Measuring Health Outcomes Flashcards
Proximal determinants include…
Include individual behaviors
Proximal determinants are viewed as having the most direct effect on health status. Examples include individual behavior such as hand-washing, alcohol and cigarette use), beliefs, attitudes, genetics and biology.
Measuring health status of individuals or populations can be used to
Measuring the health status of individuals or populations provides valuable information that can be used for several purposes including measuring the magnitude of disease, conducting surveillance of a disease in a population, improving clinical decisions, and evaluating the effectiveness of programs.
Health indicators
Indicators are characteristics that can be measured directly or indirectly and can be used to describe health.
Indicators can be qualitative or quantitative and can measure proximal or distal determinants as well as the disease or illness or instead, positive aspects of health such as quality of life.
Composite indicators
Are useful for multi-dimensional concepts that can’t be captured by a single indicator
Indicators can also be indices or composite measures which are formed when individual indicators are compiled into a single index. This is useful for multi-dimensional concepts that can’t be captured by a single indicator
Interval scales have no…
Have no true zero point
Interval scales are quantitative in nature, and do not have a true zero point where there is nothing left of the attribute. One advantage of interval scales is that statistical calculations can be performed.
The main concepts related to quality of measurement are
Validity and reliability
If a scale is incorrectly calibrated and systematically subtracts 10 kg from each weight reading but gives consistent readings, this scale would be
Reliable and invalid
A weight scale that systematically subtracts 10 kg from each respondent’s true weight is reliable because it measures the weight consistently but is not valid because each measurement is 10 kg lower than the true weight.
Internal validity is
The degree of certainty that changes in variables in an experimental study are caused by the treatment or intervention
Internal validity refers to the degree to which we are certain that changes in variables in a study are caused by the treatment or intervention. It refers to how well the study was run, including the appropriateness and rigor of research design, how indicators were measured, what wasn’t measured as well as factors such as selection bias and attrition.
Criterion validity
Criterion validity measures how well one instrument compares against another instrument or a “gold standard” that is known to be close to the truth. To assess criterion validity of a biological or clinical screening test, one would conduct a sensitivity and specificity assessment where the test is compared against a gold standard method of diagnosis.
Sensitivity is
The ability of the test to correctly identify true positives
Sensitivity and specificity assements are an approach to measuring validity and not reliability.
Specificity is
Specificity is defined as the ability of a test to correctly identify those who DO NOT actually have the disease, otherwise known as the true negatives.
A false negative is
Someone who truly has the disease and who tests negative with the test of interest
The negative predictive value
& equation
The negative predictive value (NPV) is the proportion of people who tested negative who actually don’t have the disease
= TN/(TN+FN)
Will the sensitivity and specificity of a test always be consistent, regardless of context/population?
No, it will vary in different settings and populations
Reliability refers to
Reliability refers to the consistency of a measure, or the degree of stability exhibited when a measure is repeated under identical conditions.
Test-retest reliability assesses
Variation in measurements taken by a single person or instrument under the same conditions
Inter-observer reliability
Is a measure of agreement between two or more observers measuring the same concept
An individual or population’s overall level of health is an example of
Health status
Health status is an individual or population’s overall level of health.
Public health concerns
The overall health of groups and the distribution of health within the group
Which of the following is true regarding mortality?
Mortality rate is not a useful indicator.
Neonatal mortality rate is the number of deaths of liveborn children under 28 days of age per 100,000 live births per year.
The first known mortality data was collected by John Snow in London.
Neonatal mortality rate is the number of deaths of liveborn children under 28 days of age per 1,000 live births per year.
The crude mortality rate takes into account age.
Neonatal mortality rate is the number of deaths of liveborn children under 28 days of age per 1,000 live births per year.
Traditionally mortality has been the most important indicator of the health status of a population. The first known systematic collection of mortality data was developed by John Graunt, and crude mortality rate does not take into account age. Neonatal mortality rate is typically expressed per 1,000 live births
What is a key fact to remember about incidence?
Incidence focuses only on new cases of a disease that occur within a certain period of time
Incidence refers to the number of new cases of a disease that occur during a specified period of time in a population at risk for developing the disease. The denominator, therefore, only includes people who have the potential to develop the disease in the future.
