|
|
Assessment of Iodine
Deficiency Disorders and Monitoring their Elimination : A guide for programme
managers, Second Edition. ICCIDD/UNCF/ WHO, 2001
Continued . . .
Method
3 - an extremely large number of schools
In
very large populations, it may not be possible or efficient to select schools using either the PPS or the
systematic selection method. For
example, Szechwan Province in China has a
population of approximately 100 million.
Even if a list of schools were available at the provincial level, it would take
much time and effort to select schools using either of
these methods.
Accordingly, another
approach may be more
appropriate. First, select
districts using th PPS method. Develop a listing of the districts, their populatioins, and
cumulative populations similar to
the PPS selection described
earlier. Next, determine
the number of schools to survey,
based on the cumulative population using
PPS.
For districts with one or more clusters to
be selected, select schools in
each district using a random
number table. For example,
if a district has 200 schools,
number them from 1 to 200.
Then, randomly select a number
from 1 tp 200 using
the table. If two schools are to
be selected, then randomly select
two numbers. Finally, and while not technically
correct, it would be acceptable to analyse the school-based
data as through the schools were selected
using PPS methodology.
Other possibilities
In situations
where male and female children
attend the same school,
the selection of schools and
pupils would be same
as discussed above. In situations where males and females
attend separate schools, when a
school of one sex is
selected the nearest school of
the opposite sex should also be surveyed.
For example, a survey is to be performed in an
area where males and females attend separate schools. Thirty schools are to be
selected, and twenty pupils sampled in
each. When an
all-male school is visited,
inforamtion should be collected on
ten male pupils. Then,
the nearest female
school is visited,
and information collected on ten female pupils.
Reference
Adapted from:
Sullivan KM, May S, Maberly G.
Urinary iodine assessment: a manual on survey and laboratory
methods, 2nd ed. UNICEF, PAMM, 2000.
Annex
5 Summarizing urinary iodine data: a
worked example
Some
actual urinary iodine data from schoolchildren in Cameroon, following the implementation of universal salt iodization,
are presented in the first (left) column of Table 14. The data
have been entered into a
spreadsheet on a personal computer for
ease of calculation. However,
with small numbers such as these, the calculations are relatively easily
performed by hand.
Steps in processing the data
1. Before proceeding, carefully check the data
entered against the original. Ensure that the same number of data points
(n) are present, and look for any anomalous results.
2. Next, sort the data from highlight to lowest,
or vice-versa. The spreadsheet will do this
automatically. (In Microsoft Excel,
use the
Data Analysis function on the Tools
menu, and select "Rank and Percentile".) The sorted data are shown under "value" in Table 14, starting with the highest value.
The next
columns show the rank and percentile for
each data point.
3. The median is the middle value of the ranked
data. In
other words, it is the value of the (n + 1) / 2th value. In this case, there are 98 data points, so
the median is the value of (98 + 1)
divided by 2 = 49.5th data point. Accordingly, use the
middle point between the
49th and 50th values: 122
and 121 ug/l, respectively. The
mid-point is 121.5 ug/l, so
the median is 121.5 ug/l.
4. Next, calculate the number of values below
100, 50, and 20 ug/l, respectively. The ranking will allow this to be done very easily.
In this case, there are 33 values
below 100 ug/l,
6 below 50 ug/l, and one below 20 ug/l. These should be calculated as percentages: 33 of 98 is 33.7%,
5 of 98 is 5.1% and 1 of 98 is 1.0%.
5. Check if any values are above 500 ug/l. There is one (1.0%).
6. The 20th and 80th percentiles may are
readily observed, or automatically displayed
using the PERCENTILE function [PERCENTILE (range of cells, 0.2)]. The 20th percentile (P20) is 82.4 ug/l and P80 is 191.8 ug/l.
