identify, diagnose, or manage behavioral, emotional, or substance-related problems. At T1, the interview captured service use from birth until age 4, and at T2–T6, service use after the previous visit. A binary variable reflecting service use in the previous 2 years (4 years at T1) was created at each time point. A child could have received more than one service within one period.
Symptoms of psychiatric disorders.
The Preschool Age Psychiatric Assessment (PAPA) (13) was completed by the parents at T1–T2. The PAPA is a semistructured, interviewer-based tool designed to obtain information about psychiatric disorders in young children defined according to the DSM-IV (14). Information about the frequency, intensity, duration, and onset of each symptom is recorded. At T3–T6, the Child and Adolescent Psychiatric Assessment (CAPA) (15) was administered separately to children and parents. A symptom was considered present if it was reported by the child or the parent. For attention-deficit hyperactivity disorder (ADHD), only parent-reported symptoms were recorded (15). Symptoms of behavioral disorders (i.e., oppositional defiant disorder, conduct disorder, and ADHD) and emotional disorders (i.e., major depression, dysthymia, depression not otherwise specified, separation anxiety disorder, generalized anxiety disorder, social phobia, specific phobia, agoraphobia, selective mutism, and obsessive-compulsive disorder) were summed at each time point. The PAPA and CAPA apply a 3-month primary period. The developers of the PAPA and the CAPA trained the interviewers. Raters who were blind to conditions recoded 9% of the PAPA (intraclass coefficient [ICC]=0.96) and 15% of the CAPA interviews (T3 and T4, ICC=0.94).
Impairment.
The PAPA and CAPA also record the presence of impairment in the past 3 months related to the reported symptoms in 19 different areas of functioning (e.g., relationships with parents, school function, and play) on the basis of the World Health Organization’s International Classification of Functioning, Disability and Health (16). The number of areas with impaired functioning was summed.
Social skills.
Teacher-reported social skills were obtained by using the Social Skills Rating Systems (SSRS-T) preschool (T1) and school (T2–T6) versions (17). The SSRS-T includes a set of statements covering three domains—cooperation, assertion, and self-control—and responses are recorded by using the following scale: 1, never; 2, sometimes; 3, often; and 4, very often. Each domain includes 10 statements, allowing a maximum score of 120, with higher scores indicating higher levels of social skills. Scores from each domain were summed (Cronbach’s α=0.93–0.94).
Confounders.
Variables entered into the propensity score for service use were based on previous work regarding predictors of referral (5, 18–20). These included a child having a DSM-IV–defined emotional or behavioral disorder. In addition, all diagnoses were summed to create a continuous variable reflecting the extent of comorbid conditions. Parents’ perceived need for help was recorded with the PAPA and CAPA and was coded 0, no, or 1, yes. A binary variable for previous service use was created by using information from the CASA. The summed score on the general functioning subscale of the McMaster Family Assessment Device (Cronbach’s α=0.87–0.90) was included as a measure of family functioning (21). Occupational status, as reported by the parent and coded according to the International Labour Organization’s scheme for classifying occupations (22), was used as a measure of socioeconomic status. Parental mental health was measured with summed scores of the Beck Depression Inventory (23), Beck Anxiety Inventory (24) (T1–T2), and Hopkins Symptom Checklist–25 (25) (T3–T6). Finally, living with a biological parent was included as a variable.
Statistical Analysis
Propensity score modeling.
Propensity score modeling was used to control for observed initial differences between children who received treatment and those who did not. We used binary logistic regression to determine the propensity for service use at each time point. The log of the odds of the probability of exposure was used as the propensity score (26). To avoid including variables whose variation could be affected by usual care (for which we were determining the propensity), the score for T1 included only living with biological parents and socioeconomic status.
Random intercept, cross-lagged panel model (RI-CLPM).