In a population of 5,000,000 people at a given time 33,000 people have Disease X. The prevalence is
0.66%
The prevalence is calculated by dividing 33,000 by 5,000,000 and multiplying by 100. This yields a prevalence of 0.66%.
Disease Y occurred at a rate of 149/1000 per year in a population of women aged 15-49 who did not previously have Disease Y. This is an example of..
A case fatality rate An incidence rate A crude disease rate A point prevalence rate A prevalence rate
An incidence rate
The information describes the rate of occurance of a disease within a population who is at risk for developing the disease in the future and thus is an incidence rate.
In City A, which has a population of 100,000, there are 200 new cases of tuberculosis reported between January 1, 1998 and December 31, 1998. What is the annual incidence of tuberculosis for this population? Assume that the number of existing cases on January 1 1998 was negligible.
2 per 1,000 per year
There are 200 new cases per year in a population of 100,000. This is equivalent to 2 cases per 1,000 per year.
Which of the following is a true statement?
The risk of maternal death is most often measured by the neonatal mortality rate
The severity of a disease is measured via incidence and prevalence
Neonatal mortality rate is usually higher than infant mortality rate
A disease’s burden relies on knowing only the proportion of people who have that disease in the population
All of the statements are false. Burden of diseases in a population depends on their frequency (incidence or prevalence), severity (mortality and extent of serious morbidity), consequences (health, social, economic) and type of people affected (gender, age). Risk of maternal death is most often assessed by the maternal mortality ratio. Neonatal mortality, which is defined as the probability of dying within the first month of life, is always less than the infant mortality rate, which is defined as the probability of dying within the first year of life.
Prevalence includes:
Existing cases present at a single point in time
Prevalence is the number of affected individuals in a population at a specific time divided by the number of persons in a population at that time. Prevalence is essentially a snapshot of the proportion of the population who has the disease at a given moment in time.
In Town B, the prevalence of disease Z is 4.6%, and the average duration of the disease is 4 years. What is the incidence rate of the disease?
11.5 per 1,000
In a steady state where rates are not changing and in-migration equals out-migration, prevalence and incidence can be related by the following equation: prevalence = incidence x duration of disease. To calculate incidence rate with the given information, 0.046 is divided by 4 and then multiplied by 1,000 to give a rate per 1,000.
If the prevalence of disease J has been roughly constant for the last 10 years, and assuming no other changes, what would be the impact on prevalence of a program that reduces incidence of disease J?
It would decrease prevalence of the disease
In a steady state where prevalence is stable, a decrease in incidence of a disease would mean that there are fewer new cases of the disease within a population. This would ultimately lead to a reduction in the prevalence over time.
Which of these is a distal determinant of health?
Individual behaviors
Diet
Environmental Factors
Tobacco use
Environmental factors
Distal determinants have a more indirect effect on a person’s health and tend to be background factors in the sociocultural and environmental context that can predispose people to greater or lower health risks. Examples include cultural factors, education level and the physical environment.
DALYs?
DALYs (Disability Adjusted Life Year) assess the number of years lost due to ill health, disability or early death. 1 DALY is equivalent to 1 year of healthy life lost.
QALYs
QALYs (Quality Adjusted Life Year), on the other hand, focus on the quality and quantity of life lived.
population attributable fraction
Population attributable fraction assesses the percent reduction in incidence rate of a disease that would occur if a certain risk factor were to be eliminated. The population attributable fraction of type 2 diabetes due to physical inactivity is the proportion of incident cases that would be reduced if physical inactivity was eliminated from the population
Which of the following statements is true regarding questionnaires?
Instruments exist but not all have been validated for use in every setting
Instruments exist and can be used in any setting
Questionnaires must always be created from scratch
A wide variety of questionnaires for health research exist and are available for use. However, not all of these tools have been validated for every context, so it is important to review evidence for the existing tools before making a decision.
Which of the following statements is true?
Vital signs are measurements that are invasive.
Interview guides are used for collecting qualitative data
The Patient Health Questionnaire (PHQ) is used for assessing overall quality of life
Field notes include questions and responses the enumerator can either check off or fill in
Interview guides are used for collecting qualitative data
Field Notes are the least structured tool for observational data collection and do not include any preset questions or responses. The PHQ is a tool for assessing depression. Vital signs are a non-invasive measurement.
