I am trying to collect between- and within- person variance data for common environmental and medical biomarkers. Right now, I am focussing on hydroxy PAHs such oh-pyrene in urine, cytokines, lipid (cholesterol) in blood, pesticides in various biological media, volatiles (chloroform, BTEX) in breath, etc.
Surprisingly, these data are difficult to find in the literature. Would love some assistance.
Gregory - I agree 100%. In our own studies we always try to incorporate a repeat measures data structure. However, right now I'm in the process of trying to develop guidance for interpreting large databases of biomarker measurements that (sadly) do not contain repeat measures. We would like to get to some confidence about an individual mean; an impossible task without at least a hint about the ICC or some other form of variance component meta-data.
Something of interest, even if it is a bit late.
George Box and his friends performed an experiment where they ran a Gauge R&R study on blood samples and cholesterol levels. George and a few others gave blood samples, and a lot of them, and sent 2-3 samples from each person out to different medical testing labs. The results showed that the cholesterol levels of each participant were similar within any one lab. However, some labs showed a large positive bias. Others showed a large negative bias. So, according to the results, some labs reported really high cholesterol levels that would require medication. Others showed cholesterol levels that were so low, they were unhealthy.
Having worked in commercial testing facilities, I know the same thing happens with environmental samples too.
I would try to set up a Gauge R&R study with repeated measures.
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Modeling
- Analyses of Variance
- ANOVA