A practical algorithm to reduce false critical ECG alarms using arterial blood pressure and/or photoplethysmogram waveforms.

There has been a high rate of false alarms for the critical electrocardiogram (ECG) arrhythmia events in intensive care units (ICUs), from which the 'crying-wolf' syndrome may be resulted and patient safety may be jeopardized. This article presents an algorithm to reduce false critical arrhythmia alarms using arterial blood pressure (ABP) and/or photoplethysmogram (PPG) waveform features. We established long duration reference alarm datasets which consist of 573 ICU waveform-alarm records (283 for development set and 290 for test set) with total length of 551 patent days. Each record has continuous recordings of ECGs, ABP and/or PPG signals and contains one or multiple critical ECG alarms. The average length of a record is 23 h. There are totally 2408 critical ECG alarms (1414 in the development set and 994 in the test set), each of which was manually annotated by experts. The algorithm extracts ABP/PPG pulse features on a beat-by-beat basis. For each pulse, five event feature indicators (EFIs), which correspond to the five critical ECG alarms, are generated. At the time of a critical ECG alarm, the corresponding EFI values of those ABP/PPG pulses around the alarm time are checked for adjudicating (accept/reject) this alarm. The algorithm retains all (100%) the true alarms and significantly reduces the false alarms. Our results suggest that the algorithm is effective and practical on account of its real-time dynamic processing mechanism and computational efficiency.

Alcohol-attributable mortality among American Indians and Alaska Natives in the United States, 1999-2009.

OBJECTIVES: We describe the relative burden of alcohol-attributable death among American Indians/Alaska Natives (AI/ANs) in the United States. METHODS: National Death Index records were linked with Indian Health Service (IHS) registration records to identify AI/AN deaths misclassified as non-AI/AN. We calculated age-adjusted alcohol-attributable death rates from 1999 to 2009 for AI/AN and White persons by sex, age, geographic region, and leading causes; individuals of Hispanic origin were excluded. RESULTS: AI/AN persons had a substantially higher rate of alcohol-attributable death than Whites from 2005 to 2009 in IHS Contract Health Service Delivery Area counties (rate ratio = 3.3). The Northern Plains had the highest rate of AI/AN deaths (123.8/100,000), and the East had the lowest (48.9/100,000). For acute causes, the largest relative risks for AI/AN persons compared with Whites were for hypothermia (14.2) and alcohol poisoning (7.6). For chronic causes, the largest relative risks were for alcoholic psychosis (5.0) and alcoholic liver disease (4.9). CONCLUSIONS: Proven strategies that reduce alcohol consumption and make the environment safer for excessive drinkers should be further implemented in AI/AN communities.

Estimation of the patient monitor alarm rate for a quantitative analysis of new alarm settings.

In many critical care units, default patient monitor alarm settings are not fine-tuned to the vital signs of the patient population. As a consequence there are many alarms. A large fraction of the alarms are not clinically actionable, thus contributing to alarm fatigue. Recent attention to this phenomenon has resulted in attempts in many institutions to decrease the overall alarm load of clinicians by altering the trigger thresholds for monitored parameters. Typically, new alarm settings are defined based on clinical knowledge and patient population norms and tried empirically on new patients without quantitative knowledge about the potential impact of these new settings. We introduce alarm regeneration as a method to estimate the alarm rate of new alarm settings using recorded patient monitor data. This method enables evaluation of several alarm setting scenarios prior to using these settings in the clinical setting. An expression for the alarm rate variance is derived for the calculation of statistical confidence intervals on the results.

The decrease in the unintentional injury mortality disparity between American Indians/Alaska Natives and non-American Indians/Alaska Natives in New Mexico, 1980 to 2009.

