{smcl}
{com}{sf}{ul off}{txt}{.-}
      name:  {res}<unnamed>
       {txt}log:  {res}/Users/steve/Documents/teaching/c2015/biostat208/labs/lab8/lab8.smcl
  {txt}log type:  {res}smcl
 {txt}opened on:  {res}26 Feb 2015, 10:01:55
{txt}
{com}. * summaries
. des

{txt}Contains data from {res}lab8.dta
{txt}  obs:{res}         3,154                          
{txt} vars:{res}            14                          2 Feb 2000 13:27
{txt} size:{res}       176,624                          
{txt}{hline}
              storage   display    value
variable name   type    format     label      variable label
{hline}
{p 0 48}{res}{bind:id             }{txt}{bind: float   }{bind:{txt}%9.0g     }{space 1}{bind:         }{bind:  }{res}{res}id{p_end}
{p 0 48}{bind:age            }{txt}{bind: float   }{bind:{txt}%9.0g     }{space 1}{bind:         }{bind:  }{res}{res}age{p_end}
{p 0 48}{bind:height         }{txt}{bind: float   }{bind:{txt}%9.0g     }{space 1}{bind:         }{bind:  }{res}{res}height (inches){p_end}
{p 0 48}{bind:weight         }{txt}{bind: float   }{bind:{txt}%9.0g     }{space 1}{bind:         }{bind:  }{res}{res}weight (lbs){p_end}
{p 0 48}{bind:sbp            }{txt}{bind: float   }{bind:{txt}%9.0g     }{space 1}{bind:         }{bind:  }{res}{res}systolic BP{p_end}
{p 0 48}{bind:dbp            }{txt}{bind: float   }{bind:{txt}%9.0g     }{space 1}{bind:         }{bind:  }{res}{res}diastolic BP{p_end}
{p 0 48}{bind:chol           }{txt}{bind: float   }{bind:{txt}%9.0g     }{space 1}{bind:         }{bind:  }{res}{res}total cholesterol{p_end}
{p 0 48}{bind:behpat         }{txt}{bind: float   }{bind:{txt}%9.0g     }{space 1}{bind:bpat4    }{bind:  }{res}{res}behavioral pattern (4 level){p_end}
{p 0 48}{bind:ncigs          }{txt}{bind: float   }{bind:{txt}%9.0g     }{space 1}{bind:         }{bind:  }{res}{res}cigarettes smoked per day{p_end}
{p 0 48}{bind:dibpat         }{txt}{bind: float   }{bind:{txt}%9.0g     }{space 1}{bind:bpat2    }{bind:  }{res}{res}behavioral pattern (A vs. B){p_end}
{p 0 48}{bind:chd69          }{txt}{bind: float   }{bind:{txt}%9.0g     }{space 1}{bind:ind      }{bind:  }{res}{res}CHD event{p_end}
{p 0 48}{bind:typchd69       }{txt}{bind: float   }{bind:{txt}%9.0g     }{space 1}{bind:chdtype  }{bind:  }{res}{res}type of event{p_end}
{p 0 48}{bind:time169        }{txt}{bind: float   }{bind:{txt}%9.0g     }{space 1}{bind:         }{bind:  }{res}{res}follow-up time (days){p_end}
{p 0 48}{bind:arcus          }{txt}{bind: float   }{bind:{txt}%9.0g     }{space 1}{bind:pamiss   }{bind:  }{res}{res}arcus senilis{p_end}
{txt}{hline}
Sorted by:  

