Help with ANCOVA analysis

[Harry Stern]

Hi,

I am trying to run an analysis where I compare change in systolic blood pressure from baseline to 12 months between participants with metabolic syndrome vs. participants without metabolic syndrome at baseline while adjusting for baseline systolic blood pressure. I believe I can do an ANCOVA analysis to do this, but I am not quite sure. I was hoping someone could provide some insight into what type of test I should run to do this analysis and how to do this in Stata.

Here is a small portion of the data that I am working with.

* Example generated by -dataex-. To install: ssc install dataex
clear
input long pid int(avgsbp1 avgsbp12) float(sbp_ch12 met_syn1)
30060 114 135  21 0
20076 134 131  -3 0
40171 153 146  -7 0
20074 111 108  -3 0
10015 123 126   3 0
20174 126 124  -2 0
20128 121 126   5 0
50068 143 123 -20 0
30047 142 131 -11 0
50011 141 126 -15 0
end
label values met_syn1 Met_syn
label def Met_syn 0 "- Metabolic Syndrome", modify

Variables:
pid - participant ID
avgsbp1 - average baseline blood pressure
avgsbp12 - average blood pressure at 12 months
sbp_ch12 - change in blood pressure from baseline to 12 months
met_syn1 - baseline metabolic syndrome (yes/no)

Thank you for your help.

[Roman Mostazir]

I suppose your group variable for metabolic syndrom with/without is "met_syn1":

reg avgsbp12 avgsbp1 i.met_syn1

margins i.met_syn1
marginsplot

[Joseph Coveney]

. . . I compare change in systolic blood pressure from baseline to 12 months . . . while adjusting for baseline systolic blood pressure. I believe I can do an ANCOVA . . .

So, something like

anova sbp_ch12 i.met_syn1 c.avgsbp1

or its equivalent*

anova avgsbp12 i.met_syn1 c.avgsbp1

ÿ
But if your research question is about change in systolic blood pressure over time and difference in it between two innately different (as opposed to randomly allocated) groups, then consider something like the following instead.

graph twoway ///
    scatter avgsbp12 avgsbp1 if !met_syn1, mcolor(red) msize(small) || ///
    scatter avgsbp12 avgsbp1 if met_syn1, mcolor(blue) msize(small) ///
        ylabel( , angle(horizontal) nogrid) legend(off)

or some variation on it, such as adding overlaid lfit or qfit plots, separately fitted by group.

And then maybe

ttest sbp_ch12, by(met_syn1) welch

ÿ
ÿ
* See, for example, Note 2 here:

.ÿ
.ÿversionÿ16.1

.ÿ
.ÿclearÿ*

.ÿ
.ÿsetÿseedÿ`=strreverse("1583080")'

.ÿ
.ÿquietlyÿdrawnormÿout0ÿout1,ÿdoubleÿcorr(1ÿ0.5ÿ\ÿ0.5ÿ1)ÿn(50)

.ÿgenerateÿbyteÿtrtÿ=ÿmod(_n,ÿ2)

.ÿ
.ÿ*
.ÿ*ÿBeginÿhere
.ÿ*
.ÿregressÿout1ÿc.out0ÿi.trt

ÿÿÿÿÿÿSourceÿ|ÿÿÿÿÿÿÿSSÿÿÿÿÿÿÿÿÿÿÿdfÿÿÿÿÿÿÿMSÿÿÿÿÿÿNumberÿofÿobsÿÿÿ=ÿÿÿÿÿÿÿÿ50
-------------+----------------------------------ÿÿÿF(2,ÿ47)ÿÿÿÿÿÿÿÿ=ÿÿÿÿÿÿ4.53
ÿÿÿÿÿÿÿModelÿ|ÿÿ7.06863153ÿÿÿÿÿÿÿÿÿ2ÿÿ3.53431576ÿÿÿProbÿ>ÿFÿÿÿÿÿÿÿÿ=ÿÿÿÿ0.0159
ÿÿÿÿResidualÿ|ÿÿ36.7061105ÿÿÿÿÿÿÿÿ47ÿÿ.780981074ÿÿÿR-squaredÿÿÿÿÿÿÿ=ÿÿÿÿ0.1615
-------------+----------------------------------ÿÿÿAdjÿR-squaredÿÿÿ=ÿÿÿÿ0.1258
ÿÿÿÿÿÿÿTotalÿ|ÿÿÿ43.774742ÿÿÿÿÿÿÿÿ49ÿÿ.893362082ÿÿÿRootÿMSEÿÿÿÿÿÿÿÿ=ÿÿÿÿ.88373

