Behind closed doors: unlocking hospital variations in the use of seclusion and mechanical restraint - a nationwide multilevel analysis in adult mental healthcare in France.
Monitoring and reducing restrictive practices in mental healthcare, including seclusion and restraint, is a global policy priority due to ethical concerns. However, research on these practices remains limited. This study aims to estimate the use of seclusion and mechanical restraint (MR) in mental healthcare at the national level in France, the variation in the use of these practices across hospitals, and associated factors considering a wide range of variables.
We used the French nationwide system for the routine collection of psychiatric hospital claims data, which includes a specific register on seclusion and MR, covering all hospitals providing involuntary psychiatric care for adults in France. Data from 204 hospitals were analyzed for the year 2022. Additional information on patient, hospital, and contextual characteristics was obtained by linking data from other national sources, including somatic hospital and community care claims, annual hospital surveys and census data. Using these combined datasets, we conducted multilevel logistic regressions to identify factors significantly associated with variations in the use of restrictive practices.
In 2022, among 98,271 involuntary psychiatric inpatient admissions across 204 hospitals, seclusion was used in 31,679 admissions (32%) and MR in 8,551 admissions (9%). Large variations in the use of restrictive practices were observed across hospitals, with some reporting no use of MR. Multilevel models showed that patient characteristics, including primary diagnosis and prisoner status, were among the strongest predictors of seclusion and MR use. Admissions to public hospitals with teaching activities or multidisciplinary services were associated with lower odds of seclusion, while a higher nurse-to-patient ratio and proportion of involuntary admissions were associated with lower odds of MR use. No contextual factors related to hospitals’ catchment areas were significantly associated with the use of restrictive practices.
Significant variations were found in the use of seclusion and MR across hospitals in France, which cannot be explained solely by patient needs. While patient characteristics were strong predictors, several hospital-level factors were also associated with these practices. These findings highlight the need for closer monitoring and targeted interventions to reduce unwarranted variations and raise concerns about equity and ethics in mental healthcare.
Not applicable.
The online version contains supplementary material available at 10.1186/s12888-025-07404-z.
We used the French nationwide system for the routine collection of psychiatric hospital claims data, which includes a specific register on seclusion and MR, covering all hospitals providing involuntary psychiatric care for adults in France. Data from 204 hospitals were analyzed for the year 2022. Additional information on patient, hospital, and contextual characteristics was obtained by linking data from other national sources, including somatic hospital and community care claims, annual hospital surveys and census data. Using these combined datasets, we conducted multilevel logistic regressions to identify factors significantly associated with variations in the use of restrictive practices.
In 2022, among 98,271 involuntary psychiatric inpatient admissions across 204 hospitals, seclusion was used in 31,679 admissions (32%) and MR in 8,551 admissions (9%). Large variations in the use of restrictive practices were observed across hospitals, with some reporting no use of MR. Multilevel models showed that patient characteristics, including primary diagnosis and prisoner status, were among the strongest predictors of seclusion and MR use. Admissions to public hospitals with teaching activities or multidisciplinary services were associated with lower odds of seclusion, while a higher nurse-to-patient ratio and proportion of involuntary admissions were associated with lower odds of MR use. No contextual factors related to hospitals’ catchment areas were significantly associated with the use of restrictive practices.
Significant variations were found in the use of seclusion and MR across hospitals in France, which cannot be explained solely by patient needs. While patient characteristics were strong predictors, several hospital-level factors were also associated with these practices. These findings highlight the need for closer monitoring and targeted interventions to reduce unwarranted variations and raise concerns about equity and ethics in mental healthcare.
Not applicable.
The online version contains supplementary material available at 10.1186/s12888-025-07404-z.
Authors
Touitou-Burckard Touitou-Burckard, Coldefy Coldefy, Bourin Bourin, Ellini Ellini, Saetta Saetta, Gandré Gandré
View on Pubmed