Funding
Provided
in part by NIOSH
Grant Number:
1 R01 0H03915
Interactive Monograph of Occupation, Disability, and Self Reported Health: The National Health Interview Survey (1986-1994)
Lora
E Fleming MD PhD, Terry Pitman BA, William LeBlanc PhD, Alberto Caban
Jr MPH,
Orlando Gómez-Marín PhD MSc,
David Lee PhD, Diane Zheng MS, Dulce Jane MA
Study Website: http://www.UMiamiORG.com
Department of Epidemiology and Public Health
University
of Miami School of Medicine
Miami, FL
33136
August 2004
The National Health Interview Survey (NHIS) is a multipurpose household survey of the US civilian non-institutionalized population conducted annually since 1957. From 1986-1994, over 450,000 US workers, age 18 years and older, participated in a probability sampling of the entire non-institutionalized US population; variables collected included a range of measures of acute and chronic disability.
The objective of this Monograph was to establish and apply a methodology to assess predictors of health status, and acute and chronic disability for US workers by occupation using the 1986-1994 NHIS data. After adjustment for sample weights and design effects using SUDAAN, the prevalence (with their standard errors) of several measures of acute and chronic disability and health status were created in tabular format. These prevalence rates have been presented by industry and occupational subgroups, as well as by gender, race, and ethnicity. Additional information has been made available to allow for extrapolation to the US worker population during the 1986-1994 time period. The Study Website (http://www.UMiamiORG.com) contains a repository of interactive tables which are available in Excel and PDF formats; additional study and NHIS documentation are also available at this Website.
Understanding the occupational risk factors and improving the health of US workforce remains paramount to the public health profession. The surveillance of occupational groups using the NHIS dataset allows for the careful monitoring of these occupational risk factors in order to prevent and minimize disability in the workplace.
Acute Disability, Chronic Disability, Health Status, Occupation, Industry, National Health Interview Survey (NHIS), Self Reported Health, Surveillance
The data utilized in this publication were made available in part by the Inter-University Consortium for Political and Social Research. The data for the National Health Interview Survey (NHIS) were originally collected and prepared by the US Dept of Health and Human Services and the National Center for Health Statistics. Neither the collector of the original data nor the Consortium bears any responsibility for the analyses or interpretations presented in this publication. This study was funded in part through the National Institute of Occupational Safety and Health (NIOSH) Grant number 1 R01 0H03915-01. Additional information on this study can be found at the Study Website located at: http://www.UmiamiORG.com.
It is well recognized that a variety of occupational and environmental risk factors interact to determine the overall health and well being of the US workforce. Occupational health surveillance is poised to analyze these interactions by systematically collecting, analyzing, and interpreting health data essential to the planning, implementation and evaluation of public health strategies to maximize workforce health. These robust surveillance systems collect and maintain health outcome information for all members of a temporally and geographically defined population at risk. The NIOSH NHIS Research Group at the University of Miami is currently funded by the National Institute for Occupational Safety and Health (NIOSH) to study the National Health Interview Survey (NHIS) dataset collected and conducted by the National Center for Health Statistics (NCHS) at the Centers for Disease Control (CDC).
The National Health Interview Survey (NHIS) is a continuous multipurpose and multistage probability area survey of the US civilian non-institutionalized population living at addressed dwellings (Kaminski 1980, Botman 1995, NCHS 2000). Each week a probability sample of households is interviewed by trained personnel to obtain information about the characteristics of each member of the household (Liao 1998). Data from the NHIS included a range of measures of acute and chronic disability collected for all participants. Recently, the NHIS conducted a Mortality Follow Up with cause of death through 1997. The NHIS database allows for longitudinal analysis of mortality data as a retrospective cohort study, as well as for cross-sectional and trend analysis of the aggregate morbidity data. Thus, the NHIS database represents a unique opportunity to explore new research hypotheses, and to use the data as a surveillance tool to evaluate time trends and occupational disease in the US for the past 2 decades in both genders and in a variety of race-ethnic subpopulations.
This Disability Monograph establishes and applies a methodology to assess predictors of health status, and acute and chronic disability for US workers by occupation using the 1986-1994 NHIS data. After adjustment for sample weights and design effects, the prevalence of several measures of acute and chronic disability and health status were created in tabular format. These prevalence rates have been presented by industry and occupation, as well as by gender, race, and ethnicity. Additional information has been made available to allow for extrapolation to the entire US worker population during the 1986-1994 time period.
The European countries, particularly England since 1837 in their Registrar General’s Decennial Supplements for England & Wales, have had a long and illustrious history of performing nationwide occupational studies (Boffetta 2999, Drever 1995). As noted in the 1995 Registrar General’s Report (Drever 1995), these data have provided a valuable means of generating hypotheses about work-related risks to health as well as insight in the effectiveness of preventive measures. The United States has had relatively few studies of equal scope and caliber to evaluate the causes of morbidity and mortality, and their trends, in US workers (Kaminski 1980, Milham 1976, Milham 1983, NIOSH 1993, Guralnick 1962, Guralnick 1963a, Guralnick 1963b, NIOSH 1997, Murphy 1996, Drever 1995, Gallagher 1989, Milham 1983). As described below, the majority of these studies have focused on special subsets of data, not truly representative national data, and they have focused on mortality rather than morbidity.
Similar to the English studies, Guralnick (1950, 1963a, 1963b) used all US death certificates to look at age-race standardized mortality ratios (SMRs) for white and black men aged 20-64 by industry and by occupation. She evaluated trends by age, and compared them to the US Census data. Acknowledging several study limitations, including what would now be called the healthy worker effect, and lack of information on important confounders such as socio-economic class. No female or other race-ethnic worker subpopulations were examined. Another major limitation was the use of death certificate-derived occupation/industry status; this has been shown to be highly inaccurate when compared to lifetime occupational histories, with misclassification estimated between 30-50% in US studies (Monson 1990, Checkoway 1989, Guralnick 1962, Guralnick 1963a, Guralnick 1963b, Kircher 1985, Schumacher 1986, Swanson 1984, Alderson 1972, Schade 1988).
Milham (1976, 1983) used data from Washington State to evaluate causes of mortality by occupation and industry using proportionate mortality ratios (PMRs) based on a large number of death certificates from 1950-1979. Limitations of these studies included no available confounders, the lack of rates (i.e. only proportions reported), socio-economic class, the lack of other race-ethnic worker subpopulations, and the use of death certificate occupation/industry status.
The National Traumatic Occupational Fatalities (NTOF) surveillance system estimates the risk of work-related fatal injuries for 50 industries and 50 occupations having the highest risks. NTOF supplies cause-specific risk estimates; it does not provide estimates for causes of death other than injury (Fosbroke 1997, Loomis 1998, Stout 1996). Similar analyses have been made using insurance databases (Toscano 1996).The National Institute of Occupational Safety and Health (NIOSH)(1993) used the death certificates with injury cause of death to evaluate fatal occupational injury rates from 1980-89 among male and female race-ethnic US worker populations. In addition to cause specific information, occupational injury rates (blacks>whites; males>females) and years of productive life lost were evaluated. Limitations to this study included no available confounders (such as socio-economic class), and the use of death certificate occupation/industry status.