Propensity scores were entered into an RI-CLPM by using Mplus, version 8.1 (27), applying a maximum likelihood estimator with robust standard errors. Missing data were handled in accordance with a full maximum-likelihood procedure. Because of oversampling, population weights were applied to obtain correct population estimates. The RI-CLPM is a structural equation model in which each observed score is decomposed into a within-person part and a between-person part by including random intercepts in a CLPM, removing stable between-person components, and effectively controlling for all time-invariant confounders (8). In turn, the within-person, time-varying components are used to estimate cross-lagged and autoregressive paths. (A simplified model is provided in the online supplement to this article.) To avoid negative degrees of freedom, our outcomes (i.e., symptoms, impairment, and social skills) were estimated in separate models. Service use, propensity score, and outcome were included at all six time points.
Results
A similar number of boys and girls participated in the study (Table 1). The prevalence of service use for mental health problems increased substantially, from 4% (N=44 of 996) at T1 to 18% (N=137 of 752) at T2, and remained at that level throughout childhood and early adolescence (T3: 17%, N=111 of 670; T4: 19%, N=127 of 678; T5: 17%, N=106 of 638; and T6: 18%, N=113 of 628). Characteristics of service use are provided in Table 2. Most children did not receive any services at each time point (T1: 96%, N=952; T2: 82%, N=615 of 752; T3: 83%, N=559 of 670; T4: 81%, N=551 of 678; T5: 83%, N=532 of 638; and T6: 82%, N=515 of 628). Among children <8 years, health nurses were the most commonly reported helpers. Older children were most frequently seen by a psychologist. Although the number of treatment sequences remained stable across all time points, the mean number of treatment sessions varied greatly. Few children had received psychotropic medications, and most of the psychotropic medication treatments were for ADHD (stimulants and nonstimulants). (A table with descriptive statistics of the study variables is available in the online supplement .) As expected, children who received services had higher log odds of receiving services at all time points (see the online supplement ).
| Characteristic | N | % |
|---|---|---|
| Gender | ||
| Boy | 504 | 51 |
| Girl | 492 | 49 |
| N of siblings | ||
| 0 | 128 | 13 |
| 1 | 535 | 54 |
| 2 | 235 | 24 |
| 3 | 65 | 7 |
| 4 | 24 | 2 |
| ≥5 | 9 | 1 |
| Biological parents’ marital status | ||
| Married | 538 | 55 |
| Cohabitating ≥6 months | 317 | 32 |
| Separated | 20 | 2 |
| Divorced | 77 | 8 |
| Widowed | 3 | <1 |
| Cohabitating <6 months | 13 | 1 |
| Never lived together | 16 | 2 |
| National origin of biological mother | ||
| Norwegian | 911 | 92 |
| Western countries | 32 | 3 |
| Non-Western countries | 42 | 4 |
| National origin of biological father | ||
| Norwegian | 890 | 90 |
| Western countries | 61 | 6 |
| Non-Western countries | 32 | 3 |
| Parent’s highest completed education | ||
| Junior high school not completed | 0 | — |
| Junior high school (10th grade) | 3 | <1 |
| Some education after junior high school | 28 | 3 |
| Senior high school (13th grade) | 128 | 13 |
| Some education after senior high school | 34 | 3 |
| Some college or university education | 71 | 7 |
| Bachelor’s degree | 33 | 3 |
| College degree (3–4 years of study) | 329 | 33 |
| Master’s degree or similar | 278 | 28 |
| Doctoral degree, ongoing | 23 | 2 |
| Doctoral degree, completed | 67 | 7 |
| Gross annual household income (US$) | ||
| 0–21,000 | 33 | 3 |
| 22,000–50,000 | 183 | 18 |
| 51,000–87,000 | 514 | 52 |
| >$87,000 | 266 | 27 |
a
Totals for some subcategories may not sum to the overall total because of missing values.
