Social Psychiatry and Psychiatric Epidemiology

, Volume 43, Issue 3, pp 251–256

Mental health laws that require dangerousness for involuntary admission may delay the initial treatment of schizophrenia


    • Private Practice
  • Olav Nielssen
    • Private Practice
    • Clinical Research Unit for Anxiety Disorders, School of PsychiatryUNSW at St Vincent’s Hospital
    • Psychological MedicineUniversity of Sydney
  • Christopher James Ryan
    • Psychological MedicineUniversity of Sydney
    • Westmead Hospital
  • Robert Hayes
    • Mental Health Research and Training Centre (MHRTC), Law SchoolUniversity of Western Sydney

DOI: 10.1007/s00127-007-0287-8

Cite this article as:
Large, M.M., Nielssen, O., Ryan, C.J. et al. Soc Psychiat Epidemiol (2008) 43: 251. doi:10.1007/s00127-007-0287-8



A long duration of untreated psychosis (DUP) is associated with a worse prognosis, an increased risk of suicide and may be linked to serious violence. Mental health laws that require patients to be dangerous to themselves or to others before they can receive involuntary psychiatric treatment may make it more difficult to treat patients in their first episode of psychosis.


The mean and median DUP reported in studies of schizophrenia related psychoses were examined. A comparison was made between the DUP reported from jurisdictions that had an obligatory dangerousness criterion (ODC) and those with other criteria for involuntary treatment.


The average mean DUP in samples from jurisdictions with an ODC was 79.5 weeks, but was only 55.6 weeks in those jurisdictions that did not have an ODC (P < 0.007).


Mental health laws that require the patient to be assessed as dangerous before they can receive involuntary treatment are associated with significantly longer DUP. As reducing DUP is an intervention that can improve the prognosis of schizophrenia, this finding suggests that mental health laws should be amended to allow treatment on grounds other than dangerousness, at least in the crucial first episode of psychosis.

Key words

schizophreniaduration of untreated psychosismental health services


The duration of untreated psychosis (DUP) is usually defined as the period between the emergence of psychosis and the initiation of adequate treatment with antipsychotic medication [50]. Longer DUP has been associated with a worse overall prognosis [43, 53] an increased risk of suicide [3, 14, 46], serious violence [47, 67] and homicide [36]. A public health approach to encourage earlier treatment may reduce DUP [38].

Since the mid 1970s the mental health laws in many jurisdictions have been amended to limit involuntary admission to those who have been assessed as dangerous to themselves or to others. This requirement is known as an obligatory dangerousness criterion (ODC). An ODC is a feature of the mental health acts of all states of the USA and Australia, five Canadian provinces and six European countries. The mental health acts of the remaining Canadian provinces and European nations (including the UK) include alternative criteria for involuntary treatment, so that dangerousness is not obligatory. Jurisdictions that do not have an obligatory dangerousness criterion usually allow involuntary treatment on the grounds of an assessed need for treatment if the patient is deemed unable to give consent.

It has been argued that the introduction of the ODC caused only a temporary change in patterns of involuntary treatment as clinicians and those who enforced the legislation soon worked around any legal changes in order to provide necessary care [5]. However, the treatment of patients in their first episode of psychosis (FEP) may be more likely to have been affected by the introduction of an ODC than patients with established illness. The assessment of risk of harm that is required in legislation with an ODC relies heavily on a history of previous episodes of violence or self-harm. Patients in their first episode of illness are less likely to have a history of dangerous behaviour arising from mental illness than patients with established illness. As a consequence it may be more difficult for clinicians to argue that first episode patients are dangerous. Although there are a number of studies that compared the characteristics of populations of hospitalised patients before and after changes in mental health law [4] none specifically examine the effect on patients in the FEP and these studies could not have detected the effect on FEP patients as they make up a small proportion of all patients because of the chronic nature of schizophrenia.

The hypothesis of this study was that FEP patients living in jurisdictions that have an ODC rather than alternative criteria for involuntary treatment would take longer to receive treatment and hence have a longer DUP.

In order to test this hypothesis, we compared the average mean and average median DUP of studies conducted in jurisdictions that had an ODC with those of studies conducted in jurisdictions that had alternative criteria for involuntary treatment. Our independent variables were patient age, sex, diagnosis and whether the study was reported from a specialist first episode psychosis service. In an attempt to take into account the effect of the overall standard of health care, we also considered the number of psychiatrists and psychiatric hospital beds per capita, and the way mental health care was funded.


The electronic databases [Medline], [Embase], [Psychlit] and [Pubmed] were searched for peer reviewed studies conducted between 1976 and 2006 and published in English, using the following search strategies and terms: “duration of untreated psychosis” or “first-episode psychosis” and “onset” and “schizophrenia”. The references of relevant articles were hand searched for other studies. We included studies that reported the mean or median DUP and studies from which we could calculate the mean DUP by subtracting the mean age of onset of psychotic symptoms from the mean age of initial treatment if both were reported to one decimal point.

