CRC Methods Course: Case and CrossSectional Studies

1
CRC Methods CourseCase and Cross-Sectional
Studies
File h\drived\epidem\crc05.ppt Revised 2/8/05
2
Purpose of Studies

• Descriptive
• Estimate disease frequency
• Describe time trends of disease
• Identify and classify diseased individual
• Analytic
• Test specific hypotheses
• Generate new hypotheses

3
Types of Studies

• Observational - Descriptive
• Case reports
• Case series
• Cross - sectional
• Observational - Analytic
• Case-control
• Cohort
• Meta-analyses

4
Types of Studies (continued)

• Experimental
• Clinical trial
• Laboratory trial

5
Establishing Causality

• Causality is something that brings about an
  effect or result

Risk Factor causes
Outcome
versus
Risk Factor associated with
Outcome
6
Criteria for Establishing Causality

• Strength of the association
• Dose-response relationship
• Temporal relationship between cause and effect
• Consistency of findings
• Biological plausibility

Hill AB Proc Royal Soc Med 196558295-300
7
Problems in Understanding Risk-Exposure
Relationship

• Latency of disease - disease may occur long after
  the exposure to the risk factor.
• Frequency of exposure - it may be difficult to
  determine.
• Incidence of disease - rare disease may make it
  difficult to develop disease-exposure
  relationships.

8
Problems (continued)

• Magnitude of risk - just how much exposure is
  required to generate disease or and adverse
  outcome.
• Multi-factorial nature of disease - disease
  (adverse outcomes) may be the cumulative result
  of multiple factors.

9
Problems (continued)

• Bias - a process at any stage of inference that
  produces results that depart systematically from
  the true values.
• Chance - (random error) - differences
  attributable to chance variation.

10
Types of Bias

• Selection Bias - error due to systematic
  differences in characteristics between those who
  are selected for study and those who are not.
• Measurement Bias - systematic error arising from
  inaccurate measurement (or classification) of
  subjects on the study variables.

11
Types of Bias (continued)

• Confounding - the distortion of the apparent
  effect of an exposure on risk brought about by
  the association of the exposure variable with
  other factors that can influence the outcome.

12
Example of Confounding
Enuresis
present
absent
Yes
17
83
Glasses
8
92
No
25
175
Odds Ratio (17)(92) (83)(8) 2.36
13
Example of Confounding (continued)

? 3 years
gt 3 years
Enuresis
Enuresis
-
-




1
19
16
64
Glasses
Glasses
4
76
-
-
4
16
5
95
20
80
Odds Ratio 1.0
Odds Ratio 1.0
14
Measures of Disease Frequency

• Prevalence - the proportion of a group possessing
  a condition, disease, or attribute at a given
  point in time.
• Incidence - is the proportion of a group
  initially free of a condition, disease, or
  attribute, that acquires it over a given period
  of time.

15
Total Population 100
Jan
March
April
Feb
16
Prevalence of otitis for entire period? 20/100
0.2 or 20
Jan
March
April
Feb
17
Prevalence in February? 11/100 0.11 or 11
Jan
March
April
Feb
18
Incidence over entire period? 16/100-4 0.167 or
16.7
Jan
March
April
Feb
19
Incidence for February? 6/100-9 0.066 or 6.6
Jan
March
April
Feb
20
Case Studies

• Definition - a case report consists of a
  description of an outcome or the condition of a
  patient who may be classified into a definite
  case category based on symptoms, signs or
  laboratory findings.

21
Purposes of Case Studies

• To describe the occurrence of new disease
  entities.
• To describe the outcome of a patient or patients
  with specific diseases.
• To formulate hypotheses of the association
  between various exposures or risk factors and the
  occurrence of an outcome or disease.

22
Advantages of Case Studies

• Allows for the description of new disease
  processes.
• Allows for the description of outcomes associated
  with rare diseases.

23
Disadvantages/Limitations of Case Studies

• Impossible to determine disease frequency.
• Cannot establish causality between exposures or
  risk factors and disease outcome.

24
Considerations in Designing Case Reports

• Is there a better way to report your observations
  that will allow you to report some estimate of
  disease frequency?
• If not
• Then how unique is your case definition
• Are there other case reports in the literature
  that substantiate your observations?
• Are you ready to propose any hypotheses about the
  relationships between exposures associated with
  the outcome of the case you are reporting?

25
Example of Case Report Leap of Faith.

• A cytogenetically normal infant with severe
  congenital granulocytopenia was treated with
  granulocyte colony-stimulating factor. After 11
  months of therapy, the infant developed acute
  non-lymphoblastic leukemia. The use of
  granulocyte colony-stimulating factor in patients
  with congenital marrow failure disorders may
  induce malignant transformation. (J Pediatr
  1995126263)

26

• Can we determine incidence or prevalence from
  this report?
• Can we determine causality from this report?

27
Case Series Pediatrics 2003111201-203

• Six children, 4 to 11 years old, were evaluated
  in a dermatology clinic for tinea corporis after
  having received treatment with a topical
  combination of clotrimazole (1)-betamethasone
  (.05) they had received from their pediatrician
  for 2-12 mos.
• Diagnosis confirmed with KOH prep.