OBJECTIVES: We tracked the unintentional injury death disparity between American Indians/Alaska Natives and non-American Indians/Alaska Natives in New Mexico, 1980 to 2009. METHODS: We calculated age-adjusted rates and rate ratios for unintentional injury deaths and their external causes among American Indians/Alaska Natives and non-American Indians/Alaska Natives. We tested trend significance with the Mann-Kendall test. RESULTS: The unintentional injury death rate ratio of American Indians/Alaska Natives to non-American Indians/Alaska Natives declined from 2.9 in 1980-1982 to 1.5 in 2007-2009. The rate among American Indians/Alaska Natives decreased 47.2% from 1980-1982 to 1995-1997. Among non-American Indians/Alaska Natives, the rate declined 25.3% from 1980-1982 to 1992-1994, then increased 31.9% from 1992-1994 to 2007-2009. The motor vehicle traffic and pedestrian death rates decreased 57.8% and 74.6%, respectively, among American Indians/Alaska Natives from 1980-1982 to 2007-2009. CONCLUSIONS: The unintentional injury death rate disparity decreased substantially from 1980-1982 to 2007-2009 largely because of the decrease in motor vehicle crash and pedestrian death rates among American Indians/Alaska Natives and the increase in the poisoning death rate among non-American Indians/Alaska Natives.

Physiologic monitoring alarm load on medical/surgical floors of a community hospital.

UNLABELLED: It has been known to the public that high frequency of false and/or unnecessary alarms from patient monitoring devices causes "alarm fatigue" in critical care. But little is known about the impact to care on the less acute patients located outside the critical care areas, such as the traditional medical/surgical (med/surg) floor. METHODS: As part of a larger population management study, we initiated continuous physiological monitoring to 79 beds of floor patients in a community hospital. In order to qualify the patient monitoring alarm load for subacute medical and surgical floor patients, we assessed alarm data from April 2009 to January 2010. A standard critical care monitoring system (Philips IntelliVue MP-5 and Telemetry) was installed and set to the default alarm limits. All waveform data available for the patient (typically ECG, RESP, PPG at 125hz 8 bit), all alarm conditions declared by the monitoring system, and 1 minute parameter trend data were saved to disk every 8 hours for all patients. A monitoring care protocol was created to determine whether the patient was monitored via the hardwired bedside or wirelessly via telemetry. Alarms were not announced on the care unit but instead notifications were the responsibility of remote telehealth center personnel. We retrospectively evaluated the frequency of alarms over specific physiologic thresholds (n= 4104 patients) and conducted adjudication of all alarms based on a smaller sampling (n=30 patients). RESULTS: For all patients, the average hours of monitoring per patient were 16.5 hours with a standard deviation (s) of 8.3 hours and a median of 22 hours. The average number of alarms (all severities) per patient was 69.7 (s =90.3, median =28) alarms. When this is adjusted to the duration of monitoring, the average per patient, per day rate was 95.6 (s =124.2, median =34.2) alarms. The adjudicated sample (n=30 patients) resulted in 34% of critical alarms (lethal arrhythmias, extreme high or low heart rate [HR], extreme desaturation, apnea) being true and 63% of the high priority alarms (high or low HR, high or low RR, Low SpO(2), pause, Missed Beat, Pair PVCs, Pacer Not Pace, Non Sustain VT, Irregular HR, Multiform) being true. Analysis of alarm history resulted in the ability to reduce the HR alarm load by more than 50% with a simple limit adjustment of high HR from 120 to 130 bpm and a 36% or 65% reduction in SpO(2) alarm load by reducing the SpO(2) limit from 90% to 85% or 80% respectively. CONCLUSION: 1) Standard critical care alarm limits appear be too sensitive for subacute care areas of the hospital. 2) For most patients these alarm limits do not create a significant alarm load; however, for a small number of patients they cause a significant alarm load. 3) Alarm loads can be controlled with alarm limit settings appropriate to the population. 4) Current technology for HR and SpO(2) appear suitable for continuous monitoring of this population.

Heuristics to determine ventilation times of ICU patients from the MIMIC-II database.

Mechanical ventilation is an important life support tool for patients in intensive care units (ICU). For various research purposes related to patient hemodynamic and cardiopulmonary monitoring, it is important to know when a patient is on a ventilator. Unfortunately, the widely used MIMIC-II database contains results from user charted data, where the user did not always store ventilation on and off times explicitly and accurately. The resulting ventilation-related data are subject to error. Therefore, there are no simple rules to define ventilation times retrospectively for this dataset. Hence, we designed a simple set of rules to determine the ventilation times using multiple sources of mechanical ventilator-related settings and physiological measurements by expert heuristics. The rules worked well in comparison with nursing notes regarding ventilation events. We conclude that our rule sets for determining ventilation times may be useful in assisting with MIMIC-II database analysis.

