Investigating Public Health Data_Answers

Investigate data about a local public health problem

Applied Epidemiology – Problem Assessment

Background

Professionals working as applied epidemiologists typically work in local, state or federal public health agencies solving problems. The kinds of problems they solve vary greatly, extending from acute events like disease outbreaks and increases in motor vehicle accidents within a community to broader issues like the rise of cardiovascular disease in a particular population. Acute problems are usually brought to the attention of the public health team by one of their stakeholders. For example, at the federal level, the Centers for Disease Control often works with Ministries of Health or State Health Departments while a local public health organization might respond to a problem observed by a local hospital.

When information arrives from a stakeholder, epidemiologists must assess whether or not to launch an investigation and, if so, teams progress through a series of actions common across the field of public health (see Public Health Wheel below) to understand the problem before designing an intervention, putting it into effect, evaluating it, and then potentially revising the intervention and beginning the cycle again.

The Process

  1. Understand the problem. Gather contextual information
  2. Formulate the question you want to answer or situation you will respond to
  3. Investigate the problem. Determine what information you need to answer the question.
    • Data Collection. Consider what’s already available or whether you will need to collect your own primary data.
  4. Develop methodology to answer the question
  5. Develop analytic plan
  6. Develop plan for action/response/intervention. Availability of resources weighs in heavily when determining how to move forward.

The exercise:

Read through the following case study “Texarkana — Epidemic Measles in a Divided City” based on an investigation by Philip Landrigan, EIS ‘70 and assess the data during the investigation.¹

On Tuesday, November 3, 1970, the Center for Disease Control (CDC) in Atlanta received the weekly telegram of surveillance data from the Texas State Health Department. The telegram reported 319 cases of measles in the state during the previous week. In contrast, Texas had reported an average of 26 cases per week during the previous four weeks. In follow-up telephone calls, CDC learned from State health officials that 295 cases of measles had been diagnosed in the city of Texarkana, including 25 in children reported to have been previously immunized. An invitation to investigate the situation was extended to the CDC on November 4, 1970. An EIS officer departed for Texarkana early on November 5.

Background

Texarkana is a city of roughly 50,000 that straddles the Texas-Arkansas state line. Texarkana, Texas (Bowie County), had a population of 29,393 in the 1960 census; the population had been stable during the 1960s. Texarkana, Arkansas (Miller County), had a 1960 population of 21,088.

Although Texarkana is divided by the state line, it is a single town economically and socially. Persons of all ages on both sides of town have frequent contact. Churches, physicians, offices, movie theatres, and stores draw people from both the Arkansas and Texas sides of town. People cross the state line to attend social functions such as football games and school dances. Many families have friends and relatives who visit back and forth on both sides of town. Private nurseries and kindergartens receive children from both sides of town. The two sides of Texarkana, however, do have separate public school systems.

The Investigation

The investigators obtained names of cases from the health departments, physicians, school and nursery records. They conducted a door-to-door survey. They also asked families of cases for names of other cases. They used the same methods of case-finding and epidemiologic investigation on both the Arkansas and Texas sides of town.

The Clinical Picture

The illness was clinically compatible with measles. Typically, the patients had a 4- to 5-day prodrome with high fever, coryza (runny nose), cough, and conjunctivitis (red, irritated eyes) followed by the appearance of a bright maculopapular (red spots and areas) rash. The temperature usually returned to normal 2 to 3 days after appearance of the rash, while the rash persisted for 5 to 7 days.

Task 1

Though infants, adolescents, and adults were involved in the epidemic, the majority of cases occurred in children 1 to 9 years of age. Measles cases were not evenly distributed within the two counties. Table 1 displays the number of measles cases and population by age group for Bowie County, Texas and in Miller County, Arkansas.

Calculate the totals and attack rates indicated in Table 1. Then write a brief explanation of the differences in attack rates for the Texas and Arkansas counties and for preschool versus school-age children.

Table 1. Number of measles cases and population (1960 census) by age group and county, Texarkana outbreak, 1970.

Answers:

Fields to fill out for Task 1 are indicated in color.

Residence Urban/Rural Age Group # Cases Population Attack Rate
Bowie Co., Texas Total 1-4 242 4933 4.9%
5-9 251 6252 4%
1-9 493 11,185 4.4%
Miller Co., Arkansas Total 1-4 19 2671 .7%
5-9 6 3345 .17%
1-9 25 6016 .4%

Texas infection rates were significantly higher than the Arkansas infection rates. In Texas, preschool-aged children had a slightly higher rate of infection than school-aged children whereas in Arkansas county the two age groups had relatively similar rates of infection, though the preschool-aged children still contracted measles as a slightly higher rate.   

Task 2

Before this outbreak, the proportion of children vaccinated against measles on the Arkansas side was substantially higher than the proportion vaccinated on the Texas side. The Texas side had never had a community or school vaccination campaign for measles. In contrast, the Arkansas side had held mass community programs against measles for school and pre-school children in 1968 and 1969. Based on health department and physician records, investigators estimated that over 99% of children aged 1-9 years in Miller County, Arkansas had received measles vaccine prior to the outbreak. The overall vaccination level in Bowie County, Texas, was estimated to be 57%. In this outbreak, 27 of the measles cases in Bowie County and all 25 of the measles cases in Miller County gave a history of prior vaccination with live attenuated measles-virus vaccine. Parental history of vaccination was corroborated for all the cases by clinic or physician records. Local health authorities in both counties were very concerned that children who had previously received measles vaccine got the disease.

