Using Electronic Health Record Data to Develop and Validate a Prediction Model for Adverse Outcomes on the Wards

Churpek, Matthew M; Yuen, Trevor C.; Park, Seo Young; Gibbons, Robert D.; Edelson, Dana P.

doi:10.1378/chest.1386290

Cited by 4 publications

(5 citation statements)

References 21 publications

(29 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This renewed investigation of hemodynamic parameters by Raimundo et al (23) is refreshing and fits with a recent call to better understand the physiology of AKI and to not solely rely on biochemical biomarkers (27). This continued exploration of hemodynamic and physiologic parameters mirrors other efforts in critical care medicine, where there is a renewed focus on vital signs and other parameters to help risk stratify and prevent adverse patient outcomes (28,29). Although it is important to stress the need for prospective validation of these data, as a field we can remain hopeful that follow-up investigations will validate the findings of Raimundo et al Additional investigation may even expand the window for intervention and modification of the AKI trajectory to a period of time longer than the initial 12 golden hours reported in this study (23).…”

mentioning

confidence: 75%

The Golden Hours of AKI

Koyner¹

2015

Clinical Journal of the American Society of Nephrology

View full text Add to dashboard Cite

Throughout various fields of scientific investigation, including neonatology and trauma medicine, the idea of the golden hour of treatment has long been debated (1-3). Defined as an interval most critical for successful emergency treatment and improved patient outcomes, the golden hour(s) is contingent on the delivery of early and frequently, protocol-driven care. Although advances in medicine have improved outcomes in these fields, the concept of the golden hour remains loosely applied and increasingly investigated in the setting of AKI. With recent advances in critical care nephrology, including consensus definitions, improved risk stratification, and biomarkers, intensivists and nephrologists are ideally positioned to investigate how best to improve patient outcomes and decrease the severity of renal damage in the golden hours of early AKI (4-6).In the past, AKI research has focused on the earlier identification of AKI before the rise in serum creatinine or drop in urine output (7,8). Recently, investigators have shifted their focus beyond simply replacing serum creatinine with a new gold standard and have sought to augment the information provided by serum creatinine and urine output, combining them with biochemical biomarkers, functional assessment of urine-making capacity, or both (9-11). Although these studies and others have shown that functional and damage biomarkers with physiologic links to the kidney are useful tools in predicting the risk of progression of early AKI, they were part of observational studies and require additional investigation.Additionally, an increasing number of controlled trials has shown efficacy when attempting to prevent severe AKI. Judicious intravenous fluid selection of nonhyperchloremic solutions (12) and remote ischemic preconditioning (13) showed renoprotective effects, which decreased the incidence of severe AKI (regardless of the need for RRT). Similarly, a meta-analysis investigating perioperative hemodynamic optimization, Brienza et al. (14) showed that trials targeting both supranormal and normal hemodynamic targets decreased the risk of postoperative AKI. However, not every controlled trial targeting hemodynamics parameters has shown a clinical benefit in preventing AKI. Two recent large-scale multicenter investigations of resuscitation protocols in the setting of severe sepsis and septic shock showed relatively low severe AKI event rates (2.8%-13.5%) and no difference in the duration of treatment in those requiring RRT (15,16). Although the ideal resuscitation strategy, intravenous fluid selection, and hemodynamic parameters to prevent the development of AKI continue to take shape, much less is known about how best to treat and optimize the patient just entering the golden hours that follow a new diagnosis of AKI.For those with newly diagnosed AKI, there is a dearth of proven interventions to prevent AKI progression. Data surrounding the use of automated, electronic medical record alerts for those with new-onset AKI seemed promising at first (17,18), but when pati...

show abstract

mentioning

confidence: 75%

The Golden Hours of AKI

Koyner¹

2015

Clinical Journal of the American Society of Nephrology

View full text Add to dashboard Cite

show abstract

“…The primary outcomes of interest were the occurrence of critical illness events (defined as a composite of ward cardiac arrest or transfer to ICU), inhospital mortality, and hospital LOS. Cardiac arrests, defined as lack of palpable pulse without resuscitation, were identified using a prospectively validated database maintained by the hospital (Churpek, Yuen, Park, Gibbons, & Edelson, 2014a; Volchenboum et al, 2016). ICU transfer information was derived from the hospital’s admission–discharge–transfer database.…”