Anthropometry
Anthropometry is the measurement of the human body. Anthropometric measures include height, weight, mid-upper arm circumference (MUAC), head circumference and waist and hip circumference. In anthropometry, inter-observer variability can be an issue.
MUAC stands for
Mid upper arm circumference
which is measured to assess malnutrition.
Which of the following statements is true about dried blood spots?
Usually 5 to 6 spots are collected per person
They involve collecting finger prick blood on a glass slide
Most clinical assays that use venous blood can be performed with dried blood spots
They require cold chain
Usually 5 to 6 spots are collected per person
Dried blood spots do not require cold chain or glass slides. They involve depositing drops of finger prick blood onto a special type of filter paper card and allowing it to air dry. Usually 5 to 6 drops are collected per person, however the number and range of assays available for dried blood spots is limited.
Point of care tests
Are tests that give rapid results on the spot and can be performed by staff with minimal training.
Health
Health is a complex multi-dimensional concept. It is usually measured in terms of the absence of physical pain, physical disability, or a condition that could lead to death or disability. As well as in terms of the presence of emotional and mental well-being, and adequate social functioning.
proximal determinants
viewed as having the most direct effect on health status. Examples include individual behavior such as hand-washing, alcohol and cigarette use, beliefs, attitudes, genetics, and biology.
Distal determinants
have a more indirect effect on a person’s health and tend to be background factors in the sociocultural and environmental context that can predispose people to greater or lower health risks.
Examples include cultural factors, education level, and the physical environment. For example, low income may lead to poor nutrition and housing.
Health status
refers to an individual or population’s overall level of health, taking into account various factors such as life expectancy, amount of disability, and levels of risk factors for diseases.
Health outcomes
Numerous similar definitions of health outcomes exist. All definitions include a change in health status. Some definitions require that the change be the result of an intervention.
health indicator
a characteristic of an individual, population, or the environment that can be measured directly or indirectly. And can be used to describe aspects of the health of individuals or populations.
Examples of health indicators include mortality rate, HIV prevalence, percent of children under five who are malnourished, percent of pregnant women delivering in a health facility, consumption of fruit, estimate of work capacity
Nominal scales
also called qualitative scales, categorize items based on their name or other quality of classification, and thus have no mathematical interpretation.
For example, the variable sex generally has two attributes or categories–male and female. We can assign females a value of 1 and males a value of 2, but these numbers only tell us that males and females are different, but nothing else about that difference. Males are not one unit more of sex than females. The numbers simply represent different categories.
Ordinal scales
include variables that have categories with a rank order and a direction, but that have unequal intervals between items. Thus, there is some situational relationship between the categories of the variable .
For example, consider a variable for education level with three categories–primary school, secondary school, and post-secondary school. Respondent with secondary school level education has completed more schooling than a respondent with primary school level of education. However, the distances between categories or items on this scale do not have any particular meaning.
Interval scales
have a meaningful order and equal intervals between values on the scale, but no true zero point where there’s nothing left of that attribute. The distances between the items on the scale have importance. A difference of one unit is the same at any point along the scale.
An example of an interval scale is the temperature on the Celsius scale. The difference between 22 and 30 degrees Celsius is the same as the difference between 2 and 10 degrees Celsius. However, because the zero value on the Celsius scale doesn’t reflect the absence of temperature, we can’t calculate ratios or proportions.
Ratio scales,
the most precise level of measurement, have all of the qualities of interval scales and in addition, have a true zero point at which nothing is left of the attribute.
Ratio scales are the most common type of measurement in the sciences and include examples such as weight, length, and grade point average or GPA.
Validity
efers to the degree at which a method, test, or measuring device measures what it intends to measure.
Internal validity
refers to the degree to which we are certain that changes in variables in a study are caused by the treatment or intervention.
It refers to how well the study was run, including the appropriateness and rigor of research design, how indicators were measured, what wasn’t measured, as well as factors, such as selection bias and attrition.
External validity
refers to the degree to which study findings are applicable to other populations and contexts. This is also called generalizability.
Face validity
an assessment of whether a measure appears to measure the target concept.
Content validity
a subjective assessment of the extent to which a measure adequately represents all facets of a concept.
Criterion validity
measures how well one instrument compares against another instrument, or a gold standard, that’s known to be close to the truth.
