I'm Leaving": Factors That Impact Against Medical Advice Disposition Post-Trauma.

2020 
Abstract Background Discharge against medical advice (AMA) is an important, yet understudied, aspect of health care—particularly in trauma populations. AMA discharges result in increased mortality, increased readmission rates, and higher health care costs. Objective The goal of this analysis was to determine what factors impact a patient's odds of leaving the hospital prior to treatment. Methods We performed a retrospective analysis of the National Trauma Data Bank on adult trauma patients (older than 14 years) from 2013 to 2015. Of the 1,770,570 patients with known disposition, excluding mortality, 24,191 patients (1.4%) left AMA. We ascertained patient characteristics including age, sex, race, ethnicity, insurance status, ETOH, drug use, geographic location, Injury Severity Score (ISS), injury mechanism, and anatomic injury location. Multivariate logistic regression models were used to determine which patient factors were associated with AMA status. Results Uninsured (odds ratio [OR] 2.72; 95% confidence interval [CI] 2.58–2.86) or Medicaid-insured (OR 2.50; 95% CI 2.37–2.63) trauma patients were significantly more likely to leave AMA than patients with private insurance. Compared to white patients, African-American patients (OR 1.06; 95% CI 1.02–1.11) were more likely, and Native-American (OR 0.62; 95% CI 0.52–0.75), Asian (OR 0.59; 95% CI 0.49–0.69), and Hispanic (OR 0.80; 95% CI 0.75–0.85) patients were less likely, to leave AMA when controlling for age, sex, ISS, and type of injury. Conclusions Insurance status, race, and ethnicity are associated with a patient's decision to leave AMA. Uninsured and Medicaid patients have more than twice the odds of leaving AMA. These findings demonstrate that racial and socioeconomic disparities are important targets for future efforts to reduce AMA rates and improve outcomes from blunt and penetrating trauma.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    26
    References
    2
    Citations
    NaN
    KQI
    []