Wagener et al (1997) attempted to use data from many available national databases (i.e., NHIS, NHANES, National Occupational Mortality Surveillance System, and NTOF, as well as the National Maternal and Infant Health Survey, National Hospital Ambulatory Medical Care Survey, the Census of Fatal Occupational Injuries, the Annual Survey of Occupational Injuries, and the Current Population Survey) to examine the health of women according to workforce and job conditions. In addition to presenting an overview of numbers and frequencies by age, race, ethnicity and educational level of various occupations, industries and presumed occupational exposures, for morbidity and mortality analyses, the risks were calculated as the observed over expected proportions, not rates or trends over time.
Reviere et al (1995) used the National Mortality Follow Up Survey of 1986 to identify Sentinel Health Events Occupational (SHE(O)s) to look at causes of death beyond injury for the US worker. The National Cancer Institute (NCI), NIOSH and others have developed a database known as the National Occupational Mortality Surveillance System based on the death certificate listed occupation to study cancer and other causes of mortality in a variety of occupations (Ma 1998, Alterman 1997). NIOSH (1993, 1997) also evaluated mortality by occupation, industry, state and cause of death as proportionate mortality ratios from 24 reporting States from 1984-1988 among male and female black and white US worker populations. In addition to trends, years of productive life lost were estimated. The results may be evaluated by researchers and used as leads for further studies and to confirm previously identified associations. Researchers may identify new occupations and industries not previously recognized as experiencing an excess risk for a known occupational disease. Finally, these databases may be used to prioritize health promotion and intervention activities to the appropriate workers for both occupational and non occupational diseases (Kaminski 1980, Drever 1995). Limitations to these studies included no available confounders, the lack of rates and the use of only selected states’ data, the lack of other race-ethnic worker subpopulations, and the use of death certificate occupation/industry.
Although laudable, all of these attempts have been biased by selective reporting or by the use of occupation at time of death as the definition of occupational exposure, or have focused purely on traumatic injury, or are not generalizable to the entire US workforce due to sampling issues. Furthermore, as noted in the 1995 Registrar General’s Report (Drever 1995), mortality data alone cannot describe the nature and scale of all occupational diseases since many of them are non-fatal.
Previous NHIS Occupational Morbidity Studies
Previous studies have used the NHIS data to explore a range of occupational issues, including: injury, smoking characteristics, health characteristics in the longest held occupation and industry, injuries in racial subgroups, cardiovascular disease and working women, impairments and chronic diseases in farmers, back injury and disability, AIDS knowledge among healthcare workers, and carpal tunnel (Kaminski 1980, la Rosa 1988, Wagener DK 1991, Zwerling 1997, Brackbill 1988, Nelson 1994, Sterling 1990, Sterling 1989, Cooper 1993, Brackbill 1994, Hurwitz 1997, Behrens 1994, Biddlecom 1992, Tanaka 1995, Zwerling 1998).
Kaminski and Spirtas (1980) analyzed data as Proportionate Morbidity Ratios (PMRs) from the 1969-74 NHIS surveys to examine the morbidity, disability, and reported healthcare use patterns for 498,580 individuals by industry. They did not look at trends over this relatively short time period. The highest specific disease conditions were reported for agriculture, furniture manufacturing, metal fabrication, railroad transport, repair services, amusement and recreational services, state and local government workers, and new workers; the highest disabilities were found for forestry and fisheries workers, certain manufacturers, medical and health services workers, and federal government employees; the greatest use of medical services was among metal industries, specific manufacturers, and railroad workers; the greatest morbidity was reported by private household service workers although they had less disability and use of medical services; overall manufacturing industries had the largest proportion of workers with work injuries and the service industries had the smallest. The authors pointed out that although some of the results confirmed previous studies, other results of their study revealed new associations of morbidity with particular industry/occupations of US workers. These new associations were possible because the NHIS data are not limited to a particular industry, occupation, or geographic area. Therefore, Kaminski and Spirtas (1980) suggested that NHIS data can be used as a surveillance system for occupational disease morbidity and mortality for US workers, and recommended that its use for this purpose be explored further.
The NHIS dataset is anonymous and publicly available through the National Center for Health Statistics (NCHS); although there are no human subject considerations in the use of this dataset per the NCHS, an official waiver was obtained from the University of Miami School of Medicine Human Subjects Committee. During the 1986-1994 study period, annual NHIS survey response rates reportedly ranged from 95-98% (Massey 1989). In the majority of cases (63%), the participants themselves answered all the questions, and for the remaining participants, the responses were obtained from their relatives or other proxies. For simplicity, in the present study either self-reported or proxy-reported data are referred to as “reported.”
Information on employment during the two weeks prior to the interview was collected for all persons 18 years or older in order to determine the person's employment status; as utilized by other investigators, all subjects age 18 years and older who had worked or reported having jobs, both paid and unpaid, during the two weeks prior to the NHIS survey were considered currently employed (Kaminski 1980, Zwerling 1997, Brackbill 1998). Standardized Occupational Codes (really US Census codes) were provided in the NHIS database (NCHS, 1989) as well as various NHIS recodes with less detailed grouped occupational codes. For the purpose of this study, only the 206 occupations employing at least 100,000 workers annually in the US during the 1986-94 study period were selected for detailed analyses.
In Table 1, the number of workers who participated in the NHIS survey pooled over the 1986-94 period (“sample N”) and the estimated number of workers these participants represented in the US worker population during this time period ("population N") are presented by gender, race and ethnicity subpopulations. Of note, tables for the 13, 41, and 206 occupational groups are available at the Study Website (www.UMiamiORG.com).
In the NHIS, a “chronic condition” was recorded if the respondent had a health problem that was detected at least 3 months before the interview, or it was a condition that would normally last at least three months as shown in Table 2; an “acute condition” was defined as an illness or injury lasting less than 3 months and involving either seeking medical attention or two or more days of restricted activity. Information for the past 2 weeks and for the past 12 months was collected on: restricted activity days, bed days, doctor visits, and days of short hospital stays; study-generated variables of the number of days, visits, and hospital episodes were also included. Restricted activity days were defined as those days during which the respondent reduced usual activities for all or most of the day due to illness; bed-disability days were counted in the restricted activity days as well as separately recorded as days during which all or most of the day was spent in bed. For each of these variables, acute and chronic conditions were coded using the ICD 9th Edition. Health status indicators included a general question rating health status as: excellent, very good, good, fair, and poor.