{com}. sum

{txt}    Variable {c |}       Obs        Mean    Std. Dev.       Min        Max
{hline 13}{c +}{hline 56}
{space 10}id {c |}{res}      3154    10477.94    5877.376       2001      22101
{txt}{space 9}age {c |}{res}      3154    46.27869    5.524045         39         59
{txt}{space 6}height {c |}{res}      3154    69.77774    2.528693         60         78
{txt}{space 6}weight {c |}{res}      3154    169.9537    21.09576         78        320
{txt}{space 9}sbp {c |}{res}      3154    128.6328    15.11773         98        230
{txt}{hline 13}{c +}{hline 56}
{space 9}dbp {c |}{res}      3154    82.01554     9.72688         58        150
{txt}{space 8}chol {c |}{res}      3142    226.3724    43.42043        103        645
{txt}{space 6}behpat {c |}{res}      3154    2.523145    .7989865          1          4
{txt}{space 7}ncigs {c |}{res}      3154    11.60051    14.51758          0         99
{txt}{space 6}dibpat {c |}{res}      3154    .5038047    .5000648          0          1
{txt}{hline 13}{c +}{hline 56}
{space 7}chd69 {c |}{res}      3154    .0814838    .2736201          0          1
{txt}{space 4}typchd69 {c |}{res}      3154    .1363348    .5097763          0          3
{txt}{space 5}time169 {c |}{res}      3154    2683.859    666.5241         18       3430
{txt}{space 7}arcus {c |}{res}      3152    .2985406    .4576905          0          1
{txt}
{com}. 
. * assoc. between CHD & arcus
. 
. * using cs
. cs chd69 arcus, or

{col 18}{txt}{c |} arcus senilis{col 43}{c |}
{col 18}{c |}   Exposed   Unexposed  {c |}      Total
{hline 17}{c +}{hline 24}{c +}{hline 12}
           Cases {c |} {res}      102         153{txt}  {c |}  {res}      255
        {txt}Noncases {c |} {res}      839        2058{txt}  {c |}  {res}     2897
{txt}{hline 17}{c +}{hline 24}{c +}{hline 12}
{col 12}Total {c |} {res}      941        2211{txt}  {c |}  {res}     3152
{txt}{col 18}{c |}{col 43}{c |}
            Risk {c |} {res} .1083953    .0691995{txt}  {c |}  {res}  .080901
{txt}{col 18}{c |}{col 43}{c |}
{col 18}{c |}      Point estimate    {c |}    [95% Conf. Interval]
{col 18}{c LT}{hline 24}{c +}{hline 24}
 Risk difference {c |} {res}{col 27} .0391959{txt}{col 43}{c |}   {res} .0166915    .0617003{txt} 
      Risk ratio {c |} {res}{col 27} 1.566419{txt}{col 43}{c |}   {res} 1.233865    1.988603{txt} 
 Attr. frac. ex. {c |} {res}{col 27} .3616011{txt}{col 43}{c |}   {res} .1895387    .4971343{txt} 
 Attr. frac. pop {c |} {res}{col 27} .1446404{txt}{col 43}{c |}
      Odds ratio {c |} {res}{col 27}  1.63528{txt}{col 43}{c |}   {res} 1.257732    2.126197{txt} (Cornfield)
{col 18}{c BLC}{hline 24}{c BT}{hline 24}
{col 22}          chi2(1) ={res}    13.64{txt}  Pr>chi2 ={res} 0.0002
{txt}
{com}. 
. * using logistic
. logistic chd69 arcus
{res}
{txt}Logistic regression{col 51}Number of obs{col 67}= {res}      3152
{txt}{col 51}LR chi2({res}1{txt}){col 67}= {res}     12.98
{txt}{col 51}Prob > chi2{col 67}= {res}    0.0003
{txt}Log likelihood = {res}-879.10783{txt}{col 51}Pseudo R2{col 67}= {res}    0.0073

{txt}{hline 13}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 1}       chd69{col 14}{c |} Odds Ratio{col 26}   Std. Err.{col 38}      z{col 46}   P>|z|{col 54}     [95% Con{col 67}f. Interval]
{hline 13}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{space 7}arcus {c |}{col 14}{res}{space 2} 1.635281{col 26}{space 2} .2195036{col 37}{space 1}    3.66{col 46}{space 3}0.000{col 54}{space 4} 1.257001{col 67}{space 3} 2.127399
{txt}{space 7}_cons {c |}{col 14}{res}{space 2}  .074344{col 26}{space 2} .0062298{col 37}{space 1}  -31.02{col 46}{space 3}0.000{col 54}{space 4} .0630839{col 67}{space 3} .0876141
{txt}{hline 13}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. * descriptive summaries of age by outcome status
. graph box age, over(chd69) name(age_box)
{res}{txt}
{com}. twoway (kdensity age if chd69==0, bw(3) area(1) lpattern(solid)) (kdensity age if chd69==1, bw(3) area(1) lpattern(longdash)), ytitle("Density") xtitle("age") legend(order(1 "no CHD" 2 "CHD")) name(age_den)
{res}{txt}
{com}. 
. * categorical version of age
. recode age 36/40=0 41/45=1 46/50=2 51/55=3 56/60=4, gen(agec)
{txt}(3154 differences between age and agec)

