Are Nested Case-Control Studies Biased?

Langholz, Bryan; Richardson, David B.

doi:10.1097/ede.0b013e31819e370b

Cited by 21 publications

(18 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…There has been some recent debate in the literature about study design bias in nested case-control studies [21]. However, methodological flaws discussed concern mainly analysis of lagged exposure data, which is not the case in the investigated association (ie, pRBC transfusion and ICUacquired BSIs).…”

Section: Limitations Of Our Studymentioning

confidence: 99%

Impact of red blood cells transfusion on ICU-acquired bloodstream infections: A case-control study

Michalia¹,

Kompoti²,

Panagiotakopoulou³

et al. 2012

Journal of Critical Care

View full text Add to dashboard Cite

Section: Limitations Of Our Studymentioning

confidence: 99%

Impact of red blood cells transfusion on ICU-acquired bloodstream infections: A case-control study

Michalia¹,

Kompoti²,

Panagiotakopoulou³

et al. 2012

Journal of Critical Care

View full text Add to dashboard Cite

“…In this study, no bias was observed when a simple random sample of controls was selected from the risk set for each case identified using attained age alone. 15 Hein et al 16 simulated occupational cohorts with time-dependent exposures to evaluate incidence-density sampling in the presence of exposure lagging. In these simulations, exposure effect estimates from analyses that matched on attained age alone were unbiased, and, in lagged analyses, including lagged-out cases and controls (assigned zero lagged exposure) did not introduce bias.…”

Section: Discussionmentioning

confidence: 99%

“…Risk-set sampling in conjunction with exposure lagging was evaluated by Langholz and Richardson 15 in a simulation study based on the Colorado Plateau uranium miners cohort. In this study, no bias was observed when a simple random sample of controls was selected from the risk set for each case identified using attained age alone.…”

Section: Discussionmentioning

confidence: 99%

“…13,14 Two simulation studies to evaluate the nested case-control study design in conjunction with exposure lagging did not observe bias. 15,16 Langholz and Richardson 15 further described the theoretical basis for nested case-control studies both with and without exposure lagging. These studies have nonetheless been criticized because they did "not fully model important associations" in the Reading cohort.…”

mentioning

confidence: 99%

See 1 more Smart Citation

Evaluating Bias From Birth-cohort Effects in the Age-based Cox Proportional Hazards Model

Hein

Schubauer‐Berigan²,

Deddens³

2011

Epidemiology

View full text Add to dashboard Cite

The nested case-control design is frequently used to evaluate exposures and health outcomes within the confines of a cohort study. When incidence-density sampling is used to identify controls, the resulting data can be analyzed using conditional logistic regression (equivalent to stratified Cox proportional hazards regression). In these studies, exposure lagging is often used to account for disease latency. In light of recent criticism of incidence-density sampling, we used simulated occupational cohorts to evaluate age-based incidence-density sampling for lagged exposures in the presence of birth-cohort effects and associations among time-related variables. Effect estimates were unbiased when adjusted for birth cohort; however, unadjusted effect estimates were biased, particularly when age at hire and year of hire were correlated. When the analysis included an adjustment for birth cohort, the inclusion of lagged-out cases and controls (assigned a lagged exposure of zero) did not introduce bias.

show abstract

“…Levy and colleagues32 proposed that controls sampled from age-based risk sets should be matched on the age at date of last observation in order to correct a supposed ‘imbalance’ between cases and controls. Theoretical considerations33 34 question this new matching criterion, and empirical simulations12 35 implementing it have found that it produces a notable downward bias in risk estimates. Wacholder36 notes that traditional risk set sampling is expected to produce unbiased results, even when exposure lagging is used.…”

Section: Discussionmentioning

confidence: 99%

Risk of lung cancer associated with quantitative beryllium exposure metrics within an occupational cohort

et al. 2010

View full text Add to dashboard Cite

Objectives Beryllium has been identified as a human carcinogen on the basis of animal and epidemiological studies. The authors recently reported updated associations between lung cancer and beryllium exposure in a large, pooled occupational cohort. The authors conducted the present study to evaluate the shape of exposureeresponse associations between different exposure metrics and lung cancer in this cohort, considering potential confounders (race, plant, professional and short-term work status, and exposure to other lung carcinogens). Methods The authors conducted Cox proportional hazards regression analyses of lung cancer risk with cumulative, mean and maximum 'daily weighted average' (DWA) exposure among 5436 workers, using age-based risk sets. Different exposureeresponse curves were fitted to the exposure metrics, including categorical, power, restricted cubic spline and piecewise log-linear fits. Results The authors found significant positive associations between lung cancer and mean (p<0.0001) and maximum (p<0.0001) exposure, adjusting for age, birth cohort and plant, and for cumulative (p¼0.0017) beryllium exposure, adjusting for these factors plus short-term work status and exposure to asbestos. The best-fitting models were generally categorical or piecewise log-linear, with the steepest increase in lung cancer risk between 0 and 10 mg/m 3 for both mean and maximum DWA exposure and between 0 and 200 mg/m 3 -days for cumulative DWA exposure. The estimated mean DWA beryllium exposure associated with 10 À3 excess lifetime risk based on the piecewise log-linear model is 0.033 mg/m 3 . Conclusion This study provides evidence that lung cancer risk is elevated at levels near the current US Occupational Safety and Health Administration beryllium exposure limit of 2.0 mg/m 3 DWA for workers.Beryllium exposure has long been associated with an immune-mediated, granulomatous lung disease termed chronic beryllium disease. 1 Toxicological and epidemiological studies have more recently led the International Agency for Research on Cancer to designate beryllium as a human carcinogen. Epidemiological evidence for the carcinogenicity of beryllium derives largely from studies within a pooled cohort of male workers 4 5 at seven US beryllium processing plants, and a nested casee control study of lung cancer at one of the plants. 6 7 We recently reported updated associations between cumulative and maximum beryllium exposure and several diseases of a priori interest (including lung cancer), within a three-plant subcohort having detailed beryllium exposure data. 5Although smoking data were available for only a subset (25%) of cohort members, the positive exposureeresponse associations observed in that study showed little evidence of confounding by smoking, as measured by indirect methods.The present study was conducted to evaluate the association of mean, maximum and cumulative beryllium exposure with lung cancer mortality, considering potential confounders, including age, race, birth cohort, exposure to other lung ca...

show abstract

Are Nested Case-Control Studies Biased?

Cited by 21 publications

References 30 publications

Impact of red blood cells transfusion on ICU-acquired bloodstream infections: A case-control study

Impact of red blood cells transfusion on ICU-acquired bloodstream infections: A case-control study

Evaluating Bias From Birth-cohort Effects in the Age-based Cox Proportional Hazards Model

Risk of lung cancer associated with quantitative beryllium exposure metrics within an occupational cohort

Contact Info

Product

Resources

About