Sorry about the attachment, here is the document.
Trying to get adjusted percentages from logistic model. 1st way-all
estimates are exactly the same.
xi: svylogit restcever i.insgrp2 i.race i.ppgender i.income3 i.edugp2
i.agecat2 i.newmeds6mos i.numdisgp, or
****Percent va ins who restricts
gen temprest=insgrp2
recode insgrp2 2=1 3=1 4=1 9=1
predict restricted
summarize restricted restcever
drop insgrp2
gen insgrp2=temprest
drop temprest restricted
gen temprest=insgrp2
recode insgrp2 1=2 3=2 4=2 9=2
predict restricted
summarize restricted restcever
drop insgrp2
gen insgrp2=temprest
drop temprest restricted
gen temprest=insgrp2
recode insgrp2 1=3 2=3 4=3 9=3
predict restricted
summarize restricted restcever
drop insgrp2
gen insgrp2=temprest
drop temprest restricted
gen temprest=insgrp2
recode insgrp2 1=4 2=4 3=4 9=4
predict restricted
summarize restricted restcever
drop insgrp2
gen insgrp2=temprest
drop temprest restricted
gen temprest=insgrp2
recode insgrp2 1=9 2=9 3=9 4=9
predict restricted
summarize restricted restcever
drop insgrp2
gen insgrp2=temprest
drop temprest restricted
*********results***************
xi: svylogit restcever i.insgrp2 i.race i.ppgender i.income3 i.edugp2
i.agecat2 i.newmeds6mos i.numdi
> sgp, or
i.insgrp2 _Iinsgrp2_1-9 (naturally coded; _Iinsgrp2_1 omitted)
i.race _Irace_1-2 (naturally coded; _Irace_1 omitted)
i.ppgender _Ippgender_1-2 (naturally coded; _Ippgender_1
omitted)
i.income3 _Iincome3_1-4 (naturally coded; _Iincome3_1 omitted)
i.edugp2 _Iedugp2_1-2 (naturally coded; _Iedugp2_1 omitted)
i.agecat2 _Iagecat2_1-3 (naturally coded; _Iagecat2_1 omitted)
i.newmeds6mos _Inewmeds6m_1-3 (naturally coded; _Inewmeds6m_1
omitted)
i.numdisgp _Inumdisgp_1-3 (naturally coded; _Inumdisgp_1
omitted)
Survey logistic regression
pweight: weight2 Number of obs =
4055
Strata: <one> Number of strata =
1
PSU: <observations> Number of PSUs =
4055
Population size =
4050.3726
F( 16, 4039) =
12.41
Prob > F =
0.0000
----------------------------------------------------------------------------
--
restcever | Odds Ratio Std. Err. t P>|t| [95% Conf.
Interval]
-------------+--------------------------------------------------------------
--
_Iinsgrp2_2 | 1.707165 .450561 2.03 0.043 1.017552
2.864143
_Iinsgrp2_3 | 1.473491 .5497295 1.04 0.299 .7090658
3.062022
_Iinsgrp2_4 | 2.591608 .7683887 3.21 0.001 1.449173
4.634665
_Iinsgrp2_9 | 3.232883 .9952096 3.81 0.000 1.767984
5.911557
_Irace_2 | 1.207308 .2069171 1.10 0.272 .8627592
1.689455
_Ippgender_2 | 1.493871 .1999075 3.00 0.003 1.149139
1.94202
_Iincome3_2 | 1.67865 .3054424 2.85 0.004 1.174983
2.398218
_Iincome3_3 | 2.320533 .4195796 4.66 0.000 1.627926
3.307813
_Iincome3_4 | 3.464592 .708747 6.07 0.000 2.319909
5.174082
_Iedugp2_2 | .8891192 .1072834 -0.97 0.330 .7018126
1.126416
_Iagecat2_2 | .7224912 .1207853 -1.94 0.052 .5205795
1.002716
_Iagecat2_3 | .3521471 .0634984 -5.79 0.000 .2472814
.5014836
_Inewmeds6~2 | 1.504512 .2657004 2.31 0.021 1.064208
2.126987
_Inewmeds6~3 | 1.560456 .3182482 2.18 0.029 1.046166
2.327567
_Inumdisgp_2 | 1.700446 .2687467 3.36 0.001 1.247368
2.318094
_Inumdisgp_3 | 2.647208 .4701574 5.48 0.000 1.868811
3.749824
----------------------------------------------------------------------------
--
.
