McCreedy, Ellen2017-10-092017-10-092016-07https://hdl.handle.net/11299/190526University of Minnesota Ph.D. dissertation. July 2016. Major: Health Services Research, Policy and Administration. Advisor: Robert Kane. 1 computer file (PDF); ix, 108 pages.Background Diabetes is characterized by high blood sugar, or hyperglycemia. In addition to diet and exercise, several classes of medications are commonly used to treat hyperglycemia in type 2 diabetes in an attempt to reduce the downstream vascular complications of the disease. Four large trials showed few benefits and significant harms from attempting to achieve near normal glycemic control in middle aged people with type 2 diabetes. Benefits of aggressive glycemic control are further reduced for older adults with longstanding disease, those who have accumulated many of the complications of diabetes, and people with other comorbid conditions that limit life expectancy. This more complex older adult population is also at greater risk of iatrogenic hypoglycemia, a harm associated with treatment. For these reasons the American Diabetes Association and the American Geriatric Society have published guidelines recommending less stringent glycemic control for older adults with multiple comorbid conditions and limited life expectancies. However, little is known about how these guidelines are being implemented by primary care clinicians who provide most of the chronic disease management in the US. Methods A factorial vignette study was used to determine the effect of the patient characteristics mentioned in the existing guidelines on a clinician’s decision to prescribe a second-line treatment to achieve tighter glycemic control. The factors varied were patient age / disease duration (65 with short disease duration, 80 with long disease duration), cardiovascular disease (no heart disease, coronary artery disease with previous bypass), and cognitive impairment (no impairment, cognitive impairment that restricted ability to drive). Two policy-relevant glycated hemoglobin (HbA1c) levels were varied in the vignettes: 7.5% or 8.5%. Independent and combined effects of patient factors (patient complexity) were considered. Primary care clinicians from around the US were asked to participate via email. Clinician information was collected, including: years in practice, familiarity with treating older adults, and clinician type (family, internal, nurse practitioner). Clinicians were also asked to predict how likely the hypothetical patient was to adhere to their medication choices. Mixed effect models were used to account for the panel nature of the data (clinicians viewing multiple vignettes). Results 366 primary care clinicians from 36 states participated, with the majority of respondents practicing in Minnesota (35% of sample) or Florida (26% of sample). While we found some sensitivity to the patient factors mentioned in the existing guidelines, we also found evidence of overtreatment of the most complex hypothetical patients. For example, an 80-year-old with longstanding diabetes, cognitive impairment, and coronary artery disease requiring bypass had a second-line treatment added 35% of the time at a HbA1c level of 7.5%, and 75% of the time at a HbA1c of 8.5%. The same patient was recommended a sulfonylurea or insulin (agents known to increase the risk of iatrogenic hypoglycemia) 36% of the time at a HbA1c level of 7.5% and 44% of the time at a HbA1c level of 8.5%. Family practice physicians were less likely to add an additional medication than internal medicine physicians or nurse practitioners. Clinicians did not incorporate their adherence predictions into their decisions to intensify medication therapy. Conclusions This work is part of a larger discussion around balancing the risks and benefits of aggressively treating hyperglycemia in older adults with type 2 diabetes for whom tight glycemic control produces few benefits and significantly increases risk for iatrogenic hypoglycemia. Clinicians may treat more aggressively than guidelines recommend because they are unfamiliar with the geriatric-specific guidelines or they may work in settings where performance incentives are tied to achieving HbA1c levels recommended for average or healthier patients. As few benefits and serious harms are associated with overtreatment, policy recommendations include: 1. creating performance incentives to reduce anti-glycemic medication therapy when appropriate; and 2. developing tools to help primary care clinicians evaluate and address complexity and life expectancy in their older patients with multiple chronic conditions.enThesis or Dissertation