library(ggplot2)
library(dplyr)
coefs<-data.frame(coefs=c(coef(fit1)[-1], coef(fit3)[-1]),
                  mod=c(rep("Non Survey Model", 11),rep("Survey Model", 11)),
                  effect=rep(names(coef(fit1)[-1]), 2))

coefs%>%
  ggplot()+
  geom_point(aes( x=effect,
                  y=coefs,
                  group=effect,
                  color=effect,
                  shape=mod),
             position=position_jitterdodge(jitter.width = 1),
             size=2)+
  ylab("Regression Coefficient")+
  xlab("Beta")+
  geom_abline(intercept = 0, slope=0)+
  theme(axis.text.x = element_text(angle = 45, hjust = 1))+
  ggtitle(label = "Comparison of Survey and Non-Survey Regression effects")

prop.table(table(dawn2011$NUMSUBS))
## 
##             1             2             3             4             5 
## 0.64194572131 0.23812953723 0.07515813460 0.02482134828 0.00900134016 
##             6             7             8             9            10 
## 0.00366252395 0.00196877033 0.00116554696 0.00088616491 0.00065480166 
##            11            12            13            14            15 
## 0.00051074530 0.00047145720 0.00034486221 0.00027938204 0.00024882463 
##            16            17            18            19            20 
## 0.00041907306 0.00006111482 0.00004801879 0.00006548017 0.00006548017 
##            21            22 
## 0.00006548017 0.00002619207
prop.table(svytable(~NUMSUBS,
                    design = des))
## NUMSUBS
##              1              2              3              4              5 
## 0.663261994472 0.212895252977 0.072157956350 0.027315062163 0.010626180557 
##              6              7              8              9             10 
## 0.004878337188 0.002273223187 0.001516999879 0.001285819703 0.000600102817 
##             11             12             13             14             15 
## 0.000557468715 0.000607745688 0.000319773604 0.000139870676 0.000560444990 
##             16             17             18             19             20 
## 0.000580488545 0.000069988831 0.000172696735 0.000015285094 0.000116820791 
##             21             22 
## 0.000042477023 0.000006010015
mean(dawn2011$NUMSUBS)
## [1] 1.584092
svymean(~NUMSUBS,
        design = des)
##          mean     SE
## NUMSUBS 1.588 0.0538
cat<-svyby(formula = ~morethanonesub,
           by = ~age,
           design = des,
           FUN = svymean,
           na.rm=T)

svychisq(~morethanonesub+age,
         design = des)
## 
##  Pearson's X^2: Rao & Scott adjustment
## 
## data:  svychisq(~morethanonesub + age, design = des)
## F = 85.242, ndf = 3.2842, ddf = 597.7269, p-value < 0.00000000000000022
cat%>%
  ggplot()+
  geom_point(aes(x=age,y=morethanonesub))+
  geom_errorbar(aes(x=age, ymin = morethanonesub-1.96*se, 
                    ymax= morethanonesub+1.96*se),
                width=.25)+
   labs(title = "Percent % of cases with ore than one substance", 
        caption = "Source: DAWN 2011",
       x = "Age",
       y = "%  More than one Substance")+
  theme_minimal()

dog<-svyby(formula = ~morethanonesub,
           by = ~SEX, 
           design = des, 
           FUN = svymean,
           na.rm=T)

svychisq(~morethanonesub+SEX,
         design = des)
## 
##  Pearson's X^2: Rao & Scott adjustment
## 
## data:  svychisq(~morethanonesub + SEX, design = des)
## F = 135.42, ndf = 1, ddf = 182, p-value < 0.00000000000000022
dog%>%
  ggplot()+
  geom_point(aes(x=SEX,y=morethanonesub))+
  geom_errorbar(aes(x=SEX, ymin = morethanonesub-1.96*se, 
                    ymax= morethanonesub+1.96*se),
                width=.25)+
   labs(title = "Percent % of cases with more than one substance by sex", 
        caption = "Source: DAWN Data 2011",
       x = "Sex",
       y = "%  more than one substance")+
  theme_minimal()

