Evaluating the Completeness of the Civil Registration System
Civil registration – Quality control Level of registration completeness - All local registration areas carried the procedures - Every vital event has a record on the system - All local offices transmit the records up the ladder
Direct methods Direct methods Matching – Using civil registration records as independent source Matching – Using administrative and social records Matching – Using population censuses and surveys lists Dual records system
M1. Civil registration records as independent source Evaluating birth register using death register Consists of matching the death record with the birth record Up to 30 years of age Not really a good measure – mobility of population Useful for infant deaths Accuracy increases if births occur in medical facilities
M1. Civil registration records as independent source Hand matching If the number of infant deaths is small If both events (birth and death) occur in the same health facility In the local registrar office, given the small numbers, a feasible option
M1. Civil registration records as independent source Electronic matching Fast, effective approach Having a PIN - advantage Requirements Computerized records Developing applications Hardware
M1. Civil registration records as independent source Edit method Identify, for example, record with low birth-weights Match with death record If not found, check with the facility to confirm child is alive
M2. Administrative and social records Involves other organizations Matching birth and death records with: School enrolment Hospital records Baptism Burial records Newborn screening programs Reporting of HIV
M3. Population census and survey records Lists of births and deaths Census enumeration areas Census questions Matching with civil registration records
M4. Dual record system Matching civil registration system (CVS) and periodical retrospective survey (PRS) When information from two sources are matched, four mutually exclusive sets of records emerge: Events recorded by both exercises Recorded by CVS but not PRS Recorded by PRS but not CVS Not recorded by either source:
Not recorded by either source: M4. Dual record system Not recorded by either source: N=C+N1+N2+((N1xN2)/C) Where C = the number of vital events recorded by both methods N1= the number of events recorded by the first method but not the second N2= the number of events recorded by the second method but not the first Thus, the events omitted by both: Y=((N1xN2)/C)
Advantages and limitations of direct methods Two sources for matching If both are quality sources and independent – accurate Pointing to the cause Feasible at all levels Limitations Really independent? Population movements undermine accuracy Costs
Indirect methods These methods were developed to assess the completeness and quality of vital statistics Statistics is coming from civil registration Statistics of low quality most likely consequence of errors in registering events Comparison of trends
Indirect methods These methods were developed to assess the completeness and quality of vital statistics Statistics is coming from civil registration Statistics of low quality most likely consequence of errors in registering events Comparison of trends Delayed registration Patterns in the sex ratio of births Comparison with census data Comparison of rates observed in similar population or previous periods Incomplete data methods: indirect techniques
Comparison of trends At small area level Expected values established by previous trends Unexpected current values Large proportion of “unknown” answers
Delayed registration Monitoring difference between date of occurrence and date of registration % of events not registered in prescribed period of time – indicator of possible deficiency Excessive proportion of delayed registration in health facilities indicates overburden – inadequate infrastructure
Comparison with census data Quality of the census itself Census content If all basic components of population are accurately captured by census – birth, deaths, immigrants, emigrants, then Intercensal population growth of population equals the sum of intercensal births and immigrants minus intercensal deaths and emigrants Major obstacle – reliable migration statistics
Comparison of rates Useful, but requires caution Many factors, not necessarily linked to registration errors Differences in age structures Population estimates Skewed population
Indirect techniques Technical approach Long history of use Complex Require specific training
Summary Checks and balances - crucial A list of techniques Best approach – a combination of direct and indirect assessment Computerization increases checking opportunities