Part II: Methods and Approaches 1: Assessing Disease Associations and Interactions Chapter 10 Tables

Human Genome Epidemiology: A Scientific Foundation for Using Genetic Information to Improve Health and Prevent Disease


Reporting and Review of Human Genome Epidemiology Studies

Julian Little

Table 10-1
Proposed checklist for reporting and appraising studies of (i) genotype prevalence, (ii) gene-disease associations, and (iii) gene-environment interactions
Genotype prevalence Gene-disease associations Genotype- environment interaction
1. Purpose of study
Detect associations or estimate magnitude of association
Describe joint effects; test specific hypotheses about interaction
2. Analytical validity of genotyping
Types of samples used
For cases and for controls
For cases and for controls
Timing of sample collection and analysis, by study group*
ethnic group
e.g. cases vs. controls
e.g. cases vs. controls
Success rate in extracting DNA, by study group*
e.g. ethnic group
e.g. cases vs. controls
e.g. cases vs. controls
Definition of the genotype(s) investigated; when there are multiple alleles, those tested for should be specified
Genotyping method used (reference; for PCR methods – primer sequences*, thermocyle profile*, number of cycles*)
Percentage of potentially eligible subjects for whom valid genotypic data were obtained, by study group
ethnic group
e.g. cases vs. controls
e.g. cases vs. controls
If pooling was used, strategy for pooling of specimens from cases and controls
Quality control measures*
Including blinding of laboratory staff
Including blinding of laboratory staff
Samples from each group of subjects compared (e.g cases and controls)included in each batch analyzed*
3. Assessment of exposures
Reproducibility and validity of exposure documented
Categories or exposure scale justified
4. Selection of study subjects
Geographical area from which subjects were recruited
The recruitment period
Recruitment methods for subjects whose genotypes were determined, such as random population-based sampling, blood donors, hospitalized subjects with reasons for hospitalization
Definition of cases and method of ascertainment
Number of cases recruited from families and methods used to account for related subjects
Recruitment rates
Where possible by sex, age and ethnic group For cases and controls For cases and controls
Mean age (±SD) or age-range of study subjects, and the distribution by sex
For cases and controls For cases and controls
If the subject were controls from a case-control study, information of the disease under investigation and any matching criteria such as age, gender, and/or risk factor levels
Ethnic group of study subjects
Similarity of socio-demographic (or other) characteristics of subjects for whom valid genotypic data were obtained with characteristics of subjects for whom such data were not obtained*
Steps taken to ensure that controls are non-cases*
5. Confounding, including population stratification
If other than a case-family control design, matching for ethnicity, or adjustment for ethnicity in analysis
6. Statistical issues
Distinguish clearly a priori hypotheses and hypotheses generated
If haplotypes used, specify how these were constructed
Number of subjects included in the analysis, by cell numbers where possible
Method of analysis, with reference, and software used to do this
Confidence intervals
Of genotype frequency Of measures of association with the genotype
For interaction analysis, 2xK presentation nused, or choice of stratified analysis justified
For interaction analysis, P value for interaction calculated and choice of Wald test or likelihood ratio test specified and justified
For interaction analysis, null interactions listed
Assessment of goodness-of-fit of the model used*

*Additional information recorded (ideally in web-based methods register)
[back to chapter]


Box 10-1
Guidelines for causal inference (modified from Hill [120] and Surgeon General [119])
  • Consistency
  • Strength
  • Dose-responsea
  • Biologic plausibility (including analogyb)a
  • Temporality
  • Experimental supporta
  • Coherence

aAdditional considerations specified by Hill (120)
bAnalogy is a variant of biologic plausibility (29, 121)

[back to chapter]

Page last reviewed: January 1, 2004 (archived document)