January 2019

Diversity Key to Patient Matching
By Susan Chapman
For The Record
Vol. 31 No. 1 P. 18

A RAND Corporation study recommends taking multiple approaches to the long-running problem.

The list of ills that ail the health care industry remains quite extensive. From prohibitive costs to disconnected information systems, the field continues to fight upstream against a range of obstacles, some self-inflicted, others not.

In the HIM arena, the plethora of duplicate medical records is top of mind at many organizations. It's a problem with far-reaching effects, including patient safety.

In 2018, RAND Corporation, an independent nonprofit research and policy organization, released the study "Defining and Evaluating Patient-Empowered Approaches to Improving Record Matching." The ambitious report, sponsored by The Pew Charitable Trusts, is designed to identify "potential 'patient-empowered' solutions to improve patient record matching, evaluate the solutions based on specific criteria, and, based on that assessment, select a promising solution (or cluster of solutions) within this range of approaches for further development and pilot testing."

How Errors Occur
In its October 1, 2018, alert "Quick Safety 45: People, Processes, Health IT and Accurate Patient Identification," The Joint Commission defines accurate patient identification as having the following three components:

• accurate information gathering (also called "catching" or "capturing");
• accurate information matching; and
• display of that information to enhance gathering and matching.

According to the Pew/RAND report, patient misidentification and errors in linking medical records across different databases "impedes interoperability and health information exchange among health care providers, increases health care costs, and hampers health care quality."

Using 11 criteria, the study evaluated the following patient-empowering solutions:

• implementing a voluntary universal identifier;
• using a public key as an identifier;
• expanding the use of existing government-issued identifiers;
• adding knowledge-based identity information;
• adding biometric data;
• having patients verify identity information;
• using consumer-directed exchange;
• using health record banks;
• having patients manually verify record matches; and
• having patients supply record location information.

"One thing we found in our work is actually what we didn't find: We were not able to find a silver bullet," says Robert S. Rudin, PhD, an information scientist at RAND and the study's lead investigator. "Instead, we recommended that we try a range of approaches. Essentially, there are strengths and weaknesses with every solution, which is why we need to get away from hoping that one solution will solve the problem and instead find a mix of solutions."

RAND researchers have proposed a three-stage solution that aims to improve the quality of identity information, establish new smartphone app functionality to facilitate bidirectional exchange of identity information and health care data between patients and providers, and create advanced functionality to further improve value and address other evaluation criteria such as likelihood of adoption and sustainability.

The study notes that additional pilot testing and development are necessary to advance the three-stage solution model, including developing best practices to verify mobile phone numbers and establish communication between health care providers and smartphone apps.

The researchers recommend the establishment of an overseeing organization "to help accelerate efforts to improve record matching," more research be conducted into record-matching errors, and the creation of "methods for health care providers to objectively benchmark their record-matching performance."

"Because patient identification has been a longstanding concern, people have been looking at it in different ways," Rudin says. "However, this is the first time that people can systematically compare possible record-matching solutions that involve the patient and give some kind of assessment to their different strengths and weaknesses. Over the years, there have been a lot of opinions on only one type of solution—for instance, biometrics.

"But it's very difficult to assess the strengths and weaknesses of proposed solutions in isolation. As a team, we took the first attempt at going through each solution and assessed its strengths and weaknesses. We also highlighted that there is a lot more work to be done on most of these solutions in order to effectively engage patients. As an example, there needs to be more analysis of how effective the workflows are. There are a lot of different options, but more work must be done to make them a reality."

Mark LaRow, CEO of Verato, a referential-matching solutions vendor, says medical record discrepancies frequently occur at patient presentation and when matching patients with their respective medical records.

"The problem at patient presentation can occur due to poor handwriting when the patient completes forms or inputting on the side of the health care facility," he says. "The other part of that problem is connecting that person to existing medical records, which could've been captured more than two decades ago. Over such a long period, so many things can change, making it difficult to connect the individual with the correct medical record."

