Do logistic regression and signal detection identify different subgroups at risk? Implications for the design of tailored interventions.

2001 ◽  
Vol 6 (1) ◽  
pp. 35-48 ◽  
Author(s):  
Michaela Kiernan ◽  
Helena C. Kraemer ◽  
Marilyn A. Winkleby ◽  
Abby C. King ◽  
C. Barr Taylor
Keyword(s):  
At Risk ◽  
Logistic Regression ◽  
Signal Detection ◽  
Tailored Interventions

A Predictive Model of Intrinsic Factors Associated with Long-Stay Nursing Home Care After Hospitalization

2020 ◽  
pp. 105477382098527
Author(s):  
Jane Flanagan ◽  
Marie Boltz ◽  
Ming Ji
Keyword(s):  
At Risk ◽  
Logistic Regression ◽  
Receiver Operating Characteristic ◽  
Predictive Model ◽  
Operating Characteristic ◽  
Resource Planning ◽  
Term Care ◽  
Intrinsic Factors ◽  
Patients At Risk ◽  
Receiver Operating

We aimed to build a predictive model with intrinsic factors measured upon admission to skilled nursing facilities (SNFs) post-acute care (PAC) to identify older adults transferred from SNFs to long-term care (LTC) instead of home. We analyzed data from Massachusetts in 23,662 persons admitted to SNFs from PAC in 2013. Explanatory logistic regression analysis identified single “intrinsic predictors” related to LTC placement. To assess overfitting, the logistic regression predictive model was cross-validated and evaluated by its receiver operating characteristic (ROC) curve. A 12-variable predictive model with “intrinsic predictors” demonstrated both high in-sample and out-of-sample predictive accuracy in the receiver operating characteristic ROC and area under the ROC among patients at risk of LTC placement. This predictive model may be used for early identification of patients at risk for LTC after hospitalization in order to support targeted rehabilitative approaches and resource planning.

Redesigning Prediction Algorithms for At-Risk Students in Higher Education

2019 ◽  
pp. 232-250 ◽  
Author(s):  
Dennis Foung
Keyword(s):  
Higher Education ◽  
At Risk ◽  
Logistic Regression ◽  
Academic Writing ◽  
Industry 4.0 ◽  
Classification Tree ◽  
At Risk Students ◽  
Higher Education Administrators ◽  
Writing Courses ◽  
Education Administrators

Use of algorithms and data mining approaches are not new to Industry 4.0. However, these may not be common for students and educators in higher education. This chapter compares various classification techniques: classification tree, logistic regression, and artificial neural networks (ANN). The comparison focuses on each method's accuracy, algorithm, and practicality in higher education. This study made use of a dataset from two academic writing courses in a university in Hong Kong with more than 5,000 records. Results suggest that classification trees and logistic regression can be easily used in the higher education context, but ANN may not be applicable in higher educational settings. The research team suggests that higher education administrators take this research forward and design platforms to realize these classification algorithms to predict at-risk students.

scholarly journals Malnutrition among older adults living in Portuguese nursing homes: the PEN-3S study

2018 ◽  
Vol 22 (3) ◽  
pp. 486-497 ◽  
Author(s):  
Teresa Madeira ◽  
Catarina Peixoto-Plácido ◽  
Nuno Sousa-Santos ◽  
Osvaldo Santos ◽  
Violeta Alarcão ◽  
...  
Keyword(s):  
Older Adults ◽  
At Risk ◽  
Logistic Regression ◽  
Nursing Homes ◽  
Nutritional Status ◽  
Functional Status ◽  
Mini Nutritional Assessment ◽  
Computer Assisted ◽  
Depression Symptoms ◽  
Symptoms Of Depression

