Consultoria Estatística
Universidade Estadual da Paraíba - UEPB
Professor: Ricardo Alves de Olinda
Alunos:
Giullber Valentim da Silva
José Lucas Costa de Oliveira
Banco de dados
Informações sobre o banco de dados:
- Dados referentes aos números de internações na Unidade de Tratamento Itensivo - UTI por Covid - 19 no estado da Paraíba.
- Possui registros de 2387 pacientes.
- Possui 19 variáveis, sendo 1 a variável dependente.
Análise Descritiva
Verificação de Valores Faltantes
Análise por variável
| Variavel | Missing | Existing | Prop_miss | Prop_exis |
|---|---|---|---|---|
| PUERPERA | 215 | 2172 | 9.01 | 90.99 |
| CARDIOPATI | 97 | 2290 | 4.06 | 95.94 |
| HEMATOLOGI | 197 | 2190 | 8.25 | 91.75 |
| SIND_DOWN | 194 | 2193 | 8.13 | 91.87 |
| HEPATICA | 201 | 2186 | 8.42 | 91.58 |
| ASMA | 193 | 2194 | 8.09 | 91.91 |
| DIABETES | 121 | 2266 | 5.07 | 94.93 |
| NEUROLOGIC | 190 | 2197 | 7.96 | 92.04 |
| PNEUMOPATI | 192 | 2195 | 8.04 | 91.96 |
| IMUNODEPRE | 204 | 2183 | 8.55 | 91.45 |
| RENAL | 199 | 2188 | 8.34 | 91.66 |
| OBESIDADE | 203 | 2184 | 8.50 | 91.50 |
| OUT_MORBI | 122 | 2265 | 5.11 | 94.89 |
Análise por Paciente
listando todos os pacientes que apresentaram valores “missing” em pelo menos um variável:
| n_paciente | sexo | idade | Q_variaveis |
|---|---|---|---|
| 10 | F | 50 | 1 |
| 20 | M | 90 | 12 |
| 40 | M | 93 | 11 |
| 47 | M | 63 | 1 |
| 53 | M | 71 | 1 |
| 56 | M | 67 | 2 |
| 68 | F | 55 | 2 |
| 83 | M | 50 | 1 |
| 88 | F | 92 | 1 |
| 101 | M | 75 | 1 |
| 104 | F | 55 | 12 |
| 139 | M | 22 | 1 |
| 162 | M | 83 | 1 |
| 176 | F | 61 | 1 |
| 201 | M | 64 | 2 |
| 203 | M | 83 | 2 |
| 219 | F | 76 | 12 |
| 242 | F | 88 | 1 |
| 258 | F | 75 | 1 |
| 276 | M | 62 | 12 |
Por Variável
| Variavel | Categoria | Frequencia | Percent |
|---|---|---|---|
| sexo | F | 1084 | 45.41 |
| sexo | M | 1303 | 54.59 |
| trimestre | Q3 | 1387 | 58.11 |
| trimestre | Q4 | 1000 | 41.89 |
| semestre | 2 | 2387 | 100.00 |
| FATOR_RISC | anao | 472 | 19.77 |
| FATOR_RISC | sim | 1915 | 80.23 |
| PUERPERA | anao | 2162 | 90.57 |
| PUERPERA | sim | 10 | 0.42 |
| CARDIOPATI | anao | 1315 | 55.09 |
| CARDIOPATI | sim | 975 | 40.85 |
| HEMATOLOGI | anao | 2171 | 90.95 |
| HEMATOLOGI | sim | 19 | 0.80 |
| SIND_DOWN | anao | 2172 | 90.99 |
| SIND_DOWN | sim | 21 | 0.88 |
| HEPATICA | anao | 2157 | 90.36 |
| HEPATICA | sim | 29 | 1.21 |
| ASMA | anao | 2149 | 90.03 |
| ASMA | sim | 45 | 1.89 |
| DIABETES | anao | 1453 | 60.87 |
| DIABETES | sim | 813 | 34.06 |
| NEUROLOGIC | anao | 2034 | 85.21 |
| NEUROLOGIC | sim | 163 | 6.83 |
| PNEUMOPATI | anao | 2091 | 87.60 |
| PNEUMOPATI | sim | 104 | 4.36 |
| IMUNODEPRE | anao | 2109 | 88.35 |
| IMUNODEPRE | sim | 74 | 3.10 |
| RENAL | anao | 2041 | 85.50 |
| RENAL | sim | 147 | 6.16 |
| OBESIDADE | anao | 1972 | 82.61 |
| OBESIDADE | sim | 212 | 8.88 |
| OUT_MORBI | anao | 1109 | 46.46 |
| OUT_MORBI | sim | 1156 | 48.43 |
| EVOLUCAO | 0 | 733 | 30.71 |
| EVOLUCAO | 1 | 1654 | 69.29 |
Histograma
Analise dos Pacientes por sexo
Regressão Logística
Modelo Inicial
##
## Call:
## glm(formula = factor(EVOLUCAO) ~ . - semestre, family = binomial(link = "logit"),
## data = dados)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -2.2682 -1.0918 0.6320 0.8242 1.9672
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -1.377513 0.211766 -6.505 7.78e-11 ***
