Original Article

Morning report decreases length of stay in trauma patients

Abstract

Background Modern acute care surgery (ACS) programs depend on consistent patient hand-offs to facilitate care, as most programs have transitioned to shift-based coverage. We sought to determine the impact of implementing a morning report (MR) model on patient outcomes in the trauma service of a tertiary care center.

Methods The University of Arkansas for Medical Sciences (UAMS) Division of ACS implemented MR in October 2015, which consists of the trauma day team, the emergency general surgery day team, and a combined night float team. This study queried the UAMS Trauma Registry and the Arkansas Clinical Data Repository for all patients meeting the National Trauma Data Bank inclusion criteria from January 1, 2011 to April 30, 2018. Bivariate frequency statistics and generalized linear model were run using STATA V.14.2

Results A total of 11 253 patients (pre-MR, n=6556; post-MR, n=4697) were analyzed in this study. The generalized linear model indicates that implementation of MR resulted in a significant decrease in length of stay (LOS) in trauma patients.

Discussion This study describes an approach to improving patient outcomes in a trauma surgery service of a tertiary care center. The data show how an MR session can allow for patients to get out of the hospital faster; however, broader implications of these sessions have yet to be studied. Further work is needed to describe the decisions being made that allow for a decreased LOS, what dynamics exist between the attendings and the residents in these sessions, and if these sessions can show some of the same benefits in other surgical services.

Level of evidence Level 4, Care Management.

Background

Communication is fundamental to efficient and effective patient hand-offs. These hand-offs from one shift to the next are essential for the continuity of care needed in team-based care models. One area in which the necessity of effective communication cannot be more clearly illustrated is the field of surgery. According to a comprehensive review of over 400 surgical malpractice claims by Greenberg et al,1 more than 13% of these claims were due to communication failures, most of which occurred preoperatively during shift change. It is in academic medical centers that these transitions of care tend to occur rapidly and inefficiently, setting the stage for medical errors to be made.2

As a means to improve patient outcomes, efficiency, and quality, other specialties have instituted daily morning conferences in which patients are handed off between night and day teams.3 Physician survey data suggest that morning report (MR) is an effective way to transition between teams, ensuring all pertinent patient information is accurately communicated.4 This collaboration leads to rapid implementation of the most effective treatment plan for a patient. It has been shown to decrease length of stay (LOS) in medical patients by 2.14±0.81 days, as well as overall hospital charges by $1068.50±$260.98.3

The MR model has yet to be fully characterized in the surgical field, particularly within the practice of acute care surgery (ACS). In a comprehensive study of ACS team hand-offs, Pringle et al 5 described the variability of ACS surgeons using a comprehensive MR model. The culture of patient hand-offs in the Division of ACS at the University of Arkansas for Medical Sciences (UAMS) has drastically changed after the discovery of ineffective time, location, and poor resident education associated with these hand-offs.6 The role of this study was to determine the impact that the newly implemented MR model has had on trauma patient outcomes in a statewide, level 1, tertiary care facility.

Methods

In October 2015, the UAMS Division of ACS implemented MR into their patient hand-off practices. This model consists of attending-supervised patient hand-offs occurring every morning between 06:00 and 07:00. The emergency general surgery service day team, the trauma surgery service day team, and a night float team shared by the two services participate in these sessions. Most recently, the surgical intensive care unit (SICU) team has also been included in this morning conference. All trauma admissions, trauma team activations, trauma consults, and SICU admissions from the previous 24 hours are discussed. A plan of care for each patient is then made for the day.

This study queried the UAMS Trauma Registry and the Arkansas Clinical Data Repository for all patients meeting the National Trauma Data Bank inclusion criteria from January 1, 2011 to April 30 2018.7 Patients were assigned into pre-MR and post-MR cohorts with the entire data set described in terms of age, gender, mode of arrival, mechanism of injury (blunt, burn, or penetrating), injury severity (Injury Severity Score (ISS), Trauma and Injury Severity Score (TRISS), Glasgow Coma Scale (GCS)), hospital disposition (% alive), LOS (overall and intensive care unit (ICU)), and days spent on ventilatory support.