| T1 (N=44)a | T2 (N=137) | T3 (N=111) | T4 (N=127) | T5 (N=106) | T6 (N=113) | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Characteristic | N | % | N | % | N | % | N | % | N | % | N | % | |
| Treatment sequences (M±SD)b | NA | — | 1.3±.7 | 1.3±.6 | 1.3±.6 | 1.4±.8 | 1.3±.5 | ||||||
| Treatment sessions (M±SD)c | NA | — | 12.8±28.8 | 28.4±85.9 | 14.8±44.2 | 38.1±137.1 | 11.5±21.0 | ||||||
| Reason for service used | NA | — | |||||||||||
| Anxiety | NA | — | 14 | 10 | 19 | 17 | 31 | 24 | 36 | 34 | 38 | 34 | |
| Sadness, depression, grief | NA | — | 5 | 4 | 12 | 11 | 23 | 18 | 29 | 27 | 34 | 30 | |
| Restlessness, concentration difficulties | NA | — | 31 | 23 | 48 | 43 | 53 | 42 | 35 | 33 | 28 | 25 | |
| Aggression, defiance, disobedience | NA | — | 34 | 25 | 20 | 18 | 19 | 15 | 10 | 9 | 10 | 9 | |
| Rituals, compulsions | NA | — | 3 | 2 | 3 | 3 | 3 | 2 | 7 | 7 | 6 | 5 | |
| Social interaction difficulties | NA | — | 38 | 28 | 17 | 15 | 22 | 17 | 13 | 12 | 11 | 10 | |
| Other | NA | — | 78 | 57 | 52 | 47 | 51 | 40 | 42 | 40 | 48 | 42 | |
| Where services were received | |||||||||||||
| Psychiatric outpatient clinic | 5 | 11 | 15 | 11 | 13 | 12 | 25 | 20 | 30 | 28 | 37 | 33 | |
| Public health center, well-child clinic | 21 | 48 | 30 | 22 | 32 | 29 | 27 | 21 | 23 | 22 | 28 | 25 | |
| Private practice | 1 | 2 | 17 | 12 | 11 | 10 | 3 | 2 | 9 | 8 | 10 | 9 | |
| Social office | 0 | — | 1 | 1 | 1 | 1 | 2 | 2 | 1 | 1 | 0 | — | |
| Child protection services | 3 | 7 | 1 | 1 | 2 | 2 | 3 | 2 | 3 | 3 | 4 | 4 | |
| Child’s home | NA | — | 16 | 12 | 4 | 4 | 5 | 4 | 2 | 2 | 4 | 4 | |
| Educational and psychological counseling service | 10 | 2 | 15 | 11 | 14 | 13 | 23 | 18 | 12 | 11 | 15 | 13 | |
| Support center | NA | — | 1 | 1 | 5 | 5 | 2 | 2 | 4 | 4 | 1 | 1 | |
| Other (e.g., general practitioner’s office, school, or day care) | 28 | 64 | 80 | 58 | 58 | 52 | 66 | 52 | 49 | 46 | 40 | 35 | |
| Service provider’s profession | |||||||||||||
| Health nurse | NA | — | 37 | 27 | 33 | 30 | 41 | 32 | 37 | 35 | 37 | 33 | |
| Family physician | NA | — | 34 | 25 | 23 | 21 | 26 | 20 | 21 | 20 | 30 | 27 | |
| Social worker | NA | — | 5 | 4 | 1 | 1 | 7 | 6 | 7 | 7 | 3 | 3 | |
| Special teacher | NA | — | 29 | 21 | 27 | 24 | 29 | 23 | 14 | 13 | 11 | 10 | |
| Psychologist | NA | — | 28 | 20 | 34 | 31 | 49 | 39 | 49 | 46 | 65 | 58 | |
| Psychiatrist | NA | — | 2 | 1 | 0 | — | 6 | 5 | 6 | 6 | 3 | 3 | |
| Alternative health practitioner | NA | — | 1 | 1 | 2 | 2 | 1 | 1 | 2 | 2 | 1 | 1 | |
| Other | NA | — | 47 | 34 | 40 | 36 | 19 | 15 | 22 | 21 | 21 | 19 | |
| Psychotropic medication treatment | |||||||||||||
| Anxiolytics | 1 | 2 | 0 | — | 2 | 2 | 1 | 1 | 1 | 1 | 0 | — | |
| Antidepressants | 0 | — | 0 | — | 0 | — | 0 | — | 0 | — | 1 | 1 | |
| Stimulants and nonstimulants for ADHD | 0 | — | 2 | 1 | 8 | 7 | 24 | 19 | 20 | 19 | 23 | 20 | |
a
A condensed version of the Child and Adolescent Service Assessment was used at T1, covering fewer options for where services were received and profession of the service provider, reason for service use, and amount of service use. T1, first assessment; T2, 2-year follow-up; T3, 4-year follow-up; T4, 6-year follow-up; T5, 8-year follow-up; T6, 10-year follow-up. NA, not available (missing data).
