DRAFT

Electronic Health Record Exchanges to Manage Healthcare Super Utilizers

1. Introduction

Around the United States there are a number of programs which have exploited the use of electronic health data and other technologies in a quest to better manage the health care of poor, uninsured or underinsured patients who are super utilizers of health care services. These programs are heeding the call to develop new models of care that keep patients healthier and out of high cost health environments, such as hospitals.[1] Many initiatives seek to better integrate community and social services withmedical case management. One novel subset of these integrated care initiatives are those that have succeeded in using electronic health records and other information technologies in their efforts to coordinate and improve care.

In this chapter we explore what technologies and organizational structuresare used by some of these novel healthcare initiatives to identify, monitor and provide better care for health care super utilizers. What technologies were used? What institutions were involved or created? What policies directly or indirectly incentivized the use of novel technologiesand organizational structures? What can local governments do to accelerate such information technology use for coordinated community care? The chapter is based on a literature review of the small number of flagship healthcare delivery modelswhich use uses information technologies to better combine medical and social services.

New approaches to healthcare delivery for super-utilizers areof course part of much broader changes in the social, economic and political landscape of health care. All health care providers, serving varied populations across the country, are being challenged to find models of health care delivery that yield higher quality care for lesser cost to the whole population, which is frequently referred to as the “Triple Aim.”[2]

How to transition the US healthcare system, or even single organizations therein, to deliver higher return on investments has been anintense policy debate in the United States for decades.[3] Proposed solutions to our health care crisis span a widegamut of policy tools,including: financial incentives for health care providers and individuals, changes to health care organizational structures and procedures, healthcareworkforce education and licensing, urban planning, individual behavioral changes related to nutrition and activity levels, and even urban planning.

Needless to say, then, this chapter will take a rather narrowfocus on a set oftechnologicaland organizational innovations adopted by some community health care providers. While we will discuss some of the health care quality metrics of these initiatives, we will not compare the effectiveness of the different models of care, or discuss their financial sustainability or scalability. The initiatives described here are simply recognized as alternatives to the standard model of care, but we are not yet in a position to evaluate them. They are promising experiments that address the needs of poor, often uninsured, patients with chronic conditions in a more continuous and coordinated manner. They are understood to behelping patients adhere to treatments and behaviors and thus helping to reduce expensive hospital emergency department use.

2. What is the problem with the over use of Emergency care?

Vulnerable populations include people with (frequentlymultiple) chronic complex conditions, such as diabetes or asthma, whose conditions are poorly managed; the frail elderly; and people without health insurance who seek medical care only when their condition becomes acute. These diverse groups of people are more likely to resort to the use of emergency departments at hospitals for their medical conditions. [Figure & cite needed.]

For over two decades, hospitals have been obligated to provide emergency care for any patient that visits the emergency department (ED), regardless of insurance coverage or ability to pay.[4] The expense of uncompensated emergency department care falls primarily on the receiving hospitals, although there are some offsetting federal funds from Medicare.[5] Emergency department visits are far more costly than regular physician visits and can result in the ordering of multiple tests and even hospital stays. Some of these ED visits might be averted if vulnerable populations had easier access to preventive, primary or specialistcare.[6]

The problem is not simply the fact that ED visits are more expensive than primary care visits. Vulnerable populations tend to include a sub population of super utilizers, patients that are repeat user of emergency care. Sometimes repeat users of ED return because the care is centralized and convenient (many hospitals, for example, have translating services). Other repeat users do so because their conditions are not controlled outside the hospital setting. They may be discharged without follow up, access to other providers of care, or a support network to help them manage their conditions. They may get redundant testing at multiple hospitals.

We are living in a period of rising numbers of people with chronic conditions including: diabetes, cardiac conditions, asthma, even cancer as “chronic” disease. About 60 percent of the adult U.S. population had at least one chronic condition in 2005. One in four people have two or more chronic conditions. Chronic conditions and obesity which is a risk factor for several chronic diseases, has added costs to the US Healthcare system.[7] But perhaps what this change in the health status of Americans signals, is the need to shift from a healthcare system that has been designed for acute episodes of care. As the president of the Center for Studying Health System Change notes, “We’re pretty good at treating a heart attack, but not so good at preventing and managing the underlying heart disease that leads to that heart attack.”[8]

How large a problem is the inappropriate use of emergency department care? One measure is the total charity and “bad debt” or uncompensated expenditures by hospitals. In 2011, according to the American Hospital Association, $41.1 billion of US hospitals’ total expenditures were for uncompensated care, (5.9% of total expenditures).[9] [Hospitals also argue that Medicare and Medicaid payments do not cover their actual costs, by a total of $28 billion in 2010.]