end of do-file
. do "C:\DOCUME~1\VHAANN~2\LOCALS~1\Temp\STD03000000.tmp"
. gen temprest=insgrp2
. recode insgrp2 2=1 3=1 4=1 9=1
(insgrp2: 3727 changes made)
. predict restricted
(option p assumed; Pr(restcever))
. summarize restricted restcever
Variable | Obs Mean Std. Dev. Min Max
-------------+--------------------------------------------------------
restricted | 4061 .1682303 .1245745 .0125673 .7723164
restcever | 4061 .1652302 .371434 0 1
. drop insgrp2
. gen insgrp2=temprest
. drop temprest restricted
.
end of do-file
. do "C:\DOCUME~1\VHAANN~2\LOCALS~1\Temp\STD03000000.tmp"
. gen temprest=insgrp2
. recode insgrp2 1=2 3=2 4=2 9=2
(insgrp2: 1318 changes made)
. predict restricted
(option p assumed; Pr(restcever))
. summarize restricted restcever
Variable | Obs Mean Std. Dev. Min Max
-------------+--------------------------------------------------------
restricted | 4061 .1682303 .1245745 .0125673 .7723164
restcever | 4061 .1652302 .371434 0 1
. drop insgrp2
. gen insgrp2=temprest
. drop temprest restricted
.
end of do-file
. do "C:\DOCUME~1\VHAANN~2\LOCALS~1\Temp\STD03000000.tmp"
. gen temprest=insgrp2
. recode insgrp2 1=3 2=3 4=3 9=3
(insgrp2: 3870 changes made)
. predict restricted
(option p assumed; Pr(restcever))
. summarize restricted restcever
Variable | Obs Mean Std. Dev. Min Max
-------------+--------------------------------------------------------
restricted | 4061 .1682303 .1245745 .0125673 .7723164
restcever | 4061 .1652302 .371434 0 1
. drop insgrp2
. gen insgrp2=temprest
. drop temprest restricted
.
end of do-file
. do "C:\DOCUME~1\VHAANN~2\LOCALS~1\Temp\STD03000000.tmp"
. gen temprest=insgrp2
. recode insgrp2 1=4 2=4 3=4 9=4
(insgrp2: 3550 changes made)
. predict restricted
(option p assumed; Pr(restcever))
. summarize restricted restcever
Variable | Obs Mean Std. Dev. Min Max
-------------+--------------------------------------------------------
restricted | 4061 .1682303 .1245745 .0125673 .7723164
restcever | 4061 .1652302 .371434 0 1
. drop insgrp2
. gen insgrp2=temprest
. drop temprest restricted
.
.
. gen temprest=insgrp2
. recode insgrp2 1=9 2=9 3=9 4=9
(insgrp2: 3779 changes made)
. predict restricted
(option p assumed; Pr(restcever))
. summarize restricted restcever
Variable | Obs Mean Std. Dev. Min Max
-------------+--------------------------------------------------------
restricted | 4061 .1682303 .1245745 .0125673 .7723164
restcever | 4061 .1652302 .371434 0 1
. drop insgrp2
. gen insgrp2=temprest
. drop temprest restricted
.
end of do-file
.
2nd way. Manually create all of the dummies.
svylogit restcever insgrp2b insgrp2c insgrp2d insgrp2e $demog, or
****Percent va ins who restricts
gen temprest=insgrp2a
recode insgrp2a 0=1
predict restricted
summarize restricted restcever
drop insgrp2a
gen insgrp2a=temprest
drop temprest restricted
****Percent private ins who restricts
gen temprest=insgrp2b
recode insgrp2b 0=1
predict restricted
summarize restricted restcever
drop insgrp2b
gen insgrp2b=temprest
drop temprest restricted
****Percent medicaid ins who is worried who restricts
gen temprest=insgrp2c
recode insgrp2c 0=1
predict restricted
summarize restricted restcever
drop insgrp2c
gen insgrp2c=temprest
drop temprest restricted
****Percent medicare ins who restricts
gen temprest=insgrp2d
recode insgrp2d 0=1
predict restricted
summarize restricted restcever
drop insgrp2d
gen insgrp2d=temprest
drop temprest restricted
****Percent no ins who restricts
gen temprest=insgrp2e
recode insgrp2e 0=1
predict restricted
summarize restricted restcever
drop insgrp2e
gen insgrp2e=temprest
drop temprest restricted
**********************results*******************
. svylogit restcever insgrp2b insgrp2c insgrp2d insgrp2e $demog, or
Survey logistic regression
pweight: weight2 Number of obs =
4055
Strata: <one> Number of strata =
1
PSU: <observations> Number of PSUs =
4055
Population size =
4050.3726
F( 16, 4039) =
12.41
Prob > F =
0.0000
----------------------------------------------------------------------------
--
restcever | Odds Ratio Std. Err. t P>|t| [95% Conf.
