Population attibutable fraction

[Kyra]

Hi,

I am trying to calculate the Population attibutable fraction ( outcome is head and neck cancer and exposure is occupation). Is there any way i can adjust for tobacco and alcohol too while calculating that?

I am pasting my code and the output below.

Neverexposed 0 means they were exposed to the risk factor occupation.

  Event_E is event in exposed( to occupation)

Count_E is number of exposed

Event_NE is event in not exposed

Count_NE is number of non-exposed

Thanks,

data red.arp;
   input  Event_E Count_E Event_NE Count_NE;
   datalines;
   6029  13218  2810  9315
;

ods graphics on;
proc stdrate data=red.arp
             refdata=red.arp
             method=indirect(af)
             stat=risk
             plots(stratum=horizontal)
             ;
   population event=Event_E  total=Count_E;
   reference  event=Event_NE total=Count_NE;
  
run;
ods graphics off;

frequency table just for your reference. Neverexposed 0 means they were exposed.

The FREQ Procedure

FrequencyPercentRow PctCol Pct

Table of neverexposed by caseneverexposed case(Status)1 2 Total01Total 6029 26.71 45.61 68.21 7189 31.85 54.39 52.36 13218 58.57     2810 12.45 30.05 31.79 6541 28.98 69.95 47.64 9351 41.43     8839 39.16 13730 60.84 22569 100.00

---

The SAS System

The STDRATE Procedure

Standardization InformationData SetReference Data SetMethodStatisticNumber of Strata

RED.ARP

RED.ARP

Indirect Standardization

Risk

1

---

The SAS System

The STDRATE Procedure

Standardized Morbidity/Mortality RatioObservedEvents ExpectedEvents SMR StandardError 95% Normal ConfidenceLimits Z Pr > |Z|

6029

3987.39

1.5120

0.0144

1.4839

1.5402

35.65

<.0001

Indirectly Standardized Risk EstimatesStudy Population ReferenceCrudeRisk ExpectedEvents SMR Standardized RiskObservedEvents Number ofObservations CrudeRisk Estimate StandardError 95% Normal ConfidenceLimits

6029

13218

0.4561

0.3017

3987.39

1.5120

0.4561

0.00433

0.4476

0.4646

Attributable Fraction EstimatesParameter Estimate 95% Confidence LimitsAttributable RiskPopulation Attributable Risk

0.33863	0.32609	0.35072
0.23098	0.21620	0.24547

[Kyra]

I think the answer to my question is the formula :  1-1/Adjusted OR.