Home  |   Subscribe  |   Resources  |   Reprints  |   Writers' Guidelines

March 2020

The Untold MPI Story: How ‘Dirty Data’ Affect a Value-Based Revenue Cycle
By Lisa A. Eramo, MA
For The Record
Vol. 32 No. 2 P. 22

The misidentification of patients is typically considered to be a safety issue, but the problem also carries financial ramifications.

When most of us think about master patient index (MPI) errors, our minds immediately go to the worst-case scenarios: A patient prescribed a medication to which they’re allergic experiences a negative outcome. A patient undergoes the wrong procedure. There are delays in a patient’s cancer diagnosis and treatment all because of a duplicate record or overlay.

In fact, the ECRI Institute identified diagnostic errors and the improper management of test results in the EHR as the top patient safety challenge in 2019. It isn’t a stretch to assume these errors and data mismanagement are largely due to an inability to match records accurately.

But experts say negative patient safety–related outcomes don’t tell the entire story of what can happen when MPI errors occur. In fact, there are other important implications that actually affect an organization’s bottom line and perhaps its ability to remain profitable.

“There’s a lot of literature on how the MPI affects patient safety, but not a lot of information on how the MPI affects reimbursement and the revenue cycle,” says Joe Lintz, DHA, RHIA, CHDA, HIM program director at Parker University in Dallas, who researched the impact of the MPI on revenue at three major hospitals as part of his doctoral dissertation at Central Michigan University.

Lintz, who manually reviewed denials and error codes at all three hospitals over a period of six months, found that although MPI error rates were fairly low, there was significant revenue loss due to improper tests, duplicate testing, and inappropriate billing and coding.

For example, one hospital had approximately 8,500 discharges over six months with an MPI error rate of 0.14% (eg, only 12 records). However, using questionnaires completed by MPI staff, he gleaned that it took an average of 10 hours per error to research and reconcile the information. At a rate of $85 per hour (total of staff and consultant time), this equates to $10,200. For organizations with higher duplicate rates, this amount is even more significant. This is in addition to revenue lost due to the actual denials, Lintz says.

A recent survey conducted by BlackBook Research found that 33% of all denied claims result from inaccurate patient identification or information, costing the average hospital $1.5 million in 2017 and the US health care system more than $6 billion annually.

As if the patient safety issues weren’t enough, why haven’t more hospitals cleaned up their MPIs, given the operational burden and lost revenue associated with these errors? One reason could be that providers underestimate their number of duplicates, says Erin Benson, senior director of market planning at LexisNexis Risk Solutions. It’s not uncommon for 10% to 20% of an organization’s records to be duplicates, and the number may be even higher when mergers and acquisitions occur, she says.

Another reason is that organizations tend to have limited budgets, so they prioritize efforts that directly affect patient care, such as new technologies and adding staff. However, Benson says the tides are turning in light of capitated payments.

“Organizations want an accurate count of overall lives when they’re trying to make decisions,” she says, adding that accurate information is also essential for preventive care, social determinants of health (SDoH), and population health management initiatives.

For example, it’s difficult to remind patients of preventive screenings and prevent readmissions when their contact information is spread across multiple files without a single source of truth. Likewise, it’s difficult to target SDoH interventions when ZIP codes and other data are inaccurate.

All of these efforts affect value-based payments either directly or indirectly, and it’s why organizations are starting to focus more proactively on MPI maintenance, Benson says.

Still, challenges remain. Mergers and acquisitions introduce errors in the MPI, says Vince Vitali, CHCIO, FCHIME, FHIMSS, vice president of strategy and business development at NextGate. “The number of registration points into a health system has increased dramatically,” Vitali says. “You have cancer centers, outpatient therapy areas, imaging centers, reference laboratories—there are so many more opportunities for error.”

A Blow to the Pocketbook
MPI errors can affect revenue in a variety of ways. Consider this example: Mary Jones lives in Chicago but moves to New York City for a few years, during which time she gets married. When she returns to her provider, she is now Mary Kimmel living at a new residence. A registrar doesn’t realize it’s the same patient, so a new record is created.

“In addition to her past medical history being orphaned, there now exist two medical records for Mary,” says Victoria Dames, senior director of product management at Experian Health. “This situation can happen very quickly with lifestyle changes such as marriage and divorce or with user manual entry error.” If the original record is used for billing purposes, the patient’s name on the claim won’t match with the name on the insurance card, leading to a denial, Dames says.

This scenario can also lead to a coordination of benefits issue, Vitali says. For example, if Mary has coverage through her employer as well as additional coverage through her spouse’s employer, the claim may be denied.

Problems also occur when a lab or other test result is sent to a duplicate record. When the provider can’t find the results because they’re in the duplicate record, they may ask the patient to repeat the lab or test, which could cause a denial, Benson says.

MPI Errors and the Patient Experience
There’s also a patient satisfaction angle to all of this, Benson says. When patients need to repeat labs and other tests, they may rate their providers lower on a Consumer Assessment of Healthcare Providers and Systems (CAHPS) survey, which measures timely access, care coordination, provider communication, and more. These quality ratings ultimately affect an organization’s ability to collect maximum reimbursement, Benson says.

In addition, when patients need to redo labs and other tests, delays in diagnosis and treatment can occur; this, in turn, affects actual health outcomes and decreases quality ratings, she says.

Organizations also need to think about MPI errors in the context of high-deductible health plans, Vitali says. “Patients are responsible for more and more of the bills these days, and you’ve got to make sure you have the correct patient information,” he says, adding that it’s not uncommon for organizations to repeatedly send bills to a wrong address because registrars fail to update the information. When the bills remain unpaid, the organization may send them to a collection agency, incurring additional costs. In some cases, these late payments may even affect patients’ own credit scores, which can lead to lower satisfaction ratings on CAHPS surveys.

