Masters of Their Domain
August 15, 2005
Masters of Their Domain
By Robbi Hess
For The Record
Vol. 17 No. 17 P. 30
A well-kept master patient index helps healthcare organizations take control of patient information.
People, processes, and technology are the three components necessary in getting-and keeping-a master patient index (MPI) viable and, more importantly, clean. The ongoing battle to maintain patient data integrity is nothing new to the healthcare industry and many have taken the necessary steps to manage the data. But in an industry rife with consolidations, mergers, and off-site practitioners, the challenge of tracking patient information has become increasingly more difficult.
Most facilities have acknowledged that there is a problem with data integrity. The most common form of dirty data is duplicate records. Stefanos Damianakis, PhD, president and CEO of Netrics, says that as MPIs get larger, the data quality suffers more.
"If you look at a small database with 500,000 records, there will likely be between a 5% and 12% duplication rate," he says. "You would expect that as things get larger that measure would stay the same, but it grows exponentially."
In a single entity-a non-multihospital setting-with a database of 1 million records, Damianakis says there would be a duplication rate of at least 10% to 20%.
"The hypothesis is that as the database gets larger, it reflects a larger organization that will have more people interacting with the database. And having more people involved in tfor the injection of human error and thus data quality issues," he explains.
Beth Just, president and CEO of Just Associates, elaborates that other forms of dirty data include overlaps (one person has more than one enterprise identifier across an enterprise MPI), overlays (one person is assigned, in the MPI, to another person's record) and erroneous, invalid, and/or default data stored in key identifying.
"There are any number of reasons that duplicate records occur," she explains. "During the patient registration process, human errors can occur, including misspellings, missing data, use of nicknames, typos, and transpositions."
Attempting to merge disparate systems as a result of facilities being bought and sold results in data sets being merged and duplicates being created. Just and Damianakis agree that as MPI data integrity becomes compromised, it is easier for more duplicates to be created.
"An MPI is a great thing but it's always going to be a moving target," says Mark Boyce, president and CEO of Scribe Healthcare Technologies. "It's an ongoing battle that needs the enlistment of people and the implementation of tools to maintain the information electronically."
Damianakis says that when a patient came to the hospital 20 years ago, a medical records clerk pulled open a file drawer and found a patient's records.
"If the name was misspelled Ôa little bit' or the address wasn't completely accurate, the paper record could still be found," he says.
Today, with most software, even if the patient is in the database, if the information is slightly off, that patient will not be found and in all likelihood a new record will be created. A larger problem with that, according to Damianakis, is that "in my last record, it showed I was allergic to penicillin, but this record doesn't have that information. With the creation of the duplicate record, patient safety is at risk."
According to Just Associates, the direct cost of leaving duplicates in an MPI database is anywhere from $20 per duplicate to several hundred dollars. The lower cost reflects the organization's labor and supply costs to identify and fix the record while the higher expense reflects the costs of repeated diagnostic tests done on a patient whose previous medical records could not be located.
"Liability exposure could be significant if patient care is compromised because a healthcare facility can't find a patient's record," Just says. "Inappropriate billing cycles can occur and revenue cycles are affected."
The AHIMA estimates that it costs between $10 and $20 per pair of duplicates to reconcile the records. If the records aren't reconciled, however, the costs are even higher.
"The real cost of a duplicate record can run up to several hundred dollars or even more in terms of repeat diagnostic tests, operational inefficiencies, revenue cycle impact, and legal liability," Just notes.
Damianakis says dirty data inefficiencies are evident in every area of healthcare, from unnecessary costs of duplicate mailings and poor customer support to fraud.
HIPAA concerns pop up as well and could cause a healthcare organization liability costs for failure to ensure patient information privacy.
Boyce says the cost of putting together a viable MPI is a daunting challenge made even more so because the records may not be in an easily accessible format.
No matter the price tag, Damianakis feels it's money well spent. "If you ask a doctor, one extra duplicate is one too many. What if that was your child who had the duplicate medical record and their care was compromised. There is no price tag that can be attached in that instance," he says.
In the Beginning: People, Processes, and Technology
The problem with dirty data is dual-pronged: poor search accuracy and poor data quality. That starting point leads to the correct record not being found, a duplicate record being created, and the cycle beginning again.
To gain control of an MPI, start with information input. To correctly link records when patients come in the door, Just says, those at the registration desks must ask for more identifiers other than name, address, and date of birth.
