Simulating approaches to emergency department pandemic physician staffing during COVID-19

Objective: During pandemics, emergency departments (EDs) are challenged by the need to replace quarantined ED staff and avoid staffing EDs with nonemergency medicine (EM) trained physicians. We sought to design and examine three feasible ED staffing models intended to safely schedule EM physicians to staff three EDs within a health system during a prolonged infectious disease outbreak.Methods: We conducted simulation analyses examining the strengths and limitations of three ED clinician staffing models: two-team and three-team fixed cohort, and three-team unfixed cohort. Each model was assessed with and without immunity, and by varying infection rates. We assumed a 12-week pandemic disaster requiring a 2-week quarantine.Main outcome: The outcome, time to staffing shortage, was defined as depletion of available physicians in both 8- and 12-hour shift duration scenarios. Results: All staffing models initially showed linear physician attrition with higher infection rates resulting in faster staffing shortages. The three-team fixed cohort model without immunity was not viable beyond 11 weeks. The three-team unfixed cohort model without immunity avoided staffing shortage for the duration of the pandemic up to an infection rate of 50 percent. The two-team model without immunity also avoided staffing shortage up to 30 percent infection rate. When accounting for immunity, all models behaved similarly initially but returned to adequate staffing during week 5 of the pandemic. Conclusions: Simulation analyses reveal fundamental tradeoffs that are critical to designing feasible pandemic disaster staffing models. Emergency physicians should test similar models based on local assumptions and capacity to ensure adequate staffing preparedness for prolonged pandemics.

Associations between emergency department crowding and perceptions of interpersonal care in patients presenting with suspected acute coronary syndrome

BackgroundEmergency department (ED) crowding is associated with numerous healthcare issues, but little is known about its effect on psychosocial aspects of patient-provider interactions or interpersonal care. We examined whether ED crowding was associated with perceptions of interpersonal care in patients evaluated for acute coronary syndrome (ACS).MethodsPatients presenting to a quaternary academic medical centre ED in New York City for evaluation of suspected ACS were enrolled between November 2013 and December 2016. ED crowding was measured using the ED Work Index (EDWIN), which incorporates patient volume, triage category, physician staffing and bed availability. Patients completed the 18-item Interpersonal Processes of Care (IPC) survey, which assesses communication, patient-centred decision-making and interpersonal style. Regression analyses examined associations between EDWIN and IPC scores, adjusting for demographics, comorbidities and depression.ResultsAmong 933 included patients, 11% experienced ED overcrowding (EDWIN score >2) at admission, 11% experienced ED overcrowding throughout the ED stay and 30% reported suboptimal interpersonal care (defined as per-item IPC score <5). Higher admission EDWIN score was associated with modestly lower IPC score in both unadjusted (β=–1.70, 95% CI –3.15 to –0.24, p=0.02) and adjusted models (β = –1.77, 95% CI –3.31 to –0.24, p=0.02). EDWIN score averaged over the entire ED stay was not significantly associated with IPC score (unadjusted β=–1.30, 95% CI –3.19 to 0.59, p=0.18; adjusted β=–1.24, 95% CI –3.21 to 0.74, p=0.22).ConclusionIncreased crowding at the time of ED admission was associated with poorer perceptions of interpersonal care among patients with suspected ACS.

The Effects of Implementing a “Waterfall” Emergency Physician Attending Schedule

Introduction: Increases in emergency department (ED) crowding and boarding are a nationwide issue resulting in worsening patient care and throughput. To compensate, ED administrators often look to modifying staffing models to improve efficiencies. Methods: This study evaluates the impact of implementing the waterfall model of physician staffing on door-to-doctor time (DDOC), door-to-disposition time (DDIS), left without being seen (LWBS) rate, elopement rate, and the number of patient sign-outs. We examined 9,082 pre-intervention ED visits and 8,983 post-intervention ED visits. Results: The change in DDOC, LWBS rate, and elopement rate demonstrated statistically significant improvement from a mean of 65.1 to 35 minutes (P <0.001), 1.12% to 0.92% (P = 0.004), and 3.96% to 1.95% (P <0.001), respectively. The change in DDIS from 312 to 324.7 minutes was not statistically significant (P = 0.310). The number of patient sign-outs increased after the implementation of a waterfall schedule (P <0.001). Conclusion: Implementing a waterfall schedule improved DDOC time while decreasing the percentage of patients who LWBS and eloped. The DDIS and number of patient sign-outs appears to have increased post implementation, although this may have been confounded by the increase in patient volumes and ED boarding from the pre- to post-intervention period.

