PubMed Health⌕ Search

Biomedical subjects

Margaret Gardner

Publications and source records attributed to Margaret Gardner.

3 recordsLinked to original sources

dGAMLSS: an exact, distributed algorithm to fit Generalized Additive Models for Location, Scale, and Shape for privacy-preserving population reference charts.

MOTIVATION: There is growing interest in estimating population reference ranges across age and sex to better identify atypical clinically-relevant measurements throughout the lifespan. For this task, the World Health Organization recommends using Generalized Additive Models for Location, Scale, and Shape (GAMLSS), which can model non-linear growth trajectories under complex distributions that address the heterogeneity in human populations.Fitting GAMLSS models requires large, generalizable sample sizes, especially for accurate estimation of extreme quantiles, but obtaining such multi-site data can be challenging due to privacy concerns and practical considerations. In settings where patient data cannot be shared, privacy-preserving distributed algorithms for federated learning can be used, but no such algorithm exists for GAMLSS. RESULTS: We propose distributed GAMLSS (dGAMLSS), a distributed algorithm that can fit GAMLSS models across multiple sites without sharing patient-level data. This includes specific considerations for the fitting of smooth functions at varying levels of communication efficiency. We demonstrate the effectiveness of dGAMLSS in constructing population reference charts across clinical, genomics, and neuroimaging settings and show that dGAMLSS is able to reproduce pooled reference charts and inference down to numerical differences. AVAILABILITY AND IMPLEMENTATION: An R package providing examples of the dGAMLSS algorithm, as well as functions for sharing and aggregating site-specific parameters, is available at https://github.com/hufengling/dGAMLSS.

Algorithms↗

Challenges to the effective use of unfractionated heparin in the hospitalized management of acute thrombosis.

BACKGROUND: Unfractionated heparin therapy is care intensive because of dose-response variability, and because of the necessity of constant intravenous infusion and frequent monitoring. We sought to assess the real-world course of transition from heparin to warfarin in hospitalized patients undergoing anticoagulation therapy for acute venous or arterial thrombosis at our medical center. METHODS: Patients were retrospectively identified from July 1998 to December 1998. Data collected included initiation and maintenance doses of heparin, frequency of monitoring and dose adjustments, time to the therapeutic range, complications and interruptions of therapy, and characteristics of heparin-to-warfarin transition. RESULTS: Of the 311 patients who met the study criteria during the 6-month period, 134 had venous thromboembolism, 122 had cerebral arterial thrombosis, and 55 had peripheral arterial thrombosis. Groups differed in use and magnitude of initial heparin bolus, frequency of monitoring, and time to the therapeutic range. Dose response to intravenous heparin was highly variable. Even when the activated partial thromboplastin time reached the therapeutic range of 55 to 85 seconds, the next 2 consecutive measurements remained in this range in only 29% of the patients. Patients received an average of 4 different heparin doses over the first 3 days of treatment, and the therapeutic range was maintained on each of 4 sequential days in only 7% of them. During the course of therapy, 54% of the patients had at least 1 prolonged interruption in heparin infusion, and 4.8% sustained a major hemorrhage. Overall, 20% of the patients met the currently recommended treatment guideline of 4 days or more of heparin and warfarin overlap, until the international normalized ratio is greater than 2.0 for 2 consecutive days. CONCLUSIONS: Multiple challenges to effective anticoagulation treatment with unfractionated heparin exist in the hospital setting. Strategies are needed to improve the overall quality of anticoagulant care, including the substitution of low-molecular-weight heparin for unfractionated heparin, where appropriate.

Adolescent↗

A one-day, hospital-wide survey of dying inpatients.

We attempted to identify and briefly follow until discharge all terminally ill patients in a large general hospital. On 1 day, nurse case managers reviewed all hospitalized patients and identified those whom they believed were likely to die in the next 6 months (Category A) or whom might be considered terminally ill but with a longer prognosis (Category B). Twelve percent of all adult and pediatric medical-surgical inpatients were detected, equally divided between the two categories. In Category A, 63% were on the medical service, 7% were receiving intensive care, 54% had cancer, and 46% had do-not-resuscitate (DNR) orders. In Category B, 40% were on the medical service, 10% were in intensive care, 52% had cancer, and only 5% had DNR orders. Case managers expected 6% of identified patients to die in the hospital. After 1 month, at least 19% of identified patients had died (2.3% of the medical-surgical inpatient census on the day of the survey). The average length of stay in both categories, excluding outliers, was 24 days or approximately 4 times the average length of stay for the hospital. Patients who actually died in the hospital had an average length of stay of 62 days. This study presents a simple method for estimating the number of dying patients in a hospital--the target population for a palliative care program--and for determining their location, principal diagnosis, length of stay, and disposition. We present information indicating that the survey underestimates the number of dying hospitalized patients. We discuss possible policy implication of this study, primarily that general hospitals should consider developing specialized palliative care services for this substantial group of inpatients.

Case Management↗