Joint crisis planning in mental health care: the challenge of implementation in randomized trials and in routine care

Joint crisis planning produces a plan for use during a future mental health crisis or relapse. Its distinguishing feature is facilitation by a mental health professional external to the treatment team, who engages a mental health service user and members of his/her treatment team in a process of shared decision making.

To date, there have been three trials of joint crisis plans, producing two key findings. First, the process of producing and using a joint crisis plan is highly appreciated by service users, can improve therapeutic relationships and reduce the rate of involuntary measures, and is likely to be cost-effective. Second, joint crisis plans are challenging to produce and use, exemplifying the widespread difficulty within medicine of adopting shared decision making.

In this context, the aim of this paper is to consider whether repeated emphasis on individualized crisis planning in policy documents will be sufficient to bring about the adoption of shared decision making in mental health care. Experience from the above-mentioned three trials provides indications of what other measures may help.

HOW IS JOINT CRISIS PLANNING DIFFERENT FROM ROUTINE CARE?

Written treatment plans are routine in many community mental health services and many contain an action plan for crisis or relapse. Their chief goal is to ensure timely, co-ordinated and effective care.

In England, the Care Programme Approach (CPA,(1)) provides a framework for care of the most vulnerable mental health service users, including those at risk for suicide and self-harm and people with a history of relapses requiring urgent intervention. Further guidance (2,3) has re-emphasized the need to undertake detailed crisis planning, and the Mental Capacity Act 2005 (4) provides for advance refusals of treatment in a crisis.

In the U.S., supporting people to create a psychiatric advance directive is viewed as a component of recovery-oriented treatment planning (5). Psychiatric advance directives promote consumer choice and prioritize the goal of autonomy.

Routine treatment plans lie at the other, more paternalistic, end of the crisis planning spectrum, as they may be produced without service user involvement, although by consensus this is not seen as good practice. Most routine crisis plans in England remain stubbornly “one size fits all” (6). Within the National Health Service organizations participating in the CRIMSON multisite randomized controlled trial of joint crisis plans (7), at baseline only 15% of participants had a crisis plan containing any information specific to that individual (6). The inference is that most community mental health teams do not consider individualized crisis plans a priority.

Joint crisis planning lies toward the centre of the above spectrum, as an application of the shared decision making model (8,9). To achieve this, it employs an external facilitator to complete the crisis plan, instead of the service user’s care co-ordinator or case manager. The facilitator aims to engage the service user and treating mental health professionals during formulation of the joint crisis plan. Developed after consultation with service user groups (10), this process aims to empower service users whilst facilitating early detection and treatment of relapse. Held by the service user, a joint crisis plan contains his/her treatment preferences for any future psychiatric emergency using first person language.

WHAT DIFFERENCE DOES THE JOINT CRISIS PLANNING PROCESS MAKE IN COMPARISON TO ROUTINE TREATMENT PLANNING?

Results published in 2004 of a single site randomized controlled trial of joint crisis plans for people with psychotic or bipolar illness showed reduced rate of involuntary hospitalization associated with their use (11) and generally positive views of the plan among service users and mental health professionals (12). Similarly, in 2006, a U.S. study of facilitated psychiatric advance directives showed an improvement in working alliance at one month (13). A more recent randomized controlled trial in the Netherlands found that crisis planning was associated with a reduction in court-ordered admission to hospital (14), but not other forms of involuntary admission. However, this intervention did not involve an external facilitator.

The CRIMSON multisite trial (N=569) sought to provide definitive evidence on the effectiveness of joint crisis plans delivered in routine practice (7). No significant treatment effect was seen for the primary outcome of involuntary hospitalization or secondary outcomes of overall psychiatric hospital admissions, length of stay, perceived coercion and engagement with services. However, there was a positive effect on service user-rated therapeutic relationships, consistent with the 2004 trial (11) and the trial of facilitated psychiatric advance directives (13). Qualitative trial data (15) supported the improvement in therapeutic relationships when clinicians engaged well in the discussion. Service users reported that the facilitator helped to address power imbalances and that clinicians listened more and were more reasonable.

