Reading

Schmidt & Brown (2021). Evidence-Based Practice for Nurses:

• Chapter 9: Epidemiologic Designs: Using Data to Understand Populations

Objectives

• Define epidemiology

• Describe the use of epidemiology in nursing practice

• Describe count data, ratios, proportions, and rates

• Define and compute prevalence

• Define and compute incidence

• Explain descriptive characteristics of person, place, and time when examining the distribution of disease in a population

• Describe descriptive study designs, including case reports and series, ecologic studies, and cross-sectional studies

• Construct a 2 × 2 data table

• Calculate a prevalence ratio for a cross-sectional study design

• Describe analytic study designs, including case-control studies, cohort studies, and intervention studies

• Calculate an odds ratio for a case-control study design

• Calculate relative risk for a cohort study design

Epidemiologic investigations can be descriptive or analytic:

• Descriptive epidemiology

  • = Examination of the distribution of disease in a population in terms of person, place, and time

  • distribution = the pattern of disease occurrence in and among populations or subgroups

• Analytic epidemiology

  • = Investigation of the determinants of disease

  • determinants = factors that are capable of bringing a change in health

Epidemiologic Measures

• Count Data

  • = the raw number of health phenomena under investigation in epidemiology

  • Examples of health events:

    • births

    • cases of a disease

    • deaths

  • Not particularly useful when comparing populations of different sizes

    • Example: A country with a large population will have more births than will a country in a small population regardless of other factors associated with birth rate

    • When population differs in size, epidemiologists use ratios to compare and contract health outcomes across populations

• Ratios

  • = The highest level of measurement, which involves numeric values that begin with an absolute zero and have equal intervals; in epidemiology, a mathematical relationship between two numbers

    • Formula = a/b

  • Example: of 1,000 patients with acute myocardial infarction (AMI), 600 were male and 400 were female. The sex ratio for AMIs is:

    • Number of male cases / Number of female cases = 600 / 400 = 1.5 : 1 male to female

• Proportions

  • = A type of ratio where the numerators is included in the dominator

    • Formula = a/(a + b) multiplied by 1,000

  • The numerator and denominator have an association

    • Numerator = the number of cases, deaths, or events

    • Denominator = the population being studies

  • Used in epidemiology to describe measures, such as:

    • prevalence

    • cumulative incidence

    • case fatality rates

    • attack rates

  • Example: of 300,000 children, 3,800 were diagnosed with an autism spectrum disorder and 296,200 were not. The proportion of children diagnosed with autism spectrum disorder is:

    • a/(a + b) * 1,000 = 3,800 / (3,800 + 296,200) * 1,000

= 3,800 / 300,000 * 1,000

= 0.01266 * 1,000

= 12.66 per 1,000 children

= meaning the proportion of autism spectrum disorder in this population of children is 12.66 per 1,000 children

• Rates

  • A measure of disease frequency in a defined population over a specified period of time

  • When calculating a rate:

    • Numerator = the number of people affected

    • Denominator = the entire population

  • Used in epidemiology to describe measures, such as:

    • density

    • crude mortality rates

    • fertility rates

  • Examples: of 5,610 babies born in Lake County in 2017, 59 babies died before reaching their fist birthday. The infant mortality rate in 2017 was:

    • Number of infant deaths during time period / Number of live births during time period * 1,000 = 59 / 5,610 * 1,000

= 10.5 per 1,000 live births

• Prevalence

  • = the number of existing cases of disease present in the population

  • An indicator of the extent of a health problem in a population and is used for planning related to healthcare needs for communities

  • May be expressed as:

    • a number

    • percentage

    • proportion

    • rate

  • Formula for prevalence rate per 100,000 at a specified time:

    • Prevalence = Number of existing cases of a disease / total population * 100,000

  • Example: In 2018, of the 75,877,700 people younger than age 19 years, 210,000 people were estimated to have been diagnosed with diabetes. The prevalence of diabetes in the United States in 2018 in this age group was:

