Utilizing electronic health records to predict acute kidney injury risk and outcomes: workgroup statements from the 15(th) ADQI Consensus Conference

Big Data and Pediatric Acute Kidney Injury: The Promise of Electronic Health Record Systems

Over the last decade, our understanding of acute kidney injury (AKI) has evolved considerably. The development of a consensus definition standardized the approach to identifying and investigating AKI in children. As a result, pediatric AKI epidemiology has been refined and the consequences of renal injury are better established. Similarly, "big data" methodologies experienced a dramatic evolution and maturation, leading the critical care community to explore potential AKI/big data synergies. One such concept with tremendous potential is electronic health record (EHR) enabled informatics. Much of the promise surrounding these approaches is due to the unique position of the EHR which sits at the intersection of data accumulation and care delivery. EHR data is generated simply via the provision of routine clinical care and should be considered "big" from the standpoint of volume, variety, and velocity as a myriad of diverse elements accumulate rapidly in real time, spontaneously generating an immense dataset. This massive dataset interfaces directly with providers which creates tremendous opportunity. AKI can be diagnosed more accurately, AKI-related care can be optimized, and subsequent outcomes can be improved. Although applying big data concepts to the EHR has proven more challenging than originally thought, we have seen much success and continue to explore its potential. In this review article, we will discuss the EHR in the context of big data concepts, describe approaches applied to date, examine the challenges surrounding optimal application, and explore future directions.

Predicting acute kidney injury at hospital re-entry using high-dimensional electronic health record data

Acute Kidney Injury (AKI), a sudden decline in kidney function, is associated with increased mortality, morbidity, length of stay, and hospital cost. Since AKI is sometimes preventable, there is great interest in prediction. Most existing studies consider all patients and therefore restrict to features available in the first hours of hospitalization. Here, the focus is instead on rehospitalized patients, a cohort in which rich longitudinal features from prior hospitalizations can be analyzed. Our objective is to provide a risk score directly at hospital re-entry. Gradient boosting, penalized logistic regression (with and without stability selection), and a recurrent neural network are trained on two years of adult inpatient EHR data (3,387 attributes for 34,505 patients who generated 90,013 training samples with 5,618 cases and 84,395 controls). Predictions are internally evaluated with 50 iterations of 5-fold grouped cross-validation with special emphasis on calibration, an analysis of which is performed at the patient as well as hospitalization level. Error is assessed with respect to diagnosis, race, age, gender, AKI identification method, and hospital utilization. In an additional experiment, the regularization penalty is severely increased to induce parsimony and interpretability. Predictors identified for rehospitalized patients are also reported with a special analysis of medications that might be modifiable risk factors. Insights from this study might be used to construct a predictive tool for AKI in rehospitalized patients. An accurate estimate of AKI risk at hospital entry might serve as a prior for an admitting provider or another predictive algorithm.

Innovative Use of Novel Biomarkers to Improve the Safety of Renally Eliminated and Nephrotoxic Medications

Over the last decade, the discovery of novel renal biomarkers and research on their use to improve medication effectiveness and safety has expanded considerably. Pharmacists are uniquely positioned to leverage this new technology for renal assessment to improve medication dosing and monitoring. Serum cystatin C is a relatively new, inexpensive, functional renal biomarker that responds more quickly to changing renal function than creatinine and is not meaningfully affected by age, sex, skeletal muscle mass, dietary intake, or deconditioning. Cystatin C has been proposed as an adjunct or alternative to creatinine for glomerular filtration rate assessment and estimation of drug clearance. Tissue inhibitor of metalloproteinase-2·insulin-like growth factor-binding protein 7 ([TIMP-2]·[IGFBP7]) is a composite of two damage biomarkers released into the urine at a checkpoint in mitosis when renal cells undergo stress or sense a future risk of damage. Concentrations of [TIMP-2]·[IGFBP7] increase before a rise in serum creatinine is evident, thus providing insightful information for evaluation in the context of other patient data to predict the risk for impending kidney injury. This article provides a brief overview of novel renal biomarkers being used as a mechanism to improve medication safety including a discussion of cystatin C, as part of drug-dosing algorithms and specifically for vancomycin dosing, and the use of [TIMP-2]·[IGFBP7] for risk prediction in acute kidney injury and drug-induced kidney disease. Select cases of clinical experience with novel renal biomarkers are outlined, and lessons learned and future applications are described.

