Mergers, MPIs, Oh My!
By Sarah Elkins
For The Record
Vol. 31 No. 7 P. 10
What happens to master patient indexes when health care organizations meld into one?
For the past decade, hospital mergers and acquisitions have been on a steady incline. In 2017, it was another record-breaking year, continuing a trend that’s been ongoing over the past decade. There were fewer mergers in total in 2018, but the size of those mergers was larger, making it the year of the “megamerger.”
The merger and acquisition (M&A) wave, spurred by rising health care costs, shrinking profit margins, and the drive to participate in value-based reimbursements, seems, at least for the moment, to have calmed a bit. An article on HealthCareFinanceNews.com called the first quarter of 2019 “sluggish.”
However, this may be the lull before the next onslaught of mergers.
The good news is the health care industry has gotten better at navigating the M&A process. For those involved in the details—hospital directors and vendors—it isn’t their first rodeo. Here, four experts lend their hard-earned expertise on perhaps the most challenging and critical part of the merger process: taking care of the master patient indexes (MPIs).
When to Prepare
The recommendations are clear—when it comes to preparing the MPI for a merger, there is no such thing as starting too early. The issues that arise during the merger of two or more MPIs are more complex and varied than a single-facility system conversion. The thought that it will take the same amount of time is misguided.
“We encourage [our clients] to discuss and develop a strategy way in advance—12, 15 months, even two years out, ideally,” says Karen Proffitt, MHIIM, RHIA, CHP, vice president of consulting services at Just Associates, a health information data integrity consulting firm. Yet, when asked how often a client seeks the help of a consultant a full two years before a merger, Proffitt says that happens “very little to never.”
For many hospital leaders involved on the front end of the merger, the MPI data is out of sight and out of mind.
“People don’t think about the MPI data. They’re so focused on the clinical data,” says Letha Stewart, MA, RHIA, director of customer relations at QuadraMed, a health care identity consulting firm. Like Proffitt, Stewart agrees the best start is an early start, one that leaves enough time for unexpected obstacles along the way.
“It’s hard to do something well if you don’t have enough time to do it,” Stewart says.
The Road to a Successful Conversion
Beyond allowing enough time to accomplish a clean data merge, there are other key steps to ensure a relatively worry-free MPI conversion. Following these steps will prevent last-minute setbacks or less-than-perfect outcomes that can’t be remedied once a system merger or conversion begins.
Map the Current State
Leadership will often begin discussing the future state of merged systems without a clear understanding of the current conditions. To rectify this situation, Proffitt recommends each facility create a current system diagram. She stresses the importance of not forgetting about the feeder systems, the EHR, and downstream systems such as radiology, lab, and cardiology.
According to Proffitt, those involved in early preparation should be asking: “How are [the integrated systems] interacting today, especially from an MPI perspective, with messaging? Who gets those ADT [admissions, discharges, and transfers] messages coming in or out? Who gets updates? Who gets a merge message? Who’s actually processing that message?”
Only after the current situation is understood and documented should the desired future state be addressed. Without a clear picture of how information flows, it’s difficult to determine which systems should remain in place after a merger and which ones should be retired, Proffitt notes.
Invite the Right People to the Table
Once all affected systems have been identified, it is imperative to identify the key stakeholders associated with their operations within each facility. From there, the team that is going to work together through the entire merger process can be established.
Ensure there is representation from the HIM and IT departments from all facilities. The formation of this taskforce will make all the difference in avoiding problems down the line.
“One of the trends that we see is that IT ends up making a lot of these decisions early on and they don’t always include people from the HIM departments who are really the stewards of the data and know more about what they will need going forward,” says Rachel Podczervinski, MA, RHIA, director of identity solutions at Just Associates.
The key analysts of the downstream systems are frequently overlooked during the creation of the MPI taskforce. That is why mapping the current state is so important before deciding who needs to be at the table. A forgotten system equals a forgotten perspective in ensuring a smooth MPI transition.
Another key element of building an effective project team is making sure each individual facility involved in the merger is represented.
“When you talk about merging vs converting systems, everything is multiplied by how many facilities you’re bringing together. All that data coming in from different sources is going to look a little bit different,” Podczervinski says.
Get an Analysis
After a path forward has been mapped by a thoughtfully built team of stakeholders, the next step is to get a clear understanding of the data within each system as well as across the systems. It’s particularly vital to be aware of system overlaps.
“The overlaps across systems are likely to be very high, and if those are not addressed prior to the conversion, then those will turn into duplicates,” says Stewart, who mentions this often occurs when a hospital acquires nearby physician practices. “Probably 80% of those physician practice patients are already in the hospital’s system.”
It is uncommon to encounter a duplicate rate of less than 50% when two or more facilities’ MPIs are analyzed. Just Associates recently completed a project in which it encountered a 70% duplicate rate.
These astronomical overlap rates don’t shock data integrity vendors because they expect to see a lot of the same patients represented in the MPI of two merging facilities. Most merging facilities are geographically proximal and typically share referrals.
