Loading
Expert Witness Insights by Area of Expertise
12 min read
Published Date:
Last Modified Date:
Epidemiology expert witnesses help evaluate whether population-level research supports an association or causal relationship between an exposure and a health outcome in litigation. This article explains how experts assess study design, bias, confounding, risk estimates, statistical uncertainty, general and specific causation, and whether epidemiological opinions satisfy Federal Rule of Evidence 702.
Epidemiology expert witnesses help attorneys and courts evaluate whether population-level research supports an association or causal relationship between an exposure and a health outcome. Their analysis may be relevant in toxic tort, pharmaceutical liability, occupational disease, environmental exposure, and infectious disease litigation.
Epidemiological studies can identify differences in disease rates between populations, but those differences do not automatically establish causation. Experts must assess study design, statistical findings, potential bias, confounding variables, and the applicability of research to the disputed exposure.
In litigation, epidemiological research can help determine whether an alleged exposure is associated with higher disease rates in a defined population.
For example, differences in disease rates between workers exposed to industrial chemicals and those not similarly exposed may indicate an association without establishing causation.
The Federal Judicial Center and the National Academies of Sciences, Engineering, and Medicine address these methods and the limits of applying population-level findings to individual causation in the Reference Manual on Scientific Evidence, Fourth Edition.
Epidemiological evidence is particularly relevant to two distinct inquiries:
General causation: Whether an agent or exposure is capable of causing a particular health outcome under relevant circumstances.
Specific causation: Whether the alleged exposure was a cause of a particular individual's condition, with its legal significance depending on the governing causation standard.
Population-level research may support general causation, but additional case-specific evidence is generally necessary to evaluate specific causation.
Epidemiology expert witnesses help attorneys and courts evaluate whether population-level research supports an association or causal relationship between an exposure and a health outcome. Their analysis may be relevant in toxic tort, pharmaceutical liability, occupational disease, environmental exposure, and infectious disease litigation.
Epidemiological studies can identify differences in disease rates between populations, but those differences do not automatically establish causation. Experts must assess study design, statistical findings, potential bias, confounding variables, and the applicability of research to the disputed exposure.
In litigation, epidemiological research can help determine whether an alleged exposure is associated with higher disease rates in a defined population.
For example, differences in disease rates between workers exposed to industrial chemicals and those not similarly exposed may indicate an association without establishing causation.
The Federal Judicial Center and the National Academies of Sciences, Engineering, and Medicine address these methods and the limits of applying population-level findings to individual causation in the Reference Manual on Scientific Evidence, Fourth Edition.
Epidemiological evidence is particularly relevant to two distinct inquiries:
General causation: Whether an agent or exposure is capable of causing a particular health outcome under relevant circumstances.
Specific causation: Whether the alleged exposure was a cause of a particular individual's condition, with its legal significance depending on the governing causation standard.
Population-level research may support general causation, but additional case-specific evidence is generally necessary to evaluate specific causation.
Epidemiology expert witnesses assess whether a study's design, measurements, statistical methods, and findings provide reliable information about the disputed exposure and health outcome.
Different epidemiological study designs provide different forms of evidence:
Cohort studies: Compare health outcomes in groups defined by exposure status. Differences in other risk factors between groups and incomplete follow-up may affect the findings.
Case-control studies: Compare exposure histories or other exposure measures between individuals with a particular condition (cases) and those without it (controls). Exposure measurement, recall, and the selection of controls may introduce bias.
Cross-sectional studies: Examine exposure and health status at a particular time. These studies often cannot establish whether exposure preceded the outcome.
Ecological studies: Analyze exposure and disease patterns using group-level data. Group-level associations may not accurately describe individual relationships.
An expert may examine how comparison groups were selected, whether the study population is relevant to the dispute, and whether the methods can reliably evaluate the proposed association.
Reliable epidemiological analysis depends on how accurately researchers measure exposure and disease.
An expert may examine:
How exposure intensity, duration, and timing were measured.