Interval]
-------------+--------------------------------------------------------------
--
insgrp2b | 1.707165 .450561 2.03 0.043 1.017552
2.864143
insgrp2c | 1.473491 .5497295 1.04 0.299 .7090658
3.062022
insgrp2d | 2.591608 .7683887 3.21 0.001 1.449173
4.634665
insgrp2e | 3.232883 .9952096 3.81 0.000 1.767984
5.911557
racea | 1.207308 .2069171 1.10 0.272 .8627592
1.689455
gendera | 1.493871 .1999075 3.00 0.003 1.149139
1.94202
incomea | 1.67865 .3054424 2.85 0.004 1.174983
2.398218
incomeb | 2.320533 .4195796 4.66 0.000 1.627926
3.307813
incomec | 3.464592 .708747 6.07 0.000 2.319909
5.174082
agecat2a | .7224912 .1207853 -1.94 0.052 .5205795
1.002716
agecat2b | .3521471 .0634984 -5.79 0.000 .2472814
.5014836
edua | .8891192 .1072834 -0.97 0.330 .7018126
1.126416
newmeds6mosa | 1.504512 .2657004 2.31 0.021 1.064208
2.126987
newmeds6mosb | 1.560456 .3182482 2.18 0.029 1.046166
2.327567
numdisgpa | 1.700446 .2687467 3.36 0.001 1.247368
2.318094
numdisgpb | 2.647208 .4701574 5.48 0.000 1.868811
3.749824
----------------------------------------------------------------------------
--
.
. ****Percent va ins who restricts
.
. gen temprest=insgrp2a
. recode insgrp2a 0=1
(insgrp2a: 3727 changes made)
. predict restricted
(option p assumed; Pr(restcever))
. summarize restricted restcever
Variable | Obs Mean Std. Dev. Min Max
-------------+--------------------------------------------------------
restricted | 4061 .1682303 .1245745 .0125673 .7723164
restcever | 4061 .1652302 .371434 0 1
. drop insgrp2a
. gen insgrp2a=temprest
. drop temprest restricted
.
. ****Percent private ins who restricts
.
. gen temprest=insgrp2b
. recode insgrp2b 0=1
(insgrp2b: 1318 changes made)
. predict restricted
(option p assumed; Pr(restcever))
. summarize restricted restcever
Variable | Obs Mean Std. Dev. Min Max
-------------+--------------------------------------------------------
restricted | 4061 .1961565 .1529727 .0212655 .8527421
restcever | 4061 .1652302 .371434 0 1
. drop insgrp2b
. gen insgrp2b=temprest
. drop temprest restricted
.
.
. ****Percent medicaid ins who is worried who restricts
.
. gen temprest=insgrp2c
. recode insgrp2c 0=1
(insgrp2c: 3870 changes made)
. predict restricted
(option p assumed; Pr(restcever))
. summarize restricted restcever
Variable | Obs Mean Std. Dev. Min Max
-------------+--------------------------------------------------------
restricted | 4061 .2181586 .1438504 .0184083 .8332825
restcever | 4061 .1652302 .371434 0 1
. drop insgrp2c
. gen insgrp2c=temprest
. drop temprest restricted
.
.
. ****Percent medicare ins who restricts
.
. gen temprest=insgrp2d
. recode insgrp2d 0=1
(insgrp2d: 3550 changes made)
. predict restricted
(option p assumed; Pr(restcever))
. summarize restricted restcever
Variable | Obs Mean Std. Dev. Min Max
-------------+--------------------------------------------------------
restricted | 4061 .2958942 .1716523 .0319308 .8978642
restcever | 4061 .1652302 .371434 0 1
. drop insgrp2d
. gen insgrp2d=temprest
. drop temprest restricted
.
.
. ****Percent no ins who restricts
.
. gen temprest=insgrp2e
. recode insgrp2e 0=1
(insgrp2e: 3779 changes made)
. predict restricted
(option p assumed; Pr(restcever))
. summarize restricted restcever
Variable | Obs Mean Std. Dev. Min Max
-------------+--------------------------------------------------------
restricted | 4061 .3422888 .1727728 .0395197 .8978642
restcever | 4061 .1652302 .371434 0 1
. drop insgrp2e
. gen insgrp2e=temprest
. drop temprest restricted
.
