GEE advice needed

[JK Neo]

Dear all

need help regarding GEE modelling.

I have collected data for subjects over several years at specific visits and would like to find out the mean value at each time point, adjusted for the different variables.

This is how my data looks like. SubID being the subject of interest, visit being the year of visit (which is year 1,3,4,5,7 but not everyone has a total of 5 visits. Some people only have 3 visits or less). The outcome variable of interest is global which is a continuous data. The rest of the variables are explanatory variable i would like to adjust for.

Essentially that is how i will like my data to look in the final summary, where each value is the mean at each time point and the different models are adjusted for the variables of interest. Can i seek your help to educate me on how to perform a GEE on this data set?

[Joseph Coveney]

Maybe something like the following? (Start at the "Begin here" comment; the stuff above is for creating an illustrative toy dataset.)

Is there something in particular in the help files or user's manual entries for xtgee, margins or testparm that is causing you trouble?

.ÿ
.ÿversionÿ17.0

.ÿ
.ÿclearÿ*

.ÿ
.ÿ//ÿseedem
.ÿsetÿseedÿ877151124

.ÿ
.ÿquietlyÿsetÿobsÿ200

.ÿgenerateÿstrÿsidÿ=ÿ"HD"ÿ+ÿstring(_n,ÿ"%03.0f")

.ÿlabelÿvariableÿsidÿParticipantID

.ÿgenerateÿdoubleÿpid_uÿ=ÿrnormal()

.ÿ
.ÿlabelÿdefineÿNYÿ0ÿNÿ1ÿY

.ÿlabelÿdefineÿSexesÿ0ÿMÿ1ÿF

.ÿgenerateÿbyteÿeduÿ=ÿ4ÿ*ÿruniformint(0,ÿ4)

.ÿgenerateÿbyteÿageÿ=ÿruniformint(50,ÿ90)

.ÿforeachÿvarÿofÿnewlistÿsexÿhtnÿlpdÿdbmÿcvdÿ{
ÿÿ2.ÿÿÿÿÿÿÿÿÿgenerateÿbyteÿ`var'ÿ=ÿrbinomial(1,ÿruniform())
ÿÿ3.ÿÿÿÿÿÿÿÿÿifÿ"`var'"ÿ==ÿ"sex"ÿlabelÿvaluesÿ`var'ÿSexes
ÿÿ4.ÿÿÿÿÿÿÿÿÿelseÿlabelÿvaluesÿ`var'ÿNY
ÿÿ5.ÿ}

.ÿlabelÿvariableÿeduÿEducation

.ÿlabelÿvariableÿageÿAge

.ÿlabelÿvariableÿsexÿSex

.ÿlabelÿvariableÿhtnÿHypertension

.ÿlabelÿvariableÿlpdÿHyperlipidemia

.ÿlabelÿvariableÿdbmÿDiabetes

.ÿlabelÿvariableÿcvdÿCeVD

.ÿ
.ÿquietlyÿexpandÿ5

.ÿbysortÿsid:ÿgenerateÿbyteÿvstÿ=ÿ_n

.ÿlabelÿvariablÿvstÿVisit

.ÿquietlyÿrecodeÿvstÿ(5=7)ÿ(4=5)ÿ(3=4)ÿ(2=3)

.ÿquietlyÿdropÿifÿvstÿ>ÿ1ÿ&ÿruniform()ÿ>ÿ0.9

.ÿ
.ÿgenerateÿdoubleÿscoÿ=ÿpid_uÿ+ÿrnormal()

.ÿlabelÿvariableÿscoÿGlobal

.ÿ
.ÿforeachÿvarÿofÿvarlistÿ_allÿ{
ÿÿ2.ÿÿÿÿÿÿÿÿÿlocalÿvarnameÿ:ÿvariableÿlabelÿ`var'
ÿÿ3.ÿÿÿÿÿÿÿÿÿcharÿdefineÿ`var'[varname]ÿ`varname'
ÿÿ4.ÿ}