{com}. label define agelab 0 "36-40" 1 "41-45" 2 "46-50" 3 "51-55" 4 "56-60" 
{txt}
{com}. label val agec agelab
{txt}
{com}. 
. * CHD by categorical age - frequency tables
. tabulate chd69 agec, col
{txt}
{c TLC}{hline 19}{c TRC}
{c |} Key{col 21}{c |}
{c LT}{hline 19}{c RT}
{c |}{space 5}{it:frequency}{col 21}{c |}
{c |}{space 1}{it:column percentage}{col 21}{c |}
{c BLC}{hline 19}{c BRC}

           {c |}                  RECODE of age (age)
 CHD event {c |}     36-40      41-45      46-50      51-55      56-60 {c |}     Total
{hline 11}{c +}{hline 55}{c +}{hline 10}
        no {c |}{res}       512      1,036        680        463        206 {txt}{c |}{res}     2,897 
           {txt}{c |}{res}     94.29      94.96      90.67      87.69      85.12 {txt}{c |}{res}     91.85 
{txt}{hline 11}{c +}{hline 55}{c +}{hline 10}
       yes {c |}{res}        31         55         70         65         36 {txt}{c |}{res}       257 
           {txt}{c |}{res}      5.71       5.04       9.33      12.31      14.88 {txt}{c |}{res}      8.15 
{txt}{hline 11}{c +}{hline 55}{c +}{hline 10}
     Total {c |}{res}       543      1,091        750        528        242 {txt}{c |}{res}     3,154 
           {txt}{c |}{res}    100.00     100.00     100.00     100.00     100.00 {txt}{c |}{res}    100.00 

{txt}
{com}. tabodds chd69 agec, or

{txt}{hline 13}{c TT}{hline 61}
        agec {c |}  Odds Ratio       chi2       P>chi2     [95% Conf. Interval]
{hline 13}{c +}{hline 61}
{col 8}36-40 {c |}  {res}{col 6}  1.000000{col 31}      .{col 45}    .{col 55}         .{col 66}         .
{txt}{col 8}41-45 {c |}  {res}{col 6}  0.876822{col 31}   0.32{col 45}0.5692{col 55}  0.557454{col 66}  1.379156
{txt}{col 8}46-50 {c |}  {res}{col 6}  1.700190{col 31}   5.74{col 45}0.0166{col 55}  1.095789{col 66}  2.637958
{txt}{col 8}51-55 {c |}  {res}{col 6}  2.318679{col 31}  14.28{col 45}0.0002{col 55}  1.479779{col 66}  3.633160
{txt}{col 8}56-60 {c |}  {res}{col 6}  2.886314{col 31}  18.00{col 45}0.0000{col 55}  1.728069{col 66}  4.820876
{txt}{hline 13}{c BT}{hline 61}
Test of homogeneity (equal odds): chi2({res}4{txt})  = {res}   46.64
                                  {txt}Pr>chi2  = {res}  0.0000

{txt}Score test for trend of odds:     chi2({res}1{txt})  = {res}   40.76
                                  {txt}Pr>chi2  = {res}  0.0000
{txt}
{com}. 
. * CHD by categorical age - logistic model
. 
. logistic chd69 i.agec
{res}
{txt}Logistic regression{col 51}Number of obs{col 67}= {res}      3154
{txt}{col 51}LR chi2({res}4{txt}){col 67}= {res}     44.95
{txt}{col 51}Prob > chi2{col 67}= {res}    0.0000
{txt}Log likelihood = {res}-868.14866{txt}{col 51}Pseudo R2{col 67}= {res}    0.0252