We included samples of the DUP of schizophrenia spectrum psychosis collected after 1976 from developed western countries. We excluded studies from developing countries and from Asian countries, as these countries generally report long periods of DUP and may not have comparable mental health law. Data was extracted independently by ML and ON and several disagreements due to the conversion of time units were unambiguously resolved.

We determined the relevant jurisdiction for each sample and then ascertained which jurisdictions had an ODC for involuntary treatment [4, 5, 20, 26, 28]. Studies that sampled patients from more than one jurisdiction were excluded from the analysis if the jurisdictions differed in their use of an ODC.

We recorded the mean age and proportion of male and schizophrenic patients from each sample. We also recorded whether the studies were conducted by specialist early psychosis services. We then used data available from the 2005 WHO Project Atlas of the Department of Mental Health and Substance Abuse [70] to determine the number of psychiatrists per 100,000 people, the number of psychiatric beds per 10,000 people and whether or not mental health treatment is paid for by government funded universal health schemes in each country.

The mean DUP values were not normally distributed (skewness 1.0, P < 0.003) and were therefore Log10 transformed. We used an unpaired, two tailed t test for two group comparisons of continuous data and a chi-square test for categorical data. Stepwise linear regression weighted for sample size was used in a multivariate analysis. The statistical analysis was performed using version 15.0 of SPSS.


We examined 301 papers and located 98 publications with non-overlapping samples of the DUP. Of these, eight studies from developing countries, twelve studies from Asian countries, five multinational studies that included subjects from both ODC and non-ODC countries and 24 studies that reported samples with varying proportions of patients with bipolar disorder and psychotic depression were excluded. Two studies of treatment delay were excluded because they only reported the proportion of treated patients at fixed time intervals. Data collection in the earliest study started in 1978. There were a total of 64 non-overlapping samples from 47 published studies that met the inclusion criteria, that were from Australia [32], Canada [1, 6, 8, 10, 15], Denmark [42, 45, 54], Finland [35, 58], Germany [24, 29, 56], Ireland [13], Italy [2, 3, 44, 49], The Netherlands [39, 71], Norway [23, 37, 45, 46, 48], Spain [18, 35, 52], Sweden [11], the United Kingdom [7, 9, 19, 21, 30, 57, 59, 61, 65] and the USA [12, 16, 17, 25, 27, 33, 40, 41, 55, 62, 63]. The samples had an average of 88 subjects (SD 79) and described the DUP of a total of 5,849 subjects.

The average mean DUP of all the subjects was 65.6 weeks (SD 38.4). There was evidence of significant positive skew in many of the samples, as the average mean DUP was much longer than the average median DUP of 22.9 weeks (SD 15.5). The presence of some patients with very long DUP was also evident in the studies that reported the range, and by the standard deviation exceeding the mean in almost all the samples.

The average mean DUP in samples from jurisdictions with an ODC was 43% longer (ODC; 79.5 weeks, 95% CI 63.5–95.4 weeks, median 71 weeks) than those jurisdictions that did not have an ODC (No-ODC; 55.6 weeks, 95% CI 43.4–68.8 weeks, median 49 weeks). The weighted mean DUP for the subjects in the ODC group was 77.7 weeks and was 55.7 weeks in the non-ODC group, an average difference of more than 5 months (Table 1, Fig. 1).
Table 1

Mean DUP and Median DUP in jurisdictions that differ in their use of a danger criterion


Regions sampled

Obligatory danger-criterion samples

USA, Australia, France, Germany, The Netherlands, Ontario, Quebec

No obligatory danger-criterion samples

Denmark, Finland, Ireland, UK, Norway, Italy, Spain, British Columbia, Nova Scotia



Number of samples available to calculate mean




Number of samples available to calculate median




Weighted mean DUP




Average mean DUP weeks (SD/median)

79.5 (38.6/71)

55.6 (35.6/49)

t = 2.82


Average median DUP weeks (SD/median)

27.5 (16.3/28.0)

19.9 (14.5/16.0)

t = 1.37


Mean age at initial treatment years (SD)

24.3 (3.8)

26.8 (2.6)

t = 3.10


Percentage male (SD)

60.3 (22.9)

61.6 (18.5)

t = 0.24


Mean % of patients with schizophrenia (SD)

89.0 (15.4)

80.3 (28.6)

t = 1.67


Specialized early psychosis service



χ2 = 0.04


Government funded health scheme



χ2 = 20.5


Psychiatrists per 100,000 population

10.34 (1.75)

11.1 (4.21)

t = 0.85


Psychiatric beds per 10,000 population

13.7 (4.6)

11.0 (5.4)

t = −2.01

Fig. 1

The distribution of mean and median DUP values from studies in regions with an ODC and with No-ODC. Line and diamond two standard deviations and 95% CI. Box median, 25th and 75th percentile values. Cross most extreme outlier

The average median DUP was 38% longer in the ODC samples (ODC; 27.5 weeks, 95% CI 17.3–37.3 weeks, median 28 weeks) than the non-ODC samples (No-ODC; 19.9 weeks, 95% CI 12.9–26.9 weeks, median 16 weeks) a result that was not significant. The lack of significance may have been due to the smaller number of samples for which a figure for median DUP was reported.

