# Systematic Review and Meta-Analysis in Veterinary Science ## Key Takeaways - Systematic reviews and meta-analyses are critical for synthesizing fragmented veterinary evidence, requiring a predefined, reproducible protocol that specifies search strategies, eligibility criteria, and quality appraisal methods to minimize bias across species and production systems. - A robust literature search strategy is foundational, necessitating the use of multiple databases (e.g., MEDLINE/PubMed, Embase, Scopus, CAB Abstracts, regional databases) and inclusion of grey literature (conference proceedings, theses) to avoid selection bias. - Risk of bias assessment is crucial for internal validity, employing specific tools like the Cochrane Risk of Bias tool for RCTs, Newcastle-Ottawa Scale for observational studies, and SYRCLE for animal studies, with independent assessment by two reviewers. - Data extraction must be standardized and piloted, with outcome data extracted in duplicate, capturing detailed study characteristics, animal demographics, intervention specifics, and numerical results for synthesis. - Meta-analysis is appropriate only when studies are sufficiently similar; heterogeneity is quantified using I² and Cochran Q, and publication bias is assessed via funnel plots and Egger regression when at least 10 studies are available. - Reporting must adhere to the PRISMA 2020 statement, including a flow diagram detailing study selection, risk of bias summaries, and forest plots presenting individual and pooled effect estimates with confidence intervals. --- Systematic review and meta-analysis have become essential instruments for veterinary researchers who must synthesize a fragmented and methodologically heterogeneous evidence base. This article provides procedural guidance for designing, conducting, and reporting systematic reviews and meta-analyzes in veterinary medicine, with emphasis on literature search strategy, study selection, risk of bias assessment, data extraction, and quantitative synthesis. It serves the veterinary researcher who is planning an evidence synthesis project, the graduate student preparing a thesis chapter, and the clinician who must appraise published systematic reviews critically. The methods described apply across species and production systems, from companion animal oncology to food animal infectious disease, and they extend to wildlife health and One Health investigations. The distinction between a systematic review and a narrative review is fundamental. A systematic review answers a focused clinical or epidemiological question using a predefined, reproducible protocol that specifies search strategies, inclusion and exclusion criteria, and methods for quality appraisal and synthesis. A meta-analysis is the statistical combination of results from two or more separate studies that address the same question. Not every systematic review contains a meta-analysis, when studies are too heterogeneous in design, outcome definition, or population, a narrative synthesis within a systematic review framework may be the only defensible approach. The methodological logic rests on minimizing bias at every stage, because the conclusions of an evidence synthesis are only as trustworthy as the decisions made during its construction. ## At a Glance | Parameter | Decision or Standard | |---|---| | Protocol registration | Register the review protocol before data extraction begins, specify search dates, databases, and eligibility criteria | | Primary databases | MEDLINE/PubMed, Embase, Scopus, CAB Abstracts, and regional databases relevant to the species or region | | Grey literature | Include conference proceedings, theses, and institutional reports, document their sources explicitly | | Study selection | Two independent reviewers screen titles and abstracts, then full texts, disagreements resolved by consensus or third reviewer | | Risk of bias tools | Use the Cochrane Risk of Bias tool for randomized trials, Newcastle-Ottawa Scale for cohort and case-control studies, SYRCLE for animal studies | | Data extraction | Use a piloted, standardized extraction form, extract outcome data in duplicate | | Effect measure | Choose odds ratio, risk ratio, or mean difference based on outcome type and study design | | Heterogeneity | Report I² and Cochran Q, explore sources with subgroup analysis or meta-regression when I² exceeds 50% | | Publication bias | Assess with funnel plots and Egger regression when at least 10 studies are available | ## The Rationale for Systematic Methods in Veterinary Science Veterinary medicine has historically lagged behind human medicine in adopting formal evidence synthesis methods, but the need is acute. Clinical questions in veterinary practice often have answers scattered across case series, small randomized trials, and observational studies of variable quality. A systematic review imposes structure on this disorder. The [methodological quality assessment tools for preclinical and clinical studies, systematic review and meta-analysis, and clinical practice guideline](https://pubmed.ncbi.nlm.nih.gov/25594108/) identified 