AbstractObjectiveTo characterise the nutritional status and to identify malnutrition-associated variables of older adults living in Portuguese nursing homes.DesignCross-sectional study. Data on demographic and socio-economic characteristics, self-reported morbidity, eating-related problems, nutritional status, cognitive function, depression symptoms, loneliness feelings and functional status were collected by trained nutritionists through a computer-assisted face-to-face structured interview followed by standardised anthropometric measurements. Logistic regression was used to identify factors associated with being at risk of malnutrition/malnourished.SettingPortuguese nursing homes.SubjectsNationally representative sample of the Portuguese population aged 65 years or over living in nursing homes.ResultsA total of 1186 individuals (mean age 83·4 years; 72·8 % women) accepted to participate. According to the Mini Nutritional Assessment, 4·8 (95 % CI 3·2, 7·3) % were identified as malnourished and 38·7 (95 % CI 33·5, 44·2) % were at risk of malnutrition. These percentages increased with age and were significantly higher for women. Logistic regression showed (OR; 95 % CI) that older adults reporting no or little appetite (6·5; 2·7, 15·3), those revealing symptoms of depression (2·6; 1·6, 4·2) and those who were more dependent in their daily living activities (4·7; 2·0, 11·1) were also at higher odds of being malnourished or at risk of malnutrition.ConclusionsMalnutrition and risk of malnutrition are prevalent among nursing home residents in Portugal. It is crucial to routinely screen for nutritional disorders, as well as risk factors such as symptoms of depression and lower functional status, to prevent and treat malnutrition.

scholarly journals S241. FACTORS ASSOCIATED WITH EARLY RISK OF DISENGAGEMENT FROM EARLY PSYCHOSIS INTERVENTION SERVICES

2020 ◽  
Vol 46 (Supplement_1) ◽  
pp. S130-S131
Author(s):  
Nicole Kozloff ◽  
Aristotle Voineskos ◽  
George Foussias ◽  
Alexia Polillo ◽  
Sean Kidd ◽  
...  
Keyword(s):  
At Risk ◽  
Substance Use ◽  
Logistic Regression ◽  
Young People ◽  
Family Involvement ◽  
The Body ◽  
Early Psychosis ◽  
First Episode ◽  
Factors Associated ◽  
Race Ethnicity

Abstract Background Despite the body of evidence supporting early psychosis intervention (EPI) programs for young people with psychotic disorders, approximately 30% of individuals with first-episode psychosis disengage from care. To date, two factors, lack of family involvement and presence of a substance use disorder, have emerged as robust predictors of EPI disengagement. Several factors associated with service disengagement in mental health care more broadly have not been well-studied in EPI; some of these, such as homelessness and ethnicity, may be of particular importance to urban, multicultural populations, and ethnicity in particular has been shown to affect pathways into EPI services. Early missed appointments may signal risk for subsequent service disengagement. We sought to identify early predictors of disengagement risk in an urban EPI program. Methods We conducted a prospective chart review of consecutive patients accepted for services in a large, urban EPI program in Toronto, Canada in a 3-month period from July 4-October 3, 2018. Patients were observed in their first 3 months of treatment. The primary outcome of interest was risk of disengagement, defined as having missed at least 1 appointment without cancellation. Extracted data included a variety of demographic and clinical information. The principal investigator trained 2 data abstractors on the first 50 charts; subsequent agreement on the next 5 charts was 88%. Based on previous literature, we hypothesized that risk of disengagement would be increased in individuals with problem substance use, experiences of homelessness, and nonwhite race/ethnicity and decreased in individuals with family involvement in their care. We used logistic regression to examine the odds of disengagement associated with univariate predictors individually, and then together in a multivariate model. Results Seventy-three patients were consecutively admitted to EPI services in the 3-month period. Of these individuals, 59% (N=43) were identified as being at risk of disengagement based on having missed at least 1 appointment without cancellation. In the full sample, 71% (N=52) identified as nonwhite, 23% (N=17) had a documented experience of homelessness, 52% (N=38) had problem substance use, and 73% (N=53) had family involved in their care. In univariate logistic regression, only problem substance use was associated with risk of disengagement (OR=2.91, 95% CI 1.11–7.66); no significant associations were identified with race/ethnicity, experience of homelessness, or family involvement. In multivariate logistic regression, once we controlled for these other factors, the association between risk of disengagement and problem substance use was attenuated and no longer statistically significant (OR=2.15, 95% CI 0.77–5.97). Discussion In this small study of early disengagement in an urban EPI program, only problem substance use was associated with increased odds of missing an appointment, but not when we controlled for other factors thought to be associated with disengagement. Larger studies may be required to identify factors with small but important effects. These factors may be used to identify young people at risk of disengagement from EPI services early in care in order to target them for increased engagement efforts.