Criterion validity is generally quantified by assessing correlation coefficients between the survey measure and the criterion or gold standard. The higher the correlation, the greater the validity of the instrument or measure.
predictive validity.
Criterion validity can be assessed by comparing both instruments at the same point in time, which is called concurrent validity, or by giving the target test first and measuring whether it predicts the findings of the gold standard tests at a later time.
construct validity
refers to how well the scale or instrument measures the construct that it is intending to measure.
It requires assessing relationships between a study variable and other variables that are being measured in the study that are known via theory or other evidence to be related. The more often these relationships are confirmed, the greater the construct validity of the survey variables.
Sensitivity & equation
defined as the ability of a test to correctly identify those who actually have the disease. These are also known as true positives.
calculated as the number of true positives divided by the sum of the true positives and false negatives.
TP/ (TP + FN)
Specificity & equation
defined as the ability of a test to correctly identify those who do not actually have the disease. These are also known as the true negatives.
calculated as the number of true negatives divided by the sum of the true negatives and false positives.
TN/(TN + FP)
positive predictive value
the proportion of people who tested positive who actually have the disease.
Or in other words, it’s the probability that someone with a positive test result actually has the disease
negative predictive value
the proportion of people who tested negative who actually don’t have the disease.
Or in other words, the probability that someone who tests negative actually doesn’t have the disease.
Tradeoff between sensitivity and specificity
It is important that both characteristics are high. However, sometimes it may be more important to have very high sensitivity, such as when a missed diagnosis may lead to death of a patient, even at the expense of high specificity, where false negatives may mean drug may be prescribed and wasted and some other diagnosis may be delayed.
Sensitivity and specificity in different populations
Sensitivity and specificity may vary in different settings and populations. And it’s important to look for assessments in a population very similar to your population of interest. If no published literature on sensitivity and specificity of the test of interest exist, consider assessing it yourself before deciding to use it.
Reliability
refers to the consistency of a measure or the degree of stability exhibited when a measure is repeated under identical conditions. Essentially, this refers to how consistent the answers people give to a question are when they’re asked at different points in time, assuming that no real changes have occurred.
three approaches to assess reliability
test/re-test reliability
inter-observer or inter-rater reliability
internal consistency.
test/re-test reliability
assessing whether the same question or measure yields consistent results at different points in time
It can, for example, be measured by administering the same survey to the same individuals at two different points in time to assess the correlation between the two sets of scores. This is usually quantified with a correlation coefficient. If the correlation between the repeated measures is higher than 0.7, than the test retest reliability is high.
inter-observer or inter-rater reliability
assessing whether different people collecting data on the same question get similar answers - inter-observer or inter-rater reliability
could occur due to variations in measurement techniques between observers, as well as variations in interpretations of measurements.
Inter-observer reliability can be assessed with co-relation coefficients. In general, a correlation between or among enumerators of 0.8 or higher is targeted.
internal consistency.
assessing whether different questions that are trying to measure the same concept are correlated.
Some ways to improve reliability include having clear and detailed measurement protocols, ensuring thorough training and sufficient practice, conducting regular evaluations, using appropriate measurement tools or equipment, and ensuring that equipment is properly maintained.
Intra-individual variability
defined as variations in an individual’s physiological or behavioral characteristics over time.
Intra-observer variability
random error in measurement made by the same observer on the same subject
three approaches to measuring population health.
○ Aggregating health outcome measurements made on individuals into summary statistics, such as population averages or medians.
○ Assessing the distribution of individual health outcome measures in a population and among specific population subgroups.
○And measuring the function and well-being of the population or society itself, as opposed to individual members.
The relative importance, or burden, of different diseases in a population depends on several key factors.
○ Their frequency, which refers to incidence or prevalence.
○ Severity, which depends on mortality and the extent of serious morbidity.
○ Consequences, which could be health, social, or economic consequences.
○ And the type of people affected, their gender, and age, and so on.
mortality
The number of deaths in a population over a specific period of time, usually a year, divided by the size of the population is the population’s crude mortality rate.
The mortality rate can serve as a measure of disease severity, as well as an indicator of whether interventions have been successful over time.
Common mortality indicators include
include annual mortality rate, infant mortality rate, child mortality rate, neonatal mortality rate, and the maternal mortality ratio. Other indicators include case fatality and proportionate mortality.
life expectancy at birth
Life expectancy at birth is the average number of additional years a newborn baby could expect to live if there is no change in the rate of death during the rest of the newborn’s life.