{txt}{hline 13}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 1}       chd69{col 14}{c |} Odds Ratio{col 26}   Std. Err.{col 38}      z{col 46}   P>|z|{col 54}     [95% Con{col 67}f. Interval]
{hline 13}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{space 8}agec {c |}
{space 6}41-45  {c |}{col 14}{res}{space 2} .8768215{col 26}{space 2} .2025406{col 37}{space 1}   -0.57{col 46}{space 3}0.569{col 54}{space 4} .5575563{col 67}{space 3} 1.378903
{txt}{space 6}46-50  {c |}{col 14}{res}{space 2}  1.70019{col 26}{space 2} .3800504{col 37}{space 1}    2.37{col 46}{space 3}0.018{col 54}{space 4} 1.097046{col 67}{space 3} 2.634935
{txt}{space 6}51-55  {c |}{col 14}{res}{space 2} 2.318679{col 26}{space 2} .5274963{col 37}{space 1}    3.70{col 46}{space 3}0.000{col 54}{space 4} 1.484545{col 67}{space 3} 3.621494
{txt}{space 6}56-60  {c |}{col 14}{res}{space 2} 2.886314{col 26}{space 2} .7462298{col 37}{space 1}    4.10{col 46}{space 3}0.000{col 54}{space 4} 1.738895{col 67}{space 3} 4.790864
{txt}{space 12} {c |}
{space 7}_cons {c |}{col 14}{res}{space 2} .0605469{col 26}{space 2} .0111989{col 37}{space 1}  -15.16{col 46}{space 3}0.000{col 54}{space 4} .0421358{col 67}{space 3} .0870026
{txt}{hline 13}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. testparm i.agec

{p 0 7}{space 1}{text:( 1)}{space 1} {res}[chd69]1.agec = 0{p_end}
{p 0 7}{space 1}{text:( 2)}{space 1} [chd69]2.agec = 0{p_end}
{p 0 7}{space 1}{text:( 3)}{space 1} [chd69]3.agec = 0{p_end}
{p 0 7}{space 1}{text:( 4)}{space 1} [chd69]4.agec = 0{p_end}

{txt}{col 12}chi2(  4) ={res}   44.08
{txt}{col 10}Prob > chi2 =  {res}  0.0000
{txt}
{com}. contrast q(1).agec, noeffects
{res}
{txt}Contrasts of marginal linear predictions

{txt}{p2colset 1 14 16 2}{...}
{p2col:Margins}:{space 1}{res:asbalanced}{p_end}
{p2colreset}{...}

{col 1}{text}{hline 13}{c TT}{hline 11}{hline 12}{hline 11}
{col 14}{text}{c |}         df{col 26}        chi2{col 38}     P>chi2
{col 1}{text}{hline 13}{c +}{hline 11}{hline 12}{hline 11}
{space 8}agec {col 14}{text}{c |}{result}{space 2}        1{col 26}{space 3}    31.45{col 38}{space 2}   0.0000
{col 1}{text}{hline 13}{c BT}{hline 11}{hline 12}{hline 11}

{com}. test -1.agec + 3.agec + 2*4.agec=0

{p 0 7}{space 1}{text:( 1)}{space 1}{space 1}{res}- [chd69]1.agec + [chd69]3.agec + 2{res}*{res}[chd69]4.agec = 0{p_end}

{txt}{col 12}chi2(  1) ={res}   31.45
{txt}{col 10}Prob > chi2 =  {res}  0.0000
{txt}
{com}. * evaluate evidence for departue from linearity
. contrast q(2/4).agec, noeffects
{res}
{txt}Contrasts of marginal linear predictions

{txt}{p2colset 1 14 16 2}{...}
{p2col:Margins}:{space 1}{res:asbalanced}{p_end}
{p2colreset}{...}