A vasopressor advisability index for hemodynamic monitoring.

To predict the need for vasopressor therapy, arterial blood pressure and heart rate are taken as inputs to a fuzzy-logic based algorithm that generates a 'vasopressor advisability index'. We evaluated the algorithm against patient records from a large intensive care database. The algorithm achieved a sensitivity of 69.7% and a specificity of 84.2% on a test set that was independent of the training set used.

Development and evaluation of predictive alerts for hemodynamic instability in ICU patients.

This paper describes an algorithm for identifying ICU patients that are likely to become hemodynamically unstable. The algorithm consists of a set of rules that trigger alerts. Unlike most existing ICU alert mechanisms, it uses data from multiple sources and is often able to identify unstable patients earlier and with more accuracy than alerts based on a single threshold. The rules were generated using a machine learning technique and were tested on retrospective data in the MIMIC II ICU database, yielding a specificity of approximately 0.9 and a sensitivity of 0.6.

Reducing false alarm rates for critical arrhythmias using the arterial blood pressure waveform.

BACKGROUND: Over the past two decades, high false alarm (FA) rates have remained an important yet unresolved concern in the Intensive Care Unit (ICU). High FA rates lead to desensitization of the attending staff to such warnings, with associated slowing in response times and detrimental decreases in the quality of care for the patient. False arrhythmia alarms are commonly due to single channel ECG artifacts and low voltage signals, and therefore it is likely that the FA rates may be reduced if information from other independent signals is used to form a more robust hypothesis of the alarm's etiology. METHODS: A large multi-parameter ICU database (PhysioNet's MIMIC II database) was used to investigate the frequency of five categories of false critical ("red" or "life-threatening") ECG arrhythmia alarms produced by a commercial ICU monitoring system, namely: asystole, extreme bradycardia, extreme tachycardia, ventricular tachycardia and ventricular fibrillation/tachycardia. Non-critical ("yellow") arrhythmia alarms were not considered in this study. Multiple expert reviews of 5386 critical ECG arrhythmia alarms from a total of 447 adult patient records in the MIMIC II database were made using the associated 41,301 h of simultaneous ECG and arterial blood pressure (ABP) waveforms. An algorithm to suppress false critical ECG arrhythmia alarms using morphological and timing information derived from the ABP signal was then tested. RESULTS: An average of 42.7% of the critical ECG arrhythmia alarms were found to be false, with each of the five alarm categories having FA rates between 23.1% and 90.7%. The FA suppression algorithm was able to suppress 59.7% of the false alarms, with FA reduction rates as high as 93.5% for asystole and 81.0% for extreme bradycardia. FA reduction rates were lowest for extreme tachycardia (63.7%) and ventricular-related alarms (58.2% for ventricular fibrillation/tachycardia and 33.0% for ventricular tachycardia). True alarm (TA) reduction rates were all 0%, except for ventricular tachycardia alarms (9.4%). CONCLUSIONS: The FA suppression algorithm reduced the incidence of false critical ECG arrhythmia alarms from 42.7% to 17.2%, where simultaneous ECG and ABP data were available. The present algorithm demonstrated the potential of data fusion to reduce false ECG arrhythmia alarms in a clinical setting, but the non-zero TA reduction rate for ventricular tachycardia indicates the need for further refinement of the suppression strategy. To avoid suppressing any true alarms, the algorithm could be implemented for all alarms except ventricular tachycardia. Under these conditions the FA rate would be reduced from 42.7% to 22.7%. This implementation of the algorithm should be considered for prospective clinical evaluation. The public availability of a real-world ICU database of multi-parameter physiologic waveforms, together with their associated annotated alarms is a new and valuable research resource for algorithm developers.

Predicting ICU hemodynamic instability using continuous multiparameter trends.

