Are multidisciplinary teams an effective tool to reduce emergency hospital activity?

With the second wave of coronavirus (COVID-19), it’s easy to focus on urgent needs and forget long-term plans in the NHS, such as expanding integrated working across health and social care. Community multidisciplinary teams (MDTs) bring together different health and care professionals to discuss and plan the care of individual patients. Their aim is to provide more proactive, personalised and holistic care and thereby improve the timeliness, flexibility and suitability of people’s care – as well as, in the process, reducing preventable hospital admissions.

The evidence about the impact of MDTs indicates that, although they may improve patients’ satisfaction and experience of care, they may not reduce – and may even increase emergency hospital use. This latest finding seems counterintuitive, given the experience of patients, and has sparked seminars and publications on the topic.

More evidence that multidisciplinary teams may increase emergency hospital activity, at least in the short term

The Extensive Care Service (ECS) and Enhanced Primary Care (EPC) are two complimentary MDT models in the Fylde Coast NHS vanguard. Both use risk stratification to help identify adults with complex chronic care needs who are at risk of hospitalisation.

ECS, modelled on the US CareMore model, is aimed at the highest risk patients, who are 60 years or older and have two or more specific long-term conditions. This model temporarily replaces usual GP care with a dedicated health care team led by a consultant extensivist.

EPC targets a somewhat lower risk group, who are 16 years or older and who could benefit from increased multidisciplinary support because of a long-term condition or other factors such as mental ill health and/or difficult social circumstances.

The Improvement Analytics Unit (IAU), an innovative partnership between the Health Foundation and NHS England and NHS Improvement, found that neither model of care reduced emergency hospital use compared with their carefully selected comparison groups during the first years of the intervention.

In fact, we found that enrolled patients had higher rates of emergency hospital use: for example, ECS patients were admitted to hospital in an emergency 27% more often than their comparison groups and EPC patients 42% more often. However, this may be due to differences between MDT and comparison patients that we couldn’t account for in our analysis, so we weren’t able to conclude if this higher rate of emergency hospital use was due to the MDT interventions or not.

Why might patients enrolled in MDTs have higher emergency hospital use in the short term?

We think there could be several explanations:

1. Implementing changes takes time. Establishing good cross-professional relationships and new ways of working may require more time than the 33 months and 18 months (respectively) covered in our evaluations of ECS and EPC.
2. MDTs may initially identify unmet need that might otherwise not have been identified or addressed during the period of the studies. Early diagnosis may allow patients to avoid more serious health problems later, but if the study period isn’t long enough, we might not pick up on these longer term benefits.
3. MDTs typically target high-risk, high-need individuals. But there may be limited scope to reduce hospital use for these patients. It may be easier to improve the health outcomes of patients who are less acutely ill.
4. The MDT models that we evaluated supported patients with a wide range of conditions. Yet different patient groups may benefit differently from MDTs, with these group effects hidden when looking at the overall effect. For example, there may be conditions where input from several different professionals is needed and there is particular benefit from collaborative working.
5. Although we used robust statistical methods to compare MDT patients with a comparison group of non-MDT patients with similar characteristics, there may still be unobserved differences between the groups. These could account for some or all of the difference in outcomes (unobserved confounding).
6. Hospital use may not be the best measure of the benefits of MDTs. Their value may lie in improving patients’ experience of care or quality of life, but these are difficult to evaluate robustly as they are not routinely recorded.

MDT effectiveness is more relevant than ever, but further analysis is needed

MDTs feature prominently in the NHS long term plan and are integral to the rationale for how primary care networks can deliver better care and relieve pressure on hospitals and social care. There is also emerging evidence that MDTs are playing a crucial role in local areas' response to the pandemic, for example in terms of care home support or broader community activity.

It is important to understand why MDTs are not having the anticipated effect on emergency hospital admissions. For example, should MDTs target different patient groups? Should policymakers have more realistic expectations of what MDTs can achieve in the short term? When evaluating MDTs, do we need to allow more time to capture their full effect? Should health care systems routinely collect data to better reflect the outcomes which matter most to patients?

Clearly, further research is needed to understand and maximise the benefits of MDTs. Within the IAU, we believe there is more that we can do with the quantitative data we already have available, so we are planning further analyses in the coming months. Watch this space.

Therese Lloyd (@ThereseTHF) is a Senior Statistician in the Improvement Analytics Unit