Feasibility
Rudin believes that assessing the feasibility of implementing the study's suggestions is difficult to know until there is more information available. "But none of what we're describing requires a major technological breakthrough. It's more about people rolling up their sleeves and figuring out what can be done," he says. "These are things that can improve over time. Version 1.0 is the hardest. We're more in the development stage and working out the kinks. We're not prepared to make the recommendation that we should take these patient-empowered solutions and implement them in health care across the country."

There is also the question of how viable the report's recommendations are in terms of who is able to use the technology, even those who may not be technologically savvy, particularly as it pertains to smartphone usage. "One of our recommendations was to make better use of mobile phones and mobile phone numbers, which would not require a smartphone," Rudin responds. "Not everyone has a smartphone—more people have mobile phones than they did in the past but not everyone. But there are limitations with every technology—biometrics, for instance. With elderly individuals, fingerprints aren't necessarily effective because fingerprints are not as discernible as we age."

Still, approximately 75% of American residents use smartphones, including 50% of older adults. "That is a significant percentage," Rudin says. "Even if there is lower adoption among, for example, the elderly population, the number among senior citizens using smartphones is still increasing."

Elaborating on the example of older adults and smartphone technology, Rudin notes that those who are not using smartphones may have caregivers or relatives who do use the technology. "By the time a solution involving smartphones is adopted by the health care industry, smartphone adoption will become even greater. If we have the processes in place to help with that, not only for managing senior care but also for people who manage the care for their own children, if we have a smart phone app, the value is there," he says. "The potential is clear, which is why it was a particularly promising solution. It would be a shame then if we make decisions based solely on current adoption patterns because it all takes time. We would be wise to create the infrastructure now."

Dan Cidon, chief technology officer for NextGate, concurs. "While the RAND study thoughtfully takes a pragmatic approach in asserting that there is no silver bullet solution, the use of smartphone technology is probably the most feasible," he says. "As smartwatches and other wearable devices with the same capability become more of a commodity, patients will expect these devices to be involved in the process. Smartphones can also act as an additional source of demographic data by engaging with the patient directly. The EMPI [enterprise master patient index] can manage this external data just as it manages external reference data by using its rules engine to enforce a trust policy while it automatically constructs a single best record for the patient."

Associated Costs
Patient-empowered matching solutions are promising, but they do bear some costs.

"Our recommendations are to develop the technology, workflows, and best practices, but we don't know for sure how much that will cost," Rudin says. "Hardware and software will need to be upgraded and, with the latter, there are development costs, which are likely minimal. The hardware, though, can simply be an off-the-shelf camera and a reader for your phone; an existing credit card reader may also work. The most immediate costs are essentially developing the technology a little bit more, which are not high. Greater expenses may come when this process is scaled up to a whole health care system."

Cidon offers, "With good usage of an existing infrastructure that the industry has been investing in over the last 20 years such as integration engines and EMPI technology, the other things are just incremental costs. They have to be used in an intelligent way to leverage the infrastructure. But that can be the danger of a new technology—it has to be utilized in the context of what is available infrastructure-wise."

Referential Matching
Rather than directly comparing the demographic data from two patient records to assess a match, referential matching matches demographic data that are input at the health care facility to its comprehensive and continuously updated reference database of identities. "Referential matching is almost the opposite of what we included in this study," Rudin explains. "It's a process that currently doesn't involve the patient at all."

However, LaRow believes that referential matching could complement many of the solutions proposed in the RAND study. "The beautiful thing about the RAND study is it is simple and practical, and there are many promising things that have come out of it," he says. "For instance, RAND recommends having a phone number for the health care facility to call back to verify at registration, two-factor authentication to ensure it's really you. Or to convey your identity to the registrar even more assuredly, a mobile app can be used that can automatically transfer your information to the registrar."

LaRow explains that since medical records are tied to demographics, referential matching can synchronize the billions that are scattered across health care facilities.

In a referential matching system, a medical facility that would ordinarily store patient demographic information in a conventional master patient index (MPI) would instead store its information in a referential matching MPI. Theoretically, this is a more accurate method linking patient data and would correct any demographic data errors found in the records.

