On stability issues in deriving multivariable regression models

2014 ◽  
Vol 57 (4) ◽  
pp. 531-555 ◽  
Author(s):  
Willi Sauerbrei ◽  
Anika Buchholz ◽  
Anne-Laure Boulesteix ◽  
Harald Binder
Keyword(s):  
Regression Models ◽  
Multivariable Regression ◽  
Stability Issues ◽  
Multivariable Regression Models

scholarly journals The Impact of Weather and Seasons on Falls and Physical Activity among Older Adults with Glaucoma: A Longitudinal Prospective Cohort Study

Sensors ◽  
2021 ◽  
Vol 21 (10) ◽  
pp. 3415
Author(s):  
Hursuong Vongsachang ◽  
Aleksandra Mihailovic ◽  
Jian-Yu E ◽  
David S. Friedman ◽  
Sheila K. West ◽  
...  
Keyword(s):  
Physical Activity ◽  
Older Adults ◽  
Fall Prevention ◽  
Regression Models ◽  
Activity Promotion ◽  
Injurious Fall ◽  
Multivariable Regression ◽  
Multivariable Regression Models ◽  
The Impact ◽  
The Relationship

Understanding periods of the year associated with higher risk for falling and less physical activity may guide fall prevention and activity promotion for older adults. We examined the relationship between weather and seasons on falls and physical activity in a three-year cohort of older adults with glaucoma. Participants recorded falls information via monthly calendars and participated in four one-week accelerometer trials (baseline and per study year). Across 240 participants, there were 406 falls recorded over 7569 person-months, of which 163 were injurious (40%). In separate multivariable regression models incorporating generalized estimating equations, temperature, precipitation, and seasons were not significantly associated with the odds of falling, average daily steps, or average daily active minutes. However, every 10 °C increase in average daily temperature was associated with 24% higher odds of a fall being injurious, as opposed to non-injurious (p = 0.04). The odds of an injurious fall occurring outdoors, as opposed to indoors, were greater with higher average temperatures (OR per 10 °C = 1.46, p = 0.03) and with the summer season (OR = 2.69 vs. winter, p = 0.03). Falls and physical activity should be understood as year-round issues for older adults, although the likelihood of injury and the location of fall-related injuries may change with warmer season and temperatures.

Abstract 3572: Clinical Outcomes of the “Stroke100 Club”

Stroke ◽  
2012 ◽  
Vol 43 (suppl_1) ◽  
Author(s):  
Amy K Guzik ◽  
Rema Raman ◽  
Kain Ernstrom ◽  
Dawn M Meyer ◽  
Ajeet Sodhi ◽  
...  
Keyword(s):  
Regression Models ◽  
Young Patients ◽  
Older Age ◽  
Advanced Age ◽  
Stroke Severity ◽  
Old Patients ◽  
Multivariable Regression ◽  
Independent Variable ◽  
Poor Outcomes ◽  
Multivariable Regression Models

Background: Patients with advanced age or high NIHSS have poorer tPA outcomes. When combined, old age (≥80yo) and elevated NIHSS (≥20) may have an even worse outcome. Patients who are also in this “Stroke100 Club” (any combination of age and NIHSS ≥100) by other means, have not been fully assessed. We evaluated discharge destination, 90-day mRS, sICH and death in treated and untreated Stroke100 Club patients. We further compared patients with age ≥ 80 and NIHSS ≥ 20 (“80/20s”), those who reached 100 without both characteristics (“non80/20s”) and ‘controls’. Methods: The UCSD SPOTRIAS prospectively collected database was analyzed for AIS patients (with and without tPA). Multivariable regression models including the Stroke100 group as an independent variable was used. Outcomes were adjusted for baseline mRS. For comparing categorical outcomes between controls, “80/20s” and “non80/20s” subgroups, a Fisher’s exact was used. Results: The IV tPA subset included 257 patients (mean age 71, 52% male, 85% white, mean NIHSS 12). 53 were in the “Stroke100 Club” (28 80/20, 25 non80/20), with more women (68% p= 0.002), higher NIHSS (22.5 p<0.0001), older age (mean age 86.4 p<0.0001), higher pre stroke mRS (34.6% mRS 3-6 vs 7.84%, p<0.0001), more HTN (p=0.045) and more afib (p= 0.008). There were 284 non tPA patients (mean age 69.52, 54% male, 85% white, mean NIHSS 5.92). 21 were in the “Stroke100 Club” (14 80/20, 7 non80/20), with higher NIHSS (23 p<0.0001), older age (mean 86.2 p<0.0001), higher pre stroke mRS (45.5% 3-6 vs 9.5%, p= 0.0001), and more afib (p= 0.0002). Stroke100 Club 90day mRS(3-6) outcomes were worse in both tPA treated patients (OR=6.77, p= 0.0001) and nontreated patients (OR 31.57, p= 0.001). sICH rates (in tPA subjects) were not different (3.8% vs 3.4%, p> 0.99). Conclusions: There is a question of treatment outcome in patients with various permutations of stroke severity and advanced age. Our data corroborates the concern of poor outcomes for Stroke100 Club patients, but notes no increased sICH with tPA. Though outcome may be poor, withholding tPA should be discouraged as worse outcomes were not due to sICH. Young patients with severe strokes or old patients with mild strokes may have outcomes similar to the standard “80/20” Stroke100 patients, however further adjusted analysis is ongoing. In addition, further analyses are being done to compare tPA to non tPA patients.

