Lecture 21. Demographic measures and Pop.Health Flashcards
What is population data needed for?
Measuring (trends in)
• Births
• Mortality (deaths) (all-cause, cause-specific)
• Morbidity (illness) (general, specific condition)
• Migration
More applied work …
• Unemployment / benefit claimants / pensions
• Crime (broad & detailed classes of offence)
• Health service utilisation (where to provide services, and who uses
them
• Voter turnout, political party voted for
• Education pathways
population structure vs composition
population % counts by age & sex- population structure
population % counts by other attributes- population composition
sources of population data
-Census
The census asks questions of people in homes and group living situations, including how many people live or stay in each home, and the sex, age and race of each person. The goal is to count everyone once, only once, and in the right place.
• Estimated Resident Populations (ERP) –
The estimated resident population of New Zealand is an estimate of all people who usually live in New Zealand at a given date(residency, visa) – Does not typically break down by an ethnic group
• Vital events – Births, deaths, and marriages
Department of Internal Affairs maintains, but Stats NZ prepares reports -ethnic effects/patterns
• Health service utilization and outcomes (HSU) – Ministry of Health records and reports publically funded health information
E.g. hospitalizations, blood tests, pharmaceutical dispensing
• The Integrated Data Infrastructure (IDI)
– Large data repository that links de-identified data about people that have used Government services( looks at a range of health services)
• Nationally representative surveys – e.g. the NZ Health Survey(~15,000 people/annum)
Ministry of Health manages surveys with key topics and ‘spotlight’ coverage on less common issues
Self-reported health and health behaviors
• Ad hoc surveys
– Student satisfaction surveys, market research companies, etc
May not always be generalizable to the wider population
Main difference between IDI and Census
In IDI in order to be “counted”, you need to have had an interaction with one or more of these agencies
– i.e. health, education, tax, police, social development, ACC
IDI
• The IDI uses routinely collected information
from many government (and some other
agencies)
• All information is de-identified and strict
rules are in place to preserve confidentiality
• Data from many sources can be linked ( eg educational attainment can be linked to income or can look at census and link to health data like smoking and health services used )
issue with IDI
“deficit data set”- in order to be “counted” need to have a negative event/encounter with the agencies- eg police
IDI benefits and risks
Benefits:
Link data from multiple sources to gain system-wide insights
View longitudinal, life-course information
Identify risk & protective factors
Perform predictive risk modeling
Evaluate the effectiveness of particular interventions
Identify characteristics of groups with positive and negative outcomes
Risks/disadvantages:
- CANNOT Follow individuals who are using services, i.e case management. CANNOT ask about individuals
- CANNOT identify specific individuals who are at risk and would benefit from specific interventions
- CANNOT identify individuals who are abusing systems and take enforcement action.
- CANNOT have a specific count of resident population
- concerns about data quality?
- “resident population” definition varies-cannot capture everybody unlike census
what is the importance of nominators and denominators and their impact on pop,
health? (Exam thinking)
Nominators and denominators need to come from the same data set/sources.
If they come from different sources- the prevalence will not reflect the reality( eg if the denominator is too big, the prevalence will be smaller than it actually is)
If the prevalence data is inaccurate this can affect the allocation of resources
How are denominators and age-structure related?
It matters what source is used to determine the denominator to accurattely represent the risk in specific age groups. Sometimes the risks can be underestimated
Data considerations
• Ethics and data privacy/confidentiality
• Purpose of data collection vs use in analysis
• Population vs population samples
• Are the participants representative of the NZ
population?
• Objective vs subjective measures of health
what determines population structure?
Age-sex structure is a function of previous patterns / trends of fertility, migration &
mortality events
- Vital events affect structure in different ways/extents
- Changes in fertility rates, can be dramatic but will have a time lag
- Changes in infant mortality rates have similar effect
- Changes in adult mortality rates less dramatic & less variable over time.
- Spread over wider age range
- Migration can have dramatic effect, especially if the trend is age & sex specific
- E.g. post Christchurch earthquakes, diseases(COVID)
relationship betwen population structure and events
works both ways
events influence population structure
population structure influences events( fertility, mortality, migration are all affected by the population structure)
e.g. an area/country with very youthful population is likely to have fewer
deaths than a very elderly population of the same size but may be more fertile
and more likely to move regularly
Dependency ratios
a crude measure
can calculate what % of children/elderly or (children+elderly) are dependant on the working population(taxes) for support
(Youth+ elderly)/ working population x 100
not always accurate, thus a very crude measure
can be affected by ethnicity( eg Maori population is much younger-> require more support)
Numenator denominator bias
different sources and measures of population can ask different questions and therefore get different measurements
What does ethnic composition in NZ depend on?
– Data sources for numerator and denominator
• Census
• Health Service Utilisation
• IDI
• National Health Survey
– Ethnicity coding protocol used:
• Total Response
• Prioritised
• Sole/Combination
Advantages and disadvanatages of prioritised vs total response output
The demographic transition
