Cathy Conolly of the MRC, Natal for her expert assistance with the statistical aspects of the project. In addition, people from the community, such as community health workers (CHWs), were involved in all phases of the study, including manual analysis of the results, after being properly trained.
COMPARISON OF MANUAL ANALYSIS TO ELECTRONIC ANALYSIS
EVALUATION OF THE LQAS METHODOLOGY .1 Coverage Estimates
LIMITATIONS OF STUDY .1 LQAS Methodology
AND RECOMMENDATIONS 6.1 INTRODUCTION
CONCLUSION
RECOMMENDATIONS
YEARS QUESTIONNAIRE MAP OF UTHUKELA DISTRICT
LIST OF TABLES
Coverage of OPV doses administered in the first year of life in uThukela district in 2002. Hepatitis B vaccine (HBV) coverage in the first year of life in uThukela district in 2002.
CHAPTER!
BACKGROUND TO THE STUDY
- INTRODUCTION
- OVERVIEW OF THE uTHUKELA DISTRICT CHILD SURVIVAL PROJECT
- Background statistical information of the uThukela District
- The Role of Evaluation in the TDCSP
- BACKGROUND INFORMATION ABOUT THE RESEARCH
- PURPOSE OF THE SECONDARY STUDY
- OBJECTIVES OF THE STUDY
- ASSUMPTIONS UNDERLYING THE STUDY
- SIGNIFICANCE OF THE STUDY
- ORGANISATION OF THE REPORT
- SUMMARY
- INTRODUCTION
Some questions were the same as from the previous KPC survey conducted in the year 2000. To assess the accuracy of the manual analysis using the electronic analysis as the gold standard.
LITERA TURE REVIEW
- PURPOSE OF THE LITERATURE REVIEW
- CLASSIFICATION OF HEALTH RESEARCH
- Health Systems Research
- EVALUATION TOOLS THAT CAN BE USED IN ASSESSING HEALTH PROGRAMS THROUGH OPERATIONAL RESEARCH
- Rapid Epidemiological Assessment
- Lot Quality Assurance Sampling
- Previous Studies Conducted, Using the LQAS Approach
- SUMMARY
It can rely on monitoring service level data and varies in depth of analysis. Experience with the analysis of the LQAS survey has shown that it is simple and does not require a sophisticated statistical package.
RESEARCH METHODOLOGY
INTRODUCTION
RESEARCH DESIGN
SAMPLING DESIGN
STUDY POPULATION
ETHICAL APPROVAL
PATIENT CONFIDENTIALITY
DATA CAPTURING
FREQUENCY GENERATION
STATISTICAL ANALYSIS
WEIGHTING OF RESULTS
SA is the municipality, n is the sample size used in each surveillance area, N is the size of the population in SA, wt is the weighting factor, and p is the minicoverage proportion. Adding the results of the last column gives the overall coverage for the district, represented as a fraction.
SUMMARY
INTRODUCTION
RESULTS
- COMPARISON OF MANUAL ANALYSIS TO ELECTRONIC ANALYSIS
- Maternal Health (0-11 months)
- Diarrhoeal Disease and Acute Respiratory Infection (0-23 months)
- Integrated Management of Childhood Illnesses (12-23 months)
- HIV/AIDS and Well-being (Women 15-49 years)
- COMPARISON OF UNWEIGHTED ELECTRONIC RESULTS AND WEIGHTED ANALYSIS
- Maternal Health (0-11 months)
- Integrated Management of Childhood Illnesses (12-23 months questionnaire)
- COVERAGE VALUES FOR KNOWLEDGE AND PRACTICES
A comparison was made between the unweighted coverage calculated by the manual analysis and the electronic analysis to determine how accurate the manual analysis was. In most comparisons there was a very small difference between the weighted and unweighted results. Although this implied a statistically significant difference between the weighted result and the unweighted result, the graphical comparison (Fig. 4.5) showed that the weighted results were consistently higher than the unweighted results.
The Pearson correlation of 0.99752 and a p-value of 0.842877 in the two-tail test confirmed no statistically significant difference between the weighted result and the unweighted result. A graphical comparison of unweighted results with weighted results for diarrhea and acute respiratory infections (0-23 months questionnaire) is shown in Figure 4.6. In most comparisons there was a very small difference between the weighted result and the unweighted result.
This did not mean a statistically significant difference between the unweighted and weighted scores.
AROUND CIDLD HEALTH, MATERNAL HEALTH AND HIV/AIDS AND WELL-BEING
Child Health
Mothers of children in the age group from 0 to 11 months were asked which types of complementary foods they gave their children most often. All children aged 0-23 months were included to determine the incidence of diarrhea in the last 2 weeks in uThukela district. Seventy-five of the 240 children surveyed reported diarrhea in the past 2 weeks. gave anything or did not know.
All indicators where a municipality was below the district average were underlined in the relevant annexes (Annex B and Annex C). 3 The child has been weighed once within the last two months 11 At what age should a mother stop breastfeeding completely. None of the municipalities had indicators of diarrheal diseases and acute respiratory infection that fell below the average coverage in uThukela district in 2002.
Only Mbabazane and Mtshezi had indicators that fell below the average coverage in the HIV I AIDS and well-being technical areas.
SUMMARY
INTRODUCTION
CHAPTERS DISCUSSION
EVALUATION OF OVERALL DISTRICT COVERAGE
- Maternal Health
- Knowledge and Practices Regarding HIV/AIDS
Based on an overall analysis of the age distribution of mothers in the 0-11 month and 12-23 month questionnaires, it was positive to note that 86.3% of the mothers were between 18 and 35 years old. None of the municipalities fell below the average decision rule, but this is because the percentage coverage is so low. With a target of 80.0%, all municipalities fell below the target decision rule of 16 (Appendix Cl a indicator 4).
