Moving Biosurveillance Beyond Coded Data Using AI for Symptom Detection From Physician Notes: Retrospective Cohort Study.

BACKGROUND: Real-time surveillance of emerging infectious diseases necessitates a dynamically evolving, computable case definition, which frequently incorporates symptom-related criteria. For symptom detection, both population health monitoring platforms and research initiatives primarily depend on structured data extracted from electronic health records. OBJECTIVE: This study sought to validate and test an artificial intelligence (AI)-based natural language processing (NLP) pipeline for detecting COVID-19 symptoms from physician notes in pediatric patients. We specifically study patients presenting to the emergency department (ED) who can be sentinel cases in an outbreak. METHODS: Subjects in this retrospective cohort study are patients who are 21 years of age and younger, who presented to a pediatric ED at a large academic children's hospital between March 1, 2020, and May 31, 2022. The ED notes for all patients were processed with an NLP pipeline tuned to detect the mention of 11 COVID-19 symptoms based on Centers for Disease Control and Prevention (CDC) criteria. For a gold standard, 3 subject matter experts labeled 226 ED notes and had strong agreement (F1-score=0.986; positive predictive value [PPV]=0.972; and sensitivity=1.0). F1-score, PPV, and sensitivity were used to compare the performance of both NLP and the International Classification of Diseases, 10th Revision (ICD-10) coding to the gold standard chart review. As a formative use case, variations in symptom patterns were measured across SARS-CoV-2 variant eras. RESULTS: There were 85,678 ED encounters during the study period, including 4% (n=3420) with patients with COVID-19. NLP was more accurate at identifying encounters with patients that had any of the COVID-19 symptoms (F1-score=0.796) than ICD-10 codes (F1-score =0.451). NLP accuracy was higher for positive symptoms (sensitivity=0.930) than ICD-10 (sensitivity=0.300). However, ICD-10 accuracy was higher for negative symptoms (specificity=0.994) than NLP (specificity=0.917). Congestion or runny nose showed the highest accuracy difference (NLP: F1-score=0.828 and ICD-10: F1-score=0.042). For encounters with patients with COVID-19, prevalence estimates of each NLP symptom differed across variant eras. Patients with COVID-19 were more likely to have each NLP symptom detected than patients without this disease. Effect sizes (odds ratios) varied across pandemic eras. CONCLUSIONS: This study establishes the value of AI-based NLP as a highly effective tool for real-time COVID-19 symptom detection in pediatric patients, outperforming traditional ICD-10 methods. It also reveals the evolving nature of symptom prevalence across different virus variants, underscoring the need for dynamic, technology-driven approaches in infectious disease surveillance.

Emergency department visits and boarding for pediatric patients with suicidality before and during the COVID-19 pandemic.

OBJECTIVE: To quantify the increase in pediatric patients presenting to the emergency department with suicidality before and during the COVID-19 pandemic, and the subsequent impact on emergency department length of stay and boarding. METHODS: This retrospective cohort study from June 1, 2016, to October 31, 2022, identified patients ages 6 to 21 presenting to the emergency department at a pediatric academic medical center with suicidality using ICD-10 codes. Number of emergency department encounters for suicidality, demographic characteristics of patients with suicidality, and emergency department length of stay were compared before and during the COVID-19 pandemic. Unobserved components models were used to describe monthly counts of emergency department encounters for suicidality. RESULTS: There were 179,736 patient encounters to the emergency department during the study period, 6,215 (3.5%) for suicidality. There were, on average, more encounters for suicidality each month during the COVID-19 pandemic than before the COVID-19 pandemic. A time series unobserved components model demonstrated a temporary drop of 32.7 encounters for suicidality in April and May of 2020 (p<0.001), followed by a sustained increase of 31.2 encounters starting in July 2020 (p = 0.003). The average length of stay for patients that boarded in the emergency department with a diagnosis of suicidality was 37.4 hours longer during the COVID-19 pandemic compared to before the COVID-19 pandemic (p<0.001). CONCLUSIONS: The number of encounters for suicidality among pediatric patients and the emergency department length of stay for psychiatry boarders has increased during the COVID-19 pandemic. There is a need for acute care mental health services and solutions to emergency department capacity issues.

A computable case definition for patients with SARS-CoV2 testing that occurred outside the hospital.

Objective: To identify a cohort of COVID-19 cases, including when evidence of virus positivity was only mentioned in the clinical text, not in structured laboratory data in the electronic health record (EHR). Materials and Methods: Statistical classifiers were trained on feature representations derived from unstructured text in patient EHRs. We used a proxy dataset of patients with COVID-19 polymerase chain reaction (PCR) tests for training. We selected a model based on performance on our proxy dataset and applied it to instances without COVID-19 PCR tests. A physician reviewed a sample of these instances to validate the classifier. Results: On the test split of the proxy dataset, our best classifier obtained 0.56 F1, 0.6 precision, and 0.52 recall scores for SARS-CoV2 positive cases. In an expert validation, the classifier correctly identified 97.6% (81/84) as COVID-19 positive and 97.8% (91/93) as not SARS-CoV2 positive. The classifier labeled an additional 960 cases as not having SARS-CoV2 lab tests in hospital, and only 177 of those cases had the ICD-10 code for COVID-19. Discussion: Proxy dataset performance may be worse because these instances sometimes include discussion of pending lab tests. The most predictive features are meaningful and interpretable. The type of external test that was performed is rarely mentioned. Conclusion: COVID-19 cases that had testing done outside of the hospital can be reliably detected from the text in EHRs. Training on a proxy dataset was a suitable method for developing a highly performant classifier without labor-intensive labeling efforts.