Electronic Health Record-Enabled Big-Data Approaches to Nephrotoxin-Associated Acute Kidney Injury Risk Prediction

Nephrotoxin-associated acute kidney injury (NTx-AKI) has become one of the most common causes of AKI among hospitalized adults and children; across acute and intensive care populations, exposure to nephrotoxins accounts for 15-25% of AKI cases. Although some interventions have shown promise in observational studies, no treatments currently exist for NTx-AKI once it occurs. Thus, nearly all effective strategies are aimed at prevention. The primary obstacle to prevention is risk prediction and the determination of which patients are more likely to develop NTx-AKI when exposed to medications with nephrotoxic potential. Historically, traditional statistical modeling has been applied to previously recognized clinical risk factors to identify predictors of NTx-AKI. However, increased electronic health record adoption and the evolution of "big-data" approaches to predictive analytics may offer a unique opportunity to prevent NTx-AKI events. This article describes prior and current approaches to NTx-AKI prediction and offers three novel use cases for electronic health record-enabled NTx-AKI forecasting and risk profiling.

Predicting Inpatient Acute Kidney Injury over Different Time Horizons: How Early and Accurate?

Incidence of Acute Kidney Injury (AKI) has increased dramatically over the past two decades due to rising prevalence of comorbidities and broadening repertoire of nephrotoxic medications. Hospitalized patients with AKI are at higher risk for complications and mortality, thus early recognition of AKI is crucial. Building AKI prediction models based on electronic medical records (EMRs) can enable early recognition of high-risk patients, facilitate prevention of iatrogenically induced AKI events, and improve patient outcomes. This study builds machine learning models to predict hospital-acquired AKI over different time horizons using EMR data. The study objectives are to assess (1) whether early AKI prediction is possible; (2) whether information prior to admission improves prediction; (3) what type of risk factors affect AKI prediction the most. Evaluation results showed a good cross-validated AUC of 0.765 for predicting AKI events 1-day prior and adding data prior to admission did not improve model performance.

Acute Kidney Injury in Children

Acute kidney injury (AKI) has become one of the more common complications seen among hospitalized children. The development of a consensus definition has helped refine the epidemiology of pediatric AKI, and we now have a far better understanding of its incidence, risk factors, and outcomes. Strategies for diagnosing AKI have extended beyond serum creatinine, and the most current data underscore the diagnostic importance of oliguria as well as introduce the concept of urinary biomarkers of kidney injury. As AKI has become more widespread, we have seen that it is associated with a number of adverse consequences including longer lengths of stay and greater mortality. Though effective treatments do not currently exist for AKI once it develops, we hope that the diagnostic and definitional strides seen recently translate to the testing and development of more effective interventions.

Hospital acquired Acute Kidney Injury is associated with increased mortality but not increased readmission rates in a UK acute hospital

Background

Acute Kidney Injury (AKI) has evoked much interest over the past decade and is reported to be associated with high inpatient mortality. Preventable death and increased readmission rates related to AKI have been the focus of considerable interest.

Methods

We studied hospital acquired AKI in all emergency hospital admissions, except transfers from ICU to ICU or patients known to renal services, to ascertain mortality and readmission rates, and trackable modifiable factors for death, using cox regression and Kaplan Meier survival curves. Data was extracted from the electronic patient records and a series of case notes reviewed. Admissions were included between April 2006 and March 2010 (and patients followed up until September 2011).

Results

Overall incidence of AKI was 2.2%, (AKI stage 1, 61%, stage 2,27% and stage 3, 12%). In patients who sustain in-hospital AKI, 34% die in hospital, 42% are dead at 90 days and 48% at 1 year post discharge, compared to 12% 1 year mortality in patients without AKI. In multivariable analyses, AKI is an independent risk factor for in-hospital mortality (Hazard Ratio 1.6: 95% confidence intervals 1.43–1.75: P < 0.001), death within 90 days of discharge (Hazard Ratio 1.5: 95% confidence intervals 1.3–1.9: P < 0.001) and subsequent mortality beyond 90 days (Hazard Ratio 2.9: 95% confidence intervals 2.7–3.1: P < 0.001) after adjustment for co-morbidities and peak C-reactive protein.

Thirty percent of the patients who died in the first 90 days post discharge and had AKI, also had malignancy. Readmission rates at 30 and 90 days were not increased by AKI after adjustment for co-morbidities and peak C-reactive protein.

