** -----------------------**; ** this program fits one pooled logistic model to an indicator of an outcome of interest and a second pooled logistic model to an indicator of all competing events and then takes the coefficients of the resultant models to calculate Hazards of both events that are employed in an Excel spreadsheet to calculate Absolute Risk per the desired characteristics ** -----------------------**; libname temP '\\Dmic06\poa\Biostatistics_Core\AbsRisk\SAS'; libname temC 'c://godlove/AbsRisk'; data state0model; set temP.state0model; ADLdis = 0; DeathMonth = 0; if CurrState = . then ADLdis = .; if CurrState = . then DeathMonth = .; if 1<= CurrState <= 2 then ADLdis = 1; if CurrState = 3 then DeathMonth = 1; run; data state0model; set state0model (keep = studyid ; ******** results needed to calc. Absolute Risk for occurrence of any disability; *** now use pooled logistic model, i.e., as approximation of simple exponential over entire 10 year period, for parameters for ADLdis; ** model of ADLdis without interaction; proc logistic data=state0model descending; model ADLdis = /* pd1 pd2 pd3 pd4 pd5 */ /* yr1 yr2 yr3 yr4 yr5 yr6 yr7 yr8 yr9 yr10 */ SlowGait Hmod RAnonH female gtHS nonwhite ageGTE85 cesdGTE20 gt2cc LiveAlone MMSE24 /* NumCCgt1 edu_cont LiveAlone */ /*SGbyHmod SGbyRAnonH*/ / lackfit; ods output parameterEstimates=outbeta; run; ** model of ADLdis with interactions; proc logistic data=state0model descending; model ADLdis = /* pd1 pd2 pd3 pd4 pd5 */ /* yr1 yr2 yr3 yr4 yr5 yr6 yr7 yr8 yr9 yr10 */ SlowGait Hmod RAnonH female gtHS nonwhite ageGTE85 cesdGTE20 gt2cc LiveAlone MMSE24 /* NumCCgt1 edu_cont LiveAlone */ SGbyHmod SGbyRAnonH / lackfit; ods output parameterEstimates=outbeta; run; ** model with interactions; proc logistic data=state0model descending; model DeathMonth= /* pd1 pd2 pd3 pd4 pd5 */ /* yr1 yr2 yr3 yr4 yr5 yr6 yr7 yr8 yr9 yr10 */ SlowGait Hmod RAnonH female gtHS nonwhite ageGTE85 cesdGTE20 gt2cc LiveAlone MMSE24 /* NumCCgt1 edu_cont LiveAlone */ SGbyHmod SGbyRAnonH / lackfit; ods output parameterEstimates=outbeta; run; ************************************** ************************************** **************************************