In the NHIS, disability was defined as any reported temporary or long-term reduction of a person’s activity as a result of an acute or chronic condition. Restricted activity, bed disability and work loss all denote varying degrees of disability and are not statistically independent. As Kaminski and Spirtas (1980) noted using earlier NHIS data for occupational disability evaluation, a day of bed disability or work loss would also be considered a day of restricted activity, but the converse is not necessarily true. For example, a person’s activity may be restricted in the sense that s/he may not be able to do heavy lifting, but this may not necessarily keep her/him in bed or away from work.
For the present analyses, measures of acute and chronic health and disability status were defined by variables adapted by the Investigators (Table 2). Acute disability was defined in terms of restricted bed days (0 vs. ≥ 1 restricted activity days in the prior 2 weeks), bed days (0 vs. ≥ 1 bed days in the prior 2 weeks) and lost work days (0 vs. ≥ 1 lost work days in the prior 2 weeks); chronic disability was defined as doctor visits (0-3 vs. ≥ 4 doctor visits in the prior 12 months) and hospitalizations (0 vs. ≥ 1 days short-term stay hospitalizations in the prior 12 months); and health status was defined as self rated health status (0 if health self-rated as excellent, very good, or good vs. 1 if health self-rated fair or poor) and report of one or more health conditions (0 vs. ≥ 1 health condition reported). For each of the above health and disability variables, analyses were performed by occupation and other factors available (i.e. gender, race, and ethnicity).
Statistical Methods
Because of the multi-stage sampling design, all analyses were performed with adjustment for sample weights and design effects using the SUDAAN statistical package (RTI, 2001). The sample weights used were those required for the analysis of data from combined survey years and were calculated as specified by Botman and Jack (1995).
Morbidity data is presented in tabular format for each of the 3 levels of major US occupational groupings (i.e. 13, 41, and 206). For each of these occupational groups, there is an initial table showing the overall estimated prevalence (and its corresponding standard error) for each of the disability and health measures. Subsequent tables present data on the estimated prevalence of those participants not reporting a particular disability or health measure, followed by the average number of reports of the particular disability measure among those participants that did report the particular disability or health measure within the specific time period.
These unique data tables have been made available in two contemporary file formats: Portable Document Format (PDF) and Excel Spread Sheets (Excel) at the Study Website (URL:http://www.UMiamiORG.com). The PDF format allows researchers to quickly view and print the current table layouts, while the Excel format can be utilized to download the files to a remote computer and manipulate the data table locally.
To save an Excel file to a local computer system, the Study Website user can right click with their mouse over a link, at which time a dialog box will appear. Select the “Save target as” option to save the file from the Study website to the local system. A file viewer for PDF files is readily and freely available at the following website http://www.adobe.com/products/acrobat/readstep2.html. Researchers can utilize these additional data tables to further explore disability and health report among this population based sample and extrapolate to the general US workforce. As discussed above, the Standard Errors (SEs) presented in the Tables for each disability and health measure are presented. These SEs can be used to generate estimates of national level numbers as well as to generate confidence intervals. For example, for a particular disability measure, the reader can take (1.96 x SE)±Prevalence Rate to generate the estimated range of that particular disability measure among US workers.
The results presented below summarize the disability measures and health indicators by describing the highest and lowest of the particular measure for each of the occupational groupings (i.e. 13, 41, and 206). All the tables only present the data for the 13 occupations, however more detailed tables which include standard errors, for the 41 and 206 occupations are available at the Study Websites and Appendices as described above; all tables report the data by gender, race and ethnicity as well as total populations.
Restricted Days
13 Occupations : In Table 3, the overall prevalence of the acute disability measure, restricted days, by the different occupations is reported by gender, race and ethnicity as well as the estimated US population numbers. Among the 13 occupations, service workers (not including protective or household) reported the highest prevalence of ≥ 1 restricted days in the past 2 weeks (10.14±0.17) while farming, forestry and fishing workers reported the lowest (6.52±0.19). Among the gender-race-ethnicity subgroups, female protective services workers reported the highest prevalence of ≥ 1 restricted days in the past 2 weeks (11.93±1.17) while Hispanic farming, forestry and fishing workers reported the lowest (3.81±0.51).
In Table 4, the average number of restricted days (≥ 1) in the past 2 weeks by different occupations is reported, as well as the prevalence of participants not reporting any restricted days in the past 2 weeks. Among the 13 occupations, transportation and material moving workers reported the highest average number of restricted days (5.08±0.13) and professional specialty workers reported the lowest average number (4.05±0.05). Among the gender-race-ethnicity subgroups, farming, forestry and fishing workers in the ‘other races’ category reported the highest average number of restricted days (8.96±1.37) and private household workers in the ‘other races’ category reported the lowest average number (1.87±0.46).
41 Occupations : In Table 3, the overall prevalence of the acute disability measure, restricted days, by the different occupations is reported by gender, race and ethnicity as well as the estimated US population numbers. Among the 41 occupations, health services workers reported the highest prevalence of ≥ 1 restricted days in the past 2 weeks (13.15±0.41) while health diagnosing workers reported the lowest (5.09±0.45). Among the gender-race-ethnicity subgroups, Architects and surveyors in the ‘other races’ category reported the highest prevalence of ≥ 1 restricted days in the past 2 weeks (21.55±8.70) while female other transportation (except motor vehicles workers) reported the lowest (3.60±3.55).
In Table 4, the average number of restricted days (≥ 1) in the past 2 weeks by different occupations is reported, as well as the prevalence of participants not reporting any restricted days in the past 2 weeks. Among the 41 occupations, other transportation except motor vehicles workers reported the highest average number of restricted days (7.58±0.71) and natural mathematical/computer scientists reported the lowest average number (3.50±0.18). Among the gender-race-ethnicity subgroups, Hispanic forestry and fishing workers reported the highest average number of restricted days (14.00±0.0) and other transportation workers (excluding motor vehicles) in the ‘other races’ category reported no restricted days.
206 Occupations : In Table 3, the overall prevalence of the acute disability measure, restricted days, by the different occupations is reported by gender, race and ethnicity as well as the estimated US population numbers. Among the 206 occupations, nursing aides, orderlies, and attendants reported the highest prevalence of ≥ 1 restricted days in the past 2 weeks (13.99±0.52) while dentists reported the lowest (3.11±0.70). Among the gender-race-ethnicity subgroups, other not specified mechanics and repairers reported the highest prevalence of ≥ 1 restricted days in the past 2 weeks (29.65±13.34) while forty-four occupations reported no restricted days.
In Table 4, the average number of restricted days (≥ 1) in the past 2 weeks by different occupations is reported, as well as the prevalence of participants not reporting any restricted days in the past 2 weeks. Among the 206 occupations, managers and farmers except horticultural workers reported the highest average number of restricted days (6.99±0.73) and sales workers, radio, TV, hi-fi and appliances workers reported the lowest average number (2.62±0.49). Among the gender-race-ethnicity subgroups, black sales workers, motor vehicles, and boats workers reported the highest average number of restricted days (14.00±0.0) and forty-four occupations reported no restricted days.