7. The
"Descriptive Statistics" function of Data Analysis
in Excel provides all statistics shown: select "Summary Statistics" in the dialogue box. Note that the mean is much higher than the median, indicating that the distribution
is heavily skewed to the right. This is
also shown by the much greater distance
between P80 and the median, compared to that between P20 and the median.
8. In addition, the data can be shown as a
histogram using the
"Histogram" function of
Data Analysis in Excel.
Convenient ranges need to be chosen
for making the frequency distribution,
which will be
reflected in the height of
each bar of the
histogram. 50 ug/l is suggested (i.e., the first bar is
0-49 ug/l, the second 50-99 ug/l, the third 100-149 ug/l, and so on).
Appropriate modifications can be
made using "Chart Options" and
related functions. The histogram is
shown in Figure 4. A fully detailed description for constructing
that histogram is not given here.
Table
13: Summary of results
|
Number |
98 |
|
Median |
121.5 ug/l |
|
20th percentile |
82.4 ug/l |
|
80th percentile |
191.8 ug/l |
|
|
|
|
Distribution: |
|
|
<100 ug/l |
33.7% |
|
<50 ug/l |
5.1% |
|
<20 ug/l |
1.0% |
|
>500 ug/l |
1.0% |
Table
14: Urinary iodine data in Cameroon schoolchildren following salt iodization
|
Ul(ug/l) |
Value |
Rank |
Percent |
Descriptive |
Statistics |
|
141 |
535 |
1 |
100.00% |
|
|
|
138 |
480 |
2 |
98.90% |
Mean |
142.7449 |
|
138 |
395 |
3 |
97.90% |
Standard error |
8.877338 |
|
154 |
340 |
4 |
96.90% |
Median |
121.5 |
|
162 |
320 |
5 |
95.80% |
Mode |
138 |
|
26 |
295 |
6 |
94.80% |
St. dev. |
87.88117 |
|
63 |
273 |
7 |
92.70% |
Sample variance |
7723.099 |
|
111 |
273 |
7 |
92.70% |
kurtosis |
5.463542 |
|
120 |
264 |
9 |
91.70% |
Skewness |
1.970291 |
|
65 |
261 |
10 |
90.70% |
Range |
525 |
|
190 |
240 |
11 |
89.60% |
Minimum |
10 |
|
142 |
232 |
12 |
87.60% |
Maximum |
535 |
|
138 |
232 |
12 |
87.60% |
Sum |
13989 |
|
95 |
224 |
14 |
86.50% |
Count |
98 |
|
273 |
208 |
15 |
85.50% |
Confidence |
|
|
132 |
200 |
16 |
83.50% |
level (95.0%) |
17.61905 |
|
164 |
200 |
16 |
83.50% |
|
|
|
66 |
198 |
18 |
82.40% |
|
|
|
158 |
193 |
19 |
80.40% |
|
|
|
114 |
193 |
19 |
80.40% |
|
|
|
118 |
190 |
21 |
79.30% |
|
|
|
232 |
188 |
22 |
78.30% |
|
|
|
145 |
180 |
23 |
77.30% |
|
|
|
94 |
174 |
24 |
76.20% |
|
|
|
90 |
164 |
25 |
75.20% |
|
|
|
122 |
162 |
26 |
