India EIS

Data Analysis Plan

 The Surveillance Cycle: Analysis
Diagnosis / Detection

Reporting /
Evaluation Data Collection

Action! Analysis,
Interpretation

2
Communicating Information
 Getting from Data to Report

▪ Need to know what data you have

▪ Need to decide which data to include in report
▪ Need plan for how to analyze / summarize data
Analysis plan
▪ Need to summarize each important variable
Measures of central location, frequency
▪ Need to display summary data effectively
Tables, graphs, maps
▪ Need to package into report 3
 What are the goals of the analysis?

What are the questions you want to answer?

For example,
– How many cases of each notifiable disease?
– What is time trend?
– How are cases distributed by place?
– What are their age and sex characteristics?
– How many reporting sources reported on
time?

4
 Analysis Plan = Roadmap for Analysis

▪ How to get from data to final report

▪ Analysis plan describes:
– What data you are using
• Source(s) of data, variables

– How you will look at and analyze the data

• Summary measures, tables, figures

– How you will compile the analyzed data into a

report
▪ Table shell = table with title and category labels,
5
statistics (if any), but with no data
 Analysis Plan – Steps

▪ Determine the question(s) you need to answer

and what information you want to communicate
▪ Consider the kind of data you have
▪ Determine type of analysis you would like to do
▪ Create table shells
▪ Draft results

6
 Data available

What data do you have or you want to analyze?

For example,
– S form (malaria data)
– P form (diarrhea data)
– L form (malaria positive cases)
– Weekly reports
– House-to-house survey data

7
 Type of analysis
▪ Identify variables that you would like to analyze /
summarize
– Male-female
– Age groups
– Trends
▪ For each variable, determine variable type
– Quantitative
– Qualitative
▪ Decide on how you will summarize each variable,
based on variable type
– Mean, median, mode
– Rates, proportions
8
 Table Shell — Example
Intestinal parasite positivity by age and sex
Ref: Okodua M, et al. Age and sex distribution of intestinal parasite
infection among HIV-infected subjects in Abeokuta, Nigeria. JHAS Online
2003;4:3

Table. Proportion with Intestinal parasite infection by age and sex

Males Females Total
No. No. % No. No. % No. No. %
Age Tested Positive Positive Tested Positive Positive Tested Positive Positive
≤ 20 6 1 16.7% 7 2 28.6% 13 3 23.1%
20 – 29 40 12 30.0% 42 15 35.7% 82 27 32.9%
30 – 39 38 11 28.9% 42 10 23.8% 80 21 26.3%
40 – 49 16 3 18.8% 19 6 31.6% 35 9 25.7%
50+ 1 0 0.0% 4 1 25.0% 5 1 20.0%
Total 101 27 26.7% 114 34 29.8% 215 61 28.4% 9
 Table Shell — Example
Intestinal parasite positivity by age and sex
Ref: Okodua M, et al. Age and sex distribution of intestinal parasite
infection among HIV-infected subjects in Abeokuta, Nigeria. JHAS Online
2003;4:3

Table. Proportion with Intestinal parasite infection by age and sex

Males Females Total
No. No. % No. No. % No. No. %
Age Tested Positive Positive Tested Positive Positive Tested Positive Positive
≤ 20 % % %
20 – 29 % % %
30 – 39 % % %
40 – 49 % % %
50+ % % %
Total % % % 10
 Steps for Getting to Table Shells

1. Choose variables to analyze / summarize

2. For each variable, determine variable type
3. Decide on how you will summarize each
variable, based on variable type
4. For quantitative variables, decide on groupings
(categories), if necessary
5. Create table shells and statistics without
numbers

11
 1. Choose the variables to analyze
Reported Cases of Leptospirosis, District K, August 20xx
House Date of onset of
ID Sex Number Age (yr) symptoms Date of hospitalization Fever Myalgia Jaundice
01 M 23 54 09 Aug 09 Aug Y n Y
02 M 57 69 05 Aug 07 Aug Y n n

Which variables are needed to answer the questions you

want to answer?
Among cases,
– clinical features? – fever, myalgias, jaundice
– time pattern? – date of onset, hospitalization
– place? – house
– age and sex characteristics? – age, sex
12
 2. For each variable, determine type

▪ Qualitative / categorical data

– Descriptions
– Non-numeric information
– Examples: Illness (yes / no), sex, district

▪ Quantitative data
– Measurements
– Numeric data
– Examples: Age, height, number of children
13
3. How to Analyze / Summarize Variables

Variable Type Tools for Summarizing

Qualitative Ratios, proportions, frequency

distributions, rates

Quantitative Mean, median, mode, range, et al.

14
 4. For quantitative data, decide on groupings
(categories), if necessary
▪ Recommended Method 1 — use established
categories (MOH, WHO, etc.)
▪ Recommended Method 2 — use equal width
categories, e.g., 10-year age groups

15
 Some Standard Age Categories
Used by CDC
Standard 10-year Pneumonia & Influenza
age groups Mortality HIV/AIDS
0–9 <28 days < 5 years
10 – 19 28 days – <1 year 5 – 12
20 – 29 1 – 14 13 – 19
30 – 39 15 – 24 20 – 24
40 – 49 25 – 44 25 – 29
50 – 59 45 – 64 30 – 34
60 – 69 65 – 74 35 – 39
70 – 79 75 – 84 40 – 44
80 – 89 ≥85 45 – 49
≥ 90 Unknown 50 – 54
Unknown 55 – 59
60 – 64
≥65
16
 5. Create table shells and
statistics without numbers

▪ Table shell = table with title and category

labels, statistics (if any), but with no data

17
 Table shells: example
Table. Clinical and demographic characteristics of leptospirosis
case-patients, District K, August 20xx (n=xx)
Characteristics Number Percent

Clinical
Fever
Myalgia
Jaundice
Demographic
Male sex
Age (years ) ≤ 9
10 – 19
etc.
18
 Table shells

▪ Create table shells or empty tables: table with title

and category labels, with no data
– Example: Table 1. Proportion male-female with
malaria
Male Female Total
Number tested
Number positive
% positive
Total

▪ Create empty graphs/figures

19
 Malaria positive cases in block A, district X, Rajasthan, 2017

Cases

Months 20
Diarrhea cases by age groups, block A, district X, July 2018

21
 Draft results

▪ Determine what would you like to report

– xx% positive were females/males
– xx% positive were xx years old
– High number of cases were in xx month, year
– xx number of reporting units submitted
incomplete reports in May, 2018

22
 Summary of Making an Analysis Plan

▪ Analysis plan is based on what question(s) you

need to answer and what information you want to
communicate
▪ Consider the kind of data you have
▪ Create table shells that are ready for analysis,
except for the data

23