Classifying early infant feeding status from clinical notes using natural language processing and machine learning.

The objective of this study is to develop and evaluate natural language processing (NLP) and machine learning models to predict infant feeding status from clinical notes in the Epic electronic health records system. The primary outcome was the classification of infant feeding status from clinical notes using Medical Subject Headings (MeSH) terms. Annotation of notes was completed using TeamTat to uniquely classify clinical notes according to infant feeding status. We trained 6 machine learning models to classify infant feeding status: logistic regression, random forest, XGBoost gradient descent, k-nearest neighbors, and support-vector classifier. Model comparison was evaluated based on overall accuracy, precision, recall, and F1 score. Our modeling corpus included an even number of clinical notes that was a balanced sample across each class. We manually reviewed 999 notes that represented 746 mother-infant dyads with a mean gestational age of 38.9 weeks and a mean maternal age of 26.6 years. The most frequent feeding status classification present for this study was exclusive breastfeeding [n = 183 (18.3%)], followed by exclusive formula bottle feeding [n = 146 (14.6%)], and exclusive feeding of expressed mother's milk [n = 102 (10.2%)], with mixed feeding being the least frequent [n = 23 (2.3%)]. Our final analysis evaluated the classification of clinical notes as breast, formula/bottle, and missing. The machine learning models were trained on these three classes after performing balancing and down sampling. The XGBoost model outperformed all others by achieving an accuracy of 90.1%, a macro-averaged precision of 90.3%, a macro-averaged recall of 90.1%, and a macro-averaged F1 score of 90.1%. Our results demonstrate that natural language processing can be applied to clinical notes stored in the electronic health records to classify infant feeding status. Early identification of breastfeeding status using NLP on unstructured electronic health records data can be used to inform precision public health interventions focused on improving lactation support for postpartum patients.

Validation of rapid detection methods for Salmonella enterica in green chile.

The objective of this study is to validate the US Food and Drug Administration (FDA) rea-time polymerase chain reaction (qPCR) assay, the Neogen Amplified Nucleic Single Temperature Reaction (ANSR) assay, and the Vitek ImmunoDiagnostic Assay System (VIDAS) SLM procedure against the FDA cultural procedure for Salmonella detection in green chile pepper. Green chile was artificially contaminated with Salmonella according to the FDA guidelines (FDA. Guidelines for the Validation of Microbiological Methods for the FDA Foods Program, 3rd Edition. 2019. www.fda.gov/media/83812/download?attachment (17 March 2024, date last accessed)) at a fractional recovery level (where 50%-25% tests positive and at a level +1 log greater for each organism tested). Enriched samples were tested directly by the ANSR Salmonella test and by qPCR, and were subcultured into Rappaport-Vassiliadis and tetrathionate brilliant green broth for cultural detection and qPCR. For the VIDAS-SLM assay, the selective enrichments were further cultured in M broth before testing. Presumptive salmonellae were confirmed with biochemical tests, serology, and qPCR. All three rapid assays were compared favorably with the FDA-BAM (Bacteriological Analytical Manual) method. No significant differences at P < .05 were found between the procedures using McNemar's χ2 test. The three procedures were found to be rapid and reliable alternatives to cultural detection of Salmonella enterica in green chile.

Machine learning-based prediction models for home discharge in patients with COVID-19: Development and evaluation using electronic health records.