b
A treatment sequence is defined as a running service use series that is not formally concluded. T2, N=131; T3, N=111; T4, N=127; T5, N=106; T6, N=113.
c
T2, N=132; T3, N=100; T4, N=128; T5, N=107; T6, N=116.
d
Children could have multiple reasons for service use.
Table 3 shows the results of the RI-CLPM with added propensity scores, with controls for all time-invariant and measured time-varying confounders. Usual care received in the past 2 years did not predict changes in impairment due to mental health problems or social skills. Similarly, we did not find changes in psychiatric symptoms associated with previous service use among 7- to 14-year-olds. Receiving usual care in early childhood (ages 0–4 years) and in late preschool (ages 5–6) predicted an increase in the number of psychiatric symptoms 2 years later.
| Ages 0–4 | Ages 5–6 | Ages 7–8 | Ages 9–10 | Ages 11–12 | ||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Variable | β | 95% CI | β | 95% CI | β | 95% CI | β | 95% CI | β | 95% CI |
| Symptoms | 2.65* | .32, 4.98 | 2.00** | .83, 3.16 | .86 | −.42, 2.13 | .91 | −.59, 2.42 | 1.45 | −.64, 3.55 |
| Impairment | .46 | −.32, 1.25 | .30 | −.31, .92 | −.30 | −1.10, .50 | .22 | −.67, 1.11 | .52 | −.68, 1.72 |
| Social skills | −.81 | −5.43, 3.81 | −1.83 | −4.66, 1.00 | .87 | −2.27, 4.02 | −2.29 | −5.36, .78 | −1.83 | −5.31, 1.65 |
a
Random intercept, cross-lagged panel model fit estimates of symptoms (comparative fit index [CFI]=0.963, Tucker-Lewis Index (TLI)=0.933, and root mean square error of approximation (RMSEA)=0.041 [95% CI=0.035–0.048], χ2=224.51, df=84, p<0.001), impairment (CFI=0.964, TLI=0.934, and RMSEA=0.039 [95% CI=0.032–0.046], χ2=211.40, df=84, p<0.001), and social skills (CFI=0.963, TLI=0.932, and RMSEA=0.033 [95% CI=0.026–0.040], χ2=176.90, df=84, p<0.001). Concurrent measure of the outcome and propensity for service use were included as covariates.
*p<0.05, **p<0.005.
Discussion
We investigated whether receiving usual care for mental health problems at ages 0–12 years predicted changes in the number of symptoms, impairments, and social skills 2 years following care receipt, after adjustment for the propensity of using services and all unmeasured time-invariant confounders. Receiving usual care was not associated with a decreasing number of psychiatric symptoms, reduced impairment, or better social skills among the children included in this study. In contrast, receiving usual care for mental health issues at ages 0–6 years predicted a slight increase in psychiatric symptoms 2 years later, compared with children who did not receive any services.
