Setting

The study took place in the Jonathan Swift Clinic in St James's Hospital, a 51-bed general adult and old age psychiatry unit serving the Dublin South City catchment, which had a total population of 136 000 in 2006 ⁶ and 135 000 in 2007. ⁷ The clinic is a designated approved centre for treatment of mental disorders under the Mental Health Act. ⁷ All admissions to the unit from January to October 2006 and January to October 2007 were included. We allowed 2 months following the introduction of the new Act as a period of adjustment. The same months were chosen in each year to control for potential seasonal variations in admissions. We obtained ethical approval for this study from the Internal Audit and Research Department of St James' Hospital.

Data collection and analysis

A retrospective case-note review was undertaken. We used a pro forma adapted from that used by Smith & White. ^{Reference Smith and White5} The legal status on admission and discharge was noted and the person's length of stay as an in-patient was calculated, both for periods under the Act and for periods during the same admission for which voluntary status applied. The proportion of each involuntary admission spent detained was calculated by dividing the duration of detention by the overall duration of admission. We used a t-test to determine if this proportion varied from year to year, but we transformed the proportions using arcsine/square root transformation, as t-tests are not appropriate for simple proportions. To determine whether there was a difference between the rates of involuntary admission year on year, we used a z-test, with 95% confidence intervals for the difference in rates. Otherwise, we used t-tests and χ²-tests, with continuity correction as appropriate, to detect differences between groups.

Characteristics of all admissions

There were 770 admissions during the study period. In total, 671 admissions (87.1%) were voluntary and 99 (12.9%) were involuntary. The characteristics of all admissions are in Table 1, and of involuntary admissions in Table 2. For involuntary patients, the mean duration of detention was 45.4 days (s.d. = 41.6, median = 32.0, range 1-168) and the mean duration of all admissions that included an involuntary period was 52.1 days (s.d. = 46.7, median = 38.0, range 1-181). The discrepancy between duration of detention and duration of admission (mean = 6.8 days, s.d. = 15.4) is accounted for by periods that each involuntary patient spent as a voluntary patient, either before involuntary status was commenced or after involuntary status was revoked.

Table 1 Characteristics of all admissions in 2006 (Mental Treatment Act 1945) and 2007 (Mental Health Act 2001)

Variable	Mental Treatment Act 1945 (n = 432)	Mental Health Act 2001 (n = 338)
Male, n (%)	228 (52.8)	170 (50.3)
Duration of admission, days: mean (s.d.)	31.7 (49.8)	28.1 (52.0)
Age on admission, years: mean (s.d.)	43.9 (16.6)	43.0 (16.1)
Involuntary status during admission, n (%)a	46 (10.6)	53 (15.7)

Table 2 Characteristics of involuntary admissions in 2006 (Mental Treatment Act 1945) and 2007 (Mental Health Act 2001)

Variable	Mental Treatment Act 1945 (n = 46)	Mental Health Act 2001 (n = 53)
Male, n (%)	31 (67.4)	29 (54.7)
Age on admission, years: mean (s.d.)a	40.5 (14.2)	48.4 (17.7)
Duration of admission, days: mean (s.d.)	52.1 (44.4)	52.2 (49.0)
Duration of detention, days: mean (s.d.)	49.5 (42.0)	41.8 (41.2)
Diagnosis of schizophrenia or other non-affective psychosis, n (%)	22 (47.8)	32 (60.4)
Comorbid alcohol or substance misuse, n (%)	27 (58.7)	22 (41.5)

Rate of involuntary admission per 100 000 population

In 2006 the total catchment area population was 136 000 and the number of involuntary admissions for the 10-month study period was 46, so that the rate of involuntary admissions was 33.8 per 100 000 population. In 2007, the rate was 39.3 per 100 000 population, as there were 53 admissions in a catchment population of 135 000. This was an increase of 16.3%. We found a z value of −0.74 and a corresponding P value of 0.46 (95% CI for the difference in the rates: −19.83 to 8.96). This indicated that the difference in rates of involuntary admission year on year was not significant. So as to facilitate comparison with other studies, we multiplied the 10-month rate in each year by 1.2, to approximate the rate per 100 000 population per year. The estimated rate for 2006 was 40.6 per 100 000 population per year and the estimated rate for 2007 was 47.1 per 100 000 population per year.

