Worker Health Chartbook 2004

National Electronic Injury Surveillance System (NEISS)

Case Definition

Cases are defined as follows:

• Medical: Nonfatal injuries and illnesses treated in an emergency department
• Population: Civilian, noninstitutionalized workers
• Work: Doing work for pay or other compensation (including arriving at or leaving work, but on the employer’s premises); being transported between locations as a part of the job (excluding commuting to or from home); engaging in agricultural production activities; or working as a volunteer for an organized group (e.g., volunteer fire department)
• Demographics: All workers without restriction by age, type of employer or industry (e.g., self-employed, private industry, or government), or employer size
  

• Injuries or illnesses to active duty military, national guard, and State militia
• Injuries or illnesses to institutionalized persons including prisoners or mental health patients
• Common illnesses (e.g., colds and flu)
• Routine drug and alcohol screening
• Revisits to the same emergency department for a previously treated injury or illness

For the purposes of this Chartbook, injuries and illnesses to workers under age 15 were excluded in all graphs.

Case Identification

Work-related cases are identified from admissions information and emergency department chart review by hospital coders. A workers’ compensation claim is not required for inclusion.

Injury and Illness Estimates

NEISS is designed to produce national estimates and is not suitable for regional, State, or local injury and illness estimates. A national estimate is obtained by extrapolating the number of cases seen in the approximately 67 hospitals by using the statistical weight of each case. The statistical weight varies depending on the size of the hospital and the number of patients typically treated in their emergency department. Each case captured in a sample hospital may represent 20 to more than 100 cases seen in other U.S. hospitals. By adding the weights for similar cases, a national estimate is obtained for a demographic group, type of injury, injury circumstances, or all injuries and illnesses. The public may access NEISS work-related injury and illness data through the NIOSH online Work-Related Injury Statistics Query System (Work-RISQS at www2a.cdc.gov/risqs). At Work-RISQS, national estimates are expressed in thousands of injuries and illnesses for a given year. For example in 1999, an estimated 3,930.9 thousand or 3,930,900 occupational injuries and illnesses were treated in U.S. hospital emergency departments for workers who were aged 15 and older. Other surveillance systems may produce different estimates.

Injury and Illness Estimate Confidence Bounds

Because the NEISS system is a sample of all U.S. hospitals and does not count every injury or illness treated in all hospitals, the national estimates may be in error. However, because the sample was statistically selected, we can calculate what the error due to sampling is likely to be. Each estimate produced for NEISS data through Work-RISQS has confidence bounds listed. The error estimates are based on the 95% confidence interval and are expressed as a value that should be subtracted from the estimate to get the lower confidence bound and added to it to get the upper bound. For example, in 1999, there were an estimated 3,930.9 thousand work-related injuries and illnesses with 95% confidence bounds of ±793.7 thousand cases for workers aged 15 and older. In other words, we expect the true value from this system is likely to be in the range of 3,137,200 to 4,724,600 injuries and illnesses. Simply put—about 3 to 5 million nonfatal occupational injuries and illnesses were treated in U.S. emergency departments in 1999.

The confidence bounds do not account for biases (nonsampling error) in the estimates that arise from the way in which data are collected or defined, ability to identify all work-related cases, or mistakes in data collection or coding.

The 95% confidence bounds are an approximation based on the classical formula for variance of a total from a stratified sample. These confidence bounds are an approximation of the general magnitude of error about an estimate and are not precise values.

Worker Population

The employment estimates used by NIOSH with NEISS data and Work-RISQS are derived from the U.S. Bureau of Labor Statistics Current Population Survey (CPS) of civilian noninstitutionalized workers. Employment estimates are calculated for Work-RISQS as full-time equivalent workers (FTE), where one FTE equals 2,000 hours worked per year. The FTE estimates account for hours worked in all jobs reported to CPS. At Work-RISQS, FTE estimates may be obtained by month, quarter, or year for workers aged 15 and older. FTE estimates may be obtained by age group or sex.

Volunteer workers are not included in CPS estimates, although they are included in the NEISS and Work-RISQS injury/illness counts. This creates a small but insignificant bias in rate estimates.

Approximate standard errors are calculated for CPS employment estimates by using parameters for selected demographic characteristics in a generalized variance function provided by the Bureau of Labor Statistics (see BLS [2001]). Parameters are independent of the time period selected within a given year, although they may change from year to year. In the calculation of error in the employment estimate, the selection of the variance parameters for the year are chosen based on the appropriate employment characteristics. When multiple characteristics apply for subsets of sex or age groups, the parameters are chosen to provide the most conservative error estimate (i.e., the largest error estimate). The errors generally decrease with longer time periods and when a selected sex or age group spans fewer than 20 years. However, the errors resulting from these calculations are only an approximation of the true sampling error and do not account for systematic biases in the data.