74.20% |
|
|
|
114 |
160 |
27 |
73.10% |
|
|
|
340 |
158 |
28 |
72.10% |
|
|
|
193 |
154 |
29 |
71.10% |
|
|
|
135 |
150 |
30 |
70.10% |
|
|
|
261 |
146 |
31 |
68.00% |
|
|
|
75 |
146 |
31 |
68.00% |
|
|
|
63 |
145 |
33 |
67.00% |
|
|
Table 14:
Urinary ioidine salt in Cameroon
schoolchildren following salt iodization (continued)
|
Ul (ug/l) |
Value |
Rank |
Percent |
Descriptive Statistics |
|
246 |
144 |
34 |
65.90% |
|
|
142 |
142 |
35 |
63.90% |
|
|
174 |
142 |
35 |
63.90% |
|
|
121 |
141 |
37 |
62.80% |
|
|
395 |
140 |
38 |
60.88% |
|
|
320 |
140 |
38 |
60.80% |
|
|
240 |
138 |
40 |
57.70% |
|
|
140 |
138 |
40 |
57.70% |
|
|
66 |
138 |
40 |
57.70% |
|
|
146 |
135 |
43 |
56.70% |
|
|
115 |
133 |
44 |
55.60% |
|
|
82 |
132 |
45 |
54.60% |
|
|
82 |
128 |
46 |
53.60% |
|
|
535 |
124 |
47 |
52.50% |
|
|
74 |
122 |
48 |
50.50% |
|
|
35 |
122 |
48 |
50.50% |
The median lies half-way between these two values |
|
83 |
121 |
50 |
49.40% |
|
|
104 |
120 |
51 |
46.30% |
|
|
64 |
120 |
51 |
46.30% |
|
|
208 |
120 |
51 |
46.30% |
|
|
49 |
118 |
54 |
45.30% |
|
|
89 |
117 |
55 |
44.30% |
|
|
109 |
115 |
56 |
42.20% |
|
|
106 |
115 |
56 |
42.20% |
|
|
32 |
114 |
58 |
40.20% |
|
|
128 |
114 |
58 |
40.20% |
|
|
232 |
111 |
60 |
39.10% |
|
|
88 |
110 |
61 |
38.10% |
|
|
115 |
109 |
62 |
37.10% |
|
|
144 |
108 |
63 |
36.00% |
|
|
86 |
106 |
64 |
35.00% |
|
|
150 |
104 |
65 |
34.00% |
|
|
224 |
96 |
66 |
32.90% |
<100 UG/L |
|
92 |
95 |
67 |
30.90% |
|
|
180 |
95 |
67 |
30.90% |
|
|
193 |
94 |
69 |
29.80% |
|
Table 14:
Urinary iodine data in Cameroon schoolchildren following salt iodization
(concluded)
|
Ul (ug/l) s |
Value |
Rank |
Percent |
Descriptive Statistics |
|
133 |
92 |
70 |
29.80% |
|
|
80 |
90 |
71 |
26.80% |
|
|
87 |
90 |
71 |
26.80% |
|
|
96 |
89 |
73 |
25.70% |
|
|
120 |
88 |
74 |
24.70% |
|
|
146 |
87 |
75 |
22.60% |
|
|
160 |
87 |
75 |
22.60% |
|
|
124 |
86 |
77 |
21.60% |
|
|
90 |
83 |
78 |
20.60% |
|
|
10 |
82 |
79 |
18.50% |
|
|
55 |
82 |
79 |
18.50% |
|
|
108 |
80 |
81 |
16.40% |
|
|
480 |
80 |
81 |
16.40% |
|
|
80 |
75 |
83 |
15.40% |
|
|
122 |
74 |
84 |
14.40% |
|
|
198 |
66 |
85 |
12.30% |
|
|
200 |
66 |
85 |
12.30% |
|
|
87 |
65 |
87 |
11.30% |
|
|
200 |
64 |
88 |
10.30% |
|
|
188 |
63 |
89 |
8.20% |
|
|
54 |
63 |
89 |
8.20% |
|
|
273 |
55 |
91 |
7.20 |
|
|
120 |
54 |
92 |
6.10% |
|
|
140 |
49 |
93 |
5.10% |
<50 ug/l |
