Lecture_3

Lecture_3 - Global Health Challenges Social Analysis 76:...

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: Global Health Challenges Social Analysis 76: Lecture 3 Harvard University Initiative for Global Health Core Health and Health System Measurements Definitions of Mortality Rates and Probabilities Measuring Mortality Causes of Death Definitions of Incidence and Prevalence Measuring Diseases and Risk Factors The Politics of Measurement Harvard University Initiative for Global Health Core Measurements Understanding health problems and how health systems respond to these problems is based on some core health and health system measurements. Controversies and alternative interpretations over these measurements underlie a major fraction of global health debates. Harvard University Initiative for Global Health sceptic 1. person inclined to doubt accepted opinions. Oxford English Dictionary Harvard University Initiative for Global Health Core Health and Health System Measurements Definitions of Mortality Rates and Probabilities Measuring Mortality Causes of Death Definitions of Incidence and Prevalence Measuring Diseases and Risk Factors The Politics of Measurement Harvard University Initiative for Global Health Definition of Mortality Rate (deaths in an age group in a year) (population in an age group at the midpoint of the year) Harvard University Initiative for Global Health Commonly Reported Probabilities of Death 1q0 infant mortality `rate', the probability of death between birth and exact age 1. 5q0 child mortality, the probability of death between birth and exact age 5. 45q15 adult mortality, the probability of death between age 15 and exact age 60 conditional on being alive at age 15. Harvard University Initiative for Global Health Core Health and Health System Measurements Definitions of Mortality Rates and Probabilities Measuring Mortality Causes of Death Definitions of Incidence and Prevalence Measuring Diseases and Risk Factors The Politics of Measurement Harvard University Initiative for Global Health 250 Years of Child and Adult Mortality SWEDEN 0.5 0.4 Value Female 5q0 Male 5q0 1 0.8 Value Female 45q15 Male 45q15 0.3 0.2 0.1 0 0.6 0.4 0.2 0 1750 177518001825185018751900192519501975 2000 Year 1750 1775180018251850187519001925195019752000 Year Harvard University Initiative for Global Health Coverage of Death Registration. Mortality Data (1995 Onwards) by Cause Available in WHO Coverage 90+ 75-89 50-74 <50 (sample) <50 na Harvard University Initiative for Global Health Alternative Mortality Measurement Methods Complete birth histories Sibling survival Household deaths in the last 12 months Demographic surveillance systems Harvard University Initiative for Global Health Child Mortality 5Q0:India 300 Nat.Fam.Plan.Surv.70 (ind) Inf.&Child Mort.Surv. 79 (ind.) probability of dying by age 5 per 1000 live births 250 Nat.Fam.Plan.Surv.80 (ind.) Census 81 (ind.) Nat. Fam.Health Surv.92 (dir.) Nat. Fam.Health Surv.92 (ind.) 200 Derived from vit.regis. imr 150 100 50 0 1955 1960 1965 1970 1975 year 1980 1985 1990 1995 2000 Harvard University Initiative for Global Health Core Health and Health System Measurements Definitions of Mortality Rates and Probabilities Measuring Mortality Causes of Death Definitions of Incidence and Prevalence Measuring Diseases and Risk Factors The Politics of Measurement Harvard University Initiative for Global Health International Classification of Diseases and Injuries First revision 1893 Now maintained by WHO and revised every 1015 years, current version is 10th revision Comparisons overtime complicated by changing understanding of disease causation and classification Harvard University Initiative for Global Health Japan, Ischaemic Heart Disease Deaths WHO Mortality Database 80000 70000 60000 IHD deaths 50000 40000 30000 20000 10000 0 1965 1970 1975 1980 1985 Year 1990 1995 2000 2005 Harvard University Initiative for Global Health Variation in Medical Culture and Cause of Death Attribution In addition to changes in the classification system, different countries and regions within countries appear to vary in how the same clinical entity is assigned a cause of death. Harvard University Initiative for Global Health Rank of the 50 US States According to Cardiovascular