    • Number of existing cases / Total population * 100,000 = 210,000 / 75,877,700 * 100,000

= 0.0027 * 100,000

= 270 per 100,000

• • Two different terms used to describe prevalence:

    • Point Prevalence

      • = The number of existing cases of disease in a population at a particular point in time

      • Example: Calculate the point prevalence of flu on a college campus today and then compare that to the point prevalence of flu on the same day in the previous year

    • Period Prevalence

      • = The number of existing cases of disease in a population during a specified period of time

      • The time period might be a week, a month, a year, or a number of years

      • Example: Calculate the period prevalence of flu from the previous year's flu season to make decisions about how many flu shots to purchase for the current year

• Incidence

  • = The number of new cases of a disease in a population during a specified period of time

  • The incidence rate is a measure of disease occurrence and is used for investigating the causes of disease

  • Use incidence rates to investigate risk of disease in populations or subgroups related to many factors, such as age, gender, occupation, and exposures

  • Formula: Divide he number of new cases that occur during a time period by the number of individuals in the population

    • Incidence = Number of NEW cases of a disease / total population * multiplier

  • Example: In 2017, of the 6,666,818 people living in Indiana, there were 143 confirmed new cases of Lyme disease. The incidence of Lyme disease in Indiana in 2017 was:

    • Number of new cases / total population * 100,000 = 143 / 6,666,818 * 100,000

= 0.000021 * 100,000

= 2.14 per 100,000

• The example that the 4th edition textbook used was: Between 2000 and 2016, the prevalence of persons living with Lyme disease was 970. In 2016, of the 6,483,802 people living in Indiana, there were 81 confirmed new cases of Lyme disease. The incidence of Lyme disease was:

  • Number of new cases / Total population at risk * 100,000 = 81 / (6,483,802 – 907) * 100,000

= 0.000012 * 100,000

= 1.25 per 100,000

• • You may notice that the formula is different. I asked Dr. Schmidt, the author of the textbook, about this, and here is her response (N. Schmidt, personal communication, October 22, 2021):

Hello, Dr. Niitsu: Thanks so much for using our textbook. You have asked an interesting question about something that the chapter author and I have discussed over the years.

In the literature, there seems to be two formulas for calculating incidence. In the past edition, the denominator reflected the "individuals at risk." This is calculated by subtracting the current prevalence from the population in the denominator. The rationale for using this formula was that it provided a more accurate number for incidence because the number of people who already have a disease would not be at risk because they already have the disease. This is also the formula that is more commonly used in epidemiology practice and at the graduate level.

In the current edition, we decided to go with a formula that uses the "total population" (rather than the population at risk). We based this decision on a few reasons. First, it is an easier formula for students to use. It is more like the formula for calculating prevalence. Second, it is the formula that is usually found in more general books about epidemiology, such as global health books. For example, the introductory global health textbook by Richard Skolnik, a respected author, uses the simplified formula. But a third reason we considered is the fact that, regarding COVID-19, it appears that those who have already had the infection can become reinfected. Thus, they are still at risk and it probably is a bit misleading to remove them from the denominator.

From what I see in the literature, either formula can be used. For some diseases, like heart disease, once you have it you are no longer at risk because you now have a chronic disease, thus using the 4th edition formula may be more appropriate. But for other diseases, such as COVID-19, people can become reinfected. Having the disease doesn't necessarily mean that one is not at risk for getting it again at a later time. In light of what we are learning about COVID-19, perhaps that makes the 5th edition formula more appropriate.

Dr. Buckenmeyer and I have had interesting conversations about which formula to use. And if you go back to other editions, you'll see that we have used both. In the end, we chose the formula for this edition because we wanted to keep things simple. This is a chapter that just introduces epidemiology, and the simpler formula is sufficient for our purposes. For your purposes, you could use whichever you prefer. I think that professors are able to say they disagree with something in a text and teach students other content.

I hope this explanation helps. If you have any more questions, please let me know.

