Download as xls, pdf, or txt
Download as xls, pdf, or txt
You are on page 1of 10

TWOXTWO - calculates sensitivity, specificity, probabilities and

predictive values from 2X2 tables. Enter data in C8, C9, D8 & D9.
[Press F9 to calculate]
Actual Health Status
Test Result
D+
D------------- ------------- ------------- ------------- ------------- ------------T+
|
40
10
|
50
T|
10
490
|
500
------------- ------------- ------------- ------------- ------------- ------------50
500
|
550
Prior Prob. of dis. (True Prevalence) = p(D+) = ...................
Sensitivity (ability ot detect dis.) = p(T+/D+) = ...................
Specificity (ability to detect non-dis.) = p(T-/D-) = ..........
Post-test Prob. of dis. given T+ = p(D+/T+) = ................
(Predictive Value of a Positive Test)
Post-test Prob. of dis. given T- = p(D+/T-) = .................
(1 - Predictive Value of a Negative Test)
Apparent Prevalence = p(T+) = ............................

0.091
0.800
0.980
0.800
0.020
0.091

CHISQ - Calculates rates and chisquare statistic for cross-sectional,


case-control and cohort studies.
Enter the study type in cell L4 & data in 2x2 TABLE cells D12, E12,
D14 & E14. PRESS F9 to calculate rates and chiquare statistic.
------------- ------------- - ------------- ------------- - ------------- ---- ---- ---------- ------------------- ------------- ------------What is the study type ? ...................................
3
(Enter 1 for cross-sectional, 2 for case-control, 3 for cohort)
======== ======== ========= ======== = ======== == == ====== =========== ======== ========
2x2 TABLE
DISEASE
F
|
+ve
-ve |
Totals
Rates p(D+/F)
A
------------- | ------------- ------------- | --------------------C
+ve
|
140
100 |
240
0.583
T
-ve
|
60
100 |
160
0.375
O
------------- | ------------- ------------- | --------------------R
Totals
|
200
200 |
400
The corrected chi-square
statistic is :
Odds Ratio :

15.844
2.333

ASSOC - MEASURES OF ASSOCIATION


------------- ------------- ------------- ------------- ------------- --- ------------- ------------CASE-CONTROL STUDY:
Disease
|
-----------------+ve
-ve | Totals
------------- ------------- |- -------Factor
+ve
44
706 |
750
-ve
28
2382 |
2410
------------- ------------- |- -------Totals
72
3088 |
3160
Rates p(F+/D)
0.611
0.229 |
[Press F9 to calculate / Press <Alt>A to return to the menu]
MEASURES OF ASSOCIATION
------------- ------------- ------------- ------------Odds Ratio
5.302
Pop. OR
1.984
Est. AF
Est. PAF

CHISQ (X)
--- ------------- ------------Uncorrected
56.865
Corrected
54.772

0.811
0.496

[see below for confidence intervals ]


------------- ------------- ------------- ------------- ------------- --- ------------- ------------- ------------CONFIDENCE INTERVALS
-------------------a) Precision based
| b) Test based
se(Q)or(ln(Q))
lower
upper
| X(uncor) lower
upper
------------- ------------- ------------- ------------- --- ------------- ------------- ------------OR
0.245515
3.280
8.570 |
56.865
3.437
8.179
Pop. OR 0.224381
1.279
3.076 |
1.660
2.370
|
Est. AF
n/a
0.695
0.883 |
0.600
1.022
Est. PAF 0.074644 **
0.350
0.642 |
0.367
0.625
** based on Walter, SD., Int J Epi, 1978; 7:175-182.
Note: test based intervals are reasonable if .2<OR<4.
------------- ------------- ------------- ------------- ------------- --- ------------- ------------- -------------

COHORT STUDIES:
-------------Factor

+ve
-ve
Totals

Disease
|
+ve
-ve |
------------- ------------- |120
80 |
30
270 |
------------- ------------- |150
350 |

Totals
Rates p(D+/F)
------------- ------------200
0.6000
300
0.1000
------------- ------------500

[Press F9 to calculate / Press <Alt>A to return to the menu]


MEASURES OF ASSOCIATION
----------------------------------Relative risk =
6.000
Odds ratio =
13.500
Attributable rate =
Attributable fraction =

CHISQ (X)
------------------Uncorrected
Corrected

142.857
140.486

0.500
0.833

[see below for confidence intervals - press <Pg Dn>]


