2006 HSR&D National Meeting Abstract
1049 — Providers Vary Substantially in their Propensity to Intensify Blood Pressure Treatment
Hofer TP (Center for Practice Management and Outcomes Research, Ann Arbor)
Klamerus ML (Center for Practice Management and Outcomes Research, Ann Arbor)
Zikmund-Fisher BJ (Center for Practice Management and Outcomes Research, Ann Arbor)
Kerr EA (Center for Practice Management and Outcomes Research, Ann Arbor)
For people with diabetes, blood pressure (BP) control is the single most important intervention in preventing cardiovascular mortality. Differences in often difficult treatment intensification decisions made by providers may explain the variable levels of blood pressure control seen in diabetic patients. We developed a “propensity to intensify treatment” scale from responses to brief clinical scenarios.
370 VA primary care providers, selected from a random national sample, completed mailed surveys about hypertension and diabetes (a 74% response rate). Respondents were presented with four scenarios involving a patient with a clinic BP of 145/90 on 2 medications (Scenario 1), and then in 3 alternate scenarios with the same patient already on 3 medications (Scenario 2), or with a new comorbid condition (depression) (Scenario 3), or with a report of normal home measurements (Scenario 4). The responses included accept the BP, recheck within 3 months, or change or add an antihypertensive medication. Multilevel ordinal two parameter item response models were used to analyze the results and to produce a provider scale.
We found substantial variation in the provider propensity to intensify treatment across the scenarios. For providers 1 standard deviation (sd) above the mean on the propensity to intensify scale the probabilities of changing or adding a medication are .99, .87, .66 and .25 for Scenarios 1-4 respectively. For providers one sd. below the mean, the probabilities of intensification are .87, .45, .16 and .03. The reliability of the scale score was .75. Independently we asked providers to rate the importance of intervening on a DBP=88 or SBP=145 and these ratings explained over a third of the variance in the propensity to intensify scale score. However the scale score was not explained by provider characteristics like gender, years in practice, or provider type.
Provider propensity to intensify treatment can be measured reliably using brief scenarios. For the same BP level it varies substantially across providers and is correlated with the importance attributed to meeting BP goals. It does not appear to vary substantially across provider demographic characteristics.
This measure can be used to explore reasons for lack of intensification of BP treatments and clinical inertia.