Injury and Illness Rate Estimates

In workplace safety, the injury/illness rate is commonly calculated as the frequency of injuries divided by the number of workers or the number of hours worked within a time period. Work-RISQS only provides rates as the number of injuries and illnesses per 200,000 hours worked (i.e., 100 FTEs) for time periods of a month, quarter, or year. A rate estimate query at Work-RISQS produces a table that includes the number of injuries and illnesses treated in an emergency department (the rate numerator), the FTE (the rate denominator), the rate per 100 FTEs, and the confidence bounds expressed as a value to be added and subtracted from the rate value.

Minimum Reporting Requirements

NEISS data reported through Work-RISQS have three criteria to determine whether a national estimate is reportable. First, estimates must be based on at least 40 cases among the small sample of U.S. hospitals within NEISS. This means that on average, a type of injury was treated at least once in about two-thirds of the 67 hospitals. Second, the national estimate must be greater than or equal to 2,500 injuries/illnesses. Third, the error associated with an estimate must be small enough to indicate that the estimate would be reasonably reproducible (i.e., the coefficient of variation is less than or equal to 33%).

Selected Data Elements

Sex

Definition: The distinction between male and female. A worker’s sex is indicated as male, female, or not stated. In any given year’s data, the number of cases with sex listed as not stated is very small (e.g., typically < 0.01% of all cases).

Age group

Definition: Inclusive range of ages of patients (in years). Injured or ill workers are grouped by age ranges (typically a 5-year range). One or more age groups may be selected. No age restrictions exist for cases captured in NEISS provided they meet the definition of a work-related case. However, injury rates are not available for youths under age 15 because data on hours worked are not available. Age is unknown for a small proportion of cases.

Race/Ethnicity

Definition: Construct for classifying people with similar biological, social, and cultural heritage into four race groups (White, Black, American Indian/Alaska Native, and Asian/Pacific Islander) and one ethnicity group (Hispanic or Latino) as recorded on the emergency department chart.

Within the original NEISS data, race and ethnicity of an injured worker are collected as a two-part question. In the first part, race is coded as white, black, other, or not stated. In the second part which is free text, other races may be described such as Asian or American Indian, or ethnicity may be entered such as Hispanic or Latino. Because NEISS hospitals do not collect race and ethnicity uniformly (if they collect or report it at all) and because of the two-part nature of the NEISS race/ethnicity question, consistent reporting of race and ethnicity is not always achieved. For example, a Hispanic patient might be indicated as (1) white or black with the free text blank (only race reported); (2) white or black with Hispanic entered into the free text (both race and ethnicity reported); and (3) other, with Hispanic in the free text field (only ethnicity reported).

Within Work-RISQS, race and ethnicity were recoded under a single combined data structure where Hispanic ethnicity was indicated preferentially as opposed to white or black race. Thus, for the purposes of estimating injuries and illnesses, workers who were identified as white or black and Hispanic or Latino were counted under their Hispanic ethnicity. If originally race was listed as other, and the free text contained a country of origin or ancestry, those cases were recoded to the race or ethnicity category predominant for the region (e.g., Chinese was recoded as Asian and Mexican was recoded as Hispanic). Race was recoded as not stated if the free text listed Indian without an indication that the patient was American Indian or Native American. However, Indian was recoded to American Indian if the NEISS hospital where the case was seen generally treated a large population of American Indians.

Note: Because of these and other problems in collecting race and ethnicity information and missing information for approximately 20% of cases, Work-RISQS estimates for specific race or Hispanic worker populations should be used with caution.

Data Objectives and Limitations

The primary objective of NEISS data collection and the online Work-RISQS is to provide public access to occupational injury and illness data for use in workplace safety and injury prevention activities for which understanding the general magnitude of injuries, worker characteristics, and injury events are important (e.g., estimating the number of falls from elevation). Use of these data for other purposes should be done with caution. NEISS is not suitable for examining rare events or events involving very detailed specification of the incident characteristics, workplace setting, personal protective equipment, or tools, machinery, or objects that were associated with the injury (e.g., estimating the number of falls from Type IA aluminum extension ladders made by XYZ Manufacturing or the number injured workers wearing fall protection).

Worker Health Chartbook 2004

Chapter Index: