Overtriage, Undertriage, and Value of Care after Major Surgery: An Automated, Explainable Deep Learning-Enabled Classification System

BACKGROUND

In single-institution studies, overtriaging low-risk postoperative patients to ICUs has been associated with a low value of care; undertriaging high-risk postoperative patients to general wards has been associated with increased mortality and morbidity. This study tested the reproducibility of an automated postoperative triage classification system to generating an actionable, explainable decision support system.

STUDY DESIGN

This longitudinal cohort study included adults undergoing inpatient surgery at two university hospitals. Triage classifications were generated by an explainable deep learning model using preoperative and intraoperative electronic health record features. Nearest neighbor algorithms identified risk-matched controls. Primary outcomes were mortality, morbidity, and value of care (inverted risk-adjusted mortality/total direct costs).

RESULTS

Among 4,669 ICU admissions, 237 (5.1%) were overtriaged. Compared with 1,021 control ward admissions, overtriaged admissions had similar outcomes but higher costs ($15.9K [interquartile range $9.8K to $22.3K] vs $10.7K [$7.0K to $17.6K], p < 0.001) and lower value of care (0.2 [0.1 to 0.3] vs 1.5 [0.9 to 2.2], p < 0.001). Among 8,594 ward admissions, 1,029 (12.0%) were undertriaged. Compared with 2,498 control ICU admissions, undertriaged admissions had longer hospital length-of-stays (6.4 [3.4 to 12.4] vs 5.4 [2.6 to 10.4] days, p < 0.001); greater incidence of hospital mortality (1.7% vs 0.7%, p = 0.03), cardiac arrest (1.4% vs 0.5%, p = 0.04), and persistent acute kidney injury without renal recovery (5.2% vs 2.8%, p = 0.002); similar costs ($21.8K [$13.3K to $34.9K] vs $21.9K [$13.1K to $36.3K]); and lower value of care (0.8 [0.5 to 1.3] vs 1.2 [0.7 to 2.0], p < 0.001).

CONCLUSIONS

Overtriage was associated with low value of care; undertriage was associated with both low value of care and increased mortality and morbidity. The proposed framework for generating automated postoperative triage classifications is reproducible.

Speech Pathology Services Are Integral, but Underutilized in Tracheostomy Rehabilitation

STUDY DESIGN

Retrospective case series.

OBJECTIVE

Speech language pathology (SLP) is an underutilized but important component in rehabilitation after tracheostomy. The purpose of this study was to determine rates of SLP utilization and to streamline tracheostomy decannulation to be more efficient and safer through increased utilization of SLP.

METHODS

Adult patients who underwent tracheostomy from April 2016 to December 2018 were evaluated. The primary outcome was completion of any SLP evaluation after tracheostomy, and secondary outcomes were duration from surgery to evaluation, speaking valve and swallow study utilization, downsize and decannulation rates, mean duration of cannulation, and complications.

RESULTS

A total of 255 subjects were included, where 197 (77.3%) underwent SLP evaluation. A minority received a speaking valve (33.7%), while approximately half underwent a swallow study (52.9%). There was a delay in SLP evaluation, with mean duration from surgery to SLP evaluation of 5.9 ± 8.0 days. There was consistent improvement in downsize and decannulation rates in all cohorts that utilized SLP services. Tracheostomy indication of head and neck cancer, trauma, completing a successful swallow study conferred increased odds of eventual decannulation, while obesity and tracheostomy history conferred lower odds. An interdisciplinary decannulation pathway was created, based on literature review and results, to assist in decision-making while progressing toward decannulation.

CONCLUSION

Speech language pathologists are underutilized for rehabilitation of tracheostomy patients, where they are able to offer many skills to diagnose, treat, manage, and troubleshoot, as patients advance through the decannulation process.

The Emergency Surgery Score accurately predicts the need for postdischarge respiratory and renal support after emergent laparotomies: A prospective EAST multicenter study

BACKGROUND

The Emergency Surgery Score (ESS) was recently validated as an accurate mortality risk calculator for emergency general surgery. We sought to prospectively evaluate whether ESS can predict the need for respiratory and/or renal support (RRS) at discharge after emergent laparotomies (EL).

METHODS

This is a post hoc analysis of a 19-center prospective observational study. Between April 2018 and June 2019, all adult patients undergoing EL were enrolled. Preoperative, intraoperative, and postoperative variables were systematically collected. In this analysis, patients were excluded if they died during the index hospitalization, were discharged to hospice, or transferred to other hospitals. A composite variable, the need for RRS, was defined as the need for one or more of the following at hospital discharge: tracheostomy, ventilator dependence, or dialysis. Emergency Surgery Score was calculated for all patients, and the correlation between ESS and RRS was examined using the c-statistics method.

