********************************************************
*    MAI ANH TUAN PhD5 ID:567110018-8                            * 
* Achievement of Treatment Goals (ABC) during 3 years         *
********************************************************

*-------------------------------------------------*
clear
set more off
cd "D:\My Documents\Google Drive\AAAPHD5\18 jun\EDA - DMHT_2553_54_55\"
use "10_DMHT2553_54_55_Data_Final_20130522.dta", clear


*====================== Step 1: Prepare outcome ABC (ADA taget)===*
*= a1c <7%  
*= ldl < 110  
*== sbp Systolic <130 mm Hg 
*== dbp Diastolic <85 mm Hg
*= combine bp_gr : 1. sbp <130 or dbp<85; 2. both (<130/85)
*= combine ABC: 
   * HbA1c(<7%) and LDL-C (<100 mg/dL)
   * HbA1c(<7%) and bp(<130/85 mm Hg)
   * HbA1c(<7%) and bp(<130/85 mm Hg) and dm_ldl (<100 mg/dL)
*======================================
drop if dm_ht==2                    /*95035 observations deleted*/
drop ht_*                           /* drop all var of Hypertension*/
recode dm_hba1c (-9 = .)            /* ? range ?*/
recode dm_hba1c (-73 = .)
recode dm_ldl (-9 = .)
recode dm_ldl (1000/2000 = .)
recode sbp (-9 = .)
recode dbp (-9 = .)

save "D:\My Documents\Google Drive\AAAPHD5\18 jun\EDA - DMHT_2553_54_55\ABC3year.dta",replace

* GEN ABC: 4 step 
* 1. a1cgr : gen a1c 2 group
gen a1cgr=. /* 1=<7  0=>=7  */
replace a1cgr=1 if dm_hba1c <7 & dm_hba1c !=.
replace a1cgr=0 if dm_hba1c >=7 & dm_hba1c <.
lab var a1cgr "a1c 2 group"
tab a1cgr, mis

* 2. ldlgr : gen ldl 2 group
gen ldlgr=. /* 1=<110 0=>=110 */
replace ldlgr=1 if dm_ldl <110 & dm_ldl !=.
replace ldlgr=0 if dm_ldl >=110 & dm_ldl <.
lab var ldlgr "ldl 2 group"
tab ldl, mis

* 3. bp2gr: gen BP 2 group
gen testbp=sbp-dbp  /*clear data*/
gen bpcl=.
replace bpcl=1 if testbp>20 & testbp <.
replace bpcl=2 if testbp<=20 & testbp <.
tab bpcl
sum sbp dbp
replace sbp =. if bpcl==2
replace dbp =. if bpcl==2
sum sbp dbp

count if sbp !=. & dbp==. /* dbp or sbp missing on 1 record*/
count if sbp ==. & dbp!=.
replace sbp =. if sbp!=. & dbp==.
replace dbp =. if sbp==. & dbp!=.

gen bp3gr=. /* 2. sbp<130&dbp<85;   1. >+130<85 or <130>=85;  0. else*/
replace bp3gr=2 if sbp<130 & dbp<85
replace bp3gr=0 if sbp>=130 & dbp>=85 & dbp<.
replace bp3gr=1 if sbp<130 & dbp>=85 & dbp<.
replace bp3gr=1 if sbp>=130 & dbp<85 & dbp<.
lab var bp3gr "bpgr 3 group"

gen bp2gr=. /* 1. sbp<130&dbp<85;   0. other */
replace bp2gr=0 if bp3gr<2
replace bp2gr=1 if bp3gr==2 /* */
lab var bp2gr "bpgr 2 group"
tab1 bp3gr bp2gr, mis

/* 4. gen ABC 
a1cgr(<7%) and ldlgr (<100 mg/dL)
a1cgr(<7%) and bp2gr (<130/85 mm Hg)
a1cgr(<7%) and bp2gr (<130/85 mm Hg) and ldl (<100 mg/dL)*/

tab1 a1cgr ldlgr bp2gr, mis /*test*/
gen totalabc=a1cgr+bp2gr+ldlgr
tab totalabc, mis /**/
 
gen abc6gr=.
replace abc6gr=5 if a1cgr==1&ldlgr==1&bp2gr==1&totalabc==3/*both3*/
replace abc6gr=4 if a1cgr==1&bp2gr==1&totalabc==2 /*a1c+bp*/
replace abc6gr=3 if a1cgr==1&ldlgr==1&totalabc==2 /*a1c+ldl*/
replace abc6gr=2 if ldlgr==1&bp2gr==1&totalabc==2
replace abc6gr=1 if (totalabc==1&a1cgr==1)|(totalabc==1&ldlgr==1)|(bp2gr==1&totalabc==1)
replace abc6gr=0 if a1cgr==0&ldlgr==0&bp2gr==0
tab abc6gr totalabc, mis


*====================== Prepare covariate ===*
* sex, age, dm_ht, treat (2 year), dmdura(2 year)
 
 
destring dmdate, gen(dmdate1)
recode dmdate1 (-9 =.)
recode dmdate1 (2556/2566=.)
gen dmdura=year-dmdate1     /*the duration time get dm */
sum dmdura,d                  /*non-standard distribution*/

  

*======================================= tables 1==================*


 







*======================================= tables 2==================*
*======================================= tables 3==================*










































drop abc6gr

replace abc6gr=1 if (abc==.&a1cgr==1)|(abc==.&ldlgr==1)|(abc==.&bp2gr==1)

qnorm dm_ldl
pnorm dm_ldl
sktest dm_ldl //
sum dm_ldl dm_hba1c, d
hist dm_ldl, norm
help sum

*================================


*======================== Factor of interest ==> dm_ht
tab dm_ht, nolab mis
drop if dm_ht==2     //(95035 observations deleted)
drop ht_*            // drop all var of Hypertension

*Covariate ==> sex, age, dm_dura



label variable type "financy"
































































































//1. Test distribution, variance 
sum dm_hdl,d
sum dm_hba1c,d

graph box dm_hba1c, over(year) over(sex)

*** A1c sex#year

oneway dm_hba1c year,t

table sex year, cont(freq mean age) mis

anova dm_hba1c year sex sex#year

anovaplot, scatter(ms(i))

graph box dm_hba1c, over(year) over(sex)


*** A1c dm_ht#year
*graph box dm_hba1c, over(year) over(dm_ht)

table dm_ht year, cont(mean dm_hba1c)
anova dm_hba1c year dm_ht dm_ht#year
regress
anovaplot, scatter(ms(i))


table dm_ht year, freq mean dm_hba1c)
???anova dm_hba1c year sex sex#year, sequential





*===================================
// exxample anova//

use http://www.ats.ucla.edu/stat/stata/examples/kirk/crf33, clear
tabdisp order b a, cellvar(y)
table a b, cont(sum y)
table a b, cont(mean y)
anova y b a b#a
anovaplot scatter(msymbol(i))
quietly anova y a b a#b
anovaplot, scatter(msymbol(i))
sme a b
sme b a