Cross Sectional Studies and Public Health Surveillance

1
Cross Sectional Studies and Public Health
Surveillance

• Introduction to Epidemiology
• Sharon Cooper, Ph.D.
• June 14, 2001

2
Cross-Sectional Studies

• Often used to investigate non-fatal diseases
• May necessitate special studies
• Distinct from other study designsprevalence
• Some studies may be designed to estimate burden
  of disease in a population others only exposure
• Can be difficult to assess time sequence between
  exposure and disease

3
Cross-Sectional Studies
2 uses descriptive and analytic
4
Features of Cross-Sectional Studies

• Measures disease prevalence-gt of existing cases
  in population (disease, symptoms, physiologic
  states)
• Point prevalence
• Period prevalence (less often)
• Prevalence rate
• cases/ of people, therefore a proportion,
  not a rate
• If measure prevalence in exposed and unexposed,
  can calculate prevalence odds ratio
• Can also calculate means, correlation
  coefficients

5
Population Surveys

• Focus is on some real population that can be
  defined geographically and temporally
• Purposes include
• Describing the extent of a condition, by
  determining the prevalence, mean value, or
  distribution
• Testing a hypothesis about an association
• Finding cases for treatment, or participation in
  a study

6
Example Determine the prevalence of
disease/condition

• What is the prevalence of NOT having received
  prenatal counseling among women delivering at LBJ
  Hospital, Houston, Texas in 1999?
• Possible purposes
• Determine need for services
• Evaluate effect of a program
• Among adults in California, what is the
  prevalence of elevated blood lead levels?
• Identify high risk workers and workplaces
• Design interventions

7
Example Test a hypothesis about an association

• Among smoking pregnant women, is the average
  number of cigarettes smoked associated with
  hemoglobin levels?
• Is there a difference in mean serum cholesterol
  between males and females?

8
Population Definitions

• Target Population The collection of individuals
  about which it is desired to make inferences.
• Study Population The collection of individuals
  selected for study.
• Examined Population The collection of
  individuals who were actually studied by
  investigators (subset of study population)

9
Steps in Population Survey

• Identify target population?
• Determine sampling frame?
• Draw sample to identify study pop.?
• Examine study population?
• Characterize examined population

10
Designing an Epidemiologic StudySkin Cancer
Background

• Most common type of cancer in U.S. 40-50 of
  Americans who live to age 65 will have skin
  cancer at least once.
• Types
• Non-melanoma (basal and squamous), more than 1/2
  million new cases/year in U.S.
• Melanoma (27,000 new cases/year in U.S.)

11
Skin Cancer Cont.

• Primary risk factor--UV radiation from sun
• Occupational risk factors--ionizing radiation,
  PAH, arsenic
• Risk varies with age, hair and eye color, skin
  type, ease of freckling, and geographic location

12
Keratoses

• Keratosis--dysplastic growth of keratin producing
  cells of skin
• Risks for keratoses appear to be same as for SC
  and BC carcinoma
• Natural history of keratosis--??, perhaps 10-20
  progress to squamous cell carcinoma

13
Background for Study

• Two reports (Bowra et al., 1982 Wang et al.,
  1987) suggested risk of keratoses among workers
  in production of 4,4 bi-pyridyl
• No good background data, clinical impressions,
  and clustering within job titles
• Both reports involved production using HTS or MAG
  processes, but new cases also diagnosed after
  switching to LTS process

14
Design the Study

• Texas plant (100 employees)--LTS process only
  does excess exist among workers attributable to
  occupational exposure?
• What types of expertise would you desire study
  team to have?
• Considerations study design, definition and
  ascertainment of exposure/health outcome, bias.

15
SURVEILLANCE

• Surveillance--French word keeping a close
  watch over an individual. . .in order to detect
  subversive tendencies (monitoring/intervention)
• CDC defines surveillance
• ongoing, systematic collection, analysis, and
  interpretation of health data essential to
  planning, implementation, and evaluation of
  public health practice. . .and dissemination of
  these data. . .for the purposes of prevention and
  control.
• Ultimately, a single element, a single addition
  to the strategy of the eradication process was
  responsible for its success--that change was
  the incorporation of the principle of
  surveillance.
• D.A. Henderson, 1980

16
TWO CONNOTATIONS OF SURVEILLANCE

• Public health surveillance--activities by
  federal, state, local agencies to monitor or
  follow-up on occupational diseases/injuries
  (usually population-based)
• Medical surveillance--refers to application of
  medical tests and procedures to individuals to
  detect particular disorder (medical screening is
  subset)

17
PURPOSES OF SURVEILLANCE

• to recognize a problem
• to define the scope of the problem (descriptive
  epidemiology, priority setting, trends)
• to activate epidemiologic research (not
  synonymous with)
• evaluate effectiveness of interventions

18
STEPS IN PUBLIC HEALTH SURVEILLANCE

• Gather information on cases of disease and injury
  or on exposures
• analyze data
• disseminate data to medical community, unions,
  government agencies, researchers, public
• intervene to reduce or eliminate health
  events/hazards
• (to count, evaluate,act)

19
TWO AXES OF SURVELILLANCE

• Health effect surveillance (public health,
  medical)
• Hazard surveillance (freq., distribution, and
  trends in hazard responsible for disease, e.g.,
  lead, behavioral risk factors, noise, ozone
  levels, inspection of kitchens by health
  departments)

20
TWO TYPES OF SURVELILLANCE

• Passive--data collected for other purposes, e.g.,
  insurance data, hospital discharge
• Active--data collected specifically for program
  needs

21
EXAMPLES OF SOURCES OF SURVELILLANCE DATA

• National Notifiable Disease Surveillance System
  (NNDSS)--notifiable diseases in the U.S.
• Vital statistics
• Sentinel surveillance--based on an event that is
  a preventable disease, disability, or untimely
  death whose occurrence serves as a warning signal
  that the quality of preventative and/or
  therapeutic medical care may need to be improved
  (Rutstein, 1983)
• Registries (exposure, cancer, birth defects)
• Surveys (e.g., National Health Interview Survey
  100,000 people annually)
• Administrative data collection systems (hospital
  discharge, Medicare Provider Analysis and Review
  MEDPAR, ERs, public health clinics, etc.)

22
EXAMPLES OF CRITERIA FOR EVALUATING SURVELILLANCE
SYSTEMS(MMWR 37(No. S-5)1-18, 1988)

• Public health importance
• usefulness/cost
• acceptability to persons recording data
• timeliness
• representativeness of cases in population
• simplicity of data collection and management
• flexibility (to include additional conditions)