Conclusions

A significant proportion of deaths in the first 90 days post-discharge may not be avoidable, due to malignancy and other end-stage disease. Readmission rates were not higher in patients who had had AKI.

Electronic supplementary material

The online version of this article (10.1186/s12882-017-0729-9) contains supplementary material, which is available to authorized users.

Biomarkers for Early Detection of Acute Kidney Injury

BACKGROUND

Acute kidney injury (AKI) is common in hospitalized patients and is associated with increased morbidity, mortality, and cost. Currently, AKI is diagnosed after symptoms manifest; available diagnostic tests (e.g., serum creatinine, urine microscopy, urine output) have limited ability to identify subclinical AKI. Because of the lack of treatment strategies, AKI typically is managed with supportive measures. However, strategies exist that may prevent renal insults in critically ill patients; therefore, early recognition of AKI is crucial for minimizing damage propagation.

CONTENT

Experimental and clinical studies have identified biomarkers that may facilitate earlier recognition of AKI or even identify patients at risk of AKI. Such biomarkers might aid in earlier implementation of preventive strategies to slow disease progression and potentially improve outcomes. This review describes some of the most promising novel biomarkers of AKI, including neutrophil gelatinase-associated lipocalin (NGAL), kidney injury molecule 1 (KIM-1), interleukin 18 (lL-18), liver-type fatty-acid-binding protein (L-FABP), insulin-like-growth-factor-binding protein 7 (IGFBP7), and tissue inhibitor of metalloproteinase 2 (TIMP-2).

SUMMARY

We discuss biomarker test characteristics, their strengths and weaknesses, and future directions of their clinical implementation.

Technologic Distractions (Part 1): Summary of Approaches to Manage Alert Quantity With Intent to Reduce Alert Fatigue and Suggestions for Alert Fatigue Metrics

OBJECTIVE

To provide ICU clinicians with evidence-based guidance on tested interventions that reduce or prevent alert fatigue within clinical decision support systems.

DESIGN

Systematic review of PubMed, Embase, SCOPUS, and CINAHL for relevant literature from 1966 to February 2017.

PATIENTS

Focus on critically ill patients and included evaluations in other patient care settings, as well.

INTERVENTIONS

Identified interventions designed to reduce or prevent alert fatigue within clinical decision support systems.

MEASUREMENTS AND MAIN RESULTS

Study selection was based on one primary key question to identify effective interventions that attempted to reduce alert fatigue and three secondary key questions that covered the negative effects of alert fatigue, potential unintended consequences of efforts to reduce alert fatigue, and ideal alert quantity. Data were abstracted by two reviewers independently using a standardized abstraction tool. Surveys, meeting abstracts, "gray" literature, studies not available in English, and studies with non-original data were excluded. For the primary key question, articles were excluded if they did not provide a comparator as key question 1 was designed as a problem, intervention, comparison, and outcome question. We anticipated that reduction in alert fatigue, including the concept of desensitization may not be directly measured and thus considered interventions that reduced alert quantity as a surrogate marker for alert fatigue. Twenty-six articles met the inclusion criteria.

CONCLUSION

Approaches for managing alert fatigue in the ICU are provided as a result of reviewing tested interventions that reduced alert quantity with the anticipated effect of reducing fatigue. Suggested alert management strategies include prioritizing alerts, developing sophisticated alerts, customizing commercially available alerts, and including end user opinion in alert selection. Alert fatigue itself is studied less frequently, as an outcome, and there is a need for more precise evaluation. Standardized metrics for alert fatigue is needed to advance the field. Suggestions for standardized metrics are provided in this document.

Electronic health record-based predictive models for acute kidney injury screening in pediatric inpatients

BackgroundAcute kidney injury (AKI) is common in pediatric inpatients and is associated with increased morbidity, mortality, and length of stay. Its early identification can reduce severity.MethodsTo create and validate an electronic health record (EHR)-based AKI screening tool, we generated temporally distinct development and validation cohorts using retrospective data from our tertiary care children's hospital, including children aged 28 days through 21 years with sufficient serum creatinine measurements to determine AKI status. AKI was defined as 1.5-fold or 0.3 mg/dl increase in serum creatinine. Age, medication exposures, platelet count, red blood cell distribution width, serum phosphorus, serum transaminases, hypotension (ICU only), and pH (ICU only) were included in AKI risk prediction models.ResultsFor ICU patients, 791/1,332 (59%) of the development cohort and 470/866 (54%) of the validation cohort had AKI. In external validation, the ICU prediction model had a c-statistic=0.74 (95% confidence interval 0.71-0.77). For non-ICU patients, 722/2,337 (31%) of the development cohort and 469/1,474 (32%) of the validation cohort had AKI, and the prediction model had a c-statistic=0.69 (95% confidence interval 0.66-0.72).ConclusionsAKI screening can be performed using EHR data. The AKI screening tool can be incorporated into EHR systems to identify high-risk patients without serum creatinine data, enabling targeted laboratory testing, early AKI identification, and modification of care.