{col 1}{text}{hline 13}{c TT}{hline 11}{hline 12}{hline 11}
{col 14}{text}{c |}         df{col 26}        chi2{col 38}     P>chi2
{col 1}{text}{hline 13}{c +}{hline 11}{hline 12}{hline 11}
{space 8}agec {c |}
(quadratic)  {col 14}{text}{c |}{result}{space 2}        1{col 26}{space 3}     0.33{col 38}{space 2}   0.5647
{txt}    (cubic)  {col 14}{text}{c |}{result}{space 2}        1{col 26}{space 3}     3.67{col 38}{space 2}   0.0555
{txt}  (quartic)  {col 14}{text}{c |}{result}{space 2}        1{col 26}{space 3}     1.62{col 38}{space 2}   0.2032
{col 1}{text}      Joint {col 14}{c |}{result}{space 2}        3{col 26}{space 3}     5.29{col 38}{space 2}   0.1518
{col 1}{text}{hline 13}{c BT}{hline 11}{hline 12}{hline 11}

{com}. 
. lincom 3.agec-2.agec

{p 0 7}{space 1}{text:( 1)}{space 1}{space 1}{res}- [chd69]2.agec + [chd69]3.agec = 0{p_end}

{txt}{hline 13}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 1}       chd69{col 14}{c |} Odds Ratio{col 26}   Std. Err.{col 38}      z{col 46}   P>|z|{col 54}     [95% Con{col 67}f. Interval]
{hline 13}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{space 9}(1) {c |}{col 14}{res}{space 2} 1.363777{col 26}{space 2} .2488687{col 37}{space 1}    1.70{col 46}{space 3}0.089{col 54}{space 4}    .9537{col 67}{space 3}  1.95018
{txt}{hline 13}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. logistic chd69 ib2.agec
{res}
{txt}Logistic regression{col 51}Number of obs{col 67}= {res}      3154
{txt}{col 51}LR chi2({res}4{txt}){col 67}= {res}     44.95
{txt}{col 51}Prob > chi2{col 67}= {res}    0.0000
{txt}Log likelihood = {res}-868.14866{txt}{col 51}Pseudo R2{col 67}= {res}    0.0252

{txt}{hline 13}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 1}       chd69{col 14}{c |} Odds Ratio{col 26}   Std. Err.{col 38}      z{col 46}   P>|z|{col 54}     [95% Con{col 67}f. Interval]
{hline 13}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{space 8}agec {c |}
{space 6}36-40  {c |}{col 14}{res}{space 2} .5881696{col 26}{space 2}  .131476{col 37}{space 1}   -2.37{col 46}{space 3}0.018{col 54}{space 4}  .379516{col 67}{space 3} .9115388
{txt}{space 6}41-45  {c |}{col 14}{res}{space 2} .5157198{col 26}{space 2} .0963491{col 37}{space 1}   -3.54{col 46}{space 3}0.000{col 54}{space 4} .3575926{col 67}{space 3} .7437707
{txt}{space 6}51-55  {c |}{col 14}{res}{space 2} 1.363777{col 26}{space 2} .2488687{col 37}{space 1}    1.70{col 46}{space 3}0.089{col 54}{space 4}    .9537{col 67}{space 3}  1.95018
{txt}{space 6}56-60  {c |}{col 14}{res}{space 2} 1.697642{col 26}{space 2} .3734367{col 37}{space 1}    2.41{col 46}{space 3}0.016{col 54}{space 4} 1.103073{col 67}{space 3} 2.612692
{txt}{space 12} {c |}
{space 7}_cons {c |}{col 14}{res}{space 2} .1029412{col 26}{space 2} .0129216{col 37}{space 1}  -18.11{col 46}{space 3}0.000{col 54}{space 4} .0804902{col 67}{space 3} .1316544
{txt}{hline 13}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. 
. * lowess fit & linear logistic model for age as continuous
. 
. lowess chd69 age, bw(0.5) logit generate(chdsm) name(age_lws)
{res}{txt}
{com}. logistic chd69 age
{res}
{txt}Logistic regression{col 51}Number of obs{col 67}= {res}      3154
{txt}{col 51}LR chi2({res}1{txt}){col 67}= {res}     42.89
{txt}{col 51}Prob > chi2{col 67}= {res}    0.0000
{txt}Log likelihood = {res}-869.17806{txt}{col 51}Pseudo R2{col 67}= {res}    0.0241

