CAP NewsDoing Good with Good O.R. competition
Lisa Maillart and Itai Ashlagi, co-chairs of INFORMS’ 2015 Doing Good with Good O.R. student paper competition, named the following six finalists for this year’s competition:Read More
Industry NewsFrontline Solvers V2015-R2 offers Excel users access to big data, advanced analytics, data visualization and sharing on the Web
Frontline Systems is shipping a new product line release, Version 2015-R2 of its Solvers for Excel, including its flagship integrated product, Analytic Solver Platform, and its “professional entry-level” integrated product, Analytic Solver Pro.Read More
Special ArticlesStudy incorporates predictability into airport ground delay programs
For air traffic managers coping with delays due to adverse weather, it’s good to know early how late a flight will be. Ground delay program (GDP) decisions that recognize this are superior to those that do not, substantially reducing the cost of GDPs to flight operators. In a case study based on San Francisco International Airport, researchers reduced the cost of delays by 13 percent by incorporating predictability into the GDP.Read More
Health Care & Analytics
From ‘sick’ care to ‘health’ care, courtesy of analytics
| || || |
By (Left to right) Rajib Ghosh, Atanu Basu and Abhijit Bhaduri
“In health there is freedom. Health is the first of all liberties.”
- Henri Frederic Amiel
Access to affordable health care should be a basic right of all citizens. Health care refers to everything from diagnosis to treatment of physical and mental injury, including prevention of disease. The value chain of stakeholders in health care is a complex one involving multiple players from producers at one end to purchasers and providers who need to go through fiscal intermediaries, such as HMOs and insurers, to reach the payers. Health care contributes $4 trillion to the global economy. While this article focuses on the U.S. health care system, many of the recommendations can also be broadly applicable to health care systems in other countries.
About 10 years ago, a Wharton School Study defined the U.S. health care value chain as shown in Figure 1.
Figure 1: U.S. health care value chain.
However, during the last few decades the U.S. health care system went through some transformations both vertically and horizontally. Provider organizations, such as hospitals or hospital systems, vertically integrated with fiscal intermediaries to create what is known as Integrated Delivery Networks (IDNs) – a private sector example is Kaiser Permanente. The Veterans Health Administration (VHA) is a similar example in the government sector. IDN organizations typically also include downstream care delivery entities such as physician practices. During this time, consolidation also occurred horizontally at different points of the value chain with the merger of hospitals to create health systems, the merger of physician practices to create scaled-up multi-specialty practices, and the merger of drug and/or device manufacturers to improve R&D cost and outcome (see Figure 2).
Figure 2: Providers merge with fiscal intermediaries to create IDNs.
For health care to work in any society and be inclusive, two elements need to be addressed – quality and affordability. In all the shifts the industry has seen, the U.S. consumer has been left out. Health care costs will be about 19 percent of GDP by 2019 , and more than 32 million Americans will enter the U.S. health care system during the next few years. In a September 2011 HBR article, Robert S. Kaplan and Michael E. Porter of Harvard Business School present compelling arguments on how the current system of measuring costs in health care is fundamentally broken. On top of the cost crisis, a shortage of 150,000 physicians is expected within 15 years at the current rates of graduating and retiring physicians, according to the Association of American Medical Colleges.
Following are some of the opportunities we see (not a comprehensive list, of course) for applied analytics in the health care sector.
From Sick Care to Predictive Care
The focus of care needs to shift from the current approach of “sick care” to “predictive and preemptive” care that will be delivered across the health care value chain. This is a major opportunity to address the quality as well as the cost element of the health care equation. A study by RAND Corporation projected that 48.3 percent of the U.S. population will have one or more chronic conditions by 2020 . This will continue to be a huge burden on health care resources, including cost.
The current U.S. model is based on a “fee for service” principle that encourages higher utilization, but it does not encourage better quality of health management. Health reimbursement plans in United States today are predicated on activity and not outcome. The payment models for the future need to revisit the delivery architecture, examine who they will deliver this care to and at what time, and also determine how much care is needed. For instance, in chronic conditions the cost increases with the worsening of the condition. If such disease/cost flare-ups can be preempted or minimized, the cost of chronic care patients can be controlled.
