Published on 30-Jul-2010
Validated on 21 Jul 2010
"With IBM SPSS Modeler, you can build models in less than a week.This allows us to keep our prices low and it makes me more productive. In this field, that’s really important." - David Mould, Ph.D., Predictive Analytics Scientist, MedeAnalytics
BA - Business Analytics, BA - Business Intelligence, Smarter Planet
Hospitals across the nation are facing a tough dilemma: a growing number of the patients they treat lack health insurance or are under-insured — but turning these patients away isn’t an option. The result: more and more patients are being classified as “self-pay” and hospitals increasingly have to contend with collections headaches and budget shortfalls.
Hospitals face the difficult challenge of recovering payments from under-insured and uninsured self-pay populations. To improve the collections success rate, hard data were needed to predict which patients would be in a better position to pay.
Hospitals turned to MedeAnalytics for a cost-effective analytics solution, utilizing IBM SPSS Modeler, which helps them prioritize and focus collection efforts, boost revenue and optimize collections staff.
By using data mining and statistical techniques, hospitals become more efficient and accurate in their collection activities, minimizing the loss of time and money from haphazard methods. One hospital saw a 30 percent reduction in bad-debt write-offs, a 12 percent increase in self-pay collection rates, and $25,000 per month reduction in agency fees.
Hospitals across the nation are facing a tough dilemma: a growing
number of the patients they treat lack health insurance or are
underinsured—but turning these patients away isn’t an option.
The result: more and more patients are being classified as “self-pay”
and hospitals increasingly have to contend with collections headaches
and budget shortfalls. “The rising uninsured and under-insured are
among the top issues plaguing healthcare,” researchers at Deloitte LLP
noted in a recent study. “Not only does this issue impact access to care
and hospitals’ ability to fulfill their mission, it impacts their ability to
protect margins and keep their doors open.”
In an effort to better understand and manage this segment of the
patient population, many hospitals are turning to MedeAnalytics’
Self-Pay Analytics solution. Based on predictive analytics software
from SPSS, an IBM company, the solution is enabling hospitals with
large self-pay populations to measure and predict patient payment
behavior, reduce risk from bad debt, and boost collection rates.
A better way to tackle self-pay collections
Most hospitals operate according to a treat-first, seek-payment-later
protocol—an approach that works well ethically but not necessarily
financially. In fact, research has shown that only about 15 percent of
self-pay patients end up paying for their services, with the balance being
written off by healthcare organizations as bad debt. The collection
process itself can be time-consuming and expensive, requiring hours
of phone calling and letter writing. Often, hospitals turn to outside
collection agencies to help with the process, but that is a costly alternative.
MedeAnalytics, however, provides a better way to tackle the self-pay
problem. Among other services, the healthcare performance analytics
company develops statistical models that help hospitals prioritize which
self-patients are likely to pay, and focus collections efforts on this high-
yield segment of the population. “We want to maximize the productivity
of collectors by giving them a list of patients who are more likely to pay
the hospital back and put the people who are unlikely to pay down at
the bottom of the list,” says David Mould, Ph.D., predictive analytics
scientist for MedeAnalytics and the principal developer of self-pay models
used by more than 100 hospitals throughout the U.S.
Mould relies on IBM SPSS Modeler to build and run the predictive
models, which analyze 20 or more variables that typically influence payment
behavior, such as the admission source (elective vs. an emergency),
the type of treatment or procedure, and the income and demographics
of the patient. “There is a lot you can tell by the procedures performed.”
Mould says. “Among people who have elective surgery, for example,
there is a higher chance of collecting,” In other cases —for instance, in
many emergency room visits—the likelihood of payment is extremely
low. “In those cases, we tell them to send it to a collection agency, don’t
even try to work it,” Mould added.
Detailed patient profiles
To improve the accuracy of the models, Mould builds a detailed profile
of a hospital’s patient population, drawing on admissions and treatment
records going back more than a year. A typical history file will contain
information on as many 80,000 admission records (although patient
names are never included in the data set). Mould regularly adjusts the
variables and algorithms in the model to increase its “lift”— a measure
of the predictive success of the model’s prioritized list compared to a
random sampling of patients.
Using the MedeAnalytics’ solution, collectors and financial administrators
can identify which self-pay patients have a high probability of collection,
which qualify for charity care, and which would be best to refer to
financial counseling or directly to a collections agency. “Using our
model, the collectors can work the patients that have a high probability
of paying just by giving them a phone call or sending them a letter,”
Mould says. As a result, hospitals improve collection rates, increasing
revenue while reining in costs associated with patient outreach.
Florida hospital turns to predictive analytics
For example, one of Florida’s most highly regarded private, not-for-profit
health networks turned to MedeAnalytics’ solution, motivated primarily
by self-preservation. The hospital, which treats more than two million
people each year at its regional network of facilities, was struggling with
bad debt write-offs and poor recovery rates among its self-pay patients.
