scholarly journals A Novel Weighting Method to Remove Bias from Within-subject Exposure Dependency in Case-crossover Studies

2021 ◽  
Author(s):  
Kiyoshi Kubota ◽  
Lan Kelly ◽  
Tsugumichi Sato ◽  
Nicole Pratt ◽  
Elizabeth Roughead ◽  
...  
Keyword(s):  
Logistic Regression ◽  
Crossover Study ◽  
Odds Ratio ◽  
Time Trends ◽  
Rate Ratio ◽  
Conditional Logistic Regression ◽  
Time Varying ◽  
Weighting Method ◽  
Case Crossover ◽  
Crossover Studies

Abstract Background:Case-crossover studies have been widely used in various fields including pharmacoepidemiology. Vines and Farrington indicated in 2001 that when within-subject exposure dependency exists, conditional logistic regression can be biased. However, this bias has not been well studied.Methods:We have extended findings by Vines and Farrington to develop a weighting method for the case-crossover study which removes bias from within-subject exposure dependency. Our method calculates the exposure probability at the case period in the case-crossover study which is used to weight the likelihood formulae presented by Greenland in 1999. We simulated data for the population with a disease where most patients receive a cyclic treatment pattern with within-subject exposure dependency but no time trends while some patients stop and start treatment. Finally, the method was applied to real-world data from Japan to study the association between celecoxib and peripheral edema and to study the association between selective serotonin reuptake inhibitor (SSRI) and hip fracture in Australia.Results:When the simulated rate ratio of the outcome was 4.0 in a case-crossover study with no time-varying confounder, the proposed weighting method and the Mantel-Haenszel odds ratio reproduced the true rate ratio. When a time-varying confounder existed, the Mantel-Haenszel method was biased but the weighting method was not. When more than one control period was used, standard conditional logistic regression was biased either with or without time-varying confounding and the bias increased (up to 9.4) when the study period was extended. In real-world analysis with a binary exposure variable in Japan and Australia, the point estimate of the odds ratio (around 2.5 for the association between celecoxib and peripheral edema and around 1.6 between SSRI and hip fracture) by our weighting method was equal to the Mantel-Haenszel odds ratio and stable compared with standard conditional logistic regression. Conclusion:Case-crossover studies may be biased from within-subject exposure, even without exposure time trends. This bias can be identified by comparing the odds ratio calculated by the Mantel-Haenszel method and that by standard conditional logistic regression. Our proposed method will remove bias from within-subject exposure dependency and can account for time-varying confounders.

scholarly journals A novel weighting method to remove bias from within-subject exposure dependency in case-crossover studies

2021 ◽  
Vol 21 (1) ◽  
Author(s):  
Kiyoshi Kubota ◽  
Thu-Lan Kelly ◽  
Tsugumichi Sato ◽  
Nicole Pratt ◽  
Elizabeth Roughead ◽  
...  
Keyword(s):  
Logistic Regression ◽  
Crossover Study ◽  
Odds Ratio ◽  
Time Trends ◽  
Rate Ratio ◽  
Conditional Logistic Regression ◽  
Time Varying ◽  
Weighting Method ◽  
Case Crossover ◽  
Crossover Studies

Abstract Background Case-crossover studies have been widely used in various fields including pharmacoepidemiology. Vines and Farrington indicated in 2001 that when within-subject exposure dependency exists, conditional logistic regression can be biased. However, this bias has not been well studied. Methods We have extended findings by Vines and Farrington to develop a weighting method for the case-crossover study which removes bias from within-subject exposure dependency. Our method calculates the exposure probability at the case period in the case-crossover study which is used to weight the likelihood formulae presented by Greenland in 1999. We simulated data for the population with a disease where most patients receive a cyclic treatment pattern with within-subject exposure dependency but no time trends while some patients stop and start treatment. Finally, the method was applied to real-world data from Japan to study the association between celecoxib and peripheral edema and to study the association between selective serotonin reuptake inhibitor (SSRI) and hip fracture in Australia. Results When the simulated rate ratio of the outcome was 4.0 in a case-crossover study with no time-varying confounder, the proposed weighting method and the Mantel-Haenszel odds ratio reproduced the true rate ratio. When a time-varying confounder existed, the Mantel-Haenszel method was biased but the weighting method was not. When more than one control period was used, standard conditional logistic regression was biased either with or without time-varying confounding and the bias increased (up to 8.7) when the study period was extended. In real-world analysis with a binary exposure variable in Japan and Australia, the point estimate of the odds ratio (around 2.5 for the association between celecoxib and peripheral edema and around 1.6 between SSRI and hip fracture) by our weighting method was equal to the Mantel-Haenszel odds ratio and stable compared with standard conditional logistic regression. Conclusion Case-crossover studies may be biased from within-subject exposure dependency, even without exposure time trends. This bias can be identified by comparing the odds ratio by the Mantel-Haenszel method and that by standard conditional logistic regression. We recommend using our proposed method which removes bias from within-subject exposure dependency and can account for time-varying confounders.

