While we used regression to model seasonality of outcomes, controlling for maternal characteristics, this was not an exhaustive model and was missing potential factors, which were unmeasured in our study such as maternal infections and nutritional deficiencies.
In this paper, we go beyond the use of control outcomes to detect possible unobserved confounding and propose to use control outcomes in a simple but formal counterfactual-based approach to correct causal effect estimates for bias due to unobserved confounding.
Although the original study considered a variety of chromosome abnormalities, we proceed as in the report by Rosenbaum (2), who focused on these particular outcomes to illustrate the use of negative control outcomes to detect the presence of unobserved confounding.
The use of control dyads partially controls for the Hawthorne effect on outcome measures.
To assess the relationship between adherence and health resource use, a logistic regression model was fitted for each of the resource use outcomes, controlling for age, gender, marital status, ethnicity, education, household income, insurance and the number of comorbidities.
Besides the appropriate outcome measures, the proper use of a control group is also a critical issue in designing a high-quality clinical trial.
In the context of randomized controlled trials, one must consider how the population of interest is sampled and randomly assigned, the use of a control group and the outcome measures to be employed.
However, the methods of these follow-up studies vary widely, with sources of variability including inclusion criteria, timing of assessments, outcome measures and use of appropriate control groups.
The major variables that will be assessed include presence of randomization, use of a control group, use of masked outcome assessment, and in an active-controlled trial, use of an adequate comparator and adequate enrollment to detect a difference if one exists.
Therefore, we aimed to systematically review the literature to search for evidence of: (1) the rate of infection from invasive lines placed in the ED, (2) the duration of catheterization, and (3) process outcomes such as the use of infection control procedures.
