I am not able to see the estimate of theta or the se of theta using e-returned results. It appears as if they are not being estimated. If the estimate of theta is not being returned, does that mean that the coefficients are not accounting for the shared option? Is there somewhere else I need to look? Noisily option with the MI estimate returns output similar to the MI estimate without noisily. Is there another command I need to appropriately use the shared option with MI?
When I run the complete case model without MI I see the estimated theta and SE of theta in the output, as well as the likelihood ratio test of theta =0. I am using STATA 15.1, updated in February 2019. Very grateful for your assistance.
Code:
mi set flong mi register imputed lastcd4cat ill_at_enrollment //vars not missing values mi register regular Male ageatinterview_cat_2 facility_cat /// b110_disclosed_hivstatus_r visittointv_weeks1 visittointv_weeks2 hourstoclinic_cat i_clinic_contact_for_return_c i_b19a_month_away /// i_i_cat_anticipated i_b150_confronted_r i_b15h_any_structural i_b15h_any_psychosocial i_b15h_any_clinic i_b15h_any_medical /// i_b15i_any_structural i_b15i_any_psychosocial i_b15i_any_clinic i_b15i_returning b04_clinic // vars not missing values mi describe capture mi stset, clear *impute mi impute chained (ologit) lastcd4cat ill_at_enrollment = Male ageatinterview_cat_2 facility_cat /// b110_disclosed_hivstatus_r visittointv_weeks1 visittointv_weeks2 hourstoclinic_cat i_clinic_contact_for_return_c i_b19a_month_away /// i_i_cat_anticipated i_b150_confronted_r i_b15h_any_structural i_b15h_any_psychosocial i_b15h_any_clinic i_b15h_any_medical /// i_b15i_any_structural i_b15i_any_psychosocial i_b15i_any_clinic i_b15i_returning b04_clinic, add(10) rseed(11111) chaindots dots augment mi stset days_return_or_censor, failure(returner_SC==1) mi estimate, hr: stcox Male i.ageatinterview_cat_2 i.lastcd4cat ill_at_enrollment i.facility_cat /// b110_disclosed_hivstatus_r visittointv_weeks1 visittointv_weeks2 i.hourstoclinic_cat i.i_clinic_contact_for_return_c i_b19a_month_away /// i.i_i_cat_anticipated i.i_b150_confronted_r i_b15h_any_structural i_b15h_any_psychosocial i_b15h_any_clinic i_b15h_any_medical /// i_b15i_any_structural i_b15i_any_psychosocial i_b15i_any_clinic i_b15i_returning, shared(b04_clinic) ereturn list
Multiple-imputation estimates | Imputations | = 10 | |
Cox regression: Breslow method for ties | Number of obs | = 571 | |
Average RVI | = 0.0199 | ||
Largest FMI | = 0.1820 | ||
DF adjustment: Large sample | DF: min | = 289.33 | |
avg | = 5401508.53 | ||
max | = 5.42e+07 | ||
Model F test: Equal FMI | F( 30,688602.0) | = 1.56 | |
