How likely are the calculators to improve the design of epidemiological studies involving biomonitoring of exposure?

Answer Explanations

• Likely

  Expert 9
  As presented, these calculators seem likely to be used and if that happens, could improve the design of these studies.  However, to answer this question adequately requires a comparison to other calculators and resources already available to researchers, which I recommend in my other comments.
• Unlikely

  Expert 3
   I do not see a substantial improvement over existing calculators. The presented calculators do not seem to address current needs effectively. Moreover, I do not expect most epidemiologists to use them, as the information on exposure sample size calculations is several decades old. The issue seems more psychological than technical. 
• Likely

  Expert 4
  The classical measurement error framework should be used at design stage, but unfortunately it was seldom used in the literature. Hopefully the current calculator can promote this framework to non-statisticians.
  
  
• Likely

  Expert 6
  They would be even more useful with the changes I recommended above.  Inclusion of within-subject variability in outcome would also be helpful.
• Likely

  Expert 7
  I think sample size calculations are rarely useful, because sample sizes are often driven by the available data or resources.   I think these calculators are more useful to understand the trade-offs between sample sizes (number of individuals sampled, n), number of repeats (samples per individual, m). As well as evaluations of   biases in estimated regression coefficients.   
• Very likely

  Expert 1
  I think researchers, clinicians, and students would be very likely to use these calculators. Not only do they provide the individual with the calculation process and formulas, but they importantly provide graphical representation of the calculation itself. This will be very useful when trying to work out a sample size, Power, Bias, etc... for a given budget and what the optimal data output will be for a given exposure plan. Based on my own experience with online epidemiological calculators, when I find one that provides all the information I am looking for, is easy to use, and is updated regularly with references and bug fixers I use them all the time. However, it can be very frustrating when these calculators for no reason disappear only to leave me with no other good option! I think an online site with these calculators available for biomonitoring studies would be an excellent resource that many would use daily! 
• Likely

  Expert 8
  In line with my previous answer: particularly useful for those who would otherwise not incorporate it. But I would also use them, or share them with my students and postdocs. 
  
  Important in study proposals; in making sure that the measurement methods, frequency of measurements, and sample sizes are adequate to answer the research question. Very important in avoiding more studies with a conclusion 'we did not find any associations, but that may be because of measurement error / being underpowered'. The latter is unfortunately very common and a waste of research time/energy/materials, it's crucial to think of this before starting any studies and improving methods or sample size to make sure methods and sample sizes are adequate. 
• Neutral

  Expert 5
  They could possibly as they are very user friendly and accessible.  However, there are a number of other calculators and programs published in the literature, various websites, SAS or R programming guides, that may do the same type of power and sample size calculations. 
• Likely

  Expert 2
  I recognize that the calculators provide a handy toolkit for improving epidemiology study designs by addressing key aspects that are often overlooked, such as the need for repeated measurements and the trade-offs between sample size and measurement error. However, these calculators are focused on classical measurement error and do not account for all the complexities typically encountered in real-world studies.
  Overall, while these calculators are likely to improve the design of epidemiological studies by helping researchers better account for measurement error in biomonitoring data, leading to more robust and valid findings, they are limited in scope. In practice, the design of epidemiological studies involving biomonitoring for exposure assessment is far more complex than the scenarios considered by the calculators. Factors such as the number of covariates/confounders, skewed error distributions, uncertainty in the relationships between covariates (covariance issues), and non-linear exposure-response relationships (which have become more evident in recent years) are critical elements that these tools do not fully address. Therefore, while useful, the calculators can only partly assist in considering the full complexity of epidemiology study designs.
  
  

  .