Home: Practices: Statistical & Data Sciences: Risk Analysis

Risk Analysis

All living humans continually bear a certain degree of risk of injury or death. Exponent’s Statistical & Data Sciences practice specializes in determining whether a particular activity or product poses an unreasonable risk. We maintain one of the largest in-house collections of accident and incident data in the country, which we use to quantify how machines, vehicles, consumer products, and components behave in their real-world environments.

Judging whether the risk of an activity or product is “unreasonable” is typically not as straightforward as merely estimating its risk. In some instances, there may be an absolute standard or risk threshold that cannot be exceeded. When no standard exists, the approach developed at Exponent—and subsequently adopted by such federal agencies as the U.S. Consumer Product Safety Commission—involves comparing the risk of the activity or product in question to other activities or products judged to be similar in key respects.

Information on the frequency of adverse events is usually obtained from observational data collected in reporting systems maintained by government or private sources. Exposure is usually determined from administrative records or sample surveys. Because most risk analyses are conducted using observational studies rather than controlled experiments, care must be taken in making comparative risk judgments. Specifically, before declaring a difference in risk to be significant, one should consider whether any factors other than the factor of primary interest could also have influenced (or “confounded”) the outcome and, if so, adjust for such an effect. This consideration has stimulated the development of increasingly sophisticated statistical approaches to risk analysis.

Exponent developed graphical methods of expressing the ideas of relative risk and the concept of rare event. These methods help to put risk into perspective.

Risk of fatality associated with an activity or a consumer product may be difficult to understand in the abstract. Comparison with the fatality risk associated with familiar products and activities allows the reader to put the risk into context.

Risk estimation involves establishing a reference period and then collecting information about the number of injuries (or other adverse events) suffered and the amount of exposure during this period. Because risk is expressed as a rate, measures of both frequency (numerator) and exposure (denominator) are required.

We apply our expertise in risk analysis in a number of areas including:

Motor Vehicles
Consumer Products
Fires
Medical Devices
Health & Environment
Engineering Reliability
Economics
Law

Motor Vehicles

For most people in developed countries, the greatest risk assumed on a daily basis is travel by motor vehicle. Injury data on persons involved in U.S. motor vehicle accidents are collected by several systems operated by the National Center for Statistics and Analysis (NCSA), which is part of the National Highway Traffic Safety Administration (NHTSA), an agency of the Department of Transportation. The Fatality Analysis Reporting System contains data on all traffic crashes within the 50 states, District of Columbia, and Puerto Rico that result in fatality within 30 days to a vehicle occupant or non-motorist from injuries suffered in the crash. These data can be used to answer many questions concerning the safety of vehicles, drivers, traffic situations, and roadways.

The NCSA also operates the National Automotive Sampling System, which has two components: the Crashworthiness Data System (CDS) and the General Estimates System (GES). The CDS contains detailed data on a representative, random sample of crashes severe enough for the vehicle to have been towed from the scene. Crashes with all levels of injury as well as those resulting in property damage only are included. Data are collected by 24 field research teams that each year investigate about 5,000 crashes involving passenger cars, light trucks, vans, and utility vehicles. This information is used to document crash performance, evaluate vehicle safety systems and designs, and examine relationships between the type and severity of a crash and resultant injuries.

Data for the GES come from a nationally representative sample of police-reported motor vehicle crashes of all types and severities. GES data collectors make weekly visits to about 400 police jurisdictions in 60 geographic areas across the U.S., where they randomly sample 50,000 reports each year. The system was created to identify traffic safety problem areas, provide a basis for regulatory and consumer initiatives, and form a foundation for cost-benefit analyses of proposals for improving traffic safety.

State governments also maintain traffic accident data. These systems vary considerably in structure and content, but they provide substantial additional information on traffic safety and support more localized risk analyses. Exponent maintains the largest collection of state motor vehicle accident data in the country. Additional relevant event data are also occasionally made available by other public or private sources.

Data on exposure, the other factor determining risk, are typically more difficult to obtain. Counts of registered vehicles are available from the National Vehicle Population Profile compiled annually by the Polk Company. These data files contain information derived from the vehicle identification number concerning make, model, model year, body style, and engine size, as well as the geographic location of registrants.

More detailed characterizations of exposure—accounting for such variables as vehicle mileage, driver demographics, and environmental factors—might be desirable for many applications of risk analysis, but often cannot be supported by available data. Yet one potential source for some applications is the National Household Travel Survey (NHTS), sponsored by NHTSA, the Bureau of Transportation Statistics, and the Federal Highway Administration. NHTS data are collected from a national sample of about 26,000 U.S. households and are used to provide national estimates of trips and miles by travel mode, trip purpose, and a host of household attributes. NHTS records also contain information on the make and model of each vehicle in a responding household.

Consumer Products

Risk analysts at Exponent have performed studies for a wide range of consumer products—from dishwashers to toilet bowl cleaners. The necessary inputs to such analyses are counts of adverse events (e.g., injuries or fatalities) associated with a product and a measure of the population’s collective exposure to that product. The most widely used source of U.S. data on product-related injuries is the National Electronic Injury Surveillance System (NEISS). Operated by the U.S. Consumer Product Safety Commission, NEISS is a national probability sample of hospitals from which patient information is collected for every emergency visit involving an injury associated with consumer products. From this sample, which includes more than 375,000 cases per year, the total number of product-related injuries treated in hospital emergency rooms nationwide can be estimated.