Flexible derivative time-varying model in matched case-crossover studies for a small number of geographical locations among the participants

2020 ◽  
pp. 096228022096817
Author(s):  
Ana M Ortega-Villa ◽  
Inyoung Kim
Keyword(s):  
Logistic Regression ◽  
Conditional Logistic Regression ◽  
Effect Modification ◽  
Fixed Number ◽  
Time Varying ◽  
Model Framework ◽  
Conditional Logistic Regression Model ◽  
Case Crossover ◽  
Crossover Studies ◽  
Matched Case

In matched case-crossover studies, any stratum effect is removed by conditioning on the fixed number of case–control sets in the stratum, and hence, the conditional logistic regression model is not able to detect any effects associated with matching covariates. However, some matching covariates such as time and location often modify the effect of covariates, making the estimations obtained by conditional logistic regression incorrect. Therefore, in this paper, we propose a flexible derivative time-varying coefficient model to evaluate effect modification by time and location, in order to make correct statistical inference, when the number of locations is small. Our proposed model is developed under the Bayesian hierarchical model framework and allows us to simultaneously detect relationships between the predictor and binary outcome and between the predictor and time. Inference is proposed based on the derivative function of the estimated function to determine whether there is an effect modification due to time and/or location, for a small number of locations among the participants. We demonstrate the accuracy of the estimation using a simulation study and an epidemiological example of a 1–4 bidirectional case-crossover study of childhood aseptic meningitis with drinking water turbidity.

scholarly journals A data visualisation method for assessing exposure misclassification in case-crossover studies: the example of tricyclic antidepressants and the risk of hip fracture in older people

2021 ◽  
Vol 21 (1) ◽  
Author(s):  
Michael J. Leach ◽  
Elizabeth E. Roughead ◽  
Nicole L. Pratt
Keyword(s):  
Hip Fracture ◽  
Confidence Interval ◽  
Crossover Study ◽  
Odds Ratio ◽  
Tricyclic Antidepressants ◽  
Data Visualisation ◽  
Exposure Misclassification ◽  
Case Crossover ◽  
Crossover Studies ◽  
Control Window

Abstract Background The case-crossover design is suited to medication safety studies but is vulnerable to exposure misclassification. Using the example of tricyclic antidepressants and the risk of hip fracture, we present a data visualisation tool for observing exposure misclassification in case-crossover studies. Methods A case-crossover study was conducted using Australian Government Department of Veterans’ Affairs claims data. Beneficiaries aged over 65 years who were hospitalised for hip fracture between 2009 and 2012 were included. The case window was defined as 1–50 days pre fracture. Control window one and control window two were defined as 101–150 and 151–200 days pre fracture, respectively. Patients were stratified by whether exposure status changed when control window two was specified instead of control window one. To visualise potential misclassification, each subject’s tricyclic antidepressant dispensings were plotted over the 200 days pre fracture. Results The study population comprised 8828 patients with a median age of 88 years. Of these subjects, 348 contributed data to the analyses with either control window. The data visualisation suggested that 14% of subjects were potentially misclassified with control window one while 45% were misclassified with control window two. The odds ratio for the association between tricyclic antidepressants and hip fracture was 1.18 (95% confidence interval = 0.91–1.52) using control window one, whereas risk was significantly increased (odds ratio = 1.43, 95% confidence interval = 1.11–1.83) using control window two. Conclusions Exposure misclassification was less likely to be present with control window one than with an earlier control window, control window two. When specifying different control windows in a case-crossover study, data visualisation can help to assess the extent to which exposure misclassification may contribute to variable results.

scholarly journals Control of Weekly Time Trend in Time-Stratified Case-Crossover Design