21 distinct assessment tools available to reviewers, a number that reflects both the maturity of the field and the complexity of choosing correctly. The same review noted that tools for randomized controlled trials were the most abundant, while instruments for observational designs were fewer and often less rigorously developed. The consequences of unsystematic synthesis are not theoretical. When the [Acute Dialysis Quality Initiative group conducted a systematic review of acute renal failure definitions](https://pubmed.ncbi.nlm.nih.gov/15312219/), they found more than 30 different definitions in active use, making comparison across studies nearly impossible. Their consensus process, built on a systematic literature search followed by structured expert deliberation, produced a standardized definition that transformed subsequent research. This example illustrates a broader principle: the review process itself can expose foundational problems in a field, such as inconsistent outcome definitions or inadequate reporting of adverse events, that no single primary study could reveal. ## Defining the Research Question and Protocol The research question must be structured before any literature search begins. The PICO framework, adapted for veterinary contexts, provides the standard architecture. Population specifies the species, breed, age class, production type, or disease status. Intervention names the therapeutic, preventive, or diagnostic procedure under study. Comparison identifies the alternative, which may be placebo, no treatment, a different intervention, or a reference standard. Outcome lists the clinically relevant endpoints, with primary and secondary outcomes distinguished in advance. A well-constructed question determines every subsequent decision. Consider the difference between asking "Does antimicrobial therapy improve outcomes in canine pyoderma?" and asking "In dogs with superficial bacterial pyoderma, what is the comparative efficacy of systemic versus topical antimicrobial therapy for clinical cure at 14 days?" The second question specifies the population, the competing interventions, and the outcome with enough precision to guide search terms and eligibility criteria. The first question is too broad to yield a manageable or interpretable evidence base. The protocol should be written and, where possible, registered before screening begins. It must state the databases to be searched, the search date range, the language restrictions, and the eligibility criteria in operational terms. The protocol also specifies the methods for risk of bias assessment, data extraction, and synthesis. A protocol that is modified after the search has begun risks introducing bias, because the reviewer may unconsciously adjust criteria to include favored studies or exclude inconvenient ones. The [WOAH animal health surveillance standards](https://www.woah.org/en/what-we-do/animal-health-and-welfare/disease-data-collection/) emphasize the same principle in a different context: methods must be documented before data collection so that the process is auditable and reproducible. ## Literature Search Strategy The search strategy is the foundation of the systematic review, and its quality determines the completeness of the evidence base. A search that misses relevant studies introduces selection bias that no statistical technique can correct. The search should combine controlled vocabulary terms, such as Medical Subject Headings, with free-text terms and synonyms. For veterinary questions, the search must account for species-specific terminology, breed names, and production-system descriptors that may not appear in human medical databases. Multiple databases are required because no single database indexes the full veterinary literature. MEDLINE and PubMed cover the biomedical literature broadly, but veterinary-specific journals are more completely indexed in CAB Abstracts and Scopus. Embase adds European and pharmaceutical literature. Regional databases, such as Agricola for North American agricultural research or the African Journals Online for African veterinary science, may be necessary for questions with geographic scope. The search should also include grey literature: conference proceedings, graduate theses, government reports, and institutional publications. The [updated systematic review and meta-analysis of bariatric surgery](https://pubmed.ncbi.nlm.nih.gov/24352617/) searched six databases including ClinicalTrials.gov and identified 25,060 initial articles, of which only 259 met inclusion criteria. This ratio, roughly one percent, is typical for systematic reviews and underscores the importance of a sensitive search strategy that casts a wide net before applying restrictive eligibility criteria. Search filters for study design can improve precision, but they carry the risk of excluding relevant studies if the filter is too restrictive. A filter limited to randomized controlled trials will miss observational studies that may be the only evidence available for many veterinary interventions. The search strategy should be