Calculate incidence and attack rates for the vaccinated populations in both counties and comment briefly on your findings.

Answers:

Fields to fill out for Task 2 are indicated in color.

Residence Urban/Rural Age Group # Cases Population Attack Rate for entire population Vaccinated population Immunized Cases Attack Rates for vaccinated
Bowie Co., Texas Total 1-4 242 4933 4.9%
5-9 251 6252 4%
1-9 493 11,185 4.4% 6375.45 27 .42%
Miller Co., Arkansas Total 1-4 19 2671 .7%
5-9 6 3345 .17%
1-9 25 6016 .4% 5955.84 25 .41%

Because of the population differences in the two counties, the attack rates for the vaccinated populations were very similar (.42% and .42% respectively) even though the incidence rate was dramatically different.

Task 3

Consider the use of a vaccine with 90% efficacy in four different hypothetical populations of 100 people each, with vaccine coverage of 0%, 20%, 60%, and 100%, respectively.

Complete table 2 below assuming that every unvaccinated person will be exposed to, and will develop, measles. What do you conclude about the relationship between coverage and number of cases vaccinated? What might your public health message be for these data?

Table 2. Hypothetical populations with vaccine coverage of 0%, 20%, 60%, and 100%

Answers:

Fields to fill out for Task 3 are indicated in color.

Population
A B C D
A. # persons in population 100 100 100 100
B. Vaccine efficacy (VE) 90% 90% 90% 90%
C. % population vaccinated (PPV) 0% 20% 60% 100%
D. # vaccinated (a x c) 0 20 60 100
E. # unvaccinated (a ! d) 100 80 40 0
F. # protected (d x b) 0 18 54 90
G. Number vaccinated but ill  (d ! f) 0 2 6 10
H. Total number ill  (e + g) 100 82 46 10
  1. % cases vaccinated (PCV) (g/h)
0 2.4% 13% 100%

The higher the coverage rate, the higher the percentage of vaccinated cases.

The public health message would encourage everyone to get vaccinated and to report any cases to the medical community so they can study the resistant strains and eventually increase vaccine efficacy.

Vaccine efficacy

The ability of a vaccine to prevent disease depends on its potency and proper administration to an individual capable of responding. The success of vaccination performed under field conditions may be assessed by measuring protection against clinical disease. Such field assessments can be very useful, particularly when doubt is cast on the efficacy of the vaccine because of the occurrence of disease among vaccinated persons.

Vaccine efficacy is measured by calculating the incidence (attack rates) of disease among vaccinated and unvaccinated persons and determining the percentage reduction in incidence of disease among vaccinated persons relative to unvaccinated persons. The greater the percentage reduction of illness in the vaccinated group, the greater the vaccine efficacy. The basic formula is written as:

VE = ( ARU – ARV / ARU ) x 100

where

VE = vaccine efficacy

ARU = attack rate in the unvaccinated population

ARV = attack rate in the vaccinated population.

Using the basic formula and previously recorded data, calculate vaccine efficacy for Bowie County, Texas.

Residence Urban/Rural Age Group # Cases Population Attack Rate for entire population Vaccinated population Immunized Cases Attack Rates for vaccinated
Bowie Co., Texas Total 1-4 242 4933 49%
5-9 251 6252 40%
1-9 493 11,185 44% 6375.45 27 4.2%
Miller Co., Arkansas Total 1-4 19 2671 7%
5-9 6 3345 17%
1-9 25 6016 4% 5955.84 25 4.2%

ARU = 96.9%
[466 unvaccinated cases / Unvaccinated population 4809.55 ]

ARV = 4.2%
[27 cases / Vaccinated population 6375.45]

Therefore, VE = ((.0969-.0042)/0.0969)*100

= 95% vaccine efficacy

Resources:

Screen Shot 2018-06-12 at 1.35.53 PM.png

Principles of Epidemiology in Public Health Practice, Third Edition
An Introduction to Applied Epidemiology and Biostatistics
Lesson 3: Measures of Risk
Section 2: Morbidity Frequency Measures

This section provides information on calculating incidence / attack rate.

https://www.cdc.gov/ophss/csels/dsepd/ss1978/lesson3/section2.html

Other Epidemiological Case Studies available for download

http://www.aptrweb.org/?page=CaseStudies_Epi

“Developing Competencies for Applied Epidemiology: From Process to Product”
Public Health Rep. 2008; 123 (Suppl 1): 67–118.

This article provides an overview of the professional competencies for applied epidemiologic practice.

“Charting a Future for Epidemiologic Training” Ann Epidemiol. 2015 Jun; 25(6): 458–465. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4646613/
This article outlines the trends shaping the future of epidemiologic research and practice and discuss their impact on training.

Skills used to perform this task:

  • Interpretation of data
  • Data analysis

Skills used in the field of public health:

  • Managing data and resources
  • Complying with rules and regulations
  • Time and project management
  • Broad base of science knowledge
  • Communicating clearly

Citation:

¹ The investigation used in this simulation is authored by Landrigan PJ. Epidemic measles in a divided city. JAMA1972; 221: 567-570. This case study was original developed by Philip Landrigan, Lyle Conrad and John Witte in 1971. The current version was updated by Richard Dicker in 2001 and 2003.

You are viewing a job simulation. To get started, set up SMART Goals to perform this simulation in a reasonable timeline. If you have completed the task, fill out the Self-Reflection Sheet.