Li-Wei H. Lehman

Proton-pump inhibitor use is associated with low serum magnesium concentrations.

Although case reports link proton-pump inhibitor (PPI) use and hypomagnesemia, no large-scale studies have been conducted. Here we examined the serum magnesium concentration and the likelihood of hypomagnesemia (<1.6 mg/dl) with a history of PPI or histamine-2 receptor antagonist used to reduce gastric acid, or use of neither among 11,490 consecutive adult admissions to an intensive care unit of a tertiary medical center. Of these, 2632 patients reported PPI use prior to admission, while 657 patients were using a histamine-2 receptor antagonist. PPI use was associated with 0.012 mg/dl lower adjusted serum magnesium concentration compared to users of no acid-suppressive medications, but this effect was restricted to those patients taking diuretics. Among the 3286 patients concurrently on diuretics, PPI use was associated with a significant increase of hypomagnesemia (odds ratio 1.54) and 0.028 mg/dl lower serum magnesium concentration. Among those not using diuretics, PPI use was not associated with serum magnesium levels. Histamine-2 receptor antagonist use was not significantly associated with magnesium concentration without or with diuretic use. The use of PPI was not associated with serum phosphate concentration regardless of diuretic use. Thus, we verify case reports of the association between PPI use and hypomagnesemia in those concurrently taking diuretics. Hence, serum magnesium concentrations should be followed in susceptible individuals on chronic PPI therapy.

Methods of Blood Pressure Measurement in the ICU

Objective:Minimal clinical research has investigated the significance of different blood pressure monitoring techniques in the ICU and whether systolic vs. mean blood pressures should be targeted in therapeutic protocols and in defining clinical study cohorts. The objectives of this study are to compare real-world invasive arterial blood pressure with noninvasive blood pressure, and to determine if differences between the two techniques have clinical implications. Design:We conducted a retrospective study comparing invasive arterial blood pressure and noninvasive blood pressure measurements using a large ICU database. We performed pairwise comparison between concurrent measures of invasive arterial blood pressure and noninvasive blood pressure. We studied the association of systolic and mean invasive arterial blood pressure and noninvasive blood pressure with acute kidney injury, and with ICU mortality. Setting:Adult intensive care units at a tertiary care hospital. Patients:Adult patients admitted to intensive care units between 2001 and 2007. Interventions:None. Measurements and Main Results:Pairwise analysis of 27,022 simultaneously measured invasive arterial blood pressure/noninvasive blood pressure pairs indicated that noninvasive blood pressure overestimated systolic invasive arterial blood pressure during hypotension. Analysis of acute kidney injury and ICU mortality involved 1,633 and 4,957 patients, respectively. Our results indicated that hypotensive systolic noninvasive blood pressure readings were associated with a higher acute kidney injury prevalence (p = 0.008) and ICU mortality (p < 0.001) than systolic invasive arterial blood pressure in the same range (⩽70 mm Hg). Noninvasive blood pressure and invasive arterial blood pressure mean arterial pressures showed better agreement; acute kidney injury prevalence (p = 0.28) and ICU mortality (p = 0.76) associated with hypotensive mean arterial pressure readings (⩽60 mm Hg) were independent of measurement technique. Conclusions:Clinically significant discrepancies exist between invasive and noninvasive systolic blood pressure measurements during hypotension. Mean blood pressure from both techniques may be interpreted in a consistent manner in assessing patients’ prognosis. Our results suggest that mean rather than systolic blood pressure is the preferred metric in the ICU to guide therapy.

PCoord: network position estimation using peer-to-peer measurements

Several recently emerged Internet services make use of application-level or overlay networks. Examples of such services include overlay multicast, structured peer-to-peer lookup services, and peer-to-peer file sharing. Many of these services could benefit from enabling participating end hosts to estimate their relative network locations within the overlay. We present PCoord, a peer-to-peer network coordinate system for overlay topology discovery and distance prediction. The goal of PCoord is to allow participating peer nodes in an overlay network to collaboratively construct an accurate geometric model of the overlay network topology in a completely decentralized peer-to-peer fashion. We evaluate the PCoord approach through extensive simulations using both real network measurements and simulated topologies. Our results indicate that the constructed geometric model can give accurate pair-wise distance prediction and nearest neighbor discovery. In particular, using a simulated overlay network consisting of over 3,400 peer nodes, our results indicate that over 90% of the peers can predict their closest peers by probing only a small fraction of the global peer population.

Discovering shared cardiovascular dynamics within a patient cohort

Cardiovascular variables such as heart rate (HR) and blood pressure (BP) are robustly regulated by an underlying control system. Time series of HR and BP exhibit distinct dynamical patterns of interaction in response to perturbations (e.g., drugs or exercise) as well as in pathological states (e.g., excessive sympathetic activation). A question of interest is whether “similar” dynamical patterns can be identified across a heterogeneous patient cohort. In this work, we present a technique based on switching linear dynamical systems for identification of shared dynamical patterns in the time series of HR and BP recorded from a patient cohort. The technique uses a mixture of linear dynamical systems, the components of which are shared across all patients, to capture both nonlinear dynamics and non-Gaussian perturbations. We present exploratory results based on a simulation study of the cardiovascular system, and real recordings from 10 healthy subjects undergoing a tilt-table test. These results demonstrate the ability of the proposed technique to identify similar dynamical patterns present across multiple time series.