Impact of e-alert systems on the care of patients with acute kidney injury

With the recent advancement in electronic health record systems and meaningful use of information technology incentive programs (i.e., the American Recovery and Reinvestment Act, the Health Information Technology for Economic and Clinical Health Act, and the Centers for Medicare & Medicaid Services), interest in clinical decision support systems has risen. These systems have been used to examine a variety of different syndromes with variable reported effects. In recent years, electronic alerts (e-alerts) have been implemented at various institutions to decrease the morbidity associated with acute kidney injury (AKI). AKI is common, accounting for 1 in 7 hospital admissions, and is associated with increased length of hospital stay and mortality. AKI is often underrecognized, causing delayed intervention. The use of e-alerts may result in earlier recognition and intervention, as well as decreased morbidity and mortality. This must be balanced with the possibility of increased resource utilization that e-alerts may cause. Before widespread implementation, the ethical and legal consequences of not following e-alert recommendations must be established, and the optimal algorithm for AKI e-alert detection must be determined.

Continuous Renal Replacement Therapy: Forty-year Anniversary

In 1977 Peter Kramer performed the first CAVH (continuous arteriovenous hemofiltration) treatment in Gottingen, Germany. CAVH soon became a reliable alternative to hemo- or peritoneal dialysis in critically ill patients. The limitations of CAVH spurred new research and the discovery of new treatments such as CVVH and CVVHD (continuous veno-venous hemofiltration and continuous veno-venous hemodialysis). The alliance with industry led to development of new specialized equipment with improved accuracy and performance in delivering continuous renal replacement therapies (CRRTs). Machines and filters have progressively undergone a series of technological steps, reaching a high level of sophistication. The evolution of technology has continued, leading to the development and clinical application of new membranes, new techniques and new treatment modalities. With the progress of technology, the entire field of critical care nephrology moved forward, expanding the areas of application of extracorporeal therapies to cardiac, liver and pulmonary support. A great deal of research made extracorporeal therapies an interesting option for the treatment of sepsis and intoxication and the additional use of sorbents was explored. With the progress in understanding the pathophysiology of acute kidney injury (AKI), new guidelines were developed, driving indications, modalities of prescription, monitoring techniques and quality assurance programs. Information technology and precision medicine have recently contributed to further evolution of CRRT, with the possibility of collecting data in large databases and evaluating policies and practice patterns. This is likely to ultimately result in improved patient care. CRRTs are 40 years old today, but they are still young and full of potential for further evolution.

Leveraging Big Data and Electronic Health Records to Enhance Novel Approaches to Acute Kidney Injury Research and Care

While acute kidney injury (AKI) has been poorly defined historically, a decade of effort has culminated in a standardized, consensus definition. In parallel, electronic health records (EHRs) have been adopted with greater regularity, clinical informatics approaches have been refined, and the field of EHR-enabled care improvement and research has burgeoned. Although both fields have matured in isolation, uniting the 2 has the capacity to redefine AKI-related care and research. This article describes how the application of a consistent AKI definition to the EHR dataset can accurately and rapidly diagnose and identify AKI events. Furthermore, this electronic, automated diagnostic strategy creates the opportunity to develop predictive approaches, optimize AKI alerts, and trace AKI events across institutions, care platforms, and administrative datasets.