Bed Days
In Table 3, the overall prevalence of acute disability measure, bed days, by the different occupations is reported by gender, race and ethnicity as well as the estimated US population numbers. Among the 13 occupations, service (not including protective or household) workers reported the highest prevalence of ≥ 1 bed days in the past 2 weeks (5.91±0.13) while farming, forestry and fishing workers reported the lowest (3.20±0.17). Among the gender-race-ethnicity subgroups, private household workers in the ‘other races’ category reported the highest prevalence of ≥ 1 bed days in the past 2 weeks (7.26±2.41) while Hispanic farming, forestry and fishing workers reported the lowest (1.80±0.46).
In Table 5, the average number of bed days (≥ 1) in the past 2 weeks by different occupations is reported, as well as the prevalence of participants not reporting any bed days in the past 2 weeks. Among the 13 occupations, farming, forestry, fishing workers reported the highest average number of bed days (3.84±0.23) and technicians/related support workers reported the lowest average number (2.48±0.09). Among the gender-race-ethnicity subgroups, farming, forestry and fishing workers in the ‘other races’ category reported the highest average number of bed days (7.46±1.58) and private household workers in the ‘other races’ category reported the lowest average number (1.15±0.15).
41 Occupations : In Table 3, the overall prevalence of the acute disability measure, bed days, by the different occupations is reported by gender, race and ethnicity as well as the estimated US population numbers. Among the 41 occupations, health service workers reported the highest prevalence of ≥ 1 bed days in the past 2 weeks (7.61±0.33) while health diagnosing workers reported the lowest (2.27±0.31). Among the gender-race-ethnicity subgroups, architects and surveyors in the ‘other races’ category reported the highest prevalence of ≥ 1 bed days in the past 2 weeks (15.57±8.17) while Hispanic forestry and fishing workers, female and other transportation (except motor vehicles) workers in the ‘other races’ category reported no bed days.
In Table 5, the average number of bed days (≥ 1) in the past 2 weeks by different occupations is reported, as well as the prevalence of participants not reporting any bed days in the past 2 weeks. Among the 41 occupations, farm operators and managers reported the highest average number of bed days (4.40±0.37) and natural mathematical/computer scientists reported the lowest average number (2.09±0.12). Among the gender-race-ethnicity subgroups, architects and surveyors in the ‘other races’ category reported the highest average number of bed days (14.00±0.0) and forestry and fishing workers, female and other transportation (except motor vehicles) workers in the ‘other races’ category reported no bed days.
206 Occupations : In Table 3, the overall prevalence of the acute disability measure, bed days, by the different occupations is reported by gender, race and ethnicity as well as the estimated US population numbers. Among the 206 occupations, bill and account collectors reported the highest prevalence of ≥ 1 bed days in the past 2 weeks (8.66±2.36) while dentists reported the lowest (1.43±0.49). Among the gender-race-ethnicity subgroups, mechanics and repairers in the ‘other races’ category reported the highest prevalence of ≥ 1 bed days in the past 2 weeks (29.65±10.31) while workers in one hundred different occupational groups reported no bed days.
In Table 5, the average number of bed days (≥ 1) in the past 2 weeks by different occupations is reported, as well as the prevalence of participants not reporting any bed days in the past 2 weeks. Among the 206 occupations, butchers and meat cutters reported the highest average number of bed days (4.63±0.77) and one hundred occupations reported no bed days. Among the gender-race-ethnicity subgroups, five groups of workers reported the highest average number of bed days (14.00±0.0) and one hundred occupations reported no bed days.
Lost Work
In Table 3, the overall prevalence of the acute disability measure, lost workdays, by the different occupations is reported by gender, race and ethnicity as well as the estimated US population numbers. Among the 13 occupations, administrative service (including clerical) workers reported the highest prevalence of ≥ 1 lost work days in the past 2 weeks (7.16±0.12) while farming, forestry and fishing workers reported the lowest (4.43±0.20). Among the gender-race-ethnicity subgroups, female farming, forestry, fishing workers reported the highest prevalence of ≥ 1 lost work days in the past 2 weeks (8.94±1.03) while Hispanic farming, forestry, fishing workers reported the lowest (2.94±0.46).
In Table 6, the average number of lost workdays (≥ 1) in the past 2 weeks by different occupations is reported, as well as the prevalence of participants not reporting any lost workdays in the past 2 weeks. Among the 13 occupations, transportation and material moving workers reported the highest average number of lost work days (4.90±0.14) and professional specialty workers reported the lowest average number (2.71±0.05). Among the gender-race-ethnicity subgroups, farming, forestry and fishing workers in the ‘other races’ category reported the highest average number of lost work days (8.28±1.46) and private household workers in the ‘other races’ category reported the lowest average number (2.00±0.0).
41 Occupations : In Table 3, the overall prevalence of the acute disability measure, lost workdays, by the different occupations is reported by gender, race and ethnicity as well as the estimated US population numbers. Among the 41 occupations, health service workers reported the highest prevalence of ≥ 1 lost work days in the past 2 weeks (10.17±0.37) while health diagnosing workers reported the lowest (2.49±0.29). Among the gender-race-ethnicity subgroups, architects and surveyors in the ‘other races’ category reported the highest prevalence of ≥ 1 lost work days in the past 2 weeks (21.55±8.70) while three occupations reported no lost workdays.
In Table 6, the average number of lost workdays (≥ 1) in the past 2 weeks by different occupations is reported, as well as the prevalence of participants not reporting any lost work in the past 2 weeks. Among the 41 occupations, other transportation except motor vehicles workers reported the highest average number of lost workdays (7.37±0.83) and Natural mathematical/computer scientists reported the lowest average number (2.16±0.12). Among the gender-race-ethnicity subgroups, forestry and fishing workers in the ‘other races’ category reported the highest average number of lost workdays (10.00±0); three occupations reported no lost workdays.
206 Occupations : In Table 3, the overall prevalence of the acute disability measure, lost workdays, by the different occupations is reported by gender, race and ethnicity as well as the estimated US population numbers. Among the 206 occupations, bill and account collectors reported the highest prevalence of ≥ 1 lost work days in the past 2 weeks (11.32±2.43) while dentists reported the lowest (1.73±0.49). Among the gender-race-ethnicity subgroups, black sheet metal workers reported the highest prevalence of ≥ 1 lost work days in the past 2 weeks (25.84±8.00) while seventy-six occupations reported no lost workdays.