In 2010, the year the Affordable Care Act was enacted, over 42 million Americans and 11 million illegal aliens were without health insurance;in other words, over 15% of the population did not have regular access to affordable healthcare. In some southern and western states, over 1 in 5 people were not covered.[10] With the passage of the ACA the number of uninsured figures should be reduced substantially. However, the ACA, by design, expands the number of Medicaid recipients, who statistically use emergency care at significantly higher rates than the insured and even the uninsured.[11] So, even with expanded population insurance coverage, it is not clear that vulnerable populations will quickly migrate to non-ED care.[12]

The ACA is an important policy because it extends insurance coverage, encourages pay for performance and incentivizes new models of accountable care. But alone itcannot guarantee that the health care system will suddenly be able to absorb care for the expanded Medicaid rolls or the remainder of the vulnerable populationswith inadequately managedchronic conditions. The over reliance on emergency care is thus likely to continue if there are not also accompanying social and organizational changesthat make outpatient care more affordable and accessible.

3. What can be done to try to reduce the use of Emergency care?

There are many possible approaches to reducing the inappropriate use of emergency departments. Community care centers, mobile doctors’ offices, medical homes, tele-health, e-health, better hospital discharge instructionsand follow up… [Expand.]

One novelsolution, has been to use data about patient visits to ED in order to identify those categories of people that might benefit from some form of intervention, be it case management, referral to a community care center, or targeted healthcare or social care follow up. Ideally, electronic health records could be used to identify health care super utilizers or more broadly any patient who’s care could be better managed. Electronic health recordsincludeinformation about a patient’s demographic data, conditions, medications, test results, and care received.[13] But using identifiable health records is complicated because of privacy restrictions, amongst discussed below.

In the early 2000s, Jeffrey Brenner, the founder of the Camden Coalition of Healthcare Providers located in the poorest city in New Jersey, used billing data from hospitals to create“hot spot” maps identifying the health characteristics of individuals who ended up in Camden hospitals. Taking a page from police department targeting of crime hotspots, Brenner identified a number of interesting types of patients. For example, by looking at “locations where ambulances picked up patients with fall injuries” he discovered that “a single building in central Camden sent more people to the hospital with serious falls… than any other in the city, resulting in almost three million dollars in health-care bills.”[14] Of those buildings in the city that accounted for the highest number and cost of hospital visits one was a nursing home, another a low income housing tower.

The citywide health database Brenner created from the claims data of three Camden hospitals showed that “50 percent of Camden residents visited a local emergency department or hospital in a single year, twice the rate for the United States overall. The majority of the visits were for preventable conditions that are treatable by aprimary care provider.”[15]Most often people visited the ED for head colds, viral infections, ear infections, and sore throats. Thirteen percent of the patients accounted for 80 percent of hospital costs; 20 percent of the patients accounted for 90 percent of the costs.[16]Just one percent the Camden population accounted for thirty percent of its costs: out of one seventy seven thousand people who accessed care, there less than800 “super-utilizers.”[17]

What Brenner wanted to do was better understand the needs of the super utilizer patients in Camden. Because of privacy restrictions that limit the transfer of identified personal data, Brenner simply asked local doctors to refer their most problematic patients to him and he visited them in order to find out why they accumulated such high medical costs. In 2007 he founded the Camden Coalition of Healthcare Providers whose mission was to improve care by keeping vulnerable patients out of the hospital. The Camden Coalition provided a team based approach to the management of care for individual patients, including doctors, nurses medical assistants and social workers. They delivereda combination of personalized health care consultation and coordination,as well as help applying for government benefits, transportation, shelter and other social services. Using this holistic, coordinated and preventive approach to care, the Camden Coalition was able to reduce hospital visits by 40% and hospital bills by 56%.[18]The Camden Coalition demonstrated was that it was possible to transition super utilizers into an environment where most of their care was provided by the primary care system and a more robust social network, thus lowering the overall cost of their healthcare.

4. Other examples of successful data analysis and patient identification for reduced health care expenditures in vulnerable populations

[To be completed.]

