Interpreting results: Confounding
Please refer to the following text and tables to carry out this exercise.
In a study of the epidemiology of tuberculosis in Denmark, Horwitz (Am Rev Respir Dis 1974;104:22-31) showed that tuberculosis was more common in men than in women and that it was much more common in those who were unmarried than in those who had been married. The results of a study determining the relation of marital status to tuberculosis notification in Canada from 1970-1972 is shown in table 1.
Table 1. Active tuberculosis (number of cases and average annual rate per 100 000) by gender and marital status, Canada 1970-1972
|
|
Women
|
|
Men
|
Age group
|
|
Single
|
Married / Separated
|
Widowed / Divorced
|
TOTAL
|
|
Single
|
Married / Separated
|
Widowed / Divorced
|
TOTAL
|
All
|
No
|
1 122
|
3 140
|
703
|
4 965
|
|
2 113
|
4 641
|
640
|
7 394
|
|
Rate
|
19,5
|
21,4
|
26,3
|
21,5
|
|
29,7
|
31,7
|
80,0
|
32,8
|
Grzybowski S. Evaluation of the Tuberculosis Problem and Control Measures in Canada. National Sanatorium Association, 1977
In table 2, the same data are displayed, in this case showing the distribution by age in addition to gender and marital status.
Table 2. Active tuberculosis by age, gender and marital status, Canada 1970-1972
|
|
Women
|
|
Men
|
Age group
|
|
Single
|
Married / Separated
|
Widowed / Divorced
|
ALL
|
|
Single
|
Married / Separated
|
Widowed / Divorced
|
ALL
|
15-24
|
No
|
549
|
281
|
4
|
834
|
|
638
|
129
|
|
767
|
|
Rate
|
13,3
|
15,5
|
13,8
|
14,0
|
|
12,6
|
13,5
|
|
12,7
|
25-34
|
No
|
229
|
732
|
44
|
983
|
|
372
|
565
|
13
|
950
|
|
Rate
|
42,6
|
20,1
|
21,3
|
22,9
|
|
42.3
|
16,4
|
23,1
|
21,7
|
35-44
|
No
|
104
|
720
|
47
|
871
|
|
300
|
818
|
35
|
1 153
|
|
Rate
|
39,3
|
21,8
|
29,7
|
23,4
|
|
79,0
|
24,1
|
44,1
|
29,9
|
45-54
|
No
|
74
|
607
|
77
|
758
|
|
300
|
1 038
|
85
|
1 423
|
|
Rate
|
29,0
|
20,8
|
25,5
|
21,8
|
|
99.0
|
34,7
|
83,3
|
41,9
|
55-64
|
No
|
71
|
432
|
116
|
619
|
|
247
|
1 049
|
110
|
1 406
|
|
Rate
|
28,3
|
23,2
|
22,3
|
23,5
|
|
102,2
|
47,9
|
84,1
|
54.9
|
65+
|
No
|
95
|
368
|
415
|
878
|
|
256
|
1 042
|
397
|
1 695
|
|
Rate
|
30,8
|
32,5
|
28.7
|
30,4
|
|
102,9
|
61,9
|
96,3
|
72,3
|
-
In table 1, is there a difference in notification rate according to marital status?
-
-
Which group has the highest rate (specify one group by the 3 variables)?
_____________(age)__________(gender)__________(marital)
-
Why do you think this is so?
-
-
In table 2, are the conclusions you drew from table 1 valid? Yes [ ] No [ ]
-
Which group in the table contributes the largest number of cases?
_____________(age)__________(gender)__________(marital)
-
Which group has the highest risk for tuberculosis?
_____________(age)__________(gender)__________(marital)
-
Why is there a difference?
-
-
How would you modify the results displayed in table 1 to reflect reality?
-
One decade later, we further examined this association in the city of Vancouver. The results are presented in table 3.
-
In table 3, are the conclusions you drew from table 2 valid? Yes [ ] No [ ]
-
What is the cause of the difference in rates between ever married and never married men?
-
Table 3. Active tuberculosis(per 100 000 per year) in Vancouver 1980-1982 in adult Canadian-born men according to socioeconomic status and marital status.
|
Ever married
|
|
Never married
|
Socioeconomic level
|
Number
|
Rate
|
|
Number
|
Rate
|
TOTAL
|
43
|
22
|
|
42
|
37
|
High
|
3
|
3
|
|
5
|
12
|
Middle
|
21
|
22
|
|
13
|
20
|
Low
|
19
|
299
|
|
24
|
326
|
Enarson DA, et al. Am J Epidemiol 1989;129:1268
Practical exercise
Interpreting results: Standardization of rates
The following table gives the situation concerning prevalence of significant skin reactions to tuberculin in residents of Honduras and of Canada, along with their populations:
Prevalence of significant tuberculin reactions in Honduras and Canada
-
|
|
Canada
|
|
|
|
Honduras
|
| Age |
population
|
cases
|
rate (%)
|
|
population
|
cases
|
Rate
(%)
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
TOTAL
|
27,000,000
|
8,100,000
|
30
|
|
3,510,000
|
1,072,275
|
31
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
0-14
|
1,890,000
|
0
|
0
|
|
1,120,000
|
43,971
|
4
|
|
|
|
|
|
|
|
|
15-24
|
5,130,000
|
0
|
0
|
|
910,000
|
115,379
|
13
|
|
|
|
|
|
|
|
|
25-34
|
6,210,000
|
0
|
0
|
|
700,000
|
295,400
|
42
|
|
|
|
|
|
|
|
|
35-44
|
4,320,000
|
540,000
|
13
|
|
390,000
|
262,681
|
67
|
|
|
|
|
|
|
|
|
45-54
|
3,780,000
|
1,890,000
|
50
|
|
210,000
|
183,204
|
87
|
|
|
|
|
|
|
|
|
55-64
|
3,510,000
|
3,510,000
|
94
|
|
110,000
|
101,640
|
92
|
|
|
|
|
|
|
|
|
65+
|
2,160,000
|
2,160,000
|
100
|
|
70,000
|
70,000
|
100
|
|
|
|
|
|
|
|
|
Calculate the rate of tuberculin skin reactivity in the two populations, standardized for age and enter into the table.
What is the explanation of the difference between the crude rate and the standardized rate?
What are the advantages of the standardized rates over the crude rates?
What are the disadvantages of presenting only the standardized rates?
Dostları ilə paylaş: |