Improving the Identification of Patients at Risk of Postoperative Renal Failure after Cardiac Surgery

2006 ◽  
Vol 104 (1) ◽  
pp. 65-72 ◽  
Author(s):  
Duminda N. Wijeysundera ◽  
Keyvan Karkouti ◽  
W Scott Beattie ◽  
Vivek Rao ◽  
Joan Ivanov
Keyword(s):  
At Risk ◽  
Logistic Regression ◽  
Renal Function ◽  
Cardiac Surgery ◽  
Renal Insufficiency ◽  
Care Center ◽  
Tertiary Care ◽  
Receiver Operating Curve ◽  
Alternative Measure ◽  
Patients At Risk

Background Preoperative renal insufficiency is an important predictor of the need for postoperative renal replacement therapy (RRT). Serum creatinine (sCr) has a limited ability to identify patients with preoperative renal insufficiency because it varies with age, sex, and muscle mass. Calculated creatinine clearance (CrCl) is an alternative measure of renal function that may allow better estimation of renal reserve. Methods Data were prospectively collected for consecutive patients who underwent cardiac surgery requiring cardiopulmonary bypass at a tertiary care center. The relation between CrCl (Cockcroft-Gault equation) and RRT was initially described using descriptive statistics, logistic regression, and receiver operating curve analysis. Based on these analyses, preoperative renal insufficiency was defined as CrCl of 60 ml/min or less. Preoperative renal function was classified as moderate insufficiency (sCr > 133 microM), mild insufficiency (100 microM < sCr < or = 133 microM), occult insufficiency (sCr < or = 100 microM and CrCl < or = 60 ml/min), or normal function (sCr < or = 100 microM and CrCl > 60 ml/min). The independent association of preoperative renal function with RRT was subsequently determined using multiple logistic regression. Results Of the 10,751 patients in the sample, 137 (1.2%) required postoperative RRT. Approximately 13% of patients with normal sCr had occult renal insufficiency. Occult renal insufficiency was independently associated with RRT (odds ratio, 2.80; 95% confidence interval, 1.39-5.33). The magnitude of this risk was similar to patients with mild renal insufficiency (P = 0.73). Conclusions The inclusion of a simple CrCl-based criterion in preoperative assessments may improve identification of patients at risk of needing postoperative RRT.

scholarly journals Symptoms in first-degree relatives of patients with rheumatoid arthritis: evaluation of cross-sectional data from the symptoms in persons at risk of rheumatoid arthritis (SPARRA) questionnaire in the PRe-clinical EValuation of Novel Targets in RA (PREVeNT-RA) Cohort

2021 ◽  
Vol 23 (1) ◽  
Author(s):  
R. E. Costello ◽  
J. H. Humphreys ◽  
J. C. Sergeant ◽  
M. Haris ◽  
F. Stirling ◽  
...  
Keyword(s):  
Rheumatoid Arthritis ◽  
At Risk ◽  
Logistic Regression ◽  
Joint Pain ◽  
Patient Characteristics ◽  
Cross Sectional ◽  
First Degree Relatives ◽  
Increased Risk ◽  
Large Joint ◽  
Small Joint