reported in full in the final review, including the exact date of each database search, so that readers can assess its currency and reproducibility. ## Screening and Study Selection Screening converts the raw search output into a set of studies eligible for synthesis. The process should follow the protocol exactly as registered, and any deviation must be documented and justified. Title and abstract screening removes obviously irrelevant records. Full-text screening then applies the eligibility criteria to the remaining records. Both stages should be performed independently by at least two reviewers. Disagreements are resolved by consensus or by a third reviewer. The [methodological quality assessment tools for preclinical and clinical studies, systematic reviews, and meta-analyzes](https://pubmed.ncbi.nlm.nih.gov/25594108/) described by Zeng and colleagues provide a structured basis for this workflow. A screening form should be piloted on a sample of 20 to 30 records before full application. The form lists each inclusion and exclusion criterion as a separate item with yes, no, or unclear response options. A record fails screening if any inclusion criterion is unmet or any exclusion criterion is met. The number of records excluded at each stage must be reported in a flow diagram. The diagram should state the number of records identified from each database, the number of duplicates removed, the number screened, the number excluded at title and abstract stage, the number sought for full text, the number not retrieved, the number assessed for eligibility, and the number excluded at full text with reasons. Common reasons for full-text exclusion include no original data, no relevant outcome, wrong population, and duplicate publication. ## Risk of Bias Assessment Risk of bias assessment evaluates the internal validity of included studies. The choice of tool depends on study design. The Cochrane Risk of Bias tool is the preferred instrument for randomised controlled trials. The Newcastle-Ottawa Scale is recommended for cohort and case-control studies. The Methodological Index for Non-Randomized Studies (MINORS) is suitable for non-randomised interventional studies. A systematic comparison of these instruments is available in the [review of methodological quality assessment tools for preclinical and clinical studies](https://pubmed.ncbi.nlm.nih.gov/25594108/). For animal intervention studies, the SYRCLE Risk of Bias tool is the standard instrument. It adapts the Cochrane domains to the specific threats in animal research, including random housing, blinded outcome assessment, and allocation concealment. The [methodological quality assessment tools for primary and secondary medical studies](https://pubmed.ncbi.nlm.nih.gov/32111253/) review by Ma and colleagues provides guidance on matching tools to study types. Risk of bias assessment must be performed independently by two reviewers. Each domain is rated as low risk, high risk, or unclear risk. The results should be presented in a table and in a risk of bias summary figure. Studies with high risk of bias in critical domains should not simply be excluded. Instead, sensitivity analysis should examine whether conclusions change when high-risk studies are removed. ## Data Extraction Data extraction captures all information needed for synthesis and for assessment of heterogeneity. A standardized extraction form should be developed and piloted before use. The form records study characteriztics, participant or animal characteriztics, interventions or exposures, comparators, outcomes, and numerical results. Study characteriztics include author, year, country, study design, and funding source. Animal characteriztics include species, breed, sex, age, weight, and health status. Intervention details include agent, dose, route, frequency, duration, and co-interventions. Outcome data include the number of events, sample sizes, means, standard deviations, and effect estimates with confidence intervals. Extraction should be performed independently by two reviewers. Disagreements are resolved by consensus. Where outcome data are missing, the review authors should contact the original study authors. If data cannot be obtained, the study is included in the narrative synthesis but excluded from the relevant meta-analysis. For continuous outcomes, record the measure of central tendency and dispersion, the number of animals or participants per group, and the direction of effect. For dichotomous outcomes, record the number of events and the total number per group. For time-to-event outcomes, record the hazard ratio and its confidence interval. ## Data Synthesis and Meta-Analysis Meta-analysis is appropriate only when studies are sufficiently similar in population, intervention, comparator, and outcome. Clinical and methodological heterogeneity must be assessed before statistical pooling. If studies differ substantially in any of these domains, narrative synthesis is preferred. The choice of effect measure depends on the outcome type. Risk ratios or odds ratios are used for dichotomous outcomes. Mean differences