  • Austin CommUnity Care, Travis County Texas; asthma project
  • Alliance for Health in western Michigan;
  • Better Health of Great Cleveland
  • English T. Community health workers, Temple University Health System, Philadelphia, PA
  • WellSpan Bridges to Health, York, PA
  • Alliance for Health in western Michigan;
  • Better Health of Great Cleveland
  • Crescent City Beacon Community in New Orleans;
  • The Greater Cincinnati Beacon Community; and
  • Southeast Michigan Beacon Communities in Detroit.

5. How do you get data-driven community care solutions?

The interventions by the Camden Coalition or Austin Community Care are driven by an understanding of the vulnerable patient population derived from health records. Obviously, the technological infrastructure that these community based care organizations are now using does not in itself guarantee better health care outcomes. The models of care provision matter enormously.

However, in this chapter the focus is on how to build the necessary information infrastructures. Surprisingly,the electronic health records, databases, information exchanges amongst hospital and providers are either still very new or in many instances do not yet exist. Moreover, there are several impediments to their use, some of which local organizations and governments might be able to help surmount.

What information and the technologies do the Camden Coalition and other community groups need to identify high use, vulnerable patients?

First, there must be electronic health data for all relevant patients so that data can be analyzed andindividuals with complicated medical and social needs identified. The reality is that while electronic health records have existed in some (usually partial or fragmented) form for over thirty years, it was not until the last few years that their use permeated the large majority of hospitals. Adoption of EHRs only accelerated in 2009 with the passage of the federal HITECH Act which, for all intents and purposes, mandated the use of electronic health records for any provider that services the Medicare and Medicaid populations.[19] Identifying super utilizer patients across a city or region requires each health care provider to be able to generate an electronic health records. If a patient uses multiple health care providers, or goes to multiple hospitals, each of those institutions will likely have separate EHRs for that person.

Secondly, the EHRs from separate providers must be sharable. It is important for healthcare providers to be able to see all the relevant clinical information about their patient, regardless of where they were treated in the past. A comprehensive view of patient information helpsto avoidthe unnecessary duplication of tests, to improve diagnostic and prescription accuracy, and to facilitate care coordination.

There is a distinction to be made between the ability to transmit data across organization and the ability to seamlessly incorporate data from different sources into EHRs. In the US (and internationally) there are standards for diagnostic and procedural codes, laboratory tests, and prescription medications which facilitates the ability to understand the content of different EHRs. Unfortunately, the interoperability standards for merging EHRcontent are not widely used. In other words, it is currently unlikely that hospitals have the capacity to automaticallyincorporate data generated by another EHR system.If all a patient’s care is within a single healthcare system, the data within that EHR is easily be kept up to date. But across healthcare systems, there are over 1000 different EHR products in use, which makes the merging of data difficult and thus care coordination and management is more complicated.[20]

Thirdly, organizations need consent or permissionto transmit and use personally identifiable health data. Health information exchanges must be set up across the local health care providers to facilitate the flow of data between them. Health information exchanges make it possible to share data about what patients they have (i.e., a master patient index) and what sort of data they hold (i.e., adocument locator/registry). HIEs establish the rules of exchange amongst health data holders and with patients. Barring an HIE, data will remain siloed in separate health care provider organizations for both privacy and commercial reasons.

Depending on the state, patients will either need to consent to participate in an HIE or they might be assumed to consent unless they opt out of the HIE. A few HIEs have the capacity to let patients decide what categories of health information they are willing to share within the HIE; most have an all or none approach to data exchange. There are several different models of health information exchanges and much experimentation as many HIEs are still being developed (in part to meet the data exchange requirements of the HITECH Act). Broadly speaking one can identify three models:

(1)Centralized HIEs use a single Clinical Data Repository that is maintained by an HIE authority. HIE members (e.g. hospitals) electronically transmit agreed upon patient health information to the CDR, where it is securely stored and continually updated.[21]

(2)Federated HIEs do not have a central data repository and do not normalize or standardize the data which are remotely stored in each separate organization’s EHR. The internet is used as the vehicle for moving data. To retrieve patient data, a query is sent to the HIE’s patient registry which then provides a “virtual roadmap” of where patient health records are located.

(3)Private HIEs can be created by a hospital system. Essentially a centralized model, it includes internal databases and referring physicians. Many ACO structures are supported with private HIEs. [22]