Centerstone Research Institute embraces artificial intelligence for clinical decision-making

Finding new ways to reduce the cost of care and improve patient outcomes with Big Data analytics

Published on 31-May-2013

"With our expertise in Big Data management, and IBM’s leading-edge analytics technologies, we’re well positioned to help shift the paradigm for mental health services." - Tom Doub, CEO, Centerstone Research Institute

Customer:
Centerstone Research Institute

Industry:
Healthcare

Deployment country:
United States

Solution:
Big Data, BA - Business Analytics, Business Integration, Business Performance Transformation, Cloud Computing, BA - Predictive Analytics, Small & Medium Business, Smarter Planet, Transforming Business

Smarter Planet:
Smarter Healthcare

Overview

Medical research is one of the fastest-paced and most innovative fields in modern science – but when there’s a wide gap between cutting-edge research and general clinical practice, how can you ensure that patients will actually receive the benefits of new breakthroughs?

Business need:
To advance clinical practice in the mental health sector, Centerstone Research Institute (CRI) wanted to use emergent analytics technologies to bridge the gap between researchers and healthcare providers.

Solution:
CRI is building a national data warehouse that helps mental health organizations share information and develop new techniques – for example, using artificial intelligence to support clinical decision-making.

Results:
42 percent anticipated improvement in patient outcomes by using artificial intelligence to support decision-making. 58 percent anticipated reduction in cost per unit of outcome change – which could help to offset the nation’s rising healthcare costs.

Benefits:
400,000 patient records centralized in a single data warehouse, with capacity for up to 20 million records.

Case Study

Smarter Healthcare

Instrumented
Data for over 400,000 patients is collected from electronic patient records, pharmaceutical prescriptions and clinical systems across the country.

Interconnected
A national data sharing and research collaborative, the Knowledge Network, connects physicians, scientists and policy-makers to the information they need to advance both mental health research and clinical practice.

Intelligent
Groundbreaking research at CRI based on “Big Data” suggests that the use of artificial intelligence in clinical decision-making could significantly improve patient outcomes while reducing cost of patient care.

Medical research is one of the fastest-paced and most innovative fields in modern science – but when there’s a wide gap between cutting-edge research and general clinical practice, how can you ensure that patients will actually receive the benefits of new breakthroughs?

Centerstone Research Institute (CRI), a non-profit organization that works in the mental and behavioral health sector, aims to find new ways to connect research with practice – improving care and reducing costs by getting new insights and techniques out of the research lab and into hospitals and clinics across the country.

Challenges for the mental health sector
Recent studies indicate that patients receive correct diagnoses and treatment less than 50 percent of the time on their first pass through the US health system. Diagnostic accuracy could potentially be improved with new, more effective methods – but there is also evidence that it can take up to 17 years for a new treatment to be widely adopted within the industry. This is partly because the sheer volume of new studies produced each year makes it almost impossible for general practitioners to keep up-to-date with the latest research.1

Casey Bennett, Research Fellow in the Department of Informatics at CRI, explains: “One of the biggest problems we’re seeing is that there’s just too much information for an individual clinician to be able to make the right choice. Take depression. At the moment, there are 20 different possible medications to choose from and almost no clear way to predict which one will work for a specific patient. So clinicians are forced to make an educated guess. If it works, great; if not, try another one. But that’s expensive and not really the best way to proceed.”

With healthcare costs in the US likely to reach 30 percent of GDP by 2050, the need to bridge the gap between medical science and patient service has never been more acute.2

Building the Knowledge Network
Tom Doub, CEO of CRI, comments: “Technology, and particularly analytics technology, is absolutely key to transforming our mental health system. In partnership with mental health organizations across the country, we are building the Knowledge Network – a learning collaborative that is contributing to a cloud-based data warehouse, which will allow us to explore every aspect of mental health treatment on a national scale. This will help us and our partners play a vital role in advising government and shaping mental health policy and research in the years to come.”

The Knowledge Network will ultimately scale to encompass thousands of healthcare providers, including hospitals and local healthcare providers across the US, and will potentially store data on up to 20 million patients. At present, it contains 400,000 electronic patient records, 20 million service records, 3 million medication prescriptions, as well as a large fund of information on patient outcomes.

As an example to illustrate the data challenges involved, each electronic patient record represents a complex plethora of related information assets: patient medical history including physician notes, service records, medications, insurance coverage and so forth. Although not all related files are physically stored within the database, with 70,000 unique patients seen annually, the complexity and volume increase exponentially every year.

CRI’s work on data warehousing was recognized by a TDWI Best Practices Award in 2010.3

Importance of Big Data
April Bragg, VP for Research Advancement, comments: “Centerstone serves 70,000 patients each year, but that’s across the whole spectrum of mental and behavioral health. If you’re researching a very specific diagnosis, there might only be a few hundred people in our dataset who suffer from it. This is why the scale of the Knowledge Network is so important: for the first time, we have enough data to extract reliable, actionable insights about the state of mental health care across the country.”

As well as improving insight on a national level, the Knowledge Network also enables research to be focused on specific areas and communities, and to compare differences between them. The ability to analyze localized, community-based information is one of the keys to Centerstone’s success with patients in its own network of more than 130 non-profit mental health locations in Indiana and Tennessee.

Casey Bennett, Research Fellow in the Department of Informatics at CRI, says: “A lot of people are doing what they call clinical decision support, but it’s often based on embedded, hard-coded rules from 10 years ago,” Bennett says. “Those rules are not personalized to individuals. They’re based on what works for 60 percent of people, based on average populations. That’s not the model we wanted.”