In Table 6, the average number of lost workdays (≥ 1) in the past 2 weeks by different occupations is reported, as well as the prevalence of participants not reporting any lost workdays in the past 2 weeks. Among the 206 occupations, roofers reported the highest average number of lost workdays (6.00±0.77) and chemists (except biochemists) reported the lowest average number (1.78±0.22). Among the gender-race-ethnicity subgroups, eight occupations reported the highest average number of lost workdays (14.00±0.0) and seventy-six reported no lost workdays.
Doctor Visits
In Table 3, the overall prevalence and population estimates of doctor visits by the different occupations is reported by gender, race and ethnicity. Among the 13 occupations, administrative service (including clerical) workers reported the highest prevalence of ≥ 4 doctor visits in the past year (27.37±0.12) while farming, forestry and fishing workers reported the lowest (16.51±0.47). Among the gender-race-ethnicity subgroups, female professional specialty workers reported the highest prevalence of ≥ 4 doctor visits in the past year (33.28±0.32) while Hispanic farming, forestry and fishing workers reported the lowest (8.72±0.85).
In Table 7, the average number of doctor visits (≥ 4) in the past 12 months by different occupations is reported, as well as the prevalence of participants reporting 3 or less doctor visits in the past 12 months. Among the 13 occupations, transportation and material moving workers reported the highest average number of doctor visits (11.30±0.33) and farming, forestry and fishing workers reported the lowest average number (9.97±0.36). Among the gender-race-ethnicity subgroups machine operators, assemblers and inspectors in the ‘other races’ category reported the highest average number of doctor visits (12.91±1.98) and farming, forestry and fishing workers in the ‘other races’ category reported the lowest average number (8.00±1.02).
41 Occupations : In Table 3, the overall prevalence of the chronic disability measure, doctor visits, by the different occupations is reported by gender, race and ethnicity as well as the estimated US population numbers. Among the 41 occupations, teachers, librarians, and counselors reported the highest prevalence of ≥ 4 doctor visits in the past 12 months (29.86±0.38) while forestry and fishing workers reported the lowest (12.49±1.45). Among the gender-race-ethnicity subgroups, Hispanic officials and administrators public administration workers reported the highest prevalence of ≥ 4 doctor visits work in the past 12 months (39.57±6.84) while other transportation (except motor vehicles) workers in the ‘other races’ category reported the lowest (0.0±0.0).
In Table 7, the average number of doctor visits (≥ 4) in the past 12 months by different occupations is reported, as well as the prevalence of participants reporting 3 or fewer doctor visits in the past 12 months. Among the 41 occupations, health diagnosing workers reported the highest average number of doctor visits (≥ 4) in the past 12 months (14.85±1.42) and farm operators and managers reported the lowest average number (8.99±0.31). Among the gender-race-ethnicity subgroups, black other transportation (except motor vehicles) workers reported the highest average number of doctor visits (≥ 4) in the past 12 months (19.67±8.94) and other farming and fishing workers (8.00±0.94).
206 Occupations : In Table 3, the overall prevalence and estimated number of workers reporting doctor visits in the different occupations is reported by gender, race and ethnicity as well as the estimated US population numbers. Among the 206 occupations, psychologists reported the highest prevalence of doctor visits (≥ 4) in the past 12 months (36.39±1.84) while drywall installers reported the lowest (10.51±1.42). Among the gender-race-ethnicity subgroups, Sheriffs/bailiffs/other law enforcement officers in the ‘other races’ category reported the highest prevalence of doctor visits (≥ 4) in the past 12 months (56.93±12.18) while Hispanic sales and parts workers reported the lowest (1.42±1.42).
In Table 7, the average number of doctor visits (≥ 4) in the past 12 months by different occupations is reported, as well as the prevalence of participants reporting 3 or fewer doctor visits in the past 12 months. Among the 206 occupations, supervisors, related agricultural workers reported the highest average number of doctor visits (≥ 4) in the past 12 months (17.11±7.62) and sales and parts workers reported the lowest average number (7.62±0.55). Among the gender-race-ethnicity subgroups, dispatchers in the ‘other races’ category reported the highest average number of doctor visits (≥ 4) in the past 12 months (57.39±31.73) and seven occupational groups reported no doctor visits in last 12 months.
Hospital Days
In Table 3, the overall prevalence of the chronic disability measure, hospital days, by the different occupations is reported by gender, race and ethnicity as well as the estimated US population numbers. Among the 13 occupations, transportation and material moving workers reported the highest prevalence of ≥ 1 days of hospitalization in the past year (5.97±0.19) while professional specialty workers reported the lowest (4.55±0.08). Among the gender-race-ethnicity subgroups, female transportation and material moving workers reported the highest prevalence of ≥1 days of short term hospitalization in the past year (6.93±0.61) while male private household workers reported the lowest (1.13±0.81).
In Table 8, the average number of days in hospitalizations (≥ 1) in the past 12 months by different occupations is reported, as well as the prevalence of participants not reporting any days of hospitalization in the past 12 months. Among the 13 occupations, protective service workers reported the highest average number of days in hospitalization (1.29±0.08) and precision production, craft, repair workers reported the lowest average number (1.18±0.01). Among the gender-race-ethnicity subgroups, private household workers in the ‘other races’ category reported the highest average number of days in hospitalizations (2.37±0.61) and protective service workers in the ‘other races’ category reported the lowest average number (1.00±0.0).
41 Occupations : In Table 3, the overall prevalence of the chronic disability measure, hospital days, by the different occupations is reported by gender, race and ethnicity as well as the estimated US population numbers. Among the 41 occupations, health service workers reported the highest prevalence of ≥ 1 days of hospitalization in the past year (7.17±0.30) while natural mathematical/computer scientists reported the lowest (2.94±0.24). Among the gender-race-ethnicity subgroups, farm operators and managers in the ‘other races’ category reported the highest prevalence of ≥ 1 days of short term hospitalization in the past year (10.05±3.59) while five occupations reported no days of short term hospitalization.
In Table 8, the average number of ≥ 1 days of hospitalization in the past year by different occupations is reported, as well as the prevalence of participants not reporting any hospitalizations in the past year. Among the 41 occupations, engineers and other protective service workers reported the highest average number of ≥ 1 days of hospitalization in the past year (1.39±0.13 and 1.39±0.18) and health diagnosing workers reported the lowest average number (1.12±0.04). Among the gender-race-ethnicity subgroups, female other transportation except motor vehicles workers reported the highest average number of ≥ 1 days of hospitalization in the past year (4.12±1.41) and five occupations reported no hospitalizations.
206 Occupations : In Table 3, the overall prevalence of chronic disability measure, hospital days, by the different occupations is reported by gender, race and ethnicity as well as the estimated US population numbers. Among the 206 occupations, bill and account collectors reported the highest prevalence of ≥ 1 days of hospitalization in the past year (8.43±1.24) while operations/systems researchers and analysts reported the lowest (2.04±0.44). Among the gender-race-ethnicity subgroups, supervisors and related agricultural workers in the ‘other races’ category reported the highest prevalence of ≥ 1 days of short term hospitalization in the past year (26.86±16.01) while one hundred and eleven different occupations reported no hospitalization.