{txt}{hline 13}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 1}       chd69{col 14}{c |} Odds Ratio{col 26}   Std. Err.{col 38}      z{col 46}   P>|z|{col 54}     [95% Con{col 67}f. Interval]
{hline 13}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{space 9}age {c |}{col 14}{res}{space 2} 1.077262{col 26}{space 2} .0121757{col 37}{space 1}    6.58{col 46}{space 3}0.000{col 54}{space 4}  1.05366{col 67}{space 3} 1.101392
{txt}{space 7}_cons {c |}{col 14}{res}{space 2} .0026333{col 26}{space 2} .0014465{col 37}{space 1}  -10.81{col 46}{space 3}0.000{col 54}{space 4} .0008973{col 67}{space 3} .0077283
{txt}{hline 13}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. logistic chd69 age, coef
{res}
{txt}Logistic regression{col 51}Number of obs{col 67}= {res}      3154
{txt}{col 51}LR chi2({res}1{txt}){col 67}= {res}     42.89
{txt}{col 51}Prob > chi2{col 67}= {res}    0.0000
{txt}Log likelihood = {res}-869.17806{txt}{col 51}Pseudo R2{col 67}= {res}    0.0241

{txt}{hline 13}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 1}       chd69{col 14}{c |}      Coef.{col 26}   Std. Err.{col 38}      z{col 46}   P>|z|{col 54}     [95% Con{col 67}f. Interval]
{hline 13}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{space 9}age {c |}{col 14}{res}{space 2} .0744226{col 26}{space 2} .0113024{col 37}{space 1}    6.58{col 46}{space 3}0.000{col 54}{space 4} .0522702{col 67}{space 3} .0965749
{txt}{space 7}_cons {c |}{col 14}{res}{space 2}-5.939516{col 26}{space 2} .5493226{col 37}{space 1}  -10.81{col 46}{space 3}0.000{col 54}{space 4}-7.016168{col 67}{space 3}-4.862863
{txt}{hline 13}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. predict lp, xb
{txt}
{com}. 
. * generate spline basis and fit model
. mkspline agesp=age, cubic
{txt}
{com}. logistic chd69 agesp*
{res}
{txt}Logistic regression{col 51}Number of obs{col 67}= {res}      3154
{txt}{col 51}LR chi2({res}4{txt}){col 67}= {res}     51.10
{txt}{col 51}Prob > chi2{col 67}= {res}    0.0000
{txt}Log likelihood = {res}-865.07152{txt}{col 51}Pseudo R2{col 67}= {res}    0.0287

{txt}{hline 13}{c TT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{col 1}       chd69{col 14}{c |} Odds Ratio{col 26}   Std. Err.{col 38}      z{col 46}   P>|z|{col 54}     [95% Con{col 67}f. Interval]
{hline 13}{c +}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}
{space 6}agesp1 {c |}{col 14}{res}{space 2} .7544048{col 26}{space 2} .1099386{col 37}{space 1}   -1.93{col 46}{space 3}0.053{col 54}{space 4} .5669696{col 67}{space 3} 1.003805
{txt}{space 6}agesp2 {c |}{col 14}{res}{space 2} 32.42387{col 26}{space 2} 52.31911{col 37}{space 1}    2.16{col 46}{space 3}0.031{col 54}{space 4} 1.372041{col 67}{space 3}  766.236
{txt}{space 6}agesp3 {c |}{col 14}{res}{space 2} .0003586{col 26}{space 2} .0014684{col 37}{space 1}   -1.94{col 46}{space 3}0.053{col 54}{space 4} 1.17e-07{col 67}{space 3} 1.096005
{txt}{space 6}agesp4 {c |}{col 14}{res}{space 2} 140.6726{col 26}{space 2} 454.3082{col 37}{space 1}    1.53{col 46}{space 3}0.126{col 54}{space 4} .2507358{col 67}{space 3} 78922.78
{txt}{space 7}_cons {c |}{col 14}{res}{space 2} 4251.682{col 26}{space 2} 24861.72{col 37}{space 1}    1.43{col 46}{space 3}0.153{col 54}{space 4} .0447876{col 67}{space 3} 4.04e+08
{txt}{hline 13}{c BT}{hline 11}{hline 11}{hline 9}{hline 8}{hline 13}{hline 12}

