When I started working on acute mental health wards in the mid-1990s, the ward doors on my unit were never locked, occasionally nursing staff would be posted on the door if there was a particularly high risk of a patient wanting to leave the ward to hurt themselves, the remainder of the time staff were anxiously vigilant (Bowers et al., 2008). Over time, doors in the UK became locked, driven by incidents, inquiries and policy makers. The doors got broken and they became reinforced with air-locks making wards increasingly secure.

There has long been a suspicion that locking the doors and the imposition of other blanket restrictions on wards, has led to care become increasingly coercive with less attention being paid to the therapeutic milieu. Despite the debate, the evidence for or against the locking of doors is largely weak (Steinert et al., 2019), and like most things associated with acute mental health care, more research is needed as most previous studies were based largely on observational data. For example, see Hubers et al., (2016) which was blogged about by the Mental Elf back in 2016.

The present study by Indregard et al., (2024) is a unique pragmatic, randomised controlled study of the effect of an open-door policy vs locked doors (treatment as usual) on the levels of coercion patients’ experience.

Methods

This was a pragmatic, randomised controlled, non-inferiority trial (based on the hypothesis that opening ward doors would be no worse than having them locked). It compared two wards with an open-door policy to three locked wards (treatment as usual – TAU) in a single psychiatric unit in Norway.

The open-door policy was co-created, and preparatory activities included workshops, introduction of peer-support workers to increase therapeutic dialogue. Doors were open from 9am to 9pm unless locking would ensure safety.

The allocation sequence was a simple binomial list allocating participants to either group in a 2:3 ratio for open-door policy and TAU (respectively). Clearly staff and patients not blinded to the intervention. The authors analysed the data based on intention to treat analysis.

The primary outcome focused on coercive measures which included involuntary medication, isolation or seclusion, and physical and mechanical restraints. Secondary outcome measures included Experience of Coercion Scale (ECS) and Essen Climate Evaluation Scale (EssenCES). See ISRCTN16876467 for registry.

Results

Over approximately one year, 556 patients were randomised to either open-door wards (n=245) or TAU (n=311). Patients were broadly matched in terms of demographics, and about three-quarters of both groups were there involuntarily. About half the patients were diagnosed with psychotic disorders.

• The doors stayed open 73% of the time across the two open-door wards.

The open-door policy was non-inferior (not worse) to treatment as usual (TAU) on all outcomes largely focused on coercion:

• The proportion of patient stays with exposure to coercion was 65 (26.5%) in open-door wards and 104 (33.4%) in the TAU wards (risk difference 6.9%; 95% CI -0.7 to 14.5);
• Reported incidents of violence against staff were 0.15 per patient stay in open-door wards and 0.18 in treatment-as-usual wards;
• There were no suicides during the trial period;
• The median length of stay was significantly shorter in the open-door policy group (16 days; IQR 7–31) than in the TAU wards;
• Patients in open-door wards rated their experience of coercion significantly lower than those in the TAU wards, (mean difference of 0.5 on the ECS (95% CI 0.8 to -0.2; range 0-4));
• Those admitted to open-door wards reported a significantly higher score on therapeutic holding (mean difference 2.4; 95% CI 1.2 to 3.5) and experienced safety (3.5; 95% CI 1.8 to 5.2).

Conclusion

The authors concluded:

The open-door policy could be safely implemented without increased use of coercive measures. Our findings underscore the need for more reliable and relevant randomised trials to investigate how a complex intervention, such as open-door policy, can be efficiently implemented across health-care systems and contexts.

Discussion

According to this study, it would seem that you can open the doors of acute mental health wards without seeing an increase in coercion, but many unanswered questions remain. For example, absconding data (despite being in the original protocol) was not reported, the trial design means that definite conclusions cannot be made, and no serious incidents occurred that would have stopped the trial.

It was interesting to see that this wasn’t solely a trial of the open-door policy, the intervention appeared multi-facetted with a focus on increasing therapeutic dialogue, addition of peer-support workers and was a result of 12 months of input before the doors were even opened. This might make replication tricky. All wards had robust staffing ratios two patients per member of staff during the day and evening, and four patients per member of staff at night, plus there was an additional admission ward and PICU supporting the five trial wards. I suspect that the unit contained more beds per population than a UK context, but in the absence of a robust measure of acuity comparisons remain tricky.

That the authors have been able to undertake a trial in this area is an important milestone, we need more evidence to support the clinical and managerial decisions that are made across mental health services. I do wonder if such a study would be funded in the UK, and whether the required ethical and governance procedures could be agreed. There remains a dearth of evidence about how to provide interventions with patients across inpatient and community services, which are of actual benefit and take account of service designs.

Statement of interest

None.

Links

Primary paper

Indregard A, Nussle H, Hagen M, Vandvik P, Tesli M, Gather J, Kunøe N (2024) Open-door policy versus treatment-as-usual in urban psychiatric inpatient wards: a pragmatic, randomised controlled, non-inferiority trial in Norway. The Lancet Psychiatry, Published: March 06, 2024 DOI:https://doi.org/10.1016/S2215-0366(24)00039-7

Other references

Bowers L, Allan T, Haglund K, Mir-Cochrance E, Nijman H, Simpson A, Van Der Merwe M, (2008) The City 128 extension: locked doors in acute psychiatry, outcome and acceptability. National Co-ordinating Centre for NHS Service Delivery and Organisation R&D (NCCSDO ).

Huber CG, Schneeberger AR, Kowalinski E, Fröhlich D, von Felten S, Walter M, Zinkler M, Beine K, Heinz A, Borgwardt S, Lang UE. (2016) Suicide risk and absconding in psychiatric hospitals with and without open door policies: a 15 year, observational study. Lancet Psychiatry 2016, Published Online July 28, 2016 http://dx.doi.org/10.1016/ S2215-0366(16)30168-7

Steinert, T., Schreiber, L., Metzger, F.G. et al. Offene Türen in psychiatrischen Kliniken. Nervenarzt 90, 680–689 (2019). https://doi.org/10.1007/s00115-019-0738-y