Aetna Us Health Care Articles & Resources

Building Blocks of Change

James A. Cowan, ., .H., heads the clinical consulting team, Aetna Integrated Informatics. The power of Aetna's data, he says, is precisely that it is claims data. "As a payer, we are in a position to collect and analyze data that allow us to generate specific answers to specific questions, ICD-9 codes indicate diagnoses, and CPT codes indicate treatment procedures. We know the age, sex and gender of the patient, and where and when the procedure was performed. We have in our data repository, again from claims data, information about lab tests and medications the patient uses."

Analysis of that data enables Aetna to answer two fundamental questions: What is the patient's condition and what treatment is being prescribed?

"To concretely identify a condition or disease, we require multiple 'hits' in our database," Cowan says. "With common diseases, we are right at least 95 percent of the time--up to 99 percent of the time. We are able to track adherence to a drug treatment plan, but we are limited in our ability to determine the efficacy of that treatment because often the best measures are biometric, such as lowered blood pressure, which we don't see in claims data."

The next step is to incorporate quality-of-care standards into the equation. For this, Aetna turns to external, authoritative sources such as National Institutes of Health, the American Academy of Cardiologists, the American Hospital Association, the American Diabetes Association and other standards-based bodies from which emanate accepted, evidenced-based medicine (EBM) guidelines for treating specific conditions.

Claims data help Aetna validate that contracted physicians are treating in accordance with quality standards and EBM practices. In fact, says Cowan, it's usually cut and dried. A heart attack patient should receive betablockers. Women over 50 need a mammogram every two years. Aetna uses accepted medical science from external sources, comparing it against claims data, to determine if quality treatment measures have been prescribed. It's even possible, from claims data, to identify situations where a recommended medication is not appropriate--for example, if the member has a second disease that makes the use of the medication more dangerous than using it to treat the first disease.

Added Advantage

Cowan is quick to describe an additional gain. "We see in claims data what physicians don't see because they usually don't share data.

The patient gets one drug from one physician, another drug from another physician, and tells neither what she is taking. Neither doctor knows--but we know, because we see it in the analyses. We can identify situations where there may be dangerous drug interactions, or possible situations of 'doctor-shopping' for excessive narcotic use, and bring this information to the physician's attention."

So far, so good, but not rocket science, right?

Hold on a minute. On top of its wellness, pharmacy, case management and disease management programs, Aetna has laid MedQuery, a program designed to present "opportunities for improved care" to physicians, but with no heavy hand behind it.

The MedQuery process starts with a collection of 24 months of medical and pharmacy claims data for participants in the program. MedQuery next applies complex, EBM-based algorithms to claims' and demographic data to identify potential gaps or errors in a patient's treatment. Aetna's success with the program is partially attributable to its technology, and partially attributable to the propriety Aetna has wrapped the program in.

Potentials gaps or errors are "care considerations." According to Cowan, "These are opportunities for consideration and change--not instances of a health plan dictating or directing treatment. With the extensive claims data in our data repository, essentially, we have a subset of a patient's electronic health record. Much of our data would be in the physician's database if he utilized all EHR. But without a practice-based EHR, we provide that foundation. When analyzed and mined, that foundation presents opportunities for improvement. The physician has information that we don't have, such as a knowledge of drug allergies, and is in the best position to use the information we provide to make a final decision on the best treatment for the member."

Beyond identifying care considerations, MedQuery stratifies improvement opportunities by severity.

Level 1 represents potentially serious medical situations where communication with the patient's physician should be immediate and may produce an important improvement. Level 1 care considerations require an Aetna physician to call the treating physician within 24 hours and to follow up with a letter or fax.

Level 2 represents non-urgent but potentially serious issues where communication can safely be delayed for a few days. These care considerations require a phone call or letter from an Aetna physician or nurse within one week to the treating physician, followed with a letter or fax.

Level 3 represents bona fide, yet non-urgent, concerns for which a letter to the treating physician is routine and generally sufficient.

Effecting Change

In establishing MedQuery, Aetna officials understood that physicians might demonstrate varied reactions to its interventions. "When provided comparative information about a standard of care, a physician might be receptive," says Cowan. To boost physician receptivity, Aetna has developed strong internal protocols about what information is offered to the treating physician--and how it is delivered.

"We think it's useful to bring information to a doctor's attention," says Cowan, "to point out a member who is a candidate for a particular drug, or who may experience drug interaction. Our dialog is carefully constructed. We might say, 'From analyzing our claims data, it appears that your patient has ...' and we might cite the literature. We would invite the doctor, if appropriate, to review the situation and 'consider a change' in treatment. Working from claims data, we can't always grasp the nuances of the situation, so the final decision is certainly left to the doctor."

Aetna covers more than 13 million lives, and al ready more than 1 million are covered by MedQuery. Last year, the program identified about 100,000 situations where there was an opportunity for care consideration and change. Level 1 cases, the most severe, represented between 2 percent and 5 percent of all cases--but for the Level 1 cases, patient treatment changed about 70 percent of the time within three months of Aetna's intervention. Even at the low end of the percentages spectrum, that represents at least 1,400 Aetna members whose care was improved because of a MedQuery intervention.

Think doctors don't change? Think claims data analysis doesn't have an impact on improved care? Aetna has at least 1,400 good reasons to disagree.