In Table 8, the average number of ≥ 1 days of hospitalization in the past year by different occupations is reported, as well as the prevalence of participants not reporting any ≥ 1 days of hospitalization in the past year. Among the 206 occupations, dispatchers reported the highest average number of ≥ 1 days of short term hospitalization in the past year (1.57±0.25) and drywall installers, electronic repairers, communication/industrial equipment, and dietitians reported the lowest average number (1.00±0.0). Among the gender-race-ethnicity subgroups, childcare workers and private household workers in the ‘other races’ category reported the highest average number of ≥ 1 days of short term hospitalization in the past year (3.00±0.0) and one hundred and eleven different occupations reported no hospitalization.
Self-rated Health
13 Occupations : In Table 3, the overall prevalence of the health indicator, self-rated health, by the different occupations is reported by gender, race and ethnicity as well as the estimated US population numbers. Among the 13 occupations, private household workers reported the highest prevalence of poor or fair health (16.38±0.74) while professional specialty workers reported the lowest (3.30±0.08). Among the gender-race-ethnicity subgroups, black private household workers reported the highest prevalence of poor or fair health in the past 2 weeks (25.33±1.65) while protective service workers in the ‘other races’ category reported the lowest (1.73±0.93).
In Table 9, the prevalence of self-reported health by different occupations is reported by each of the 5 different categories: poor, fair, good, very good, and excellent. Among the 13 occupations, professional specialty workers reported the highest prevalence of excellent health (48.58±0.33) and private household workers reported the lowest prevalence (24.02±0.88). Among the gender-race-ethnicity subgroups, male professional specialty workers reported the highest prevalence of excellent health (52.59±0.36) and black private household workers reported the lowest prevalence (17.66±1.22). Administrative support (including clerical) workers reported the highest prevalence of very good health (33.24±1.58) and private household workers reported the lowest prevalence (27.39±0.90). Among the gender-race-ethnicity subgroups, female protective service workers reported the highest prevalence of very good health (34.14±1.58) and black private household workers reported the lowest prevalence (21.99±1.54). Private household workers reported the highest prevalence of good health (32.21±0.93) and professional specialty workers reported the lowest prevalence (16.36±0.21). Among the gender-race-ethnicity subgroups, black machine operator, assembler, and inspectors reported the highest prevalence of good health (36.46±0.99) and male professional specialty workers reported the lowest prevalence (14.53±0.25). Private household workers reported the highest prevalence of fair health (14.6813±0.69) and professional specialty workers reported the lowest prevalence (2.90±0.07). Among the gender-race-ethnicity subgroups, black private household workers reported the highest prevalence of fair health (20.07±1.52) and protective service workers in the ‘other races’ category reported the lowest prevalence (1.73±0.93). Private household workers reported the highest prevalence of poor health (2.26±0.28) and technicians and related support workers reported the lowest prevalence (0.40±0.05). Among the gender-race-ethnicity subgroups, black private household workers reported the highest prevalence of poor health (4.63±0.71) and Hispanic technician and related support service workers reported the lowest prevalence (0.19±0.14). Of note, there were no persons reporting poor health among the protective service workers.
41 Occupations : In Table 3, the overall prevalence of the health indicator, self-rated health, by the different occupations is reported by gender, race and ethnicity as well as the estimated US population numbers. Among the 41 occupations, private household workers reported the highest prevalence of poor or fair health (16.38±0.74) while health diagnosing workers reported the lowest (2.23±0.26). Among the gender-race-ethnicity subgroups, black private household workers reported the highest prevalence of poor or fair health in the past 2 weeks (25.33±1.65) while Hispanic health diagnosing workers reported the lowest (1.26±0.96).
In Table 9, the prevalence of self-reported health by 41 different occupations is reported by each of the 5 different categories: poor, fair, good, very good, and excellent. Among the 41 occupations, health diagnosing workers reported the highest prevalence of excellent health (65.27±0.95) and private household workers reported the lowest prevalence (24.02±0.88). Among the gender-race-ethnicity subgroups, Hispanic private household workers reported the highest prevalence of excellent health (66.60±4.31) and black private household workers reported the lowest prevalence (17.66±1.22). Secretaries, stenographers, and typists reported the highest prevalence of very good health (33.96±0.42) and health diagnosing occupation workers reported the lowest prevalence (23.06±0.89). Among the gender-race-ethnicity subgroups, Hispanic officials and administrators and public administration workers reported the highest prevalence of very good health (38.70±5.17) and Hispanic forestry and fishing workers reported the lowest prevalence (12.15±6.62). Private household workers reported the highest prevalence of good health 32.21±0.93) and health diagnosing workers reported the lowest prevalence (9.44±0.62). Among the gender-race-ethnicity subgroups, black forestry and fishing workers reported the highest prevalence of good health (39.84±5.09) and black health diagnosing workers reported the lowest prevalence (7.78±2.24). Private household workers reported the highest prevalence of fair health (14.13±0.69) and health diagnosing workers reported the lowest prevalence (2.10±0.25). Among the gender-race-ethnicity subgroups, black forestry and fishing workers reported the highest prevalence of fair health (20.77±5.34) and private household workers reported the lowest prevalence (1.26±0.92). In contrast, Forestry and fishing workers reported the highest prevalence of poor health (2.52±0.65) and health diagnosing workers reported the lowest prevalence (0.13±0.06). Among the gender-race-ethnicity subgroups, female other transportation except motor vehicles workers reported the highest prevalence of poor health (7.92±7.69) and male health diagnosing workers reported the lowest prevalence (0.10±0.06). Of note, there were no persons reporting poor health among 19 occupations.
206 Occupations : In Table 3, the overall prevalence of the health indicator, self-rated health, by the different occupations is reported by gender, race and ethnicity as well as the estimated US population numbers. Among the 206 occupations, private household cleaners and servants reported the highest prevalence of poor or fair health (17.99±0.94) while dentists reported the lowest (1.23±043). Among the gender-race-ethnicity subgroups, female brick masons and stonemasons reported the highest prevalence of poor or fair health in the past 2 weeks (35.24±26.42) while black computer programmers reported the lowest (0.32±0.32).
