Regresion de Poisson para muestra unificada (14 de octubre)

El diseño muestral es el que se logró armar considerando los diseños de cada país Finalmente se trabaja con un diseño equivalente a 2 etapas, primera donde las psu son las escuelas, el pais es la variable de estratficacióny se usan como pesos muestrales los que se consideraron en Uruguay (ver diseño de Uruguay y para POA se creó un nuevo vector de peso usando como variables de ajuste distribucion por sexo y por tipo de escuela)

LOs analisis son unificados (no por pais)

names(disenio_urubra$variables) [1] “id” “pais” “ficha”
[4] “eta” “tipoesc” “sexoinv”
[7] “weight” “weight.rec” “numesc”
[10] “numer_esc” “idade” “socioecon4cat”
[13] “socioecon3cat” “escolmae” “escolmae13cat”
[16] “escolmaerecat23cat” “freqescov” “freqescovacat”
[19] “usofio” “freqfio” “usocreme”
[22] “idadcreme” “visidentcatonde” “visiquando”
[25] “fluorprof” “idadefluor” “isg.”
[28] “isg45” “isg20” “cposoms”
[31] “cpodoms” “cpodoms” “cposbere”
[34] “cpodbere” “prevbere” “cposicdas”
[37] “cpodicdas” “previcdas”


# 28 de octubre 2013

load("~/Dropbox/odontologia/maestria_anunziatta/BancoUnidoBrasUru/BRASURU(28102013).RData")

library(survey)
## Attaching package: 'survey'
## 
## The following object(s) are masked from 'package:graphics':
## 
## dotchart
levels(disenio_urubra$variables$prevoms)
## [1] "1-SinCavi" "2-conCavi"
table(disenio_urubra$variables$cpodoms, disenio_urubra$variables$prevoms)
##     
##      1-SinCavi 2-conCavi
##   0       1105         0
##   1          0       495
##   2          0       404
##   3          0       259
##   4          0       217
##   5          0       108
##   6          0        38
##   7          0        22
##   8          0        14
##   9          0        12
##   10         0         3
##   11         0         3
##   14         0         1
##   15         0         1



modelo0.poi <- svyglm(cpodoms ~ pais, disenio_urubra, family = quasipoisson())
summary(modelo0.poi)
## 
## Call:
## svyglm(formula = cpodoms ~ pais, disenio_urubra, family = quasipoisson())
## 
## Survey design:
## svydesign(id = ~numer_esc, strata = ~pais, weights = ~round(weight.rec, 
##     1), data = brasuru, nest = TRUE)
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)   0.3328     0.0623    5.34  8.5e-07 ***
## paisUru       0.1613     0.0924    1.75    0.085 .  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1 
## 
## (Dispersion parameter for quasipoisson family taken to be 2.224)
## 
## Number of Fisher Scoring iterations: 5
confint(modelo0.poi)
##                2.5 % 97.5 %
## (Intercept)  0.21066 0.4550
## paisUru     -0.01974 0.3423
exp(modelo0.poi$coefficients)
## (Intercept)     paisUru 
##       1.395       1.175


modelo1.poi <- svyglm(cpodoms ~ sexoinv, disenio_urubra, family = quasipoisson())
summary(modelo1.poi)
## 
## Call:
## svyglm(formula = cpodoms ~ sexoinv, disenio_urubra, family = quasipoisson())
## 
## Survey design:
## svydesign(id = ~numer_esc, strata = ~pais, weights = ~round(weight.rec, 
##     1), data = brasuru, nest = TRUE)
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)   0.3915     0.0648    6.04  4.6e-08 ***
## sexoinv2-F    0.0732     0.0735    1.00     0.32    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1 
## 
## (Dispersion parameter for quasipoisson family taken to be 2.233)
## 
## Number of Fisher Scoring iterations: 5
confint(modelo1.poi)
##                2.5 % 97.5 %
## (Intercept)  0.26442 0.5186
## sexoinv2-F  -0.07076 0.2172
exp(modelo1.poi$coefficients)
## (Intercept)  sexoinv2-F 
##       1.479       1.076