OBJECTIVE: This study aimed to develop and validate predictive models using electronic health records (EHR) data to determine whether hospitalized COVID-19-positive patients would be admitted to alternative medical care or discharged home. METHODS: We conducted a retrospective cohort study using deidentified data from the University of Florida Health Integrated Data Repository. The study included 1,578 adult patients (≥18 years) who tested positive for COVID-19 while hospitalized, comprising 960 (60.8%) female patients with a mean (SD) age of 51.86 (18.49) years and 618 (39.2%) male patients with a mean (SD) age of 54.35 (18.48) years. Machine learning (ML) model training involved cross-validation to assess their performance in predicting patient disposition. RESULTS: We developed and validated six supervised ML-based prediction models (logistic regression, Gaussian Naïve Bayes, k-nearest neighbors, decision trees, random forest, and support vector machine classifier) to predict patient discharge status. The models were evaluated based on the area under the receiver operating characteristic curve (ROC-AUC), precision, accuracy, F1 score, and Brier score. The random forest classifier exhibited the highest performance, achieving an accuracy of 0.84 and an AUC of 0.72. Logistic regression (accuracy: 0.85, AUC: 0.71), k-nearest neighbor (accuracy: 0.84, AUC: 0.63), decision tree (accuracy: 0.84, AUC: 0.61), Gaussian Naïve Bayes (accuracy: 0.84, AUC: 0.66), and support vector machine classifier (accuracy: 0.84, AUC: 0.67) also demonstrated valuable predictive capabilities. SIGNIFICANCE: This study's findings are crucial for efficiently allocating healthcare resources during pandemics like COVID-19. By harnessing ML techniques and EHR data, we can create predictive tools to identify patients at greater risk of severe symptoms based on their medical histories. The models developed here serve as a foundation for expanding the toolkit available to healthcare professionals and organizations. Additionally, explainable ML methods, such as Shapley Additive Explanations, aid in uncovering underlying data features that inform healthcare decision-making processes.

Feasibility of a Smartwatch Platform to Assess Ecological Mobility: Real-Time Online Assessment and Mobility Monitor.

BACKGROUND: Early detection of mobility decline is critical to prevent subsequent reductions in quality of life, disability, and mortality. However, traditional approaches to mobility assessment are limited in their ability to capture daily fluctuations that align with sporadic health events. We aim to describe findings from a pilot study of our Real-time Online Assessment and Mobility Monitor (ROAMM) smartwatch application, which uniquely captures multiple streams of data in real time in ecological settings. METHODS: Data come from a sample of 31 participants (Mage = 74.7, 51.6% female) who used ROAMM for approximately 2 weeks. We describe the usability and feasibility of ROAMM, summarize prompt data using descriptive metrics, and compare prompt data with traditional survey-based questionnaires or other established measures. RESULTS: Participants were satisfied with ROAMM's function (87.1%) and ranked the usability as "above average." Most were highly engaged (average adjusted compliance = 70.7%) and the majority reported being "likely" to enroll in a 2-year study (77.4%). Some smartwatch features were correlated with their respective traditional measurements (eg, certain GPS-derived life-space mobility features (r = 0.50-0.51, p < .05) and ecologically measured pain (r = 0.72, p = .01), but others were not (eg, ecologically measured fatigue). CONCLUSIONS: ROAMM was usable, acceptable, and effective at measuring mobility and risk factors for mobility decline in our pilot sample. Additional work with a larger and more diverse sample is necessary to confirm associations between smartwatch-measured features and traditional measures. By monitoring multiple data streams simultaneously in ecological settings, this technology could uniquely contribute to the evolution of mobility measurement and risk factors for mobility loss.

Comparisons Between GPS-based and Self-reported Life-space Mobility in Older Adults.

Assessments of Life-space Mobility (LSM) evaluate the locations of movement and their frequency over a period of time to understand mobility patterns. Advancements in and miniaturization of GPS sensors in mobile devices like smartwatches could facilitate objective and high-resolution assessment of life-space mobility. The purpose of this study was to compare self-reported measures to GPS-based LSM extracted from 27 participants (44.4% female, aged 65+ years) who wore a smartwatch for 1-2 weeks at two different site locations (Connecticut and Florida). GPS features (e.g., excursion size/span) were compared to self-reported LSM with and without an indicator for needing assistance. Although correlations between self-reported measures and GPS-based LSM were positive, none were statistically significant. The correlations improved slightly when needing assistance was included, but statistical significance was achieved only for excursion size (r=0.40, P=0.04). The poor correlations between GPS-based and self-reported indicators suggest that they capture different dimensions of life-space mobility.