{com}. predict lps, xb
{txt}
{com}. 
. * test for nonlinearity of log outcome odds in age
. test agesp2 agesp3 agesp4

{p 0 7}{space 1}{text:( 1)}{space 1} {res}[chd69]agesp2 = 0{p_end}
{p 0 7}{space 1}{text:( 2)}{space 1} [chd69]agesp3 = 0{p_end}
{p 0 7}{space 1}{text:( 3)}{space 1} [chd69]agesp4 = 0{p_end}

{txt}{col 12}chi2(  3) ={res}    8.07
{txt}{col 10}Prob > chi2 =  {res}  0.0446
{txt}
{com}. 
. twoway (connected chdsm age, sort msymbol(none)) (connected lp age, sort msymbol(none)) (connected lps age, sort msymbol(none)), legend(order(1 2 3) label(1 "lowess") label(2 "linear") label(3 "spline") pos(6) ring(0)) name(chdage_pr)
{res}{txt}
{com}. 
. tabodds chd69 age, ciplot
{res}
{txt}{hline 12}{c TT}{hline 61}
       age  {c |}      cases     controls       odds      [95% Conf. Interval]
{hline 12}{c +}{hline 61}
{col 3}       39 {c |}  {res}       19          247    0.07692        0.04824   0.12266
{txt}{col 3}       40 {c |}  {res}       12          265    0.04528        0.02539   0.08075
{txt}{col 3}       41 {c |}  {res}       12          221    0.05430        0.03037   0.09707
{txt}{col 3}       42 {c |}  {res}        5          217    0.02304        0.00949   0.05592
{txt}{col 3}       43 {c |}  {res}       13          202    0.06436        0.03673   0.11276
{txt}{col 3}       44 {c |}  {res}       13          222    0.05856        0.03347   0.10244
{txt}{col 3}       45 {c |}  {res}       12          174    0.06897        0.03842   0.12379
{txt}{col 3}       46 {c |}  {res}       11          159    0.06918        0.03755   0.12746
{txt}{col 3}       47 {c |}  {res}        8          139    0.05755        0.02822   0.11737
{txt}{col 3}       48 {c |}  {res}       19          145    0.13103        0.08123   0.21138
{txt}{col 3}       49 {c |}  {res}       21          113    0.18584        0.11665   0.29608
{txt}{col 3}       50 {c |}  {res}       11          124    0.08871        0.04788   0.16435
{txt}{col 3}       51 {c |}  {res}       17          106    0.16038        0.09611   0.26763
{txt}{col 3}       52 {c |}  {res}       15           98    0.15306        0.08889   0.26355
{txt}{col 3}       53 {c |}  {res}       12           93    0.12903        0.07073   0.23539
{txt}{col 3}       54 {c |}  {res}       11           96    0.11458        0.06140   0.21383
{txt}{col 3}       55 {c |}  {res}       10           70    0.14286        0.07365   0.27712
{txt}{col 3}       56 {c |}  {res}       12           64    0.18750        0.10121   0.34735
{txt}{col 3}       57 {c |}  {res}        7           56    0.12500        0.05697   0.27425
{txt}{col 3}       58 {c |}  {res}        6           50    0.12000        0.05145   0.27986
{txt}{col 3}       59 {c |}  {res}       11           36    0.30556        0.15554   0.60026
{txt}{hline 12}{c BT}{hline 61}
Test of homogeneity (equal odds): chi2({res}20{txt})  = {res}   72.98
                                   {txt}Pr>chi2  = {res}  0.0000