Infant mortality rate
It refers to the number of deaths of infants less than one year of age per 1,000 live births in a given year. Thus, it measures how many infants younger than one will die for every 1,000 who are born in a given year
The lower the rate, the better the health status of the population.
Under 5 mortality rate
the number of deaths of children under five years of age per 1,000 live births in a given year, is also a commonly used indicator.
Maternal mortality ratio
captures the risk of death associated with pregnancy, labor, and the postpartum period
morbidity
refers to sickness or departure from a psychological or physiological state of well-being, whether subjective or objective.
Morbidity indicators
Morbidity indicators show the burden that a disease has on a population. It can be measured using rates which tell us how fast the disease is occurring in a population and proportions, which tell us what fraction of the population is affected.
Incidence
refers to the number of new cases of a disease that occur during a specified period of time in a population at risk for developing the disease.
it captures disease in a population of individuals who don’t have the disease at baseline but have the potential to develop the disease within a specified time period.
The critical points to remember about incidence include that the focus is only on new cases. And therefore, this indicator is a measure of risk.
Incidence rate
Number of new cases of disease in a population in a specified period of time divided by the total person-time x 1000
*the answer will be written as # per 1000 person years
total person-time
Sum of time periods of observation of each person who has been observed for all or part of the entire time period
How do researchers deal with individuals lost in the follow-up when estimating the total person-time?
Typically, researchers assume that individuals lost to follow up were on average disease-free for half of the specific follow-up period.
Prevalence
the number of affected individuals in a population at a specific time divided by the number of persons in a population at that time.
Prevalence is essentially a snapshot of the proportion of the population who has the disease at a given moment in time. Prevalence does not tell us when the disease developed, so it comprises people who are newly diagnosed along with people who’ve had the disease much longer.
Prevalence equation
Number of cases divided by the number of people in the population
(vs. Incidence rate = number of NEW cases / total person time x 1000)
Point prevalence
the proportion of people who had the disease at a specific point in time, such as today
period prevalence
refers to the proportion of people who had the disease at any time during a certain period, such as a calendar year.
Typically, for the period prevalence calculation, the denominator consists of the average or mid-interval population.
Prevalence & incidence relationship (equation)
Prevalence equals incidence times duration of disease.
How are the sum of DALY’s used?
The sum of DALYs across a population is the overall burden of disease in that population, or put another way, the difference between the population’s current health status and an ideal where people live free of disease and disability.
DALYs were designed to allow international comparisons. DALYs for a specific disease or health condition are calculated by adding the years of life lost due to premature mortality, or the YLL, and the years of life lost due to disability, or the YLD.
years of life lost due to premature mortality, or the YLL
takes into account the age at which the death occurs by giving higher weights to earlier deaths
the years of life lost due to disability, or the YLD
YLD is a measure of healthy years lost due to disease or injuries.
To calculate the YLD for a specific disease during a specific time period, multiply the number of incident cases of that disease by the average duration of the case until death or cure, and also by a factor that reflects the severity of the disease.
QALY equation
we multiply the length of life, usually in years, by the quality of life.
Quality of life is measured by…
Quality of life is measured on a scale of 0-1 with 1 being equivalent to perfect health and 0 being equivalent to death. These qualities of life valuations, or health utilities, can be generated using a number of approaches, including patient questionnaires.
GINI index
The GINI index measures the extent to which income distribution among a population deviates from the theoretical line of equality in which each segment of the population earns the same proportion of the total income.
Concentration index
Adapted version of the GINI index. It indicates the extent to which a health indicator is concentrated among disadvantaged or advantaged groups in a population - to capture the distribution of health inequality across populations
Population Attributable Fraction
assesses the percent reduction in incidence rate of a disease that would occur if a certain risk factor were to be eliminated.
self-reported measures
respondents provide information on their own knowledge, attitudes, and behaviors.
Observational measures
involve careful systematic observation of behaviors, factors related to health, events, or physical characteristics of the environment.
Biological and clinical data
involves collection of physical measures such as weight, physiological measures, as well as testing for presence or absence of diseases.
Electronic measures
involve utilizing technologies to record occurrence of specific target behaviors or events