No abstract

BackgroundTypically, algorithms to classify phenotypes using electronic medical record (EMR) data were developed to perform well in a specific patient population. There is increasing interest in analyses which can allow study of a specific outcome across different diseases. Such a study in the EMR would require an algorithm that can be applied across different patient populations. Our objectives were: (1) to develop an algorithm that would enable the study of coronary artery disease (CAD) across diverse patient populations; (2) to study the impact of adding narrative data extracted using natural language processing (NLP) in the algorithm. Additionally, we demonstrate how to implement CAD algorithm to compare risk across 3 chronic diseases in a preliminary study.Methods and ResultsWe studied 3 established EMR based patient cohorts: diabetes mellitus (DM, n = 65,099), inflammatory bowel disease (IBD, n = 10,974), and rheumatoid arthritis (RA, n = 4,453) from two large academic centers. We developed a CAD algorithm using NLP in addition to structured data (e.g. ICD9 codes) in the RA cohort and validated it in the DM and IBD cohorts. The CAD algorithm using NLP in addition to structured data achieved specificity >95% with a positive predictive value (PPV) 90% in the training (RA) and validation sets (IBD and DM). The addition of NLP data improved the sensitivity for all cohorts, classifying an additional 17% of CAD subjects in IBD and 10% in DM while maintaining PPV of 90%. The algorithm classified 16,488 DM (26.1%), 457 IBD (4.2%), and 245 RA (5.0%) with CAD. In a cross-sectional analysis, CAD risk was 63% lower in RA and 68% lower in IBD compared to DM (p<0.0001) after adjusting for traditional cardiovascular risk factors.ConclusionsWe developed and validated a CAD algorithm that performed well across diverse patient populations. The addition of NLP into the CAD algorithm improved the sensitivity of the algorithm, particularly in cohorts where the prevalence of CAD was low. Preliminary data suggest that CAD risk was significantly lower in RA and IBD compared to DM.

Purpose: A tumor-derived proteolysis-inducing factor (PIF) is suggested to be a potent catabolic factor in skeletal muscle of mice and humans.We aimed to establish the clinical significance of PIF in cancer patients and to elucidate its structural features. Experimental Design: PIF was detected in human urine using a monoclonal antibody (mAb) and related to clinicaloutcomes. PIFimmunoaffinity-purifiedusing the mAbwaspurified/separated using reverse-phase high-performanceliquid chromatography and two-dimensional electrophoresis.Tenhuman cancer cell lines were tested for expression of mRNA encoding PIF core peptide. Results: PIF immunoreactivity was present in 160 of 262 patients with advanced cancers of the lung, esophagus/stomach, and other organs. In a Kaplan-Meier survival analysis of 181 lung cancer patients, PIF was unrelated to survival; PIF status was also unrelated to skeletal muscle loss confirmed by computed tomography imaging. PIF was seen in 16 of 24 patients with chronic heart failure and thus is not exclusive to malignant disease. In-gel digestion and mass spectrometric analysis of immunoaffinity purified PIF from cancer patients consistently identified human albumin and immunoglobulins.We showed nonspecific binding of purified albumin and immunoglobulins to the anti-PIF mAb, which is thus not a useful tool for PIF detection or purification in humans. Finally, the human PIF core peptide was detected in human cancer cell lines using reverse transcription-PCR and nucleotide sequencing; however, none of the amplified products had a site for the glycosylation critical to the proteolysis-inducing activity of murine PIF. Conclusions: A putative human homologue of murine PIF and its role in human cancer cachexia cannot be verified.A proteolysis-inducing glycoprotein [proteolysis-inducing factor (PIF)] of tumor origin mediates muscle wasting in mice bearing the MAC16 adenocarcinoma (1). PIF elicits intense skeletal muscle catabolism in muscle cells or animals (1 -4). Purified PIF has a mass of f24,000 Da and consists of a short polypeptide containing both N-linked (f10 kDa) and O-linked (f6 kDa) sulfated oligosaccharides (5). Preliminary evidence (6, 7) suggested that an identical factor is associated with weight loss in cancer patients (6 -10). Additionally, PIF was absent in cancer patients without weight loss or weight-losing patients with benign disease (7). These results suggested discovery of a critical factor underlying human cancer cachexia, but attempts at further confirmation suggested that PIF was not necessarily associated with clinical outcomes (i.e., survival and weight loss) in cancer patient populations (11,12).Regulation of expression of the human PIF core peptide, the sites at which glycosylation occurs to form the functional glycoprotein, and the structure of the oligosaccharides (5, 7), which confer the proteolysis-inducing activity to this unusual molecule, remain unresolved. A peptide sequence (7) and two patents describing the human cachexia-associated protein (HCAP) gene enc...

BACKGROUND/OBJECTIVES: Hospice care confers well-documented benefits to patients and their families, but it is underutilized. One potential reason is inadequate family support to make end-of-life decisions and care for older adults on hospice at home. We assessed the association between amount of family support and hospice use among a population of decedents and among specific illness types. DESIGN: Prospective cohort study using the National Health and Aging Trends Study waves 2011 to 2017, linked to Medicare claims data. SETTING: Contiguous United States. PARTICIPANTS: A total of 1,868 NHATS decedents. MEASUREMENTS: Outcome variable was 1 day or longer of hospice. Family caregiving intensity was measured by self-reported hours of care per week and number of caregivers. Covariates included probable dementia status and other demographic, clinical, and functional characteristics. RESULTS: At the end of life, hours of family caregiving and numbers of helpers vary widely with individuals with dementia receiving the most hours of unpaid care (mean = 64.5 hours per week) and having 2.4 unpaid caregivers on average. In an adjusted analysis, older adults with cancer receiving 40 hours and more of unpaid care/week as compared with fewer than 6 hours per week were twice as likely to receive hospice care at the end of life (odds ratio = 2.0; 95% confidence interval = 1.0-4.1). This association was not seen among those with dementia or among decedents in general. No significant association was found between number of caregivers and hospice use at the end of life. CONCLUSION: Older adults at the end of life receive a high number of hours of help at the end of life, many from more than one caregiver, which may shape hospice access. Better understanding of disparities in hospice use can facilitate timely access to care for older adults with a serious illness.

The 2014 i2b2/UTHealth Natural Language Processing (NLP) shared task featured a new longitudinal corpus of 1,304 records representing 296 diabetic patients. The corpus contains three cohorts: patients who have a diagnosis of coronary artery disease (CAD) in their first record, and continue to have it in subsequent records; patients who do not have a diagnosis of CAD in the first record, but develop it by the last record; patients who do not have a diagnosis of CAD in any record. This paper details the process used to select records for this corpus and provides an overview of novel research uses for this corpus. This corpus is the only annotated corpus of longitudinal clinical narratives currently available for research to the general research community.