are used for continuous outcomes measured on the same scale. Standardized mean differences are used when different scales measure the same construct. Hazard ratios are used for time-to-event outcomes. A random-effects model is generally preferred over a fixed-effect model in veterinary meta-analysis. Random-effects models incorporate between-study variance and produce wider confidence intervals when heterogeneity is present. The choice should be specified in the protocol and justified. Heterogeneity is quantified using the I-squared statistic. I-squared values of 25%, 50%, and 75% correspond to low, moderate, and high heterogeneity. The Cochrane Handbook cautions that I-squared is a descriptive statistic and should not be used as the sole basis for deciding whether to pool studies. The Q statistic tests the null hypothesis that all studies share a common effect size. Publication bias is assessed using funnel plots and statistical tests such as Egger's regression test. Funnel plot asymmetry can indicate publication bias, but it can also arise from true heterogeneity or from poor methodological quality in small studies. The [updated systematic review and meta-analysis of bariatric surgery outcomes](https://pubmed.ncbi.nlm.nih.gov/24352617/) by Chang and colleagues illustrates the importance of comprehensive literature searching and explicit exclusion criteria for reducing the risk of publication bias. Sensitivity analyzes test the robustness of the pooled estimate. Common sensitivity analyzes include removing studies at high risk of bias, removing outliers, changing the effect measure, and changing the model from random effects to fixed effect. Subgroup analyzes examine whether treatment effects differ by species, breed, sex, age, or disease severity. ## Reporting the Review The PRISMA 2020 statement provides the reporting framework for systematic reviews and meta-analyzes. The checklist covers title, abstract, introduction, methods, results, discussion, and funding. Each item should be addressed explicitly, and the checklist should be submitted with the manuscript. The abstract should follow the PRISMA for Abstracts format. It reports the background, objectives, eligibility criteria, information sources, risk of bias assessment, synthesis methods, main results, and interpretation. The abstract is often the only part read by clinicians, so it must be accurate and complete. The results section reports the flow of studies through the review, the characteriztics of included studies, the risk of bias within studies, the results of individual studies, and the results of syntheses. Forest plots present the effect estimate and confidence interval for each study and the pooled estimate. The forest plot should include the weight assigned to each study and the heterogeneity statistics. The discussion interprets the results in the context of the broader evidence base. It addresses the strengths and limitations of the review, the certainty of the evidence, and the implications for practice and research. The certainty of the evidence should be assessed using the GRADE approach, which considers risk of bias, inconsistency, indirectness, imprecision, and publication bias. The [WOAH animal health surveillance standards](https://www.woah.org/en/what-we-do/animal-health-and-welfare/disease-data-collection/) and the [WOAH terrestrial animal health code](https://www.woah.org/en/what-we-do/standards/codes-and-manuals/terrestrial-code-online-access/) provide context for how systematic reviews inform international standard setting in animal health. Reviews that address surveillance, diagnostic test accuracy, or intervention effectiveness can directly support these standards. ## Protocol and Reporting Checklist | Review stage | Minimum reporting requirement | Common failure mode | |---|---|---| | Protocol | Registration number, date, and amendments | No protocol registered | | Search | Full search strategy for one database, date of search, limits applied | Search strategy not reproducible | | Screening | Number screened, excluded, and reasons for exclusion | Reasons for exclusion not stated | | Risk of bias | Tool used, domains assessed, reviewer independence | Tool not matched to study design | | Data extraction | Items extracted, process for resolving disagreements | Single reviewer extraction | | Synthesis | Model choice, heterogeneity statistics, sensitivity analyzes | Pooling despite high heterogeneity | | Reporting | PRISMA checklist submitted, flow diagram included | Checklist not submitted | The checklist above summarizes the minimum reporting requirements for each stage of a systematic review. Each item should be addressed in the final manuscript, and the completed checklist should be available to editors and reviewers. ## Recognized Complications and Failure Modes Systematic reviews in veterinary science fail in characteriztic patterns. The most consequential is the retrieval gap, where relevant studies never enter the screening pool because the search strategy omitted species-specific terminology, grey literature sources, or non-English