modelo2.poi <- svyglm(cpodoms ~ socioecon4cat, disenio_urubra, family = quasipoisson())
summary(modelo2.poi)
## 
## Call:
## svyglm(formula = cpodoms ~ socioecon4cat, disenio_urubra, family = quasipoisson())
## 
## Survey design:
## svydesign(id = ~numer_esc, strata = ~pais, weights = ~round(weight.rec, 
##     1), data = brasuru, nest = TRUE)
## 
## Coefficients:
##                           Estimate Std. Error t value Pr(>|t|)    
## (Intercept)                 -0.227      0.158   -1.44    0.154    
## socioecon4cat2-medio-alto    0.564      0.231    2.45    0.017 *  
## socioecon4cat3-medio-bajo    0.795      0.173    4.60  1.6e-05 ***
## socioecon4cat4-bajo          0.922      0.170    5.41  6.6e-07 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1 
## 
## (Dispersion parameter for quasipoisson family taken to be 2.148)
## 
## Number of Fisher Scoring iterations: 6
confint(modelo2.poi)
##                             2.5 %  97.5 %
## (Intercept)               -0.5358 0.08197
## socioecon4cat2-medio-alto  0.1122 1.01644
## socioecon4cat3-medio-bajo  0.4562 1.13317
## socioecon4cat4-bajo        0.5885 1.25613
exp(modelo2.poi$coefficients)
##               (Intercept) socioecon4cat2-medio-alto 
##                     0.797                     1.758 
## socioecon4cat3-medio-bajo       socioecon4cat4-bajo 
##                     2.214                     2.515

modelo3.poi <- svyglm(cpodoms ~ socioecon3cat, disenio_urubra, family = quasipoisson())
summary(modelo3.poi)
## 
## Call:
## svyglm(formula = cpodoms ~ socioecon3cat, disenio_urubra, family = quasipoisson())
## 
## Survey design:
## svydesign(id = ~numer_esc, strata = ~pais, weights = ~round(weight.rec, 
##     1), data = brasuru, nest = TRUE)
## 
## Coefficients:
##                      Estimate Std. Error t value Pr(>|t|)    
## (Intercept)           -0.1135     0.0825   -1.38     0.17    
## socioecon3cat2-medio   0.6183     0.0986    6.27  1.8e-08 ***
## socioecon3cat3-bajo    0.8089     0.1034    7.82  1.9e-11 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1 
## 
## (Dispersion parameter for quasipoisson family taken to be 2.138)
## 
## Number of Fisher Scoring iterations: 5
confint(modelo3.poi)
##                        2.5 %  97.5 %
## (Intercept)          -0.2751 0.04817
## socioecon3cat2-medio  0.4250 0.81158
## socioecon3cat3-bajo   0.6062 1.01157
exp(modelo3.poi$coefficients)
##          (Intercept) socioecon3cat2-medio  socioecon3cat3-bajo 
##               0.8927               1.8557               2.2454

modelo4.poi <- svyglm(cpodoms ~ escolmaerecat23cat, disenio_urubra, family = quasipoisson())
summary(modelo4.poi)
## 
## Call:
## svyglm(formula = cpodoms ~ escolmaerecat23cat, disenio_urubra, 
##     family = quasipoisson())
## 
## Survey design:
## svydesign(id = ~numer_esc, strata = ~pais, weights = ~round(weight.rec, 
##     1), data = brasuru, nest = TRUE)
## 
## Coefficients:
##                                       Estimate Std. Error t value Pr(>|t|)
## (Intercept)                             0.0114     0.1356    0.08   0.9330
## escolmaerecat23cat2-high school         0.4003     0.1337    2.99   0.0037
## escolmaerecat23cat3-elementary school   0.6250     0.1513    4.13  8.9e-05
##                                          
## (Intercept)                              
## escolmaerecat23cat2-high school       ** 
## escolmaerecat23cat3-elementary school ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1 
## 
## (Dispersion parameter for quasipoisson family taken to be 2.208)
## 
## Number of Fisher Scoring iterations: 6
confint(modelo4.poi)
##                                         2.5 % 97.5 %
## (Intercept)                           -0.2544 0.2772
## escolmaerecat23cat2-high school        0.1382 0.6623
## escolmaerecat23cat3-elementary school  0.3286 0.9215
exp(modelo4.poi$coefficients)
##                           (Intercept) 
##                                 1.012 
##       escolmaerecat23cat2-high school 
##                                 1.492 
## escolmaerecat23cat3-elementary school 
##                                 1.868