"In this way, we offer a more precise patient database for the hospital. We can help that hospital translate the patient demographic information into an accurate and complete patient medical record," LaRow explains.

LaRow says that referential matching has a higher accuracy rate than currently exists in a typical MPI. "Conventional MPI systems can give you 70% success rates. When a hospital can't decide if they have the correct record, they send the record to data stewards, something that occurs about 30% of the time. These data stewards use intuition or perform research like internet searches, which can bring the accuracy closer to 90%. However, [referential matching] boasts an even higher success rate of 95%."

LaRow believes the referential matching process and the patient-empowered solutions suggested by the Pew/RAND study complement each other. "With the solutions that the study offers, the errors that happen at the registration desk can be greatly reduced, whether patients use biometrics or an app to transmit the information to an offsite database for matching. Just the change in the process at patient check-in will reduce the errors by 50%," he says.

The Patient's Role and Response
Rudin and his team discovered early on in the study's progress that most patients are not thinking about medical record matching. "If they did, they might want to be more involved," Rudin says. "We tried to think of what we could give patients that would give them more value, like convenience, and better record matching would come as a byproduct of that."

The team learned that if there is no energy toward developing patient-empowered solutions, the tools likely will not come to fruition. "We'll then have to try approaches that don't involve the patient as much. Health care capabilities don't just appear. Someone has to pick up the ball and move them forward or they will sit there as an idea in concept and never make it to reality," Rudin explains.

"I think the technology the report suggests would go a long way to improve patient satisfaction, but it's a bit of a mixed issue," Cidon says. "In terms of biometric solutions, user anxiety among patients is common. For instance, palm and fingerprint scanners require individuals to touch or interact with devices that other patients have been touching, evoking hygiene and other concerns. With social media, though, patients are more comfortable exposing themselves to a larger extent. Asking for phone numbers is mild compared to what individuals generally post on social media sites."

Patients have their own opinions on the initial solutions proposed in the Pew/RAND study. One patient expressed mixed feelings about using a smartphone for health care purposes. "I feel like since everything we do is through our smartphones—money handling, media, there are so many things that you can do now—I think that if doctors asked us to do that, it's not bizarre to me. It's the way society is progressing," she said. "But I would have questions about whether or not they have access to other things on my smartphone. Is it possible for someone to hack or impersonate me? What happens if I lose my smartphone or it breaks vs being able to write down my Social Security number at the reception desk? Will the old way just be gone completely? Or would losing my smartphone cause me to not be able to go to the doctor? If it's just checking in, then it's OK. But if it's something that also requires credit cards, for example, then it's different."

Another patient was more enthusiastic about the prospect of having her smartphone be an integral part of keeping her health care record as accurate as possible. "I would love to use a cell phone to check in at a doctor's office," she says. "I use Kaiser [Permanente] now and everything is accessible through an app, and I think it is great."

Clear in her preference to avoid using technology in a health care setting, another patient opined, "I think I would prefer the personal touch when visiting the doctor's office. Check-in with a smartphone may risk patients feeling like they are being reduced to a digital entity."

Another patient expressed his frustration at the use of technology across the medical experience. "I used to go to my doctor, and I would tell him all these private things about myself I never told anyone else," he said. "He would watch me with interest, occasionally jot down a note, and encourage me along. The last few visits, however, he sat facing his computer, clacking away at the keyboard, not looking at me anymore. It was very disconcerting. I wondered if he still cared about what I was saying, if our discussions were private or, even more frightening, what might be done with what I said. Needless to say, I resent being itemized on a smartphone, computer, or otherwise. As I get older, I want to walk into a clinic and talk to a face with a name and to be shown some measure of human compassion. It gives me comfort; I feel private, important, and somewhat dignified."

For those patients who are not ready to embrace technology as a necessary step in ensuring accurate medical records, Rudin offers hope. "This study, we hope, will not be the end of the conversation. Instead, we want it to be a catalyst for solutions," he says. "We want people to publish their views and argue about this issue; the study is meant to spark further debate."

— Susan Chapman is a Los Angeles-based freelance writer.