## sexoM 0.018734 0.104733 0.179 0.85803
## idade 0.034441 0.002935 11.735 < 2e-16 ***
## trimestreQ4 -0.523527 0.105596 -4.958 7.13e-07 ***
## FATOR_RISCsim 0.485626 0.175383 2.769 0.00562 **
## PUERPERAsim -0.497798 0.716382 -0.695 0.48713
## CARDIOPATIsim -0.363295 0.125571 -2.893 0.00381 **
## HEMATOLOGIsim 0.210337 0.602683 0.349 0.72709
## SIND_DOWNsim -0.123319 0.527201 -0.234 0.81505
## HEPATICAsim 0.465119 0.526446 0.884 0.37696
## ASMAsim 0.303306 0.395458 0.767 0.44310
## DIABETESsim -0.152809 0.123486 -1.237 0.21592
## NEUROLOGICsim -0.261552 0.212357 -1.232 0.21808
## PNEUMOPATIsim -0.369237 0.248720 -1.485 0.13766
## IMUNODEPREsim 0.063816 0.289836 0.220 0.82573
## RENALsim 0.539884 0.250486 2.155 0.03114 *
## OBESIDADEsim 0.042508 0.184197 0.231 0.81749
## OUT_MORBIsim 0.182081 0.123883 1.470 0.14162
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 2456.5 on 2015 degrees of freedom
## Residual deviance: 2240.1 on 1998 degrees of freedom
## (371 observations deleted due to missingness)
## AIC: 2276.1
##
## Number of Fisher Scoring iterations: 4Modelo Ajustado p < 0.20
##
## Call:
## glm(formula = factor(EVOLUCAO) ~ idade + trimestre + FATOR_RISC +
## CARDIOPATI + PNEUMOPATI + RENAL + RENAL + OUT_MORBI, family = binomial(link = "logit"),
## data = dados)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -2.2724 -1.0977 0.6388 0.8233 1.9569
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -1.339169 0.192590 -6.953 3.56e-12 ***
## idade 0.033919 0.002805 12.092 < 2e-16 ***
## trimestreQ4 -0.517848 0.102243 -5.065 4.09e-07 ***
## FATOR_RISCsim 0.443149 0.155457 2.851 0.00436 **
## CARDIOPATIsim -0.366534 0.120638 -3.038 0.00238 **
## PNEUMOPATIsim -0.338375 0.241346 -1.402 0.16091
## RENALsim 0.457318 0.231470 1.976 0.04819 *
## OUT_MORBIsim 0.142990 0.120128 1.190 0.23392
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 2560.9 on 2102 degrees of freedom
## Residual deviance: 2343.9 on 2095 degrees of freedom
## (284 observations deleted due to missingness)
## AIC: 2359.9
##
## Number of Fisher Scoring iterations: 4Envelope de simulação
## Binomial model
## Simulation 1 out of 99
## Simulation 2 out of 99
## Simulation 3 out of 99
## Simulation 4 out of 99
## Simulation 5 out of 99
## Simulation 6 out of 99
## Simulation 7 out of 99
## Simulation 8 out of 99
## Simulation 9 out of 99
## Simulation 10 out of 99
## Simulation 11 out of 99
## Simulation 12 out of 99
## Simulation 13 out of 99
## Simulation 14 out of 99
## Simulation 15 out of 99
## Simulation 16 out of 99
## Simulation 17 out of 99
## Simulation 18 out of 99
## Simulation 19 out of 99
## Simulation 20 out of 99
## Simulation 21 out of 99
## Simulation 22 out of 99
## Simulation 23 out of 99
## Simulation 24 out of 99
## Simulation 25 out of 99
## Simulation 26 out of 99
## Simulation 27 out of 99
## Simulation 28 out of 99
## Simulation 29 out of 99
## Simulation 30 out of 99
## Simulation 31 out of 99
## Simulation 32 out of 99
## Simulation 33 out of 99
## Simulation 34 out of 99
## Simulation 35 out of 99
## Simulation 36 out of 99
## Simulation 37 out of 99
## Simulation 38 out of 99
## Simulation 39 out of 99
## Simulation 40 out of 99
## Simulation 41 out of 99
## Simulation 42 out of 99
## Simulation 43 out of 99
## Simulation 44 