Within acute psychiatric inpatient services, patients exhibiting severely disturbed behaviour can be transferred to a psychiatric intensive care unit (PICU) and/or secluded in order to manage the risks posed to the patient and others. However, whether specific patient groups are more likely to be subjected to these coercive measures is unclear. Using robust methodological and statistical techniques, we aimed to determine the demographic, clinical and behavioural predictors of both PICU and seclusion.

Data were extracted from an anonymised database comprising the electronic medical records of patients within a large South London mental health trust. Two cohorts were derived, (1) a PICU cohort comprising all patients transferred from general adult acute wards to a non-forensic PICU ward between April 2008 and April 2013 (N = 986) and a randomly selected group of patients admitted to general adult wards within this period who were not transferred to PICU (N = 994), and (2) a seclusion cohort comprising all seclusion episodes occurring in non-forensic PICU wards within the study period (N = 990) and a randomly selected group of patients treated in these wards who were not secluded (N = 1032). Demographic and clinical factors (age, sex, ethnicity, diagnosis, admission status and time since admission) and behavioural precursors (potentially relevant behaviours occurring in the 3 days preceding PICU transfer/seclusion or random sample date) were extracted from electronic medical records. Mixed effects, multivariable logistic regression analyses were performed with all variables included as predictors.

PICU cases were significantly more likely to be younger in age, have a diagnosis of bipolar disorder and to be held on a formal section compared with patients who were not transferred to PICU; female sex and longer time since admission were associated with lower odds of transfer. With regard to behavioural precursors, the strongest predictors of PICU transfer were incidents of physical aggression towards others or objects and absconding or attempts to abscond. Secluded patients were also more likely to be younger and legally detained relative to non-secluded patients; however, female sex increased the odds of seclusion. Likelihood of seclusion also decreased with time since admission. Seclusion was significantly associated with a range of behavioural precursors with the strongest associations observed for incidents involving restraint or shouting.

Whilst recent behaviour is an important determinant, patient age, sex, admission status and time since admission also contribute to risk of PICU transfer and seclusion. Alternative, less coercive strategies must meet the needs of patients with these characteristics.

Clinicians need guidance to address the heterogeneity of treatment responses of patients with major depressive disorder (MDD). While prediction schemes based on symptom clustering and biomarkers have so far not yielded results of sufficient strength to inform clinical decision-making, prediction schemes based on big data predictive analytic models might be more practically useful.

We review evidence suggesting that prediction equations based on symptoms and other easily-assessed clinical features found in previous research to predict MDD treatment outcomes might provide a foundation for developing predictive analytic clinical decision support models that could help clinicians select optimal (personalised) MDD treatments. These methods could also be useful in targeting patient subsamples for more expensive biomarker assessments.

Approximately two dozen baseline variables obtained from medical records or patient reports have been found repeatedly in MDD treatment trials to predict overall treatment outcomes (i.e., intervention v. control) or differential treatment outcomes (i.e., intervention A v. intervention B). Similar evidence has been found in observational studies of MDD persistence-severity. However, no treatment studies have yet attempted to develop treatment outcome equations using the full set of these predictors. Promising preliminary empirical results coupled with recent developments in statistical methodology suggest that models could be developed to provide useful clinical decision support in personalised treatment selection. These tools could also provide a strong foundation to increase statistical power in focused studies of biomarkers and MDD heterogeneity of treatment response in subsequent controlled trials.

Coordinated efforts are needed to develop a protocol for systematically collecting information about established predictors of heterogeneity of MDD treatment response in large observational treatment studies, applying and refining these models in subsequent pragmatic trials, carrying out pooled secondary analyses to extract the maximum amount of information from these coordinated studies, and using this information to focus future discovery efforts in the segment of the patient population in which continued uncertainty about treatment response exists.