Analytics – a discipline that applies mathematical sciences to data in order to answer what, why, when and how questions – can lower costs and bring predictability to payers and fiscal intermediaries, who in turn can share the benefits with providers in a “fee for performance” model. The predictive care model will give rise to higher drug adherence and device utilization in home care setting, thus making it lucrative for device and drug manufacturers. Analytics can address the quality as well as the cost aspect for not just patients but also the other stakeholders in the health care value chain.
Clinical Decision Support Systems
Health care related data is increasingly being used to manage patients. New models of health care delivery will make it easy to collect specific data for each patient at different points of the delivery network. Once genomic data becomes commonplace, the data availability will explode. Medical literature is now doubling every seven years. The clinician will be challenged to analyze these huge volumes of data and medical insights before making decisions regarding diagnosis. Analytics has a major role to play here by providing intelligent decision support. Twenty percent of medical errors in the United States today are caused by diagnostic errors, which also include delayed diagnosis. A robust clinical decision support system, powered by a strong analytical engine, can digest many sources of data and convert them into meaningful information for clinicians in real time (even possibly in a predictive manner). Such a tool can be great aide for clinicians in a world where demand for health care is far outstripping the supply.
Population Health management
A study, conducted in London between 1967 and 1977, found that social status of people affects their health conditions. Another famous study, conducted by Nicholas Christakis and James Fowler, found that in a population that is connected in social network, the network impacts health of individuals in the population. Hence what used to be known as disease management now includes distribution of the overall health outcomes within the entire population under the purview of a care provider organization. The scope has been broadened because there is a strong relationship between socioeconomic status and health.
A key objective of population health management is to reduce inequities among population groups. Therefore, population health management needs to take a holistic view of the population health with various parameters and various data sources. Identifying the root causes for a disease that an individual has will have to go beyond the traditional symptomatic approach toward medicine. To achieve success in population health management at a scale, strong analytics support is essential. Analytics can help find patterns in the data to ensure that the benefits of positive health outcomes get replicated over a large geographical area housing the population.
Remote Health Monitoring and Telehealth
Telehealth allows patients to stay at home by managing their health conditions through networked equipment. This modality of care delivery allows health care professionals to monitor vital signs and manage health risk assessments without being present physically in that location. Telehealth also aims to empower patients to manage their own health conditions through appropriate and timely feedback and education. Analytics expands the reach of such value-added services.
Telehealth, as it is designed today, can reach only a small segment of the population. Until we find a way to scale the reach of telehealth, it will not achieve economies of scale – and thus not lower costs substantially. Yet, telehealth is an extremely important component of the future health care delivery model since the health care safety nets in U.S., funded by Medicare and Medicaid, are dangerously close to bankruptcy in some years. It is a huge opportunity to help people whose health conditions require specialized health care. Telehealth for these patients is a mechanism to avoid repeated hospitalization – especially to ERs that significantly drives up cost.
The fitness industry has already adopted telehealth, albeit in a different form driven by “worried well” consumers. It is also called “mHealth” or mobile health. A segment of such consumers in the United States has adopted health monitoring to receive warnings and/or health related feedback, without direct involvement of any health care professional. Device manufactures have captured the imagination of this segment, and we are at the cusp of an exponential growth in “personal health monitoring” devices in the market. Such an early adopter population segment can create a new wave of “prevention and wellness” in the United States – which in the longer term can reduce sudden flare-ups in diseases.
It is, however, difficult for the care provider and/or the payer organization to identify the group of patients who will respond to telehealth and provide best return on investment. Such a decision can be facilitated by analytics since it involves making sense of large volumes of data. For example, it may involve combining patient data from insurance claims, lab results, hospital information systems, pharmacies, etc. to derive a target population using sophisticated analytical algorithms. Analytics can also predict a possible flare up based on the data collected, while consciously minimizing false positives and false negatives. This can help minimize the number of people for whom a visit to the ER is needed. Analytics can help effectively address the quality and the scale priorities of telehealth.