RESULTS

From a total of 1,649 patients, 1,347 were included. Median age was 60 years, 49.4% were men, and 70.9% were White. The most common diagnoses were hollow viscus organ perforation (28.1%) and small bowel obstruction (24.5%); 87 patients (6.5%) had a need for RRS (4.7% tracheostomy, 2.7% dialysis, and 1.3% ventilator dependence). Emergency Surgery Score predicted the need for RRS in a stepwise fashion; for example, 0.7%, 26.2%, and 85.7% of patients required RRS at an ESS of 2, 12, and 16, respectively. The c-statistics for the need for RRS, the need for tracheostomy, ventilator dependence, or dialysis at discharge were 0.84, 0.82, 0.79, and 0.88, respectively.

CONCLUSION

Emergency Surgery Score accurately predicts the need for RRS at discharge in EL patients and could be used for preoperative patient counseling and for quality of care benchmarking.

LEVEL OF EVIDENCE

Prognostic and epidemiological, level III.

The emergency surgery score (ESS) and outcomes in elderly patients undergoing emergency laparotomy: A post-hoc analysis of an EAST multicenter study

INTRODUCTION

We sought to evaluate whether the Emergency Surgery Score (ESS) can accurately predict outcomes in elderly patients undergoing emergent laparotomy (EL).

METHODS

This is a post-hoc analysis of an EAST multicenter study. Between April 2018 and June 2019, all adult patients undergoing EL in 19 participating hospitals were prospectively enrolled, and ESS was calculated for each patient. Using the c-statistic, the correlation between ESS and mortality, morbidity, and need for ICU admission was assessed in three patient age cohorts (65-74, 75-84, ≥85 years old).

RESULTS

715 patients were included, of which 52% were 65-74, 34% were 75-84, and 14% were ≥85 years old; 51% were female, and 77% were white. ESS strongly correlated with postoperative mortality (c-statistic:0.81). Mortality gradually increased from 0% to 20%-60% at ESS of 2, 10 and 16 points, respectively. ESS predicted mortality, morbidity, and need for ICU best in patients 65-74 years old (c-statistic:0.81, 0.75, 0.83 respectively), but its performance significantly decreased in patients ≥85 years (c-statistic:0.72, 0.64, 0.67 respectively).

CONCLUSION

ESS is an accurate predictor of outcome in the elderly EL patient 65-85 years old, but its performance decreases for patients ≥85. Consideration should be given to modify ESS to better predict outcomes in the very elderly patient population.

Serial hemoglobin monitoring in adult patients with blunt solid organ injury: less is more

Background

Patients who sustain blunt solid organ injury to the liver, spleen, or kidney and are treated nonoperatively frequently undergo serial monitoring of their hemoglobin (Hb). We hypothesized that among initially hemodynamically stable patients with blunt splenic, hepatic, or renal injuries treated without an operation, scheduled monitoring of serum Hb values may be unnecessary as hemodynamic instability, not merely Hb drop, would prompt intervention.

Methods

We performed a retrospective review of patients admitted to our urban Level 1 trauma center following blunt trauma with any grade III, IV, or V liver, spleen, or kidney injury from January 1, 2016 to December 31, 2016. Patients who were hemodynamically unstable and went directly to the operating room or interventional radiology were excluded. Patients who required any urgent or unplanned operative or angiographic intervention were compared with patients who did not require an intervention. Routine demographic and outcome variables were obtained and bivariate and multivariate regression statistics were performed using Stata V.10.

Results

A total of 138 patients were included in the study. Age (39.3 vs 41.4, p=0.51), mean injury severity score (26.7 vs 22.1, p=0.12), and admission Hb (11.9 vs 12.8, p=0.06) did not differ significantly between the two groups. The number of Hb draws (9.2 vs 10, p=0.69) and the associated change in Hb (3.7 vs 3.5, p=0.71) did not differ significantly between the two groups. Only splenic grade predicted need for urgent intervention (3.5 vs 2, p<0.001). All patients who required an operative or radiologic intervention did so based on change in hemodynamics or severity of splenic grade, per our institutional protocol, and not Hb trend.

Discussion

Among patients with blunt solid organ injury, a need for emergent intervention in the form of laparotomy or angioembolization occurs within the first hours of injury. Routine scheduled Hb measurements did not change management in our cohort.

Level of evidence

Level III.

Prolonged Antibiotics for Drains After Spine Injury Instrumentation for Trauma: Not Prophylactic or Necessary

BACKGROUND

Antibiotics after spine instrumentation are often extended while the surgical drain is in place, particularly for traumatic injuries. We sought to study if continuing antibiotics past 24 hours affected outcomes.

METHODS

We performed a retrospective observational study of all patients who underwent spine fixation with hardware and surgical drains for trauma at our institution. We compared the effect of perioperative (≤24 hours of antibiotics) versus prolonged (>24 hours) antibiotics on surgical outcomes. Bivariate and multivariable logistic and linear regression statistics were performed.

RESULTS

Three hundred and forty-six patients were included in the analysis. On multivariate analysis, antibiotic duration >24 hours did not predict surgical site infection (odds ratio, 2.68; 95% confidence interval, 0.88-8.10, P = 0.08) or mortality (odds ratio, 0.59; 95% confidence interval, 0.10-3.44; P = 0.56).

