Cooperative effort helps Dallas-Fort Worth area hospitals ease data burden
Cooperative effort helps Dallas-Fort Worth area hospitals ease data burden
Data Initiative helps ensure clean, consistent data
It seems the data requirements of hospitals keep growing and growing. But follow the lead of a group of 51 Dallas-area hospitals, and you might see that burden ease.
Along with the clinical data for Medicare and Medicaid, hospitals have to submit quality data to the Joint Commission on Accreditation of Healthcare Organizations for the ORYX program. There are state data requirements and even local initiatives that put a paper burden on hospitals, eating into time and money pools. While coming to grips with this growing problem and the need for high-quality, standardized data, the Dallas-Fort Worth Hospital Council (DFWHC) created its Data Initiative. The council represents 83 hospitals in the area, 51 of which participate in the Data Initiative.
The council was created to promote responsibility to patients and communities served by its member institutions, in part by using data to measure value, facilitate evaluation of health care quality, and promote quality improvements.
Prior to the Data Initiative, one of the key projects of the DFWHC and member hospitals was the Healthcare Value Initiative, a collaborative effort of the purchasers and providers to measure, track, and improve the quality and cost-effectiveness of health care in the community.
"We have a very unique market," says Denise Remus, PhD, RN, vice president of the Data Initiative. "What works in other areas isn’t seen as automatically working here. So when we wanted to have cleaner data, and open, nonproprietary risk-adjusted data at that, we felt we had to start from scratch."
The Data Initiative was created so the hospitals involved could be sure that their council data, as well as state and ORYX data, were clean and consistent. A side benefit would be that members could use each other’s data to compare themselves without worrying that they were relating apples to oranges.
After several approaches to quality measurement were attempted, including contracting with an outside vendor for data analysis and reporting, the Data Initiative has taken on every aspect of data processing for participating hospitals. Members said this was the best way to ensure access to accurate and timely data, says Remus.
To enable a smooth transition and make sure the DFWHC could continue to meet members’ data needs, Remus says she looked for a vendor to collaborate on development of a software product to validate all aspects of data quality. That alone was difficult, she notes, because so many companies are anxious to provide Internet-based information.
Until there is better security of the sensitive data incorporated in most health care reports, Remus says the council will stick with the old-fashioned desktop versions.
A little give, a little take
"I knew that the standard of data quality is something that other states are struggling with, so I felt that if we could bring this to the table, the right vendor could go out and sell this to others. Then we would also have access, for a fee, of course, to data from other regions that would have the same level of quality as the data from our own council," Remus says.
Remus says she didn’t have a lot of money, but could provide her time, knowledge, and expertise as well as that of the councils and member hospitals.
"Whoever we chose would have the programming skills and related experience with hospital discharge data. We would have the rest, and they would benefit because what results from the collaborative effort should be in every hospital across the nation. The product will enhance the quality of data, and that means that you can do more with it," she explains.
Remus met with representatives of four companies and discussed her vision of collaborative product development with each one. She chose the MEDSTAT Group of Ann Arbor, MI. "The primary reasons were their expertise in product development, experience with large data administrative and clinical data sets, and interest in promoting the quality of administrative data," she says. And it didn’t hurt that they were willing to partner with the council.
After a few conceptual discussions, Remus provided MEDSTAT with technical specifications in late 1999. Through conference calls, Internet-based exchanges, and meetings, MEDSTAT was able to develop a product. The project development has been both challenging and exciting, says Remus.
Tool validates data quality
"It is a pleasure to work with such a talented, creative group of individuals," she says. Currently, the product is in beta testing, with plans to release it to all of the Data Initiative participants late this summer. The resulting product has created a standardized format for data that provides the information required for the Joint Commission, the state of Texas, and the DFWHC with zero errors, Remus explains.
Included in the data are all of the administrative data that are on a typical patient bill. "We have discharge data from 1997 forward and hands-on knowledge of it," she says.
The data tool validates the format and quality of the data submitted, says Wendy Richardson, product manager at MEDSTAT. "It relieves the hospital of having to resubmit data that are formatted incorrectly or are wrong. And the resource consumption also moves away from the hospital."
The DFWHC board and executive committee, as well as hospital chief financial officers, chief executive officers, and physicians help the council prioritize needs and resources. For instance, in the last nine months, the majority of staff time was spent working through issues associated with the newly mandated Texas hospital discharge data program requirements.