BACKGROUND: Identifying hemodynamically unstable patients in a timely fashion in intensive care units (ICUs) is crucial because it can lead to earlier interventions and thus to potentially better patient outcomes. Current alert algorithms are typically limited to detecting dangerous conditions only after they have occurred and suffer from high false alert rates. Our objective was to predict hemodynamic instability at least two hours before a major clinical intervention (e.g., vasopressor administration), while maintaining a low false alert rate. STUDY POPULATION: From the MIMIC II database, containing ICU minute-by-minute heart rate (HR) and invasive arterial blood pressure (BP) monitoring trend data collected between 2001 and 2005, we identified 132 stable and 104 unstable patients that met our stability-instability criteria and had sufficient data points. METHOD: We first derived additional physiological parameters of shock index, rate pressure product, heart rate variability, and two measures of trending based on HR and BP. Then we developed 220 statistical features and systematically selected a small set to use for classification. We applied multi-variable logistic regression modeling to do classification and implemented validation via bootstrapping. RESULTS: Area under receiver-operating curve (ROC) 0.83+/-0.03, sensitivity 0.75+/-0.06, and specificity 0.80+/-0.07; if the specificity is targeted at 0.90, then the sensitivity is 0.57+/-0.07. Based on our preliminary results, we conclude that the algorithms we developed using HR and BP trend data may provide a promising perspective toward reliable predictive alerts for hemodynamically unstable patients.

Predicting respiratory instability in the ICU.

Acute lung injury (ALI) and acute respiratory distress syndrome (ARDS) contribute to the morbidity and mortality of intensive care patients worldwide, and have large associated human and financial costs. We identified a reference data set of 624 mechanically-ventilated patients in the MIMIC-II intensive care database with and without low PaO(2)/FiO(2) ratios (termed respiratory instability), and developed prediction algorithms for distinguishing these patients prior to the critical event. In the end, we had four rule sets using mean airway pressure, plateau pressure, total respiratory rate and oxygen saturation (SpO(2)), where the specificity/sensitivity rates were either 80%/60% or 90%/50%.

Expression of CINC-2beta is related to the state of differentiation of alveolar epithelial cells.

Alveolar epithelial cells are among the first cells to encounter inhaled particles or organisms. These cells likely participate in the initiation and modulation of the inflammatory response by production of chemokines. However, there is little information on the extent or regulation of chemokine production by these cells. Rat type II cells were studied under differentiated and dedifferentiated conditions to determine their ability to express and secrete CXC chemokines. Both differentiated and dedifferentiated type II cells secreted MIP-2, MCP-1, and CINC-2 in response to a cytokine mixture of IL-1beta, TNF-alpha, and IFN-gamma or to IL-1beta alone. The cytokine mixture also induced iNOS expression and nitrite secretion. Both differentiated and dedifferentiated type II cells expressed CINC-1 (GRO), CINC-2alpha, CINC-3 (MIP-2), and MCP-1 mRNA, and their expression was increased by the cytokine mixture or by IL-1beta alone. However, CINC-2beta, a splice variant of CINC-2, was only expressed under differentiated conditions stimulated by KGF and was not increased by the cytokine mixture or by IL-1beta. In situ hybridization of normal lung and lung instilled with Ad-KGF demonstrated that CINC-2beta was expressed by alveolar and bronchiolar epithelial cells in vivo. We conclude that CINC-2beta is regulated differently from most other chemokines and that its expression is related to the state of alveolar type II cell differentiation.

Transforming growth factor-beta antagonizes alveolar type II cell proliferation induced by keratinocyte growth factor.

