Hi all,
I have a big data set and am trying to QC some variables such as blood pressure and cholesterol measurement. I am trying to see if there is any drift pattern/consistency of the measurements taken by the devices we have overtime. Basically we have multiple daily (1 year worth of data) measurements for each variable and devices. Eg:
device date sbp dbp cholesterol
1 22/8/18 108
1. 22/8/18 110
2 22/8/18 115
2 23/8/18 118
2 23/8/18 115
My questions:
1. What is the best way to analyse these dataset? Do you think CUMSUM plot is applicable? Any examples?
2. I tried taking the weekly average reading and then do a basic scatter plot ? But I have problem adding the mean line, and +/- SD line into the plot to identify outlier measurements. I keep getting error message when using the common yline? Eg of my command:
global mean1= r(mean)
global lowerCL1= r(mean) - 1*r(sd)
global upperCL1= r(mean) + 1*r(sd)
twoway (scatter ave_height weekly_date, ///
legend(off) mcolor(black) ///
yline($mean1, lpattern(shortdash) lcolor(gray))///
yline($lowerCL1, lpattern(dash) lcolor(gray)) ///
yline($upperCL1, lpattern(dash) lcolor(gray)) ///
graphregion(color(white)) ylabel(, grid glcolor(gs14)) ///
ylabel(150(1)154) xlabel(1(2)60) ///
aspectratio(1.09) //
Please help!
Many thanks!
0 Response to Data quality control check with CUMSUM
Post a Comment