Background: Multi-problem families face problems in several domains that are often found to be chronic and intergenerational. Effective mental health care for youth from these families is currently lacking, urging research on new methods. The InConnection approach is an integrated care program to improve resilience in multi-problem families by connecting the professional expertise from multiple disciplines with the informal social network of the youth. Specifically, youth are asked to nominate a youth initiated mentor (YIM) from among the supportive adults in their network. The aim of this protocol is to describe the design of a mixed-methods study to examine the effectiveness and working mechanisms of the InConnection approach. Method/design: The effectiveness of the InConnection approach is studied in a quasi-experimental questionnaire study using propensity score matching, with N = 300 families with youth aged 10-23 years receiving treatment in either the intervention group (InConnection approach) or the control group (care as usual). The main outcome variables include youth resilience (primary), youth mental health, parental functioning, and the number, duration and types of out-of-home placements. Mediators, moderators, and predictors of effectiveness are examined. Assessments take place at the start of the care program and after three, nine and 15 months. Additionally, semistructured interviews are conducted with families who have and have not nominated a YIM to understand why some families successfully nominate a YIM, whereas others do not. Discussion: Effective care for youth in multi-problem families is urgently needed. Given its flexibility and accessibility to suit all youth aged 10-23 years from multi-problem families, and its low costs compared to out-ofhome placements, the InConnection approach seems an appealing approach to support these families. The current study will provide information on the effectiveness of the InConnection approach. Strengths of this study include its robust design, the ecological validity, and the inclusion of possible mediators, predictors, and moderators of treatment effects.