Introduction
The ‘PRemIum for aDolEscents’ (PRIDE) project was initiated in 2016 to address the large burden of adolescent mental health problems in India, home to 20% of the global population of adolescents (United Nations, Department of Economic and Social Affairs, Population Division, 2019). Designed initially for urban, low-income secondary schools, the PRIDE intervention model incorporates three design innovations. First, a core set of practice elements have been systematically identified by matching evidence-based practices to common adolescent problems in the local context (Boustani et al., Reference Boustani, Daleiden, Bernstein, Michelson, Gellatly, Malik, Patel and Chorpita2020). These practices are combined within a structured transdiagnostic protocol that targets elevated presentations of anxiety, depression and conduct problems, identified through screening for symptoms and associated distress/impairment (Chorpita et al., Reference Chorpita, Daleiden, Malik, Gellatly, Boustani, Michelson, Knudsen, Mathur and Patel2020; Michelson et al., Reference Michelson, Malik, Krishna, Sharma, Mathur, Bhat, Parikh, Roy, Joshi, Sahu, Chilhate, Boustani, Cuijpers, Chorpita, Fairburn and Patel2020a). Second, PRIDE employs lay counsellors as the primary delivery agents, in line with evidence for the cost-effectiveness of task-sharing in mental health care in diverse low-resource settings (Raviola et al., Reference Raviola, Naslund, Smith and Patel2019). Third, a stepped care architecture (Bower and Gilbody, Reference Bower and Gilbody2005) allows for further resource efficiency, such that a broadly applicable problem-solving intervention (‘Step 1’) is delivered as a brief first-line intervention, followed by a more tailored, higher-intensity second step (‘Step 2’) for non-responders. The blueprinting of these intervention steps has been detailed elsewhere (Chorpita et al., Reference Chorpita, Daleiden, Malik, Gellatly, Boustani, Michelson, Knudsen, Mathur and Patel2020; Michelson et al., Reference Michelson, Malik, Krishna, Sharma, Mathur, Bhat, Parikh, Roy, Joshi, Sahu, Chilhate, Boustani, Cuijpers, Chorpita, Fairburn and Patel2020a), while Step 1 has also been trialled as a standalone intervention (Michelson et al., Reference Michelson, Malik, Parikh, Weiss, Doyle, Bhat, Sahu, Chilhate, Mathur, Krishna, Sharma, Sudhir, King, Cuijpers, Chorpita, Fairburn and Patel2020b; Malik et al., Reference Malik, Michelson, Doyle, Weiss, Greco, Sahu, Jose, Mathur, Sudhir, King, Cuijpers, Chorpita, Fairburn and Patel2021). The latter showed small but sustained effects over 12 months on multiple outcome domains, including mental health symptoms, self-defined problems, functional impairment and perceived stress. Nonetheless, around half of the Step 1 participants did not meet remission criteria in the short term, suggesting the need for extended and more differentiated treatment among early non-responders. Here we report the first formal evaluation of the integrated steppedcare programme, examining acceptability and feasibility with respect to quantitative process indicators, indicative outcome measures and qualitative feedback from intervention participants and providers.
Methods
Design and setting
A mixed-method design was used to assess the acceptability and feasibility of the stepped care programme in three government-aided and two charity-aided secondary schools in New Delhi, India. Three of the schools were co-educational and two were all girls' schools, with sampling frames in grades 8–12 ranging in size from 560 to 1250 across the five schools. None of the schools had pre-existing individual counselling provision. Approvals were obtained from the partner schools and Institutional Review Boards of Sangath (the implementing organisation) and Harvard Medical School (the sponsor).