Whether diagnoses or health outcomes were independently verified.
Whether exposure classifications were consistent across participants.
Whether missing data or measurement errors could influence the results.
Whether the study accounted for relevant differences between comparison groups.
These considerations are especially important in occupational and environmental exposure litigation, where historical exposure records may be incomplete.
Common concerns include:
Selection bias: Systematic differences in how participants enter or remain in a study.
Information bias: Errors in recording or classifying exposures and health outcomes.
Confounding: Distortion of an association by another factor related to both exposure and disease.
Random error: Variability in study results attributable to chance.
For example, smoking may be a confounding factor when researchers investigate an occupational exposure associated with lung disease.
Experts assess how well researchers addressed these potential sources of error and whether any remaining limitations materially affect the findings.
Epidemiological studies use statistical measures to describe the strength and uncertainty of associations. These measures help experts evaluate scientific findings but do not independently establish causation.
Relative risk, also called a risk ratio, compares the risk of an outcome in an exposed population with the risk in an unexposed population.
A relative risk of 1.0 indicates equal observed risk between the groups, while a relative risk of 1.5 indicates a 50% higher observed risk in the exposed group.
An odds ratio compares odds rather than risks and is commonly used in case-control studies. It can approximate relative risk when the outcome is rare, but the measures are not generally interchangeable.
Experts also consider absolute risk because a substantial relative increase may correspond to a small absolute increase when the underlying condition is uncommon.
Confidence intervals describe the precision of statistical estimates under the assumptions of the method used.
A wide confidence interval indicates greater statistical uncertainty, while a narrow interval generally suggests greater precision. Neither establishes that a study is free from bias or confounding.
Statistical significance is commonly assessed using a p-value, which measures how incompatible the observed data are with a specified null hypothesis under the assumptions of the statistical test. It does not establish whether an association is causal. A statistically nonsignificant result also does not establish the absence of an effect.
Experts should consider effect estimates, confidence intervals, statistical power, and methodological limitations rather than relying on statistical significance alone.
An epidemiological association does not necessarily establish that an exposure causes a disease. Causal interpretation requires evaluating alternative explanations and the broader scientific evidence.
Epidemiologists often draw on the considerations outlined in Austin Bradford Hill's 1965 publication, The Environment and Disease: Association or Causation?
Relevant considerations include:
Temporality: Whether exposure occurred before the health outcome.
Strength: The magnitude of the observed association.
Consistency: Whether similar findings appear across different studies and populations.
Dose-response relationship: Whether differences in exposure correspond to differences in disease risk.
Biological plausibility: Whether scientific knowledge supports the proposed relationship.
Coherence: Whether the proposed relationship is consistent with other relevant scientific evidence.
These considerations are not a mandatory checklist, and no fixed number establishes causation. However, temporality is necessary: an alleged causal exposure must precede the health effect it is claimed to cause. The relevance of the remaining considerations depends on the scientific question and the broader evidence.
When several epidemiological studies address the same exposure-disease relationship, an expert may compare their populations, methods, findings, and limitations.
Systematic reviews and meta-analyses may help synthesize research when the studies are sufficiently comparable. However, combining results does not automatically resolve weaknesses in the original studies.
Causal analysis may also incorporate toxicological research, biological mechanisms, exposure measurements, and other relevant scientific evidence.
Population-level research may support a general causal relationship without establishing that an exposure caused a particular individual's condition.
This distinction is particularly important when a disease has several recognized risk factors or can occur without the alleged exposure.
For example, a study may identify elevated disease risk among workers exposed to a chemical. Applying that finding to a plaintiff requires assessing whether the plaintiff's actual exposure and other relevant characteristics are sufficiently comparable to those studied.
Relevant case-specific materials may include:
Employment histories and occupational exposure records.
Environmental measurements and available exposure estimates.
Medical records, diagnostic findings, and disease chronology.
Exposure intensity, duration, and relevant latency periods.
Other medical, occupational, environmental, or lifestyle risk factors.