.ÿ
.ÿlistÿsidÿvstÿscoÿageÿsexÿeduÿÿhtnÿlpdÿdbmÿcvdÿinÿ1/3,ÿnoobsÿsubvarname

ÿÿ+---------------------------------------------------------------------------------------------+
ÿÿ|ÿParticipÿÿÿVisitÿÿÿÿÿÿGlobalÿÿÿAgeÿÿÿSexÿÿÿEducatioÿÿÿHypertenÿÿÿHyperlipÿÿÿDiabetesÿÿÿCeVDÿ|
ÿÿ|---------------------------------------------------------------------------------------------|
ÿÿ|ÿÿÿÿHD001ÿÿÿÿÿÿÿ1ÿÿÿ1.4045535ÿÿÿÿ82ÿÿÿÿÿFÿÿÿÿÿÿÿÿÿÿ4ÿÿÿÿÿÿÿÿÿÿYÿÿÿÿÿÿÿÿÿÿNÿÿÿÿÿÿÿÿÿÿYÿÿÿÿÿÿYÿ|
ÿÿ|ÿÿÿÿHD001ÿÿÿÿÿÿÿ3ÿÿÿ1.7867626ÿÿÿÿ82ÿÿÿÿÿFÿÿÿÿÿÿÿÿÿÿ4ÿÿÿÿÿÿÿÿÿÿYÿÿÿÿÿÿÿÿÿÿNÿÿÿÿÿÿÿÿÿÿYÿÿÿÿÿÿYÿ|
ÿÿ|ÿÿÿÿHD001ÿÿÿÿÿÿÿ4ÿÿÿ.11078023ÿÿÿÿ82ÿÿÿÿÿFÿÿÿÿÿÿÿÿÿÿ4ÿÿÿÿÿÿÿÿÿÿYÿÿÿÿÿÿÿÿÿÿNÿÿÿÿÿÿÿÿÿÿYÿÿÿÿÿÿYÿ|
ÿÿ+---------------------------------------------------------------------------------------------+

.ÿ
.ÿ*
.ÿ*ÿBeginÿhere
.ÿ*
.ÿencodeÿsid,ÿgenerate(pid)ÿlabel(ParticipantIDs)

.ÿ
.ÿxtgeeÿscoÿc.ageÿi.(sexÿeduÿhtnÿlpdÿdbmÿcvdÿvst),ÿi(pid)ÿt(vst)ÿ///
>ÿÿÿÿÿÿÿÿÿfamily(gaussian)ÿlink(identity)ÿcorr(unstructured)ÿnolog

GEEÿpopulation-averagedÿmodelÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿNumberÿofÿobsÿÿÿÿ=ÿÿÿÿ919
Groupÿandÿtimeÿvars:ÿpidÿvstÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿNumberÿofÿgroupsÿ=ÿÿÿÿ200
Family:ÿGaussianÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿObsÿperÿgroup:ÿÿ
Link:ÿÿÿIdentityÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿminÿ=ÿÿÿÿÿÿ3
Correlation:ÿunstructuredÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿavgÿ=ÿÿÿÿ4.6
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿmaxÿ=ÿÿÿÿÿÿ5
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿWaldÿchi2(14)ÿÿÿÿ=ÿÿ18.27
Scaleÿparameterÿ=ÿ1.947632ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿProbÿ>ÿchi2ÿÿÿÿÿÿ=ÿ0.1945