Participants, measures and procedures
Study referrals were accepted from teachers and directly from students over a 7-month period (July 2019–February 2020). Referrals were generated using a combination of whole-school and classroom-based sensitisation activities (see Parikh et al., Reference Parikh, Hoogendoorn, Michelson, Ruwaard, Sharma, Bhat, Malik, Sahu, Cuijpers and Patel2021 for a detailed description). The former included poster displays on school notice boards and information sessions with teachers and principals. Classroom activities included counsellor-led presentations involving an explanatory video and interactive discussion. Referred students were followed up individually by the research team and assessed for eligibility using the Strengths and Difficulties Questionnaire (SDQ) and its Impact Supplement (Goodman, Reference Goodman2001), which are internationally validated measures of youth mental health symptoms and distress/impairment respectively. Eligibility thresholds are set out in Fig. 1 and elaborated elsewhere (Parikh et al., Reference Parikh, Michelson, Malik, Shinde, Weiss, Hoogendoorn, Ruwaard, Krishna, Sharma, Bhat, Sahu, Mathur, Sudhir, King, Cuijpers, Chorpita, Fairburn and Patel2019). Individual assent was obtained from participants along with parent/guardian consent for adolescents aged below 18 years. At baseline (T0), participants completed the Youth Top Problems (YTP, Weisz et al., Reference Weisz, Chorpita, Frye, Ng, Lau, Bearman, Ugueto, Langer and Hoagwood2011), a self-reported idiographic measure of prioritised psychosocial problems. Participants were then offered a lay counsellor-delivered problem-solving intervention (Step 1). Following Step 1, adolescents were re-assessed (T1) and those who were non-remitted were offered further intervention (Step 2) by a psychologist. As described in Fig. 1, the Step 1 providers included six lay counsellors (one per school and one supplementary counsellor deployed in co-educational schools). Step 2 was delivered by three psychologists (each shared between two schools). All providers received five days of office-based training by intervention developers, followed by weekly supervision for the study duration. Further details about the intervention steps are provided in Fig. 1.
Fig. 1. Description of stepped care components used in the current study
Participants who completed the T1 assessments were followed up at T2 (approximately 12 weeks from T0, or immediately following Step 2 if this extended beyond 12 weeks). Participants who received both intervention steps were additionally invited to take part in semi-structured individual interviews. Qualitative data were also obtained from separate focus group discussions with Step 1 and Step 2 providers. All outcome assessments, individual interviews and focus groups were conducted by researchers who were not involved in intervention development or delivery.
Analysis
Acceptability was examined using quantitative indicators of uptake, completion, and reasons for non-completion for each intervention step. These data were triangulated with themes from qualitative interviews with adolescents (n = 9). Feasibility was examined using quantitative indicators of dosage, duration of intervention steps and transitions, as well as qualitative feedback from Step 1 providers (n = 5) and Step 2 providers (n = 3). Indicative clinical outcomes (SDQ Total Difficulties and YTP scores) were analysed descriptively for those participants who completed assessments across the three time points. The remission rate, operationalised as scoring below initial eligibility cut-offs on both the SDQ Total Difficulties and SDQ Impact scales, was assessed at T2.
Findings
Acceptability
A total of 476 referrals were received, accounting for 43% of the student population in 32 classes sensitised during the study period. As shown in Fig. 2, 337 (71%) of the referred adolescents were assessed for eligibility, from which 80 (24%) completed consent procedures and were enrolled in the study (age: M = 15.3 years, s.d. = 1.4; n = 44, 55% females). From this enrolled sample, 72 (90%) started Step 1 and 49 (61%) attended 4 or 5 sessions, as recommended by the intervention protocol. Sixty-one (76%) study participants were assessed at T1, of which 33 (54%) met non-remission criteria and were thus eligible for Step 2. Only 12 (36%) non-remitted participants opted for Step 2 (age: M = 14.7 years, s.d. = 1.3; n = 4, 33% females) and five (42%) completed the full course (i.e., attended at least five sessions, including final maintenance and relapse prevention session). The remainder of the non-remitted cases (n = 21, 64%) opted out from further treatment, most commonly because of perceived resolution of the primary problem (n = 9, 43%) (see Fig. 2 for other reasons for opting out and non-completion of intervention across both steps).
Fig. 2. Flowchart of research activities with adolescents.
Adolescent interviews suggested that the relatively shorter duration of Step 1 was preferred because it minimised interference with participants' other regular commitments. There was also consensus around the benefits of a strong therapeutic bond and the undesirability of transitioning to a new counsellor between steps. In addition, most interviewees preferred the illustrated comic-book format of the Step 1 booklets compared with the text-based handouts given during Step 2. Nevertheless, most participants felt that Step 2 provided a useful complement to Step 1 by teaching additional coping skills.
Feasibility
Step 1 completers attended a mean of 4.7 sessions (s.d. = 0.5) spread over 22.0 days (IQR = 17.0, 34.0). Step 2 completers attended an additional 7.4 sessions (s.d. = 2.7), spread over 70 days (IQR = 37.0, 74.0), with a median gap of 31 days (IQR = 22.0, 67.3) between steps. These feasibility indicators were longer than planned for Step 2 and for the transition between the two steps (Fig. 1).
In line with qualitative feedback from adolescents, providers endorsed a shorter intervention and greater continuity between steps. To this end, they recommended (i) more regular availability of school-based counsellors to enable flexibility in scheduling, (ii) use of the same trained lay counsellors to deliver both intervention steps, and (iii) counsellors (rather than researchers) conducting step-up assessments to enable more dynamic and collaborative decision-making.