We note that heterogeneity may have been present in the results with respect to the content and duration of services, where services were provided, the type of service provider (e.g., psychiatrist, psychologist, or nurse), or nature of the child’s mental health problems, among other factors. If some children benefited from services, the observed lack of an effect of services on children’s mental health may be explained by any positive outcomes having been offset by negative effects of the services. It is also possible that receiving usual care did not alter vital outcomes for most children or that any changes that did occur were too modest to be detected in this study. Our results contradict the findings of Jörg and colleagues (7), who concluded that service provision during preadolescence predicted an increase in symptoms, and those of Angold and colleagues (4), who found that service provision reduced symptoms. Neither of these studies adjusted their analyses for time-invariant confounders, and the study by Angold and colleagues adjusted analyses only for the initial values of the outcome. Although we adjusted for confounders in our analyses, in the absence of randomization, unmeasured time-varying confounders (e.g., time-varying parenting practices) could still have influenced the results.
The CASA definition of service provision included not only treatment but also screening for and diagnosis of mental health. Furthermore, significant heterogeneity existed in the types of services. Distinguishing between different types of services and content would have enabled a more precise analysis. Such more detailed analysis was not possible because a further division of our sample into subgroups would have reduced statistical power. However, cases where individuals received only an initial assessment should not cancel any real effects among individuals who received treatment. Similarly, it is possible that the 2-year interval between measurements could have obscured temporary improvements that had occurred before the 3-month period assessed at each time point.
Findings from some studies suggest that evidence-based psychotherapies have smaller effects when implemented as part of usual care, compared with the effects seen in efficacy studies of these interventions in highly controlled settings. Ng and Weisz (28) ascribed these observations to potential mismatches between evidence-based psychotherapies and usual care. These mismatches may arise because patients in usual care are a more heterogeneous group with more comorbid conditions; moreover, their conditions follow a less predictable course, resulting in shifting treatment needs during therapy. Although many evidence-based approaches provide flexibility, they usually lack approaches to manage comorbid conditions, crises, and other treatment challenges. This lack could explain why evidence-based approaches are frequently not implemented in usual care and may not be as well suited for many children seen in this setting. Furthermore, logistic barriers, such as insufficient time and training of providers, could be important obstacles to implementing new treatments (29).
We found that the number of psychiatric symptoms increased more among preschoolers who received services than among those who did not. Before it may be concluded that service provision to this age group does more harm than good, several alternative explanations should be considered. First, we could not carry out a comprehensive propensity adjustment for children receiving treatment at ages 0–4. As indicated by previous research, factors included in the propensity score at later ages predict increased referrals to treatment during early childhood (5) and worsening of symptoms (30). Not including these factors in the propensity score for T1 could have produced bias suggesting a detrimental effect of service provision. However, we adjusted for the propensity to use services between age 5 and 6 and still found that service use predicted increased symptoms. Second, some time-varying factors might be responsible for worsening symptoms and the initiation of services that were not included in our propensity score. Third, behavioral problems are more strongly associated with referrals than are emotional problems (5, 18, 20). Addressing such overt behavioral symptoms may leave co-occurring emotional problems untreated, and these emotional problems may persist or worsen over time. Fourth, children receiving services at ages 0–4 could be in the initial phase of more chronic conditions, implying that the worsening of symptoms would have been more severe without intervention. Finally, receiving services may moderate the negative effects of adverse life events (e.g., parental divorce or bullying) and thus potentially reduce the negative impact of such events among vulnerable children. In any event, the potential negative effects of service use observed in this study require validation in future studies.
Importantly, these results may not generalize to health care systems outside of Norway. However, Norway consistently ranks as one of the wealthiest countries in the world per capita, with a well-developed public health care system (31). We have no reason to believe that Norwegian services provide particularly poor care compared with services in other countries.