28
Case series (continued)

• Treated with one of several oral or topical
  antifungal agents with complete resolution.
• Conclusion The use of combination
  antifungal-steroid therapy may be associated with
  persistent/recurrent infection.
• Hypotheses
• The suppression of inflammation may be associated
  with premature discontinuation of therapy.
• Intact local immunity is necessary for the
  eradication of dermatophytes

29
Case series (continued)

• Can we say anything about the prevalence or
  incidence of recurrent or persistent fungal
  infections in children?
• Can we say anything about the strength of the
  association between the use of the
  antifungal-steroid combination and
  recurrent/persistent disease?
• Can we say anything about the consistency of
  these findings?

30
Case series (continued)

• Did the authors describe the outcome of a patient
  or patients with specific diseases?
• Did the authors attempt to formulate hypotheses
  of the association between various exposures or
  risk factors and the occurrence of the outcome or
  disease?

31
Example of Case Series(MMWR 198130250)
33 y.o. HSM CMV
PC X
30 y.o. HSM CMV
PC
30 y.o. HSM CAN
PC CMV
29 y.o. HSM
PC X(cmv)
36 y.o. HSM CMV
PCCAN
6 7 8 9 10 11
12 1 2 3 4 5 6
1980
1981
32
Case Series Example(N Engl J Med 19813051425)
33
Hypotheses for Acquired Cellular Immunodeficiency
(N Engl J Med 19813051425)

• Sexually transmitted infectious agent.
• Exposure to a common environment
• New strain of cytomegalovirus or reinfection with
  CMV following immune suppression from initial CMV
  infection
• CMV infection may be result of T-cell defect than
  cause.

34
Case Series Example (continued)
J Infect Dis 1983148339
35
Case Series Example (continued)
36
Intravenous Drug Abusers 236
Hemophiliacs 8
7
Others
1
642
61
81
154
1
3
50
Haitians 54
Homosexual or Bisexual Men 727
37
Case Series Example (continued)
38

• Can we determine incidence or prevalence with
  this case series study?
• Can we determine causality?
• Did the authors describe the characteristics of a
  unique disease process?
• Were their initial hypotheses eventually shown to
  be true?

39
Cross-Sectional Study

• Definition - also called a survey or
  prevalence study - usually involves a random
  sampling of a target population and the reporting
  of the sampled populations disease status
  current or past exposure level and the recording
  of any other relevant variables.

40
Purposes of a Cross-Sectional Study

• To describe characteristics of the target
  population - important in planning for health
  services, facilities, and programs.
• To generate new etiologic hypotheses between
  exposure factors and diseases.

41
Advantages of Cross-Sectional Studies

• Relatively inexpensive and simple to carry out,
  because no follow-up is required.
• No one is exposed to a putative causal agent or
  is denied a potentially beneficial therapy.

42
Disadvantages/Limitations

• Cannot determine causation, because the disease
  status and the exposure status are determined
  simultaneously.
• Exposure and disease status may be dependent on
  recall and thus subject to bias.
• Exposure groupings may have unequally distributed
  population characteristics that produce
  confounding of the disease-exposure relationship.

43
Disadvantages/Limitations (continued)

• Cases with early deaths and those in which
  evidence or recall of exposure have disappeared
  may be missed.
• Rare diseases may not be well represented.
• Arbitrary groupings based on disease status or
  exposure status may be of unequal size resulting
  in loss of statistical efficiency.

44
Designing Cross-Sectional Studies

• Make the sampling process random to ensure
  representativeness among the target population.
• Make sure the case, outcome, and exposure
  categories are defined as specifically as
  possible
• Define your target population.

45
Designing Study (continued)

• Think carefully about why you are doing the study
  and focus the study as narrowly as possible on
  that reason.

46
Planning for the Cross-Sectional Study

• Determine how you will acquire a registry of your
  target population
• Detail the selection or randomization process
• How will you obtain consent for acquiring the
  target population registry?
• How will you obtain consent for gathering the
  information?

47
Planning (continued)

• How will you collect the information?
• Mail out questionnaires
• Telephone interview
• Personal interview
• Physical exam
• Laboratory information
• How valid are the measurement instruments you are
  using?

48
Planning (continued)

• How do you plan to store the data?
• How do you plan to computerize the data?
• Will you develop a data verification plan?
• What are the specific hypotheses you plan to
  test?
• Have you calculated the sample size required to
  test the hypothesis?
• What statistical methods will you use?

49
Planning (continued)

• Who will write the manuscript?
• Who will pay for the costs of the study?
• How long will the study take from start to finish?

50
Example of a Cross-Sectional Study(Am J Dis
Child 1988142640)

• Sampling scheme for the National Health Interview
  Survey
• Primary Sampling Units (PSU) 1900
• Random Sample of PSUs 200
• Each PSU divided into segments and
• each segment40 households, 8 house-
• holds from each segment randomly sampled.

51
1 Segment with 40 households
xxx
xxx
xxx
8 Households randomly selected
xxx
xxx
xxx
xxx
xxx
52
Respondents for Syrup of Ipecac Study using the
National Health Interview Study Data
53
Percent of Respondents having Ipecac in Household
54
Percent of Respondents Answering Affirmatively to
the Following Questions
55

• Have we determined the prevalence of ipecac use
  in this survey?
• Have we determined incidence of ipecac use with
  this survey?
• Have we determined why only 8 of the Black
  population have ipecac in the home compared to
  30 of the White population?