Involuntary admissions as proportion of all admissions, by year

In 2006, the total number of admissions was 432, and 46 (10.6%) were involuntary. In 2007, there were 338 admissions, and 53 (15.7%) were involuntary. Involuntary admissions formed a higher proportion of all admissions in 2007 than in 2006, (χ² = 3.8, P<0.05). Of 46 involuntary admissions in 2006, 6 (13.0%) were readmissions of people who had been discharged from a previous involuntary admission within the year; of 53 involuntary admissions in 2007, 7 (13.2%) were such readmissions. In 2006, two admissions followed less than 1 month after the previous discharge. In 2007, two readmissions followed within 8 days of the previous admission being revoked by tribunal; one came 4 weeks after a pretribunal revocation by the consultant.

Duration of involuntary admissions and duration of detentions

There was no difference between 2006 and 2007 with respect to the mean duration of admissions including an involuntary period (52.1 v. 52.2. days). The mean duration of detention in 2006 (49.5 days, s.d. = 42.0) was longer than in 2007 (41.8 days, s.d. = 41.2), but not significantly so (t = 0.91, P = 0.36).

Periods of voluntary status during involuntary admissions, by year

The mean proportion of each involuntary admission spent actually detained in 2006 was 0.98 (s.d. = 0.07), and in 2007 this proportion was 0.86 (s.d. = 0.21). That is, under the 1945 Act, periods of voluntary status comprised a mean of 2% of the duration of involuntary admissions, whereas under the 2001 Act, this figure was 14%. The difference was significant (t = 3.8, P<0.001).

We determined then whether the difference in duration of voluntary status during involuntary admissions was because individuals under the new Act were more likely to be admitted voluntarily and detained later in the admission, or because they were more likely to have initial involuntary status re-graded to voluntary before discharge. Involuntary patients in 2007 were no more likely to have been voluntary on admission than in 2006 (χ² = 1.3, P = 0.25), but people who were initially involuntary in 2007 were more likely to become voluntary before discharge. A total of 37 involuntary admissions in 2006 were involuntary when the admission began, and of these, 3 (8.1%) were re-graded to voluntary before discharge. In 2007, 47 admissions were involuntary at commencement, of which 21 (44.7%) were re-graded. The difference in the likelihood of regrading was significant (χ² = 11.8, P<0.001).

Principal findings

We hypothesised that there would be fewer involuntary admissions under the Mental Health Act 2001 than under the less rigorous Mental Treatment Act 1945, and that involuntary admissions would be shorter under the new Act than under the old Act. Neither hypothesis was supported. In fact, there were 16% more involuntary admissions to the Jonathan Swift Clinic in 2007 than in 2006. However, this difference was only seven admissions in absolute numbers and the increase was not statistically significant. As such, our findings indicate that the introduction of the Mental Health Act made no difference to the rate of involuntary admissions. We present this interpretation of our findings cautiously: another potential explanation is that there was an actual increase year on year but, because of our small catchment population and the rarity of involuntary admissions, we were underpowered to detect it. In fact, the trend appeared to continue upward after our study period ended. According to the Mental Health Commission, there were 54.8 involuntary admissions per 100 000 population in the Dublin South City catchment in 2008, ⁸ which was 16% higher than the estimated yearly value we calculated for 2007 and 26% higher than the estimated value for 2006.

We also found that the proportion of all admissions that were involuntary was higher in 2007 than in 2006. This could be interpreted as an increased propensity on the part of Jonathan Swift Clinic staff to admit involuntarily rather than voluntarily under the new Act. However, a change in this proportion (involuntary admissions/all admissions) depends as much or more on the number of voluntary admissions as on the number of involuntary admissions. As such, an increase in the proportion of all admissions that are involuntary could be a function of a reduction in voluntary admissions through better out-patient care, rather than reflecting on involuntary admissions at all. Against the latter interpretation in this case is the fact that community supports in Dublin South City did not change appreciably between 2006 and 2007. The duration of involuntary admissions was no different in 2007 than in 2006. However, periods of voluntary status during involuntary admissions were over three times longer under the 2001 Act than under the 1945 Act.

Involuntary admissions under the Mental Health Act 2001

Whether or not there was an increase in involuntary admissions in 2007, there was certainly no decrease, which we expected because of the improved protective measures that are integral to the new Act. In this respect, our results differed from those of Smith & White in Scotland, ^{Reference Smith and White5} and from the findings of the Mental Health Commission itself in its report on activity in 2007. ⁷ Murray and colleagues, in a recent study from the West of Ireland, also found no change in the rate of involuntary admission during the transition to the new Act. ^{Reference Murray, Hallahan and McDonald9} Other authors have found that low rates of compulsory admission require a very specific safeguard: automatic legal representation at the stage of commitment. ^{Reference Dressing and Salize10} In Ireland, automatic legal representation is not mandated until the first review of detention, up to 21 days after involuntary admission comes into force. The timing of legal representation in Ireland is consistent with our finding of higher rates of re-grading to voluntary status after initial detention compared with admissions prior to November 2006.

Although the improved protections in the 2001 Act may have put downward pressure on rates of detention, a number of factors may have had an opposing effect. Additional involuntary admissions under the new Act could have resulted from readmissions after revocation at tribunal. A survey of psychiatrists in Ireland found that 9% of orders revoked by tribunals were followed by immediate readmission, as these orders were deemed to have been revoked on ‘technical grounds’ rather than in the patients' best interests. ^{Reference O'Donoghue and Moran11} Currier reported a similar finding after the introduction of the Mental Health (Compulsory Assessment and Treatment) Act of 1992 in New Zealand. ^{Reference Currier12} In our study, one person was readmitted 1 day after a tribunal revoked an admission order, and one person was re-detained 8 days after revocation. Otherwise, there was little evidence of premature revocation.

Also, it is possible that consultants felt comfortable with the decision to detain an individual under the Mental Health Act because they felt that they would soon have an independent opinion to support that decision, which they did not have under the 1945 Act. Under the 1945 Act, after a reception order was completed, an individual was essentially detained for as long as the consultant, alone, decided, for up to 6 months. Arguably, under the new Act, the automatic review process paradoxically has given consultants greater freedom to detain people, knowing that the decision is less permanent and personal than previously. Against this explanation for increased rates of detention is the dislike stated by many consultants for the ‘adversarial’ tribunal process. ^{Reference Jabbar, Kelly and Casey13} In 2001, Obomanu & Kennedy presciently raised concerns about the negative effects that adversarial mental health tribunals could have on patient care. ^{Reference Obomanu and Kennedy14}

Periods of voluntary status under the Mental Health Act 2001

We found that involuntary patients had longer periods of voluntary status as part of their admissions under the new Act than under the 1945 Act. A plausible reason for this finding is that under the Mental Health Act, automatic review of all involuntary admissions is required within 3 weeks after involuntary status commences rather than 6 months, as was required under the 1945 Act. With this timeline to consider, consultants may have become more proactive in reviewing patients' status. Responses to the survey by O'Donoghue & Moran ^{Reference O'Donoghue and Moran11} suggested that up to 20% of involuntary admissions are revoked with avoidance of mental health tribunals in mind; but widespread premature revocation would likely lead to a pattern of discharge and readmission that we did not find. Indeed, there is reason to believe that re-grading has benefits for the therapeutic alliance that extend beyond the admission. A study in Georgia found that participants who changed status from involuntary to voluntary were more likely to attend for follow-up post-discharge than those who were involuntary throughout their admissions and, surprisingly, than those who were voluntary throughout their admissions. ^{Reference Craw and Compton15}

Limitations and external validity

Limitations of our study include its retrospective nature, depending heavily on availability and correctness of recorded data. Relevant data on seven individuals were obtained from routine records of admission and discharge activities in the unit, as their case notes could not be traced. Additionally, the St James's service has as yet limited access to the kind of community-based treatments envisioned in recent mental health service planning documents. ¹⁶ This limits the extent to which our findings can be generalised to services with more highly developed community-based activity. For instance, the neighbouring Dublin West/Southwest mental health service has a rehabilitation team and several dedicated home-care teams, as well as generic community mental health teams. ⁷ Arguably, rates of involuntary admission might be higher in our service than in those with a greater capacity for community management. Indeed, this may be reflected in the discrepancy between our findings and those of the Mental Health Commission report, although the Mental Health Commission did not provide rates of involuntary admission by service. ⁷ The lack of a less restrictive treatment option is a key factor driving the decision to admit an individual involuntarily. ^{Reference Lorant, Depuydt, Gillain, Guillet and Dubois17}

By January 2007, psychiatrists had only limited training and experience on the practical workings of the Mental Health Act 2001, and this may have affected our findings. It might be expected that ongoing training and experience will result in improvements in the parameters we have evaluated, such as frequency of involuntary admission.

Implications and future research

The reform of the Mental Treatment Act 1945 was overdue. Its replacement, the Mental Health Act 2001, has been founded on the core principles of respect for fundamental human rights and the best interests of the individual. We found that there was a 16% increase in the rate of involuntary admission to our service after implementation of the 2001 Act, but although this increase was not significant, involuntary admissions formed a higher proportion of all admissions under the new Act than under the 1945 Act. There were longer periods of voluntary status during those admissions. This suggests an effective review process. Continuing, prospective research over the next 5-10 years will help determine the longer-term effects of the Act in our centre and elsewhere in Ireland.