Conclusions
In the context of the uncertainty of results obtained from observational research, the effect of usual care on young children may be small or even absent, a possibility that should be investigated in a variety of countries. There is a need to develop and implement evidence-based approaches to treat children with mental health problems in routine clinical practice.
Supplementary Material
File (appi.ps.20220079.ds001.docx)
- Download
- 138.33 KB
References
1.
Weisz JR, Kuppens S, Ng MY, et al: What five decades of research tells us about the effects of youth psychological therapy: a multilevel meta-analysis and implications for science and practice. Am Psychol 2017; 72:79–117
2.
Polanczyk GV, Salum GA, Sugaya LS, et al: Annual research review: a meta-analysis of the worldwide prevalence of mental disorders in children and adolescents. J Child Psychol Psychiatry 2015; 56:345–365
3.
Zwaanswijk M, Verhaak PFM, Van Der Ende J, et al: Change in children’s emotional and behavioural problems over a one-year period: associations with parental problem recognition and service use. Eur Child Adolesc Psychiatry 2006; 15:127–131
4.
Angold A, Costello EJ, Burns BJ, et al: Effectiveness of nonresidential specialty mental health services for children and adolescents in the “real world.” J Am Acad Child Adolesc Psychiatry 2000; 39:154–160
5.
Wichstrøm L, Belsky J, Jozefiak T, et al: Predicting service use for mental health problems among young children. Pediatrics 2014; 133:1054–1060
6.
Angold A, Messer SC, Stangl D, et al: Perceived parental burden and service use for child and adolescent psychiatric disorders. Am J Public Health 1998; 88:75–80
7.
Jörg F, Ormel J, Reijneveld SA, et al: Puzzling findings in studying the outcome of “real world” adolescent mental health services: the TRAILS study. PLoS One 2012; 7:e44704
8.
Hamaker EL, Kuiper RM, Grasman RPPP: A critique of the cross-lagged panel model. Psychol Methods 2015; 20:102–116
9.
Steinsbekk S, Wichstrøm L: Cohort profile: the Trondheim Early Secure Study (TESS)—a study of mental health, psychosocial development and health behaviour from preschool to adolescence. Int J Epidemiol 2018; 47:1401–1401i
10.
Goodman R, Ford T, Simmons H, et al: Using the Strengths and Difficulties Questionnaire (SDQ) to screen for child psychiatric disorders in a community sample. Br J Psychiatry 2000; 177:534–539
11.
Indergård PJ, Fuglset AS, Krogh F, et al: Activity Data for Specialized Child and Adolescent Mental Health Services [in Norwegian]. Trondheim, Norway, 2019. www.helsedirektoratet.no. Accessed Oct 11, 2021
12.
Ascher BH, Farmer EMZ, Burns BJ, et al: The Child and Adolescent Services Assessment (CASA): description and psychometrics. J Emot Behav Disord 1996; 4:12–20
13.
Egger HL, Erkanli A, Keeler G, et al: Test-retest reliability of the Preschool Age Psychiatric Assessment (PAPA). J Am Acad Child Adolesc Psychiatry 2006; 45:538–549
14.
Diagnostic and Statistical Manual of Mental Disorders, 4th ed. Washington, DC, American Psychiatric Association, 1994
15.
Angold A, Costello EJ: The Child and Adolescent Psychiatric Assessment (CAPA). J Am Acad Child Adolesc Psychiatry 2000; 39:39–48
16.
International Classification of Functioning, Disability and Health. Geneva, World Health Organization, 2001
17.
Gresham FM, Elliot SN: Social Skills Rating System: Manual. Circle Pines, MN, American Guidance Services, 1990
18.
Costello EJ, He JP, Sampson NA, et al: Services for adolescents with psychiatric disorders: 12-month data from the National Comorbidity Survey–Adolescent. Psychiatr Serv 2014; 65:359–366
19.