Samples from ODC jurisdictions also reported a significantly lower mean age of initial treatment and a non-significantly larger number of patients diagnosed with schizophrenia. Jurisdictions with an ODC were less likely to have universal mental health care funded by taxation and had a slightly higher number of psychiatric hospital beds. However, a stepwise least squares regression, weighted for sample size found that the ODC was the only significant factor associated with Log10 mean DUP after one step (Table 2).
Table 2

The associations between ODC, covariates and Log10 Mean DUP*


Included variables






Presence of ODC













Variables excluded after one step

Beta In



Partial correlation

Colinearity (tolerance)

Proportion of patients with schizophrenia






Mean age at first treatment






Percentage male






Specialized early psychosis service






Tax based mental health funding






Psychiatrists per 100,000 population






Psychiatric beds per 10,000 population






*Stepwise least squares regression weighted for sample size


The mean DUP for all of the studies was over a year. However, in those jurisdictions with an ODC, the mean DUP was on average about 5-months longer. The longer mean DUP in regions with an ODC may have been partly due to differences in the distributions of DUP samples, as the longer mean DUP may have been due to a minority of patients with very long DUP who were not initially considered to be dangerous, but who received treatment at a later date.

In addition to the association with a long DUP, the presence of an ODC was associated with a slightly lower mean age at presentation. The most likely explanation for this is that it reflects the different demographics of some of the countries with an ODC. For example the USA and Australia have a larger proportion of young people than the UK and many other European countries that do not have an ODC in their mental health law [66]. Furthermore, several of the studies conducted in the jurisdictions with an ODC were providing specific services for younger people [32, 39, 56] and had a lower mean age. However, this result might also be due to younger patients receiving treatment earlier because they are more aggressive while some older patients may have missed out on treatment in ODC regions because they were not assessed as dangerous. If younger more dangerous patients are treated earlier in ODC jurisdictions an ecological fallacy could arise in which the mean age of treatment is lower, but the very delayed treatment in a proportion of patients significantly prolongs the mean DUP of the whole group.

Several limitations of the study should be acknowledged. First we were unable to estimate the proportion of first episode patients whose treatment was delayed as a result of the presence of an ODC, partly because the median DUP was reported in fewer samples than the mean. If the median DUP had been reported in more of the studies it may have been possible to assess if an ODC prolongs the DUP of the majority of patients, rather than increasing the number of subjects with very long DUP.

It is possible that the ODC prolongs the DUP of a significant proportion of patients. As many as 80% of FEP patients require inpatient treatment early in their illness [60, 69], and half of those are admitted within one week of contact with services [60]. In many cases admission has to be involuntary as few patients recognise that their symptoms are due to an illness [22, 64]. A significant proportion of these patients might be considered to be dangerous during their FEP. About a third of first episode patients have some suicidal ideas [51] and as many as 20% of patients may make an actual suicide attempt [3]. Aggressive behaviour is also common. About 20% of first presentation patients threaten serious violence [34], and about 10% commit a serious act of aggression [47] or an actual assault [68]. These patients and those with actual or threatened self-harm would be legally regarded as dangerous and could receive treatment under mental health law with an ODC. However, it is not known how many patients are not initially dangerous but later become so, and what proportion of patients are not dangerous and initially refuse treatment but accept treatment later in their illness.

A second limitation of the study is that we did not consider the DUP of two groups of patients with psychosis. Studies that included patients with affective psychosis were not included so that we could compare samples with similar patients and only a few of the studies recorded the DUP of those who were unwilling to participate in research. Furthermore, areas in which FEP research is conducted may have better services and hence a shorter DUP than other areas in the same jurisdictions.

A third limitation stems from the fact that ODC were mostly introduced before studies of DUP were conducted or published, which precluded a direct examination of the effect of the changes in legislation on DUP in any jurisdiction.

Finally it is also possible that the association between the presence of an ODC and long DUP is not causal. DUP is thought to be associated with a range of illness, patient, family and cultural factors. Some of these factors, although they may be difficult to quantify on a population basis, might by coincidence prolong DUP in the regions with an OCD. However, it is likely that at least some of the increase in DUP in regions with an OCD is a direct result of differences in mental health law, as we found no independent association between longer DUP and a range of patient and health service factors. In addition the finding is consistent with an a priori hypothesis that was formed on the basis of both a plausible mechanism and clinical observations made in a jurisdiction with an ODC.

The widespread adoption of the ODC in mental health laws arose from the desire to balance the rights of the mentally ill with the need to protect the public [31]. In reality, the introduction of the ODC may have had the unwanted and harmful effect of delaying the initial treatment of patients with psychosis.

Copyright information

© Springer-Verlag 2007