Regarding indicators 1 and 5 of the 0-11 month category (in terms of reported and indicated vitamin A supplements), as shown in Table CIa of Appendix C, it is evident that more mothers reported being given vitamin A A immediately after birth (indicator 1) than what was shown in RTHC (indicator 5). Even with the set target of 90.0%, as reflected in the LQAS Table (Table C3a of Appendix C), none of the municipalities fell under the target decision-making rule. Of the children surveyed, 31.3% had been sick with a cough or difficulty breathing in the two weeks before the survey.
The unweighted results calculated coverage assuming equal population sizes in each of the SAs.
EVALUATION OF THE LQAS METHODOLOGY
- Coverage estimates
The average decision rule is determined by using the total percentage coverage of an indicator. If an indicator has a very low overall percentage coverage, the average decision rule will consequently be very low, and municipalities can reflect as performing above the average decision rule. It is thus important to ensure that targets are set so that target decision-making rules can be determined and municipalities can be assessed on this basis.
However, it is still important to identify the average decision rule for an indicator so that prioritization is possible. This greater precision is due to the fact that LQAS is rooted in the principles of stratified sampling, which generally yields estimates with narrower confidence intervals than estimates derived from cluster samples of the same size. Due to the small sample size used in LQAS, the ability to further disaggregate data based on dichotomous variables results in a loss of statistical power.
This method can be further adapted to build capacity, strengthen partnerships, improve project effectiveness and ultimately translate these changes into beneficial outcomes at the beneficiary level.
LIMITATIONS OF STUDY
To improve these performance areas, PGS managers will need to target specific indicators that may be difficult to achieve in isolation. Questions that had a number of correct options aimed at determining how many correct options the informant chose (eg, Table 16) were grouped into "2 or more correct options" and "3 or more correct options." This overlap implies that informants who knew 3 options would fit into the "3 or more correct options" category as well as the "2 or more correct options" category. In the category "2 or more correct options" it is therefore not known how many informants knew only 2 options.
For some of the indicators and performance results measured, targets for district coverage have not been set. As this is one of the most reliable sources to measure performance results, in light of some of the limitations already discussed, it would be worthwhile to define coverage targets.
CONCLUSION AND RECOMMENDATIONS
INTRODUCTION
These indicators may not require prioritization at this stage (as determined by LQAS), but it may still be valuable to try to improve overall coverage, especially when they are significantly lower than national standards or pre-defined district targets. With regard to municipalities that perform below the average coverage decision rule, it may be more appropriate to establish the indicators that require prioritization based on the target decision rule, as municipalities may fall within the acceptable coverage decision rule, but still be. These indicators in the manual analysis can then be re-evaluated to determine possible reasons why a significant difference between the manual analysis and the electronic analysis resulted.
In terms of setting targets for indicators or performance outcomes that do not have district coverage targets, the most appropriate source of targets would be from previous KPC surveys in the district or from national targets. These indicators may not need to be prioritized at this stage (as defined by LQAS), but it may still be useful to try to improve overall coverage, especially if it is significantly lower than national standards or pre-defined district targets. With respect to municipalities operating under the average coverage decision rule, it may be more appropriate to identify indicators that require prioritization based on the target decision rule, as municipalities may fall within an acceptable coverage decision rule, but still are. performs poorly against predetermined district coverage targets or national standards.
With regard to setting targets for indicators or performance outcomes that do not have district coverage targets, the most appropriate source of targets would be from previous KPC surveys in the district, or frolD national targets.
Municipality 2 Municipality 3 Municipality 4 Municipality 5
Municipality 2 Municipality 3 Municipality 4 Municipality 5
No indicator number is correct (or %) in each municipality or SA sample total. Total Capture- The average average target for the target decision. No number of the indicator is correct (or %) in any total sample total in the municipality or SA.
None Jndier Number correct (or %) in each Total Total in municipality or SA sample Total Average Average coverage Target for the Target Decision. No indicator Number correct (or %) in each Total no Total in municipality or SA sample Total catch Average Average Target for the Target decision. Mother municipality/decision rule) correct ment area coverage assessment rule. HIV/AIDS municipality/decision rule (circle the no area coverage the rule . indicators) under SA's standard) correct sample Coverage decision assessment.
None Indicator Number correct (or %) in each Total Total in municipality or SA Total catch Average An average coverage Target for the Target Decision.
No indicator Number correct (or %) in each Total none Total in municipality or SA sample Total catch- Average Average Target for the target decision. No indicator Number correct (or %) in each Total no Total in municipality or SA sample Total catch- Average average target for the target decision.
WEIGHTING ELECTRONIC RESULTS WITH 95% CONFIDENCE INTERVALS Each indicator used corresponds to the indicators used in the LQAS Tables in Appendix B and Appendix C. If you had danger signs during pregnancy or after delivery, where to seek medical attention. What dangerous signs of a respiratory infection would cause you to take your child for medical help?
Participation in this survey is voluntary and you can choose not to answer any individual question or all of the questions. However, we hope you will participate in this survey as your opinions are important.
PLEASE CIRCLE THE RESPONDENT'S ANSWERS
BREASTFEEDING -
B Have you given (NAME) teas or juices or any other liquids in the last 24 hours. Have you given (NAME) any other milk, such as canned milk, D powder, (Nespray) or fresh animal milk.
DIARRHEA - Knowledge
MARK ALL RESPONSES
RESPIRATORY INFECTIONS AND GENERAL DANGER SIGNS Knowledge and Behavior
4 6 When (NAME) became ill, and he/she was taken to a health Mother 1 facility who decided that the child needed treatment.
Growth Monitoring (12-23 Months)
END OF QUESTIONNAIRE
