280 likes | 289 Views
Promoting Rational Drug Use in the Community. Data analysis. Objectives: Session on Analysis. Describe in what ways quantitative and qualitative data can be processed Describe how quantitative and qualitative data can best be analysed
E N D
Promoting Rational Drug Use in the Community Data analysis
Objectives: Session on Analysis • Describe in what ways quantitative and qualitative data can be processed • Describe how quantitative and qualitative data can best be analysed • Understand the differences between analysis of quantitative and qualitative data
Why plan for data-processing and analysis? • To make sure that all data needed to answer research questions are indeed collected • To avoid collecting superfluous data • To make sure you plan enough time and resources for processing and analysis • To make sure your research tools are adequate and easily processed
How to plan for data-processing and analysis? • Review research questions and data-collection tools • Decide how you want to present data: - qualitative: as texts - quantitative: as numbers • Make a list of variables for quantitative analysis • Decide on key drug use measures/indicators • Make dummy tables • Decide on data-master sheets for analysis of quantitative data • Make a list of key themes for qualitative analysis
Processing of quantitative data • Check if each questionnaire/interview form is complete • Sort data according to study populations (e.g. women – men; intervention community – control community) • Review all responses to categorical variables and refine the list of values for the categorical variable (you may need to add values you had not foreseen) • Assign codes to responses in questionnaires/interview forms
Variables: • Are defined as characteristics of persons or objects which can take on different values • Categorical variables are expressed in words/categories • Numerical variables are expressed in numbers • When planning for analysis of quantitative data, make a list of all variables and their values • Assign codes to categorical variables
Analysis of quantitative data • Summarise data on data master sheet • Determine missing values • Check data master sheet for consistency/mistakes • Calculate drug use measures/indicators • Make relevant frequency distributions • Fill in tables • Do statistical tests to test hypothesis on associations between variables
Examples of drug use measures • Percentage of illness episodes not treated • Percentage of illness episodes treated with traditional medicines • Percentage of illness episodes treated on health worker advice • Percentage of illness episodes treated in self-care with medicines • Percentage of fever episodes treated with chloroquine • Percentage of diarrhoea self-medicated with antibiotics
Examples of frequency distributions as way of presenting data: • Ten most commonly used medicines: calculated as relative percentage of total medications used • Main sources of medicines, calculated as the number of times medications are obtained from specific sources divided by total number of medications • Five most commonly used medicines for diarrhoea, expressed as percentage of total number of medications used to treat diarrhoea.
Activity 1 • Review the two data master-sheets in pairs • Are any data missing: if yes, how will you deal with it? Delete the record? • How can you check if mistakes have been made during data-entry? • Have mistakes been made? • Is the data master-sheet well-designed? • How could the data master-sheets be improved?
Activity 2 • The data in the master-sheet allow for a comparison between men and women of types of drugs taken to the PRDUC course • Design a dummy table to present the data
Processing of qualitative data • Expand notes/transcribe tapes everyday • Add comments on non-verbal communication • Order data by type/group of informants • Read notes/transcriptions, read again
Qualitative analysis: an ongoing process • Read your notes, reflect, reflect more • Review your research questions: have they been answered: what do you still need to ask? • What unexpected issues/problems emerged? • Do you have sufficient data for each question; can you triangulate? Are there inconsistencies in data: do interviews confirm your observations or not? • Write down preliminary conclusions and queries • Go back to your informants: probe, ask them to explain and respond to your preliminary conclusions.
Rapid qualitative analysis • Review your list of themes for qualitative analysis, read your notes and find out if new issues emerged • Make matrices to summarise the data by theme. • Check if you have data on all your research questions • Beware of generalising: your data are not representative. • Describe your study population using key demographic variables (age, marital status, etc.)
Analysis of textual data • Make a list of codes • Apply codes to texts • Add codes as you go along • Make analytical notes on the relation between factors; how things work • Make methodological notes: observations on how the methods influenced the results; ideas on new questions to ask
Typ-fev Cause-fev Tx-fev P.eff-Tx Type of fever Cause of fever Treatment of fever Perceived efficacy treatment Coding of transcripts
Summarizing qualitative data • Matrix • Flow charts • Diagrams
Type of treatment Perceived effect Perceived side-effect Example of a treatment matrix
Source of medicines Perceived advantages Perceived disadvantages Example of a medicine source matrix
Drawing and verifying conclusions Continuous process, based on: • Summary of data • Identifying trends • Identifying associations - causations • Consider confounding factors • Validation in group and individual discussions with informants
Cite your informants to illustrate • Select case-histories which are typical and illustrate findings • Use quotes to illustrate findings
Strategies to confirm findings • Check for representativeness • Check for observer bias • Use multi-method • Compare and contrast data • Do additional research, include surveys to test hypothesis • Get feedback from communities and key informants
Activity 3 Community sub-groups: • Review the illness-recall data in the SSI forms. • If you had collected 20 of such illness-recalls: how can you summarize these data in one or two data master-sheet(s)?
Activity 3 Health institution sub-groups • Review the simulated client visit guidelines. • If you had done 20 such visits, how could you have summarized the data in a data-master-sheet?