Software breakthrough makes data sharing easy
Program standardizes indicators
Scaling the Great Wall of China may be easier in some cases than removing the technical barriers to sharing outcomes information. Even if the information is comparable, different formats preclude data sharing and pooling. Those walls may be tumbling down. Quality and utilization departments developing databases and surveys can turn to new software to make sharing and storing data much easier.
The software, Outcomes Data Conversion Utilities, was developed in collaboration with Detroit-based Henry Ford Health System’s Center for Clinical Effectiveness; Alexandria, VA-based American Medical Group Association (AMGA); and Minneapolis-based Health Outcomes Institute (HOI).
Here’s how it works:
The software integrates a database of 2,000 coded outcomes research questions. A graphical user interface allows users to link their research questions, such as patient comorbidities associated with joint replacement, to the coded equivalent in the database. The result is a standardized format that can be transferred to other health care systems, third-party data pooling centers, and specialty registries. The data structure is not like a typical spreadsheet, where each column is a variable name and each row is a patient record or patient observation. Data are transmitted at the question level instead of at the form level. This is similar to how medical record data are stored and transmitted.
"It is a narrow and long file, instead of a flat and wide file," says Reuben Richards, MA, data analyst for AMGA.
As new measures are developed, HOI assigns them standardized numbers unique to those questions and puts them in the database. So far, outcomes measures for hip and knee replacement, hypertension, diabetes, asthma, cataracts, sleep disorders, and patient satisfaction are in the database. (See sample graphical interfaces, p. 148.)
The challenge to data sharing
A long standing problem faced by data collection professionals in general has been dealing with the various data entry and database management technologies of hospitals and clinics. "The main objective of this software is to provide some standardization to the growing field of data collection," says Reuben Richards, MA, data analyst for the AMGA.
Many hospitals face the same situation the AMGA faced. "We have all sorts of members who are competitors and noncompetitors all over the country, and they are all using their own internal data collection systems," Richards says. "It became difficult to merge these databases together with confidence."
The AMGA had been collecting uniform sets of data from its members on condition-specific areas, such as diabetes, which was one of the conditions that precipitated the software’s development. In the diabetes project, eight to 12 medical groups agreed to collect data on various aspects of diabetes such as diabetic functioning, quality of life, and other such indicators. Each one of those groups has its own information system. Some developed their own data collection systems, some bought vendor software, and others linked it to their lab data system.
"We were spending all our time trying to merge these data sets together," Richards says.
Two years ago, the AGMA started working on the data standardization project so the groups could:
• submit the questions they were asking and have them standardized to avoid extensive conversions after the data were collected;
• store the incoming data in the original form for future comparisons.
Now data collection projects progress more smoothly and the AMGA can focus on more important work, such as comparative analysis of compiled data. In a similar vein, hospital QI coordinators can collect data from various departments, affiliated clinics, and outpatient facilities with less time spent on converting data to one format and more time on evaluating the information.
Richards adds that quality improvement and patient satisfaction survey development are aided with Outcomes Data Conversion Utilities. "You can search the question database using key words such as leg’ and pain’ and it will return all questions matching the criteria you are looking for," he says.
To further complicate data collection efforts, outcomes data collection instruments frequently are updated after the initial round of data collection. Treatment technologies and methodologies change too, which alters what type of information is collected.
"Data set modifications can wreak havoc with data collection software, which is typically set up assuming that instruments would remain constant," Richards says.
Merging the old with the new is easy because the software views old versions of surveys as it views surveys developed by a vendor or hospital information systems department. For example, the AMGA started collecting data on hip replacement patients in 1989. In 1995, Richards started revising the data collection forms. Questions were eliminated, and some were combined.
Each question is assigned a standardized number. For example, the question on patients with a comorbidity of ischemic heart disease is numbered 104.353.3. (See sample graphical interface, p. 149.) When the subset of questions comes in, regardless of the form, questions concerning ischemic heart disease cormorbidities are matched with the same questions on the original data collection tool.
"It all comes down to storing the data in the original forms they are asked," Richards says. "When forms are modified or response categories are added to questions, you can still access the old data under the old format. . . . You can still trend data. You can still go back and look at change over time. Most of our outcomes studies go out three or so years. This at least allows you to maintain data continuity. This way, you don’t have to have a programmer come in and map those 25 items over to this new database."
When the content of individual indicators changes, the reliability of those indicators comes into question. Outcomes Data Conversion Utilities software allows users to come to their own conclusion about comparing a reformatted question with an old question or keeping that data segregated. "You can still make the decisions at the point of analysis as to whether it is comparable," Richards says. "Sometimes it is not, but you have that option, and you are not losing information."
[Editor’s note: For more information on Outcomes Data Conversation Utilities ($395 for members, $595 for nonmembers), contact the AMGA, 1422 Duke St., Alexandria, VA. Telephone: (703) 838-0033.]
You have reached your article limit for the month. Subscribe now to access this article plus other member-only content.
- Award-winning Medical Content
- Latest Advances & Development in Medicine
- Unbiased Content