catdog<-svyby(formula = ~morethanonesub,
              by = ~age+SEX,
              design = des,
              FUN = svymean,
              na.rm=T)


catdog%>%
  ggplot()+
  geom_errorbar(aes(x=age,y = morethanonesub,
                    ymin = morethanonesub-1.96*se, 
                   ymax= morethanonesub+1.96*se,
                   color=SEX,
                   group=SEX),
                width=.25,
                position="dodge")+
  labs(title = "Percent % of cases with more than one substance by Sex and Age", 
        caption = "Source: DAWN Data 2011",
       x = "age",
       y = "%  more than one substance")+
  theme_minimal()

fit.logit<-svyglm(morethanonesub ~ age + SEX,
                  design = des,
                  family = binomial)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
summary(fit.logit)
## 
## Call:
## svyglm(formula = morethanonesub ~ age + SEX, design = des, family = binomial)
## 
## Survey design:
## svydesign(ids = ~PSU, strata = ~STRATA, weights = ~CASEWGT, nest = TRUE, 
##     data = dawn2011)
## 
## Coefficients:
##                              Estimate Std. Error t value             Pr(>|t|)
## (Intercept)                  -2.00855    0.08218 -24.441 < 0.0000000000000002
## age(02) 6 TO 11:(2)           0.42767    0.13411   3.189               0.0017
## age(03) 12 TO 17:(3)          1.27097    0.11974  10.615 < 0.0000000000000002
## age(04) 18 TO 20:(4)          1.35699    0.11944  11.361 < 0.0000000000000002
## age(05) 21 TO 24:(5)          1.97817    0.09840  20.104 < 0.0000000000000002
## age(06) 25 TO 29:(6)          1.92360    0.11194  17.184 < 0.0000000000000002
## age(07) 30 TO 34:(7)          1.95377    0.12997  15.032 < 0.0000000000000002
## age(08) 35 TO 44:(8)          1.76244    0.09800  17.984 < 0.0000000000000002
## age(09) 45 TO 54:(9)          1.72504    0.08674  19.888 < 0.0000000000000002
## age(10) 55 TO 64:(10)         1.34163    0.11630  11.536 < 0.0000000000000002
## age(11) AGE 65 OR OLDER:(11)  0.83656    0.14227   5.880          0.000000021
## SEX(2) FEMALE:(2)            -0.33897    0.02688 -12.612 < 0.0000000000000002
##                                 
## (Intercept)                  ***
## age(02) 6 TO 11:(2)          ** 
## age(03) 12 TO 17:(3)         ***
## age(04) 18 TO 20:(4)         ***
## age(05) 21 TO 24:(5)         ***
## age(06) 25 TO 29:(6)         ***
## age(07) 30 TO 34:(7)         ***
## age(08) 35 TO 44:(8)         ***
## age(09) 45 TO 54:(9)         ***
## age(10) 55 TO 64:(10)        ***
## age(11) AGE 65 OR OLDER:(11) ***
## SEX(2) FEMALE:(2)            ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1.002533)
## 
## Number of Fisher Scoring iterations: 5
library(gtsummary)
fit.logit%>%
  tbl_regression(exponentiate=TRUE )
Characteristic OR1 95% CI1 p-value
age
(01) AGE 5 OR YOUNGER:(1) — —
(02) 6 TO 11:(2) 1.53 1.18, 2.00 0.002
(03) 12 TO 17:(3) 3.56 2.81, 4.51 <0.001
(04) 18 TO 20:(4) 3.88 3.07, 4.92 <0.001
(05) 21 TO 24:(5) 7.23 5.95, 8.78 <0.001
(06) 25 TO 29:(6) 6.85 5.49, 8.54 <0.001
(07) 30 TO 34:(7) 7.06 5.46, 9.12 <0.001
(08) 35 TO 44:(8) 5.83 4.80, 7.07 <0.001
(09) 45 TO 54:(9) 5.61 4.73, 6.66 <0.001
(10) 55 TO 64:(10) 3.83 3.04, 4.81 <0.001
(11) AGE 65 OR OLDER:(11) 2.31 1.74, 3.06 <0.001
SEX
(1) MALE:(1) — —
(2) FEMALE:(2) 0.71 0.68, 0.75 <0.001