------------------------------------------------------------------------------
ÿÿÿÿÿÿÿÿout1ÿ|ÿÿÿÿÿÿCoef.ÿÿÿStd.ÿErr.ÿÿÿÿÿÿtÿÿÿÿP>|t|ÿÿÿÿÿ[95%ÿConf.ÿInterval]
-------------+----------------------------------------------------------------
ÿÿÿÿÿÿÿÿout0ÿ|ÿÿÿ.3920431ÿÿÿ.1304409ÿÿÿÿÿ3.01ÿÿÿ0.004ÿÿÿÿÿ.1296298ÿÿÿÿ.6544564
ÿÿÿÿÿÿÿ1.trtÿ|ÿÿÿ.0496016ÿÿÿ.2514762ÿÿÿÿÿ0.20ÿÿÿ0.844ÿÿÿÿ-.4563032ÿÿÿÿ.5555064
ÿÿÿÿÿÿÿ_consÿ|ÿÿÿ-.161906ÿÿÿ.1788433ÿÿÿÿ-0.91ÿÿÿ0.370ÿÿÿÿ-.5216922ÿÿÿÿ.1978802
------------------------------------------------------------------------------

.ÿtempnameÿaÿd

.ÿscalarÿdefineÿ`a'ÿ=ÿ_b[1.trt]

.ÿscalarÿdefineÿ`d'ÿ=ÿ_b[out0]

.ÿ
.ÿgenerateÿdoubleÿdeltaÿ=ÿout1ÿ-ÿout0

.ÿregressÿdeltaÿc.out0ÿi.trt

ÿÿÿÿÿÿSourceÿ|ÿÿÿÿÿÿÿSSÿÿÿÿÿÿÿÿÿÿÿdfÿÿÿÿÿÿÿMSÿÿÿÿÿÿNumberÿofÿobsÿÿÿ=ÿÿÿÿÿÿÿÿ50
-------------+----------------------------------ÿÿÿF(2,ÿ47)ÿÿÿÿÿÿÿÿ=ÿÿÿÿÿ11.12
ÿÿÿÿÿÿÿModelÿ|ÿÿ17.3622909ÿÿÿÿÿÿÿÿÿ2ÿÿ8.68114544ÿÿÿProbÿ>ÿFÿÿÿÿÿÿÿÿ=ÿÿÿÿ0.0001
ÿÿÿÿResidualÿ|ÿÿ36.7061105ÿÿÿÿÿÿÿÿ47ÿÿ.780981074ÿÿÿR-squaredÿÿÿÿÿÿÿ=ÿÿÿÿ0.3211
-------------+----------------------------------ÿÿÿAdjÿR-squaredÿÿÿ=ÿÿÿÿ0.2922
ÿÿÿÿÿÿÿTotalÿ|ÿÿ54.0684014ÿÿÿÿÿÿÿÿ49ÿÿ1.10343676ÿÿÿRootÿMSEÿÿÿÿÿÿÿÿ=ÿÿÿÿ.88373

------------------------------------------------------------------------------
ÿÿÿÿÿÿÿdeltaÿ|ÿÿÿÿÿÿCoef.ÿÿÿStd.ÿErr.ÿÿÿÿÿÿtÿÿÿÿP>|t|ÿÿÿÿÿ[95%ÿConf.ÿInterval]
-------------+----------------------------------------------------------------
ÿÿÿÿÿÿÿÿout0ÿ|ÿÿ-.6079569ÿÿÿ.1304409ÿÿÿÿ-4.66ÿÿÿ0.000ÿÿÿÿ-.8703702ÿÿÿ-.3455436
ÿÿÿÿÿÿÿ1.trtÿ|ÿÿÿ.0496016ÿÿÿ.2514762ÿÿÿÿÿ0.20ÿÿÿ0.844ÿÿÿÿ-.4563032ÿÿÿÿ.5555064
ÿÿÿÿÿÿÿ_consÿ|ÿÿÿ-.161906ÿÿÿ.1788433ÿÿÿÿ-0.91ÿÿÿ0.370ÿÿÿÿ-.5216922ÿÿÿÿ.1978802
------------------------------------------------------------------------------

.ÿassertÿfloat(_b[1.trt])ÿ==ÿfloat(`a')

.ÿassertÿ`d'ÿ==ÿ1ÿ+ÿ_b[out0]

.ÿ
.ÿexit

endÿofÿdo-file


.