Van der Post and colleagues report on the links between opinions about prior psychiatric treatment and risk of civil detentions in Amsterdam, the Netherlands. The authors conclude that a history of previous involuntary admission is associated with low treatment satisfaction and that more satisfaction seems to reduce the risk of civil detention. This is another way of saying that the three- to fourfold increase in the risk of repeated compulsory admission is mediated by patients’ perspectives. Unfortunately, these conclusions are not supported by the data from the Amsterdam Study of Acute Psychiatry.
First of all, evidence for an association of a history of involuntary hospitalisation and treatment satisfaction is weak at best. Logistic regression analysis was performed including age and previous involuntary admission as predictors. But the modeling strategy did not follow textbook guidelines as developed by Hosmer and Lemeshow (2000), among others. Moreover, the validity of the estimated association between previous admission and treatment satisfaction (Odds Ratio = 2.2) is conditional on the model being correct. However, model fit was poor. The authors reported the percentage of correctly predicted cases: 66%, which is not much of an improvement on a base-rate of 58% dissatisfied patients. Poor model fit is also reflected in a low Nagelkerke’s pseudo r-square (0.09).
Secondly, the association of treatment satisfaction and compulsory admission during follow-up is in fact non-existent in the final logistic regression model. Cox regression analyses would have been more appropriate in order to take into account that the outcome is censored at the end of the two-year follow-up period (Kleinbaum & Klein, 2005). But leaving that aside, the analysis is flawed because bi-variate association and multiple regression analysis are mixed up. The authors claim that more satisfaction with prior treatment reduced the risk of civil detention (OR = 0.3). However, this is a crude OR concerning only a very satisfied subgroup next to patients with medium scores (OR=1.4) and contrasted with a very dissatisfied group of patients (OR=1). In the final model, including previous involuntary admission, treatment satisfaction was non-significant. An impressive 82% of correctly predicted cases was reported together with r-square = 0.25 (Nagelkerke). But this only means that the model is not very helpful because an unadjusted 82% of patients was not admitted involuntarily during follow-up.
By picking favourable odds ratios, Van der Post and colleagues constructed an interpretation of the data to “open up a new perspective for diminishing the risk of (new) civil detention by trying to enhance satisfaction with treatment”. It may be sound advice to ask patients under detention about their previous experiences with mental health care. But this advice is not grounded on thorough statistical analysis. The conclusion is experience-based, not evidence based. Doctor knows best?
Hosmer, D. W., & Lemeshow, S. (2000) Applied Logistic Regression (SE). New York: John Wiley & Sons, INC.
Kleinbaum, D. G., & Klein, M. (2005) Survival Analysis (SE). New York: Springer.
van der Post, L. F. M., Peen, J., Visch, I., Mulder, C. L., Beekman, A. T. F., & Dekker, J. J. M. (2013) Patient perspectives and the risk of compulsory admission: The Amsterdam Study of Acute Psychiatry V. International Journal of Social Psychiatry, 0, 1-9.