Virginia Research Day 2021

A CONTINUOUS PROCESS IMPROVEMENT PROJECT TO REDUCE THE INCIDENCE OF INPATIENTS LEAVING AGAINST MEDICAL ADVICE Seavon Munoz, DO, Adam Huryta, DO, Michael Moore, MD Sovah Health Family Medicine Residency Program, Danville, VA RR-13

Background: Each year, 2% of inpa � ents leave against medical advice (LAMA) i.e. a discharge not recommended by the trea �ng p hysician, 2 which increases the poten �al for adverse health outcomes and increased treatment cost 3 . The 30- day readmission rate of LAMA pa � ents is about 30%. LAMA pa � ents display characteris �cs including male gender, younger age, fi nancial issues, sickness within the family (pa � ent is needed at home), pa � ents feeling be � er, psychiatric illness, social issues at home, no primary care physician and illicit drugs or alcohol problems. Two years ago, we ini � ated a con �nuous process improvement (CPI) project with nursing leadership at SOVAH Health Danville, to iden � fy pa � ents at risk for LAMA in order to implement preven �on strategies. Over a 3-month period in 2018, 46 pa � ents le � AMA with a mean age of 49 years, 46% were males, 52% were Caucasian and there was a mean length of stay (LOS) of 1.91 and 91% had third party payor insurance. There was not data available then on why these pa � ents le �. Following the ini�a l study, a LAMA screen was added to the inpa � ent electronic medical record (EMR) discharge process asking the provider for the LAMA reason(s). This follow-up study was done over 3 months in 2020. Methods: An inpa � ent EMR retr ospec� ve study of LAMA pa � ents was done July 1 to September 30, 2020, to analyze LAMA reasons. Data collected included: ethnicity, reason for admission, age, gender, length of stay, history of substance abuse, LOS, if they had a primary care physician, reasons for leaving AMA, and readmission. This study was granted a waiver for individual informed consent by the SOVAH Health IRB since de-iden �fi ed data was used. Results: Over the 3 months, 28 pa � ents le � AMA out of 1,801 total. Demographics include a mean age of 43.7 years, 15 were Caucasian, 13 were African American, half were male (Figure 1 & 2)). Hospital length of stay was similar in 2018, with a mean of~2 days for those who LAMA, compared to 4 days for the others (Figure 3). Readmission rate for individuals LAMA was ~32% compared to overall readmission rate of the hospital of ~15.8% (Figure 4). Reasons for leaving AMA included social issues (9), concerns with care being received (10), need to return to caregiving responsibili � es (1), and 8 stated no reason. None were due to the pa � ent’s perceived cost of hospitaliza �on or needing to care for a pet (Figure 5).

AMA

Non- AMA

Figure 3: Visual representa �on of the mean length of stay for AMA and non-AMA discharges over a three-month period in 2018 and 2020 at SOVAH Danville

Figure 2: Visual representa �on of race dis tribu�on for individuals LAMA

Sovah Totals

Pa �e nts LAMA

15%

32%

Figure 4: Overall readmission rate (15%) compared to those individuals LAMA (32%) over a 3-month period in 2020 at SOVAH Danville

Figure 5: Reasons for LAMA for 28 individuals over a 3-month period in 2020 at SOVAH Danville

Discussion: This study compared the demographics of pa � ents LAMA from Sovah Health-Danville in 2020 compared to 2018 and included the ra �onale for LAMA in 2020. As was the case in 2018, LAMA pa � ents were rela � vely younger, half males and half Caucasian. The most common reasons for LAMA were concerns with their hospital care or social issues. Their readmission rate was 32%, which is in line with 2018 and other studies. There were fewer overall pa � ents LAMA over 3- months in 2020 compared to 2018. However, the reason for this r educ�on is not known since no interven �ons to reduce LAMA were introduced. However, improved awareness of the poten �al detriments to LAMA by those involved in the current and previous studies, could have played a role. The new understanding of the ra �onale behind those LAMA will allow implementa �on of s peci fi c admission screening for those at risk, coupled with the applica �on of specific interven �ons to address pa � ent concerns, such as se �ng p a � ent’s expecta �ons of care. References: (1) Alfandre DJ. Mayo Clin Proc. 2009; 84(3):255-260 (2) Glasgow, J.M.. Gen Intern Med . 2010; 25:926 (3) Gautam, N. Indian J Cri ti cal Care Med . 2018; 22(8) 591-596 (4) Choi M. Plos One . 2011; 6(9) 24459

Figure 1: Visual representa �on of gender dis tribu�on of individuals LAMA July 1-September 20, 2020

33

2 0 2 1 R e s e a r c h R e c o g n i t i o n D a y