ER Overuse: Leveraging Analytics To Improve Care, Reduce Visits

By Eric Grossman, CEO, NextHealth Technologies, and Terri Mayne-Jarman, Point B Management Consulting

A mother takes her asthmatic child to the ER for treatment. An elderly man with a stomach ache visits the ER for an assessment. A 30-something gets bitten by a spider and promptly heads to the ER. The common thread among each of these patients was a visit to the ER. Yet each could just have easily visited the pediatrician, primary care provider or an urgent care facility.

Unnecessary visits to the ER, or visits in which patients could have received treatment through less costly and more optimal sites for care, is a $38 billion problem in the U.S. each year according to a 2010 NEHI report. Many states, such as Washington, are giving providers the opportunity to earn 1 percent incentive payment under the Medicaid Quality Incentive Program, and reducing unnecessary ER visits is one of the measures. Finding products and services which allow providers and payers to reduce cost and increase payments is key to their financial success.

There are numerous long-held beliefs about reasons patients rely on the ER for their medical care. According to Robert Wood Johnson Clinical Scholar Research conducted in 2013 which sought to debunk some commonly held myths about frequent ER use among low-income patients, insurance status was not the key. Study respondents offered a variety of reasons, including convenience and the ability to get in to see a doctor faster than a PCP, cost and lack of co-pays, and a perceived higher quality of care.

Yet since that report, while Medicaid and Affordable Care Act coverage of previously uninsured populations have expanded, so too has reliance on the ER. A randomized trial in Oregon showed expanding Medicaid coverage increased ER use by recipients by about 40 percent, including visits for conditions that might best be treated by a primary care physician (Taubman et al., 2014). In addition to imposing needless costs on premium payers and taxpayers, such overuse increases waiting times, threatens service quality, and reduces the availability of ER medical personnel for those who most need them (NPP 2010).

There are many causes for avoidable ER utilization, and in an effort to control costs and reduce ER utilization, many healthcare organizations recommend that many of the ER patient follow up with their primary care provider in three to five days. However, this strategy is not effective — patients may not have a PCP or they just don’t follow up. Organizations need effective strategies which help the patient find a PCP or schedule the appointment immediately upon discharge or even before they leave the ER.

As a result, reducing inappropriate use of the ER has become a top priority for many healthcare risk managers. And as the healthcare landscape evolves from fee for service to value-based care, the problem takes on even greater importance as payers and providers look to improve outcomes, focus on preventative measures, and reduce costs by identifying and intervening with at-risk populations.

Solutions To Reduce ER Visits

Many approaches have been tried in an effort to reduce ER overuse and appropriately avoidable costs, some involving expensive commitments of resources. Some examples include providing a regular medical checkup schedule and transportation for very heavy users (NPP, 2010), and others using telemedicine or clinics to try to divert members from inappropriate use of the ER (ibid).

Despite efforts to reduce ER reliance, changing behavior has proven difficult. In theory, leveraging a combination of channels such as hospital-based urgent care clinics, telehealth services, and retail clinics should create an impact on reducing trips to the ER. In reality, the problem persists on a large scale.

Several market imperfections make it difficult to change the behaviors of real patients enough to achieve large-scale, lasting reductions in ER overuse. For example, Medicaid recipients may lack both information and incentives to avoid needlessly costly ER visits. They may be unaware not only of the costs to others of their choices, but also of the availability of lower-cost, comparable or better quality of care, and possibly more convenient alternatives. On the provider side, reimbursements that reward hospitals for busy ERs have led many hospitals to actively seek to attract more Medicaid patients to their ERs by advertising their convenience and excellence. Thus, unnecessary ER visits can be expected to continue to increase unless and until countervailing strategies are developed to increase patient and provider awareness of and willingness to use lower-cost alternatives.

However, the economics really depend on the patient demographics. For example, if the patient population is heavy Medicare/Medicaid, then offering lower-cost alternatives makes sense. But, if Medicare or Medicaid reward and/or penalize health systems for the higher cost ER visits, commercial payers typically will follow suit.

