Utilization of the Signature Method to Identify the Early Onset of Sepsis From Multivariate Physiological Time Series in Critical Care Monitoring.

OBJECTIVES: Patients in an ICU are particularly vulnerable to sepsis. It is therefore important to detect its onset as early as possible. This study focuses on the development and validation of a new signature-based regression model, augmented with a particular choice of the handcrafted features, to identify a patient's risk of sepsis based on physiologic data streams. The model makes a positive or negative prediction of sepsis for every time interval since admission to the ICU. DESIGN: The data were sourced from the PhysioNet/Computing in Cardiology Challenge 2019 on the "Early Prediction of Sepsis from Clinical Data." It consisted of ICU patient data from three separate hospital systems. Algorithms were scored against a specially designed utility function that rewards early predictions in the most clinically relevant region around sepsis onset and penalizes late predictions and false positives. SETTING: The work was completed as part of the PhysioNet 2019 Challenge alongside 104 other teams. PATIENTS: PhysioNet sourced over 60,000 ICU patients with up to 40 clinical variables for each hour of a patient's ICU stay. The Sepsis-3 criteria was used to define the onset of sepsis. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: The algorithm yielded a utility function score which was the first placed entry in the official phase of the challenge.

Colorectal Cancer Screening Patient Navigation for Patients with Mental Illness and/or Substance Use Disorder: Pilot Randomized Control Trial.

OBJECTIVE: Colorectal cancer (CRC) is the second leading cause of cancer death in the US. Screening has decreased CRC mortality. However, disadvantaged patients, particularly those with mental illness or substance use disorder (SUD), are less likely to be screened. The aim of this trial was to evaluate the impact of a patient navigation program on CRC screening in patients with mental illness and/or SUD. METHODS: A pilot randomized nonblinded controlled trial was conducted from January to June 2017 in an urban community health center serving a low-income population. We randomized 251 patients aged 50-74 years with mental illness and/or SUD diagnosis overdue for CRC screening to intervention (n = 126) or usual care (n = 125) stratified by mental illness, SUD, or dual diagnosis. Intervention group patients received a letter followed by a phone call from patient navigators. Navigators helped patients overcome their individual barriers to CRC screening including: education, scheduling, explanation of bowel preparation, lack of transportation or accompaniment to appointments. If patient refused colonoscopy, navigators offered fecal occult blood testing. The main measure was proportion of patients completing CRC screening in intervention and usual care groups. RESULTS: Navigators contacted 85 patients (67%) in the intervention group and 26 declined to participate. In intention-to treat analysis, more patients in the intervention group received CRC screening than in the usual care group, 19% versus 10.4% (p = .04). Among 56 intervention patients who received navigation, 19 completed screening (33.9% versus 10.4% in the control group, p = .001). In the subgroup of patients with SUD, 20% in the intervention group were screened compared to none in the usual care group (p = .05). CONCLUSIONS: A patient navigation program improved CRC screening rates in patients with mental illness and/or SUD. Larger studies in diverse care settings are needed to demonstrate generalizability and explore which modality of CRC screening is most acceptable and which navigator activities are most effective for this vulnerable population. TRIALS REGISTRATION NUMBER: 2016P001322.

Subtle variation in sepsis-III definitions markedly influences predictive performance within and across methods.

Early detection of sepsis is key to ensure timely clinical intervention. Since very few end-to-end pipelines are publicly available, fair comparisons between methodologies are difficult if not impossible. Progress is further limited by discrepancies in the reconstruction of sepsis onset time. This retrospective cohort study highlights the variation in performance of predictive models under three subtly different interpretations of sepsis onset from the sepsis-III definition and compares this against inter-model differences. The models are chosen to cover tree-based, deep learning, and survival analysis methods. Using the MIMIC-III database, between 867 and 2178 intensive care unit admissions with sepsis were identified, depending on the onset definition. We show that model performance can be more sensitive to differences in the definition of sepsis onset than to the model itself. Given a fixed sepsis definition, the best performing method had a gain of 1-5% in the area under the receiver operating characteristic (AUROC). However, the choice of onset time can cause a greater effect, with variation of 0-6% in AUROC. We illustrate that misleading conclusions can be drawn if models are compared without consideration of the sepsis definition used which emphasizes the need for a standardized definition for sepsis onset.

A Machine Learning Methodology for Identification and Triage of Heart Failure Exacerbations.