Making Accountable Care Organizations Work
In 2010, the 111th United States Congress passed the Patient Protection and Affordable Care Act (PPACA), which was made into a law by President Obama. This was part of the health care reform agenda designed to improve access to health care for all Americans. As part of the entire initiative, a distributed delivery architecture known as the Accountable Care Organization (ACO) was proposed. The concept is similar to HMOs, but it additionally puts emphasis on population health management and improving the quality of care without increasing cost. By 2012, a large of number of hospital systems, payers, physician practices and pharmacies are expected to join forces to create their own ACO.
Such a distributed architecture works well when patient data can be shared across a broader group of stakeholders to improve outcome and quality of care. Federal financial incentives are encouraging the use of Electronic Health Records (EHRs) and expect full adoption by 2015. This will drive the discussion on data interoperability and build traction for Health Information Exchanges (HIEs).
While data hub architecture is available in many other countries such as Singapore, the United Kingdom, Canada and Spain, its adoption has been sporadic in United States. The recent public-private partnerships are helping to shape this move in the United States. It is easy to see the increased use of analytics in health care in the visible future. In a connected network where individual electronic health records and other health care information systems share data, analytics can be used to generate much more specific information for the use of the physician. In particular, analytics can help answer :
- What treatment regimen to use for a patient population with a certain disease pattern and genetic profile (assuming genetic data will also become part of the HIE at some point)?
- What are the risk factors that need to be managed for a patient and what kind of drugs can be used considering different drug constituents and interactions?
- What protocol produces the best results during rehabilitation for a target population?
- Which therapies or combination can produce best results for patients undergoing certain procedures?
The Future of Health Care
We are optimists – we are energized to see that the health care industry is finally realizing the power of data and analytics to transform health care as we know it. A May 2011 McKinsey report estimated that “big data” – referring to the growing volumes of data (pharmaceutical data, claims data, clinical data, and patient data), lacking integration, in possession of different stakeholders in the health care industry – and analytics can generate $300 billion in annual value to U.S. health care.
Personalized medicine is the next frontier of health care. If the cost of the human genome analysis drops to about $1,000 in the not so distant future , it will trigger increased application of analytics in health care.
The future of health care lies in being able to address two elements. One is lowering the cost of health care with a new care delivery model that, we hope, would be much more focused on outcomes than activities. This is dependent on building better economies of scale – transparency with respect to outcomes and cost should help providers become more competitive and more effective, and help patients select the best available care. The other element lies in improving the quality of health care. Analytics can make personalize medicine a reality by intelligently combining genomics, imaging, pharmaceutical and diagnostic data. Data sharing and interoperability will play a significant role in making this all happen, and different governments will address the policy issues in their own ways. Yet what remains clear is health care will be driven more by data in the future than just by the knowledge of medicine.
Rajib Ghosh (email@example.com) is the director of Global Product Management for Robert Bosch Healthcare, based in Palo Alto, Calif. Ghosh has 18 years of experience in technology business including almost a decade in the medical device industry. At Robert Bosch Healthcare, he is focused on creating product strategy for the future in telehealth, remote patient monitoring, and chronic disease management.
Atanu Basu (firstname.lastname@example.org) is the CEO of DataInfoCom, an analytics software company headquartered in Austin, Texas. DataInfoCom’s customers include Dell, Microsoft, Cisco Systems and Juniper Networks. Basu has 17 years of experience in the semiconductor and the software industry, almost all of it in new technology development and sales. He holds a master’s degree in engineering.
Abhijit Bhaduri (email@example.com) is the chief learning officer for Wipro Ltd. He runs a popular blog at http://abhijitbhaduri.com. Prior to Wipro, Bhaduri led global teams at Microsoft, PepsiCo, Colgate and Tata Steel and has worked in India, Southeast Asia and the United States.
- A. M Sisko, C. Truffer et. al., “National Health Spending Projections: The Estimated Impact of Reform Through 2019,” Health Affairs, Vol. 30, No. 4, April 2011.
- Wu, Shin Yi, and Green, Anthony, “Projection of Chronic Illness Prevalence and Cost Inflation,” RAND Corporation.
- SAS whitepaper, “Analytics in Healthcare.”
- Kevin Davies, “The $1,000 Genome: The Revolution in DNA sequencing and the new Era of Personalized Medicine,” Free Press, N.Y.