• Purpose

  • To identify subgroups in populations that may have the highest risk for a specific disease or outcome

• Used to:

  • Find clues about potential causes of disease so that nurses can generate hypotheses about the relationship between exposures and diseases or other health-related outcomes

  • Measure disease frequency by person, place, and time to determine why disease occurs more frequently under certain conditions

• • • 3. Short-term changes

      • = Brief, unexpected changes in disease distribution

      • 3 key terms

        • Endemic

          • The expected occurrence of a particular disease within a community or population

        • Epidemic

          • A widespread occurrence of a disease in a community or population that is in excess of what is expected

        • Pandemic

          • An epidemic that has spread worldwide

      • Example: rubella in Japan

        • Endemic until the early 2000s

        • The number of reported rubella cases remained at record low levels until 2010 (n = 87)

        • A few outbreaks reported in the workplace among adult males in 2011 (n = 378)

        • The number of rubella cases sharply increased in 2012 (n = 2,392)

        • A total of 5,442 rubella cases reported from 01/01/2013 to 05/01/2013

        • Health officials deemed that Japan was experiencing a nationwide rubella epidemic

      • Example: Coronavirus

        • Article by The New York Times on 02/02/2020: "Wuhan Coronavirus Looks Increasingly Like a Pandemic, Experts Say"

        • Announcement by the University of Washington on 01/14/2022: Winter 2022 Designated Extraordinary Circumstances Quarter

          • The Lancet: COVID-19 will continue but the end of the pandemic is near

            • Institute for Health Metrics and Evaluation (IHME)

        • Announcement by the University of Washington on 01/24/2022: The UW is returning to largely in-person classes Jan. 31

Descriptive Study Designs

Descriptive study designs are hypothesis-generating and are used to examine different types of phenomena by using three study designs: case reports or series, ecologic studies, and cross-sectional studies:

• 1. Case Series Studies

  • = Used to describe rare diseases or outcomes

  • Purpose: to describe new diseases, explain a change in disease patterns, or alert the healthcare community to unusual signs and symptoms in:

    • An individual patient (case report)

    • Rare findings among a few patients (case series)

      • Example: The first official report of what later became known as the AIDS epidemic in the Morbidity and Mortality Weekly Report

        • In the period October 1980 - May 1981, 5 young men, all active homosexuals, were treated for biopsy-confirmed Pneumocystis carinii pneumonia at 3 different hospitals in Los Angeles, California. Two of the patients died. All 5 patients had laboratory-confirmed previous or current cytomegalovirus (CMV) infection and candidal mucosal infection... The diagnosis of Pneumocystis pneumonia was confirmed for all 5 patients antemortem by closed or open lung biopsy. The patients did not know each other and had no known contacts or knowledge of sexual partners who had had similar illnesses. Two of the 5 reported having frequent homosexual contacts with various partners. All 5 reported using inhalant drugs, and 1 reported parenteral drug abuse. Three patients had profoundly depressed in vitro proliferative responses to mitogens and antigens. Lymphocyte studies were not performed on the other 2 patients.

  • Advantage

    • Describe unusual signs and symptoms so that healthcare providers may be able to identify commonalities among patients when a new disease appears in a population

    • Helpful in identifying when a current disease mutates

  • Disadvantage

    • The lack of a comparison group

      • Without a comparison group, there is no mechanism to test hypotheses

• • Advantage

    • Efficient and relatively inexpensive

    • Can be used to examine a number of different phenomena, including behaviors, symptoms, diseases, and health status

    • Can be not only a single exposure and disease but can also examine multiple exposures and diseases simultaneously

  • Disadvantage

    • Temporal ambiguity

      • Unable to ascertain if the exposure preceded the disease because data about exposure and disease status are collected at the same time

    • Unable to determine whether the specific exposure caused the specific diseases

      • Because prevalence is a function of incidence and duration of time, it's difficult to distinguish determinants of the cause of the disease from determinants of survival with the disease