Pediatric Emergency Department Sepsis Screening Tool Accuracy During the COVID-19 Pandemic.

BACKGROUND: Automated sepsis alerts in pediatric emergency departments (EDs) can identify patients at risk for sepsis, allowing for earlier intervention with appropriate therapies. The impact of the COVID-19 pandemic on the performance of pediatric sepsis alerts is unknown. METHODS: We performed a retrospective cohort study of 59 335 ED visits before the pandemic and 51 990 ED visits during the pandemic in an ED with an automated sepsis alert based on systemic inflammatory response syndrome criteria. The sensitivity, specificity, negative predictive value, and positive predictive value of the sepsis algorithm were compared between the prepandemic and pandemic phases and between COVID-19-negative and COVID-19-positive patients during the pandemic phase. RESULTS: The proportion of ED visits triggering a sepsis alert was 7.0% (n = 4180) before and 6.1% (n = 3199) during the pandemic. The number of sepsis alerts triggered per diagnosed case of hypotensive septic shock was 24 in both periods. There was no difference in the sensitivity (74.1% vs 72.5%), specificity (93.2% vs 94.0%), positive predictive value (4.1% vs 4.1%), or negative predictive value (99.9% vs 99.9%) of the sepsis alerts between these periods. The alerts had a lower sensitivity (60% vs 73.3%) and specificity (87.3% vs 94.2%) for COVID-19-positive versus COVID-19-negative patients. CONCLUSIONS: The sepsis alert algorithm evaluated in this study did not result in excess notifications and maintained adequate performance during the COVID-19 pandemic in the pediatric ED setting.

Population-based surveillance of severe microcephaly and congenital Zika syndrome in Canada.

PURPOSE: To estimate the minimum incidence of congenital Zika syndrome (CZS) and severe microcephaly in Canada and describe key clinical, epidemiological, aetiological and outcome features of these conditions. METHODS: Two separate national surveillance studies were conducted on CZS and severe microcephaly using the well-established Canadian Paediatric Surveillance Program from 2016 to 2019. Over 2700 paediatricians across Canada were surveyed monthly and asked to report demographic details, pregnancy and travel history, infant anthropometry, clinical features and laboratory findings of newly identified cases. Reports were reviewed to assign an underlying aetiology of severe microcephaly. Incidence rates were estimated using monthly live birth denominators. RESULTS: Thirty-four infants met the case definition for severe microcephaly and <5 met the case definition for CZS. The associated minimum incidence rates were 4.5 per 100 000 live births for severe microcephaly and 0.1-0.5 per 100 000 live births for CZS. Of severe microcephaly cases, 53% were attributed to genetic causes, 15% to infectious or ischaemic causes and 32% to unknown causes. The median head circumference-for-age Z-score at birth was -3.2 (IQR -3.8 to -2.6), and catch-up growth was often not achieved. Common clinical features included intracranial abnormalities (n=23), dysmorphology (n=19) and developmental delays (n=14). Mothers of infants with non-genetic aetiologies travelled during pregnancy more often (10/16) than mothers of infants with genetic aetiologies (<5/18; p<0.01). CONCLUSION: Severe microcephaly and CZS are both rare in Canada. Minimum incidence rates can be used as a baseline against which novel or re-emergent causes of severe microcephaly or CZS can be compared.

Central Line-Associated Blood Stream Infections and Non-Central Line-Associated Blood Stream Infections Surveillance in Canadian Tertiary Care Neonatal Intensive Care Units.

OBJECTIVE: To determine if the reported reduction in hospital-acquired infections is due to reduced central line-associated blood stream infections (CLABSI) or non-CLABSIs. STUDY DESIGN: A retrospective cohort study design was used to describe the change in organism pattern and incidence of hospital-acquired infections (CLABSIs and non-CLABSIs) in neonates <33 weeks of gestation admitted to tertiary neonatal intensive care units in the Canadian Neonatal Network between January 1, 2010, and December 31, 2016. Hospital-acquired infection was diagnosed when a pathogenic organism was isolated from blood or cerebrospinal fluid in a neonate with suspected sepsis. CLABSI was diagnosed when a central venous catheter was present at the time or removed in the 2 days before a hospital-acquired infection diagnosis. Cochran-Armitage and Mann-Kendall trend tests and linear regression models were used for statistical analyses. RESULTS: Of 28 144 eligible neonates from 30 Canadian Neonatal Network neonatal intensive care units, 3306 (11.7%) developed hospital-acquired infections. There was a significant decrease in the rate of hospital-acquired infections (14.2% in 2010 and 9.2% in 2016; P < .01), and the rate of both CLABSIs and non-CLABSIs (P < .01) over the study period concomitant with a significant decrease in the duration of central line use (P = .01). The rates of meningitis also decreased during the study period (1.2% in 2010 and 0.9% in 2016; P < .01). Infections owing to gram-positive cocci significantly decreased, but infections owing to gram-negative organisms remained unchanged. CONCLUSION: Although there was a significant decrease in CLABSIs and non-CLABSIs, hospital-acquired infections in preterm neonates remained high. Infections owing to gram-negative organisms remained unchanged and are a target for future preventative efforts.