modelo5.poi <- svyglm(cpodoms ~ escolmae13cat, disenio_urubra, family = quasipoisson())
summary(modelo5.poi)
## 
## Call:
## svyglm(formula = cpodoms ~ escolmae13cat, disenio_urubra, family = quasipoisson())
## 
## Survey design:
## svydesign(id = ~numer_esc, strata = ~pais, weights = ~round(weight.rec, 
##     1), data = brasuru, nest = TRUE)
## 
## Coefficients:
##                                  Estimate Std. Error t value Pr(>|t|)    
## (Intercept)                        -0.275      0.124   -2.21     0.03 *  
## escolmae13cat2-high school          0.507      0.213    2.38     0.02 *  
## escolmae13cat3-Elementary School    0.884      0.134    6.62  3.9e-09 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1 
## 
## (Dispersion parameter for quasipoisson family taken to be 2.128)
## 
## Number of Fisher Scoring iterations: 6
confint(modelo5.poi)
##                                     2.5 %   97.5 %
## (Intercept)                      -0.51843 -0.03082
## escolmae13cat2-high school        0.08954  0.92480
## escolmae13cat3-Elementary School  0.62255  1.14621
exp(modelo5.poi$coefficients)
##                      (Intercept)       escolmae13cat2-high school 
##                           0.7599                           1.6606 
## escolmae13cat3-Elementary School 
##                           2.4215


modelo6.poi <- svyglm(cpodoms ~ freqescovacat, disenio_urubra, family = quasipoisson())
summary(modelo6.poi)
## 
## Call:
## svyglm(formula = cpodoms ~ freqescovacat, disenio_urubra, family = quasipoisson())
## 
## Survey design:
## svydesign(id = ~numer_esc, strata = ~pais, weights = ~round(weight.rec, 
##     1), data = brasuru, nest = TRUE)
## 
## Coefficients:
##                                Estimate Std. Error t value Pr(>|t|)    
## (Intercept)                      0.6176     0.0473   13.07  < 2e-16 ***
## freqescovacat2-veces al d\xeda  -0.1737     0.0845   -2.05    0.043 *  
## freqescovacat3-veces al d\xeda  -0.3166     0.0726   -4.36  3.9e-05 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1 
## 
## (Dispersion parameter for quasipoisson family taken to be 2.197)
## 
## Number of Fisher Scoring iterations: 5
confint(modelo6.poi)
##                                  2.5 %    97.5 %
## (Intercept)                     0.5250  0.710275
## freqescovacat2-veces al d\xeda -0.3394 -0.008004
## freqescovacat3-veces al d\xeda -0.4590 -0.174279
exp(modelo6.poi$coefficients)
##                    (Intercept) freqescovacat2-veces al d\xeda 
##                         1.8545                         0.8406 
## freqescovacat3-veces al d\xeda 
##                         0.7286

modelo7.poi <- svyglm(cpodoms ~ usocreme, disenio_urubra, family = quasipoisson())
summary(modelo7.poi)
## 
## Call:
## svyglm(formula = cpodoms ~ usocreme, disenio_urubra, family = quasipoisson())
## 
## Survey design:
## svydesign(id = ~numer_esc, strata = ~pais, weights = ~round(weight.rec, 
##     1), data = brasuru, nest = TRUE)
## 
## Coefficients:
##              Estimate Std. Error t value Pr(>|t|)    
## (Intercept)    0.4225     0.0499    8.46    1e-12 ***
## usocreme2-No   0.1866     0.1920    0.97     0.33    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1 
## 
## (Dispersion parameter for quasipoisson family taken to be 2.239)
## 
## Number of Fisher Scoring iterations: 5
confint(modelo7.poi)
##                2.5 % 97.5 %
## (Intercept)   0.3246 0.5204
## usocreme2-No -0.1898 0.5629
exp(modelo7.poi$coefficients)
##  (Intercept) usocreme2-No 
##        1.526        1.205