The intensive care medicine agenda on acute kidney injury

Acute kidney injury (AKI) is a common complication in the critically ill. Current standard of care mainly relies on identification of patients at risk, haemodynamic optimization, avoidance of nephrotoxicity and the use of renal replacement therapy (RRT) in established AKI. The detection of early biomarkers of renal tissue damage is a recent development that allows amending the late and insensitive diagnosis with current AKI criteria. Increasing evidence suggests that the consequences of an episode of AKI extend long beyond the acute hospitalization. Citrate has been established as the anticoagulant of choice for continuous RRT. Conflicting results have been published on the optimal timing of RRT and on the renoprotective effect of remote ischaemic preconditioning. Recent research has contradicted that acute tubular necrosis is the common pathology in AKI, that septic AKI is due to global kidney hypoperfusion, that aggressive fluid therapy benefits the kidney, that vasopressor therapy harms the kidney and that high doses of RRT improve outcome. Remaining uncertainties include the impact of aetiology and clinical context on pathophysiology, therapy and prognosis, the clinical benefit of biomarker-driven interventions, the optimal mode of RRT to improve short- and long-term patient and kidney outcomes, the contribution of AKI to failure of other organs and the optimal approach for assessing and promoting renal recovery. Based on the established gaps in current knowledge the trials that must have priority in the coming 10 years are proposed together with the definition of appropriate clinical endpoints.

AKIpredictor, an online prognostic calculator for acute kidney injury in adult critically ill patients: development, validation and comparison to serum neutrophil gelatinase-associated lipocalin

PURPOSE

Early diagnosis of acute kidney injury (AKI) remains a major challenge. We developed and validated AKI prediction models in adult ICU patients and made these models available via an online prognostic calculator. We compared predictive performance against serum neutrophil gelatinase-associated lipocalin (NGAL) levels at ICU admission.

METHODS

Analysis of the large multicenter EPaNIC database. Model development (n = 2123) and validation (n = 2367) were based on clinical information available (1) before and (2) upon ICU admission, (3) after 1 day in ICU and (4) including additional monitoring data from the first 24 h. The primary outcome was a comparison of the predictive performance between models and NGAL for the development of any AKI (AKI-123) and AKI stages 2 or 3 (AKI-23) during the first week of ICU stay.

RESULTS

Validation cohort prevalence was 29% for AKI-123 and 15% for AKI-23. The AKI-123 model before ICU admission included age, baseline serum creatinine, diabetes and type of admission (medical/surgical, emergency/planned) and had an AUC of 0.75 (95% CI 0.75-0.75). The AKI-23 model additionally included height and weight (AUC 0.77 (95% CI 0.77-0.77)). Performance consistently improved with progressive data availability to AUCs of 0.82 (95% CI 0.82-0.82) for AKI-123 and 0.84 (95% CI 0.83-0.84) for AKI-23 after 24 h. NGAL was less discriminant with AUCs of 0.74 (95% CI 0.74-0.74) for AKI-123 and 0.79 (95% CI 0.79-0.79) for AKI-23.

CONCLUSIONS

AKI can be predicted early with models that only use routinely collected clinical information and outperform NGAL measured at ICU admission. The AKI-123 models are available at http://akipredictor.com/ . Trial registration Clinical Trials.gov NCT00512122.

Drug-associated acute kidney injury: who's at risk?

The contribution of nephrotoxic medications to the development of acute kidney injury (AKI) is becoming better understood concomitant with the increased incidence of AKI in children. Treatment of AKI is not yet available, so prevention continues to be the most effective approach. There is an opportunity to mitigate severity and prevent the occurrence of AKI if children at increased risk are identified early and nephrotoxins are used judiciously. Early detection of AKI is limited by the dependence of nephrologists on serum creatinine as an indicator. Promising new biomarkers may offer early detection of AKI prior to the rise in serum creatinine. Early detection of evolving AKI is improving and offers opportunities for better management of nephrotoxins. However, the identification of patients at increased risk will remain an important first step, with a focus on the use of biomarker testing and interpretation of the results.

Clinical decision support for drug related events: Moving towards better prevention

Clinical decision support (CDS) systems with automated alerts integrated into electronic medical records demonstrate efficacy for detecting medication errors (ME) and adverse drug events (ADEs). Critically ill patients are at increased risk for ME, ADEs and serious negative outcomes related to these events. Capitalizing on CDS to detect ME and prevent adverse drug related events has the potential to improve patient outcomes. The key to an effective medication safety surveillance system incorporating CDS is advancing the signals for alerts by using trajectory analyses to predict clinical events, instead of waiting for these events to occur. Additionally, incorporating cutting-edge biomarkers into alert knowledge in an effort to identify the need to adjust medication therapy portending harm will advance the current state of CDS. CDS can be taken a step further to identify drug related physiological events, which are less commonly included in surveillance systems. Predictive models for adverse events that combine patient factors with laboratory values and biomarkers are being established and these models can be the foundation for individualized CDS alerts to prevent impending ADEs.