Section: Methodsmentioning

confidence: 99%

“…We adjusted for a number of potential confounders (all calculated using data at the time of admission), including patient severity of illness using the electronic Cardiac Arrest Risk Triage (eCART score), a previously published early warning score calculated from vital sign and laboratory result data (Churpek et al, 2014a, 2014b), time of day (day: 7 a.m. to 5 p.m., evening: 5 p.m. to 10 p.m., and night: 10 p.m. to 7 a.m.), weekday versus weekend, year, and the specific ward to which the patient was admitted.…”

Section: Methodsmentioning

confidence: 99%

“…Conversely, patients who are admitted farther away from the nurse’s station or other high-traffic areas (e.g., entrance of the ward) may suffer from delays in care, which could increase their risk of adverse outcomes. However, studies involving patient location within a ward been limited, with most work focusing on patient-level physiological risk factors for the development of critical illness (Churpek et al, 2014a; Knaus et al, 1992; Smith, Prytherch, Schmidt, Featherstone, & Higgins, 2008). Care-level factors have also been extensively studied where decreased staffing, off-hours shift periods, and increased workload have all been associated with poor patient outcomes (Aiken, Clarke, Sloane, Sochalski, & Silber, 2002; Laupland, Shahpori, Kirkpatrick, & Stelfox, 2008; Weissman et al, 2007).…”

mentioning

confidence: 99%

See 1 more Smart Citation

Association Between Room Location and Adverse Outcomes in Hospitalized Patients

Mayampurath

Ward

Fahrenbach

et al. 2018

HERD

View full text Add to dashboard Cite

Objective: To investigate whether a patient’s proximity to the nurse’s station or ward entrance at time of admission was associated with increased risk of adverse outcomes. Method: We conducted a retrospective cohort study of consecutive adult inpatients to 13 medical–surgical wards at an academic hospital from 2009 to 2013. Proximity of admission room to the nurse’s station and to the ward entrance was measured using Euclidean distances. Outcomes of interest include development of critical illness (defined as cardiac arrests or transfer to an intensive care unit), inhospital mortality, and increase in length of stay (LOS). Results: Of the 83,635 admissions, 4,129 developed critical illness and 1,316 died. The median LOS was 3 days. After adjusting for admission severity of illness, ward, shift, and year, we found no relationship between proximity at admission to nurse’s station our outcomes. However, patients admitted to end of the ward had higher risk of developing critical illness (odds ratio [ OR] = 1.15, 95% confidence interval [CI] = [1.08, 1.23]), mortality ( OR = 1.16, 95% CI [1.03, 1.33]), and a higher LOS (13-hr increase, 95% CI [10, 15] hours) compared to patients admitted closer to the ward entrance. Similar results were observed in sensitivity analyses adjusting for isolation room patients and considering patients without room transfers in the first 48 hr. Conclusions: Our study suggests that being away from the nurse’s station did not increase the risk of these adverse events in ward patients, but being farther from the ward entrance was associated with increase in risk of adverse outcomes. Patient safety can be improved by recognizing this additional risk factor.

show abstract

“…A significant limitation in this setting is the lack of any electronically recorded vital signs in the source data. All identified comparison deterioration models (both traditional [7, 17, 18, 19, 20, 21] and deep-learning [10, 15, 16]) rely on patient vital signs and physiological observations as key predictors. We are therefore also aiming to establish the viability of an alternative for predicting short-term patient deterioration where vital signs observations are not available.…”

Section: Introductionmentioning

confidence: 99%

Developing a deep learning system to drive the work of the critical care outreach team

Kennedy

Rihari-Thomas

Dras

et al. 2020

Preprint

View full text Add to dashboard Cite

Care of patients at risk of deterioration on acute medical and surgical wards requires timely identification, increased monitoring and robust escalation procedures. The critical care outreach role brings specialist-trained critical care nurses and physicians into acute wards to facilitate these processes. Performing this role is challenging, as the breadth of information synthesis required is both high and rapidly updating. We propose a novel automated `watch-list' to identify patients at high risk of deterioration, to help prioritise the work of the outreach team. This system takes data from the electronic medical record in real-time and creates a discrete tokenized trajectory, which is fed into a recurrent neural network model. These models achieve an AUROC of 0.928 for in-patient death and 0.778 for unplanned ICU admission (within 24 hours), which compares favourably with existing early warning scores and is comparable with proof of concept deep learning systems requiring significantly more input data.

show abstract

Using Electronic Health Record Data to Develop and Validate a Prediction Model for Adverse Outcomes on the Wards

Cited by 4 publications

References 21 publications

The Golden Hours of AKI

The Golden Hours of AKI

Association Between Room Location and Adverse Outcomes in Hospitalized Patients

Developing a deep learning system to drive the work of the critical care outreach team

Contact Info

Product

Resources

About