azprocedure: azprocedure

In COUNT: Functions, Data and Code for Count Data

Description

Data come from the 1991 Arizona cardiovascular patient files. A subset of the fields was selected to model the differential length of stay for patients entering the hospital to receive one of two standard cardiovascular procedures: CABG and PTCA. CABG is the standard acronym for Coronary Artery Bypass Graft, where the flow of blood in a diseased or blocked coronary artery or vein has been grafted to bypass the diseased sections. PTCA, or Percutaneous Transluminal Coronary Angioplasty, is a method of placing a balloon in a blocked coronary artery to open it to blood flow. It is a much less severe method of treatment for those having coronary blockage, with a corresponding reduction in risk.

Usage

data(azprocedure)

Format

A data frame with 3589 observations on the following 6 variables.

los

length of hospital stay

procedure

1=CABG;0=PTCA

sex

1=Male; 0=female

admit

1=Urgent/Emerg; 0=elective (type of admission)

age75

1= Age>75; 0=Age<=75

hospital

encrypted facility code (string)

Details

azprocedure is saved as a data frame. Count models use los as response variable. 0 counts are structurally excluded

Source

1991 Arizona Medpar data, cardiovascular patient files, National Health Economics & Research Co.

References

Hilbe, Joseph M (2014), Modeling Count Data, Cambridge University Press Hilbe, Joseph M (2007, 2011), Negative Binomial Regression, Cambridge University Press Hilbe, Joseph M (2009), Logistic Regression Models, Chapman & Hall/CRC

Examples

library(MASS)
library(msme)

data(azprocedure)

glmazp <- glm(los ~ procedure + sex + admit, family=poisson, data=azprocedure)
summary(glmazp)
exp(coef(glmazp))

nb2 <- nbinomial(los ~ procedure + sex + admit, data=azprocedure)
summary(nb2)
exp(coef(nb2))

glmaznb <- glm.nb(los ~ procedure + sex + admit, data=azprocedure)
summary(glmaznb)
exp(coef(glmaznb))

Example output

Loading required package: msme
Loading required package: MASS
Loading required package: lattice
Loading required package: sandwich

Call:
glm(formula = los ~ procedure + sex + admit, family = poisson, 
    data = azprocedure)

Deviance Residuals: 
    Min       1Q   Median       3Q      Max  
-3.1987  -1.1451  -0.4756   0.5331  12.4784  

Coefficients:
            Estimate Std. Error z value Pr(>|z|)    
(Intercept)  1.49140    0.01539   96.91   <2e-16 ***
procedure    0.95738    0.01218   78.61   <2e-16 ***
sex         -0.13022    0.01179  -11.04   <2e-16 ***
admit        0.33307    0.01210   27.52   <2e-16 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

(Dispersion parameter for poisson family taken to be 1)

    Null deviance: 16265.0  on 3588  degrees of freedom
Residual deviance:  8968.9  on 3585  degrees of freedom
AIC: 22483

Number of Fisher Scoring iterations: 5

(Intercept)   procedure         sex       admit 
  4.4433339   2.6048728   0.8779003   1.3952480 
There were 50 or more warnings (use warnings() to see the first 50)

Call:
ml_glm2(formula1 = formula1, formula2 = formula2, data = data, 
    family = family, mean.link = mean.link, scale.link = scale.link, 
    offset = offset, start = start, verbose = verbose)

Deviance Residuals:
   Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
-2.0474 -0.8062 -0.3166 -0.1451  0.3387  6.4689 

Pearson Residuals:
     Min.   1st Qu.    Median      Mean   3rd Qu.      Max. 
-1.505133 -0.709226 -0.301576 -0.002558  0.354577 13.195139 

Coefficients (all in linear predictor):
              Estimate      SE     Z         p    LCL     UCL
(Intercept)      1.451 0.02318 62.61         0  1.406  1.4968
procedure        0.978 0.01847 52.97         0  0.942  1.0143
sex             -0.132 0.01915 -6.91  4.75e-12 -0.170 -0.0948
admit            0.379 0.01913 19.79  3.83e-87  0.341  0.4160
(Intercept)_s    0.163 0.00657 24.78 1.63e-135  0.150  0.1757

Null deviance: 6645.529  on  3587 d.f.
Residual deviance: 3527.73  on  3584 d.f.
Null Pearson: 7994.057  on  3587 d.f.
Residual Pearson: 4928.398  on  3584 d.f.
Dispersion: 1.375111
AIC:  19992.18

Number of optimizer iterations:  187 

  (Intercept)     procedure           sex         admit (Intercept)_s 
    4.2691376     2.6593510     0.8760067     1.4601076     1.1768777 

Call:
glm.nb(formula = los ~ procedure + sex + admit, data = azprocedure, 
    init.theta = 6.140143318, link = log)

Deviance Residuals: 
    Min       1Q   Median       3Q      Max  
-2.0474  -0.8062  -0.3166   0.3387   6.4690  

Coefficients:
            Estimate Std. Error z value Pr(>|z|)    
(Intercept)  1.45141    0.02293   63.31  < 2e-16 ***
procedure    0.97808    0.01837   53.25  < 2e-16 ***
sex         -0.13238    0.01913   -6.92 4.52e-12 ***
admit        0.37851    0.01906   19.86  < 2e-16 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

(Dispersion parameter for Negative Binomial(6.1401) family taken to be 1)

    Null deviance: 6645.6  on 3588  degrees of freedom
Residual deviance: 3527.8  on 3585  degrees of freedom
AIC: 19992

Number of Fisher Scoring iterations: 1


              Theta:  6.140 
          Std. Err.:  0.248 

 2 x log-likelihood:  -19982.184 
(Intercept)   procedure         sex       admit 
   4.269141    2.659351    0.876007    1.460107 

COUNT documentation built on May 2, 2019, 2:37 a.m.