Post on 14-Dec-2015
transcript
The road map to problem solving
The analysis plan
FETP India
Competency to be gained from this lecture
Plan the analysis of a study on the basis of the study objectives
Key areas
• Objectives of the study • Design and indicators• Study parameters• Analysis• Sample size
The ad-hoc approach to conducting an epidemiological study
Before data collection• I want to do a study
I am not clear about the objectives
• I prepare a questionnaire I am not clear about
what information I need
• I collect data I am not clear what I
will use for what
After data collection• I come back with data
I realize they are difficult to analyse
• I analyse the data I realize it is difficult to
interpret the results
• I interpret the results I realize it is difficult to
use them
Sound familiar?
The life cycle of an epidemiological investigation
Identifying data needs
Spelling out the research question
Formulating the study objectives
Planning the analysis
Preparing data collection instruments
Analysing data
Drawing conclusions
Formulating recommendations
Involving the programme
Collecting data
Analysis plan
The analysis plan: A road map to making sense of data
1. Formulate the objectives of the study 2. Choose a design to identify key
indicators3. Identify parameters needed for
indicators 4. Prepare the analysis5. Estimate sample size
Objectives
The analysis plan: A road map to making sense of data
1. Formulate the objectives of the study 2. Choose a design to identify key
indicators3. Identify parameters needed for
indicators 4. Prepare the analysis5. Estimate sample size
Objectives
The study objectives
• Formulated in limited number • Sorted out as primary and secondary• Focused• No more than one verb each• Clear about whether:
Hypothesis testing Quantity measuring
• Epidemiological termsObjectives
Estimating versus testing
• Estimating a quantity Use the verb “Estimate” • E.g., Estimate the prevalence of diabetes
• Testing a hypothesis Use the verb “Determine” • E.g., Determine whether a contaminated
well caused an outbreak
Objectives
Good and bad examples of study objectives
• Determine the importance of Kala Azar Estimate the prevalence of Kala Azar in the
community
• Assess vitamin A deficiency and tuberculosis Estimate the effect of vitamin A
supplementation over the cure rate of tuberculosis patients
• Evaluate iodine deficiency and equity Determine whether iodine deficiency is more
common among poorer people Objectives
From testing a hypothesis to estimating a quantity
• Determine whether iodine deficiency is more common among poorer people Hypothesis testing Crude objective, smaller sample size
• Estimate the relative frequency of iodine deficiency among poorer people Quantity estimating More elaborate objective, larger sample size
Objectives
The analysis plan: A road map to making sense of data
1. Formulate the objectives of the study 2. Choose a design to identify key
indicators3. Identify parameters needed for
indicators 4. Prepare the analysis5. Estimate sample size
Design and indicators
Elements to consider to choose a study design
• Is the study descriptive or analytical? Is there a need to compare groups? Is there just a need to estimate a frequency?
• Is the outcome (e.g., disease) acute or chronic Need of prevalence data for chronic outcomes Need of incidence data for acute outcomes
• Is the outcome common or rare? Case control design for rare outcomes Cohort / cross sectional designs for common
outcomes
Design and indicators
Choosing a study design adapted to the objective to identify the indicator
Example: Estimating the relative frequency of iodine deficiency
among people below poverty line (BPL)
• Elements deducted from the objective: Analytical approach: Compare two groups Chronic condition: Prevalence data Common condition: Survey
• Study design: Analytical cross sectional study
• Indicator: Ratio of prevalence of iodine deficiency among BPL
persons
Design and indicators
The analysis plan: A road map to making sense of data
1. Formulate the objectives of the study 2. Choose a design to identify key
indicators3. Identify information needed for
indicators 4. Prepare the analysis5. Estimate sample size
Parameters
Identification of information needed to calculate the indicator
• List the indicators that the study will generate Rates, ratio, proportions or quantitative variables
• Example: Measles coverage
• Identify the information elements that will be needed to calculate the indicators Numerators an denominators
• Example: Number of children vaccinated / total children
• Information elements may address: Outcome variable (s) “Covariate”, including