To assess risk however, one must also gather exposure data, the source of which will depend on the application. For example, Exponent investigations of all-terrain vehicles have used exposure data from both industry- and government-sponsored surveys. We have conducted comparative risk analyses of other recreational products using information collected by the National Sporting Goods Association, as well as from a supplement to the National Health Interview Survey. In a study of burn risk from home appliances, we obtained data on the number of U.S. households with these appliances from the Energy Information Administration’s Residential Energy Consumption Survey.

Fires

Risk analysis can also be used to complement engineering and scientific studies of the causes of fires and the combustion process. Data on the incidence of fires are provided by the National Fire Incident Reporting System, which is maintained by the U.S. Fire Administration’s Office of Fire Data and Analysis. Over 15,000 fire departments from all 50 states and more than 40 major metropolitan areas have voluntarily submitted data to the system, collected on a uniform set of key elements using standardized terminology. More than 600,000 incidents are reported each year, constituting an estimated 44 percent of all U.S. fires to which fire departments respond. This information can be used to improve understanding of the relative contribution to fire risk from various sources, as well as the geographic distribution of risk.

Related Case Study/Studies:
Fire Risk, Property Loss and Insurability – Cost/Benefit Analysis

Related Links:
Fires & Explosions

Medical Devices

Manufacturers of medical devices are obligated under U.S. regulations to collect information on the performance of certain critical products after they have been approved for use. Several large federal surveys and administrative record systems produce data that can support post-market surveillance. The Nationwide Inpatient Sample (NIS) is the largest all-payer inpatient care database in the United States, containing records from approximately 7 million hospital stays. Sponsored by the Agency for Healthcare Research and Quality, the NIS contains all discharge data from about 1,000 hospitals located in 35 states, approximating a 20 percent stratified sample of U.S. community hospitals. Other relevant sources include the National Hospital Discharge Survey conducted by the National Center for Health Statistics. Information contained in these systems can help to answer questions about the comparative costs and benefits of alternative treatments for various medical conditions. Examination of such databases can also support market studies early in the product development process by identifying and characterizing profitable opportunities to improve health outcomes.

Exponent’s Statistical & Data Sciences practice recently teamed with other researchers at Exponent to study the experiences of patients undergoing hip and knee replacement surgeries. By analyzing public-use Medicare records, obtained from the Centers for Medicare and Medicaid Services, we were able to estimate the proportion of all such surgeries that are secondary, or “revision,” in nature and to judge whether this proportion (“revision burden”) has changed significantly over time. We have also used these data sources to assess the economic impact of vertebral compression fractures, a prominent risk to continuing quality of life for elderly Americans.

Related Case Study/Studies:
Quality Assurance – Medical Syringes

Health & Environment

Exponent statisticians, programmers, and analysts support epidemiological and environmental studies of various potential health and environmental risks. We recently assisted the company’s health scientists in assessing the strength of evidence allegedly linking perchlorate to thyroid function and development in infants. We have collaborated with hydrogeologists in designing water quality monitoring systems and with ecologists in quantifying uncertainties associated with the measurement of ecosystem performance.

Engineering Reliability

Risk analysis and related methodologies find application beyond studies concerned with human injury or death. In engineering reliability studies, we apply similar techniques in modeling the lifetimes of systems and components. One recent investigation of alleged construction defects involved assessing the susceptibility of piping in a large residential development to corrosion-related failure using historical data from the site.

System Reliability and Failure Probability
Statistical methods can be applied to assist in understanding system reliability and improving production quality. Product failure probability is often a function of product life or age. Statistical modeling of product failure probability can often be used to compare the reliability of different product designs or production process and to forecast future failures.

Product Safety Assessment
Exponent consults with product safety engineers to estimate the risk of accident and injury associated with product design. Patterns of complaints and returns of product are used to model the failure process, to estimate current and future risk and to provide context to the decision to recall or not recall a product.

Related Case Study/Studies:
Product Safety Assessment – Recall Scope

Related Links:
Risk Management & Reliability

Economics

Statistical methods can also be applied to assist in understanding and managing financial risks. Economic damage assessments, for example, are commonly associated with class action lawsuits. In such cases, Exponent statisticians have collected and analyzed sample data to make defensible extrapolations to larger populations of interest.

Related Case Study/Studies:
Economic Damage Assessment – Worker’s Compensation Claims

Law

The Statistical & Data Sciences practice has capabilities and project experience in related applications of statistics to the law in such areas as antitrust, employment discrimination, and forensic science.

Intellectual Property and Copyright Infringement
Statistical sampling and data analysis are useful tools in researching intellectual property issues and in resolving disputes concerning allegations of copyright infringement, deceptive advertising, and trademark dilution. Evidence may be collected from focus groups or from market research surveys taken at malls and other shopping areas. Typically, the goals in such studies are to identify potential consumers or purchasers of the product, to ascertain their beliefs, and to determine how the trademark or advertising influenced those beliefs.

Related Case Study/Studies:
Intellectual Property – Dolls

Labor and Employment Statistics
Statistical issues arise in the context of fair labor practices, EEOC and affirmative action and in manpower planning. Analyses may be based upon labor market information from published sources or upon the employer’s internal statistics.

Related Case Study/Studies:
Labor and Employment – Work Standards and Disciplinary Rules