An epidemiology expert may evaluate whether the published studies reasonably apply to the individual's circumstances. Clinical specialists, toxicologists, occupational medicine physicians, or exposure assessment professionals may provide complementary analysis.
A population-level risk estimate does not automatically represent the probability that an exposure caused an individual's disease.
Some courts consider reliable epidemiological evidence showing a relative risk greater than 2.0 when evaluating the sufficiency of specific causation evidence. The Reference Manual on Scientific Evidence, Fourth Edition explains that courts differ in their treatment of this risk-doubling evidence.
Relevant considerations include the governing causation standard, whether the association is causal, the reliability of the risk estimate, the research's applicability to the individual, and competing risk factors. A relative risk greater than 2.0 is neither a universal legal requirement nor an automatic determination of individual causation.
Epidemiology experts may review studies reporting associations between medications or medical products and adverse outcomes. They may examine differences between patient populations and whether researchers accounted for relevant medical conditions and other risk factors.
In toxic tort and occupational disease disputes, experts may evaluate occupational cohorts, exposure intensity and duration, dose-response findings, and research relevant to alleged exposure-related health outcomes.
When environmental litigation involves allegations of elevated community disease rates linked to pollution or contamination, an expert may compare observed and expected disease patterns and assess potential exposures.
However, an apparent disease cluster does not independently establish a shared cause. In its guidance on unusual cancer patterns, the Centers for Disease Control and Prevention identifies limitations involving incomplete exposure information, competing risk factors, population characteristics, and random variation.
In infectious disease disputes, epidemiologists may examine infection patterns, transmission pathways, and surveillance data to assess whether the evidence is consistent with a proposed outbreak source or transmission scenario. Such analysis may not establish how every individual infection occurred.
Relevant expertise may include epidemiology, public health, biostatistics, occupational epidemiology, environmental epidemiology, or pharmacoepidemiology, depending on the scientific and legal questions involved.
Attorneys may consider an expert's experience with:
The study designs and statistical methods involved in the dispute.
The disease, population, or exposure being investigated.
Critical appraisal of scientific literature and causal inference.
Analysis of bias, confounding, and uncertainty.
Explaining technical findings and methodological limitations.
An expert's qualifications should correspond to the opinions offered. Depending on the case, related disciplines such as biostatistics, toxicology, and clinical medicine may be necessary to address questions outside the epidemiologist's expertise.
In federal proceedings governed by the Federal Rules of Evidence, Rule 702 establishes the admissibility requirements for epidemiological expert testimony. As amended effective December 1, 2023, the rule requires the party offering the testimony to show that those requirements are more likely than not satisfied.
These requirements address whether:
The expert is appropriately qualified and the testimony will assist the factfinder.
The opinion rests on sufficient facts or data.
The testimony uses reliable principles and methods.
The opinion reflects a reliable application of those methods to the case.
As clarified by the 2023 amendment, the court must assess whether the expert has sufficient facts or data and whether the methods were reliably applied to support the opinions offered. The court need not resolve every factual dispute or determine that the conclusions are ultimately correct. Once Rule 702's admissibility requirements are satisfied, challenges that do not implicate those requirements may affect the weight assigned to the testimony.
In Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579 (1993), the Supreme Court explained that federal trial judges must evaluate the reliability and relevance of scientific expert testimony.
The Court identified flexible considerations, including testing, peer review, potential error rates, and general acceptance, that may inform the evaluation of epidemiological opinions.
In General Electric Co. v. Joiner, 522 U.S. 136 (1997), the Supreme Court held that appellate courts review decisions admitting or excluding expert scientific testimony for abuse of discretion. Applying that standard, the Court upheld the exclusion of opinions linking the plaintiff's alleged PCB exposure to his cancer. The district court had reasonably found that the studies did not adequately support those opinions.
The decision illustrates why experts must explain the connection between their scientific evidence and conclusions, particularly when studies involve different populations, exposures, or outcomes.
