Part one of a two-part series
Good system-level data depend on effective stakeholder alliances
Useful metrics within reach, but not global solutions
The ease of converting numbers into charts and graphs belies the difficulty of creating data that we can believe, explains Michael Victoroff, MD, medical director of the Colorado Division of Aetna U.S. Healthcare in Englewood, CO. "The better doctor is the doctor with the information. And, in this day and age, that’s computer information. Period!" Most health care professionals would agree that Victoroff’s statement could accurately be expanded by switching the word doctor to organization.
As with all things technical, computer-generated health care information is only as good as the human intelligence behind it. Solid support from the key players will make your project succeed. To engage them as competent partners takes skill building and rewards that speak to their needs. Data collection must serve the users, and not vice versa. That entails good judgments on where to compromise and where to hold the line.
Friends, foes — you need them all
To launch a successful data-collection initiative, rally the key stakeholders early. It’s this group that will determine whether you can collect your data in the first place and what you can do with it once you have it. "Involve those who want you to succeed, as well as those who do not want you to succeed," advises Denise Remus, RN, PhD, vice president of the Data Initiative at the Dallas-Fort Worth Hospital Council in Irving, TX.
If you get your allies and opponents to work side by side from the start, you’re less likely to hit an impasse six months into the project. "Ask your opponents for their opinions," she urges. Their resistance might stem from ruffled pride at not being asked for their expertise, especially if they’re experts or if they perceive that the outcome of the initiative will affect them.
To engage supporters and opponents in data collection efforts, Remus uses a straightforward technique: Find out what bugs them. "I don’t know one clinician or quality manager who couldn’t tell you at least one concern they have about their work," she says. If you can help them see how the process will address their concerns, it might give you a handle by which to pull them on board.
To quality managers, Remus offers this advice: "Before you ever organize a meeting, do a literature review on your subject. You don’t want to have people quoting a key study that you haven’t bothered to read."
While Victoroff observes that automated data collection can eliminate some of the painstaking process of alliance building, it will never be a complete solution. "The best data come from systems that collect them automatically without adding a second step, such as a lab order entry system," he says. The electronic medical record (EMR) is the obvious and "most needed missing element in medical record keeping today," Victoroff continues, but a universally useful design is yet to be found. Even when we reach that point, the EMR will remain a permanently evolving technology. "We will never have a perfect or complete EMR," he explains. There will always be extra measurement needs for specific studies because many automatic collection processes are not designed to study outcomes.
The key to fulfilling those extra needs is to "give immediate and compelling value to the reporting activity," he says. Show, for example, how it might improve the care of congestive heart failure patients. (For news of a project that offers compelling value for accurate, on-time reporting, see the related story, "Regional data initiative stretches hospital dollars," p. 78.)
Global issues still haunt experts
While the design of useful measurement tools at the organizational level is well-refined, Victoroff predicts that dilemmas at the macro level will dog the industry for some time. One issue is the absence of integrated information systems encompassing functions from inpatient, ambulatory, pharmacy, and lab to business office and supply. Even within the same organization, information tools are generally incompatible across functions. Even more exasperating, he notes, "to this day, the majority of systems were designed for financial billing. They were never designed to capture clinical data accurately. While existing systems do capture some clinical data, it’s flawed. And it’s enough to drive epidemiologists mad."
Take clinical errors, for example. The definition varies from system to system. To date, there is no standard for describing or reporting them in any useful sense, contends Victoroff. (For more information, see Quality Talk, p. 80.)
Other problems lie in the complexity of clinical data and the definitions of health and illness. "These are not flaws," he points out. "They’re legitimate scientific challenges in building useful systems of data collection, analysis, and reporting. The biggest problem with monitoring outcomes of care is in measuring the severity or burden of illness. It’s a difficult challenge in biostatistics. How do you assess how sick someone is?"
Some would measure it by functional capacity as shown in the ability to perform activities of daily living. Others would say it’s more accurate to assess the subjective impact on the individual’s personal sense of well-being because that often drives the motivation to adapt to health deficits. Other measures might include the number of dollars an individual spends on health care. "They all have some value," Victoroff says, "it depends on who wants to know and for what purpose."
One way to look at the affordability question is to consider how competently you could drive a car without a dashboard, or fly an airplane without an instrument panel. "How can we not afford to collect data?" asks Charles Kilo, MD, senior fellow with the Institute for Healthcare Improvement, Boston, and chief medical officer for MyHealthBank, Portland, OR. He adds that we must steer clear of an all-or-nothing approach to measurement. Stick to the middle ground where measurement is a value-added activity.
Among the typical objections to data collection is "We’re too busy." That in itself could be a clue to opportunities for process simplification. If, for example, staff are too busy fielding phone calls, Kilo suggests it might help to track the content of calls to determine whether employees inadvertently contribute to the volume:
• If many calls pertain to directions or parking information, a recorded set of instructions might handle them.
• For requests to speak to the nurse or physician, it might help to record the times the nurse or physician will return calls each day. Update the message as schedules change, and invite callers to leave their names, phone numbers, and questions.
Victoroff offers another perspective on the affordability question: "It’s generally accepted that most industries spend 15% to 25% of their budgets on informatics [information technology], while health care spends 3% to 4%. Whether it’s fair to compare health care with other industries is a matter of opinion, but there are other clues that health care vastly under-funds informatics."
Some of those clues are described earlier in this article in the discussion of fragmented information systems and the generally primitive state of the EMR, he notes.
In the past 12 to 18 months, the perennial question of who could or should fund health care informatics has drawn a brand new set of volunteers. A number of very large companies have stepped up to the plate, Victoroff explains. They include companies like WebMD/Healtheon in Atlanta; Medem, sponsored by the American Medical Association in Chicago; and Medscape in New York City. While he predicts that the field will shake down to perhaps two of the largest and strongest, the information offered will not be proprietary to any health care system. It will be available over the Internet to clinicians, researchers, regulators, consumers, and others yet to be determined.
As the revolution of widely accessible health information gains speed, however, it will pose new ethical and legal quandaries. And whether it will untie the knots in the validity and reliability of health care data remains to be seen, but only the naive would count on it.
Need More Information?
For information on getting best value from your measurement activities, contact:
o Judy Homa-Lowry, Homa-Lowry Healthcare Consulting, 7245 Provincial Court, Canton, MI 48187-2121. Telephone: (734) 459-9333. Fax: (734) 459-4234. E-mail: 105255.42@ compuserve.com.
o Charles Kilo, MD, Chief Medical Officer, MyHealthBank, Portland, OR. Telephone: (503) 228-3435, ext. 124. E-mail: [email protected].
o Denise Remus, Data Initiative, Dallas-Fort Worth Hospital Council, 250 Decker Court, Irving, TX 75062. Telephone: (972) 717-4279. E-mail: [email protected].
o Michael Victoroff, MD, Medical Director, Colorado Division, Aetna U.S. Healthcare, Englewood, CO. E-mail: [email protected].
o Duke Rohe, Performance Improvement Specialist, M.D. Anderson Cancer Center, Houston. E-mail: [email protected].
o Institute for Healthcare Improvement, 135 Francis St., Boston, MA 02215. Telephone: (617) 754-4800. Web site: www.ihi.org.
For more on software applications and research services, contact:
o MEDSTAT Group, 777 E. Eisenhower Parkway, Ann Arbor, MI 48108. Telephone: (734) 913-3000. Web site: www.medstat.com.
To browse the internet-based health care data providers as mentioned in the article, visit:
o Medem, contact through www. AmericanMedicalAssociation.com.
o Medscape, www.medscape.com.
o WebMD, www.webmd.com.
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