Is this a good exclusion criteria or not?

[yog55677]

TL;DR Summary

I read a study with limited exclusion criteria, but does it do its job correctly?

I read a study that linked cell phone use to infertility. They monitored 2,110 men from 1993-2007. However, they picked all of their subjects from an infertility clinic, therefore their subjects were already experiencing fertility issues. They tried to correct for this bias by having these exclusion criteria:

Although smoking or alcohol consumption as well as systemic diseases, orchitis or varicocele were exclusion criteria, no other confounders were investigated

Is this a good exclusion critera, or is the study still flawed because they haven't excluded all possible factors that make someone infertile? (It sounds like they didn't measure key things, like BMI, stress, bathing habits, etc.)

 

[PeroK]

Just a minor point. Criteria is the plural; criterion is the singular.

 

[Ygggdrasil]

Summary:: I read a study with limited exclusion criteria, but does it do its job correctly?

I read a study that linked cell phone use to infertility. They monitored 2,110 men from 1993-2007. However, they picked all of their subjects from an infertility clinic, therefore their subjects were already experiencing fertility issues. They tried to correct for this bias by having these exclusion criteria:
Is this a good exclusion critera, or is the study still flawed because they haven't excluded all possible factors that make someone infertile? (It sounds like they didn't measure key things, like BMI, stress, bathing habits, etc.)

No. Ultimately, the study suffers from the same biases as other observational studies. The study compares people who use cell phones vs people who don't use cell phones. There are many differences between these groups, so if the study found any differences between the groups, the study is not able to attribute that difference to cell phone use. Yes, the study can attempt to control for some differences, but as you note, it is nearly impossible to control for all factors. The study can show correlation between cell phone use and fertility, but it cannot definitiely prove any link. To get more definitive evidence of a link, one would need to perform a randomized controlled trial which would help to ensure that the only difference between the two groups studied is cell phone use, and not some other confounding factor.

 

[yog55677]

No. Ultimately, the study suffers from the same biases as other observational studies. The study compares people who use cell phones vs people who don't use cell phones. There are many differences between these groups, so if the study found any differences between the groups, the study is not able to attribute that difference to cell phone use. Yes, the study can attempt to control for some differences, but as you note, it is nearly impossible to control for all factors. The study can show correlation between cell phone use and fertility, but it cannot definitiely prove any link. To get more definitive evidence of a link, one would need to perform a randomized controlled trial which would help to ensure that the only difference between the two groups studied is cell phone use, and not some other confounding factor.

Thanks! I just realized that they found no link between phone use and sperm count, but rather a link between phone use and a rise in certain hormones (Testosterone, etc.) that can harm fertility and sperm morphology was worse among cell phone users. But I guess the bias of all of their subjects being from a fertility clinic (As all subjects will have had to experience some form of fertility problem to attend a clinic, therefore they all already have bad fertility) and the bad exclusion criteria means no conclusions should be made, am I right?

 

[berkeman]

Since the OP has left us, this thread can be closed now.