Keratinocyte growth factor (KGF) is a mitogen for rat type II cells and also stimulates differentiation in vitro. Administration of KGF also protects the lung from a variety of injuries and subsequent development of fibrosis. Because transforming growth factor (TGF)-beta has been shown to inhibit epithelial cell proliferation and surfactant protein gene expression in other systems and is thought to be a major effector in pulmonary fibrosis, we sought to determine if TGF-beta would antagonize the effects of KGF in primary cultures of alveolar type II cells. Type II cells were cultured on a matrix of type I collagen and Matrigel in the presence or absence of KGF and/or TGF-beta. KGF alone greatly stimulated proliferation and increased cyclin-dependent kinase (cdk) 2 kinase activity and Retinoblastoma susceptibility gene product (Rb) phosphorylation. Cyclin D1, cdk2, and cdc25A protein levels were increased, and p15(Ink4b) and p27(Kip1) protein levels were decreased. TGF-beta markedly inhibited alveolar epithelial cell proliferation induced by KGF. TGF-beta inhibited cdk2 enzyme activity and Rb phosphorylation and increased p15(Ink4b) protein levels. TGF-beta also inhibited differentiation induced by KGF as measured by secretion of surfactant protein-A into the apical media. In summary, TGF-beta inhibits the proliferative effect of KGF in vitro and may be a biologic antagonist of KGF.

Lipogenesis in fetal rat lung: importance of C/EBPalpha, SREBP-1c, and stearoyl-CoA desaturase.

Alveolar type II cells increase lipogenesis and convert glycogen into the phospholipids of surfactant in the late term fetal lung. Recent studies suggest that CCAAT/enhancing-binding protein (C/EBP) isoforms and sterol regulatory element binding protein (SREBP)-1c regulate fatty acid synthesis in adult type II cells in vitro. To define the temporal relationships and enzymes involved in lipogenesis in fetal rat lung, the mRNA levels of selected transcription factors and enzymes were determined. There was an increase in the mRNA levels of C/EBPalpha, C/EBPbeta, C/EBPdelta, peroxisomal proliferator-activated receptor gamma (PPARgamma), and SREBP-1c, but not SREBP-1a or SREBP-2 from fetal Days 19-21. There was also an increase in the mRNA levels of fatty acid synthase, stearoyl-CoA desaturase 1 (SCD-1), fatty acid translocase, glycerol-3-P acyl transferase, and phosphatidate cytidylyltransferase. By in situ hybridization, there was detectible expression of fatty acid synthase, SCD-1, and C/EBPalpha along the alveolar septae with the same distribution pattern as surfactant protein-C, whereas PPARgamma expression appeared to be restricted to macrophages. Regulation of lipogenesis at the mRNA level is predominately on enzymes of fatty acid synthesis and appears to be regulated by C/EBPalpha and SREBP-1c. SCD-1 and phosphatidate cytidylyltransferase are important components of the lipogenic response in the fetal lung that have not been recognized previously.

Serum levels of surfactant protein D are increased in mice with lung tumors.

Most murine lung tumors are composed of differentiated epithelial cells. We have reported previously that surfactant protein (SP)-D is expressed in urethane-induced tumors. Serum levels of SP-D are increased in patients with interstitial lung disease and acute respiratory distress syndrome and in rats with acute lung injury but have not been measured in mice. In this study, we sought to determine whether SP-D could be detected in murine serum and discovered that it was increased in mice bearing lung tumors. Serum SP-D concentration was 5.0 +/- 0.2 ng/ml in normal C57BL/6 mice, essentially absent in SP-D nulls, and 63.6 +/- 9.0 ng/ml in SP-D-overexpressing mice. SP-D in serum was verified by immunoblotting. Serum SP-D was increased in mice bearing tumors induced by three different protocols, and the SP-D level correlated with tumor volume. However, in mice with a single adenoma or a few adenomas, SP-D levels were usually within the normal range. SP-D was expressed by the tumor cells, and there was also a field effect whereby type II cells near the tumor expressed more SP-D than type II cells in the remainder of the lung. Serum SP-D was also increased by lung inflammation. In airway inflammation induced by aerosolized ovalbumin in sensitized BALB/c mice, the serum levels were elevated after challenge. In conclusion, serum SP-D concentration is increased in mice bearing lung tumors and generally reflects the tumor burden but is also elevated during lung inflammation.

Keratinocyte growth factor and the transcription factors C/EBP alpha, C/EBP delta, and SREBP-1c regulate fatty acid synthesis in alveolar type II cells.