Lavigne JV, Arend R, Rosenbaum D, et al: Mental health service use among young children receiving pediatric primary care. J Am Acad Child Adolesc Psychiatry 1998; 37:1175–1183
20.
Wu P, Hoven CW, Bird HR, et al: Depressive and disruptive disorders and mental health service utilization in children and adolescents. J Am Acad Child Adolesc Psychiatry 1999; 38:1081–1090
21.
Byles J, Byrne C, Boyle MH, et al: Ontario Child Health Study: reliability and validity of the general functioning subscale of the McMaster Family Assessment Device. Fam Process 1988; 27:97–104
22.
International Standard Classification of Occupations: ISCO-88. Geneva, International Labour Office, 1990
23.
Beck AT, Steer RA, Ball R, et al: Comparison of Beck Depression Inventories -IA and -II in psychiatric outpatients. J Pers Assess 1996; 67:588–597
24.
Beck AT, Epstein N, Brown G, et al: An inventory for measuring clinical anxiety: psychometric properties. J Consult Clin Psychol 1988; 56:893–897
25.
Mollica RF, Wyshak G, de Marneffe D, et al: Indochinese versions of the Hopkins Symptom Checklist–25: a screening instrument for the psychiatric care of refugees. Am J Psychiatry 1987; 144:497–500
26.
Lunt M: Selecting an appropriate caliper can be essential for achieving good balance with propensity score matching. Am J Epidemiol 2014; 179:226–235
27.
Muthén LK, Muthén BO: Mplus User’s Guide, 8th ed. Los Angeles, Muthén and Muthén, 2017
28.
Ng MY, Weisz JR: Personalizing evidence-based psychotherapy for children and adolescents in clinical care; in Evidence-Based Psychotherapies for Children and Adolescents. Edited by Weisz JR, Kazdin AE. New York, Guilford Press, 2017
29.
Stewart RE, Stirman SW, Chambless DL: A qualitative investigation of practicing psychologists’ attitudes toward research-informed practice: implications for dissemination strategies. Prof Psychol Res Pr 2012; 43:100–109
30.
Wichstrøm L, Berg-Nielsen TS, Angold A, et al: Prevalence of psychiatric disorders in preschoolers. J Child Psychol Psychiatry 2012; 53:695–705
31.
Schneider EC, Sarnak DO, Squires D, et al: Mirror, Mirror 2017: International Comparison Reflects Flaws and Opportunities for Better US Health Care. New York, The Commonwealth Fund, 2017. https://www.commonwealthfund.org/publications/fund-reports/2017/jul/mirror-mirror-2017-international-comparison-reflects-flaws-and. Accessed Feb 18, 2020
Information & Authors
Information
Published In
History
Received: 13 February 2022
Revision received: 26 March 2023
Accepted: 5 April 2023
Published online: 31 May 2023
Published in print: December 01, 2023
Keywords
Authors
Competing Interests
The authors report no financial relationships with commercial interests.
Funding Information
This research was funded by the Research Council of Norway (grant ES611813) and a grant from the liaison committee between the Central Norway Regional Health Authority and the Norwegian University of Science and Technology (to Dr. Wichstrøm).
Metrics & Citations
Metrics
Citations
Export Citations
If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download.
For more information or tips please see 'Downloading to a citation manager' in the Help menu.
View Options
View options
PDF/EPUB
View PDF/EPUBLogin options
Already a subscriber? Access your subscription through your login credentials or your institution for full access to this article.
Personal login Institutional Login Open Athens loginNot a subscriber?
PsychiatryOnline subscription options offer access to the DSM-5-TR® library, books, journals, CME, and patient resources. This all-in-one virtual library provides psychiatrists and mental health professionals with key resources for diagnosis, treatment, research, and professional development.
Need more help? PsychiatryOnline Customer Service may be reached by emailing [email protected] or by calling 800-368-5777 (in the U.S.) or 703-907-7322 (outside the U.S.).