The power of analytics in mental health from operational efficiency to suicide risk

Dr. G. Oana Costea is Clinical Associate Professor of Psychiatry and Human Behavior at Brown University and Unit Chief and Psychiatrist at Lifespan/Bradley Hospital in Providence, RI. She also is a member of the MIT EMBA class of 2018.

There is a significant role for analytics in driving the drastic changes needed in healthcare to address the significant waste and astronomical costs prevalent in the industry. While data science has begun to make an impact in hospital operational efficiency and other aspects of healthcare delivery, the industry would benefit from wide-spread continued focus on analytics.

There are several reasons for its slow start. First, many healthcare organizations struggle across the entire analytics value chain. Due to limited infrastructure suitable for data collection, the capture of raw data and storage is an issue. However, this has improved with the implementation of electronic health records. Other obstacles for analytics include access to data due to privacy issues and the overall absence of an analytics culture of being data driven and seeing data as a source of value.

Due to these challenges, the main role of data in healthcare today is in the areas of reporting, regulatory compliance and governance rather than decision making.

The course work at MIT Sloan, from Data, Models and Decisions to Healthcare Lab and the Analytics Edge highlighted the benefits of analytics in healthcare in terms of providing key insights to optimize decision making and to improve organizational performance. Furthermore, the Analytics Edge course allowed the opportunity to use analytics methods and experience their impact in answering critical questions faced by the industry.

For example, the hospital length of stay(LOS) is one of the critical metrics in healthcare as it is linked to healthcare cost and operational efficiency, including resource utilization and bed availability. This is especially critical in Child Psychiatry due to a significant mismatch between the high demand for services and the limited supply. There is evidence that the limited resources are not being used in the most optimum fashion, which accentuates the supply/demand mismatch. Predictive models for hospital LOS have the potential to greatly increase operational efficiency by allowing adequate resource allocation and planning based on demand for services and patient needs. This in turn decreases healthcare cost both directly by eliminating waste and indirectly by increasing quality of care and therefore decreasing the risk of re-admission.

Analytics tools also can be used to predict re-admission rate, which is a useful metric both from the patient and the healthcare facility perspective. For patients, predicting re-admission rate could assist in determining illness prognosis and drivers of such prognosis, therefore highlighting possible interventions to shift the illness course. For healthcare facilities, re-admission rate could help with resource planning and population health management.

In the mental health space, research has emerged on the development of predictive models to estimatesuicide risk. Detecting individuals at increased suicide risk is a major clinical challenge due to the complexity of the phenomenon and often the absence of a disclosure of suicidal thoughts or intent by those that go on to complete suicide. Predictive models could play a major role in identifying high risk patients that could not be detected solely based on clinical assessment. They could also allow for an earlier detection of patients at risk and therefore adequate interventions. In the context of rising suicide rates in the recent years among the youth and the shortage of services and providers specialized in Child Psychiatry, artificial intelligence toolscould become the linchpin in managing this major public health issue.

The bottom line is that analytics is starting to make improvements in healthcare, but there is still a lot of work to be done. There is a need to embrace analytics as part of the culture in healthcare and make it more of a norm across the industry to overcome the challenges and realize its benefits.

How can you use analytics to improve decision making in your industry?