Appropriate Use of Nuclear Myocardial Perfusion Imaging in University-Based Veteran Administration Institution.

Background:
Appropriate Use Criteria (AUC) forNuclear myocardial perfusion imaging (NMPI)was written for judicious use of cardiovascular testing. There is paucity of data in Veteran Administration (VA) population.

Methods:
Retrospective NMPI at Dayton VA Medical Center were evaluated (Dec 2016 to May 2017). Patients with incomplete medical record or who did not complete testing were excluded. AUC guidelines for NMPI were applied with interobserver variability.

Results:
Of 465 patients 93.5% were male, 6.5% females, mean age 64.05 (range 30 to 89 years). 2.4% had zero, 9.9% one, 24.9% two, 40.4% three, 20.2% four and 2.2% five cardiovascular disease (CVD) risk factors. 112 (24.1%) NMPI were ordered for evaluation of known CVD, 48 (10.3%) for perioperative evaluation, 253 (54.4%) and 52 (11.2%) for risk stratification of symptomatic and asymptomatic patients without CVD (non-CVD). Primary Care 252 (54.2%), Cardiology 125 (26.9%), Primary Care Nurse Practitioners 73 (5.7%), Surgery 1(1.5%), Internal Medicine residents 5 (1.1%) and Hospitalists 2 (0.4%) ordered the studies.71.4% NMPI for perioperative evaluation were ordered prior to intermediate risk surgery. In known CVD NMPI was ordered >2 year from stenting and >5 years after cardiac surgery in 50% and 33.3% patients. In symptomatic patients without CVD 164 (54.5%) NMPI was ordered to assess ischemic symptoms. 81.6% had intermediate risk for CVD. 31 (63.3%) of perioperative, 78 (76.5%) known CVD, 132 (80.4%) symptomatic without prior NMPI and 23 (22.1%) symptomatic and 22 (45.7%) asymptomatic non-CVD with prior NMPI were classified as appropriate. 18 (36.7%) perioperative, 24 (23.5%) known CVD, 32 (19.6%) symptomatic without prior NMPI and 80 (77.6%) symptomatic and 26 (54.2%) asymptomatic with prior non-CVD were classified as inappropriate or uncertain. Inter-observer agreement was noted.

Conclusions:
In general three in ten NMPI were deemed not appropriate. One-half asymptomatic non-CVD NMPI were deemed inappropriate or uncertain. four in ten perioperative, two in ten known CVD and symptomatic non-CVD without prior NMPI were inappropriate or uncertain. For patients with prior NMPI one in ten symptomatic were inappropriate.