We’ve heard the stories and read our facility’s reports on patients receiving treatments based on information in someone else’s health record. The frequent occurrence of duplicate, mismatched or overlaid patient records has lead to medical errors of varying severity. Despite the proliferation of electronic health record (EHR) systems, advances in integration capabilities and numerous technical and procedural checkpoints to avert medical errors, problems still occur at a disturbing rate, emphasizing the need for more sophisticated approaches to match and link the right patient to the right data.
A client shared a real situation that illustrates the point: A provider searching for a patient record chooses one from a list of candidate records on a screen. If the patient is present, the provider can ask a few questions to verify that the record belongs to that patient. Based on the information provided in the record, a medical care decision is made and medication is prescribed. However, that particular patient record, though accurate, is incomplete. The patient’s allergy information is contained in a separate record created at the time of registration and assigned a different medical record number. When the patient takes the newly prescribed medication, it causes a reaction that requires a visit to the emergency room.
The reliability and consistency of patient information is also a contributing factor to the success of today’s efforts to improve patient care through greater collaboration. The challenge, however, is that different settings use different patient identifiers. As entities share more EHRs with accountable care organizations (ACOs), health information exchanges (HIEs), patient centered medical homes (PCMHs) and other connected care initiatives, the opportunity for data mismatches increases. What was once an error limited to a single setting can now be propagated across the continuum of care and corrupt data in other systems. Healthcare entities must therefore undertake a rigorous review of their current patient identification procedures and data governance policies to ensure data integrity.
The accurate identification of a patient is not a trivial task, and it becomes more complex as the number of applications, facilities and data exchange partners increases. One reason for the identity challenge is the multiple points at which patient information is collected. Each entry point is an opportunity for data entry errors, such as misspellings and transposition of characters. Another reason is the normal change in demographics, such as residential moves and changes in employment and insurance. The various permutations of names can be an especially difficult issue, including aliases, nicknames, hyphens, spaces and unfamiliar spellings. Senior and junior designations and multiple births require extra attention.
The result of these complexities is often duplicate patient records. In one scenario, multiple medical record numbers are assigned to the same patient, essentially fragmenting the patient’s clinical information among several, disparate records. In another scenario, two or more patients can be assigned the same medical record number, causing clinical information for both patients to be combined.
The most common approach to patient identity management utilizes the master patient index (MPI) of an existing core application, such as an EHR, hospital information system (HIS) or practice management system. The MPI assigns a patient identifier and keeps track of patient demographic information known to that system. Sometimes, the MPI can also store information on a patient that originated in another system, but that depends on the interface capabilities of the applications.
Since these MPIs are most often just an embedded infrastructure component designed for the host application, they provide only basic capabilities and are not sophisticated enough to handle the requirements of cross-application or multi-facility patient identity management. When trying to determine if a patient has presented before, the organization lacks the advanced query, match and data stewardship features needed and, as a result, suffers from the convenient creation of duplicate patient records and multiple patient identifiers.
Additionally, these embedded MPIs are inadequate to manage more complex situations, such as the addition of patient populations from an organizational merger, the intentional obfuscation of personal details by drug seekers, imposters seeking care under someone else’s credentials and the accidental hijacking of an existing patient record through a perfectly normal interaction.
In the latter case, a staff member pulls up a record they believe is associated with the patient, but in actuality it is not, and makes a change to the record. For example, a patient named Eva Rodriguez calls her doctor’s office to schedule an appointment. The scheduler locates a record for an Ava Rodriguez, and verifies her phone number and insurance. The insurance happens to be the same, but Eva provides a new phone number, which is corrected on the patient record. When Eva arrives at the appointment, the same record is retrieved, and more verification questions are asked, resulting in a change to the address field. When Eva meets with her physician, the physician asks detailed clinical questions based on the information in the record, and they both discover that the record is not Eva’s, but Ava’s.
The record has been “hijacked” and unintentionally changed over time. Fortunately, the misidentification was caught in time to avoid any adverse treatment, but there are still consequences: the patient-provider relationship was breached when Ava’s personal health information was discussed, the physician’s confidence in the technology erodes, the health information management (HIM) staff spends hours returning the record to its original state and the patient questions the integrity of the organization, impacting customer satisfaction.
In addition to the medical repercussions of failing to identify patients accurately, there are other ramifications that carry significant costs in time, resources and money. The plague of duplicate patient records triggers repeat testing, redundant administrative activities, time wasted to investigate and resolve the records, inaccurate billing, malpractice settlements and public relations issues.
My provider organizations have between 8 and 13 percent of their medical records duplicated – and sometimes as high as 22 percent. In multi-facility environments, where disparate application systems are integrated, the percentage of duplicates can surpass 30 percent.
Fortunately, there is an approach that has proven successful over the years to address the challenges in patient identity management and reduce the creation of duplicate records. A vendor-neutral enterprise master patient index (EMPI) is designed to compare patient records from disparate systems and indicate the probability that two records are duplicates, unique or potentially duplicate. By using probabilistic matching algorithms that incorporate phonetic similarities, variances in typographical entries and dates, aliases and the relative “weight” of specific comparators, an EMPI significantly improves the patient identity process.
The EMPI assigns a unique identifier for the patient and maintains a list of all other identifiers associated with that patient, such as multiple medical record numbers (MRNs), enabling a translation or a “map” from one identifier to another. This feature allows external systems to share information for the same patient, even though each system is only familiar with its “local” identifier. Additionally, the EMPI can aggregate data from the patient’s different records into a single best record, creating a source for consistent, reliable patient demographic information for use across the enterprise.
Organizations derive maximum value from an EMPI when it is incorporated into front-end processes, such as patient registration. Instead of searching only the local MPI for a patient record, a registrar can search the EMPI, taking advantage of its advanced query capabilities and larger patient population, and determine if the patient has been seen at any facility previously. The patient’s MRN is available, which eliminates the need to create a new one. As an added bonus, demographic information can be updated while the patient is present.
Imprecise patient identity management leads to fragmented data, mismatched and corrupted patient records, and diminished data integrity, all of which contribute to patient safety concerns. Without a sophisticated method to compare records from any system and provide a consistent and reliable source of patient information across the enterprise, the business and effectiveness of delivering quality healthcare is significantly impacted. The use of an EMPI, however, provides precise matching technology that identifies a higher rate of potential duplicate records, enables a patient-centric view and reduces a major cause of unnecessary complications and poor quality of patient care. HMT
About the Author
Richard Garcia is industry solutions strategist, NextGate. For more on NextGate Solutions: www.rsleads.com/303ht-204