Clinical effectiveness of drop-in mental health services in paediatric healthcare settings: a non-randomised multi-site study for children, young people and their families | BMC Health Services Research

Participant flow

Figure 1 illustrates the flow of participants through the study.

One hundred and twenty families consented to take part in the study across six paediatric healthcare settings. Sixty-six families (55%) were referred by their clinician, and 54 (45%) self-referred to the service. These proportions varied across the different sites.

Participant characteristics are presented in Table 1. The age of CYP in the study ranged from two to 24 years with a mean age of 13.25. Parent-reported measures were completed for CYP age 5 or under, or if requested by the family. The majority of CYP were White British and used English as their first language. Nearly 1/3 of CYP described themselves as having a disability. CYP’s LTC was categorised using the ICD11 coding tool [30].

Parents’ average age was 42.94 and the majority were female. Most parents were also White British with English as first language. Over half of the parents in the study were employed with the majority working part time.

Intervention characteristics

The majority of CYP in the study attended an assessment appointment during which information about the individual’s primary presenting problem was collected. 58% of participants were offered 6–8 low intensity CBT sessions, and 65% of this low intensity treatment was for adolescent anxiety. The average wait for an initial assessment was 13.42 days (range 4–34 days).

Only one site offered assessments and intervention for parents (UCLH). Eight parents sought help for their own mental health. Three were referred to existing services and five signposted to external organisations.

Missing data

There were consistent significant differences in 6 month follow up completion between families who attended an assessment and the intervention that was received. Since the data were not missing at random, CCA using data only from participants who completed follow up measures, was conducted [28]. Reasons for non-completion are displayed in Fig. 1.

Main analysis

Primary outcome

Emotional and behavioural problems demonstrated a statistically significant decrease from an estimated mean score of 17.68 (6.08) pre-intervention to 15.67 (7.29) at 6 months, post-baseline, a mean decrease of 2.01, 95%CI (0.87, 3.18), t(81) = 3.50, p < 0.001, d = 0.39 (Table 2). Mean total scores at baseline (17.68) were above the clinical threshold whereas 6 moth post-baseline scores (15.67) were below clinical threshold.

Secondary outcomes

There was also a statistically significant decrease on specific SDQ subscales including emotional symptoms and hyperactivity. All results from the secondary outcomes are displayed in Table 2.

Total quality of life scores on the PedsQL significantly increased from a mean score of 52.44 (18.48) at baseline to 59.15 (21.93), 95%CI (− 10.09, − 3.33), t(81) = − 3.95, p < 0.001, d = 0.44 (Table 2). In addition, when looking at parent and CYP reported outcomes separately (in Supplementary Material 3) all subscales on the PedsQL showed significant increases from baseline to 6 month follow up.

The non-parametric Wilcoxon signed rank test was used to determine whether there was a significant change in depression scores. For CYP over the age of 12 (n = 46), PHQ- 9 depression scores decreased from 13.28 (6.12) to 9.98 (6.48), Z=− 3.96, p < 0.001, d = 0.65 from baseline to 6 month follow up. Scores moved from the “moderate” clinical range to “mild”.

There was also a significant decrease in parental depression (n = 64) from 8.17 (6.35) to 6.36 (6.00), Z= − 2.42, p = 0.02, d = 0.30 from baseline to 6 month follow up. Scores remained within the “mild” range.

There was a significant decrease in anxiety symptoms for CYP with child-reported anxiety scores (n = 46) decreasing from 10.91 (5.40) to 7.70 (5.98), 95% CI (1.73, 4.70), t(45) = 4.37, p < 0.001, d = 0.64. As scores above 10 are considered to be within the clinical range, average follow-up anxiety scores for CYP have been found to no longer meet clinical threshold.

Parental self-reported anxiety scores (n = 64) decreased from 6.89 (5.71) to 5.16 (4.72), 95% CI (0.44, 3.03), t(63) = 2.68, p = 0.01, d = 0.34.

As the results indicated a significant difference between baseline and 6-month post consent outcomes for those who accessed the drop-in services, post-hoc t-tests for parametric data were conducted to investigate if this change was being driven by the specific intervention that CYP received.

Participants receiving low intensity CBT showed a significant difference in SDQ scores from a mean score of 15.64 (5.60) at baseline to 13.75 (6.56) at 6 month post-consent, a mean difference of 1.89, 95% CI (0.33, 3.46), t(47) = 2.26, p = 0.02, d = 0.35 (see Table 3). For those referred to existing services, there was also a significant improvement when comparing mean scores.

CSQ- 8

Seventy-seven families completed the modified CSQ- 8. 30 reports were from the same family, with 18 CYP-report only and 29 parent-report only. Responses and proportions split by CYP and parent-report are shown in Supplementary Material 4. Both CYP and parents indicated they would “Totally” (median = 5) recommend the project to a friend. The only question with a lower median score was “whether the information and support received made any difference to you/your child’s physical health”, which CYP reported, 2, “only a little” and parent reported, 3, “somewhat”.

Cost per patient

Intervention costs were calculated overall and per site and baseline and 6 month follow-up SDQ scores were converted to CHU9D health utilities. The total cost of the drop-in service per patient was £358.56, and analysis revealed an average cost per unit increase on the CHU9D of £9,743.60, and an average cost per unit decrease on the SDQ of £183.88. Full analysis including cost per site are presented in Supplementary Material 5.

Harms

None were reported.