Excessively high suicide rates among psychiatric inpatients

May 1 2015

People with mental disorders have a higher risk of mortality compared with the general population. Among those with severe mental disorders, life expectancy is reduced by around 15 to 20 years, which is equivalent, or worse than, the detrimental effect of heavy smoking (Chesney et al., 2014).

Suicide is an important contributor to the premature mortality experienced by those with mental disorders. Those with conditions such as depression, bipolar disorder and schizophrenia are more than 10 times more likely to take their life compared to the general population.

When individuals are a considered to be a risk to themselves they are sometimes hospitalised in order to prevent suicide from occurring. However, it is well-known that suicides still take place within psychiatric inpatient settings and the rate at which they occur can be an important indicator of how mental health services are performing.

A recent study by Walsh et al (2015) highlights the scarcity of information relating to what might be considered a high or low rate of inpatient suicide. They therefore aimed to calculate estimates of:

The number of psychiatric admissions per inpatient suicide
The number of suicides per 100,000 inpatient years

They also examined whether variation in these estimates could be explained by:

The year in which suicides occurred
The background population rate of suicide
Average length of stay in the psychiatric hospital
Whether coronial records were used to establish the number of suicides

Research shows that, taking all mental disorders into account, suicide risk is highest in people with borderline personality disorder, anorexia, depression and bipolar.

Methods

The authors conducted a systematic review and meta-analysis of the literature which reported inpatient suicides taking place within general mental health settings
They defined a general mental health setting as a psychiatric ward or hospital, or groups of these within a designated geographical area
Inpatient suicides included those completed by registered inpatients, which occurred on hospital grounds, or whilst on approved or unapproved leave
They searched 3 medical databases and conducted a hand-search of relevant reviews and references of included studies
Studies published between January 1945 and December 2013 were considered
They used a random-effects meta-analysis to calculate the pooled suicide rates

The study reported on inpatient suicides that took place within general mental health settings.

Results

The search generated 834 relevant hits and 32 papers were identified via the hand-search. After reviewing the abstracts and full-texts, 44 papers were selected for inclusion. Of these:

39 reported data on the number of suicides and the number of admissions
27 reported on either the number of suicides per 100,000 in-patient years or contained sufficient information allowing calculation
22 papers included data to answer both questions

Included studies were grouped into five regions: Australasia, Continental Europe, Nordic countries, UK and Ireland, and the USA.

Psychiatric admissions per inpatient suicide

The 39 studies included information on 6,832,071 psychiatric admissions
The estimated number of inpatient admissions per suicide was 676 (95% CI: 604 to 755)
Examining trends over time, the authors found that the number of admissions per suicide dropped up until 1999, indicative of an increasing risk of inpatient suicide
However, after the year 2000 the risk of inpatient suicide fell as the number of admissions per suicide increased
The lowest number of admissions per suicide was found in Australasia and the highest in UK and Ireland
There was a suggestion that as the population suicide rate increased the risk of inpatient suicide also increased, but this was not statistically significant

Suicides per 100,000 inpatient years

The estimated number of suicides per 100,000 in-patient years was 147 (95% CI: 138 to 156), which was drawn from 27 studies including 5,766 suicides
However, analysis suggested publication bias may be an issue and the estimated rate was probably higher than expected
The number of suicides per 100,000 inpatient years increased substantially from 1980
Studies from the USA reported the highest rate, with those from Nordic countries exhibiting the lowest
A higher population rate of suicide and use of coronial records was related to a higher number of suicides per 100,000 inpatient years, whereas, a longer length of stay was associated with a decreased rate

Rates of in-patient suicide in psychiatric hospitals vary remarkably and are disturbingly high

Inpatient suicide rates varied considerably across different psychiatric hospitals, but overall rates were disturbingly high.

Conclusions

Substantial variation in the number of admissions per suicide and the number of suicides per 100,000 inpatient years was apparent between studies, which the authors suggest is the more important finding compared to the estimated rates. They also found that studies that rely on hospital records are likely to underestimate the number of cases. The authors conclude that the number of admissions per suicide may be a more important measure of inpatient suicide as it was less influenced by the length of stay and the rate of suicide among the general community.

The estimated rates of inpatient suicides are excessively high when compared with the general population. Inpatient suicide therefore represents a significant public health issue that deserves more attention. Reducing inpatient suicide could be one way to narrow the inequality in mortality between those with and without mental disorders.

Studies that don’t use coroners’ records may be under-reporting the true level of inpatient suicides.

Strengths and limitations

The study has a number of strengths including a comprehensive search of the literature and its consideration of suicides rates using two methods.

Limitations include the exclusion of non-English language articles and the fact that no studies were identified from low and middle income countries. Thus, the calculated figures may be an under-estimate of global trends.

Although the authors considered data quality by examining suicides established by coronial records, the review lacked a comprehensive assessment of the quality of studies included.

In addition, although the authors only use it once, it was disappointing to read the phrase “committing suicide”. This language increases stigma, which ultimately reduces help-seeking behaviour (Beaton et al., 2013).