Potential Solutions
Experts agree that solving the problem of dirty MPI data requires a multiprong approach that includes people, process, and technology. Consider the following strategies:

Teach registrars about the revenue cycle. Vitali says registrars need to understand the connection between the MPI and reimbursement, including all of the ways in which the data they input affect their organization’s ability to collect payment.

Their role is critical because in most cases they meet face to face with patients and can dig more deeply when questions arise about identity, he says. “After patients leave the hospital or clinic, it’s difficult to get ahold of them and get information in a timely fashion,” Vitali notes.

To prevent unnecessary duplicates, HIM directors must provide registrars with thorough training not only in terms of what data to capture—and how—but also on how to conduct searches in the EHR and what processes to follow when questions arise, he adds.

Focus on consistent data input enterprisewide. “Anything you can do to make sure the initial data coming in are entered in a standardized format will improve match rates,” Benson says.

For example, standardizing patient addresses using the United States Postal Service (USPS) format in EHRs has been shown to improve match rates by up to 3%, according to a recent study published in the Journal of the American Medical Informatics Association.

Likewise, anyone entering patient data into the EHR should populate as many name fields as possible—first name, last name, nickname, alias, other registered name, maiden name, and preferred name. “That way, you can cover every angle when you search,” says James Hoover, cofounder and executive vice president of Medarcus. “Leveraging all the possible name fields greatly increases your chance of finding the right patient.”

Asking patients to spell their names is important, as is spelling it back to them for confirmation, Hoover adds.

For patient lookup, Lintz says registrars should be consistent in their approach. For example, a surgery unit might use Social Security numbers to link patient records while radiology and pharmacy collect information from driver’s licenses. Ideally, all three departments would use the same look-up methodology since many duplicate medical record errors can be traced to small miscues and inconsistencies that arise during the patient registration process.

According to Lintz, using consistent methods for patient lookup not only ensures better care coordination for patients who see multiple health care providers but also ultimately reduces costs associated with patient misidentification. Recommendations include the following:

Make patient photos accessible to the registrars. This may seem obvious, but Hoover says most organizations don’t display patient photos in the EHR at the time of registration. “We’re losing our last best chance to verify a patient,” he notes.

Establish a denial management strategy. Organizations need an enterprisewide view of denials and the ability to trace each denial to its root cause—specifically errors in the MPI, Vitali says. Ideally, they should also use back-end business analytics to correlate billing errors with patient satisfaction scores. “It’s not an easy process, and most organizations don’t have a handle on how to do it, but it’s definitely something that gives good payback,” he says.

Implement an enterprisewide MPI (EMPI). An EMPI aggregates information from multiple systems across an organization to create a single patient identity. Vitali says an EMPI is critical because it enables organizations to integrate outpatient clinics and physician practices, places where patient information tends to be more accurate because there are more frequent points of contact.

Capitalize on technology. Referential-matching technology uses unique identifiers and third-party data (eg, data from public records and other external agencies) to provide continuous updates to the MPI, handing registrars the tools they need to be successful, according to Dames.

Why is this technology necessary? In theory, patients should be able to give registrars all the information they need to find the right record, but in reality, it doesn’t always work that way, Vitali says. Patients forget information or they simply may not be in the right frame of mind, he says. “Patients are caught up in their own situations, and they just want to get in and see their provider,” Vitali notes.

One limitation of referential data is that some people may have only a few public records (eg, minors, undocumented immigrants, those without a credit history), though Benson says sophisticated linking algorithms can often help overcome this hurdle.

Geocoding, a form of location intelligence that converts addresses into geographic coordinates, is an additional element that enables organizations to authenticate address information in real time to avoid duplicate record creation and identify fraud, Vitali says.

Hoover says facial recognition and soft biometrics can also help organizations match patients more effectively. Soft biometrics, for example, measure the speed at which someone signs their name as well as the pressure of the pen—details that are nearly impossible to replicate.

In addition to these technologies, Hoover—a trained computer scientist—says organizations can capitalize on “modern” search—which converts text-based queries to mathematical-based searches—to improve match rates. “You can do precision searches with unclean data either on input or in the database and still find the exact match,” he explains, adding that this technique is used by Google and other search engines.

Modern search helps overcome instances of compound errors—situations where there are multiple errors either on input or during the patient look-up process. This can easily occur with patients who use a European date of birth format (ie, day then month then year) or observe Chinese traditions (ie, first and last names are reversed).

When registrars make errors with the name or date of birth—as well as address or other identifying information—the chances of finding the right patient are slim, Hoover says. “These compound errors are where EHRs fail miserably. They’re just not programmed for that kind of uncertainty,” he says. “If we put modern search in front of EHRs, we’ll prevent a significant number of errors. This is what we, as an industry, need to do.”

Perform quarterly MPI audits. Even a rudimentary data analysis of records with matching Social Security numbers is worthwhile, according to Hoover, who once found 85,000 instances of two or more records sharing the same number at one hospital, including 11 separate records sharing a single number.

Other audit types focus on records with addresses not in the USPS standardized format and records with a middle initial rather than a middle name spelled out. “A certain percentage of your data has errors in the spelling or addresses, and that’s a proxy for how many other spelling errors you have in other data fields,” Hoover says.

No Time to Waste
Experts agree that solving the problem of MPI errors will be critical as organizations move forward with value-based payment models. “We focus on patient safety because that’s why health care organizations exist—to help patients and keep them safe,” Benson says. “But the revenue angle is critical as well.”

— Lisa A. Eramo, MA, is a freelance writer and editor in Cranston, Rhode Island, who specializes in HIM, medical coding, and health care regulatory topics.