KEYSTONE, COLO. — The most popular method of assessing patient adherence to type 1 diabetes management instructions consists of a look at glycosylated hemoglobin values.
But this is also one of the least reliable means of doing so, “and I don't recommend it,” said Suzanne Bennett Johnson, Ph.D., professor and chair of the department of medical humanities and social service at Florida State University, Tallahassee.
She argued that hemoglobin A1c (HbA1c) provides only limited and often misleading information about patient behavior. Although it's widely assumed that patients with high HbA1c levels are poorly compliant with diabetes care instructions across the board, data from numerous studies indicate otherwise. In reality, patient behavior and HbA1c are only weakly correlated, according to Dr. Johnson, who spoke at a conference on the management of diabetes in youth.
Moreover, HbA1c values yield no specific information about which behaviors need to change. In clinical practice, reliance upon HbA1c to assess patient adherence often results in a blame-the-patient dynamic which, in turn, promotes noncompliance, she said.
Many physicians fall into the trap of thinking of compliance as a sort of personality trait. In fact, compliant behavior with one component of the diabetes regimen—say, blood glucose testing—doesn't predict behavior with regard to any other component, according to Dr. Johnson.
“The bottom line is if you want to know what patients are doing, you're going to have to assess all the components of the regimen you consider important. We'd like to think there's a simple answer, like just measuring glycosylated hemoglobin, but the data really don't support that. So you're going to have to do some extra work if you're really interested in how patients manage their disease in an effort to help them manage it better,” Dr. Johnson continued.
She urged clinicians to perform an annual formal assessment of patient behavior as part of routine diabetes care. The importance of doing so in order to keep patients on track is illustrated by a classic study of patient-provider miscommunication in which, following a visit to a diabetes clinic, patients were asked to recall the recommendations made by the health care team.
Patients were able to recall only 18% of what providers had recommended; however, they supplemented that with a considerably larger body of recommendations they remembered the providers making, but which in fact hadn't been recommended at all, she explained.
The second most popular method of assessing patient behavior, next to HbA1c, is by physician ratings. But the data suggest that physician rating is the most unreliable method of all, said Dr. Johnson.
What, then, is the best way to assess diabetic patient behavior? Dr. Johnson advocated patient self-reports, which she said are underutilized.
The key to getting good quality, reliable data by self-report, she said, is to do a 24-hour recall interview in which the patient describes yesterday's activities in sequence from awakening to retiring.
“We find that by age 6 years, patients can do this quite well,” Dr. Johnson added.
The interview must be conducted in a nonjudgmental fashion. Prompting the patient for missing information is okay; in fact, it's often necessary, particularly in the dietary area.
A couple of interviews are a good idea, one covering a weekday and another a weekend, since the behaviors are often quite different. A separate interview with a family member can be conducted for confirmation of the patient's behavior.
For patients having major problems in carbohydrate counting, Dr. Johnson suggested interviewing them about what they most frequently eat.
The meeting was sponsored by the Barbara Davis Center for Childhood Diabetes, the University of Colorado, and the Children's Diabetes Foundation at Denver.
Data suggest that physician rating is the most unreliable assessment method of all. DR. JOHNSON