Reproducible Research Literature Review

  1. Main conclusion and study background information

  2. Article Information/Identifier
    1. journal abbreviation from PubMed
      • JAMA --- Journal of the American Medical Association
      • NEJM --- New England Journal of Medicine
      • AJE --- American Journal of Epidemiology
      • EPI --- Epidemiology
    2. four digit year
    3. journal volume number
    4. article first page number
    5. first author last name and first/middle initials (e.g. Peng RD)

    Example:

    journal year volume first page # first author last name first initials
    AJE 2005 161 585 Peng RD
    JAMA 2004 292 2372 Bell ML
  3. Type of study
    1. randomized trial
    2. observational study
    3. methodologic study
    4. literature review/meta-analysis
    5. other

  4. If study is a randomized trial, methodologic study, or literature review/meta-analysis, stop here.

  5. Main concluding sentence(s) from the abstract (author's interpretation of the results)
  6. Epidemiology subspeciality (check all that apply, comma-separated list)
    1. environmental
    2. nutritional
    3. reproductive
    4. cardiovascular
    5. infectious diseases
    6. genetics
    7. other
  7. Study design
    1. cross-sectional
    2. case-control (including variants, e.g. case-cohort, case-crossover)
    3. cohort
    4. other
  9. Total sample size (NA if not reported)
  10. Main outcome variable --- record the following:
  11. Main exposure/risk variable (at most 2 from main conclusion) --- record the following
  12. Are the results supporting the main conclusion adjusted for potential confounders? (Yes/No)
  13. Are effect modifiers or interactions explored in the results supporting the main conclusion? (Yes/No)

  14. Data supporting main conclusion

  15. Which table supports the main conclusion? (Table name, or NA)
  16. Which figure supports the main conclusion? (Figure name, or NA)

  17. Analytic data are defined as any data that were inputted directly into a statistical algorithm to produce a derived statistic reported in the table supporting the main conclusion. For example, if a table reports a mean and standard error, the mean and standard error are the derived statistics and the list of numbers used to calculate that mean and standard error constitute the analytic data.

    Measured data are any other data that were used to create the analytic data via a computer program or hand calculation. In some cases it may be that the measured data and the analytic data are the same. For example, if a table reports a regression coefficient from running a Poisson regression of daily hospital admissions and air pollution exposures, then the analytic data would be the data frame inputted to the regression program and the measured data would be all of the data that came before that data frame (e.g. original Medicare records, individual pollution monitor data, etc.)

    Repeat the next 3 questions for the (a) outcome, (b) exposure, (c) any confounders, and (d) any effect modifiers.

  18. Where are the analytic data for generating the table/figure supporting the conclusions reported to be available?
    1. printed in the paper (e.g. in a separate table)
    2. from the journal publisher
    3. from a website (not journal affiliated)
    4. by request from the author
    5. by request from a third party
    6. other
    7. data not reported to be available
  19. Where are the measured data for generating the analytical data reported to be available?
    1. printed in the paper
    2. from the journal publisher
    3. from a website (not journal affiliated)
    4. by request from the author
    5. by request from a third party
    6. other
    7. data not reported to be available
    8. measured data and analytic data are the same
  20. What is the source of the measured data? (comma-separated list if more than one)
    1. original data set (e.g. a new cohort study)
    2. ongoing study
    3. government data sources (e.g. Medicare, NHANES, EPA, cancer registries)
    4. other

  21. Analysis supporting main conclusion

  22. What is the conclusion a statement about?
    1. the distribution of the outcome variable
    2. the distribution of the outcome across groups or levels of another explanatory variable
    3. the distribution of the outcome across groups or levels of another explanatory variable taking into account of potential confounders
    4. differences in the relationship between the outcome and an explanatory variable by levels of an effect modifier (possibly taking into account of potential confounders)
    5. other
  23. Is the presence of measurement error in the exposure or other explanatory variables accounted for in the analysis? (Yes/No)
  24. What type of statistical model was employed?
    1. standard 1- or 2-sample method, e.g. t-test, chi-square test of proportions, Kaplan-Meier curve
    2. regression method such as linear regression, generalized linear model (logistic or Poisson), or Cox proportional hazards model
    3. other method (write in)
  25. How was the analysis implemented?
    1. not reported
    2. by hand
    3. use of software package(s)
  26. What software package(s) is reported to have been used for analysis? (comma separated list, or NA)
  27. Is the computer code or instructions used for analysis reported to be available? (Yes/No/NA)
  28. How were the measured data processed into analytical data?
    1. by hand
    2. use of software package(s)
    3. not reported
    4. NA (measured and analytic data are the same)
  29. Is the computer code or instructions used for processing the measured data into analytical data reported to be available? (Yes/No/NA)

  30. Other

  31. Notes --- write in any notes regarding the study.