------------- ------------- ------------- ------------- ------------- ----- ------------- ------------- ------------CONFIDENCE INTERVALS
-------------------a) Precision based
|
b) Test based
se(Q)or(ln(Q))
lower
upper |
X(uncor) lower
upper
------------- ------------- ------------- ------------- ----- ------------- ------------- ------------RR
0.18
4.20
8.58 |
142.86
4.47
8.05
OR
0.24
8.42
21.63 |
8.81
20.69
|
AR
0.04
0.42
0.58 |
0.42
0.58
AF
N/A
0.71
1.18 |
0.70
0.97
Note: test based intervals are reasonable if .2<OR<4.
------------- ------------- ------------- ------------- ------------- ----- ------------- ------------- -------------

CROSS-SECTIONAL STUDIES:
----------------------Factor

+ve
-ve
Totals

Rates p(F+/D)

Disease
+ve
-ve
------------- ------------140
100
60
100
------------- ------------200
200
0.70

|
|
||
|
||

Totals Rates p(D+/F)


------------- ------------240
0.5833
160
0.3750
------------- ------------400
0.5000

0.50

MEASURES OF ASSOCIATION
CHISQ (X)
-----------------------------------------------------------Relative risk =
1.556
Uncorrected =
16.667
Odds ratio =
2.333
Corrected =
15.844
Pop relative risk =
1.333
Pop odds ratio =
1.667
[Press F9 to calculate]
Attributable rate =
0.208
[Press <Alt>A to
Attributable fraction =
0.357
return to menu]
Pop Att. rate =
0.125
[See below for confidence
Pop Att. fraction =
0.250
intervals - <Pg Dn>]
------------- ------------- ------------- ------------- ------------- ---- ------------- ----------------------------CONFIDENCE INTERVALS
-------------------a) Precision based
b) Test based
se(Q)or(ln(Q))
lower
upper
X(uncor)
lower
upper
------------- ------------- ------------- ------------- ---- ------------- ------------- ------------RR
0.12
1.24
1.95
16.67
1.26
1.92
OR
0.21
1.55
3.52
1.55
3.50
Pop RR
0.11
1.07
1.67
1.16
1.53
Pop OR
0.19
1.15
2.43
1.30
2.13
AR
AF
Pop AR
Pop AF

0.05
N/A
0.05
#N/A

0.11
0.19

0.31
0.49

0.11
0.19

0.31
0.53

0.04
#N/A

0.21
#N/A

0.06
0.13

0.19
0.37

Confidence intervals for measures of association with explicit SEs are


calculated in the usual manner. Walter's formula (I. J. Epi. 7:175:182)
for SE(PAF) was used. Other CIs were calculated by mapping.
If the upper CL exceeded 100% for "fractions", 100% is used.
Note: test based intervals are reasonable if .2<OR<4.
------------- ------------- ------------- ------------- ------------- ---- ------------- ------------- -------------

KAPPA - measures the agreement between two tests.


Enter the data in cells C8, D8, C10, D10.
[Press F9 to calculate]
------------- ------------- ------------- ------------- ---- ---------- -- ------------- ------------TEST 2
T+
T- |
|
TEST 1 T+
3
2 |
5
T0
995 |
995
------------- ------------- |
---------3
997 |
1000
% Observed agreement .........................................
% Expected agreement ...........................................
% Actual agreement beyond chance ..................................
% Potential agreement beyond chance ............................
Kappa ..............................................

99.8
99.2
0.6
0.8
0.75

%
%
%
%

COMBINE - Sensitivity, specificity and positive predictive value of


combined tests with test results interpreted in SERIES and in PARALLEL.
Enter data in cells D8 to D11 and F8 to F11.
Enter True Prevalence, if known (as a proportion) : ...............
0
PRESS F9 TO CALCULATE.
PRESS ALT<E> TO ERASE.
------------- ------------- ------------- ------------- ------------- ------------- ------ ------------- --- ------------Test 1
Test 2
Diseased
Not Diseased
+
3
15
+
0
0
+
+
61
1
9
525
------------- ------------- ------------73
541
------------- ------------- ------------- ------------- ------------- ------------- ------ ------------- --- ------------Sensitivity
Specificity
PV+ve
Test 1
(3+61)/73 =
87.7%
(0+525)/541
=
97.0%
n/a
Test 2
(0+61)/73 =
83.6%
(15+525)/541
=
99.8%
n/a
Parallel
(3+0+61)/73
=
87.7%
(525)/541 =
97.0%
n/a
Series
(61)/73
=
83.6%
(15+0+525)/541
=
99.8%
n/a
------------- ------------- ------------- ------------- ------------- ------------- ------ ------------- --- -------------