Within VCE type: OIM | Prob > F | = 0.0262 | |
_t Haz. Ratio | Std. Err. t | P>t [95% Conf. | Interval] |
Male .8055849 | .1321687 -1.32 | 0.188 .5840559 | 1.111139 |
ageatinterview_cat_2 | |||
1=25-34 .9374358 | .2519998 -0.24 | 0.810 .5535065 | 1.58767 |
2=35-44 1.106355 | .301895 0.37 | 0.711 .648073 | 1.888709 |
3=45+ 1.407724 | .43825 1.10 | 0.272 .7647535 | 2.591276 |
lastcd4cat | |||
101-350 1.321381 | .5567874 0.66 | 0.509 .5772341 | 3.024852 |
351-500 1.077576 | .4716436 0.17 | 0.865 .4563089 | 2.544701 |
>500 1.047709 | .4714006 0.10 | 0.918 .4321626 | 2.540005 |
ill_at_enrollment 1.021365 | .1925792 0.11 | 0.911 .7055878 | 1.478465 |
facility_cat | |||
urban .7115683 | .1610734 -1.50 | 0.133 .4565993 | 1.108914 |
hospital .5191499 | .1495675 -2.28 | 0.023 .2951623 | .9131133 |
b110_disclosed_hivstatus_r 1.111266 | .2597305 0.45 | 0.652 .7028632 | 1.756975 |
visittointv_weeks1 .9891243 | .0070703 -1.53 | 0.126 .9753633 | 1.003079 |
visittointv_weeks2 .9963817 | .0038809 -0.93 | 0.352 .9888043 | 1.004017 |
hourstoclinic_cat | |||
1 - | .1941369 0.29 | 0.773 .7352222 | 1.512843 |
2 hours + .6917431 | .1326421 -1.92 | 0.055 .4750351 | 1.007312 |
i_clinic_contact_for_return_c | |||
Contacted 1-3 times .8985222 | .2178359 -0.44 | 0.659 .5586827 | 1.445082 |
Contacted 4-12 times 3.164457 | 1.386466 2.63 | 0.009 1.340776 | 7.468653 |
i_b19a_month_away .9333822 | .1457113 -0.44 | 0.659 .6873491 | 1.267482 |
i_i_cat_anticipated | |||
1=Medium 1.046974 | .2084225 0.23 | 0.818 .7087385 | 1.546628 |
2=High 1.296827 | .2390717 1.41 | 0.159 .9035699 | 1.861239 |
i_b150_confronted_r | |||
1=Once 1.600127 | .4375052 1.72 | 0.086 .9363106 | 2.734571 |
2=More than once .5441514 | .1763496 -1.88 | 0.060 .2883107 | 1.02702 |
i_b15h_any_structural 1.05646 | .2009187 0.29 | 0.773 .7277308 | 1.533683 |
i_b15h_any_psychosocial .9372346 | .1603545 -0.38 | 0.705 .670213 | 1.310641 |
i_b15h_any_clinic 1.084263 | .1910634 0.46 | 0.646 .767609 | 1.531543 |
i_b15h_any_medical 1.14473 | .2058022 0.75 | 0.452 .8047707 | 1.628299 |
i_b15i_any_structural 1.1655 | .2489362 0.72 | 0.473 .7668401 | 1.771412 |
i_b15i_any_psychosocial .7464311 | .1507536 -1.45 | 0.148 .5024305 | 1.108928 |
i_b15i_any_clinic .9912523 | .1788625 -0.05 | 0.961 .6959732 | 1.411809 |
i_b15i_returning .9223598 | .165134 -0.45 | 0.652 .6493906 | 1.310071 |
. | |||
. ereturn list | |||
scalars: | |||
e(df_m) = . | |||
e(se_theta) = . | |||
e(chi2_c) = . | |||
e(ll) = . | |||
e(g_min) = 5 | |||
e(risk) = 271374 | |||
e(converged) = 1 | |||
e(theta) = . | |||
e(N) = 571 |
Below is my code and output for a complete case analysis without using MI. Here, theta is estimated.