• Potential risk factors• Potential confounders
Parameters
From information element to variables
• Identify the variables that may be used to reflect the information element The information element “Measles vaccination
status” can be assessed by review of cards or interview of the mother
• Choose the best possible variable Review standardized guidelines (e.g., WHO, CDC)
• Plan data collection methods for each variable Observation Interview Laboratory methods
Parameters
Example: Outcome measurement for iodine deficiency study
Information element
Variable
Data collection method to obtain the variable
Past exposure to iodine deficiency
•Goitre •Physical examination
Current iodine intake
•Urine iodine excretion
•Laboratory
Access to iodized salt
•Test of house salt for iodization
•Field spot testParameters
Example: Covariate measurement for iodine deficiency
• Potential risk factors Income
(Validated field methods) Community (e.g., minorities) Caste Education Residence
• Potential confounding factors Age Sex
Parameters
The analysis plan: A road map to making sense of data
1. Formulate the objectives of the study 2. Choose a design to identify key
indicators3. Identify parameters needed for
indicators 4. Prepare the analysis5. Estimate sample size
Analysis
Rationale for preparing the data analysis in advance
• Focus on the objectives of the study• Limit multiple comparisons• Avoid comparisons for which the study was
not designed• Ensure that data collected can be analyzed
“Other, specify: _____” kind of data that create minuscule groups that cannot be analyzed
• Save time Filling dummy tables accelerates data analysis
Analysis
Preparing the analysis, stage by stage
• Recoding stage Example: Transform age into age groups
• Descriptive stage Calculate prevalence or incidence
• Analytical stage Univariate, stratified and multivariate analysis Prepare empty (dummy) table shells upfront
• Dichotomize all variables for simple dummy tables– Using the median (e.g., Income > median)– Using a value known to be important (e.g., 200
CD4)Analysis
Example: Initial stage of the analysis of the study on iodine deficiency
according to income• Recoding stage
Create outcome data with laboratory results Recode income data
• Dichotomize quantitative income variable to create a “BPL” Yes/ No variable
• Descriptive stage Calculate prevalence of the thee outcomes
• Goitre, urinary excretion and salt spot test
Adjust confidence intervals for design effect
Analysis
Example: Analytical stage of the analysis of the study on iodine
deficiency according to income
• Univariate analysis Prevalence of three outcomes by age, sex
and residence Prevalence of three outcomes by income
(potentially examine dose response effect)
• Stratified analysis Prevalence of three outcomes by income,
stratified for age, sex and residence
• Multivariate analysis Logistic regression model
Analysis
Dummy table for iodine deficiency study
(Analytical stage) *Prevalence
Prevalence ratio (95% confidence
interval)Exposures Exposed Unexposed
Female sex XX/XX (XX%) XX/XX (XX%) XX (XX-XX)
Muslim XX/XX (XX%) XX/XX (XX%) XX (XX-XX)
Age > median XX/XX (XX%) XX/XX (XX%) XX (XX-XX)
Below poverty line
XX/XX (XX%) XX/XX (XX%) XX (XX-XX)
Schedule caste XX/XX (XX%) XX/XX (XX%) XX (XX-XX)
* All variables dichotomized for the sake of simplicity
The analysis plan: A road map to making sense of data
1. Formulate the objectives of the study 2. Choose a design to identify key
indicators3. Identify parameters needed for
indicators 4. Prepare the analysis5. Estimate sample size
Sample size
The analysis plan determines the sample size
• Choose the study design Cohort, case control or survey
• Determine the level Descriptive or analytical
• Common mistake: Designing a descriptive study Trying comparisons for which the sample
size is insufficient
Sample size
Sample size for study on iodine deficiency among people below
poverty• Study design
Analytical cross sectional survey
• Level Analytical Need to:
• Use prevalence ratio for sample size estimation• OR• Use prevalence but multiply final sample size by
two to allow comparisons
Sample size
Take home messages
• Clarify precise, focused objectives• Choose a design to identify the indicator• Know the parameter you want before you
think about how to get information about it • Know where you go with the analysis
The planed analysis drives the data needs and not the reverse
• Deduct your sample size from all of the above
Additional resources on analysis plan
• Dummy tables for field epidemiology• Case study on protocol writing (Scrub
Typhus in Darjeeling, Volume 2)