databases. Detection requires a documented search protocol with transparent reporting of database coverage, date ranges, and search strings, as recommended in the [WOAH animal health surveillance standards](https://www.woah.org/en/what-we-do/animal-health-and-welfare/disease-data-collection/), which emphasize reproducibility in evidence collection. Publication bias distorts effect estimates when negative or null studies remain unpublished. Veterinary researchers can detect this through funnel plot asymmetry, Egger regression, or trim-and-fill methods, though these diagnostics lose power with fewer than ten studies. A second failure mode is clinical heterogeneity masquerading as statistical homogeneity. Pooling studies that differ in species, production system, diagnostic criteria, or outcome definition produces a precise but meaningless summary estimate. The [CDC principles of epidemiology in public health practice](https://www.cdc.gov/csels/dsepd/ss1978/index.html) describe how case definitions shape every downstream comparison, and the same logic applies to veterinary outcomes. Data extraction errors propagate silently. Double extraction with independent verification, followed by reconciliation of disagreements, remains the standard safeguard. Early detection of extraction drift requires periodic rechecking of a random sample of completed forms against source documents. | Observation | Likely cause | Discriminating check | |---|---|---| | Funnel plot asymmetry | Publication bias, small-study effects, or true heterogeneity | Compare fixed-effect and random-effects estimates, examine study characteriztics by size | | Excessive heterogeneity (I² > 75%) | Clinical or methodological diversity | Subgroup analysis by species, design, or outcome definition, sensitivity analysis excluding outliers | | Few studies pass full-text screening | Overly restrictive eligibility criteria or narrow search | Re-run search with broadened terms, verify database coverage | | Effect estimate changes markedly after excluding one study | Single-study leverage or fragile result | Influence diagnostics (leave-one-out analysis), assess study quality | | Disagreement between extractors | Ambiguous outcome definitions | Refine extraction manual, pilot extraction on three studies before full screening | ## Common Errors and Corrective Actions Less experienced reviewers frequently confuse risk of bias with reporting quality. A study can be poorly reported yet methodologically sound, and vice versa. The distinction matters because risk of bias tools assess internal validity, while reporting checklists assess completeness of description. Selecting the appropriate tool depends on study design, as summarized in the [systematic review of methodological quality assessment tools](https://pubmed.ncbi.nlm.nih.gov/25594108/), which recommends the Cochrane Risk of Bias tool for randomised trials and the Newcastle-Ottawa Scale for observational studies. A second recurring error is post hoc modification of eligibility criteria to accommodate inconvenient studies. This undermines the protocol's protective function. Corrective action is to pre-specify criteria, register the protocol, and report any deviations with justification. A third error involves narrative synthesis of studies that were never designed to answer the same question. When meta-analysis is inappropriate, the review should state this explicitly and provide a structured narrative synthesis organized by outcome, population, and intervention characteriztics instead of a chronological listing. Students often conflate statistical significance with clinical importance. A meta-analysis may show a statistically significant pooled effect that is biologically trivial for the target species or production context. Reporting should include absolute measures, such as risk differences or number needed to treat, alongside relative measures. ## Limitations of Current Evidence The veterinary evidence base is thinner than its human medical counterpart. Many veterinary systematic reviews identify fewer than twenty eligible studies, and randomised controlled trials are scarce in fields such as wildlife medicine, exotic animal practice, and production medicine under field conditions. The [updated systematic review and meta-analysis of bariatric surgery](https://pubmed.ncbi.nlm.nih.gov/24352617/) screened over 25,000 records to include 259 studies, veterinary reviews rarely approach this volume of available literature. Expert opinion still differs on several methodological points. Whether to include conference abstracts, which may contain incomplete outcome data, remains contested. Some reviewers exclude them to avoid unreliable estimates, others include them to reduce publication bias. Similarly, no consensus exists on how to handle studies with unit-of-analysis errors, such as litter-based randomisation analyzed at the individual animal level. The [methodological quality assessment tools review](https://pubmed.ncbi.nlm.nih.gov/32111253/) notes that no single tool fits all study types, and veterinary reviewers must often adapt