Hospital Factors, Performance on Process Measures After Transient Ischemic Attack, and 90-Day Ischemic Stroke Incidence

Background and Purpose: We determined the association between hospital factors, performance on transient ischemic attack (TIA) process measures, and 90-day ischemic stroke incidence. Methods: Longitudinal analysis of retrospectively obtained data on 9168 veterans ≥18 years with TIA presenting to the emergency department or inpatient unit at 69 Veterans Affairs hospitals with ≥10 eligible patients per year in fiscal years 2015 to 2018. Process measures were high/moderate potency statin within 7 days of discharge, antithrombotic by day 2, and hypertension control (<140/90 mm Hg) at 90 days. The outcome was 90-day stroke incidence. Results: During the 4-year study period, hospitals significantly increased statin use (adjusted odds ratio [aOR] per 1-year increase, 1.24 [95% CI, 1.17–1.32]; P <0.001), whereas neither hypertension control ( P =0.44) nor antithrombotic use ( P =0.82) improved over time. Hospitals that admitted a higher proportion of TIA patients versus emergency department discharge had significantly greater use of statins (aOR per 10-percentage point increase in the proportion of TIA patients admitted, 1.09 [1.03–1.16]; P =0.003) and antithrombotics (aOR per 10-percentage point increase in TIA patients admitted, 1.14 [1.06–1.23]; P <0.001). Hospitals with higher emergency physician staffing and lower TIA patient volume had greater use of antithrombotics (aOR per 1 full-time physician increase, 1.05 [1.01–1.08]; P =0.008 and aOR per 10-patient decrease in volume, 1.09 [1.01–1.16]; P =0.02). Higher emergency physician staffing was associated with lower 90-day stroke incidence (aOR per 1 full-time physician increase, 0.96 [0.92–0.99]; P =0.02) but other hospital factors were not. Conclusions: Hospitals admitting higher percentages of TIA patients and having higher emergency physician staffing have greater performance on select guideline-concordant process measures for TIA. Higher emergency physician staffing was associated with improved outcomes 90 days after TIA.

Characteristics of the ideal hospitalist inpatient care program: perceptions of Canadian health system leaders

Background: Despite the growing prevalence of hospitalist programs in Canada, it is not clear what program features are deemed desirable by administrative and medical leaders who oversee them. We aimed to understand perceptions of a wide range of healthcare administrators and frontline providers about the necessary characteristics of a hospitalist service. Methods: We conducted semi-structured interviews with a range of administrators, medical leaders and frontline providers across three hospital sites in an integrated health system in Western Canada. Results: Most interviewees identified the hospitalist model as the ideal inpatient care service line, but identified a number of challenges. Interviewees identified the necessary features of an ideal hospitalist service to include considerations for program design, care and non-clinical processes, and alignment between workload and physician staffing. Conclusions: Most hospital administrators and frontline providers in our study believed the hospitalist model resulted in improvements in clinical processes and work environment.

Factors Associated With the Rate of Sepsis After Surgery

Background Sepsis is a critical illness that requires early detection and intervention to prevent disability and/or death. Objective To analyze the association between various hospital-related factors and rates of sepsis after surgery in Massachusetts hospitals. Methods The sample consisted of 53 hospitals with intensive or critical care units and 25 hospitals with step-down units. Hospital characteristics, staffing levels, and health care–acquired conditions were examined using publicly available data. Analysis of variance and linear regression were performed to explore the relationship between nurse and physician staffing levels and sepsis rates. Results Sepsis rates were significantly lower when nurses cared for fewer patients (P &lt; .001) and when intensivist hours were greater (P = .03). Linear regression for nurse staffing revealed that higher rates of catheter-associated urinary tract infection (P = .001) and higher numbers of step-down patients cared for by nurses (P = .001) were associated with a significantly higher rate of sepsis (P &lt; .001). Linear regression for physician staffing revealed that higher rates of catheter-associated urinary tract infection (P &lt; .001) and wound dehiscence after surgery (P &lt; .001), greater hospitalist hours (P = .001), and greater physician hours (P = .05) were associated with a significantly higher sepsis rate, while greater intensivist hours were associated with a lower sepsis rate (P = .002). Conclusion In this study, greater nurse staffing and intensivist hours were associated with significantly lower rates of sepsis, whereas greater physician staffing and hospitalist hours were associated with significantly higher rates. Further research is needed to understand the roles of the various types of providers and the reasons for their differing effects on sepsis rates.