.
end of do-file
.
3rd way: use adjust
xi: svylogit restcever i.insgrp2 i.race i.ppgender i.income3 i.edugp2
i.agecat2 i.newmeds6mos i.numdisgp, or
adjust, by(insgrp2) pr
*******************results**********************
. adjust, by(insgrp2) pr
----------------------------------------------------------------------------
---------------------------
Dependent variable: restcever Command: svylogit
Variables left as is: _Iinsgrp2_2, _Iinsgrp2_3, _Iinsgrp2_4, _Iinsgrp2_9,
_Irace_2, _Ippgender_2,
_Iincome3_2, _Iincome3_3, _Iincome3_4, _Iedugp2_2,
_Iagecat2_2, _Iagecat2_3,
_Inewmeds6~2, _Inewmeds6~3, _Inumdisgp_2,
_Inumdisgp_3
----------------------------------------------------------------------------
---------------------------
----------------------
insgrp2 | pr
----------+-----------
1 | .08082
2 | .12135
3 | .240155
4 | .176516
9 | .306773
----------------------
Key: pr = Probability
-----Original Message-----
From: Eby, Elizabeth
Sent: Thursday, September 25, 2003 10:58 AM
To: '[email protected]'; Copeland, Laurel
Subject: RE: adjusted percentages and adjusted odds ratios
Hi again, I have attached some more information regarding the methods I am
using to get the adjusted percentages. A couple of questions:
In method one, I use the xi.svylogit method and all of the estimates are the
same. Can anyone explain why? So, in method two, I manually created the
dummies and got different estimates on all levels. But, then in method
three, I simply ran an adjust statement after the model, and I got entirely
different estimates than the predict command in method 2. Could someone
help me understand why these two methods give different results and which is
the better method?
Note about the data. It is a nationally representative sample of adults
with at least one chronic illness. We are trying to say that when adjusted
by SES and other clinical covariates, less VA patients restrict their
medications than do people with the other types of insurance.
Thank you for your help,
Elizabeth Eby
Ann Arbor VAMC
-----Original Message-----
From: Eby, Elizabeth
Sent: Friday, September 19, 2003 11:47 AM
To: '[email protected]'
Subject: adjusted percentages and adjusted odds ratios
Hi-
Could someone help me out with understanding the relatively small difference
in adjusted percentages (20% vs 17%) compared to a more robust adjusted odds
ration of 1.7? The adjusted percentages are controlling for covariates, and
were obtained using the predict sequence.
Thank you,
Elizabeth Eby
Research Health Science Specialist
Ann Arbor VAMC
734-769-7100 X6248
-----Original Message-----
From: Copeland, Laurel
Sent: Friday, September 19, 2003 11:28 AM
To: Eby, Elizabeth
Subject: RE: adjusted percentages
I don't know. Maybe put the question to Statalist?
-----Original Message-----
From: Eby, Elizabeth
Sent: Friday, September 19, 2003 11:27 AM
To: Copeland, Laurel
Subject: RE: adjusted percentages
This hold even though those are adjusted percentages?
-----Original Message-----
From: Copeland, Laurel
Sent: Friday, September 19, 2003 11:26 AM
To: Eby, Elizabeth
Cc: VHAANN HSRD Statistical Group
Subject: RE: adjusted percentages
The raw odds ratio can be calculated...17% of 334=277, 20% of 2741=2193
cut back?
no yes
priv no 277 57
ins? yes 2193 548
OR = (277/57) / (2193/548)
= 4.86 / 4.00
= approx 1.22
The remainder of the effect in the adjusted OR must be due to the
covariates.
-----Original Message-----
From: Eby, Elizabeth
Sent: Friday, September 19, 2003 11:05 AM
To: VHAANN HSRD Statistical Group
Subject: adjusted percentages
How would you explain the small difference in the adjusted percentages (17%
vs 20%) with an odds ratio of 1.7 (p=.04)? Does it have anything to do with
the underlying sample sizes of the 2 groups (n=334 VA patients and n=2741
with private insurance)?
Thanks.
-----Original Message-----
From: [email protected] [mailto:[email protected]]
Sent: Friday, September 19, 2003 10:51 AM
To: [email protected]
Subject:
Liz - can you double check this? THe difference in %'s seem small relative
to
the AOR.
Controlling for their sociodemographic characteristics, number of chronic
diseases, and number of prescription medications, 17% of VA patients cut
back
on medication use due to cost compared to 20 % of those with private
insurance
(adjusted odds ratio [AOR]: 1.7, p=.04),
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
