Chapter 10: Epidemiology and Disease – Flashcards

A. Epidemiology a. Epidemiology is the study of disease in a population; it is considered the science of public health i. Studies the determinants and distribution of disease ii. Assumes disease does NOT occur at random 1. Therefore it assumes that causes can be identified 2. Also assumes exposure to the cause occurs BEFORE the disease iii. If causes can be identified, disease may be prevented b. It is a collection of study designs and data analysis methods for calculating disease rates i. It is used to link causes of disease to the development of disease and to determine how a disease spreads or appears in a population. c. Epidemiology looks at both determinants (causes) and distribution of disease in a population i. Epidemiology was first used to study outbreaks of infectious disease, then to study biological or behavioral causes of disease (e.g., Framingham Heart Study). A more recent branch of epidemiology is "pharmacoepidemiology." B. Pharmacoepidemiology a. Pharmacoepidemiology is a subset that focuses on medication-related disease i. So, it is the use of epidemiological methods to study medication use-related issues in a population including: 1. Medication use patterns 2. Distribution of adverse reactions in a population 3. Identifying source of a contaminated product 4. Impact of medication on disease survival rates 5. Phase IV (post-marketing surveillance) studies for rare adverse events

i. Monitor health of a population ii. Respond to emerging public health problems iii. Promote research and use of evidence-based interventions iv. Evaluate effectiveness of a program v. Develop public health policy and law vi. Set funding priorities for research and intervention programs

a. The following are measures of disease frequency: i. Prevalence (total number of cases) ii. Incidence (number of new cases) iii. Mortality (number of deaths) a. EXAMPLE: In the past month, Town A reported five new cases of HIV/AIDS. This brings the total number of HIV/AIDS cases this year to 26. In Town B, there were 10 new cases and over 100 total cases during the same time periods. • We use these terms all the time when we are referring to diseases. o It is worth noting that the counts of new or total cases or deaths by themselves provide limited information. Most of the time, we are using these terms to refer to rates (number of cases per total population)

i. Need an indication of how the number of cases relates to the population so we calculate a prevalence rate ii. Prevalence rate 1. Total number of cases during a specified time period divided by the population count

i. Count of cases (total including new and ongoing) - the numerator ii. Population count (the whole population regardless of risk category) - the denominator iii. Period of time (may be one day, one year, or something else) • Rates may include cases from a single point in time (point prevalence) or from a period of time such as one year. Annual rates are probably the most frequently used time period.

i. Cumulative incidence rate (number of new cases in a specified time period divided by number of population that is at risk of the disease) 1. Differ from prevalence rates in that they consider: a. Only new cases that appear in a given time period b. Compare that number to the sub-group in the population that is actually at risk of the disease. ii. Incidence rates can be calculated in two ways: 1. The one we commonly call the "incidence rate" is actually the Cumulative Incidence Rate. This approach to calculating rate of new cases in at risk populations assumes the risk is occurring equally across time. 2. For chronic diseases or exposures, this measure works

i. Incidence rate or density (number of new cases in a specified time period divided by the total number of person-time when at risk) 1. Like the Cumulative Incidence Rate (CIR), it will consider only the new cases in its numerator. 2. Unlike the CIR, the denominator will be adjusted to account for the length of time members in the "at risk" population are actually at risk. a. This means the time element is integrated into the denominator as a "person-time" measure like "person-month" or "person-year." ii. For a chronic exposure or cause of disease, this adjustment is not critical since a person at risk at the start of the time period will be at risk of disease at the end. 1. However, if one is looking at incidence rates for disease caused by short term exposures or diseases, then failure to adjust the denominator to account only for the days at risk will result in a falsely low incidence rate. For example, to estimate the incidence rate of anaphylaxis from penicillin or the development of complications from pneumonia, where the actual time at risk may be several days or a couple weeks, failure to adjust the denominator down will result in a low rate.

i. EXAMPLE: Five new cases of HIV/AIDS were reported. This brings the total number of HIV/AIDS cases this year to 56; total population is 100,000 and population at risk of HIV/AIDS is 20,000. • Suppose the actual time at risk for any one individual is estimated at 183 days per year (= 0.5 years per individual). • Cumulative Incidence Rate calculation for one year: o Use 5 new cases and 20,000 at-risk individuals o CI = 5/20,000 (=25/100,000) • Incidence Rate calculation for person-years: o Use 5 new cases in numerator o Adjust denominator: 20,000p x 0.5y/p = 10,000 person-years o IR = 5/10,000 (=50/100,000)

i. There are three main categories of epidemiology studies 1. Descriptive 2. Analytical 3. Interventional

ii. Descriptive study designs: 1. Case report (or case series) a. These are reports that describe the events leading up to the development of a disease in an individual person. i. A CASE REPORT would be a single person ii. CASE SERIES would combine several individual but apparently similar cases into a single report. b. PROS: i. Provides a timeline of events leading up to the development of a disease. (Can help establish cause-effect hypotheses for future studies.) ii. Another advantage is the patient-level data that includes details that will be lost when aggregated into population-level data. c. CONS: i. Does not necessarily represent the whole population or a common problem. ii. Cannot compare populations to one another with this type of study.

