Although it is not possible to completely standardize outcomes tables because of the diversity of issues among preventive services, the following are some preferred entries for authors to use in outcomes tables when possible.
- For services with an extended time frame: 10 years (5 years has typically been used, although 5 years is a short time for many consequences of screening and prophylactic interventions. Since these calculations require some assumptions, extrapolating to 10 years seems reasonable.). Alternatives: 5 years, lifetime.
- For services with a short time frame (e.g., pregnancy): 1 year.
- Express this number as per 1,000 persons targeted (e.g., per 1,000 women ages 40 to 49 years).
- Rationale: preventive services with a large magnitude of effect should have substantial numbers of outcomes when expressed per 1,000 persons; those preventive services with less than one outcome averted will clearly be interpreted as having relatively small effect.
Interventions: Interventions should be shown in columns and described. For repeated services (e.g., annual fecal occult blood testing), the number of services should be identified.
Parameter estimates: Important parameters should be provided, as appropriate:
- Screening results (sensitivity, specificity).
- Prevalence of condition.
- Adherence (to screening, treatment).
- Intermediate outcomes.
- Number identified (with and without the condition).
- Number treated.
Outcome measures (harms and benefits):
- Deaths (where relevant).
- Important health outcomes (e.g., strokes averted or cancers caused).
- Quality-adjusted life years (when possible).
- Adverse events/states.
Number needed to screen/treat/counsel:
- Express in outcome terms (e.g., number needed to screen to avert one death).
- Number needed to counsel to achieve change in behavior should only be provided if it is also provided for a health outcome.
Current as of: July 2017
Internet Citation: Appendix VIII. Standardization of Outcome Tables. U.S. Preventive Services Task Force. July 2017.