State courts may apply different expert-evidence standards, while the substantive requirements for establishing causation depend on the law governing the claim. Counsel should distinguish the admissibility of epidemiological testimony from whether the evidence is sufficient to satisfy the applicable causation standard.
Is epidemiological evidence always required to establish causation in litigation?
No. Epidemiological evidence can be important in exposure-related litigation, but it is not universally required. Depending on the jurisdiction and circumstances, other reliable scientific or medical evidence may support a causation opinion.
Does a relative risk greater than 2.0 establish specific causation?
No. Some courts consider risk-doubling evidence when evaluating the sufficiency of individual causation evidence, but it is not a universal threshold. The association must be causal, the estimate reliable, and the research sufficiently applicable to the individual.
Can an epidemiology expert rely on studies that are not statistically significant?
Potentially, yes. Statistical significance is one consideration in evaluating scientific evidence. Experts must also assess statistical power, effect estimates, confidence intervals, methodological limitations, and the broader research record. Admissibility depends on the governing legal standard.
What happens when epidemiological studies produce conflicting findings?
A qualified expert may compare study designs, populations, exposure measurements, statistical methods, and potential sources of error. These differences may explain inconsistent findings and affect the weight assigned to individual studies.
Can epidemiological research involving different populations support an expert opinion?
Yes, when a reliable scientific basis supports applying the research to the disputed circumstances. Differences in exposure levels, population characteristics, and health outcomes may limit the conclusions that can reasonably be drawn.
Epidemiology expert witnesses assess whether a study's design, measurements, statistical methods, and findings provide reliable information about the disputed exposure and health outcome.
Different epidemiological study designs provide different forms of evidence:
Cohort studies: Compare health outcomes in groups defined by exposure status. Differences in other risk factors between groups and incomplete follow-up may affect the findings.
Case-control studies: Compare exposure histories or other exposure measures between individuals with a particular condition (cases) and those without it (controls). Exposure measurement, recall, and the selection of controls may introduce bias.
Cross-sectional studies: Examine exposure and health status at a particular time. These studies often cannot establish whether exposure preceded the outcome.
Ecological studies: Analyze exposure and disease patterns using group-level data. Group-level associations may not accurately describe individual relationships.
An expert may examine how comparison groups were selected, whether the study population is relevant to the dispute, and whether the methods can reliably evaluate the proposed association.
Reliable epidemiological analysis depends on how accurately researchers measure exposure and disease.
An expert may examine:
How exposure intensity, duration, and timing were measured.
Whether diagnoses or health outcomes were independently verified.
Whether exposure classifications were consistent across participants.
Whether missing data or measurement errors could influence the results.
Whether the study accounted for relevant differences between comparison groups.
These considerations are especially important in occupational and environmental exposure litigation, where historical exposure records may be incomplete.
Common concerns include:
Selection bias: Systematic differences in how participants enter or remain in a study.
Information bias: Errors in recording or classifying exposures and health outcomes.
Confounding: Distortion of an association by another factor related to both exposure and disease.
Random error: Variability in study results attributable to chance.
For example, smoking may be a confounding factor when researchers investigate an occupational exposure associated with lung disease.
Experts assess how well researchers addressed these potential sources of error and whether any remaining limitations materially affect the findings.
Epidemiological studies use statistical measures to describe the strength and uncertainty of associations. These measures help experts evaluate scientific findings but do not independently establish causation.
Relative risk, also called a risk ratio, compares the risk of an outcome in an exposed population with the risk in an unexposed population.
A relative risk of 1.0 indicates equal observed risk between the groups, while a relative risk of 1.5 indicates a 50% higher observed risk in the exposed group.
An odds ratio compares odds rather than risks and is commonly used in case-control studies. It can approximate relative risk when the outcome is rare, but the measures are not generally interchangeable.
Experts also consider absolute risk because a substantial relative increase may correspond to a small absolute increase when the underlying condition is uncommon.
Confidence intervals describe the precision of statistical estimates under the assumptions of the method used.