Influence of paternal and maternal ethnicity and ethnic enclaves on newborn weight.

BACKGROUND: The association between maternal ethnicity and newborn weight is understood. Less is known about the additional influence of paternal ethnicity and neighbourhood ethnic composition. METHODS: We studied 692 301 singleton live births of parents of Canadian, Bangladeshi, Sri Lankan, Pakistani, Indian, Filipino, Vietnamese, Korean, Hong Kong or Chinese birthplace. We used multivariable regression to calculate mean (95% CI) birthweight differences between infants of two Canadian-origin parents and (1) foreign-born mother and Canadian-born father, (2) Canadian-born mother and foreign-born father or (3) two foreign-born parents from the same country. We also stratified by high versus low same-ethnic concentration of the parent's residence. We adjusted for gestational age at birth, maternal age, parity, marital status and income quintile. RESULTS: Compared with male and female infants of two Canadian-born parents, those of same-country foreign-born parents weighed 6.2% (-218 g, 95% CI -214 g to -223 g) and 5.6% (-192 g, 95% CI -187 g to -196 g) less, respectively. The largest mean weight difference was among male (8.4% (-297 g, 95% CI -276 g to -319 g)) and female (8.2% (-279 g, 95% CI -262 g to -296 g)) infants of two Bangladeshi parents. Infants of a foreign-born mother and Canadian-born father had weights closest to those of two Canadian-born parents. Residing in an area of high (vs low) same-ethnic concentration was associated with lower birthweight among infants of mixed union couples, but not among those of parents originating from the same country. CONCLUSIONS: Paternal and maternal ethnic origin influence newborn weight, which is modified by settlement in a high same-ethnic concentration area only among parents of mixed union.

Evaluation of brief screening tools for neurocognitive impairment in HIV/AIDS: a systematic review of the literature.

OBJECTIVE(S): To systematically review literature on brief screening tools used to detect and differentiate between normal cognition and neurocognitive impairment and HIV-associated neurocognitive disorders (HANDs) in adult populations of persons with HIV. DESIGN: A formal systematic review. METHODS: We searched six electronic databases in 2011 and contacted experts to identify relevant studies published through May 2012. We selected empirical studies that focused on evaluating brief screening tools (<20 min) for neurocognitive impairment in persons with HIV. Two reviewers independently reviewed retrieved literature for potential relevance and methodological quality. Meta-analyses were completed on screening tools that had sufficient data. RESULTS: Fifty-one studies met inclusion criteria; we focused on 31 studies that compared brief screening tools with reference tests. Within these 31 studies, 39 tools were evaluated and 67% used a comprehensive neuropsychological battery as a reference. The majority of these studies evaluated HIV-associated dementia (HAD). Meta-analyses demonstrated that the HIV Dementia Scale (HDS) has poor pooled sensitivity (0.48) and the International HIV Dementia Scale (IHDS) has moderate pooled sensitivity (0.62) in detecting a range of cognitive impairment. Five newer screening tools had relatively good sensitivities (>0.70); however, none of the tools differentiated HAND conditions well enough to suggest broader use. There were significant methodological shortcomings noted in most studies. CONCLUSION: HDS and IHDS perform well to screen for HAD but poorly for milder HAND conditions. Further investigation, with improved methodology, is required to understand the utility of newer screening tools for HAND; further tools may need to be developed for milder HAND conditions.

Pituitary volume prospectively predicts internalizing symptoms in adolescence.

BACKGROUND: Early adolescence is a critical time for the development of both internalizing and externalizing disorders. We aimed to investigate whether pituitary volume, an index of hypothalamic-pituitary-adrenal (HPA) axis function, represents a vulnerability factor for the emergence of internalizing and externalizing symptoms during adolescence using a prospective, longitudinal design. METHODS: One hundred and fifty-five adolescents completed 3T structural magnetic resonance imaging (MRI), symptom rating scales and a diagnostic interview during early adolescence (M age 12.6 years, SD .5 years); symptom rating scales were re-administered approximately three years later (M age 15.2 years). The volume of the pituitary gland was estimated by manually delineating its structure on MR images. The degree to which pituitary volumes prospectively predicted change in internalizing and externalizing symptoms across the two time-points was assessed using hierarchal linear regression, after controlling for the influence of gender, age, pubertal stage and intracranial volume. RESULTS: Larger pituitary volumes prospectively predicted an increase in internalizing, but not externalizing, symptoms from early adolescence to mid-adolescence. CONCLUSIONS: This study provides the first evidence that increased pituitary volume might represent a specific vulnerability marker for the development of internalizing symptoms during early to mid-adolescence.