Several years ago, CRI began using IBM predictive analytics technologies to create sophisticated predictive models to assess the effectiveness of various treatment options for its own patients. These models analyzed 14 variables (selected from hundreds of potential variables) across more than 9,000 patients, including socioeconomic status, demographic information, and a range of diagnostic and clinical data. Researchers also incorporated multiple sets of outcome measures from internal electronic health records, as well as statewide outcome measures. The end result was a set of predictions about the effectiveness of various treatment options for each patient based on his or her unique characteristics.

The use of these models helped to yield a 50 percent improvement in choosing the correct course of treatment during a pilot of the technology. This was the first step in helping Centerstone’s network of mental health clinics incorporate a “practice-based evidence” approach to treating patients, allowing for highly personalized care plans based on actual patient experiences with medications and modalities. Other related predictive modeling efforts provided an opportunity to streamline overall operating costs, improve clinical productivity and help clinics weather fluctuations in federal funding and insurance reimbursements.

Casey Bennett comments: “Our initial pilot project was very encouraging, and we wanted to take the next step. Instead of just modeling each patient’s data at a single point in time, we wanted to see if we could assess the patient over the course of their treatment, and recommend specific courses of action at the appropriate stages – just as a human doctor would do.”

Artificial intelligence in clinical decision support
In partnership with Indiana University, CRI used a sample set of 6,700 patients’ electronic health records from the data warehouse – all with co-occurring physical and mental illness – to develop a non-disease-specific artificial intelligence (AI) framework that could learn from clinical data and simulate decision-making at multiple stages during a course of treatment.

“Effectively, we were aiming to create an AI that could ‘think like a doctor’, and analyze the possible alternative treatment paths to make the right clinical choices for each individual patient,” says Casey Bennett. “With the help of IBM predictive analytics software, we built a framework that is able to determine optimal actions based on what has been observed so far, the beliefs about the current situation based on those observations, and what we expect might happen in the future. In other words, the system can maintain beliefs about the patient’s health status to deal with uncertainty or missing observations, and continually plan/re-plan over time as the information changes.”

AI outperforms traditional models
The results were impressive: compared to the standard “treatment as usual” model of healthcare provision, the AI framework provided a potential 42 percent improvement in patient outcomes, and a 58 percent savings in the cost per unit of outcome change.

“We’re not suggesting that our AI could – or should – replace a human physician, but the results suggest that it could be a very valuable tool for decision support,” says Casey Bennett.

April Bragg adds: “Ultimately, these kinds of analytics technologies could help us transform the way we provide mental health services by bridging the gap between researchers and physicians. If our AI models are constantly integrating and learning from new data in the Knowledge Network – whether it’s generated by researchers or supplied by patients – they can communicate new insights to doctors across the country, and help to accelerate the advancement of clinical practice.”

Finding a new way forward for mental health
Tom Doub concludes: “In the field of mental health, we have seen a slowing of innovative new treatments. Cognitive behavioral therapy, still the state of the art treatment for depression, was developed in the 1970s. Progress in pharmaceuticals has also slowed down in recent years. At the same time, the financial and regulatory pressures on the sector are increasing. Something has to change – and we believe that technology will be the main agent of that change.

“For example, mobile technology has the potential to help patients monitor their own conditions on a daily basis and send new data to their physicians without needing to visit a clinic. The ability to gather and analyze data every day rather than weekly or monthly could completely transform our understanding of mental health in the community. With our expertise in Big Data management, and IBM’s leading-edge analytics technologies, we’re well positioned to help shift the paradigm for mental health services.”

About IBM Business Analytics
IBM Business Analytics software delivers data-driven insights that help organizations work smarter and outperform their peers. This comprehensive portfolio includes solutions for business intelligence, predictive analytics and decision management, performance management, and risk management.

Business Analytics solutions enable companies to identify and visualize trends and patterns in areas, such as customer analytics, that can have a profound effect on business performance. They can compare scenarios, anticipate potential threats and opportunities, better plan, budget and forecast resources, balance risks against expected returns and work to meet regulatory requirements. By making analytics widely available, organizations can align tactical and strategic decision-making to achieve business goals.

For more information
For further information or to reach a representative please visit ibm.com/business-analytics.

Products and services used

IBM products and services that were used in this case study.

Software:
SPSS Statistics, SPSS Modeler

Footnotes and legal information

1. Bennett CC, Hauser K. Artificial intelligence framework for simulating clinical decision-making: A Markov decision process approach. Artif Intell Med (2013), 57(1): 9-19. http://dx.doi.org/10.1016/j.artmed.2012.12.003 2. Ibid. 3. http://tdwi.org/Articles/2010/08/16/TDWI-Best-Practices-Awards-Winners-2010-Summaries.aspx

© Copyright IBM Corporation 2013. IBM Corporation, Software Group, Route 100, Somers, NY 10589. Produced in the United States. May 2013. IBM, the IBM logo, ibm.com, and SPSS are trademarks of International Business Machines Corp., registered in many jurisdictions worldwide. Other product and service names might be trademarks of IBM or other companies. A current list of IBM trademarks is available on the web at: www.ibm.com/legal/copytrade.shtml. This document is current as of the initial date of publication and may be changed by IBM at any time. Not all offerings are available in every country in which IBM operates. The client examples cited are presented for illustrative purposes only. Actual performance results may vary depending on specific configurations and operating conditions. THE INFORMATION IN THIS DOCUMENT IS PROVIDED “AS IS” WITHOUT ANY WARRANTY, EXPRESS OR IMPLIED, INCLUDING WITHOUT ANY WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND ANY WARRANTY OR CONDITION OF NON-INFRINGEMENT. IBM products are warranted according to the terms and conditions of the agreements under which they are provided. The client is responsible for ensuring compliance with laws and regulations applicable to it. IBM does not provide legal advice or represent or warrant that its services or products will ensure that the client is in compliance with any law or regulation.


File Attachment Icon
Centerstone_YTC03607_approval.pdf