A wide confidence interval indicates greater statistical uncertainty, while a narrow interval generally suggests greater precision. Neither establishes that a study is free from bias or confounding.
Statistical significance is commonly assessed using a p-value, which measures how incompatible the observed data are with a specified null hypothesis under the assumptions of the statistical test. It does not establish whether an association is causal. A statistically nonsignificant result also does not establish the absence of an effect.
Experts should consider effect estimates, confidence intervals, statistical power, and methodological limitations rather than relying on statistical significance alone.
An epidemiological association does not necessarily establish that an exposure causes a disease. Causal interpretation requires evaluating alternative explanations and the broader scientific evidence.
Epidemiologists often draw on the considerations outlined in Austin Bradford Hill's 1965 publication, The Environment and Disease: Association or Causation?
Relevant considerations include:
Temporality: Whether exposure occurred before the health outcome.
Strength: The magnitude of the observed association.
Consistency: Whether similar findings appear across different studies and populations.
Dose-response relationship: Whether differences in exposure correspond to differences in disease risk.
Biological plausibility: Whether scientific knowledge supports the proposed relationship.
Coherence: Whether the proposed relationship is consistent with other relevant scientific evidence.
These considerations are not a mandatory checklist, and no fixed number establishes causation. However, temporality is necessary: an alleged causal exposure must precede the health effect it is claimed to cause. The relevance of the remaining considerations depends on the scientific question and the broader evidence.
When several epidemiological studies address the same exposure-disease relationship, an expert may compare their populations, methods, findings, and limitations.
Systematic reviews and meta-analyses may help synthesize research when the studies are sufficiently comparable. However, combining results does not automatically resolve weaknesses in the original studies.
Causal analysis may also incorporate toxicological research, biological mechanisms, exposure measurements, and other relevant scientific evidence.
Population-level research may support a general causal relationship without establishing that an exposure caused a particular individual's condition.
This distinction is particularly important when a disease has several recognized risk factors or can occur without the alleged exposure.
For example, a study may identify elevated disease risk among workers exposed to a chemical. Applying that finding to a plaintiff requires assessing whether the plaintiff's actual exposure and other relevant characteristics are sufficiently comparable to those studied.
Relevant case-specific materials may include:
Employment histories and occupational exposure records.
Environmental measurements and available exposure estimates.
Medical records, diagnostic findings, and disease chronology.
Exposure intensity, duration, and relevant latency periods.
Other medical, occupational, environmental, or lifestyle risk factors.
An epidemiology expert may evaluate whether the published studies reasonably apply to the individual's circumstances. Clinical specialists, toxicologists, occupational medicine physicians, or exposure assessment professionals may provide complementary analysis.
A population-level risk estimate does not automatically represent the probability that an exposure caused an individual's disease.
Some courts consider reliable epidemiological evidence showing a relative risk greater than 2.0 when evaluating the sufficiency of specific causation evidence. The Reference Manual on Scientific Evidence, Fourth Edition explains that courts differ in their treatment of this risk-doubling evidence.
Relevant considerations include the governing causation standard, whether the association is causal, the reliability of the risk estimate, the research's applicability to the individual, and competing risk factors. A relative risk greater than 2.0 is neither a universal legal requirement nor an automatic determination of individual causation.
Epidemiology experts may review studies reporting associations between medications or medical products and adverse outcomes. They may examine differences between patient populations and whether researchers accounted for relevant medical conditions and other risk factors.
In toxic tort and occupational disease disputes, experts may evaluate occupational cohorts, exposure intensity and duration, dose-response findings, and research relevant to alleged exposure-related health outcomes.
When environmental litigation involves allegations of elevated community disease rates linked to pollution or contamination, an expert may compare observed and expected disease patterns and assess potential exposures.
However, an apparent disease cluster does not independently establish a shared cause. In its guidance on unusual cancer patterns, the Centers for Disease Control and Prevention identifies limitations involving incomplete exposure information, competing risk factors, population characteristics, and random variation.