The data set included only those patients admitted to the hospital (LOS≥1). We used the international ISS to differentiate between mild (ISS<15) and moderate to severe (ISS>15) injuries. The ISS cut-off values for mild to moderate/severe injury are based on the increased likelihood of trauma activation with multisystem trauma in those patients scored >15.8

Using STATA V.14.2 a bivariate analysis was performed for differences between the means for patient subgroups in the pre-MR and post-MR time period. Using Student’s t-test for continuous variables and χ2 tests for categorical variables, all subgroups were compared. A generalized linear model was used to evaluate the effect of MR on LOS. Statistical significance was set at α=0.05 for all analyses.

Results

This study included a total of 11 253 patients who met the National Trauma Database inclusion criteria (pre-MR, n=6556; post-MR, n=4697). Table 1 defines the data set in terms of age, gender, mode of arrival, mechanism of injury (blunt, burn, penetrating), injury severity (ISS, New Injury Severity Score (NISS), TRISS), GCS score (<13, 13–15), mortality (% alive), LOS, and vent days.

Table 1
|
Pre versus post inpatient bivariate analysis

The pre-MR and post-MR cohorts were not significantly different in mode of arrival (p=0.932), ISS (p=0.144), NISS (p=0.091), TRISS (p=0.535), or admission GCS score (p=0.14). However, they were different in their age (p<0.0001), gender (p<0.0001), mortality rates (p=0.040), LOS (p<0.0001), ICU LOS (p<0.0001), and number of days on ventilator support (p<0.0001). Specifically, the number of patients who were alive at discharge (% alive) was increased in the post-MR cohort by less than 1%. In post-MR patients, the overall LOS decreased by ~1 day, the ICU LOS decreased by ~0.5 days, and the total vent days decreased by ~0.4 days.

Table 2 presents the results from a generalized linear regression model, which was selected since the dependent variable, LOS, has a positively skewed distribution. This model was used to determine the effect of the post-MR on LOS, while controlling for other patient characteristics. The implementation of MR was found to significantly reduce the LOS in trauma patients. TRISS was also found to significantly reduce the LOS in trauma patients, whereas age, number of comorbidities, moderate GCS (as compared with mild GCS), severe GCS (as compared with mild GCS), and ISS>15 were found to significantly increase LOS in trauma patients.

Table 2
|
Generalized linear regression

Discussion

In this study, we hypothesized that MR would improve outcomes in trauma patients. Analysis of the data showed that the most significant effect of MR was a decrease in the overall LOS between pre-MR and post-MR patients. A further analysis of these patients revealed that LOS in the ICU was also decreased, as well as how many days these patients were required to be on ventilator support.

The importance of a decreased LOS is illustrated by the impact on the patient. Although certain subsets of patients benefit from an extended stay in the ICU, the majority of both medicine and surgery patients are at an increased risk of mortality the longer they spend in the hospital.9 This increased mortality can be attributed to many different factors; however, none have been more extensively studied than infection rates. Barnett et al 10 have described the significant correlation between an increased LOS and hospital-acquired bloodstream infections.

MR also serves an important role in the systematic process of trauma care. It allows for real-time quality improvement. System issues are identified rapidly and corrections are expeditiously made to any inefficiency or error. This process encompasses the entire system of care for trauma patients, extending from the emergency medical service transfer all the way into the operating room (OR) and postoperative care.

Interdisciplinary rounds, such as MR, have been shown to improve interprofessional education and to help deliver a higher quality of patient care.11 These daily conferences allow for individuals from all aspects of trauma care to be involved in the day-to-day healthcare decisions of a patient. All parties are afforded the opportunity to contribute to patient care, which allows for a more efficient and consistent hospital stay for patients.

The main limitation of this study is the inability to stratify confounding variables. These variables include the time it took for a patient to get to the OR and the correlation of specific injuries of a patient to their LOS. One could hypothesize that MR would improve the time it took for a patient to get to the OR, which would explain the decreased LOS for patients in the post-MR cohort. In this study, we broke mechanism of injury into three categories for simplicity of statistical analysis. Injury mechanism is different for every individual, and this categorization generalized injuries for pre-MR and post-MR cohorts. This could have affected the overall LOS results. Ideally, a randomized controlled trial would be conducted in which individuals were randomized into the pre-MR or post-MR cohort. However, this would be unethical given that MR has already been shown to benefit the patient in previous studies.

This study describes an implementable approach to improving patient outcomes in the trauma surgery service of a tertiary care center. The data show how an MR model can decrease LOS; however, broader implications of these sessions have yet to be studied. Further work is needed to describe the specific decisions being made that allow for decreased LOS, what dynamics exist between the attendings and the residents in these sessions, and if these sessions can show some of the same benefits in other surgical services.