1. Dave suggests to make different “birthing surface” for different hormones to make sure it cross “only once”.
2. Have think about it and Don’t have good suggestions
3. Back to use “CRH” and work on criteria to make sure have more preterm and less term and postterm
1. Improve single parameter risk by adjusting “interval”
2. Adjust “Criteria” to improve the multiple parameters’ risk
1. Single parameter: CRH
2. Two parameters: CRH-E2
3. Three parameters:CRH-E2-AFP
4. Four Parameters: CRH-E2-AFP-ALP
5. Five Parameters: CRH-E2-AFP-ALP-Cortisol
1. David suggests to curve fitting for risk rate to get a smooth curve for before.
2. We tested and try to use “a*exp(b*x)+c*exp(d*x)”.
3. We’ve proved that CRHALPCortisol=CRHALP “AND” CRHCortiol;
CRHALPCortisol=CRHCortisol “AND” CRHALP:
Choose CRH as final study standard for two reasons:
1 . risk is the second high
2. no multiple intersection with “birthing surface”
3. Although “hCG” is highest risk, but involves multiple intersection with “birthing surface”.
1. Today has finished dynamic plotting “birthing surface” with “risk rate”
2. Tomorrow should focus on how to improve the “risk rate” by adjusting interval first
3. then start to play with criteria
1. keep the interval fixed
2. only change the criteria to find the best
3. I think it is useful to try 8 parameter’s risk seperately
4. Then choose the largest risk to be studied and start to filter the data out
1. Curve fitting for “Median Trajectory” and “CI” to smooth it.
2. Plot CI for both side to better compare
1. May test all 8 parameters to find the highest value of risk rate
2. Based on that parameter to fiter the second parameter
3. Keep testing criteria to filter the data out
4. Try to adject the “Interva” (<32, 32-37….)
1. Keep working on “Birthing Surface” with original data
2. Try: CRH
CRH+Cortisol,
CRH+ALP,
CRH+Cortisol+ALP
Two problems to care:
1. How to define the criteria to get more term patients
2. What’s the best comination sequence to test (CRH+Coritisol/ALP????)