Patient prognosis from vital sign time series: Combining convolutional neural networks with a dynamical systems approach

In this work, we propose a stacked switching vector-autoregressive (SVAR)-CNN architecture to model the changing dynamics in physiological time series for patient prognosis. The SVAR-layer extracts dynamical features (or modes) from the time-series, which are then fed into the CNN-layer to extract higher-level features representative of transition patterns among the dynamical modes. We evaluate our approach using 8-hours of minute-by-minute mean arterial blood pressure (BP) from over 450 patients in the MIMIC-II database. We modeled the time-series using a third-order SVAR process with 20 modes, resulting in first-level dynamical features of size 20×480 per patient. A fully connected CNN is then used to learn hierarchical features from these inputs, and to predict hospital mortality. The combined CNN/SVAR approach using BP time-series achieved a median and interquartile-range AUC of 0.74 [0.69, 0.75], significantly outperforming CNN-alone (0.54 [0.46, 0.59]), and SVAR-alone with logistic regression (0.69 [0.65, 0.72]). Our results indicate that including an SVAR layer improves the ability of CNNs to classify nonlinear and nonstationary time-series.

Learning outcome-discriminative dynamics in multivariate physiological cohort time series

Model identification for physiological systems is complicated by changes between operating regimes and measurement artifacts. We present a solution to these problems by assuming that a cohort of physiological time series is generated by switching among a finite collection of physiologically-constrained dynamical models and artifactual segments. We model the resulting time series using the switching linear dynamical systems (SLDS) framework, and present a novel learning algorithm for the class of SLDS, with the objective of identifying time series dynamics that are predictive of physiological regimes or outcomes of interest. We present exploratory results based on a simulation study and a physiological classification example of decoding postural changes from heart rate and blood pressure. We demonstrate a significant improvement in classification over methods based on feature learning via expectation maximization. The proposed learning algorithm is general, and can be extended to other applications involving state-space formulations.

A temporal search engine for a massive multi-parameter clinical information database

We describe a novel search engine that is capable of rapid execution of queries concerning changes in the gradients and absolute (and relative) values of multiple irregularly sampled and asynchronous physiological parameters over many time scales. The search engine enables search criteria for multiple physiological parameters using gradient bounds, rates of change, and threshold breeches over various time scales. Multiple signals can be searched and combined in a Boolean manner to form complex queries. Pre-computed ranges and multi-scale gradients are used to significantly reduce the search time for locating temporal events. We have implemented the search engine in MATLAB and tested the algorithm on a massive multi-parameter intensive care unit database (MIMIC II). To illustrate the use of our search approach, a set of numerical search criteria were developed by clinicians to locate evidence for important pathophysiological conditions.

Outcome Prediction for Patients with Traumatic Brain Injury with Dynamic Features from Intracranial Pressure and Arterial Blood Pressure Signals: A Gaussian Process Approach.

Previous work has been demonstrated that tracking features describing the dynamic and time-varying patterns in brain monitoring signals provide additional predictive information beyond that derived from static features based on snapshot measurements. To achieve more accurate predictions of outcomes of patients with traumatic brain injury (TBI), we proposed a statistical framework to extract dynamic features from brain monitoring signals based on the framework of Gaussian processes (GPs). GPs provide an explicit probabilistic, nonparametric Bayesian approach to metric regression problems. This not only provides probabilistic predictions, but also gives the ability to cope with missing data and infer model parameters such as those that control the functions shape, noise level and dynamics of the signal. Through experimental evaluation, we have demonstrated that dynamic features extracted from GPs provide additional predictive information in addition to the features based on the pressure reactivity index (PRx). Significant improvements in patient outcome prediction were achieved by combining GP-based and PRx-based dynamic features. In particular, compared with the a baseline PRx-based model, the combined model achieved over 30 % improvement in prediction accuracy and sensitivity and over 20 % improvement in specificity and the area under the receiver operating characteristic curve.

Blood Pressure and the Risk of Acute Kidney Injury in the ICU: Case-Control Versus Case-Crossover Designs

This chapter describes two different approaches—a case-control and a case-crossover design—to examine the effect of transient exposure of hypotension on the risk of acute kidney injury (AKI) in intensive care unit (ICU) patients. We highlight the key differences and the design rationale of these two approaches, and present preliminary findings from applying these techniques to study the relationship between hypotension and AKI using the MIMIC II database.

A research infrastructure for real-time evaluation of predictive algorithms for intensive care units

In the medical informatics, most algorithms for clinical settings are initially developed and evaluated using retrospective data. However, researchers often lack convenient methods to validate their algorithms with real-time patient data. To bridge the gap between research and clinical applications, we present a research infrastructure that grants researchers access to real-time clinical data without disturbing patient care. The infrastructure is based on the Health Level Seven (HL7) messaging standard. Admission/discharge/transfer and observation result messages (containing lab test results and vital signs) are de-identified in real-time on a hospital network and subsequently transmitted to a research network. Then the processing cluster on the research network can process incoming de-identified data for a wide variety of applications. In the research network, a translator for converting the HL7 data stream into MATLAB format is created for convenient algorithm evaluation. As an example, we evaluated a hypotension predictor using our realtime testing environment. Our infrastructure can easily be replicated in other institutions and has the potential to benefit many researchers in translational medicine.