{txt}Score test for trend of odds:     chi2({res}1{txt})  = {res}   44.85
                                  {txt}Pr>chi2  = {res}  0.0000
{txt}
{com}. 
. 
. * generate distinct levels of age
. levelsof age
{txt}39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59

{com}. 
. * generate log-odds estimates and pointwise 95% CI for unique ages based on the previously fitted spline model, save results as new variables for graphing ("x" represents age)
. xblc agesp*, covname(age) at(`r(levels)') generate(x losp lolb loub)

{col 1}age{col 15}xb{col 26}(95% CI)
{txt}{col 1}39{col 15}-2.64{col 25} (-3.06--2.21)
{col 1}40{col 15}-2.91{col 25} (-3.19--2.62)
{col 1}41{col 15}-3.11{col 25} (-3.43--2.79)
{col 1}42{col 15}-3.19{col 25} (-3.56--2.82)
{col 1}43{col 15}-3.10{col 25} (-3.43--2.77)
{col 1}44{col 15}-2.90{col 25} (-3.16--2.63)
{col 1}45{col 15}-2.67{col 25} (-2.94--2.40)
{col 1}46{col 15}-2.48{col 25} (-2.77--2.19)
{col 1}47{col 15}-2.33{col 25} (-2.60--2.07)
{col 1}48{col 15}-2.22{col 25} (-2.45--2.00)
{col 1}49{col 15}-2.14{col 25} (-2.35--1.93)
{col 1}50{col 15}-2.07{col 25} (-2.30--1.85)
{col 1}51{col 15}-2.02{col 25} (-2.26--1.78)
{col 1}52{col 15}-1.97{col 25} (-2.21--1.73)
{col 1}53{col 15}-1.93{col 25} (-2.16--1.70)
{col 1}54{col 15}-1.89{col 25} (-2.10--1.68)
{col 1}55{col 15}-1.86{col 25} (-2.07--1.64)
{col 1}56{col 15}-1.83{col 25} (-2.07--1.58)
{col 1}57{col 15}-1.80{col 25} (-2.10--1.49)
{col 1}58{col 15}-1.77{col 25} (-2.15--1.39)
{col 1}59{col 15}-1.74{col 25} (-2.20--1.28)

{com}. 
. * plot, with style options suitable for printing/presentation
. twoway (rarea lolb loub x, color(gray)) (connected losp x, sort msymbol(none) color(black) lpattern(solid)) , xtitle("Age") ytitle("Log Odds of CHD") legend(order(2 1) label(1 "95% CI") label(2 "Log odds") pos(5) ring(0) region(lwidth(none))) plotregion(style(none)) scheme(s1mono) name(chdage_lsp)
{res}{txt}
{com}. 
. * convert to probability scale from log odds
. gen prsp = exp(losp)/(1 + exp(losp))
{txt}(3133 missing values generated)

{com}. gen plb = exp(lolb)/(1 + exp(lolb))
{txt}(3133 missing values generated)

{com}. gen pub = exp(loub)/(1 + exp(loub))
{txt}(3133 missing values generated)

{com}. twoway (rarea plb pub x, color(gray)) (connected prsp x, sort msymbol(none) color(black) lpattern(solid)) , xtitle("Age") ytitle("Probability CHD") legend(order(2 1) label(1 "95% CI") label(2 "Probability") pos(6) ring(0) region(lwidth(none))) plotregion(style(none)) scheme(s1mono) name(chdage_psp)
{res}{txt}
{com}. graph export pr_age_sp.pdf, replace
{txt}(file /Users/steve/Documents/teaching/c2015/biostat208/labs/lab8/pr_age_sp.pdf written in PDF format)

{com}. 
. log close
      {txt}name:  {res}<unnamed>
       {txt}log:  {res}/Users/steve/Documents/teaching/c2015/biostat208/labs/lab8/lab8.smcl
  {txt}log type:  {res}smcl
 {txt}closed on:  {res}26 Feb 2015, 10:02:01
{txt}{.-}
{smcl}
{txt}{sf}{ul off}