CONCLUSIONS

Continuing antibiotics past 24 hours after traumatic spine instrumentation was not associated with improved outcomes. A prospective study to verify these findings may be warranted.

Nail Gun Injury and Endovascular Repair of Cervical Internal Carotid Artery

A male patient aged 49 years presented to the emergency room after sustaining a nail-gun injury to the left neck (Zone III). Computed tomography (CT) angiogram demonstrated retained nail traversing in close proximity to the left internal carotid artery. Catheter angiogram with three-dimensional (3D) reconstruction revealed partial left internal carotid injury without active extravasation and with preserved flow through the vessel. The nail was removed by gentle traction with the simultaneous deployment of stent-graft across the injured segment. Balloon angioplasty of the stent was performed secondary to endoleak and active extravasation. Complete vessel reconstruction with maintained blood flow was achieved. The patient was extubated the following day and was discharged home on hospital day five without neurological deficits. This case report demonstrates the usefulness of endovascular repair of high cervical arterial injuries with special attention to the unique nature of nail gun injuries. 

Traumatic pulmonary artery pseudoaneurysm with concomitant pulmonary embolism

Traumatic pulmonary artery pseudoaneurysms are rare, with only 25 previously documented cases. We present a case of traumatic pulmonary artery pseudoaneurysm in a 27-year-old male with concomitant pulmonary embolism which presented eight days after injury. He was treated with coil embolization of the pseudoaneurysm along with inferior vena cava filter placement. This unique case of concomitant traumatic pulmonary artery pseudoaneurysm and pulmonary embolism made management challenging given the competing priorities.

Nonoperative management of splenic injuries: significance of age

BACKGROUND

In the nonoperative management (NOM) of blunt splenic injuries (BSI), the clinical relevance of age as a risk factor has not been well studied.

METHODS

Using the 2011 National Trauma Data Bank data set, age was analyzed both as a continuous variable and a categorical variable (group 1 [13-54 y], group 2 [55-74 y], and group 3 [≥75 y]). BSI severity was stratified by abbreviated injury scale (AIS): group 1 (AIS ≤2), group 2 (AIS 3), and group 3 (AIS ≥4). A semiparametric proportional odds model was used to model NOM outcomes and effects due to age and BSI severity.

RESULTS

Of 15,113 subjects, 15.3% failed NOM. The odds of failure increased by a factor of 1.014 for each year of age, or factor of 1.5 for groups 2 and 3 each. BSI severity groups 2 and 3 had increases in the odds of failure by factors of 3.9 and 13, respectively, compared with those of group 1. Most failures occurred by 48 h irrespective of age. The effect of age was most pronounced in age groups 2 and 3 with the most severe BSI, where a NOM failure rate of >50% was seen. Both age and failure of NOM were independent predictors of mortality.

CONCLUSIONS

Age is associated with failure of NOM but its effect seems more clinically relevant only in high-grade BSI. Factors that could influence NOM success in elderly patients with high-grade injuries deserve further study.

Analysis of Hypoxemia in Early Ventilator-Associated Pneumonia Secondary to Haemophilus in Trauma Patients

BACKGROUND

Haemophilus species bacteria (HSB) are known pathogens responsible for early pneumonia in intubated trauma patients. The primary goal of this study was to examine the incidence and extent of hypoxemia in intubated trauma patients who develop early ventilator-associated pneumonia (VAP) secondary to HSB. On the basis of our clinical experiences, we hypothesized that patients with Haemophilus species bacteria pneumonia (HSBP) would have a high rate of hypoxemia but that the effect would be transient.

METHODS

Retrospective review of intubated trauma patients from an urban level I trauma center with HSBP diagnosed by deep tracheal aspirate or bronchoalveolar lavage from April 2007 to November 2012. Collected variables included day of HSBP diagnosis; PaO2 to FIO2 ratio (P:F) at HSBP diagnosis as well as HSBP day three and HSBP day seven; injury severity score (ISS) and its component parts; admission Glasgow Coma Scale (GCS) score; and mortality. Hypoxemia was defined as P:F <200. χ(2) Tests were utilized to assess factors that differed between hypoxemic and non-hypoxemic patients; data are presented as median (interquartile range, IQR).

RESULTS

Sixty-nine patients were identified (80% male; age, 35 y [range, 24-49]; ISS 27 [9-59]). Diagnosis of HSBP occurred early (hospital day 4 [range, 3-5]). Forty-three percent of patients had acute respiratory distress syndrome (ARDS) on HSBP day 1; this decreased to 26% on day three and to 30% on day seven. Forty patients (77%) had a tracheostomy performed. Patients with hypoxemia were significantly less likely to have a severe head injury (GCS<9), p<0.05. Patients with hypoxemia had similar hospital length of stay and mortality to patients who did not develop hypoxemia.

CONCLUSION

Haemophilus species bacteria pneumonia in trauma patients is associated with high rates of transient hypoxemia and a high tracheostomy rate, although subsequent outcomes are not affected. Patients with head injuries had a lower incidence of hypoxemia from pneumonia.