"That way, we could assure our members that the data required are something that will be meaningful to them," explains Remus.
Although having administrative data may not seem like much to work with, Remus says a lot can be learned from the information that appears on a bill. "You have clinical data including the nine diagnoses and six procedures codes, patient demographic data, and the physicians on each case," she says.
"Using these data, along with risk and severity adjustment, we can, for instance, tell a hospital what their expected and observed rates of C-section are. They can see both their actual data and how we reported it. That will help them evaluate why a patient might be at risk for a specific outcome," adds Remus.
Co-morbidities, patient race, number of previous births, and a history of previous cesareans may all be factors, she adds.
"We use a logistic regression model to predict what diagnoses and procedure codes put someone at risk. So a woman — who never had a C-section — coming in with the third baby won’t likely be at risk. But someone who has diabetes or hypertension might," Remus says. "When the hospitals use their medical records to evaluate why an observed outcome was different than predicted, they may find additional clinical data to explain the variance. They may identify a diagnosis code we hadn’t considered previously, which could be evaluated as a risk factor in the next regression model. Or it may be clinical data we cannot capture that remain a limitation of using administrative data for quality measurement."
Being aware of problem areas certainly can focus the attention of administrative and clinical staff. Remus notes that one hospital found in its last report that its acute myocardial infarction mortality rate was significantly higher than that of the other council facilities. "We looked at it closely and drilled down to find that they had a higher percentage of deaths in diabetic patients than the rest of the community."
Diabetics have become a new risk group for that facility. It is looking at its practice guidelines, diabetes management program, and plans of care. The hospital can look and see if it is one particular physician who needs assistance, or something else. "The goal is to have a better understanding of the variances. They have started to collect more data on diabetics, and we hope that the next report will show some improvement."
Even if the council reports can’t provide the nitty-gritty detail needed for quality improvement, adds Remus, the facilities can be spurred to do so.
Outpatient data may be added
The costs to the member hospitals had to be less than they would spend if they were doing the three separate projects themselves. So far, the council has been able to achieve that, even lowering prices over the last year. The price paid depends on the size of the hospital and its number of discharges. The minimum cost last year was $4,500, and the maximum was $25,000. That price will be less again this year and in 2001.
"There is a real sense of ownership among the hospitals," Remus says. "They have representatives on all of our committees, and they don’t see us as a vendor. They can really control what we do."
Although the data right now are primarily administrative, there might be the addition of outpatient data in the future, as well as additional analysis to generate longitudinal indicators to help facilities see what happens to patients they transfer out, such as premature babies. "We are driven by our membership and responsive to what they want us to do," says Remus. "But I don’t think that we will look at duplicating some of the clinical data that others do."
Remus says the kind of quality the council members get in their data now is something that can and should spread throughout the country. But she has a warning for those who want to mimic what she has done. "Technical people will tell you that data are data are data. I believed them, but it isn’t true. Health care data are different. There are idiosyncrasies of diagnoses and procedure codes. The attending physician isn’t the same as an operating physician. Sometimes getting them to understand the need to exclude certain patients from analysis is beyond their understanding."
Part of the problem, continues Remus, might be that health care came late to technology. "We really don’t have as many data experts as we need. So if you can’t find someone who has worked in health care, be sure you have the time to dedicate to the person you hire, to help him or her learn the ins and outs. You have to have a clinician or coder who understands what the data elements mean to help that [information systems] person set up the data warehouse effectively."
Remus adds that creating a good, functional data warehouse is harder than many will admit. "We still can’t easily do an ad hoc report. We can’t evaluate data as quickly as we would like. And the company we hired at the start told us what hardware we needed for the next two years, but I had to upgrade after only four months. To manage this large set of health care data is very, very difficult."
[For more information, contact:
• Denise Remus, PhD, RN, Vice President of Data Initiative, Dallas-Fort Worth Hospital Council. Tele-phone: (972) 717-4279.
• Wendy Richardson, Product Manager, MEDSTAT, Ann Arbor, MI. Telephone: (734) 913-3000.]
Subscribe Now for Access
You have reached your article limit for the month. We hope you found our articles both enjoyable and insightful. For information on new subscriptions, product trials, alternative billing arrangements or group and site discounts please call 800-688-2421. We look forward to having you as a long-term member of the Relias Media community.