1 OR = Odds Ratio, CI = Confidence Interval 

library(sjPlot)
plot_model(fit.logit,
           axis.lim = c(.1, 15), 
           title = "Odds ratios for More than one substance")

library(emmeans)
rg<-ref_grid(fit.logit)

marg_logit<-emmeans(object = rg,
              specs = c( "age",  "SEX"),
              type="more than one substance" )

knitr::kable(marg_logit,  digits = 4)
age SEX emmean SE df asymp.LCL asymp.UCL
(01) AGE 5 OR YOUNGER:(1) (1) MALE:(1) -2.0085 0.0822 Inf -2.1696 -1.8475
(02) 6 TO 11:(2) (1) MALE:(1) -1.5809 0.1268 Inf -1.8293 -1.3324
(03) 12 TO 17:(3) (1) MALE:(1) -0.7376 0.1045 Inf -0.9423 -0.5328
(04) 18 TO 20:(4) (1) MALE:(1) -0.6516 0.0933 Inf -0.8345 -0.4687
(05) 21 TO 24:(5) (1) MALE:(1) -0.0304 0.0810 Inf -0.1892 0.1284
(06) 25 TO 29:(6) (1) MALE:(1) -0.0850 0.0833 Inf -0.2482 0.0783
(07) 30 TO 34:(7) (1) MALE:(1) -0.0548 0.1202 Inf -0.2903 0.1808
(08) 35 TO 44:(8) (1) MALE:(1) -0.2461 0.0819 Inf -0.4066 -0.0856
(09) 45 TO 54:(9) (1) MALE:(1) -0.2835 0.0464 Inf -0.3744 -0.1926
(10) 55 TO 64:(10) (1) MALE:(1) -0.6669 0.0778 Inf -0.8193 -0.5145
(11) AGE 65 OR OLDER:(11) (1) MALE:(1) -1.1720 0.1199 Inf -1.4069 -0.9371
(01) AGE 5 OR YOUNGER:(1) (2) FEMALE:(2) -2.3475 0.0848 Inf -2.5137 -2.1814
(02) 6 TO 11:(2) (2) FEMALE:(2) -1.9199 0.1203 Inf -2.1557 -1.6840
(03) 12 TO 17:(3) (2) FEMALE:(2) -1.0766 0.0873 Inf -1.2477 -0.9055
(04) 18 TO 20:(4) (2) FEMALE:(2) -0.9905 0.0857 Inf -1.1586 -0.8225
(05) 21 TO 24:(5) (2) FEMALE:(2) -0.3693 0.0645 Inf -0.4957 -0.2430
(06) 25 TO 29:(6) (2) FEMALE:(2) -0.4239 0.0649 Inf -0.5511 -0.2968
(07) 30 TO 34:(7) (2) FEMALE:(2) -0.3938 0.1036 Inf -0.5968 -0.1907
(08) 35 TO 44:(8) (2) FEMALE:(2) -0.5851 0.0683 Inf -0.7190 -0.4511
(09) 45 TO 54:(9) (2) FEMALE:(2) -0.6225 0.0353 Inf -0.6916 -0.5533
(10) 55 TO 64:(10) (2) FEMALE:(2) -1.0059 0.0665 Inf -1.1363 -0.8755
(11) AGE 65 OR OLDER:(11) (2) FEMALE:(2) -1.5110 0.1088 Inf -1.7243 -1.2977 
comps<-as.data.frame(marg_logit)