In Table 9, the prevalence of self-reported health by different occupations is reported by each of the 5 different categories: poor, fair, good, very good, and excellent. Among the 206 occupations, airplane pilots and navigators reported the highest prevalence of excellent health (78.29±2.19) and private household cleaners and servants reported the lowest prevalence (22.43±0.98). Among the gender-race-ethnicity subgroups, both black and airplane pilots and navigators in the ‘other races’ category reported the highest prevalence of excellent health (100.00±0.0) and black sales workers, hardware and building supplies workers reported the lowest prevalence (7.89±6.85). Library clerks reported the highest prevalence of very good health (37.71±2.46) and airplane pilots and navigators reported the lowest prevalence (14.49±1.80). Among the gender-race-ethnicity subgroups, sales workers, radio, TV, hi-fi and appliances workers in the ‘other races’ category reported the highest prevalence of very good health (59.50±12.70) and sheet metal workers in the ‘other races’ category reported the lowest prevalence (4.49±4.25). Textile sewing machine operators reported the highest prevalence of good health (37.52±1.30) and airplane pilots and navigators reported the lowest prevalence (5.91±1.28). Among the gender-race-ethnicity subgroups, timber cutting and logging workers in the ‘other races’ category reported the highest prevalence of good health (100.00±0.0) and dentists in the ‘other races’ category reported the lowest prevalence (2.44±2.41). Private household cleaners and servants reported the highest prevalence of fair health (15.33±0.88) and airplane pilots and navigators reported the lowest prevalence (1.12±0.59). Among the gender-race-ethnicity subgroups, female brick masons and stonemasons reported the highest prevalence of fair health (35.24±26.42) and black computer programmers reported the lowest prevalence (0.32±0.32). Taxicab drivers and chauffeurs reported the highest prevalence of poor health (3.15±0.90) and firefighting workers reported the lowest prevalence (0.11±0.11). Among the gender-race-ethnicity subgroups, female roofers reported the highest prevalence of poor health (21.50±18.08) and male and non-Hispanic electrical and electronic, chemists (except biochemists) reported the lowest prevalence (0.08±0.06). Of note, there were no persons reporting poor health among the numerous occupational groups.
Health conditions
In Table 3, the overall prevalence of the health indicator, number of health conditions, by the different occupations is reported by gender, race and ethnicity as well as the estimated US population numbers. Among the 13 occupations, transportation and material moving reported the highest prevalence of ≥ 1 health conditions (45.58±1.13) while handler, equipment cleaner, helper and laborer workers reported the lowest (33.47±0.46). Among the gender-race-ethnicity subgroups, non-Hispanic private household workers reported the highest prevalence of ≥ 1 health condition (48.95±1.19) while Hispanic handler, equipment cleaner, helper and laborer workers reported the lowest (23.63±1.20).
In Table 10, the average number of self-reported health conditions (≥ 1) by different occupations is reported, as well as the prevalence of participants not reporting any health conditions. Among the 13 occupations, private household workers reported the highest average number of self reported health conditions (1.85±0.03) and handler, equipment cleaner, helper and laborer workers reported the lowest average number (1.52±0.01). Among the gender-race-ethnicity subgroups, black private household workers reported the highest average number of reported health conditions (1.89±0.06) and Hispanic precision production, craft and repair workers reported the lowest average number (1.37±0.02).
41 Occupations : In Table 3, the overall prevalence of the chronic disability measure, number of health conditions, by the different occupations is reported by gender, race and ethnicity as well as the estimated US population numbers. Among the 41 occupations, private household workers reported the highest prevalence of ≥ 1 health conditions (45.58±1.13) while construction laborers reported the lowest (28.83±1.04). Among the gender-race-ethnicity subgroups, female other transportation except motor vehicles workers reported the highest prevalence of ≥ 1 health conditions (49.25±11.27) while Hispanic forestry and fishing workers reported the lowest (15.35±6.65).
In Table 10, the average number of ≥ 1 health conditions by different occupations is reported, as well as the prevalence of participants not reporting any health conditions. Among the 41 occupations, private household workers reported the highest average number of ≥ 1 health conditions (1.85±0.03) and construction laborers reported the lowest average number (1.43±0.03). Among the gender-race-ethnicity subgroups, construction laborers in the ‘other races’ category reported the highest average number of ≥ 1 health conditions (2.30±0.32) and Hispanic other transportation except motor vehicles workers reported the lowest average number (1.17±0.13).
206 Occupations : In Table 3, the overall prevalence of chronic disability measure, number of health conditions, by the different occupations is reported by gender, race and ethnicity as well as the estimated US population numbers. Among the 206 occupations, psychologists reported the highest prevalence of ≥ 1 health conditions (48.59±1.94) while airplane pilots and navigators reported the lowest (27.91±2.58). Among the gender-race-ethnicity subgroups, correctional institution officers in the ‘other races’ category reported the highest prevalence of ≥ 1 health conditions (82.17±12.34) while four groups of workers reported the lowest no health conditions.
In Table 10, the average number of ≥ 1 health conditions by different occupations is reported, as well as the prevalence of participants not reporting any health conditions. Among the 206 occupations, street and door-to-door sales workers reported the highest average number of ≥ 1 health conditions (1.89±0.06) and airplane pilots and navigators reported the lowest average number (1.33±0.06). Among the gender-race-ethnicity subgroups, bill and account collectors in the ‘other races’ category reported the highest average number of ≥ 1 health conditions (4.47±2.07) and four groups of workers reported no health conditions.
Kaminski and Spirtas (1980) analyzed data from the 1969-74 NHIS surveys to examine the morbidity, disability, and reported healthcare use patterns for 498,580 individuals by industry. They did not look at trends over this relatively short time period. The highest specific disease conditions were reported for agriculture, furniture manufacturing, metal fabrication, railroad transport, repair services, amusement and recreational services, state and local government workers, and new workers; the highest disabilities were found for forestry and fisheries workers, certain manufacturers, medical and health services workers, and federal government employees; the greatest use of medical services was among metal industries, specific manufacturers, and railroad workers; the greatest morbidity was reported by private household service workers although they had less disability and use of medical services; overall manufacturing industries had the largest proportion of workers with work injuries and the service industries had the smallest.
The current Disability Monograph of all currently employed adults 18 years or older from the 1986-1994 NHIS surveys demonstrated continued increased acute and chronic disability among certain occupational groups more than others. For example, with respect to prevalence of disability, service workers had consistently higher prevalence of acute disability as measured by restricted days and bed days within the past 2 weeks, while (particularly Hispanic) farmers had consistently lower prevalence of these measures of acute disability. Of interest, for several acute disability measures, healthcare service workers reported high average restricted, bed and lost workdays while health diagnosing workers reported some of the lowest averages. With respect to prevalence of chronic disability, female and black transportation workers had the highest prevalence of doctor visits and hospital days in the past 12 months, while farmers and professional workers had the lowest prevalence.
There were differences between those occupations with increased prevalence of acute and chronic disability and those occupations that reported an increased average number. For example, for the acute disability measure of bed days, as noted above, service workers had the highest prevalence but the lowest average number of bed days while farmers had the lowest prevalence of bed days but the highest average of bed days. Thus, farmers rarely reported bed days but when they did report bed days, it resulted in a large number of bed days per farmer, possibly reflecting infrequent but more serious injury.
