Thursday, August 18, 2016

Open Psychiatric Units Mean Fewer Suicides and Elopements ?!!

There is a headline making the rounds in the media about whether or not locked psychiatric units are useful in preventing suicides and "absconding" behavior in inpatient psychiatric units.  Absconding is running away before the formal discharge and in the US it is referred to as elopement.  The media handling of this article is a bit less scholarly than you might expect from the average psychiatrist reading this article.  Even media circulating to psychiatrists sends out the headlines from a news service:  "Locked psychiatric wards may mean more suicide or escape attempts." Since I have spent the majority of my career on locked psychiatric units and consider myself an expert in this area - reading the article and looking at its deficiencies comes naturally to me.

The article looks at a coalition of 22 German psychiatric hospitals and their affiliated psychiatric services.  Sixteen of the hospitals had at least one locked psychiatric unit over the course of the study.  Four hospitals had no locked wards over the course of the study.  One of the hospitals started out with no locked wards but "had to introduce locked wards for legal reasons" in November 2000. organized under a central agency that looks at quality assurance and quality management.  Twenty one of the hospitals participated in data analysis by a central quality assurance/management agency the Dokumentationsverbund Psychiatrie (DVP).  The study period ran from January 1, 1998 to December 31, 2012.  This was  an entirely retrospective analysis based on anonymized data.  During the study period there were 271,128 admissions to locked wards and 78,446 admissions to open wards.

Primary outcome variable was completed suicide and secondary outcome variables were suicide attempts during treatment, elopement without return, and elopement with return.  Some of the characteristics of the populations were described and they appear to have diagnoses similar to what might be found on inpatient units in the US with major difference - some of the primary diagnoses listed would likely not be admitted - like somatoform disorder or personality disorder as a primary diagnosis, but the study says very little about admission criteria.  On American inpatients psychiatric units pure substance use disorders are actively discriminated against, by insurers and government agencies that govern hospitalizations and in the German sample. they constitute 18-25% of the primary diagnoses.  The authors do a statistical comparison between the locked ward and open ward groups across the outcome variables.               

There are two logical flaws with the study and the researchers comment on one.

The first is generalizability of the data.  The authors seem to recognize this in their use of OECD data and the rates of psychiatric bed utilization in Germany (2.8 per 1,000 population) versus the UK (0.5 per 1,000) and the US (0.3 per 1,000)  suggesting that there is greater acuity in the populations with fewer beds and that there is a greater proportion of acutely ill patients.  The other parameter that is critical in American inpatient psychiatry is the number of aggressive and homicidal patients.  At large metropolitan hospitals units comprised almost entirely of highly aggressive patients are not unusual.  Other patients are generally considered too vulnerable to be admitted to these units.

Aggressive behavior can create near riot conditions on units like this and an unlocked door would create numerous situations leading to violent confrontations with staff.  The striking part about this comparison to the German system was that this paper left out all mention of aggression, violence and homicide suggesting that these patients were not being admitted to these hospitals.  The only line containing these words in the entire paper was in one of the references.  That makes this study impossible to compare with any set of metropolitan psychiatric units in the US.  There is the associated question of what the Germans do with their aggressive patients?  Are they sent to forensic hospitals or specialized units?  It would be very unlikely to not encounter thousands of highly aggressive patients in any American sample this large even at a time when the largest psychiatric hospitals in the country are county jails.

The second is that the implicit notion about a randomized controlled trial.  For the reason I previously mentioned it is not likely to be ethical, amenable to human subjects approval or therefore doable.  The authors suggest that being under a mandate to treat all patients in a certain geographic area reduces selection bias.  That is difficult to accept if potential for aggression and overt aggressive behavior is not an admission criteria and if it not compared between the locked and unlocked units.      

That said, what can American psychiatric units learn from the German experience?  The first and most important is that unlocked units are possible.  I worked at a facility that typically had 4 psychiatric units and when we started one unit was open.  It was a transitional unit where people were sent after their acute disorder, agitation, aggressive behavior. and suicide risk was treated but they were not quite ready for discharge.  The management of psychiatric units by business managers eventually dictated that these partially stabilized people should just be discharged - frequently when there had been an almost imperceptible improvement.  This was all based on the fallacious "dangerousness" argument by managed care companies.  They decided about 20-25 years ago that the only reason anyone should be hospitalized on an inpatient psychiatric unit was if they were dangerous to themselves or others.  That also led to locked hospital wards, if not by implication by explicit managed care feedback as in: "If the patient does not need to be on a locked unit - they don't need to be in the hospital and therefore we are denying payment for this admission."  Have these managed care tactics dumbed down inpatient treatment and adversely affected the atmosphere of these units?  Of course it has.  It has created a palpable corrections-like atmosphere in many units.  The only reason people are there is to figure a way to get out.  This reinforces the thought that the people there really don't have any problems in the first place they are just being discriminated against.  So the first lesson from Germany is to restore the running of hospitals to psychiatrists and not business managers.

The second issue is infrastructure and length of stay (LOS).  Most EU countries have significantly more psychiatric beds available to their populations.  The most likely reason is that they are not rationed (nearly out of existence) by managed care companies or the government like they have been in the US.   Lengths of stay are also significantly greater.  The interesting dimensions for comparison would be the functional status of patient at discharge as well as the therapeutic milieu in comparing German to American units.  That would require a more sophisticated research approach but it might bring some science to inpatient psychiatric care.  It would also be interesting to know if the German hospitals have state of the art specialized programs for specific conditions and whether their environment is designed to emphasize the therapeutic rather than containment aspects.

There is also the opportunity to look at the administrative aspects of these units more specifically the impact of a business management approach to a more clinical or at least less of a short term profit approach.

We have all witnessed what healthcare businesses can do to inpatient care in the US - and it is never good. 

George Dawson, MD, DFAPA


1: Huber CG, Schneeberger AR, Kowalinski E, Fröhlich D, von Felten S, Walter M, Zinkler M, Beine K, Heinz A, Borgwardt S, Lang UE. Suicide risk and absconding in psychiatric hospitals with and without open door policies: a 15 year, observational study. Lancet Psychiatry. 2016 Jul 28. pii: S2215-0366(16)30168-7. doi: 10.1016/S2215-0366(16)30168-7. [Epub ahead of print] PubMed PMID: 27477886.

No comments:

Post a Comment