Daan Hulsmans

147 Case study challenging behavior 6 mono-causal if-then explanations (e.g., if phase has many familial tensions, then high aggression or if day with hallucination, then self-injury). The multitude of bivariate null-results speaks to the complex nature of these behaviors at the case-level (Ali et al., 2014; Crowell et al., 2009; Griffith & Hastings, 2014; Tevis & Matson, 2022). In the second step, we described the trajectory of challenging behaviors over time. We identified 11 distinct, relatively stable phases within the 560days timeline. These 11 phases could be narrowed down to four qualitatively different attractor states: high levels of self-injury and aggression (2 phases), average levels of self-injury and aggression (5 phases), low levels of selfinjury and aggression (3 phases), or high levels self-injury with low levels of aggression (1 phase). The mean frequency of the 11 staff-hypothesized risk- and protective factors varied by phase: no two phases were similar (Table 1). In the third step we focused on (the week before) transitions between attractors, exploring potential change-inducing mechanisms (Figure 1). Our findings suggest that the mechanism of two transitions remained unknown, two were event-induced, two were instability-induced and four could be environment- and/or instability-induced (Table 2). Six transitions were thus potentially instability-induced, which is in line with empirical evidence for instability as an early warning signal for upcoming transitions (Fartacek et al., 2016; Olthof et al., 2020b; Schreuder et al., 2020). Nevertheless, extraordinary events and/or instability did not unequivocally imply a transition, as both instability and extraordinary events occurred without transitions afterwards (Figure 5). The two unknown mechanisms were both for desirable transitions, which could mean that relatively minor events in daily life apparently were enough to elicit positive change. One possible explanation would be that her desirable attractor is stronger than the undesirable one. That is, we could perceive her undesirable basin (Figure 1) to be shallower, making this state easier malleable relatively minor everyday events. Future research could explore this further with recently developed analytical methods that quantify the stability of an attractor state (Cui et al., 2023). There were four notable limitations to this study. First, results from a case-study are obviously not generalizable. Repeating (and finetuning) our three-step-approach on different cases, will reveal the extent to which of our findings are person-specific or generalizable across cases. This will ultimately increase our understanding of challenging behaviors and consequently enable optimized care. Second, our thematic analysis was based on care professionals’ daily records. Registering relevant events in the electronic health records is a routine practice in the residential care

RkJQdWJsaXNoZXIy MTk4NDMw