instruments designed for human medicine, a process that requires justification and transparency. ## Referral, Consultation, and Reporting Triggers Systematic reviews benefit from early consultation with a librarian or information specialist trained in systematic searching. This consultation should occur before the search is executed, not after retrieval problems emerge. Statistical support is warranted when the review will include meta-analysis, particularly for complex data structures such as clustered or longitudinal outcomes. Regulatory reporting obligations arise when the review identifies evidence of adverse effects that may affect product safety or when findings bear on notifiable disease surveillance. The [WOAH terrestrial animal health code](https://www.woah.org/en/what-we-do/standards/codes-and-manuals/terrestrial-code-online-access/) sets international standards for disease notification, and reviewers whose synthesis reveals previously unrecognised patterns of disease occurrence should consult the relevant veterinary authority. Laboratory involvement may be needed to verify diagnostic methods used across included studies, especially when outcome definitions depend on laboratory confirmation. The [MSD Veterinary Manual](https://www.msdvetmanual.com/) provides species-specific guidance on diagnostic standards that can inform such verification. ## Frequently Asked Questions ### How much time and funding should we budget for a systematic review in veterinary medicine? A rigorous systematic review typically requires 6 to 18 months from protocol registration to submission. A team of at least three reviewers is needed to allow independent screening and data extraction, with a fourth member available to resolve disagreements. Budget for database access fees, reference management software, interlibrary loan charges, and statistical support if you lack meta-analytic expertise. Open-access publication fees may also apply. If funding is constrained, restrict the search to fewer databases and a narrower date range, but document these deviations in the protocol. The [methodological quality assessment tools for preclinical and clinical studies](https://pubmed.ncbi.nlm.nih.gov/25594108/) described by Zeng and colleagues can help you prioritize which quality checks are essential when resources are limited. ### What should we do when the ideal risk of bias tool is unavailable for our study design? Select the closest validated tool and adapt it transparently. For randomised controlled trials, the Cochrane Risk of Bias tool remains the preferred instrument. For observational designs, the Newcastle-Ottawa Scale serves cohort and case-control studies, while MINORS covers non-randomised interventional studies. When no tool fits, construct a bespoke appraisal form based on the domains most relevant to your question, such as allocation concealment, blinding, attrition, and outcome measurement validity. Document every modification in an appendix. The [systematic review of methodological quality assessment tools](https://pubmed.ncbi.nlm.nih.gov/32111253/) by Ma and colleagues provides a catalogue of options across study types, which helps justify your choice to peer reviewers. ### How does the review process differ when the evidence base includes multiple species? Species heterogeneity demands explicit handling at several stages. During screening, decide whether the research question requires species-specific inclusion criteria or permits cross-species synthesis. During data extraction, record species, breed, age, production system, and husbandry conditions as separate fields. During meta-analysis, consider species as a subgroup variable or as a covariate in meta-regression. If biological plausibility suggests different treatment effects across species, present separate pooled estimates instead of a single combined effect. The [WOAH terrestrial animal health standards](https://www.woah.org/en/what-we-do/standards/codes-and-manuals/terrestrial-code-online-access/) emphasize that surveillance and intervention evidence must be interpreted within species-specific contexts, a principle that applies equally to therapeutic and diagnostic systematic reviews. ### What record keeping practices are essential for reproducibility? Maintain a complete audit trail from protocol to publication. Save every search strategy exactly as executed in each database, including date stamps and the number of records retrieved. Keep screening decisions in a shared spreadsheet or dedicated software, with reasons for exclusion recorded at full-text stage. Preserve all versions of the data extraction form and the raw extracted data. Document risk of bias judgments with supporting quotations from each study. Archive the statistical code used for meta-analysis. The [CDC principles of epidemiology in public health practice](https://www.cdc.gov/csels/dsepd/ss1978/index.html) stress that systematic documentation underpins the validity of any epidemiologic investigation, and the same standard applies to evidence synthesis. ### How should we respond when a client or supervisor