2021 ◽  
Vol 10 (5) ◽  
pp. 85
Author(s):  
Xiaoming Wang ◽  
Sukun Wang ◽  
Warren Kindzierski
Keyword(s):  
Logistic Regression ◽  
Time Trends ◽  
Time Trend ◽  
Ambient Air ◽  
Effective Control ◽  
Systematic Bias ◽  
Case Crossover ◽  
Crossover Studies ◽  
Reference Selection ◽  
Regression Approach

Case-crossover designs have become widespread in biomedical investigations of transient associations. However, the most popular reference-selection strategy - the time-stratified scheme - may not be an optimum solution to control systematic bias in case-crossover studies. To prove this, we conducted a time series decomposition for daily ozone records and examined the capability of the time-stratified scheme to control for yearly, monthly, and weekly time trends; and observed its failure on the control for the weekly time trend. To solve this issue, we proposed a new logistic regression approach in which we suggest the adjustment for the weekly time trend. We compared the performance of the proposed with that of the traditional method by simulation. We further conducted an empirical study to explore the performance of the new logistic regression approach in examining potential associations between ambient air pollutants and acute myocardial infarction hospitalizations. The time-stratified scheme provides effective control for yearly and monthly time trends but not of the weekly time trend. Uncontrolled weekly time trends could be the dominant source of systematic bias in time-stratified case-crossover studies. In contrast, the proposed logistic regression approach can effectively minimize systematic bias in a case-crossover study.

scholarly journals Impact of Heat Waves on Hospitalisation and Mortality in Nursing Homes: A Case-Crossover Study

2021 ◽  
Vol 18 (20) ◽  
pp. 10697
Author(s):  
Ine Van den Wyngaert ◽  
Katrien De Troeyer ◽  
Bert Vaes ◽  
Mahmoud Alsaiqali ◽  
Bert Van Schaeybroeck ◽  
...  
Keyword(s):  
Logistic Regression ◽  
Nursing Homes ◽  
Confidence Interval ◽  
Crossover Study ◽  
Heat Waves ◽  
Conditional Logistic Regression ◽  
Odds Ratios ◽  
Mortality And Morbidity ◽  
Case Crossover ◽  
The Impact

Climate change leads to more days with extremely hot temperatures. Previous analyses of heat waves have documented a short-term rise in mortality. The results on the relationship between high temperatures and hospitalisations, especially in vulnerable patients admitted to nursing homes, are inconsistent. The objective of this research was to examine the discrepancy between heat-related mortality and morbidity in nursing homes. A time-stratified case-crossover study about the impact of heat waves on mortality and hospitalisations between 1 January 2013 and 31 December 2017 was conducted in 10 nursing homes over 5 years in Flanders, Belgium. In this study, the events were deaths and hospitalisations. We selected our control days during the same month as the events and matched them by day of the week. Heat waves were the exposure. Conditional logistic regression models were applied. The associations were reported as odds ratios at lag 0, 1, 2, and 3 and their 95% confidence intervals. In the investigated time period, 3048 hospitalisations took place and 1888 residents died. The conditional logistic regression showed that odds ratios of mortality and hospitalisations during heat waves were 1.61 (95% confidence interval 1.10–2.37) and 0.96 (95% confidence interval 0.67–1.36), respectively, at lag 0. Therefore, the increase in mortality during heat waves was statistically significant, but no significant changes in hospitalisations were obtained. Our result suggests that heat waves have an adverse effect on mortality in Flemish nursing homes but have no significant effect on the number of hospitalisations.

scholarly journals Fatal opioid overdoses during and shortly after hospital admissions in England: A case-crossover study

PLoS Medicine ◽  
2021 ◽  
Vol 18 (10) ◽  
pp. e1003759
Author(s):  
Dan Lewer ◽  
Brian Eastwood ◽  
Martin White ◽  
Thomas D. Brothers ◽  
Martin McCusker ◽  
...  
Keyword(s):  
Drug Use ◽  
Hospital Admission ◽  
Crossover Study ◽  
Hospital Admissions ◽  
Conditional Logistic Regression ◽  
Early Discharge ◽  
Odds Ratios ◽  
Drug Poisoning ◽  
Case Crossover ◽  
Increased Risk