Clinical impact
Complete outcome data were available for 41 participants (see Fig. 2). As shown in Fig. 3a and b, the cohort of 22 participants who were remitted at T1 sustained overall improvements in their SDQ Total Difficulties and YTP scores up to T2 without further treatment being provided. Among the group that was non-remitted at T1 (n = 5), completing Step 2 coincided with reduced scores from T1 to T2. Among the participants who were eligible for, but did not initiate or complete Step 2 (n = 14), outcomes were unchanged from T1 to T2. At the individual case level, 18 out of 22 (82%) Step 1 remitters remained remitted at T2. Step 2 completers were more likely to remit (3 out of 5, 60%) compared with counterparts that did not start or otherwise did not complete Step 2 (3 out of 14, 21%).
Fig. 3. Outcomes for participants on SDQ and YTP (mean and s.d.) across three time points (n = 41).
Discussion
This study examined the acceptability and feasibility of a novel stepped-care programme targeting common adolescent mental health problems in India. The results suggested that teaching additional coping skills could help to improve outcome trajectories for non-responders to a first-line intervention. However, several factors deterred uptake of Step 2. These included a mismatch between the stepping up criteria and adolescents' subjective perceptions of improvement, delays in stepping up, dissatisfaction with changing counsellors, and the prolonged overall treatment duration. Further, the relatively limited appeal of text-heavy printed handouts affected adolescents' engagement with Step 2 materials.
Following from these results, four key modifications have been incorporated into an optimised stepped care protocol. First, a two-provider model has been superseded by an arrangement where a single trained (lay) counsellor delivers both steps. Second, we have changed the decision-making procedures for stepping up. Rather than deploying binary cut-offs based on standardised symptom-based measures, counsellors focus on trajectories of change on the YTP, a person-centred outcome measure (Krause et al., Reference Krause, Edbrooke-Childs, Singleton and Wolpert2021). The new system emphasises shared decision-making, involving collaborative discussions that explore the index adolescent's expectations and perceived progress, leading to a shared understanding about the need for, and potential focus of, additional treatment (Guinaudie et al., Reference Guinaudie, Mireault, Tan, Pelling, Jalali, Malla and Iyer2020). Third, the preference for an accelerated delivered schedule has been addressed by capping the maximum number of Step 2 sessions at six and providing practice materials earlier in the intervention so that participants can use these at their own pace. Fourth, we have re-designed the adolescent-facing resource materials for Step 2 to ensure better consistency with illustrated booklets used in Step 1. A future report will provide evidence on the modified stepped care protocol, including data on its use remotely during the Covid-19 pandemic when schools were physically closed.
To our knowledge, this is the first study to evaluate a stepped care mental health intervention for adolescents in a low- or middle-income country. Strengths of the study include the triangulation of multiple data sources, and a participant sample with diverse mental health presentations that reflect real-world case mix, thus strengthening the generalisability of findings. However, the findings need to be interpreted with caution, given the lack of a control condition, the small number of participants who ultimately completed both intervention steps, and the relatively large proportion of Step 1 participants who were unavailable for follow-up. Moreover, future research is needed to assess the PRIDE stepped care model and other stepped protocols across diverse populations and resource contexts, such as out-of-school youth and in rural communities. Research is also needed to investigate how the competencies required to deliver shared decision-making can be operationalised, learned and used effectively by lay counsellors.
Conclusion
This study provides preliminary evidence for a sequential stepped-care programme addressing common adolescent mental health problems in a low-resource setting. Several modifications have been proposed to ensure that the next iteration of the stepped care protocol is more collaborative and attuned to changes in young people's prioritised problems (rather than symptoms per se), can be delivered in a time-efficient manner, and affords adequate continuity between steps. Future research is planned for the modified stepped care protocol, particularly with respect to the use of intervention in the context of the COVID-19 pandemic and the feasibility of implementing shared decision-making.
Acknowledgements
We acknowledge the contributions of adolescents, their parents and guardians, school staff, research staff and counsellors who made this work possible.
Financial support
This work was supported by the Wellcome Trust, UK (Grant number 106919/Z/15/Z). The funder had no role in study design, data collection and analysis, writing of the report, or in the decision to submit the paper for publication.
Conflict of interest
The authors declare no conflict of interest.
Ethical standards
The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki Declaration of 1975, as revised in 2008. The study was approved by the Institutional Review Boards of Sangath, India (Reference number: VP_2015_017) and Harvard Medical School, USA (Reference number: IRB17-0379). The written consent procedure is detailed in the manuscript.
Open access
