Data in the MMWR
Impact (10 of 12)
Posted to LinkedIn on 2026-02-17. The text below is the post as published.
Links to works now on this site point to their pages here; the words are as published. Where the post showed a web address, the icon stands in for it (hover over it to see the address as posted).
Data in the MMWR (§6.2): Describe potential impact to the extent possible On this site (opens in a new tab)
[2026-02-17]
Principles: Reports should approach causal claims carefully and explicitly, with attention to assumptions and context beyond statistics. Causal concepts can be supported by, but rarely flow directly from, statistical analysis. It is especially important to relate clearly the inferential support from the report’s analyses to claims regarding public health actions and impact.
Observations: Reports commonly describe a public health phenomenon over time to determine whether, when, and how to apply guidance; to monitor phenomena; or, to set a baseline for further monitoring, especially with emerging conditions like Covid-19.
Vaccine effectiveness reports compare health outcomes between vaccinated and unvaccinated persons. These reports consistently address possible uncontrolled confounding, sample size, consistency with phase 3 trials, and the relative strength of evidence by observational studies.
Before-and-after reports explicitly characterize results from relative to a reference event, with preselected or data-driven change points and continuous or discontinuous relationships.
Causal reports directly try to corroborate or to disavow causation; others address causal ideas indirectly. These reports typically distinguish being consistent with causation from evidence of causation.
Reports typically place results in the context of previous understanding, practice, and policy. Some reports use words like highlight, reinforce, and underscore to connect each report to related recommendations—often without clearly noting whether the report provides evidence for the cited guidance.
Recommendations: State and address assumptions, such as those regarding confounders and mediators, that are intended to support causal claims. These assumptions require careful justification, because they cannot be directly tested. Take care with efforts to capture, describe, and draw inferences from changes over time. Be mindful of words (e.g., protect, harm, effective, impact) that imply causal relationships and words (e.g., association, trend) that usually do not. Take care with interaction (a statistical concept) and effect modification (a causal concept). When linking to related but distinct information, be clear about whether or not the report substantively supports or refutes the outside information. Otherwise, it can appear that the link implies substantive support.
Next topic: Place interpretation in context, including limitations (§6.3)
Data in the MMWR
- Introduction · How learners read
- 1 · Sources
- 2 · Methods
- 3 · Description
- 4 · Comparison
- 5 · Adjustment
- 6 · Conclusions
- 7 · Tables
- 8 · Figures
- 9 · Interpretation
- 10 · Impact (this post)
- 11 · Context
- 12 · Additional practices