Writing while the evidence is still moving
A developing evidence base creates decisions about when to publish, which limitations to state and how to update an account as findings accumulate. Outbreak reporting makes these decisions particularly visible because reliable evidence may be limited while clinical questions are urgent.
Provisional evidence
Biomedical evidence is provisional: accumulating support can increase confidence in a practice without proving that it is correct. Evidence-based medicine calls for conscientious, explicit and judicious use of the current best evidence in decisions about individual patients. As shared evidence accumulates, views can converge and consensus tends to emerge on formerly disputed questions.
Reliable data underpin medical writing; without them, producing information useful to an audience becomes difficult. In 2016, DynaMed Plus, a point-of-care information service, synthesised the best available evidence clearly, concisely and transparently for outbreak coverage. Best available evidence meant the most methodologically rigorous study for the particular scenario, and not every topic had evidence meeting the service’s level 1 methodological criteria.
Monitoring, publication and revision
DynaMed Plus’s topic-development process used three overlapping stages:
- Identify and monitor available data.
- Write and publish the topic.
- Monitor emerging information and update the topic as needed.
In 2016, staff systematically surveyed more than 500 medical journals. During an outbreak they supplemented this with ProMED, WHO Disease Outbreak News, CDC Morbidity and Mortality Weekly Report and keyword searches in databases such as PubMed.
The Zika outbreak search focused on three signals:
- Increased numbers of cases.
- A newly affected geographical area.
- Increased severity or comorbidities.
Publication before a causal link is established
DynaMed began work on Zika in May 2015 after ProMED reported a confirmed outbreak in Brazil. As studies accumulated, their growing number made reliability assessment difficult, and none met level 1 criteria: limitations included small cohorts, no standardised diagnostic approach to microcephaly or failure to identify Zika as the infectious agent.
No study had yet demonstrated a causal link between Zika infection during pregnancy and microcephaly, but the team judged the accumulated evidence sufficient to report an association. It published the topic in January 2016 with the studies’ limitations stated.
The team critically appraised and incorporated subsequent studies and had made more than 150 updates by the late-2016 account. In October 2016 it published a separate topic on Zika infection in pregnancy and congenital Zika syndrome. New questions and studies continued to arise, leaving further monitoring and revision to do.