Preadmission Corticosteroid Therapy and the Risk of Respiratory Failure in Adults Without HIV Presenting With Pneumocystis Pneumonia

2019 ◽  
Vol 35 (12) ◽  
pp. 1465-1470 ◽  
Author(s):  
Patrick M. Wieruszewski ◽  
Erin F. Barreto ◽  
Jason N. Barreto ◽  
Hemang Yadav ◽  
Pritish K. Tosh ◽  
...  
Keyword(s):  
Respiratory Failure ◽  
Corticosteroid Therapy ◽  
Regression Models ◽  
Pneumocystis Pneumonia ◽  
Severe Respiratory Failure ◽  
Corticosteroid Administration ◽  
Multivariable Regression ◽  
Multivariable Regression Models ◽  
Reduced Risk ◽  
Hiv Negative

Background: Corticosteroid therapy is a well-recognized risk factor for Pneumocystis pneumonia (PCP); however, it has also been proposed as an adjunct to decrease inflammation and respiratory failure. Objective: To determine the association between preadmission corticosteroid use and risk of moderate-to-severe respiratory failure at the time of PCP presentation. Methods: This retrospective cohort study evaluated HIV-negative immunosuppressed adults diagnosed with PCP at Mayo Clinic from 2006 to 2016. Multivariable regression models were used to evaluate the association between preadmission corticosteroid exposure and moderate-to-severe respiratory failure at presentation. Results: Of the 323 patients included, 174 (54%) used preadmission corticosteroids with a median daily dosage of 20 (interquartile range: 10-40) mg of prednisone or equivalent. After adjustment for baseline demographics, preadmission corticosteroid therapy did not decrease respiratory failure at the time of PCP presentation (odds ratio: 1.23, 95% confidence interval: 0.73-2.09, P = .38). Additionally, after adjusting for inpatient corticosteroid administration, preadmission corticosteroid use did not impact the need for intensive care unit admission ( P = .98), mechanical ventilation ( P = .92), or 30-day mortality ( P = .11). Conclusions: Corticosteroid exposure before PCP presentation in immunosuppressed HIV-negative adults was not associated with a reduced risk of moderate-to-severe respiratory failure.

Component-resolved diagnosis and beyond: Multivariable regression models to predict severity of hazelnut allergy

Allergy ◽  
2017 ◽  
Vol 73 (3) ◽  
pp. 549-559 ◽  
Author(s):  
M. R. Datema ◽  
R. van Ree ◽  
R. Asero ◽  
L. Barreales ◽  
S. Belohlavkova ◽  
...  
Keyword(s):  
Regression Models ◽  
Component Resolved Diagnosis ◽  
Multivariable Regression ◽  
Multivariable Regression Models

scholarly journals Describing and Presenting Multivariable Regression Models

2020 ◽  
Vol 30 (4) ◽  
pp. 303-305
Author(s):  
Dmitry Tumin ◽  
Mary Hayney ◽  
Rebecca P Winsett
Keyword(s):  
Regression Models ◽  
Multivariable Regression ◽  
Multivariable Regression Models

Confidence interval reporting for measures of association in multivariable regression models in observational studies