Diseases and Ischaemic Heart Disease age standardized death rate, 2000 50 40 CVD (rank) AL NY 30 20 10 0 0 10 20 30 IHD (rank) 40 50 Harvard University Initiative for Global Health Core Health and Health System Measurements Definitions of Mortality Rates and Probabilities Measuring Mortality Causes of Death Definitions of Incidence and Prevalence Measuring Diseases and Risk Factors The Politics of Measurement Harvard University Initiative for Global Health Definition of Incidence Rate (the number of new cases of a disease) (person-time of observation) Harvard University Initiative for Global Health Definition of Prevalence Rate (number of individuals with a disease) (population) Harvard University Initiative for Global Health Approximate Relationship Between Prevalence and Incidence Prevalence=Incidence*Duration Harvard University Initiative for Global Health Core Health and Health System Measurements Definitions of Mortality Rates and Probabilities Measuring Mortality Causes of Death Definitions of Incidence and Prevalence Measuring Diseases and Risk Factors The Politics of Measurement Harvard University Initiative for Global Health Key Issues Diagnostic technology the impact of test characteristics on the interpretation of results Interpreting data collected at health facilities the problem of selection bias and coverage Using population surveys biomarkers and self-reports Harvard University Initiative for Global Health Diagnostic Test Characteristics Test Positive Negative True disease status Positive Negative A C B D Harvard University Initiative for Global Health Test Sensitivity The proportion of those with the disease who will test positive Sensitivity = A/(A+B) Harvard University Initiative for Global Health Test Specificity The proportion of those without the disease who will test negative Specificity = D/(C+D) Harvard University Initiative for Global Health Surveys and Test Characteristics Imagine a test for HIV that is 98% sensitive and 95% specific. If the survey finds 10% of the population test positive, what is the true prevalence? Answer: 5.4% Harvard University Initiative for Global Health True Prevalence p(0.98) + (1-p)(1-0.95) = 0.10 p(0.93) = 0.05 p = 0.054 Harvard University Initiative for Global Health Facility Based Data Collection For medium duration conditions in highly resourced health systems can be used for incidence measurement Can be very useful for epidemic or outbreak surveillance Global Outbreak and Response Network and SARS Harvard University Initiative for Global Health Reported TB Cases, United States, 1980-2000 28000 26000 24000 22000 20000 18000 16000 Cases Cases 80 82 84 86 88 90 92 94 96 98 00 Year Harvard University Initiative for Global Health Self-reported and Measured Height, Males and Females, 1999-2000 180 175 170 Height (cm) 165 160 M - NHANES IV F - NHANES IV M - BRFSS 1999-2000 155 F - BRFSS 1999-2000 M - NHANES IV self-report F - NHANES IV self-report 150 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80+ Age (years) Harvard University Initiative for Global Health Self-reported and Measured Weight, Males and Females, 1999-2000 100 95 90 85 Weight (kgs) 80 75 70 M - NHANES IV 65 F - NHANES IV M - BRFSS 1999-2000 60 F - BRFSS 1999-2000 M - NHANES IV self-report F - NHANES IV self-report 55 50 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80+ Age (years) Harvard University Initiative for Global Health Core Health and Health System Measurements Definitions of Mortality Rates and Probabilities Measuring Mortality Causes of Death Definitions of Incidence and Prevalence Measuring Diseases and Risk Factors The Politics of Measurement Harvard University Initiative for Global Health Three Questions to Ask When Politics and Measurement Intersect 1. What is the primary source of information and what biases are expected? 2. Have known biases been corrected and is there an explicit data audit trail? 3. Will someone who could have influenced the primary data collection or corrections for known bias stand to gain or lose from this result? Harvard University Initiative for Global Health ...
View Full Document

This note was uploaded on 07/30/2008 for the course SA 76 taught by Professor Sanes during the Fall '07 term at Harvard.

Ask a homework question - tutors are online