![[ Visit Client Website ]](images/banner.png)
BACKGROUND: The Bureau of Communicable Disease (BCD) at the New York City Department of Health and Mental Hygiene performs systematic analyses of surveillance data for >70 reportable infectious diseases to monitor trends and to identify risk factors to inform prevention efforts. Special epidemiologic studies are conducted as needed by staff, fellows, and students. An in-house consultation process is used to help ensure the scientific validity and statistical accuracy of the bureau’s epidemiologic investigations and research projects.
METHODS: For new epidemiology studies, special investigations, and analyses, the project lead completes a standard proposal form detailing objectives, data sources, analytical plans, and potential limitations. For conference abstracts and scientific manuscripts using BCD data and focusing on surveillance or epidemiology, the lead author drafts the document and incorporates feedback from co-authors and supervisors. Next, the documents are provided to a doctoral-level epidemiologist on staff for review and editing within agreed timeframes. In addition to scientific validity, documents are reviewed for consistency with agency guidance, e.g., regarding appropriate population denominator sources, the use of neighborhood-level poverty as an agency standard variable, and writing conventions. Recommendations for major revisions are discussed at in-person meetings with project leads and supervisors, and documents are iteratively revised and reviewed. After all recommendations are incorporated or waived, proposals are approved for implementation, and abstracts and manuscripts are submitted for agency clearance and review by the Bureau, Division, and Commissioner’s Office.
RESULTS: In 2013, BCD’s epidemiologic consultant reviewed 12 proposals, 17 abstracts, and 9 manuscripts. Recommendations ranged from minor wording edits to major revisions of analytic plans and relevant SAS code. Common recommendations included using multivariable regression to assess the effects of potential confounders, using sensitivity analyses to assess the robustness of findings to important assumptions, and expanding proposal and manuscript sections relating to possible biases and limitations.
CONCLUSIONS: Strengths of BCD’s epidemiologic consultation process include scientific review prior to project implementation (at a point when changes in study design are practical), the cross-pollination of ideas between projects, and the consistent application of agency guidelines. Challenges include balancing methodological rigor with feasibility and resource availability, a lack of in-house biostatistical support, and competing priorities. Epidemiologic consultations (via dedicated staff or health department/academic partnerships) can support workforce development and help to ensure that recommendations for public health action are based on high-quality science.