Detection and isolation of low levels of E. coli O157:H7 in cilantro by real-time PCR, immunomagnetic separation, and cultural methods with and without an acid treatment.

UNLABELLED: Leafy greens such as cilantro, contaminated with Escherichia coli O157:H7, have been implicated in cases of human illnesses. High levels of microflora in fresh cilantro make recovery of low numbers of E. coli O157:H7 difficult. To improve upon current methods, immunomagnetic separation (IMS) techniques in combination with real-time PCR (RTiPCR) and selective enrichment protocols were examined. Rinsates were prepared from cilantro samples inoculated with low (~0.02 CFU/g) and slightly higher (~0.05 CFU/g) levels of E. coli O157:H7. Rinsate portions were enriched in modified buffered peptone water with pyruvate (mBPWp) for 5 h at 37 °C. After 5 h, selective agents were added to samples and further incubated at 42 °C overnight. Detection and recovery were attempted at 5 and 24 h with and without IMS. IMS beads were screened by RTiPCR for simultaneous detection of stx1, stx2, and uidA SNP. Additionally, broth cultures and IMS beads were streaked onto selective agar plates (Rainbow(®) agar, R&F(®) E. coli O157 Chromogenic medium, TC-SMAC and CHROMagar™ 0157) for isolation of E. coli O157:H7. Both broth cultures and IMS beads were also acid treated in Trypticase Soy Broth pH 2 prior to plating to selective media to improve upon cultural recovery. Although E. coli O157 strains were detected in most samples by PCR after 5 h enrichment, cultural recovery was poor. However, after 24 h enrichment, both PCR and cultural recovery were improved. Acidification of the broths and the IMS beads prior to plating greatly improved recovery from 24 h enrichment broths by suppressing the growth of competing microorganisms. PRACTICAL APPLICATION: Detection and recovery of Escherichia coli O157:H7 in fresh produce matrices (e.g., cilantro) can be complicated by high background microflora present in these foods. Rapid detection by molecular methods combined with effective enrichment and isolation procedures such as using immunomagnetic separation (IMS) techniques can quickly identify potential hazards to public health. Additional techniques such as acidification of enrichment broths can exploit acid resistance characteristics of pathogens such as E. coli O157:H7, facilitating their isolation in complex food matrices.

Efficacy of a post enrichment acid treatment for isolation of Escherichia coli O157:H7 from alfalfa sprouts.

The enrichment, detection and isolation procedure in the current US FDA BAM have been shown effective for Escherichia coli O157:H7 in a wide variety of foods. Recently reported modifications to the enrichment protocol, including post-enrichment immunomagnetic separation (IMS) procedures have improved sensitivity of the method for alfalfa sprouts. However, cultural isolation on selective agar plates still presents a challenge in this food matrix. The focus of this study was to reduce levels of competing microflora and enhance isolation of E. coli O157:H7 on selective agars. We optimized the use of a short acid treatment after enrichment and with post-enrichment IMS beads. The optimized acid treatments were then evaluated in alfalfa sprouts artificially contaminated with E. coli O157:H7. After 5h enrichment of alfalfa sprouts contaminated at 0.2cfu/g, there was significant improvement in isolation on the selective plates following acid treatment of two types of IMS beads. Following 24h enrichment of alfalfa sprouts contaminated at low levels, E. coli O157:H7 was only recovered from 8/25 samples when no IMS or acid treatments were used. The use of only the acid treatment improved recovery to 19/25 samples. Following IMS of the enrichment broths, acid treatment increased isolation to 23/25 for Pathatrix™ and 24/25 for BeadRetriever™ IMS. Acid treatment (pH 2) of the enrichment broth for 30min or IMS beads for 7min is a simple and rapid way to greatly improve isolation of E. coli O157 from alfalfa sprout enrichments by reducing the interfering microflora on the selective media.