|
110 |
42 |
94 |
4.10% |
|
|
42 |
35 |
95 |
3.00% |
|
|
95 |
32 |
96 |
2.00% |
|
|
117 |
26 |
97 |
1.00% |
|
|
295 |
10 |
98 |
.00% |
<20 ug/l |
Figure
4: Frequency table and histogram to show
distribution of urinary iodine values after
iodization in
|
Urinary iodine (ug/l) |
Frequency |
|
0-49 |
6 |
|
50-99 |
27 |
|
100-149 |
35 |
|
150-199 |
13 |
|
200-249 |
7 |
|
250-299 |
5 |
|
300-349 |
2 |
|
350-399 |
1 |
|
400-449 |
0 |
|
450-499 |
1 |
|
500-549 |
1 |
|
550-599 |
0 |
Table 10:
Selection of communities in El Saba
using the PPS method
|
Name |
Population |
Cumulative population |
Cluster |
|
|
|
Utural |
600 |
600 |
|
|
|
|
Mina |
700 |
1,300 |
1 |
|
|
|
Bolama |
350 |
1,650 |
2 |
|
|
|
Taluma |
680 |
2,380 |
3 |
|
|
|
War-Yali |
430 |
2,810 |
|
|
|
|
Galey |
220 |
3,030 |
|
|
|
|
Tarum |
40 |
3,070 |
|
|
|
|
Hamtato |
150 |
3,220 |
4 |
|
|
|
Nayjaff |
90 |
3,310 |
|
|
|
|
Nuviya |
300 |
3,610 |
|
|
|
|
Cattical |
430 |
4,040 |
5 |
|
|
|
Paralai |
150 |
4,190 |
|
|
|
|
Egala-kuru |
380 |
4,570 |
|
|
|
|
Uwarnapol |
310 |
4,880 |
6 |
||
|
Hilandia |
2,000 |
6,880 |
7 |
||
|
|
|
|
8 |
||
|
Assosa |
750 |
7,630 |
9 |
||
|
|
|
|
|
||
|
Dimma |
250 |
7,880 |
|
||
|
Aisha |
420 |
8,300 |
10 |
||
|
Nam Yao |
180 |
8,480 |
|
||
|
Mai Jarim |
300 |
8,780 |
|
||
|
Pua |
100 |
8,880 |
|
||
|
Gambela |
710 |
9,590 |
11 |
||
|
Fugnido |
190 |
9,880 |
12 |
||
|
Degeh Bur |
150 |
10,030 |
|
||
|
Mezan |
450 |
0,480 |
|
||
|
Ban Vinai |
400 |
10,880 |
13 |
||
|
Puratna |
220 |
11,100 |
|
||
|
Kegalni |
140 |
11,240 |
|
||
|
Hamali-Ura |
80 |
11,320 |
|
||
|
Kameni |
410 |
11,730 |
14 |
||
|
Jiroya |
280 |
12,010 |
|
||
|
Yanwela |
330 |
12,340 |
|
||
|
Bagvi |
440 |
12,780 |
15 |
||
|
Atota |
320 |
13,100 |
|
||
|
Kogouva |
120 |
13,220 |
16 |
||
|
Ahekpa |
60 |
13,280 |
|
||
|
Yondot |
320 |
13,600 |
|
||
|
Nozop |
1,780 |
15,380 |
17 |
||
|
|
|
|
18 |
||
|
Mapazko |
390 |
15,770 |
19 |
||
|
Lotohah |
1,500 |
17,270 |
20 |
||
|
Voattigan |
960 |
18,230 |
21 |
||
|
|
|
|
22 |
||
|
Plitok |
420 |
18,650 |
|
||
|
Dopoltan |
270 |
18,900 |
|
||
|
Cococopa |
3,500 |
22,400 |
23 |
||
|
|
|
|
24 |
||
|
|
|
|
25 |
||
|
|
|
|
26 |
||
|
|
|
|
27 |
||
|
Famegzi |
400 |
22,820 |
|
||
|
Jigpelay |
210 |
22,840 |
|
||
|
Mewoah |
50 |
22,890 |
|
||
|
Odigla |
350 |
23,240 |
28 |
||
|
Sanbati |
1,440 |
24,680 |
29 |
||
|
Andidwa |
260 |
24,940 |
30 |
||