Code:
stcox Male i.ageatinterview_cat_2 i.lastcd4cat ill_at_enrollment i.facility_cat /// b110_disclosed_hivstatus_r visittointv_weeks1 visittointv_weeks2 i.hourstoclinic_cat i.i_clinic_contact_for_return_c i_b19a_month_away /// i.i_i_cat_anticipated i.i_b150_confronted_r i_b15h_any_structural i_b15h_any_psychosocial i_b15h_any_clinic i_b15h_any_medical /// i_b15i_any_structural i_b15i_any_psychosocial i_b15i_any_clinic i_b15i_returning, shared(b04_clinic)
Cox regression -- Breslow method for ties | ||
Gamma shared frailty | Number of obs = 448 | |
Group variable: b04_clinic | Number of groups = 31 | |
Obs per group: | ||
No. of subjects = 448 | min = 1 | |
No. of failures = 170 | avg = 14.451613 | |
Time at risk = 208878 | max = 41 | |
Wald chi2(30) = 37.84 | ||
Log likelihood = -975.31468 | Prob > chi2 = 0.1538 | |
_t Haz. Ratio | Std. Err. z P>z [95% Conf. | Interval] |
Male .8297834 | .1514522 -1.02 0.307 .5802341 | 1.18666 |
ageatinterview_cat_2 | ||
1=25-34 1.031318 | .3191602 0.10 0.921 .5623064 | 1.891527 |
2=35-44 1.158861 | .3619261 0.47 0.637 .6283344 | 2.137331 |
3=45+ 1.451009 | .5067411 1.07 0.286 .7318156 | 2.876989 |
lastcd4cat | ||
101-350 1.379417 | .5713796 0.78 0.437 .6125086 | 3.106553 |
351-500 1.106223 | .4902374 0.23 0.820 .4641106 | 2.636719 |
>500 1.055018 | .4620821 0.12 0.903 .4471426 | 2.489278 |
ill_at_enrollment 1.023086 | .2004435 0.12 0.907 .6968583 | 1.502035 |
facility_cat | ||
urban .747191 | .1631256 -1.33 0.182 .4870782 | 1.146211 |
hospital .4978426 | .1408213 -2.47 0.014 .2859687 | .8666937 |
b110_disclosed_hivstatus_r 1.262853 | .3447006 0.85 0.393 .7396289 | 2.156212 |
visittointv_weeks1 .9937592 | .007794 -0.80 0.425 .9786001 | 1.009153 |
visittointv_weeks2 .9953744 | .004216 -1.09 0.274 .9871455 | 1.003672 |
hourstoclinic_cat | ||
1 - | .2195736 0.53 0.598 .7531965 | 1.635638 |
2 hours + .7992391 | .1676841 -1.07 0.285 .529774 | 1.205765 |
i_clinic_contact_for_return_c | ||
Contacted 1-3 times .9347687 | .2452422 -0.26 0.797 .5589672 | 1.563227 |
Contacted 4-12 times 2.775985 | 1.213377 2.34 0.019 1.178579 | 6.538459 |
i_b19a_month_away .9791736 | .1695891 -0.12 0.903 .6973243 | 1.374943 |
i_i_cat_anticipated | ||
1=Medium 1.322599 | .2915854 1.27 0.205 .8585561 | 2.037452 |
2=High 1.637422 | .3359041 2.40 0.016 1.095326 | 2.447811 |
i_b150_confronted_r | ||
1=Once 1.416865 | .4172812 1.18 0.237 .7955014 | 2.523574 |
2=More than once .5371145 | .1738965 -1.92 0.055 .2847615 | 1.0131 |
i_b15h_any_structural 1.026157 | .210909 0.13 0.900 .6859051 | 1.535196 |
i_b15h_any_psychosocial .8947523 | .1669077 -0.60 0.551 .6207535 | 1.289694 |
i_b15h_any_clinic 1.106091 | .2097494 0.53 0.595 .7627403 | 1.604004 |
i_b15h_any_medical 1.020475 | .2054367 0.10 0.920 .68777 | 1.514126 |
i_b15i_any_structural 1.204139 | .2838235 0.79 0.431 .7586525 | 1.911219 |
i_b15i_any_psychosocial .9037175 | .1936872 -0.47 0.637 .5937463 | 1.375512 |
i_b15i_any_clinic 1.015074 | .1993901 0.08 0.939 .6907118 | 1.491758 |
i_b15i_returning 1.016642 | .194275 0.09 0.931 .699049 | 1.478523 |
theta .0364613 | .0755885 | |
LR test of theta=0: chibar2(01) = 0.27 | Prob >= chibar2 = 0.300 |
0 Response to estimate of theta = . in Cox shared frailty model using MI
Post a Comment