modelo8.poi <- svyglm(cpodoms ~ visidentcatonde, disenio_urubra, family = quasipoisson())
summary(modelo8.poi)
## 
## Call:
## svyglm(formula = cpodoms ~ visidentcatonde, disenio_urubra, family = quasipoisson())
## 
## Survey design:
## svydesign(id = ~numer_esc, strata = ~pais, weights = ~round(weight.rec, 
##     1), data = brasuru, nest = TRUE)
## 
## Coefficients:
##                                        Estimate Std. Error t value
## (Intercept)                              0.2528     0.0655    3.86
## visidentcatonde2-Publico                 0.4146     0.0723    5.73
## visidentcatonde3-Nunca fue al dentista   0.2323     0.0954    2.43
##                                        Pr(>|t|)    
## (Intercept)                             0.00023 ***
## visidentcatonde2-Publico                1.7e-07 ***
## visidentcatonde3-Nunca fue al dentista  0.01721 *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1 
## 
## (Dispersion parameter for quasipoisson family taken to be 2.235)
## 
## Number of Fisher Scoring iterations: 6
confint(modelo8.poi)
##                                          2.5 % 97.5 %
## (Intercept)                            0.12444 0.3812
## visidentcatonde2-Publico               0.27292 0.5563
## visidentcatonde3-Nunca fue al dentista 0.04521 0.4193
exp(modelo8.poi$coefficients)
##                            (Intercept) 
##                                  1.288 
##               visidentcatonde2-Publico 
##                                  1.514 
## visidentcatonde3-Nunca fue al dentista 
##                                  1.261

modelo9.poi <- svyglm(cpodoms ~ visiquando, disenio_urubra, family = quasipoisson())
summary(modelo9.poi)
## 
## Call:
## svyglm(formula = cpodoms ~ visiquando, disenio_urubra, family = quasipoisson())
## 
## Survey design:
## svydesign(id = ~numer_esc, strata = ~pais, weights = ~round(weight.rec, 
##     1), data = brasuru, nest = TRUE)
## 
## Coefficients:
##                                  Estimate Std. Error t value Pr(>|t|)    
## (Intercept)                        0.4188     0.0607    6.90  1.1e-09 ***
## visiquando2- m\xe1s de 2 a\xf1os  -0.0173     0.1448   -0.12     0.91    
## visiquando3- nunca fue             0.0736     0.0923    0.80     0.43    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1 
## 
## (Dispersion parameter for quasipoisson family taken to be 2.249)
## 
## Number of Fisher Scoring iterations: 5
confint(modelo9.poi)
##                                    2.5 % 97.5 %
## (Intercept)                       0.2999 0.5377
## visiquando2- m\xe1s de 2 a\xf1os -0.3012 0.2666
## visiquando3- nunca fue           -0.1072 0.2545
exp(modelo9.poi$coefficients)
##                      (Intercept) visiquando2- m\xe1s de 2 a\xf1os 
##                           1.5202                           0.9828 
##           visiquando3- nunca fue 
##                           1.0764

modelo10.poi <- svyglm(cpodoms ~ fluorprof, disenio_urubra, family = quasipoisson())
summary(modelo10.poi)
## 
## Call:
## svyglm(formula = cpodoms ~ fluorprof, disenio_urubra, family = quasipoisson())
## 
## Survey design:
## svydesign(id = ~numer_esc, strata = ~pais, weights = ~round(weight.rec, 
##     1), data = brasuru, nest = TRUE)
## 
## Coefficients:
##               Estimate Std. Error t value Pr(>|t|)    
## (Intercept)     0.4214     0.0597    7.05  5.6e-10 ***
## fluorprof2-No  -0.0118     0.0692   -0.17     0.86    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1 
## 
## (Dispersion parameter for quasipoisson family taken to be 2.246)
## 
## Number of Fisher Scoring iterations: 5
confint(modelo10.poi)
##                 2.5 % 97.5 %
## (Intercept)    0.3043 0.5385
## fluorprof2-No -0.1475 0.1238
exp(modelo10.poi$coefficients)
##   (Intercept) fluorprof2-No 
##        1.5240        0.9882