However, lack of engagement amongst some clinicians may have undermined the potential effect of planning (for instance, psychiatrists’ lack of attendance or engagement at the planning meeting, or lack of awareness of the joint crisis plan on the part of subsequent clinicians following staff turnover) (16). Moreover, while some clinicians believed the external facilitator was necessary for empowering service users, others feared potential interference. Finally, many clinicians believed that they already engaged in joint crisis planning, or that crisis planning was a bureaucratic exercise of little value due to lack of service user choice.

While the main outcomes from CRIMSON might support some of these views, other evidence from this trial does not. Contrary to the assertion that the joint crisis plan adds little to routine practice, an audit of routine crisis plans of the trial participants showed that individualization was infrequent (6). Further, content analysis of the joint crisis plans showed a wide range of service user choices, that were on the whole clinically reasonable, including efforts to self-manage early warning signs of relapse and some requests for hospitalization (17). Finally, while clinicians endorsed shared decision making approaches and believed that they were enacting it in routine care, reports from service users contradicted this view (15). It seems that more needs to be done to convince clinicians of the potential benefits of the approach.

ARE JOINT CRISIS PLANS RELEVANT AND HELPFUL FOR SERVICE USER GROUPS OTHER THAN THOSE WITH PSYCHOSIS?

The single site JOSHUA randomized controlled trial (18) was set up to develop and provide a preliminary test of the effectiveness of joint crisis plans for people with borderline personality disorder, who are especially vulnerable to the experience of crises and their adverse consequences, particularly in terms of self-harm. Again, participants’ views were generally strongly positive: joint crisis plans were used both during (74%) and between (44%) crises, and approximately half of intervention participants reported experiencing a greater sense of control over their mental health problems and an improved relationship with their mental health team at follow-up (19).

Nevertheless, the trial failed to demonstrate superiority for the primary outcome, self-reported self-harm, and also for all secondary outcomes. This was despite an excellent level and rate of joint crisis plan production, although subsequent problems in adherence to the contents may have reduced its effectiveness. For this trial, the production process excluded treating psychiatrists as a response to service user preference. The trial under-recruited, thus, the absence of positive significant findings in favour of joint crisis plans may partly have been explained by type II error.

THE ECONOMICS OF CRISIS PLANNING

The provision of facilitators to ensure high quality crisis planning may appear prohibitively costly. However, the 2004 trial of joint crisis plans showed that their use was cost-effective relative to the control condition (available non-individualized treatment information plus routine care planning) (20). Likewise, the JOSHUA randomized controlled trial showed that there was at least an 80% probability that the joint crisis plan plus treatment as usual was more cost-effective than treatment as usual (19).

The economic evaluation of CRIMSON (21) showed no evidence for the total sample of cost-effectiveness of the joint crisis plan. However, analysis by ethnic subgroup showed there is at least a 90% probability of the joint crisis plan intervention being more cost-effective than treatment as usual in the Black ethnic group.

CONCLUSIONS

Joint crisis plans may be cost-effective for Black people with psychotic or bipolar illness (21) and people with borderline personality disorder (19), two groups for whom mental health services have tended to provide the least satisfactory care. This suggests that any future study of joint crisis plans should target service users whom the clinical team are particularly struggling to engage in collaborative working.

In England, those who are poorly engaged with services are likely to be subject to a paternalistic approach in the form of a community treatment order. This has not been shown to be effective in reducing involuntary admissions or any other outcomes (22,23). Interventions such as the joint crisis plan are welcomed by service users when the clinical team engages with the process, and this may improve therapeutic alliance (12,13,15). However, although many clinicians endorse the general idea of shared decision making (24), the variability of adoption reflects a mixed response to this method of operationalizing it.

To date, external facilitation has not been adopted in the UK. However, without the facilitator, the application of shared decision making to crisis planning is likely to continue to be variable. One way to resolve this dilemma would be to train care co-ordinators/case managers to provide external facilitation for other teams as part of a reciprocal arrangement among teams, thus adding to their own skills in encouraging shared decision making. Whether addressing the barrier to adoption in this way leads to positive outcomes in routine care remains to be seen.