Background Hospital patients who use illicit opioids such as heroin may use drugs during an admission or leave the hospital in order to use drugs. There have been reports of patients found dead from drug poisoning on the hospital premises or shortly after leaving the hospital. This study examines whether hospital admission and discharge are associated with increased risk of opioid-related death. Methods and findings We conducted a case-crossover study of opioid-related deaths in England. Our study included 13,609 deaths between January 1, 2010 and December 31, 2019 among individuals aged 18 to 64. For each death, we sampled 5 control days from the period 730 to 28 days before death. We used data from the national Hospital Episode Statistics database to determine the time proximity of deaths and control days to hospital admissions. We estimated the association between hospital admission and opioid-related death using conditional logistic regression, with a reference category of time neither admitted to the hospital nor within 14 days of discharge. A total of 236/13,609 deaths (1.7%) occurred following drug use while admitted to the hospital. The risk during hospital admissions was similar or lower than periods neither admitted to the hospital nor recently discharged, with odds ratios 1.03 (95% CI 0.87 to 1.21; p = 0.75) for the first 14 days of an admission and 0.41 (95% CI 0.30 to 0.56; p < 0.001) for days 15 onwards. 1,088/13,609 deaths (8.0%) occurred in the 14 days after discharge. The risk of opioid-related death increased in this period, with odds ratios of 4.39 (95% CI 3.75 to 5.14; p < 0.001) on days 1 to 2 after discharge and 2.09 (95% CI 1.92 to 2.28; p < 0.001) on days 3 to 14. 11,629/13,609 deaths (85.5%) did not occur close to a hospital admission, and the remaining 656/13,609 deaths (4.8%) occurred in hospital following admission due to drug poisoning. Risk was greater for patients discharged from psychiatric admissions, those who left the hospital against medical advice, and those leaving the hospital after admissions of 7 days or more. The main limitation of the method is that it does not control for time-varying health or drug use within individuals; therefore, hospital admissions coinciding with high-risk periods may in part explain the results. Conclusions Discharge from the hospital is associated with an acute increase in the risk of opioid-related death, and 1 in 14 opioid-related deaths in England happens in the 2 weeks after the hospital discharge. This supports interventions that prevent early discharge and improve linkage with community drug treatment and harm reduction services.

scholarly journals The rate ratio of injury and aggressive incident for alcohol alone, cocaine alone and simultaneous use before the event: A case–crossover study

2015 ◽  
Vol 75 ◽  
pp. 137-143 ◽  
Author(s):  
Jinhui Zhao ◽  
Scott Macdonald ◽  
Guilherme Borges ◽  
Chantele Joordens ◽  
Tim Stockwell ◽  
...  
Keyword(s):  
Crossover Study ◽  
Rate Ratio ◽  
Case Crossover ◽  
Simultaneous Use

scholarly journals Multiple sclerosis relapse risk in the postoperative period: Effects of invasive surgery and anesthesia

2019 ◽  
Vol 26 (11) ◽  
pp. 1437-1440
Author(s):  
Lindsey B De Lott ◽  
Samantha Zerafa ◽  
Kerby Shedden ◽  
Galit Levi Dunietz ◽  
Michelle Earley ◽  
...  
Keyword(s):  
Multiple Sclerosis ◽  
Logistic Regression ◽  
Medical Records ◽  
Odds Ratio ◽  
Conditional Logistic Regression ◽  
Unadjusted Odds Ratio ◽  
Relapse Risk ◽  
Multiple Sclerosis Relapse ◽  
Period Effects ◽  
Tertiary Center

Background: Postoperative multiple sclerosis (MS) relapses are a concern among patients and providers. Objective: To determine whether MS relapse risk is higher postoperatively. Methods: Data were extracted from medical records of MS patients undergoing surgery at a tertiary center (2000–2016). Conditional logistic regression estimated within-patient unadjusted and age-adjusted odds of postoperative versus preoperative relapse. Results: Among 281 patients and 609 surgeries, 12 postoperative relapses were identified. The odds of postoperative versus preoperative relapse in unadjusted (odds ratio (OR) = 0.56, 95% confidence interval (CI) = 0.18–1.79; p = 0.33) or age-adjusted models (OR = 0.66, 95% CI = 0.20–2.16; p = 0.49) were not increased. Conclusions: Surgery/anesthesia exposure did not increase postoperative relapse risk. These findings require confirmation in larger studies.