2019 ◽  
Vol 153 (6) ◽  
pp. 239-242
Author(s):  
Adrián González-Marrón ◽  
Jordi Real ◽  
Carles Forné ◽  
Albert Roso-Llorach ◽  
Eva María Navarrete-Muñoz ◽  
...  
Keyword(s):  
Confidence Interval ◽  
Observational Studies ◽  
Regression Models ◽  
Measures Of Association ◽  
Multivariable Regression ◽  
Multivariable Regression Models

General Forms of a Class of Multivariable Regression Models

2014 ◽  
Vol 11 (2) ◽  
pp. 22-28
Author(s):  
A. Efremov
Keyword(s):  
Regression Models ◽  
Multiple Input Multiple Output ◽  
Parameters Estimation ◽  
Linear Functions ◽  
Experimental Modelling ◽  
Advantages And Disadvantages ◽  
Input Multiple Output ◽  
Multivariable Regression ◽  
Mimo Model ◽  
Multivariable Regression Models

Abstract There are two possible general forms of multiple input multiple output (MIMO) regression models, which are either linear with respect to their parameters or non-linear, but in order to estimate their parameters, at a certain stage it could be assumed that they are linear. This is in fact the basic assumption behind the linear approach for parameters estimation. There are two possible representations of a MIMO model, which at a certain level could be fictitiously presented as linear functions of its parameters. One representation is when the parameters are collected in a matrix and hence, the regressors are in a vector. The other possible case is the parameters to be in a vector, but the regressors at a given instant to be placed in a matrix. Both types of representations are considered in the paper. Their advantages and disadvantages are summarized and their applicability within the whole experimental modelling process is also discussed.

Abstract 16: EARLY: EMS Prenotification Augments Stroke Code Timepoints But Rigorous Training is Needed for Larger Yield in stroke.

Stroke ◽  
2013 ◽  
Vol 44 (suppl_1) ◽  
Author(s):  
Amy K Guzik ◽  
Rema Raman ◽  
Karin Ernstrom ◽  
Dawn M Meyer ◽  
Thomas Hemmen ◽  
...  
Keyword(s):  
Regression Models ◽  
Treatment Rate ◽  
Accurate Identification ◽  
Stroke Team ◽  
Stroke Diagnosis ◽  
Multivariable Regression ◽  
Independent Variable ◽  
Baseline Characteristics ◽  
Multivariable Regression Models ◽  
Baseline Covariates

Background: rtPA for AIS is time sensitive, requiring efficient and coordinated acute care. We evaluated time to evaluation, stroke diagnosis, treatment rate, and 90 day outcome in patients with stroke team prenotification by EMS and those identified after arrival. Methods: The UCSD SPOTRIAS prospectively collected database was analyzed for patients with stroke team prenotification by EMS and other patients seen in stroke code, excluding inpatient codes. Multivariable regression models used outcome of interest as independent variable. Models were adjusted for pre-specified covariates: pre-stroke mRS, age, gender, smoking, baseline NIHSS and glucose. Time differences between groups were analyzed using Wilcoxon Rank Sum Tests. Results: We assessed 2867 patients, with EMS prenotification in 643 (22.4%). Assessment at 90 days was obtained in 216 with prenotification and 807 others. Those with prenotification were older (mean 68 vs 66, p=0.0498), with higher pre-stroke mRS (p=0.0243), NIHSS (10.9 vs 8.5, p< 0.0001) and glucose (139 vs 135, p=0.0013). Prenotification led to shorter time to imaging, decision, and IV rtPA treatment (all p<0.0001). No difference was seen in IV rtPA treatment rate (18% EMS prenotifications vs 16% others). When controlling for baseline characteristics, stroke was diagnosed more frequently in patients without EMS prenotification (OR 1.31, 95% CI 1.08-1.58, p=0.0057). Poor outcome (mRS 3-6) was seen more frequently in prenotification patients (45.83% vs 35.32%, p=0.006, NS after adjusting for baseline covariates). Conclusions: In the UCSD experience, EMS prenotification leads to faster evaluation critical in stroke. Prenotification occurred in patients at a medically worse baseline, but did not result in higher rates of final stroke diagnosis or IV rtPA. With improved education, accurate identification of AIS patients may improve, further expediting care and improving treatment and outcomes.

Sign in / Sign up

Export Citation Format

Share Document