comps[comps$age=="(02) 6 TO 11:(2)" | comps$age== "(03) 12 TO 17:(3)" | comps$age=="(01) AGE 5 OR YOUNGER:(1)", ]
##                          age            SEX     emmean         SE  df asymp.LCL
## 1  (01) AGE 5 OR YOUNGER:(1)   (1) MALE:(1) -2.0085490 0.08217979 Inf -2.169618
## 2           (02) 6 TO 11:(2)   (1) MALE:(1) -1.5808767 0.12675949 Inf -1.829321
## 3          (03) 12 TO 17:(3)   (1) MALE:(1) -0.7375776 0.10445828 Inf -0.942312
## 12 (01) AGE 5 OR YOUNGER:(1) (2) FEMALE:(2) -2.3475236 0.08476492 Inf -2.513660
## 13          (02) 6 TO 11:(2) (2) FEMALE:(2) -1.9198513 0.12033272 Inf -2.155699
## 14         (03) 12 TO 17:(3) (2) FEMALE:(2) -1.0765522 0.08729726 Inf -1.247652
##     asymp.UCL
## 1  -1.8474795
## 2  -1.3324327
## 3  -0.5328431
## 12 -2.1813874
## 13 -1.6840035
## 14 -0.9054527
comps[comps$age=="(05) 21 TO 24:(5)"  | comps$age == "(06) 25 to 29:(6)" & comps$SEX == "(1) MALE:(1)" , ]
##                  age            SEX      emmean         SE  df  asymp.LCL
## 5  (05) 21 TO 24:(5)   (1) MALE:(1) -0.03037434 0.08101976 Inf -0.1891701
## 16 (05) 21 TO 24:(5) (2) FEMALE:(2) -0.36934892 0.06448762 Inf -0.4957423
##     asymp.UCL
## 5   0.1284215
## 16 -0.2429555
comps[comps$age=="(09) 45 TO 54:(9)" & comps$SEX == "(2) FEMALE:(2)" | comps$SEX == "(1) MALE:(1)", ]
##                          age            SEX      emmean         SE  df
## 1  (01) AGE 5 OR YOUNGER:(1)   (1) MALE:(1) -2.00854897 0.08217979 Inf
## 2           (02) 6 TO 11:(2)   (1) MALE:(1) -1.58087670 0.12675949 Inf
## 3          (03) 12 TO 17:(3)   (1) MALE:(1) -0.73757758 0.10445828 Inf
## 4          (04) 18 TO 20:(4)   (1) MALE:(1) -0.65155765 0.09331716 Inf
## 5          (05) 21 TO 24:(5)   (1) MALE:(1) -0.03037434 0.08101976 Inf
## 6          (06) 25 TO 29:(6)   (1) MALE:(1) -0.08495205 0.08328404 Inf
## 7          (07) 30 TO 34:(7)   (1) MALE:(1) -0.05477837 0.12019006 Inf
## 8          (08) 35 TO 44:(8)   (1) MALE:(1) -0.24611145 0.08189529 Inf
## 9          (09) 45 TO 54:(9)   (1) MALE:(1) -0.28350707 0.04638464 Inf
## 10        (10) 55 TO 64:(10)   (1) MALE:(1) -0.66691925 0.07775252 Inf
## 11 (11) AGE 65 OR OLDER:(11)   (1) MALE:(1) -1.17198710 0.11985176 Inf
## 20         (09) 45 TO 54:(9) (2) FEMALE:(2) -0.62248166 0.03528506 Inf
##     asymp.LCL   asymp.UCL
## 1  -2.1696184 -1.84747953
## 2  -1.8293207 -1.33243266
## 3  -0.9423120 -0.53284311
## 4  -0.8344559 -0.46865938
## 5  -0.1891701  0.12842147
## 6  -0.2481858  0.07828167
## 7  -0.2903466  0.18078981
## 8  -0.4066233 -0.08559963
## 9  -0.3744193 -0.19259485
## 10 -0.8193114 -0.51452711
## 11 -1.4068922 -0.93708196
## 20 -0.6916391 -0.55332420

Once again, I am interested in looking at age and sex as my predictors for whether or not there is more than one substance involved in the case. Are men or women more likely to use more than one substance at a time? Does the likelihood of using more than one substance increase with age?

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