Research Article: Risk of bias and confounding of observational studies of Zika virus infection: A scoping review of research protocols

Date Published: July 7, 2017

Publisher: Public Library of Science

Author(s): Ludovic Reveiz, Michelle M. Haby, Ruth Martínez-Vega, Carlos E. Pinzón-Flores, Vanessa Elias, Emma Smith, Mariona Pinart, Nathalie Broutet, Francisco Becerra-Posada, Sylvain Aldighieri, Maria D. Van Kerkhove, Marcia Edilaine Lopes Consolaro.


Given the severity and impact of the current Zika virus (ZIKV) outbreak in the Americas, numerous countries have rushed to develop research studies to assess ZIKV and its potential health consequences. In an effort to ensure that studies are comprehensive, both internally and externally valid, and with reliable results, the World Health Organization, the Pan American Health Organization, Institut Pasteur, the networks of Fiocruz, the Consortia for the Standardization of Influenza Seroepidemiology (CONSISE) and the International Severe Acute Respiratory and Emerging Infection Consortium (ISARIC) have generated six standardized clinical and epidemiological research protocols and questionnaires to address key public health questions on ZIKV.

We conducted a systematic search of ongoing study protocols related to ZIKV research. We analyzed the content of protocols of 32 cohort studies and 13 case control studies for systematic bias that could produce erroneous results. Additionally we aimed to characterize the risks of bias and confounding in observational studies related to ZIKV and to propose ways to minimize them, including the use of six newly standardized research protocols.

Observational studies of ZIKV face an array of challenges, including measurement of exposure and outcomes (microcephaly and Guillain-Barré Syndrome). Potential confounders need to be measured where known and controlled for in the analysis. Selection bias due to non-random selection is a significant issue, particularly in the case-control design, and losses to follow-up is equally important for the cohort design.

Observational research seeking to answer key questions on the ZIKV should consider these restrictions and take precautions to minimize bias in an effort to provide reliable and valid results. Utilization of the standardized research protocols developed by the WHO, PAHO, Institut Pasteur, and CONSISE will harmonize the key methodological aspects of each study design to minimize bias at different stages of the study. Biases need to be considered by researchers implementing the standardized protocols as well as by users of observational epidemiological studies of ZIKV.

Partial Text

Although the Zika virus (ZIKV) was first identified in the mid twentieth century on the African continent, only fourteen cases were documented in humans prior to the first large, documented epidemic outbreak on the Island of Yap in 2007 [1]. This was followed by the largest ZIKV outbreak ever previously reported in French Polynesia from October 2013 to April 2014 [2]. Since 2015, 76 countries and territories around the world have reported mosquito-borne ZIKV transmission, particularly the Americas where Brazil has been hit the hardest [3]. In February 2016 the World Health Organization declared the ZIKV outbreak a Public Health Emergency of International Concern [4].

The PAHO, through the creation of the Zika Virus Research Platform (, maintains a database of research protocols and primary research studies, including both observational and experimental study designs. These records contain data elements that describe the study’s purpose, recruitment status, design, eligibility criteria, and locations, as well as other key protocol details. Resources and links to additional information are inserted by the PAHO to enhance the overall usefulness of the database. Researchers, policymakers, and others can now examine features and ongoing trends of Zika virus research.

Our searches in the PAHO’s Zika Virus Research Platform found 317 research protocol titles. We applied the inclusion criteria to the titles and summaries (when available) of all records and excluded 288 of them. We contacted researchers to obtain the full text of each one of the 32 cohort and 13 case control studies that were identified. Supporting information 1–2 show our screening and study selection process of the 16 cohort protocols and 13 case control studies that were finally reviewed can be found in S2 Fig and S1 File. PRISMA 2009 Checklist is presented in S2 File.

We have enumerated the potential risks of bias and confounding in all observational studies, which are applicable to observational studies of ZIKV. We have developed six standardized protocols for observational studies whose aim is to minimize bias by specifying clear criteria for selection of participants; standardized measurement methods and definitions for both exposures (ZIKV) and outcomes (microcephaly, GBS); and specification of potential confounders, along with standardized methods to measure and control for them. Concerted efforts must be made in the design and analysis stages of observational studies related to ZIKV infection. The use of these standardized protocols will help minimize potential risk of bias of future studies on ZIKV infection. The use of the standardized protocols by researchers in the field will increase the quality of their data. In addition, it will facilitate comparability of data and thus the possibility of performing joint analyses to answer complex questions that individual studies cannot answer yielding more reliable, valid and generalizable results. Users of the ZIKV research, including public health decision-makers, should critically appraise future observational studies for known risks of bias and use this information in their decision-making.




0 0 vote
Article Rating
Notify of
Inline Feedbacks
View all comments