Boyce agrees, saying that a healthcare facility must choose what they will use as their master system to identify patients. Some argue that using a Social Security number as an identifier is the way to go because patients are accustomed to using that number. Others feel that Social Security numbers should remain out of this arena and a new identifier should be implemented.
"Will it [a new identifier] just be name, address, Social Security number, or a unique combination of that information and the birth date? Whatever a facility decides, at the end of the day when the rubber meets the road, you need to be able to accurately control information that is input," Boyce says.
The trend, he continues, is toward installation of patient portals. With a portal, he explains, a patient has online access to his or her medical records and can also preregister for visits and surgeries.
"The patient needs to play a more active role in maintaining [his or her] information," Boyce notes. "We need to put the patient in the driver seat to get the information updated. We can't keep putting more work on the people who sit at the desks who are already overworked."
Using patient portals, patients can perform such functions as address and name changes-all the things that would normally occur to maintain and update the MPI on an ongoing basis.
Damianakis says it's inevitable that humans will inject errors into databases when they are inputting information. Policies and procedures definitely have to be in place to catch and prevent as many errors as possible.
"Most times when you deal with looking up information in a database you have to learn a special language," he explains, adding that Netrics software models human similarities. "The person looking up patient information inputs their best approximation of the name, Social Security number, or date of birth and our software pulls up the information that best matches what's been input."
The correct policies working hand in hand with the right software solutions, Damianakis says, play a critical role in eliminating duplication problems.
Software, such as Netrics', can mathematically model human similarity algorithms.
"Computers can look at two records and determine they are identical but if they are only similar, not identical, it will bypass that record," Damianakis explains. "What is needed is a computer software program that is empowered to locate and pull up those similar records."
Damianakis offers an example of when the creation of a duplicate record may occur. "There are two Smiths in the database and I went to the hospital and found two records. One is invariably a duplicate and if the medical records office has to make a decision on which record to use they will likely create a third record- just to play it safe. The patient is going to the emergency room and there simply isn't time to figure out which Smith he is and he cannot speak for himself."
In that instance, Damianakis explains, the importance of data quality is critical.
Healthcare facilities need to bridge the gap between the human tendency to be less than precise and the computer's built-in intelligence that is completely precise.
"There needs to be an improvement in what happens at the intersection between the human, the computer, and the database," he says.
Healthcare facilities will likely find themselves having to start from ground zero and get their dirty data and duplicates cleaned professionally before they can truly be up and running toward interoperability-whether hat interoperability is between connected departments within their own facility, in a rural healthcare organization environment, or on a larger scale (nationwide interoperability).
Boyce says there is no "glue that holds all the systems together. You almost need a separate system to maintain how all the other systems capture the patient information and eventually bring it to a central location."
To implement a truly effective MPI, Boyce explains, a healthcare facility must take it in stages.
"Find out what area has the largest churn of patients, take that system and start there establishing benchmarks on how to input patient information," he says. "Don't try to tackle the whole beast. Take what you want to be the master list, based on the largest volume of patients. Look for clues that will help work through all of the systems and remember to bring the patient factor into the equation."
Hospitals will eventually have to deal with their duplicate records and many of them outsource the information for analysis. Once duplicates are identified, the hospital has to ultimately determine how to deal with either the merging or deletion of its duplicate records.
Once the data is cleaned, it will take both front- and back-end software, user training, and processes that the hospital either has in place or must put in place, to keep the problem from manifesting itself again.
The importance of having clean data cannot be stressed enough, Damianakis says. "It's important to have not only because of patient safety but as we are moving slowly toward a national health number and more integrated delivery networks, the processes need to be in place locally before going global," he explains.
In a regional healthcare information organization (RHIO), three or four community hospitals share patient data electronically, attempting to merge various systems into one. Avoiding and eliminating duplicate medical records play a key role in running a successful RHIO.
Installing proper software and making certain that employees who come in contact with patient information know how crucial it is to input correct data every time are steps on the way to not only cleaning up data but also to establishing interoperability. Damianakis says taking these steps could reduce an organization's duplicate records creation by a staggering 95%.
"There's always work to do, but medical records departments in hospitals are thrilled if they can get a reduction in duplicate rates by that much," he says. "It makes the problem of the last 5% manageable. What we need to do now is get to that final mile."
- Robbi Hess, a journalist for more than 20 years, is a writer/editor for a weekly newspaper and monthly business magazine in western New York.