------------------------------------------------------------------------------
ÿÿÿÿÿÿÿÿÿscoÿ|ÿCoefficientÿÿStd.ÿerr.ÿÿÿÿÿÿzÿÿÿÿP>|z|ÿÿÿÿÿ[95%ÿconf.ÿinterval]
-------------+----------------------------------------------------------------
ÿÿÿÿÿÿÿÿÿageÿ|ÿÿÿ.0041215ÿÿÿ.0064899ÿÿÿÿÿ0.64ÿÿÿ0.525ÿÿÿÿ-.0085985ÿÿÿÿ.0168415
ÿÿÿÿÿÿÿÿÿÿÿÿÿ|
ÿÿÿÿÿÿÿÿÿsexÿ|
ÿÿÿÿÿÿÿÿÿÿFÿÿ|ÿÿÿÿ.009472ÿÿÿ.1576239ÿÿÿÿÿ0.06ÿÿÿ0.952ÿÿÿÿ-.2994651ÿÿÿÿ.3184091
ÿÿÿÿÿÿÿÿÿÿÿÿÿ|
ÿÿÿÿÿÿÿÿÿeduÿ|
ÿÿÿÿÿÿÿÿÿÿ4ÿÿ|ÿÿÿ.0009721ÿÿÿ.2362719ÿÿÿÿÿ0.00ÿÿÿ0.997ÿÿÿÿ-.4621123ÿÿÿÿ.4640565
ÿÿÿÿÿÿÿÿÿÿ8ÿÿ|ÿÿÿ.4489268ÿÿÿ.2388593ÿÿÿÿÿ1.88ÿÿÿ0.060ÿÿÿÿ-.0192289ÿÿÿÿ.9170825
ÿÿÿÿÿÿÿÿÿ12ÿÿ|ÿÿÿ.3433746ÿÿÿ.2298456ÿÿÿÿÿ1.49ÿÿÿ0.135ÿÿÿÿ-.1071145ÿÿÿÿ.7938638
ÿÿÿÿÿÿÿÿÿ16ÿÿ|ÿÿÿ-.115553ÿÿÿ.2539495ÿÿÿÿ-0.46ÿÿÿ0.649ÿÿÿÿÿ-.613285ÿÿÿÿ.3821789
ÿÿÿÿÿÿÿÿÿÿÿÿÿ|
ÿÿÿÿÿÿÿÿÿhtnÿ|
ÿÿÿÿÿÿÿÿÿÿYÿÿ|ÿÿÿ.1503052ÿÿÿ.1556471ÿÿÿÿÿ0.97ÿÿÿ0.334ÿÿÿÿ-.1547575ÿÿÿÿÿ.455368
ÿÿÿÿÿÿÿÿÿÿÿÿÿ|
ÿÿÿÿÿÿÿÿÿlpdÿ|
ÿÿÿÿÿÿÿÿÿÿYÿÿ|ÿÿ-.0031624ÿÿÿ.1574289ÿÿÿÿ-0.02ÿÿÿ0.984ÿÿÿÿ-.3117173ÿÿÿÿ.3053926
ÿÿÿÿÿÿÿÿÿÿÿÿÿ|
ÿÿÿÿÿÿÿÿÿdbmÿ|
ÿÿÿÿÿÿÿÿÿÿYÿÿ|ÿÿÿÿ.260727ÿÿÿ.1589576ÿÿÿÿÿ1.64ÿÿÿ0.101ÿÿÿÿ-.0508241ÿÿÿÿ.5722781
ÿÿÿÿÿÿÿÿÿÿÿÿÿ|
ÿÿÿÿÿÿÿÿÿcvdÿ|
ÿÿÿÿÿÿÿÿÿÿYÿÿ|ÿÿ-.1951082ÿÿÿÿ.158677ÿÿÿÿ-1.23ÿÿÿ0.219ÿÿÿÿ-.5061095ÿÿÿÿ.1158931
ÿÿÿÿÿÿÿÿÿÿÿÿÿ|
ÿÿÿÿÿÿÿÿÿvstÿ|
ÿÿÿÿÿÿÿÿÿÿ3ÿÿ|ÿÿÿ.0122433ÿÿÿ.0993553ÿÿÿÿÿ0.12ÿÿÿ0.902ÿÿÿÿ-.1824895ÿÿÿÿ.2069761
ÿÿÿÿÿÿÿÿÿÿ4ÿÿ|ÿÿ-.0646945ÿÿÿ.1050752ÿÿÿÿ-0.62ÿÿÿ0.538ÿÿÿÿ-.2706381ÿÿÿÿÿ.141249
ÿÿÿÿÿÿÿÿÿÿ5ÿÿ|ÿÿÿ.0270244ÿÿÿ.0947281ÿÿÿÿÿ0.29ÿÿÿ0.775ÿÿÿÿ-.1586392ÿÿÿÿ.2126881
ÿÿÿÿÿÿÿÿÿÿ7ÿÿ|ÿÿÿ.1089386ÿÿÿ.1042782ÿÿÿÿÿ1.04ÿÿÿ0.296ÿÿÿÿ-.0954429ÿÿÿÿ.3133201
ÿÿÿÿÿÿÿÿÿÿÿÿÿ|
ÿÿÿÿÿÿÿ_consÿ|ÿÿ-.5351477ÿÿÿ.5397103ÿÿÿÿ-0.99ÿÿÿ0.321ÿÿÿÿÿ-1.59296ÿÿÿÿÿ.522665
------------------------------------------------------------------------------