To overcome these difficulties, practical solutions are needed that are both affordable and scalable to large numbers of patients. Relatively inexpensive actions, such as outbound calls or direct mail, are preferable to more-costly ones, such as counseling by a nurse or a physician, when either one would successfully change behaviors to reduce inappropriate future usage of the ER by diverting likely over users to more appropriate care settings.

Following are some guidelines to help implement your strategy. Begin by determining:

1. What problem are you trying to solve? Are you working to reduce ER usage? Provide guidance in helping specific patients more easily manage a chronic condition in which they can avoid office visits? Reduce hospital readmission? What are your key performance indicators (KPIs) for the use case?
2. Who are you targeting? Once you know the problem you want to address, use predictive analytics to pull subsets of members (populations) who are most likely to impact the KPIs — for example, those with a history of avoidable ER usage — and those who are impactable. Targeting all high-risk users versus those who are both avoidable and impactable leverages scarce resources. Overuse is determined not only from the frequency of use by members who are classified as healthy by a commercial risk grouper (which maps various types of medical claims information to health status indicators), but also by applying a standard model, the NYU avoidable visits model (Ballard et al., 2010) to diagnostic codes for patient visits to the ER to estimate the fraction of them that are avoidable, in the sense that they should be treated elsewhere.
3. What can you say to change behavior? After determining your population targets, such as women 25 to 40 years old with at least one ER visit in the past year, or parents of young children with at least one ER visit, develop messaging aimed at these populations, taking information you have learned to help change the behaviors of these populations.
4. How do you deliver the message? Determine the best channels to engage the patient, whether by phone, email, social media, direct mail, or other means. Based on what is known about these patients, determine which methods each will respond to in order to help educate and change their behaviors.
5. Measure and optimize. Finally, benchmark your communications, measure results, and determine what’s working — and what’s not. Then optimize your outcomes by redirecting resources to those campaigns and populations where you are seeing success.

Ultimately, you need to marry analytics with consumer engagement to change behavior and drive outcomes. This is true for both providers and payers, and is true across business use cases — from ER avoidance to readmissions reduction to chronic disease management.

Through the use of sophisticated predictive models and advanced algorithms, technology can identify key predictors and power combinations of predictors in which healthcare organizations can implement programs or campaigns to initiate a change in patient behavior. Data should first identify characteristics of individual members that reliably predict high risk of undesirable behaviors — in this case, ER overuse — under current conditions.

These characteristics then inform inferences about which individual members’ ER overuse behaviors can potentially be reduced the most. Armed with this data, deliver carefully designed and targeted messages across channels such as live outbound calls to members or follow-up letters that use principles of behavioral economics to motivate changes in behavior (Thaler and Sunstein, 2008). Then, then deliver messaging to individual members to optimize predicted improvement (or lift) in KPIs such as ER visit reduction. Monitor resulting changes in behaviors and adjust messaging and possibly communication channels over time to maximize the improvement in KPIs achieved by allocating available resources to nudge members.

This example reveals that relatively simple and inexpensive measures can be used to accomplish significant behavior change if messages are well crafted and well targeted to those who will most benefit from them. Crafting carefully targeted messaging and campaigns can help achieve important reductions in costs and improvements in the efficiency and convenience with which care can be provided to vulnerable populations.

In the end, healthcare organizations need to own their patient engagement process. Effective communications to change behavior require a plan for prioritizing goals, implementing communications, measuring results and adjusting messaging as needed. However, as with any other strategic program, the organization needs to align around the strategy and vision and identify the internal resources best suited to achieve the defined goals. Combining the right tools with organizational effectiveness may lead to measurable outcomes.

About The Authors

Eric Grossman is the CEO and founder of NextHealth. Previously, Eric was the Vice President and General Manager of TriZetto’s Consumerism and Analytics business unit. Prior to TriZetto, Eric was the founder and CEO of Connecture (NASDAQ: CNXR). Connecture is the leader in health plan sales automation solutions. Eric was also Chief of Staff to the Vice Chairman of Ernst and Young. Eric is currently a Governor-appointed, two-term member of the Board of Directors for the Colorado Health Benefit Exchange.

Terri Mayne-Jarman is healthcare group principal at Point B, a management consulting firm. Mayne-Jarman has extensive experience managing large systems development projects across the revenue cycle and bridging gaps between business and IT.