Strategies to stimulate endogenous surfactant production require a detailed understanding of the regulation of lipogenesis in alveolar type II cells. We developed culture conditions in which keratinocyte growth factor (KGF) stimulates fatty acid and phospholipid synthesis. KGF stimulated acetate incorporation into phosphatidylcholine, disaturated phosphatidylcholine, and phosphatidylglycerol more than 5% rat serum alone. To determine the mRNA levels of lipogenic enzymes and transport proteins, we analyzed gene expression by oligonucleotide microarrays. KGF increased the mRNA levels for fatty acid synthase, stearoyl-CoA desaturase-1 (SCD-1), and epidermal fatty acid-binding protein more than rat serum alone. In addition, KGF increased the mRNA levels of the transcription factors CCAAT/enhancer-binding protein alpha (C/EBPalpha) and C/EBPdelta as well as SREBP-1c (ADD-1), but not PPARgamma. These changes in C/EBPalpha and C/EBPdelta were confirmed by in situ hybridization. SCD-1 was also found to be highly expressed in alveolar type II cells in vivo. Furthermore, KGF increased protein levels of fatty acid synthase, C/EBPalpha, C/EBPdelta, SREBP-1, epidermal fatty acid-binding protein, and SCD. Finally, the liver X receptor agonist T0901317 increased acetate incorporation and SREBP-1 but not SREBP-2 protein levels. In summary, KGF stimulates lipogenesis in type II cells by a coordinated expression of lipogenic enzymes and transport proteins regulated by C/EBP isoforms and SREBP-1c.

Serum SP-D is a marker of lung injury in rats.

Pulmonary surfactant protein D (SP-D) is expressed in alveolar type II and bronchiolar epithelial cells and is secreted into alveoli and conducting airways. However, SP-D has also been measured in serum and is increased in patients with acute respiratory distress syndrome, pulmonary fibrosis, and alveolar proteinosis. To demonstrate that SP-D can be measured in rat serum, we instilled rats with keratinocyte growth factor, which produces type II cell hyperplasia and an increase in SP-D in bronchoalveolar lavage fluid (BALF). To evaluate serum SP-D as a biomarker of lung injury, we examined several injury models. In rats treated with 1 unit of bleomycin, serum SP-D was elevated on days 3, 7, 14, and 28 after instillation, and SP-D mRNA was increased in focal areas as detected by in situ hybridization. However, there was no increase in whole lung SP-D mRNA when the expression was normalized to whole lung 18S rRNA. After instillation of 2 units of bleomycin, the serum levels of SP-D were higher, and SP-D was also increased in BALF and lung homogenates. In another model of subacute injury, serum SP-D was increased in rats treated with paraquat plus oxygen. Finally to evaluate acute lung injury, we instilled rats with HCl; SP-D was increased at 4 h after instillation. Our data indicate that serum SP-D may be a useful indicator of lung injury and type II cell hyperplasia in rats.

Maintenance of surfactant protein A and D secretion by rat alveolar type II cells in vitro.

Secretion of surfactant proteins A and D (SP-A and SP-D) has been difficult to study in vitro because a culture system for maintaining surfactant secretion has been difficult to establish. We evaluated several growth factors, corticosteroids, rat serum, and a fibroblast feeder layer for the ability to produce and maintain a polarized epithelium of type II cells that secretes SP-A and SP-D into the apical medium. Type II cells were plated on a filter insert coated with an extracellular matrix and were cultured at an air-liquid interface. Keratinocyte growth factor (KGF) stimulated type II cell proliferation and secretion of SP-A and SP-D more than fibroblast growth factor-10 (FGF-10), hepatocyte growth factor (HGF), or heparin-binding epidermal-like growth factor (HB-EGF). Cells cultured in the presence of KGF and rat serum with or without fibroblasts had high surfactant protein mRNA levels and exhibited a high level of SP-A and SP-D secretion. Dexamethasone inhibited type II cell proliferation but increased expression of SP-B. In the presence of KGF, rat serum, and dexamethasone, the mRNAs for the surfactant proteins were maintained at high levels. Secretion of SP-A and SP-D was found to be independent of phospholipid secretion.