out of 99
## Simulation 45 out of 99
## Simulation 46 out of 99
## Simulation 47 out of 99
## Simulation 48 out of 99
## Simulation 49 out of 99
## Simulation 50 out of 99
## Simulation 51 out of 99
## Simulation 52 out of 99
## Simulation 53 out of 99
## Simulation 54 out of 99
## Simulation 55 out of 99
## Simulation 56 out of 99
## Simulation 57 out of 99
## Simulation 58 out of 99
## Simulation 59 out of 99
## Simulation 60 out of 99
## Simulation 61 out of 99
## Simulation 62 out of 99
## Simulation 63 out of 99
## Simulation 64 out of 99
## Simulation 65 out of 99
## Simulation 66 out of 99
## Simulation 67 out of 99
## Simulation 68 out of 99
## Simulation 69 out of 99
## Simulation 70 out of 99
## Simulation 71 out of 99
## Simulation 72 out of 99
## Simulation 73 out of 99
## Simulation 74 out of 99
## Simulation 75 out of 99
## Simulation 76 out of 99
## Simulation 77 out of 99
## Simulation 78 out of 99
## Simulation 79 out of 99
## Simulation 80 out of 99
## Simulation 81 out of 99
## Simulation 82 out of 99
## Simulation 83 out of 99
## Simulation 84 out of 99
## Simulation 85 out of 99
## Simulation 86 out of 99
## Simulation 87 out of 99
## Simulation 88 out of 99
## Simulation 89 out of 99
## Simulation 90 out of 99
## Simulation 91 out of 99
## Simulation 92 out of 99
## Simulation 93 out of 99
## Simulation 94 out of 99
## Simulation 95 out of 99
## Simulation 96 out of 99
## Simulation 97 out of 99
## Simulation 98 out of 99
## Simulation 99 out of 99
ODDS Ratio
dados$sexo=as.factor(dados$sexo)
dados$trimestre=as.factor(dados$trimestre)
dados$semestre=as.factor(dados$semestre)
dados$FATOR_RISC=as.factor(dados$FATOR_RISC)
dados$PUERPERA=as.factor(dados$PUERPERA)
dados$CARDIOPATI=as.factor(dados$CARDIOPATI)
dados$HEMATOLOGI=as.factor(dados$HEMATOLOGI)
dados$SIND_DOWN=as.factor(dados$SIND_DOWN)
dados$HEPATICA=as.factor(dados$HEPATICA)
dados$ASMA=as.factor(dados$ASMA)
dados$DIABETES=as.factor(dados$DIABETES)
dados$NEUROLOGIC=as.factor(dados$NEUROLOGIC)
dados$PNEUMOPATI=as.factor(dados$PNEUMOPATI)
dados$IMUNODEPRE=as.factor(dados$IMUNODEPRE)
dados$RENAL=as.factor(dados$RENAL)
dados$OBESIDADE=as.factor(dados$OBESIDADE)
dados$OUT_MORBI=as.factor(dados$OUT_MORBI)
dados$EVOLUCAO=as.factor(dados$EVOLUCAO)
OR <- glm(EVOLUCAO~idade+trimestre+FATOR_RISC+CARDIOPATI+PNEUMOPATI+RENAL+RENAL+OUT_MORBI,
family=binomial, data=dados)
logistic.display(OR)##
## Logistic regression predicting EVOLUCAO : 1 vs 0
##
## crude OR(95%CI) adj. OR(95%CI) P(Wald's test)
## idade (cont. var.) 1.03 (1.03,1.04) 1.03 (1.03,1.04) < 0.001
##
## trimestre: Q4 vs Q3 0.66 (0.55,0.8) 0.6 (0.49,0.73) < 0.001
##
## FATOR_RISC: sim vs anao 1.78 (1.44,2.21) 1.56 (1.15,2.11) 0.004
##
## CARDIOPATI: sim vs anao 1.17 (0.96,1.41) 0.69 (0.55,0.88) 0.002
##
## PNEUMOPATI: sim vs anao 0.93 (0.6,1.46) 0.71 (0.44,1.14) 0.161
##
## RENAL: sim vs anao 1.7 (1.1,2.62) 1.58 (1,2.49) 0.048
##
## OUT_MORBI: sim vs anao 1.56 (1.29,1.88) 1.15 (0.91,1.46) 0.234
##
## P(LR-test)
## idade (cont. var.) < 0.001
##
## trimestre: Q4 vs Q3 < 0.001
##
## FATOR_RISC: sim vs anao 0.004
##
## CARDIOPATI: sim vs anao 0.002
##
## PNEUMOPATI: sim vs anao 0.168
##
## RENAL: sim vs anao 0.041
##
## OUT_MORBI: sim vs anao 0.235
##
## Log-likelihood = -1171.9462
## No. of observations = 2103
## AIC value = 2359.8925