rel risk & ARR

[sam369]

Hi All,

I have following dataset. i want relative risk & absolute relative reduction? how can i achieve them using procedure?

data have;

  subgrp trt coun;

   1     1   72

   1     0    8

   2     1   74

   2     0    6

Want:

rr=(72/80)/(74/80)=0.97

ARR is by newcombe aprroach.

Output for ARR is : 0.03(-.07,0.12)

THanks

Sam

[Reeza]

Proc Freq - relrisk and riskdiff options, check the documentation for the exact methods/calcuations.

proc freq data=have;

table subgrp*trt/relrisk riskdiff;

weight count;

run;

[sam369]

Hi Reeza,

i tried both but haven't got the solution

thanks

sam

[Ksharp]

Are you sure ?

data have;
input subgrp trt count;
cards;
   1     1   72
   1     0    8
   2     1   74
   2     0    6
;
run;
proc freq data=have order=data;
table subgrp*trt/relrisk riskdiff;
weight count;
run;

                                  Column 1 Risk Estimates

                                                                (Asymptotic) 95%         (Exact) 95%

                                            Risk        ASE     Confidence Limits     Confidence Limits

                          -----------------------------------------------------------------------------

                          Row 1           0.9000     0.0335     0.8343     0.9657     0.8124     0.9558

                          Row 2           0.9250     0.0294     0.8673     0.9827     0.8439     0.9720

                          Total           0.9125     0.0223     0.8687     0.9563     0.8575     0.9513

                          Difference    -0.0250     0.0446    -0.1125     0.0625

                                                  Difference is (Row 1 - Row 2)

                                                     Column 2 Risk Estimates

                                                                (Asymptotic) 95%         (Exact) 95%

                                            Risk        ASE     Confidence Limits     Confidence Limits

                          -----------------------------------------------------------------------------

                          Row 1           0.1000     0.0335     0.0343     0.1657     0.0442     0.1876

                          Row 2           0.0750     0.0294     0.0173     0.1327     0.0280     0.1561

                          Total           0.0875     0.0223     0.0437     0.1313     0.0487     0.1425

                          Difference      0.0250     0.0446    -0.0625     0.1125

                                                  Difference is (Row 1 - Row 2)

                                              Statistics for Table of subgrp by trt

                                           Estimates of the Relative Risk (Row1/Row2)

                                Type of Study                   Value       95% Confidence Limits

                                -----------------------------------------------------------------

                                Case-Control (Odds Ratio)      0.7297        0.2412        2.2079

                                Cohort (Col1 Risk)             0.9730        0.8839        1.0711

                                Cohort (Col2 Risk)             1.3333        0.4846        3.6686

Xia Keshan

Message was edited by: xia keshan

[sam369]

TThank you ksharp!!!

now I got the relative risk and 95% ci

but how to calculate ARR(absolute relative reduced)?

expected result for arr:

.03 (-.07,0.12)

thanks

sam

[Ksharp]

Did you see :

Difference      0.0250     0.0446    -0.0625     0.1125

Xia Keshan