Its 16 collectors, operating in an information vacuum, spent much of
their time trying to collect from patients who were unable or unwilling
to pay. At the same time, patients legitimately eligible for charity care
weren’t being identified, adding to the hospital’s bad debt. The result
was fiscal uncertainly, concerns about long-term viability, and steadily
deteriorating morale among collections staff.
The hospital hoped to recoup charges by improving its collections
processes and eliminating wasteful expenditures such as returned mail
and agency fees. The hospital was already using a suite of MedeAnalytics’
statistical, financial and operational analysis tools with good results.
For this application, however, MedeAnalytics recommended adding
its Self-Pay Analytics solution.
Before the hospital could begin using the solution, however, Mould had
to design a model tailored to the Florida hospital. “We build these for
urban hospitals, rural hospitals, one-building hospitals, multi-hospitals,
with models tailored specifically for each hospital type,” Mould explained.
In short order, MedeAnalytics had developed the customized model and
distributed it to all revenue cycle managers, collectors and financial
By prioritizing and focusing its collection efforts, the hospital garnered
a host of benefits. Specifically, the hospital calculated a 30 percent
reduction in bad debt write-offs; a 12 percent increase in self-pay
collection rates; $270,000 per month savings on returned mail; a
$100,000 per month increase in Medicaid reimbursements; and $25,000
per month reduction in collection agency fees. The hospital has also
been able to reduce the number of on-site collections staff by 50 percent.
“Self-Pay Analytics addresses the most critical problem facing hospitals
today,” the hospital’s vice president of revenue management says. “With
MedeAnalytics, we have been able to leverage analytics to collect more
self-pay dollars, earlier in the revenue cycle, with fewer internal resources.”
Mould credits IBM SPSS Modeler software with enabling the company to
develop its Self-Pay Analytics solution rapidly. One of Modeler’s big advantages,
he says, is its drag-and-drop capability, which accelerates model creation.
“With SAS, it’s three to six months to build a model,” he says. “With IBM
SPSS Modeler, you can build it in less than a week. This allows us to keep
our prices low, and it makes me more productive in terms of cranking out
models. In this field, that’s really important.”
Founded in 1994, MedeAnalytics enables healthcare organizations to
improve clinical, financial and operational performance through on-demand
analytics and client services. For hospitals and health systems, MedeAnalytics
provides solutions to address revenue cycle, patient access, clinical operations,
staff productivity, regulatory compliance, finance and accounting, and
enterprise performance management. For healthcare payers, the company
provides solutions for provider network management, medical management
and operations. MedeAnalytics delivers business intelligence to over 800
healthcare organizations using a hosted, Software-as-a-Service (SaaS) model,
which reduces up-front costs and enables rapid implementation and exceptional
time-to-value. For more information, visit www.medeanalytics.com.
About SPSS, an IBM Company
SPSS, an IBM Company, is a leading global provider
of predictive analytics software and solutions. The
company’s complete portfolio of products — data
collection, statistics, modeling and deployment —
captures people’s attitudes and opinions, predicts
outcomes of future customer interactions, and then
acts on these insights by embedding analytics into
business processes. IBM SPSS solutions address
interconnected business objectives across an entire
organization by focusing on the convergence of
analytics, IT architecture and business process.
Commercial, government and academic customers
worldwide rely on IBM SPSS technology as a
competitive advantage in attracting, retaining
and growing customers, while reducing fraud
and mitigating risk. SPSS was acquired by IBM
in October 2009. For further information, or
to reach a representative, visit www.spss.com.
Products and services used
IBM products and services that were used in this case study.
© Copyright IBM Corporation 2010 SPSS Inc., an IBM Company Headquarters, 233 S. Wacker Drive, 11th floor Chicago, Illinois 60606 SPSS is a registered trademark and the other SPSS products named are trademarks of SPSS Inc., an IBM Company. © 2010 SPSS Inc., an IBM Company. All Rights Reserved. IBM and the IBM logo are trademarks of International Business Machines Corporation in the United States, other countries or both. For a complete list of IBM trademarks, see www.ibm.com/legal/copytrade.shtml. IBM, the IBM logo, ibm.com, Smarter Planet and the Planet Icon (Western View) are trademarks or registered trademarks of International Business Machines Corporation in the United States, other countries, or both. These and other IBM trademarked terms are marked on their first occurrence in this information with the appropriate symbol (® or TM), indicating U.S. registered or common law trademarks owned by IBM at the time this information was published. Such trademarks may also be registered or common law trademarks in other countries. A current list of IBM trademarks is available on the Web at ibm.com/legal/copytrade.shtml Other company, product, or service names may be trademarks or service marks of others. This case study illustrates how one IBM customer uses IBM products. There is no guarantee of comparable results. References in this publication to IBM products or services do not imply that IBM intends to make them available in all countries in which IBM operates.