In infectious disease disputes, epidemiologists may examine infection patterns, transmission pathways, and surveillance data to assess whether the evidence is consistent with a proposed outbreak source or transmission scenario. Such analysis may not establish how every individual infection occurred.
Relevant expertise may include epidemiology, public health, biostatistics, occupational epidemiology, environmental epidemiology, or pharmacoepidemiology, depending on the scientific and legal questions involved.
Attorneys may consider an expert's experience with:
The study designs and statistical methods involved in the dispute.
The disease, population, or exposure being investigated.
Critical appraisal of scientific literature and causal inference.
Analysis of bias, confounding, and uncertainty.
Explaining technical findings and methodological limitations.
An expert's qualifications should correspond to the opinions offered. Depending on the case, related disciplines such as biostatistics, toxicology, and clinical medicine may be necessary to address questions outside the epidemiologist's expertise.
In federal proceedings governed by the Federal Rules of Evidence, Rule 702 establishes the admissibility requirements for epidemiological expert testimony. As amended effective December 1, 2023, the rule requires the party offering the testimony to show that those requirements are more likely than not satisfied.
These requirements address whether:
The expert is appropriately qualified and the testimony will assist the factfinder.
The opinion rests on sufficient facts or data.
The testimony uses reliable principles and methods.
The opinion reflects a reliable application of those methods to the case.
As clarified by the 2023 amendment, the court must assess whether the expert has sufficient facts or data and whether the methods were reliably applied to support the opinions offered. The court need not resolve every factual dispute or determine that the conclusions are ultimately correct. Once Rule 702's admissibility requirements are satisfied, challenges that do not implicate those requirements may affect the weight assigned to the testimony.
In Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579 (1993), the Supreme Court explained that federal trial judges must evaluate the reliability and relevance of scientific expert testimony.
The Court identified flexible considerations, including testing, peer review, potential error rates, and general acceptance, that may inform the evaluation of epidemiological opinions.
In General Electric Co. v. Joiner, 522 U.S. 136 (1997), the Supreme Court held that appellate courts review decisions admitting or excluding expert scientific testimony for abuse of discretion. Applying that standard, the Court upheld the exclusion of opinions linking the plaintiff's alleged PCB exposure to his cancer. The district court had reasonably found that the studies did not adequately support those opinions.
The decision illustrates why experts must explain the connection between their scientific evidence and conclusions, particularly when studies involve different populations, exposures, or outcomes.
State courts may apply different expert-evidence standards, while the substantive requirements for establishing causation depend on the law governing the claim. Counsel should distinguish the admissibility of epidemiological testimony from whether the evidence is sufficient to satisfy the applicable causation standard.
Is epidemiological evidence always required to establish causation in litigation?
No. Epidemiological evidence can be important in exposure-related litigation, but it is not universally required. Depending on the jurisdiction and circumstances, other reliable scientific or medical evidence may support a causation opinion.
Does a relative risk greater than 2.0 establish specific causation?
No. Some courts consider risk-doubling evidence when evaluating the sufficiency of individual causation evidence, but it is not a universal threshold. The association must be causal, the estimate reliable, and the research sufficiently applicable to the individual.
Can an epidemiology expert rely on studies that are not statistically significant?
Potentially, yes. Statistical significance is one consideration in evaluating scientific evidence. Experts must also assess statistical power, effect estimates, confidence intervals, methodological limitations, and the broader research record. Admissibility depends on the governing legal standard.
What happens when epidemiological studies produce conflicting findings?
A qualified expert may compare study designs, populations, exposure measurements, statistical methods, and potential sources of error. These differences may explain inconsistent findings and affect the weight assigned to individual studies.
Can epidemiological research involving different populations support an expert opinion?
Yes, when a reliable scientific basis supports applying the research to the disputed circumstances. Differences in exposure levels, population characteristics, and health outcomes may limit the conclusions that can reasonably be drawn.
Expert Witness Insights by Area of Expertise
Expert Witness Insights by Area of Expertise