2. Cross-sectional: a. This type of epidemiological study design collects and uses patient-level data about exposure and disease too. i. The data are collected across a number of subjects as one point in time. ii. Can aggregate individual data into population-level information. b. PROS i. Can collect data from a larger number of people than a case report. ii. Can collect data in a short amount of time. iii. Provides patient-level data which is needed to explore exposure-to-disease relationships c. CONS i. May be difficult to establish the exposure-disease timeline if the subjects' recall of events is poor or incomplete

3. Correlational: a. These study designs focus on population-level data and are used to compare two or populations or changes in one population over time. Results may identify populations with unusual levels of a disease that is worth further study. b. PROS: i. Allows for comparisons between populations or changes in one population over time. ii. Can be useful in identifying potential risk factors such as age, gender, or location. c. CONS: i. Lacks individual level data so cause and effect (exposure and disease) cannot be explored. ii. Also cannot determine if an exposure occurred before or after the disease onset.

1. Overall, they differ in these ways: a. Ability to tie cause to effect b. Ability to allow comparisons across time or with other groups

1. These designs must be able to: a. Establish that exposure preceded disease b. Determine if risk factor is necessary and/or sufficient c. Determine if risk factor is a direct or indirect cause d. Rule out confounding factors e. Eliminate or reduce systematic bias 2. Two basic types: Cohort and Case Control

a. Cohort i. Approaches the study question from the exposure side of the exposure-disease relationship. ii. Use two groups of subjects 1. Subjects selected on basis of exposure status a. Exposed b. Not exposed iii. May be prospective or retrospective iv. Seeks to determine whether an exposure affects the likelihood that a person will get the disease v. Results usually reported as Relative Risk

b. Case control i. Will start at the disease end of the exposure-disease relationship ii. Use two groups of subjects 1. Subjects selected on basis of disease status a. Disease b. No Disease iii. Retrospective only iv. Seeks to determine whether a person with the disease was more likely exposed to the risk factor than someone without the disease v. Results usually reported as odds ratios

• These studies take the observations of potential cause-effect relationships and systematically compare risks to a control group. o It is important to be able to establish that the exposure or risk factor was present BEFORE the disease occurred. The observational studies just described in earlier slides can assist with this determination.

• It is also key to figure out whether a specific risk factor must be present in order for disease to occur (necessary) and whether it alone is enough to cause disease (sufficient) or if it requires the presence of one or more co-factors. o For example, the presence of the mosquito breeds capable of carrying the malarial microorganism is not sufficient by itself to create an outbreak of malaria. The infective organism must also be present. The mosquito are necessary however to convey and transfer the microorganism to hosts. o Another example = with regard to flu, just being in close proximity to others who are sneezing or coughing is not sufficient to spread the flu if no one in the group is infected. The virus must also be present.

• Some risk factors indirectly cause disease, so it must be determined whether the link is direct or indirect. o For example, malnutrition can be an indirect cause of death because it contributes to electrolyte imbalances that lead to heart failure and ultimately cause death.

• Confounding factors can lead to misinterpretation of results if they are not accounted for and controlled. o For example, Erroneous associations may occur if the link between risk factor and disease is not clearly identified. For example, there are more cases of seasonal flu when people travel to and from work when it is dark out. Darkness is not really related to the spread of influenza, but the darkness is related to winter months when more people spend more time indoors together which does promote the spread of the viral illness. Amount of daylight is a confounding factor in this example.

• A final consideration is systematic bias - the kind introduced by the researcher when random processes are not used.

iv. Interventional study designs 1. Use same approach as experimental design a. Random assignment to study arms b. Researcher controls the exposure 2. Indirect method for learning more about a disease a. Used to test the effects of removing risk factors or adding protective factors on subsequent disease development b. Never used to directly test whether an exposure causes a disease • It is basically an experimental design where the researcher controls the exposure and determines exposure status for each participant in the study. • These are NOT used to test whether an exposure will CAUSE a disease.....rather, these designs are used for two specific things: o Test exposure to protective factors to study effects on disease development. o Test the impact of removing a known risk factor on disease development. • These studies will be based on preliminary studies performed using observational and analytical designs. • Potential disease risks are often identified in descriptive studies then studied in analytical and interventional designs