Abstract Background First-degree relatives (FDRs) of people with rheumatoid arthritis (RA) have a fourfold increased risk of developing RA. The Symptoms in Persons At Risk of Rheumatoid Arthritis (SPARRA) questionnaire was developed to document symptoms in persons at risk of RA. The aims of this study were (1) to describe symptoms in a cohort of FDRs of patients with RA overall and stratified by seropositivity and elevated CRP and (2) to determine if patient characteristics were associated with symptoms suggestive of RA. Methods A cross-sectional study of FDRs of patients with RA, in the PREVeNT-RA study, who completed a study questionnaire, provided a blood sample measured for rheumatoid factor, anti-CCP and CRP and completed the SPARRA questionnaire. Moderate/severe symptoms and symmetrical, small and large joint pain were identified and described. Symptoms associated with both seropositivity and elevated CRP were considered suggestive of RA. Logistic regression was used to determine if symptoms suggestive of RA were associated with patient characteristics. Results Eight hundred seventy participants provided all data, 43 (5%) were seropositive and 122 (14%) had elevated CRP. The most frequently reported symptoms were sleep disturbances (20.3%) and joint pain (17.9%). Symmetrical and small joint pain were 11.3% and 12.8% higher, respectively, in those who were seropositive and 11.5% and 10.7% higher in those with elevated CRP. In the logistic regression model, seropositivity, older age and feeling depressed were associated with increased odds of small and symmetrical joint pain. Conclusions This is the first time the SPARRA questionnaire has been applied in FDRs of patients with RA and has demonstrated that the presence of symmetrical and small joint pain in this group may be useful in identifying people at higher risk of developing RA.

scholarly journals 859. Same-Day HIV Pre-exposure Prophylaxis (PrEP) Initiation During Drop-in STD Clinic Appointments Is a Safe, Feasible, and Effective Method to Engage Patients at Risk for HIV in PrEP Care

2018 ◽  
Vol 5 (suppl_1) ◽  
pp. S20-S20 ◽  
Author(s):  
Kevin Kamis ◽  
Kenneth Scott ◽  
Edward Gardner ◽  
Karen Wendel ◽  
Grace Marx ◽  
...  
Keyword(s):  
At Risk ◽  
Logistic Regression ◽  
Health Insurance ◽  
Low Income ◽  
Median Income ◽  
Navigation Support ◽  
Patients At Risk ◽  
Exposure Prophylaxis ◽  
Research Grant

Abstract Background Patients at risk for HIV generally do not have immediate access to PrEP. We hypothesized that by offering free, 30-day PrEP starter packs and navigation support during drop-in STD clinic appointments, individuals would be likely to initiate and continue PrEP. Methods Individuals aged ≥18 years presenting for drop-in appointments in the Metro Denver STD Clinic and indicated for PrEP were eligible for the study. Exclusion criteria were history of renal dysfunction, chronic hepatitis B (HBV), HIV, pregnancy, and indications for postexposure prophylaxis. Eligible individuals were provided PrEP education and offered a free, 30-day PrEP starter pack and navigation support for cost assistance. Participants were tested for creatinine, HBV, HIV, and pregnancy at enrollment, and navigated to an appointment for ongoing PrEP care. Participants’ medical records were reviewed for a minimum of 4 months after enrollment. Descriptive statistics and logistic regression were used to characterize the study population and follow-up. Results From April to October 2017, 100 individuals filled a tenofovir–emtricitabine prescription (figure). Median participant age was 28 years, 98% were male, 53% were non-Hispanic White, 8% non-Hispanic Black, and 34% Hispanic. Median annual income was $24,000, 62% had health insurance, 26% had a primary care provider (PCP), and 50% had a recent bacterial STI. No participants had abnormal baseline creatinine or HBV. 77% completed ≥1 PrEP follow-up visit during the study period; 57% completed their first visit within 31 days. 56% completed a second follow-up visit. No HIV seroconversions were detected during follow-up. Factors significantly associated with attending ≥1 follow-up appointment were age ≥ 30 years, higher income, and having health insurance or a PCP at enrollment. In multivariate logistic regression, only higher income was associated with attending ≥1 follow-up appointment (median income for those with ≥1 follow-up visit vs. no follow-up: $24,960 vs. $14,000, P <0.01). Conclusion Providing immediate access to PrEP during drop-in STD clinic visits is a safe and feasible approach to initiation of PrEP care. Additional resources are needed to support PrEP continuity care, particularly for low-income individuals. Disclosures K. Kamis, Gilead Scienes: Research Coordinator, Research grant. S. Rowan, Gilead Sciences: Investigator, Research grant.