Inadequate at-home management and self-awareness of heart failure (HF) exacerbations are known to be leading causes of the greater than 1 million estimated HF-related hospitalizations in the USA alone. Most current at-home HF management protocols include paper guidelines or exploratory health applications that lack rigor and validation at the level of the individual patient. We report on a novel triage methodology that uses machine learning predictions for real-time detection and assessment of exacerbations. Medical specialist opinions on statistically and clinically comprehensive, simulated patient cases were used to train and validate prediction algorithms. Model performance was assessed by comparison to physician panel consensus in a representative, out-of-sample validation set of 100 vignettes. Algorithm prediction accuracy and safety indicators surpassed all individual specialists in identifying consensus opinion on existence/severity of exacerbations and appropriate treatment response. The algorithms also scored the highest sensitivity, specificity, and PPV when assessing the need for emergency care. Here we develop a machine-learning approach for providing real-time decision support to adults diagnosed with congestive heart failure. The algorithm achieves higher exacerbation and triage classification performance than any individual physician when compared to physician consensus opinion.

Hepatitis B virus outreach to immigrant population in Greater Boston Area: Key to improving hepatitis B knowledge.

AIM: To characterize the understanding of hepatitis B virus (HBV) and determine if outreach improves HBV understanding among Greater Boston Area immigrants. METHODS: Six outreach sessions were held in various community venues in the Greater Boston Area. Verbal consent was obtained from participants prior to starting each session. Each session included a pre-session questionnaire, followed by a teaching session, and then a post-session questionnaire. In person interpreters were present for translation during the teaching session and assistance for questionnaire completion when needed. The questions were developed based on the HBV clinical experience of physicians who serve largely immigrant populations. Questionnaires included Likert-type scale, open-ended, and true-false questions. All results were anonymous. RESULTS: One hundred and one people participated in this study. Participants were 30% male with ages ranging from 19 to 87 years. The study population included immigrants from 21 countries, as well as seven United States-born participants. The greatest numbers of participants were from Somalia (44%), Morocco (10%), and Cameroon (8%). Pre session questionnaires revealed that 42% of participants were unaware that HBV can cause cancer, and 50% were unaware that therapies for HBV exist. Our brief teaching intervention led to improved scores on post session questionnaires. For example, at baseline, 58% of participants responded correctly to the question "HBV infection can cause scarring of the liver and liver cancer", whereas 79% of participants responded correctly after the teaching session (P = 0.01). Furthermore, the mean of total correct answers in the true or false portion of the questionnaire increased from 5.5 to 7.6 (P < 0.001). CONCLUSION: A teaching session targeting Boston Immigrants at-risk for HBV helped improve scores on HBV knowledge questionnaires. Outreach may empower at-risk patients to pro-actively seek HBV care.

Barriers to the treatment of hepatitis C. Patient, provider, and system factors.

BACKGROUND: Hepatitis C virus (HCV) infection is both prevalent and undertreated. OBJECTIVE: To identify barriers to HCV treatment in primary care practice. DESIGN: Cross-sectional study. SETTING AND PARTICIPANTS: A cohort of 208 HCV-infected patients under the care of a primary care physician (PCP) between December 2001 and April 2004 at a single academically affiliated community health center. MEASUREMENTS: Data were collected from the electronic medical record (EMR), the hospital clinical data repository, and interviews with PCPs. MAIN RESULTS: Our cohort consisted of 208 viremic patients with HCV infection. The mean age was 47.6 (+/-9.7) years, 56% were male, and 79% were white. Fifty-seven patients (27.4% of the cohort) had undergone HCV treatment. Independent predictors of not being treated included: unmarried status (adjusted odds ratio [aOR] for treatment 0.36, P=.02), female gender (aOR 0.31, P=.01), current alcohol abuse (aOR 0.08, P=.0008), and a higher ratio of no-shows to total visits (aOR 0.005 per change of 1.0 in the ratio of no-shows to total visits, P=.002). The major PCP-identified reasons not to treat included: substance abuse (22.5%), patient preference (16%), psychiatric comorbidity (15%), and a delay in specialist input (12%). For 13% of the untreated patients, no reason was identified. CONCLUSIONS: HCV treatment was infrequent in our cohort of outpatients. Barriers to treatment included patient factors (patient preference, alcohol use, missed appointments), provider factors (reluctance to treat past substance abusers), and system factors (referral-associated delays). Multimodal interventions may be required to increase HCV treatment rates.