scholarly journals The Role of Stroke as a Trigger for Incident Venous Thromboembolism: Results from a Population-based Case-Crossover Study

TH Open ◽  
2019 ◽  
Vol 03 (01) ◽  
pp. e50-e57
Author(s):  
Vânia Morelli ◽  
Joakim Sejrup ◽  
Birgit Småbrekke ◽  
Ludvig Rinde ◽  
Gro Grimnes ◽  
...  
Keyword(s):  
Venous Thromboembolism ◽  
Acute Stroke ◽  
Crossover Study ◽  
Conditional Logistic Regression ◽  
Population Based ◽  
Red Blood Cell Transfusion ◽  
Odds Ratios ◽  
Case Crossover ◽  
Increased Risk

AbstractStroke is associated with a short-term increased risk of subsequent venous thromboembolism (VTE). It is unclear to what extent this association is mediated by stroke-related complications that are potential triggers for VTE, such as immobilization and infection. We aimed to investigate the role of acute stroke as a trigger for incident VTE while taking other concomitant VTE triggers into account. We conducted a population-based case-crossover study with 707 VTE patients. Triggers were registered during the 90 days before a VTE event (hazard period) and in four preceding 90-day control periods. Conditional logistic regression was used to estimate odds ratios with 95% confidence intervals (CIs) for VTE according to triggers. Stroke was registered in 30 of the 707 (4.2%) hazard periods and in 6 of the 2,828 (0.2%) control periods, resulting in a high risk of VTE, with odds ratios of 20.0 (95% CI: 8.3–48.1). After adjustments for immobilization and infection, odds ratios for VTE conferred by stroke were attenuated to 6.0 (95% CI: 1.6–22.1), and further to 4.0 (95% CI: 1.1–14.2) when other triggers (major surgery, red blood cell transfusion, trauma, and central venous catheter) were added to the regression model. A mediation analysis revealed that 67.8% of the total effect of stroke on VTE risk could be mediated through immobilization and infection. Analyses restricted to ischemic stroke yielded similar results. In conclusion, acute stroke was a trigger for VTE, and the association between stroke and VTE risk appeared to be largely mediated by immobilization and infection.

Proton pump inhibitor-related headaches: A nationwide population-based case-crossover study in Taiwan

Cephalalgia ◽  
2014 ◽  
Vol 35 (3) ◽  
pp. 203-210 ◽  
Author(s):  
Jen-Feng Liang ◽  
Yung-Tai Chen ◽  
Jong-Ling Fuh ◽  
Szu-Yuan Li ◽  
Tzeng-Ji Chen ◽  
...  
Keyword(s):  
Proton Pump Inhibitor ◽  
Crossover Study ◽  
Proton Pump ◽  
Conditional Logistic Regression ◽  
Population Based ◽  
Female Patients ◽  
Case Crossover ◽  
Increased Risk ◽  
Tension Type Headaches ◽  
Acute Headache

Background Headaches resulting from proton pump inhibitor (PPI) use could cause discontinuation of PPI in as many as 40% of patients who experience such headaches. Previous studies focusing on acute headache risk from PPI use are rare and limited to clinical trials of a single PPI. Objectives To investigate the association between PPI use and headache with a nationwide population-based case-crossover study. Methods Records containing the first diagnosis of any headache, including migraine and tension-type headaches, were retrieved from Taiwan National Health Insurance Database (1998–2010). We compared the rates of PPI use for cases and controls during time windows of 7, 14, and 28 days. The adjusted self-matched odds ratios (ORs) and 95% confidence intervals (CIs) from a conditional logistic regression model were used to determine the association between PPI use and headache. Results Overall, 314,210 patients with an initial diagnosis of any headache during the study period were enrolled. The adjusted ORs for headache risk after PPI exposure were calculated for three time periods (within 7 days = 1.41, p = 0.002, 95% CI 1.14–1.74; within 14 days = 1.36, p < 0.001, 95% CI 1.16–1.59; within 28 days = 1.20, p = 0.002, 95% CI 1.07–1.35). Subgroup analyses showed female patients had an increased risk of headache. Among PPIs, lansoprazole and esomeprazole had the highest risks of headache incidence, which were similar to that of nitrates. Conclusion PPI usage is associated with an increased risk for acute headache. Female patients and use of lansoprazole or esomeprazole present the greatest risks of headache.

Sign in / Sign up

Export Citation Format

Share Document