Question

You want to know if there is a relationship between surgeons’ surgical experience and surgical site infection (SSI) rates. Since SSI’s are relatively rare (infections develop in roughly 1-3 of every 100 patients who have surgery), you decide to do a case control study in which you select hospital patients who experienced a SSI and hospital patients who did not, and compare patients’ exposures to surgeons with <10 years of experience, and >10 years of experience. Assume you have access to data collected by a hospital’s infection control department from 2008-2010.

You enroll a total of 400 patients in your study. From the infection control records, you select 200 patients who developed a SSI in the period 2008-2010. Out of those 200 patients, 36 had a surgeon with >10 years experience. Out of the total 400 patients, you found that 232 had a surgeon with <10 years experience.

A) Based on the above data, fill in Table 1, name and calculate the appropriate crude measure of association for this study.

Table 1.

	SSI	No SSI
Surgeon <10 yrs experience
Surgeon >10 yrs experience

After talking to the surgery department you learn that the surgeons with <10 years of experience tended to perform surgery on sicker patients. Because you are unsure whether a patients’ disease severity may influence the association between surgeon experience and SSI, you decide to stratify the results by disease severity.

The following 2x2 tables summarize the results of your stratification:

SEVERE DISEASE

	SSI	No SSI
Surgeon <10 yrs experience	162	54
Surgeon >10 yrs experience	18	6

NON-SEVERE DISEASE

	SSI	No SSI
Surgeon <10 yrs experience	2	14
Surgeon >10 yrs experience	18	126

                     

B) Calculate the appropriate measure of association for each stratum (ie. for severe disease and non-severe disease).

C) Compare the measures of association in parts a and b. Do you think that disease severity confounds the relationship between surgeon experience and SSI? Why, or why not? Do you think that this association between surgeon experience and surgical site infection is a true association or does something else explain the association? Justify your answer.

Answer 1

a)

H₀: "There is no association between Surgeon <10 yrs experience and Surgeon >10 yrs experience
H_A: "There is some association between Surgeon <10 yrs experience and Surgeon >10 yrs experience

	SSI	No SSI	Totals
Surgeon <10 yrs experience	164	68	232
Surgeon >10 yrs experience	36	132	168
Totals	200	200	400

Chi-Square Test for Association: C2, Worksheet columns

Rows: C2   Columns: Worksheet columns

	SSI	No SSI	All
Surgeon <10 yrs experience	164	68	232
	116	116
Surgeon >10 yrs experience	36	132	168
	84	84
All	200	200	400

Cell Contents
      Count
      Expected count

Chi-Square Test

	Chi-Square	DF	P-Value
Pearson	94.581	1	0.000
Likelihood Ratio	99.262	1	0.000

As p-value is less than 0.05 we reject null hypothesis and conclude that "There is some association between Surgeon <10 yrs experience and Surgeon >10 yrs experience".

the results of your stratification:

SEVERE DISEASE

Results
	Category 1	Category 2				Row Totals
Group 1	164  (163.95)  [0.00]	54  (54.05)  [0.00]				218
Group 2	18  (18.05)  [0.00]	6  (5.95)  [0.00]				24
Column Totals	182	60				242  (Grand Total)

The chi-square statistic is 0.0006. The p-value is .980297. The result is not significant at p < .05.

NON-SEVERE DISEASE :

Results
	Category 1	Category 2				Row Totals
Group 1	2  (2.00)  [0.00]	14  (14.00)  [0.00]				16
Group 2	18  (18.00)  [0.00]	126  (126.00)  [0.00]				144
Column Totals	20	140				160  (Grand Total)

The chi-square statistic is 0. The p-value is 1. The result is not significant at p < .05.