“[Overlaps across systems] aren’t technically errors. If you’ve got three hospitals and I’ve been to all three and all three have different systems, I should be in there three times,” Stewart says.
Overlap records across merging systems become a problem only if there is no plan to link the records after conversion. The one patient who has been seen at three different hospitals will become a duplicate record upon the merger if there is not a mechanism in place for linking the records.
Whose MPI Survives?
During analysis, overlaps are not the only variable to assess. Stakeholders should also be looking across systems to see whose data quality is best. In this case, “best” means cleanest and most current.
That information may be used to determine whether one MPI is integrated into another facility’s EHR or vice versa. In some cases, older data may be archived without converting them to the new system.
Daniel Flowe, MS, RHIT, CHPS, HIM manager of use and disclosure at Norton Healthcare, assists with several MPI acquisitions each year. And, while an acquisition can look a little different from a merger, he says, “In an acquisition, the acquiring organization maintains its EHR and MPI and integrates the MPI of the acquired organization. In the case of a merger, however, I would recommend using the larger MPI, provided there is a reasonable confidence level in the accuracy of the larger MPI.”
This is a best-case scenario, Podczervinski notes. “Our general recommendation is that it’s best for the whole organization to be on the same system because that allows for continuity and limits redundancy,” she says.
Flowe says, “Often—but not always—the larger organization has greater resources to leverage identity management processes and technology and implement best practices. As such, the larger organization’s MPI usually absorbs the smaller MPI.”
There are cases, Podczervinski adds, in which it’s “not viable for the merging organizations to all go onto the EHR. In that case, we recommend they make sure they have at least one place where there is a higher-level [enterprise] MPI,” Podczervinski says.
Stewart recommends creating a combined record with the best data from each system. For example, if a patient exists in two MPIs, he or she will carry the home address from the most recent encounter into the new system. Similarly, some large EHR vendors recommend not converting patient data that are older than a couple of years. Instead, the data are archived and a link is provided in the MPI.
“Now, from an HIM perspective, I have some issues with that to a certain extent because we know from the analysis we do of data that people who haven’t been seen in five years will show up tomorrow. People who were here 10 years ago will come back this year,” Stewart says.
Flowe notes that when an organization opts to create an archive database, it’s important to address access issues. “After creating a legacy database, additional options such as interface or individual user access to the legacy record should be explored to optimize operations,” he says.
Different Facilities, Different Concerns
The complexity of the MPI conversion can depend on the types of facilities being merged. For example, a children’s hospital will likely have fewer patient matching issues if it is being acquired by a nonpediatric health care organization.
“The pediatric and the adult provider services rarely overlap,” Flowe says.
Stewart adds that specialty organizations, such as a children’s hospital or a cancer clinic, have lower duplicate rates in the first place. “They have a smaller population of patients. [The patients] are often known to them, and the registration process is more focused compared with a large emergency department,” she says.
Overall, the issues that make converting a single MPI into a new EHR challenging are the same issues that make merging MPIs challenging. Having a large number of birth records without a tertiary unique identifier or being located in a region with a large immigrant population where people might be afraid to share accurate information are some of the issues that might make assessing and merging data difficult.
However, according to Stewart, “What really impacts it more is the size of the organization and how many databases they’re trying to combine into that EHR.”
Recently, she was converting a group of five hospitals and 80 physician practices, all on various systems, to Epic. The number of systems involved was the complicating factor more than the types of facilities involved. Questions such as “Who goes first?” and “While half the facilities are on the new system and half are on the old system, how do you handle billing?” became top of mind.
The human element can also pose challenges to a successful merger. For example, with the merger of disparate data comes the merger of disparate HIM departments with their own sets of rules and policies.
“We end up doing a little bit more consulting and help in strategies to bring those two groups together, help them make decisions for the organization as a whole,” Podczervinski says.
Common Conversion Errors
Once the team tasked with overseeing the merger of several MPIs into one system has completed the many months of preparation, the actual conversion can still encounter any number of unforeseen issues. In most cases, this is not a reason to panic. Most large vendors employ several test or clone environments and load the data multiple times to find errors before they occur in the live production environment.
“That’s why they want the MPI loaded six weeks or so before the actual go-live date because they want to make sure it’s right before they load the clinical data on top of it,” Stewart explains.
Podczervinki adds that it is typical for records to “error out” as they move into a new database. There are several reasons this might occur. A common case is that one system might store the complete patient name in one field while another system will have separate fields for first, middle, and last name. A seemingly small difference in data sets can spell major trouble if it’s not addressed prior to conversion.
Additionally, “autoreconciliation processes might fail because of flags,” Podczervinki says. “For example, if one system has a flag for deceased and the other doesn’t, the records may not autoreconcile properly.”
In the end, the success of an MPI merger is only as successful as the alignment of the operational structures and decision making that happens before the first line of data is converted. And while it’s important to have an engaged team composed of all the stakeholders, having a “buck-stops-here person who says, ‘We’re going to do it this way’” is equally necessary, Proffitt says.
— Sarah Elkins is a West Virginia–based freelance writer.