Optimization and evaluation of a modified enrichment procedure combined with immunomagnetic separation for detection of E. coli O157:H7 from artificially contaminated alfalfa sprouts.

Escherichia coli O157:H7 has been linked to foodborne disease outbreaks with alfalfa sprouts. Detection of the organism in sprouts by standard cultural methods can be difficult due to the high background microflora. The objective of this study was to develop and optimize an enrichment protocol with and without post-enrichment immunomagnetic separation (IMS) for the rapid detection by real-time PCR (RTiPCR) and cultural recovery of E. coli O157:H7 from artificially contaminated alfalfa sprouts. Initially we found that the FDA BAM procedure, enriching samples in modified buffered peptone water with pyruvate and at 37°C for 5h, followed by the addition of acriflavin, cefsulodin and vancomycin (mBPWp+ACV) and static incubation at 42°C gave poor results for both PCR detection and isolation for alfalfa sprouts artificially contaminated at 0.2cfu/g. The addition of post-enrichment IMS improved detection but not isolation. This procedure was modified and optimized by changing to mBPWp with cefsulodin and vancomycin at 42°C and shaking for 24h with and without IMS prior to PCR detection and cultural isolation. Using the resulting protocol we were able to detect E. coli O157:H7 in 100% of samples of alfalfa sprouts contaminated at 0.2cfu/g. This was validated for five strains of E. coli O157:H7. Isolation was 84% without added post-enrichment IMS and 100% with IMS. The optimized procedure was effective for detection and isolation of E. coli O157:H7 from this difficult food matrix.

Detection of E. coli O157:H7 in raw ground beef by Pathatrix™ immunomagnetic-separation, real-time PCR and cultural methods.

Detection of Escherichia coli O157:H7 by conventional cultural methods can be difficult in food matrices with a high background flora such as raw ground beef. Raw ground beef samples, artificially contaminated separately with five strains of E. coli O157:H7 at low (~0.2 cfu/g) and high (~2 cfu/g) levels, were enriched by two enrichment protocols; buffered peptone water (BPW) at 37 °C for 5h and 24h and modified buffered peptone water with pyruvate (mBPWp) for 5h at 37 °C followed by adding selective agents and incubating at 42 °C to 24h. Detection of added E. coli O157:H7 by real-time PCR (RTiPCR) and recovery on isolation agars was performed before and after PATHATRIX™ immunomagnetic separation (IMS). RTiPCR detection and cultural recovery of inoculated E. coli O157:H7 after 5h enrichment were poor at 0.21-0.24 cfu/g. The addition of IMS after 5h enrichment did not improve RTiPCR detection but markedly improved recovery by culturing. By extending enrichment to 24h, RTiPCR detection improved to 76% for either enrichment protocol without IMS. When 24h enrichment was followed by IMS, RTiPCR detection was also further improved. Cultural recovery after 24h enrichment was 56% and 84% without IMS and 100% and 92% after IMS for BPW and mBPWp respectively. Extended enrichment to 24h followed by IMS was found to be sensitive and reliable for detection and cultural recovery from raw ground beef using either enrichment method.

Comparison of Petrifilm Staph Express Count System with the bacteriological analytical manual direct-plating method for enumeration of Staphylococcus aureus in artificially contaminated hard cheese.

The 3M Petrifilm Staph Express Count System was compared with the U.S. Food and Drug Administration's Bacteriological Analytical Manual (BAM) direct-plate count method for the enumeration of Staphylococcus aureus in 6 types of artificially contaminated hard cheese (Asiago, Cheddar, Gruyère, Parmesan, Romano, and Swiss). Five different samples of each cheese type were inoculated with S. aureus (ATCC 25923) to achieve low, medium, and high inoculum levels. S. aureus was enumerated by the Petrifilm and BAM methods, and the results were compared. Multivariate analysis of variance revealed no significant differences (P<0.05) between the 2 methods. The Petrifilm method compared favorably with the BAM procedure. The rapid method was more convenient to use, considerably faster, and less expensive to perform than the BAM method.