LLLawrence M. Lewis

Lawrence M. Lewis

*This is an algorithmically generated author record based on publicly available information.

Lawrence M. Lewis has a total of 17 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    ChatGPT Provides Inconsistent Risk-Stratification of Patients With Atraumatic Chest Pain

    by Thomas F. Heston, Lawrence M. Lewis

    BACKGROUND ChatGPT is a large language model with promising healthcare applications. However, its ability to analyze complex clinical data and provide consistent results is poorly known. This study evaluated ChatGPT-4’s risk stratification of simulated patients with acute nontraumatic chest pain compared to validated tools. METHODS Three datasets of simulated case studies were created: one based on the TIMI score variables, another on HEART score variables, and a third comprising 44 randomized variables related to non-traumatic chest pain presentations. ChatGPT independently scored each dataset five times. Its risk scores were compared to calculated TIMI and HEART scores. A model trained on 44 clinical variables was evaluated for consistency. RESULTS ChatGPT showed a high correlation with TIMI and HEART scores (r = 0.898 and 0.928, respectively), but the distribution of individual risk assessments was broad. ChatGPT gave a different risk 45-48% of the time for a fixed TIMI or HEART score. On the 44 variable model, a majority of the five ChatGPT models agreed on a diagnosis category only 56% of the time, and risk scores were poorly correlated (r = 0.605). ChatGPT assigned higher risk scores to males and African Americans. CONCLUSION While ChatGPT correlates closely with established risk stratification tools regarding mean scores, its inconsistency when presented with identical patient data on separate occasions raises concerns about its reliability. The findings suggest that while large language models like ChatGPT hold promise for healthcare applications, further refinement and customization are necessary, particularly in the clinical risk assessment of atraumatic chest pain patients.


  • Improving Follow‐up Attendance for Discharged Emergency Care Patients Using Automated Phone System to Self‐schedule: A Randomized Controlled Trial

    by Kyla L Bauer, Omolade O Sogade, Brian F Gage, Brent Ruoff, Lawrence Lewis

    Background: Automated phone appointment reminders have improved adherence with follow‐up appointments in a variety of hospital settings, but have mixed results in patients discharged from the emergency department (ED). Increasing adherence to follow‐up care has been a priority in the ED to improve patient outcomes and reduce unnecessary future visits.Methods: We conducted a prospective randomized open, blinded end‐point (PROBE) trial of 278 adult patients discharged from the ED and referred to a provider for follow‐up care. Participants in the intervention arm received a self‐scheduling text or phone message that automatically connected them to their referral provider to schedule a follow‐up appointment and sent them appointment reminders. Those in the control arm received standard‐of‐care written instructions to contact listed referral providers. The primary outcome was time to appointment. The secondary outcome was time to return visit to the ED.Results: The automated reminders increased the cumulative incidence of keeping the referral appointment after ED discharge (p < 0.001, Gray's test). Of participants randomized to the automated phone intervention, 49.3% (n = 74) kept their follow‐up appointment versus 23.4% (n = 30) in the control arm, with a hazard ratio (HR) and 95% confidence interval (CI) over the duration of the study period of 2.4 (1.6 to 3.7; p < 0.001). In a sensitivity analysis using 30 days of follow‐up data, 42.0% (n = 63) of participants randomized to the phone intervention kept their follow‐up versus 21.1% (n = 27) in the control arm, with a HR (95% CI) of 2.2 (1.4 to 3.5; p < 0.001). There was no difference in ED revisits between the intervention and control group within 120 days postdischarge.Conclusions: An automated self‐scheduling phone system significantly improved follow‐up adherence after ED discharge, but did not decrease ED revisits.


  • Efficacy of a new intravenous β2-adrenergic agonist (bedoradrine, MN-221) for patients with an acute exacerbation of asthma

    by Stacey L. House, Kazuko Matsuda, Geoffrey O'BRien, Malath Makhay, Yuichi Iwaki, Ian Ferguson, Luis M. Lovato, Lawrence M. Lewis


  • Open Access

    Emergency Department Patients With Diabetes Have Better Glycemic Control When They Have Identifiable Primary Care Providers

    by David A. Horwitz, Evan S. Schwarz, Mitchell G. Scott, Lawrence M. Lewis

    ACADEMIC EMERGENCY MEDICINE 2012; 19:00–00 © 2012 by the Society for Academic Emergency Medicine Abstract Objectives: The objective was to determine if emergency department (ED) patients with diabetes mellitus (DM) who have primary care providers (PCPs) have better control of their DM than patients with no PCPs. Methods: This was a prospective, cross-sectional, observation study at a large, adult, urban, academic ED with 85,000 annual visits. ED patients with a history of DM were eligible. Patients with severe systemic disease, diabetic ketoacidosis (DKA), sepsis, active steroid use, pregnancy, or cognitive impairment were excluded. Consenting patients had hemoglobin A1c (HgbA1c) analysis and completed a questionnaire regarding demographics, lifestyle, medication usage, educational level attained, and health care access, including whether or not they had PCPs. HgbA1c levels were compared between subjects with and without PCPs using medians with interquartile ranges (IQRs). A continuous plot was developed to demonstrate the proportion of patients without PCPs (PCP−) compared to those with PCPs (PCP+) at every level of %HgbA1c across the entire measured range. Multivariate logistic regression analysis was used to determine which clinical and demographic factors obtained from the questionnaire were associated with improved glycemic control (increased relative risk [RR] of having a %HgbA1c < 8%). Results: A total of 284 patients were screened; 227 were enrolled, had HgbA1c analysis performed, and had complete PCP, race, and sex information. Complete demographic data (insurance status, employment status, etc.) were available on 209 subjects. Sixty-four of the 227 patients (28.2%) denied having PCPs. Median HgbA1c was 7.7% (IQR = 6.5% to 9.68%) in PCP+ versus 8.9% (IQR = 6.8% to 11.3%) in PCP− patients (p = 0.01). Ninety-one of 163 (55.8%) PCP+ subjects had a median HgbA1c < 8% versus 25 of 64 (39.1%) in the PCP− group (p = 0.02). After adjusting for multiple clinical and demographic variables, having a PCP remained significantly associated with a median HgbA1c value less than 8% (RR = 1.43; p = 0.04). Conclusions: Diabetes control was significantly better in patients with PCPs, even after adjusting for a number of potentially confounding social and demographic factors.