Weekly Article: Low level trainees and increased LOS

by Paul Mullan, Children's National

One of the Institute of Medicine’s quality indicators is timeliness. Academic hospitals like ours are training a larger number of learners to meet the demand for physician services, but the effect of learners on timeliness of care is understudied. While a cluster (groups of subjects are randomized, instead of individuals being randomized) randomized trial might be ideal to study the effect of different groups (eg no residents vs 2 residents vs 2 residents/2 med students, etc.) on flow, organizing such a trial would be difficult to accomplish and Jen/Pavan would probably put the idea into the “not a great fellow’s project” category. Two docs from a Canadian peds ED used data from thousands of visits and some time studies to perform Discrete Event Simulation (DES) modeling to study the effect of trainee presence on several process measures while controlling for the varying ED conditions.

Setting: 40,000 pts/yr academic Canadian ED with 5 level triage system; high acuity (triage 1-3) and low acuity (triage 4,5) patient tracks. In low volume season, all attendings work with trainees; in high season, some attendings do not work with trainees. Time duration by direct observation for various care delivery processes were performed for trainees (medical students and residents; they did NOT look at PEM fellows or senior EM residents).

(Adapted & rounded numbers) Table 2. Trainee timing in minutes

Acuity Track	Patient Assessment	Discussion with ED Attg	Patient Reassess	Total
High	16	6	6	28
Low	15	4	6	25

Notice - while ED attendings decreased their discussion time by about 1/3 with high vs low acuity tracks, trainees spend nearly the identical amount of time with each patient type. One thing that Jane Lavelle taught me in residency was the 5/5 rule – any triage level 5 patient should take 5 minutes to assess. This is something we can pass on to residents/students at CNMC as well to increase their number of patients seen, increase ED timeliness, and enable more time for that “discussion with ED attending” educational part.

Findings: (TPA = trainees per attending)

ED Length of Stay (arrival to dispo): increase of 7 minutes/pt for each additional TPA; avg 212 minutes with 0 TPA, avg 259 minutes with 6 TPA. No effect on Triage 1 patients

ED Wait Time (arrival to attending/trainee): Decreased as number of TPA increased in low volume season, but no effect in high volume season; avg 110 minutes with 0 TPA, avg 80 minutes with 6 TPA. No effect on Triage 1 patients.

Left without being seen: ~2.5% overall.
- Triage 4/5 pts decreased with increasing TPA (from 4.5% with 0 TPA to 3% with 6 TPA), but…
- Triage 3 pts increased (from ~1% with 0 TPA, to 2.5% with 6 TPA).
- In high volume season, though, LWBS doubled from 1.8 (0 TPA) to 3.6% (6 TPA) ; in low volume season, no effect on~1% LWBS regardless of TPA.

Busy winter night shift? This study suggest that more low level trainees will get you: higher LWBS, same ED wait time, higher ED LOS.

Solution? Priming and situational awareness. When flying solo on that busy winter shift with 2 medical students and 2 new interns… huddle your team and talk about methods to improve patient flow. Huddle with your charge nurse and let them know that we are at risk for LWBS and higher LOS, and keep communication lines open for game planning how to react when these metrics do start moving.

Other studies suggest that LWBS rates are more affected by wait time than by total LOS, as patients are less likely to leave once seen. Our LEAN program in the ED is trying to decrease wait time by ~50% and it will be exciting to (hopefully) see our LWBS rates drop further as well.