modelo11.poi <- svyglm(cpodoms ~ isg20 + pais, disenio_urubra, family = quasipoisson())
summary(modelo11.poi)
## 
## Call:
## svyglm(formula = cpodoms ~ isg20 + pais, disenio_urubra, family = quasipoisson())
## 
## Survey design:
## svydesign(id = ~numer_esc, strata = ~pais, weights = ~round(weight.rec, 
##     1), data = brasuru, nest = TRUE)
## 
## Coefficients:
##                 Estimate Std. Error t value Pr(>|t|)    
## (Intercept)        0.157      0.113    1.38  0.17095    
## isg202- 20 a 60    0.042      0.105    0.40  0.69003    
## isg203- >=60       0.421      0.102    4.12  9.3e-05 ***
## paisUru            0.282      0.082    3.44  0.00095 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1 
## 
## (Dispersion parameter for quasipoisson family taken to be 2.197)
## 
## Number of Fisher Scoring iterations: 5
confint(modelo11.poi)
##                    2.5 % 97.5 %
## (Intercept)     -0.06551 0.3787
## isg202- 20 a 60 -0.16350 0.2474
## isg203- >=60     0.22067 0.6209
## paisUru          0.12105 0.4424
exp(modelo11.poi$coefficients)
##     (Intercept) isg202- 20 a 60    isg203- >=60         paisUru 
##           1.170           1.043           1.523           1.325

modelo12.poi <- svyglm(cpodoms ~ isg45, disenio_urubra, family = quasipoisson())
summary(modelo12.poi)
## 
## Call:
## svyglm(formula = cpodoms ~ isg45, disenio_urubra, family = quasipoisson())
## 
## Survey design:
## svydesign(id = ~numer_esc, strata = ~pais, weights = ~round(weight.rec, 
##     1), data = brasuru, nest = TRUE)
## 
## Coefficients:
##                 Estimate Std. Error t value Pr(>|t|)    
## (Intercept)       0.3574     0.0701    5.10  2.3e-06 ***
## isg452- 45 a 60   0.0554     0.0849    0.65  0.51626    
## isg453- >=60      0.2938     0.0727    4.04  0.00012 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1 
## 
## (Dispersion parameter for quasipoisson family taken to be 2.229)
## 
## Number of Fisher Scoring iterations: 5
confint(modelo12.poi)
##                   2.5 % 97.5 %
## (Intercept)      0.2200 0.4947
## isg452- 45 a 60 -0.1111 0.2218
## isg453- >=60     0.1513 0.4362
exp(modelo12.poi$coefficients)
##     (Intercept) isg452- 45 a 60    isg453- >=60 
##           1.430           1.057           1.341


modelo13.poi <- svyglm(cpodoms ~ tipoesc, disenio_urubra, family = quasipoisson())
summary(modelo13.poi)
## 
## Call:
## svyglm(formula = cpodoms ~ tipoesc, disenio_urubra, family = quasipoisson())
## 
## Survey design:
## svydesign(id = ~numer_esc, strata = ~pais, weights = ~round(weight.rec, 
##     1), data = brasuru, nest = TRUE)
## 
## Coefficients:
##                  Estimate Std. Error t value Pr(>|t|)    
## (Intercept)       -0.0741     0.0747   -0.99     0.32    
## tipoesc2-Publica   0.6237     0.0818    7.62  4.4e-11 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1 
## 
## (Dispersion parameter for quasipoisson family taken to be 2.127)
## 
## Number of Fisher Scoring iterations: 5
confint(modelo13.poi)
##                    2.5 %  97.5 %
## (Intercept)      -0.2206 0.07238
## tipoesc2-Publica  0.4633 0.78404
exp(modelo13.poi$coefficients)
##      (Intercept) tipoesc2-Publica 
##           0.9286           1.8658