3016 — Getting to Goal Isn’t Everything: Seeing Risk Reductions in Absolute, Rather than Relative, Terms

Author List:
Zikmund-Fisher BJ (VA Ann Arbor Healthcare System)
Hofer TM (VA Ann Arbor Healthcare System)
Ubel PA (VA Ann Arbor Healthcare System)
Kerr EA (VA Ann Arbor Healthcare System)

Objectives:
Many chronic disease treatment guidelines are threshold-based, requiring intervention for test values (e.g., blood pressure [BP]) above a particular cutoff value. Doing so, however, ignores potentially important distinctions among people labeled as "above goal." For example, hypertension has an exponential relationship to cardiovascular risk: A systolic BP of 40 points over goal confers more than twice the cardiovascular risk as a BP only 20 points over goal. But, do VA healthcare providers understand this relationship?

Methods:
370 VA primary care providers, selected from a random national sample, completed mailed surveys regarding hypertension and diabetes (a 74% response rate). In the key question, respondents indicated which BP reduction reduced cardiovascular risk the most: lowering systolic BP from 175mm Hg to 155mm Hg, lowering systolic BP from 155 to 135, or both BP reductions yielded equal risk reductions. We used bivariate statistics to examine associations between responses to this question and physician specialty as well as perceived relative priority of hypertension treatment.

Results:
Only 14% of providers gave the correct answer (lowering systolic BP from 175mm Hg to 155mm Hg yields the largest risk reduction), while 45% said both reductions were equal and 41% gave the reverse answer. MDs were significantly more likely to answer correctly than NPs and PAs (p<0.002). While all providers attributed equal importance to intervening on elevated systolic BP (145mm Hg) in a scenario question, those correctly answering the key question assigned significantly lower importance (vs. incorrect answerers) to interventions for marginally elevated hemoglobin A1c and marginally decreased HDL (p's <0.001).

Implications:
Consistent with past psychological research, most providers surveyed appeared to evaluate different test values by their relative "distance" from normal or zero risk levels, seeing less risk reduction in actions which left patients still above goal. Providers who correctly understood the exponential risk relationship were more likely to prioritize treatment strategies based on risk of potential long-term complications.

Impacts:
Providers and patients alike may need help to understand that reducing extreme values is of great value, even when defined treatment goal levels cannot be achieved.