Analytic study designs are hypothesis-testing designs and are used to test the association between exposure and disease. The figure below describes the epidemiological study designs so intuitively (slide credit: Professor Wade):

1. Case-Control Studies

• = A type of retrospective study in which researchers begin with a group of people who already have the disease; studies that compare two groups

  • those who have a specific condition

  • those who do not have the condition

• Retrospective because individuals are recalling their past

• Key terms

  • Case

    • = Individuals who have the disease

    • Often asked to recall exposures up to 1 year prior to diagnosis

  • Control

    • = Individuals who do not have the disease of interest but who are at risk for developing the disease

    • Asked about their history of exposure up to the point of entry into the study

• Odds Ratio (OR)

  • = The statistic reported when epidemiologists conduct a case-control study

  • Formula

    • OR = AD / BC

  • Interpretation

    • OR = 1 means the probability of disease among the exposed and the nonexposed is identical

      • There is no association between the exposure and the disease

    • OR > 1 means there is a greater probability of disease among the exposed

      • There is an association between the exposure and the disease

    • OR < 1 means there is decreased probability of disease among the exposed

      • A protective effect

• Advantage

  • Expedient

  • Require a small sample size

  • Relatively inexpensive

  • Examine rare diseases and situations that involve individuals who have had many exposures

  • Low risk of attrition

    • because participants are asked to recall exposures and thus do not leave the study

• Disadvantage

  • Cannot measure incidence

  • Able to examine only one disease rather than many

  • Not conducive to measuring rare exposures

  • Recall bias is a threat because of the retrospective design

2. Cohort Studies

• = Studies using two or more groups; epidemiologic designs in which participants are selected based on their exposure to a determinant

• " natural experiment"

  • Example: Nurses' Health Study

    • Goal = to examine diet and lifestyle risk factors and their associations with cancer, heart disease, and other chronic diseases

    • Initiated in 1976 with 120,000 female nurses between the ages of 30 and 55 years

    • Nurses continue to be followed and outcomes have been added

• Case-control studies vs. Cohort studies

  • Case-control studies

    • Individuals are selected based on the presence or absence of disease

  • Cohort studies

    • Individuals are selected based on their exposure

    • Select a sample group that is representative of the target population, then assign individuals to groups based on their exposure status

• Prospective designs

  • Follow individuals from the time they enroll in the study until they develop the disease

• Retrospective designs

  • Determine a historical exposure and then follow the sample forward to the present time to determine whether the disease is present

• Relative Risk (RR)

  • = The statistics reported by epidemiologists when they conduct a cohort study

  • Formula

    • PR = [A/(A+B)] / [C/C+D)]

  • Interpretation

    • PR = 1 means the probability of disease among the exposed and the nonexposed is identical

      • There is no association

    • PR > 1 means there is a greater probability of disease among the exposed

      • There is an association

    • PR < 1 means there is a decreased probability of disease among the exposed

      • A protective effect

• Advantages

  • Measure incidence and study many outcomes

  • Large sample size --> a good design to study rare exposures and can readily establish that an exposure preceded the disease

  • Less vulnerable to recall bias

• Disadvantages

  • Tend to be expensive because of the need for a large sample size and their longitudinal nature

  • Require a significant amount of time depending on the disease of interest

  • Impractical for rare outcomes

  • Thread of mortality because participants may drop out, move, or even die during the course of hte study

  • Exposures during the course of a cohort may change

3. Intervention Studies

• = In epidemiology, a study that has a treatment that can be manipulated by the researcher

• The statistical test used for intervention studies is the RR

• Advantage

  • Control over the exposure

  • Randomization to control for potential confounding variables

  • Blinding to reduce bias

• Disadvantage

  • Cost

  • Time-consuming

    • Participants may become noncompliant or withdraw from the study, which also biases outcomes

  • Hawthorne effect

    • This effect occurs when participants in an intervention study alter their behavior because they know that they are being observed

  • Not all exposures are able to be manipulated for ethical reasons