.ÿmarginsÿvst,ÿat(dbmÿ=ÿ1)

PredictiveÿmarginsÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿNumberÿofÿobsÿ=ÿ919
ModelÿVCE:ÿConventional

Expression:ÿLinearÿprediction,ÿpredict()
At:ÿdbmÿ=ÿ1

------------------------------------------------------------------------------
ÿÿÿÿÿÿÿÿÿÿÿÿÿ|ÿÿÿÿÿÿÿÿÿÿÿÿDelta-method
ÿÿÿÿÿÿÿÿÿÿÿÿÿ|ÿÿÿÿÿMarginÿÿÿstd.ÿerr.ÿÿÿÿÿÿzÿÿÿÿP>|z|ÿÿÿÿÿ[95%ÿconf.ÿinterval]
-------------+----------------------------------------------------------------
ÿÿÿÿÿÿÿÿÿvstÿ|
ÿÿÿÿÿÿÿÿÿÿ1ÿÿ|ÿÿÿ.1339205ÿÿÿ.1339504ÿÿÿÿÿ1.00ÿÿÿ0.317ÿÿÿÿ-.1286174ÿÿÿÿ.3964583
ÿÿÿÿÿÿÿÿÿÿ3ÿÿ|ÿÿÿ.1461638ÿÿÿ.1361356ÿÿÿÿÿ1.07ÿÿÿ0.283ÿÿÿÿ-.1206571ÿÿÿÿ.4129847
ÿÿÿÿÿÿÿÿÿÿ4ÿÿ|ÿÿÿÿ.069226ÿÿÿ.1365699ÿÿÿÿÿ0.51ÿÿÿ0.612ÿÿÿÿ-.1984462ÿÿÿÿ.3368981
ÿÿÿÿÿÿÿÿÿÿ5ÿÿ|ÿÿÿ.1609449ÿÿÿ.1358893ÿÿÿÿÿ1.18ÿÿÿ0.236ÿÿÿÿ-.1053932ÿÿÿÿÿ.427283
ÿÿÿÿÿÿÿÿÿÿ7ÿÿ|ÿÿÿ.2428591ÿÿÿ.1371126ÿÿÿÿÿ1.77ÿÿÿ0.077ÿÿÿÿ-.0258767ÿÿÿÿ.5115949
------------------------------------------------------------------------------

.ÿtestparmÿi.vst

ÿ(ÿ1)ÿÿ3.vstÿ=ÿ0
ÿ(ÿ2)ÿÿ4.vstÿ=ÿ0
ÿ(ÿ3)ÿÿ5.vstÿ=ÿ0
ÿ(ÿ4)ÿÿ7.vstÿ=ÿ0

ÿÿÿÿÿÿÿÿÿÿÿchi2(ÿÿ4)ÿ=ÿÿÿÿ2.94
ÿÿÿÿÿÿÿÿÿProbÿ>ÿchi2ÿ=ÿÿÿÿ0.5671

.ÿ
.ÿ//ÿAlsoÿconsider
.ÿ/*ÿxtgeeÿscoÿi.(sexÿeduÿhtnÿlpdÿdbmÿcvdÿvst),ÿi(pid)ÿ///
>ÿÿÿÿÿÿÿÿÿfamily(gaussian)ÿlink(identity)ÿcorr(independent)ÿvce(robust)ÿnologÿ*/
.ÿ
.ÿexit

endÿofÿdo-file


.