Calculating risks: Including mortality factors

Quantifying glaucoma risk in concert with mortality predictions

When deciding whether to treat glaucoma suspects, practitioners often rely on risk calculators. A new study published in the October 2009 issue of the American Journal of Ophthalmology indicates that mortality may be another important factor to weigh in the equation, according to Beth Ann Griffin, Ph.D., association statistician, RAND Corporation, Arlington, Va.
The study began as an outgrowth of an investigation the RAND Corporation was doing to evaluate available risk calculators for glaucoma. “It is a simple calculator that a doctor can find on the web or as a handheld device that Pfizer [New York] gives out. A doctor can insert patient characteristics and come up with a number of what the predicted five-year risk is for a patient,” said Dr. Griffin. “One part of the RAND study was to consider what would happen to those calculations if we built in mortality risk.”
There was some thought that mortality risk could provide an important component. “We’re trying to expand the current tools that are available to clinicians because we feel that mortality is an important risk factor,” Dr. Griffin said. “This is something that the American Academy of Ophthalmology has said that clinicians should consider when making treatment decisions.”

Factoring in mortality

For the study, investigators used the Ocular Hypertension Treatment Study and European Glaucoma Prevention Study calculator developed at the University of Washington, Seattle. “It does not take into account mortality risk but it has been shown to be a good predictor of glaucoma risk,” Dr. Griffin said. “We took that calculator and incorporated the risk of dying.”This was combined with the Charlson Index, which predicts mortality. “It takes into account a patient’s age as well as his or her co-morbities,” Dr. Griffin said. “As a function of those two, we can come up with a predicted value for how long a patient has to live or what the likelihood is that the patient will pass away in 5 or 10 years.”
Investigators found that mortality did have an impact. “As we expected, accounting for mortality risk decreases a patient’s risk of glaucoma,” Dr. Griffin said. “This happens because once you account for mortality, you’re allowing for the fact that a patient, particularly an older patient or one who has severe co-morbities, has the possibility of dying before he or she develops glaucoma.”
What investigators felt was important was how much impact the mortality component had in some cases. “What’s important is how much this changes what we’re referring to as an unadjusted glaucoma risk once we account for mortality,” Dr. Griffin said. “The amount of reduction naturally depends on the patient’s general overall health.”
A select group of patients was found to be impacted most by including mortality risk. “We found that the patients who are most affected are those who are older, who have an increased risk of dying, and those who have more co-morbities, as well as patients who have a higher unadjusted risk of glaucoma in general,” Dr. Griffin said. “They’re the ones whose unadjusted risk of glaucoma is different than their mortality adjusted.”
For example, in a younger patient with no significant co-morbidities, the patient’s unadjusted risk might be 50%; once mortality is accounted for, the risk may only be adjusted to 45%. For a patient who is older or who has significant co-morbidity, the risk may jump considerably. “If there is something risky the adjusted risk can go all the way down to 30% or even 20%,” Dr. Griffin said. “We have the formulas to come up with the numbers depending upon the patient characteristics.”

Clinical implications

The clinical implications of this vary from practitioner to practitioner. “We’re hoping that the article will show clinicians the importance of taking mortality into account,” Dr. Griffin said. “There’s no universal rule that says depending on what number comes out of a glaucoma risk calculator, a patient should be treated—every clinician has his or her own guidelines.” If a clinician has a general rule that he expects to treat anyone with a glaucoma risk of greater than 20%, Dr. Griffin thinks this study will show that treatment decisions might change if life expectancy is accounted for.
Overall, Dr. Griffin hopes that practitioners will take away the idea that mortality risk is an important factor to consider. “I think that the manuscript provides clinicians with an explicit way of incorporating mortality risk into estimates of glaucoma risk,” she said. “I also think that this is something that’s important not just for glaucoma but for other diseases as well. We advocate wider use of explicit measures of life expectancy in risk calculations in general.”

Editors’ note: Dr. Griffin has no financial interests related to her comments.

Contact information

Griffin: 703-413-1100, bethg@rand.org