SENSITIVITY, Test 1 =
SPECIFICITY, Test 1 =
SENSITIVITY, Test 2 =
SPECIFICITY, Test 2 =
In Parallel - SENSITIVITY
- SPECIFICITY
- PV +ve
In Series - SENSITIVITY
- SPECIFICITY
- PV +ve

DETECT : Sampling to detect disease


------------- ------------- ------------- ------------- ------------- ------------- ------------- ------------1 - Determines n of N animals to be examined to be 95/99/99.9% sure
that at least one diseased animal will be detected if disease is
present.
2 - Determines the expected maximum number of cases, with confidence a,
if n of N animals are examined and found to be negative.

DETECT-1
...determining the sample size(n) required to be 95/99/99.9% sure of
detecting disease present in a population(N)...
------------- ------------- ------------- ------------- ------------- ------------- ------------- ------------Note: enter prevalence and confidence level as proportions.
[Press F9 to calculate]
======== ======== ======== ======== ======== ======== ======== ========
Population size(N): .................................................
1200
Likely MINIMUM PREVALENCE if disease is present: ...........................
0.002
Likely NUMBER OF DISEASED ANIMALS(D): .................................
2.4
Required CONFIDENCE LEVEL(a) [ie. 0.95, 0.99, 0.999] : ...............
0.95
The required SAMPLE SIZE(n) is : .....................................
855
======== ======== ======== ======== ======== ======== ======== ========
A thumb-rule approach (THE RULE OF THREE) is as follows...
Enter 3 for 95%, 4.6 for 99% and 6.9 for 99.9% CONFIDENCE :
3
The required SAMPLE SIZE(n) is : ..............................
855
Adjusting this for SAMPLING FRACTION the SAMPLE SIZE(n) is :
499
======== ======== ======== ======== ======== ======== ======== ========

DETECT-2
...determining the max. no. of diseased animals(D) when a given sample
(n) is examined and found to be disease free...
------------- ------------- ------------- ------------- ------------- ------------- ------------- ------------Note: enter confidence level as a proportion.
[Press F9 to calculate]
======== ======== ======== ======== ======== ======== ======== ========
Population size(N): ..........................................
0
Required CONFIDENCE LEVEL(a) [ie. 0.95, 0.99, 0.999] : ...................
0
SAMPLE SIZE(n) : ..............................................
0
If n animals test -ve the MAX. NO. of diseased animals(D) is :
#DIV/0!
The MAX. PREVALENCE of disease in the pop. : .............................
#DIV/0!
======== ======== ======== ======== ======== ======== ======== ========
A thumb-rule approach (THE RULE OF THREE) is as follows...
Enter 3 for 95%, 4.6 for 99% and 6.9 for 99.9% CONFIDENCE : .......
0
If n animals test -ve the MAX. PREVALENCE of disease is : ............
#DIV/0!
======== ======== ======== ======== ======== ======== ======== ========

SAMSIZE1 - SURVEY SAMPLE SIZE


|
The formulae used assume simple random sampling.
|
[Press F9 to calculate]
|
======== ======== ======== ======== ======== ======== ======== ======== ======== |
QUANTITATIVE DATA
|
QUALITATIVE DATA CI 95%
|
Enter S, L & N in cells D8,
|
Enter P & L as proportions in
|
D9 & D13 respectively.
|
cells I9 & I10 and N in cell I13.
|
------------- ------------- ------------- ------------- |-------------- ------------- ------------- ------------- |
ESTIMATED VARIANCE(S) ?
25 |
EST. MEAN PROPORTION(P) ?
0.6 |
ALLOWABLE ERROR(L) ?
0.5 |
ALLOWABLE ERROR(L) ?
0.05 |
Required SAMPLE SIZE(n) :
400 |
Required SAMPLE SIZE(n) :
384 |
------------- ------------- ------------- ------------- |-------------- ------------- ------------- ------------- |
What is N ?
0 |
What is N ?
8000 |
|
0
|
|
0
|
======== ======== ======== ======== ======== ======== ======== ======== ========

#DIV/0!

|
|
|
|-|
|
|
|-|
|
|
|

QUALITATIVE DATA
CI 99%
Enter P & L as proportions in
cells I9 & I10 and N in cell I13.
------------- ------------- ------------- ------------EST. MEAN PROPORTION(P) ?
0.075
ALLOWABLE ERROR(L) ?
0.05
Required SAMPLE SIZE(n) :
250
------------- ------------- ------------- ------------What is N ?
7500
0
0

You might also like