Abstract P186: Extent of Myocardium at Risk in Women, but Not in Men, is Associated With Major Adverse Cardiac Events Following Coronary Artery Bypass Grafting

2011 ◽  
Vol 4 (suppl_1) ◽  
Author(s):  
Karen J Buth ◽  
Maria E Currie ◽  
Alexandra M Yip ◽  
Gregory M Hirsch
Keyword(s):  
At Risk ◽  
Logistic Regression ◽  
Coronary Artery ◽  
Cardiac Events ◽  
Men And Women ◽  
Adverse Cardiac Events ◽  
Angiographic Data ◽  
Myocardium At Risk ◽  
Cabg Patients ◽  
The Impact

Objective: Many studies have reported higher rates of adverse outcomes in women than men following CABG surgery. To date, the mechanism has not been elucidated. We studied a large group of CABG patients for whom detailed angiographic data were available and investigated the impact of myocardium at risk (MAR) on major adverse cardiac events (MACE) in men and women post-CABG. Methods: For patients undergoing isolated primary CABG, a Duke Index score was calculated from angiographic data. Duke Index is a validated score that quantifies MAR using distribution and severity of coronary artery disease. Categories of MAR were defined as Low, Moderate and High based on number of diseased vessels as well as location of disease, with proximal lesions conferring a higher weight than distal stenoses. The post-CABG outcome was in-hospital MACE (1 or more of mortality, low cardiac output, stroke, MI, sepsis, deep sternal wound infection, ventilation >48 hrs or return to ICU). Logistic regression was used to examine the impact of MAR as a predictor of MACE in women and in men, after adjusting for clinical comorbidities. Results: Of 3744 CABG patients, 70% (2614) had complete angiographic data and were included in the analysis; 23% (607 of 2614) were women. Compared with men, women undergoing CABG had similar Duke Index scores but were older and more likely to have diabetes, hypertension, vascular disease, disabling angina, and require urgent surgery. Completeness of revascularization did not differ between men and women. Prevalence of MACE was higher in women than in men: 19.9% (121 of 607) versus 13.0% (262 of 2007), p=0.0001. In a logistic regression model fully adjusted for comorbidities, there was a significant interaction between female sex and increased MAR as predictors of MACE. In separate logistic regression models for each sex, increasing MAR was an independent predictor of MACE for women (High MAR: OR 3.2, 95% CI 1.4-7.6, Moderate MAR: OR 2.5, 95% CI 1.0-6.2), but not for men (High MAR: OR 1.0, 95% CI 0.6-1.7, Moderate MAR: OR 1.1, 95% CI 0.6-1.8). Conclusions: Myocardium at risk impacts post-CABG outcome differently for women than for men. This novel finding suggests that CABG surgery may not provide the same potential for recovery of ischemic myocardium for women compared with men.

Binary Logistic Regression Approach for Decision Making in Bridge Management

2021 ◽  
pp. 1-11
Author(s):  
Uditha A Wijesuriya ◽  
Adam G Tennant
Keyword(s):  
Decision Making ◽  
At Risk ◽  
Logistic Regression ◽  
Construction Material ◽  
Binary Logistic Regression ◽  
Decision Making Process ◽  
Bridge Management ◽  
Binary Logistic Regression Model ◽  
National Bridge Inventory ◽  
Load Types

Bridge management professionals need effective tools to help guide the decision-making process and maintain quality infrastructure in a region. A new binary response is herein defined by categorizing bridges as at-risk and not at-risk, based on the existing overall bridge condition scores. Fitting binary logistic regression model for the response, the probability of a bridge being at-risk is expressed in terms of the primary bridge factors age, load, types of construction material and structural design, and conditions of the deck, superstructure, and substructure. These estimated probabilities multiplied by specified consequence values are used to introduce the risk classes and their ranks. Employing the method for training and validating sets of sizes 13,540 and 3,385 in 2017, and 13,481 and 3,370 in 2018 data in National Bridge Inventory (NBI) Indiana, a statistically significant model is established containing age, load, conditions of both superstructure and substructure. Moreover, at-risk bridges are identified from Indiana NBI data in both years and for a subset from Connecticut in 2017. The novel bridge-ranking tool prioritizes bridges for maintenance purposes such as replacing or repairing and hence efficiently guides the management in the decision-making process for capital expenditures, and perhaps, for predicting the missing overall bridge condition.

Sign in / Sign up

Export Citation Format

Share Document