asks why the review is taking so long? Explain that systematic reviews differ from narrative summaries because they use explicit, reproducible methods to minimize bias. The screening phase alone often involves reviewing thousands of titles, and each full-text article must be assessed against predefined eligibility criteria. Risk of bias appraisal and data extraction are performed independently by multiple reviewers to ensure accuracy. The [WOAH animal health surveillance standards](https://www.woah.org/en/what-we-do/animal-health-and-welfare/disease-data-collection/) demonstrate that international bodies accept this investment of time because it produces conclusions that withstand scrutiny. Offer a timeline with milestones and interim progress updates. If a rapid answer is needed, propose a scoping review or a focused rapid review with a narrower question, and be explicit about the trade-offs in certainty. ### What are the options when full-text articles cannot be obtained? First, exhaust legitimate access routes: institutional subscriptions, interlibrary loan, and direct requests to corresponding authors. If an article remains unavailable, record it as an unobtainable study and report this in the PRISMA flow diagram. Sensitivity analysis should compare results with and without the missing studies to assess their likely impact. Contacting authors for unpublished data is acceptable and often productive, particularly for conference abstracts that never reached full publication. Do not rely on abstracts alone for data extraction, as they frequently omit methodological details needed for risk of bias assessment. The [MSD Veterinary Manual professional edition](https://www.msdvetmanual.com/) illustrates how clinical reference sources can supplement background understanding, but they cannot substitute for the primary study data required in a systematic review. ## Related Clinical & Scientific Guides * [Evaluating Veterinary Surveillance System Attributes](/knowledge/veterinary-medicine/veterinary-epidemiology/evaluating-veterinary-surveillance-system-attributes) * [Network Analysis for Infectious Disease Spread in Animal Populations](/knowledge/veterinary-medicine/veterinary-epidemiology/network-analysis-infectious-disease-spread-animal-populations) * [Randomized Controlled Trials in Veterinary Field Settings](/knowledge/veterinary-medicine/veterinary-epidemiology/randomized-controlled-trials-veterinary-field-settings) ## References and Further Reading - [The effectiveness and risks of bariatric surgery: an updated systematic review and meta-analysis, 2003-2012.](https://pubmed.ncbi.nlm.nih.gov/24352617/). 2014. - [A systematic review on reporting and assessment of adverse effects associated with transcranial direct current stimulation.](https://pubmed.ncbi.nlm.nih.gov/21320389/). 2011. - [Acute renal failure - definition, outcome measures, animal models, fluid therapy and information technology needs: the Second International Consensus Conference of the Acute Dialysis Quality Initiative (ADQI) Group.](https://pubmed.ncbi.nlm.nih.gov/15312219/). 2004. - [Eosinophilic esophagitis: updated consensus recommendations for children and adults.](https://pubmed.ncbi.nlm.nih.gov/21477849/). 2011. - [Methodological quality (risk of bias) assessment tools for primary and secondary medical studies: what are they and which is better?](https://pubmed.ncbi.nlm.nih.gov/32111253/). 2020. - [The methodological quality assessment tools for preclinical and clinical studies, systematic review and meta-analysis, and clinical practice guideline: a systematic review.](https://pubmed.ncbi.nlm.nih.gov/25594108/). 2015. - [WOAH Animal Health Surveillance Standards](https://www.woah.org/en/what-we-do/animal-health-and-welfare/disease-data-collection/). WOAH. - [CDC Principles of Epidemiology in Public Health Practice](https://www.cdc.gov/csels/dsepd/ss1978/index.html). CDC. - [MSD Veterinary Manual, Professional Edition](https://www.msdvetmanual.com/). MSD Veterinary Manual. ## Related Articles - [Sensitivity Analysis in Veterinary Disease Models](/knowledge/veterinary-medicine/veterinary-epidemiology/sensitivity-analysis-veterinary-disease-models) - [Risk Factor Analysis for Disease in Animal Populations](/knowledge/veterinary-medicine/veterinary-epidemiology/risk-factor-analysis-disease-animal-populations) - [Time Series Analysis for Veterinary Disease Surveillance](/knowledge/veterinary-medicine/veterinary-epidemiology/time-series-analysis-veterinary-disease-surveillance) - [Cohort Studies in Veterinary Medicine: Design and Analysis](/knowledge/veterinary-medicine/veterinary-epidemiology/cohort-studies-veterinary-medicine-design-analysis) - [Spatial Epidemiology in Veterinary Science: Mapping Disease Clusters](/knowledge/veterinary-medicine/veterinary-epidemiology/spatial-epidemiology-veterinary-science-mapping-disease-clusters) > This article is educational professional reference material for veterinary audiences. It is not a substitute for veterinary diagnosis, individual clinical judgment, current product labeling, or applicable regulatory requirements.