Requirements for systematic reviews

I am writing to express concerns with the recent review by Scotney et al¹ regarding occupational stress among personnel working in animal-related occupations. It appears that the authors failed to include controlled vocabulary terms and relevant keywords in their search strategy, which may have affected the sensitivity of their search. The “Cochrane Handbook for Systematic Reviews of Interventions”² recommends that searches for systematic reviews use subject terms from a controlled vocabulary in addition to free-text terms. The authors also neglected to search two databases with extensive coverage of the biomedical literature: MEDLINE and CAB Abstracts. An investigation³ of veterinary journal coverage in nine databases concluded that CAB Abstracts should be included in any literature search when the aim is to find all the published evidence on a veterinary topic, and MEDLINE should be included if there is a biomedical aspect to the topic. Scopus, one of the databases that Scotney et al searched, does include MEDLINE content. However, Scopus does not offer the same search capabilities as a direct search of MEDLINE via PubMed or another provider. The unique indexing, search features, and content of MEDLINE and CAB Abstracts make them important resources for reviews in veterinary medicine.

Discrepancies in the methodology and reporting of this study would make it difficult to reproduce. The total number of articles identified (2,694) and duplicates removed (14) were low, considering the number and scope of databases searched. Did 2,694 represent the total number of articles after initial duplicate removal (through, perhaps, the use of an unnamed software program) and 14 the number of additional duplicates identified during prescreening? The timing of duplicate removal described in the Results section (after prescreening) was different from that indicated in Figure 1 (before prescreening). Also, the Materials and Methods section did not specify and define the data extracted from each included article.

Ultimately, I believe the authors’ stated objectives were better suited to a narrative review than a systematic review. When attempting to differentiate between review types, O'Connor and Sargeant⁴ suggest considering the aim of the review. The goal of a systematic review is to answer a specific clinical or policy-related question while the goal of a narrative review is to integrate and interpret the literature on a topic. These goals guide the methods employed in a review. Systematic reviews require explicit, planned scientific methods to identify and select studies for inclusion and quantitative and qualitative methods to synthesize the results of those studies.⁵ Authors need to choose a review type appropriate for their objectives and resources, and journal editors and peer reviewers should consider whether authors have accurately identified their review type.

The Institute of Medicine and Journal of the American Medical Association recommend working with a medical librarian when developing narrative and systematic reviews^5,6; inclusion of a medical librarian or other information specialist in the design and execution of the study by Scotney et al may have eliminated the aforementioned concerns. Medical librarians provide expertise in review types and standards, search methods and documentation, and reference organization and review. The involvement of a medical librarian in the review process helps ensure that a review will be thorough and reproducible.

Laura L. Pavlech, DVM, MSLS Hirsh Health Sciences Library Tufts University Boston, Mass

The authors respond:

On behalf of the authors, I want to thank Dr. Pavlech for her interest in our study and constructive feedback.

In response to the initial point regarding keywords, we recognized inconsistencies in the terminology used in the current literature (one of the main driving forces for the review) and therefore carefully defined the specific key terms used throughout our report. We became convinced of the benefits of this approach during our preliminary investigations of which databases to incorporate into our review, in that articles returned from an initial MEDLINE search were found to relate to human health care and therefore were considered to not be relevant to our review.

Although we accept that there are various methods for conducting systematic reviews, we used the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines for our study and believe that the criteria were met as specified.

We thank Pavlech for pointing out the discrepancy in the timing of duplicate removal reported in the text versus Figure 1. Details given in the text are correct: there were 2,694 articles identified through our search strategy, of which 2,604 were removed after assessing the titles and abstracts. Fourteen of the remaining 90 articles were identified as duplicates, leaving 76 articles for full-text review.

Finally, we appreciate the advice to solicit the expertise of a medical librarian when developing narrative and systematic reviews and will take this into consideration for future reviews.

Rebekah Scotney, GCVSc Clinical Studies Centre School of Veterinary Science University of Queensland Gaton, QLD, Australia

Residue concerns following exposure of livestock to oil and petroleum products

Lately, there has been increased public concern regarding potential environmental impacts related to activities of the petroleum industry. Reports of contamination linked to accidental spills or leaks or to controversial extraction methods such as hydraulic fracturing seem to be increasing.¹ The most obvious concern is the effect of these environmental contaminations on general public health. An often overlooked concern, however, is the potential for these contaminants to make their way into the human food supply through contamination of water and feed resulting in chemical residues in animal meat and milk. Topical exposure of food animals, such as might occur during or after a flood, is also a concern, because of the potential for animal hide and wool products to be contaminated with petroleum products such as diesel fuel and other motor fuels and oils.

Many of the petroleum industries’ activities take place in remote areas uninhabited by humans, yet these areas are often inhabited by livestock and wildlife that humans may eventually consume.² There are many complicated steps from exploration of a potential drilling site to when the final raw material is collected. Harvested raw materials must be transported by rail car, tractor trailer, or pipeline for processing. Active drilling sites provide a potential for direct exposure of livestock and wildlife to oil and petroleum products and therefore must be properly fenced and maintained to ensure that these materials are inaccessible to animals. Water and feedstuffs can themselves become contaminated, becoming a source of indirect exposure. There are currently thousands of miles of pipelines strewn across remote areas that must be tended to and monitored regularly to avoid becoming a point source of exposure. Finally, abandoned oil and gas wells can serve as sources of exposure if livestock and wildlife have access to explore these sites.

To fully appreciate the risks associated with exposure of livestock to oil and petroleum products associated with exploration, drilling, recovery, and processing of raw materials by conventional and unconventional (hydraulic fracturing) methods of harvesting would require substantial study in a wide variety of scientific fields. A more general knowledge of these procedures, however, can provide veterinarians with the knowledge necessary to understand the risks faced by livestock and wildlife in areas where these activities are taking place.

To help practitioners interested in learning more about the risks inherent in these areas, the Food Animal Residue Avoidance Databank (FARAD) has posted on its website an article on exposure of livestock to oil and petroleum products.³ This article includes a broad background and overview of potential issues and describes what postexposure measures can be taken to mitigate the risk of residue violations and, more importantly, prevent contamination of the human food supply by harmful substances.

Practitioners are advised to minimize the risk of interactions between animals and petroleum products and byproducts and to contact FARAD immediately for guidance should such an exposure occur.

Keith D. DeDonder, MS, DVM Ronette Gehring, BVSc, MMedVet Jim E. Riviere, DVM, PhD Food Animal Residue Avoidance Databank

Institute of Computational Comparative Medicine

College of Vterinary Medicine

Kansas State University

Manhattan, Kan

Ronald E. Baynes, DVM, PhD

Food Animal Residue Avoidance Databank

Department of Population Health and Pathobiology

College of Veterinary Medicine

North Carolina State University

Raleigh, NC

Lisa A. Tell, DVM

Food Animal Residue Avoidance Databank

Department of Medicine and Epidemiology

School of Veterinary Medicine

University of California

Davis, Calif

Thomas W. Vickroy, PhD

Food Animal Residue Avoidance Databank

Department of Physiological Sciences

College of Veterinary Medicine

University of Florida

Gainesville, Fla