Data Challenges Present an Obstacle on the Road to Population Health
 |
As hospitals wade into the brave new world of population health improvement, they’re finding they need a host of new tools and resources to guide their way. To improve the health of a population, a hospital needs to have deep and rich knowledge about that population, as well as the ability to use that knowledge to effect change. The emerging challenge for hospitals: That knowledge will come from sources they haven’t typically had access to, nor do their current tools, electronic health records, and IT systems support its analysis.
It’s not that the data isn’t available. There are many sources of information about patient and community population, including registries, health information exchanges, databases, mobile applications and analytics tools that process information from electronic health records, clinical information systems and other sources. Even the venerable telephone has become an important data collection tool, as well as a line of defense against readmissions; it is standard process at 95 percent of hospitals to call select patients within 72 hours of discharge. These resources all help hospitals collect pieces to the population health puzzle, but putting the pieces together remains very challenging, especially since the puzzle seems to have no borders.
“We are continuing down the path of being data rich and information poor,” said Stephen A. Martin Jr., PhD, MPH, who is executive director of the Association for Community Health Improvement (ACHI) and former health commissioner for Cook County, Ill. “EHRs have received a lot of investment, but a lot of the key information sources are outside of the hospital. It is hard getting all the data together.”
Central to the challenge is the fact that much of the information that providers need to manage conditions (e.g., type 2 diabetes) is structured, defined and clinical, while much of the information needed to identify at-risk populations and develop outreach strategies is just the opposite. That’s why traditional information systems need to change to support new information sources. Community organizations, family members and the pharmacist at the local drug store can all provide valuable insight into the patient behavior and socioeconomic conditions that affect population health, but there needs to be a way to collect and apply this input. Population health leaders say partnering with individuals and organizations that care for and influence patients outside the hospital is essential.
“If we’re truly going to have a continuum of care then everything needs to be connected so everybody knows what’s going on and we can know how best to care for that patient,” Joan Carroll, RN, BA, CCM, CDMS, director of care transition at Lee Memorial Hospital said during an AHA Solutions Signature Learning Series™ webinar last October. Lee Memorial combined community outreach, analytics and improved patient communication to reduce its readmission rate for certain at-risk populations. It analyzed its readmitted patients and found that “silos of care” — lack of coordination and information sharing among a patient’s health care providers — presented a significant obstacle to optimizing post-discharge care. The analysis led Lee Memorial to develop a new care transition program that emphasizes more interaction with patients and their support systems outside the hospital.
“One of the things I am most excited about is that we are now charting outpatients in [our EHR]” Carroll said. “Now all the physicians in our community can see the notes from home visits.”
Care transition points clearly present great opportunities to improve both individual outcomes and population health. Medicare began paying for transitional care management and communication in 2014 (CPT), and starting on Jan. 1, 2015, will allow an exception to direct physician supervision of chronically ill patients if there is an interoperable electronic health record.1 Companies that partner with hospitals on population health management, such as Healthways and Walgreens, are developing new solutions that focus on care transition in order to reach and motivate patients at this important moment in their care experience. And hospitals are increasingly leveraging health information technology (HIT) systems to identify patient populations that represent a high readmission risk so hospitals can tailor follow-up plans.
Most hospitals that are developing information systems to support population health management are blazing their own trails, using information from disparate sources. Here are some of the resources that are available and examples of how they are being put to use.
Registries — According to the U.S. Department of Health and Human Services (HHS) Office of the National Coordinator for Health Information Technology (ONC), 40 percent of eligible providers have chosen to submit immunization data to registries. To encourage more data sharing and integration, the ONC has awarded grants to more than 1,800 providers to upgrade their EHRs and immunization information systems. Some hospitals are creating their own registries from EHR data and are subscribing to commercial registries.
Other ONC efforts — The ONC is working with public health agencies, the CDC, HIT technology developers and others to create implementation guides (IGs) for electronic transmission of immunization, syndromic surveillance, lab reports and cancer registry data. Health eDecisions (HeD) is an ONC initiative to create standards to facilitate shareable clinical decision support systems and services.
Extending the EHR — The Johns Hopkins Center for Population Health IT (CPHIT) conducts research into how EHR data can be used to improve population health. It created algorithms to search for patterns and risk factors, and some of its work was commercialized into the Johns Hopkins Adjusted Clinical Groups (ACG) Case-Mix System, which has been licensed and is used in 17 countries.
Analyzing EHR data aligns with one of the must-do strategies that Hospitals in Pursuit of Excellence (HPOE) identified for hospitals to transform. HPOE’s March 2014 report, The Second Curve of Population Health, calls for hospitals to have HIT systems that have “…capacity for sophisticated analytics for prospective and predictive modeling to support clinical and business decisions.”
Health information exchanges — HIEs are relatively new and have great promise as data sources for population health. The Institute for Health Technology Transformation notes that HIE data management is challenging because participants have different patient identification methods and data attributes. The IHTT recommends using master identification numbers and linking fields when using HIEs for population health.
Continuity of Care Documents — CCDs provide a standardized way to electronically share health information and can be used with HIEs, EHRs and other systems.
Patient Centered Medical Home (PCMH) Model — Committing to this model aligns hospitals with other providers and gives participants access to new sources of information and insight. Integration and coordination are still required, but the collaborative model provides a foundation for population health initiatives. The Role of Small and Rural Hospitals and Care Systems in Effective Population Health Partnerships, a 2013 report published by HPOE and ACHI, describes how adopting the PCMH model and joining an alliance enabled Yuma Hospital District to identify at-risk populations, and provide services to them directly and through alliance partners, which is expected to reduce readmissions.
These are just a few of the potential data sources and use cases. Using information to improve population health is part art, part science. The art is in identifying sources and forming partnerships to gain the needed data; the science is in developing systems, applications and analytics to make the input useful.
“The key is collecting the right data points. I see this as more of a partnership challenge than an IT problem,” said Martin. “Population health requires partnerships. IT leaders are often brought in too late in the planning process, and the result is that important data points and information sharing are left out of the agreement. HIT leaders should be involved early when partnerships are being planned.”
1 As reported by AIS Health Nov. 24, 2014: http://aishealth.com/archive/rmc112414-01