With regards to self-rated health, male professional workers (particularly airplane pilots and navigators) had the highest prevalence of reporting excellent health, while (particularly black) private household workers and female transportation workers (other than motor vehicle workers) had the highest prevalence of poor health. Transportation and material moving workers reported the highest prevalence of having at least one health condition while laborers reported the lowest prevalence. Among those reporting at least one health condition, (particularly black) private household cleaners reported the highest average number while laborers and craft/repair workers reported the lowest average number.
In general, among the occupations, more skilled “white collar” workers were more likely to report less disability, better self rated health and fewer health conditions, although farmers seemed to represent a consistent exception to this pattern. Within particular occupations, female and black and workers were more likely to report acute and chronic disability measures, poor self rated health and numbers of health conditions; Hispanic and male workers were more likely to report less acute and chronic disability, better health and fewer health conditions. Race-specific comparisons often revealed higher morbidity in workers in the 'other races' category.
In a separate analysis, the investigators have modeled acute and chronic disability adjusting for age, gender, race/ethnicity, and education in the entire NHIS worker database, as well as comparing disability among farmers and pesticide applicators with all other US workers (Gomez Marin in press). Among all NHIS workers, for acute disability younger age was a significant risk factor; older age was a significant risk factor in a dose response fashion for reporting chronic disability and poor health status. Of note, using the same NHIS database, Zwerling et al (1996, 1997, 1998) have found a significantly increased reported injury risk among older and disabled US workers. Gomez Marin et al (in press) also found that older female uneducated workers were at significantly higher risk of chronic disability and reported poor health status compared to all other US workers. Less educated older female workers are more likely to have poorly paying jobs with significant physical labor and hours.
Limitations
These analyses suffer from some of the data limitations seen in previous epidemiologic studies of pesticide-exposed workers. Limitations include: the self report and cross sectional nature of the data which might lead to underestimation of the true health situation since really sick people leave the work force (part of the healthy worker effect); possible farmer under-reporting bias; lack of individual exposure measures, and occupational misclassification in general, as well as misclassification of occupation related pesticide exposures. However, self-rated health has been shown in numerous national and international studies to be a significant independent predictor of morbidity and mortality (Markides 1991, Idler 1997). The use of the US worker population as the major comparison population is appropriate for controlling for the healthy worker effect and other biases (Cooper 1993, Breslow 1980, Checkoway 2004, Monson 1990, Burnett 1989). Previous work using the NHIS database has shown that certain occupational groups, such as farmers, smoke less than many other worker groups (15-29% prevalence) which may explain some but not all of the study findings (Brackbill 1998, Nelson 1994, Sterling 1990, Lee 2004). These data were based entirely on self-report (by the individual or proxy) without objective confirmation. Access to care issues, particularly among certain occupational groups such as farmers, may be reflected in some of the reporting of the disability indicators, particularly the chronic disability measures of doctor visits and hospitalizations. The lack of these data is an important limitation on any conclusions that can be drawn from these analyses.
Although research in this area is often contradictory, validation studies conducted by the NCHS and others suggest that proxy reports lead to slightly lower prevalence estimates of chronic conditions and ambulatory medical visits compared with reports obtained directly from respondents (Edwards 1994, Edwards 1996, Thornberry 1987). To address this potential bias, we repeated our morbidity analyses, including only the 63% of NHIS participants who were interviewed directly in the 1986-1994 NHIS surveys. Findings indicated that the non proxy self reported responses of the disability measures were uniformly greater than those from the proxy responses. Further evaluation of the individual disability measures of all workers demonstrated that the non proxy self reported respondents were uniformly <1% greater than the total population of proxy and non proxy self reported respondents used in this Monograph.
Ultimately the NHIS data strongly suggest that these workers, as well as older workers, should be an important target population for occupational prevention and intervention. Researchers in occupational health are urged to seek out this important and publicly available resource. The NHIS database and mortality follow up represent a probability sample of the entire US population, with the ability to compare both morbidity and mortality among US workers. Furthermore, as noted by NIOSH (1997) and the English Registrar’s Decennial Reports (Drever 1995), databases such as the NHIS surveys and mortality follow up can be used not only to target studies of work-related conditions and to add to the body of evidence generated from epidemiologic studies, but also to provide surveillance data for establishing priorities, and for tracking progress towards the elimination of preventable diseases.
Alderson MR. Some sources of error in British occupational mortality data. Br J Ind Med 1972;29:245-254.
Alterman T, Burnett C, Peipins L, Lalich N, Halperin W. Occupation and cervical cancer: an opportunity for prevention. J Womens Health 1997;6:649-57.
Behrens V, Seligman P, Cameron L, Mathias CGT, Fine L. The prevalence of back pain, hand discomfort, and dermatitis in the US working population. Am J Pub Health 1994;84:1780-5.
Biddlecom AE, LeClere FB, Hardy AM, Hendershot GE. National study of knowledge of AIDS, testing patterns, and self-assessed risk among health care workers. J AIDS 1992;5:1131-6.
Boffetta P, Kogevinas M, ed. Occupational Cancer in Europe. Env Health Perspectives 1999; 107; Suppl 2:227-306.
Botman SL, Jack SS. Combining national health interview survey datasets: issues and approaches. Stats in Med 1995;14:669-77.
Brackbill R, Frazier T, Shilling S. Smoking Characteristics of US workers, 1978-80. Am J Ind Med 1998;13:5-41.
Brackbill RM, Cameron LL, Behrens V. Prevalence of chronic diseases and impairments among US farmers, 1986-90. Am J Epid 1994;139:1055-1065.
Breslow NE, Day NE. Statistical Methods in Cancer Research. Vol I: The Analysis of Cohort Studies. Lyon, France: IARC, 1980.
Burnett CA, Crouse WE. Agreement of latest and longest occupation and industry as reported by 1980 National Health Interview Survey. JOM 1989;31(4):339-346.
Checkoway H, Pearce NE, Kriebel D. Research Methods in Occupational Epidemiology, Second Edition. New York: Oxford University Press, 2004.
Cooper SP, Buffler PA, Lee ES, Cooper CJ. Health characteristics by longest held occupation and industry of employment: US. Am J Ind Med 1993;24:25-39.
Drever F ed. The Registrar General’s Decennial Supplement for England & Wales; Occupational Health: Decennial Supplement. London: Office of Population Censuses and Surveys Health and Safety Executive, Series DS #10, 1995.
Edwards WS WD, Kurlantzick V, et al. Evaluation of National Health Interview Survey diagnostic reporting. Vital Health Stat 2. Feb 1994(120):1-116.
Edwards WS, Winn DM, Collins JG. Evaluation of 2-week doctor visit reporting in the national health interview survey. Vital Health Stat 2. Aug 1996(122):1-46.
Emanuel DA, Draves DL, Nycz GR. Occupational health services for farmers. Am J Ind Med 1990;18:149-162.
Fosbroke DE, Kisner SM, Myers JR. Working lifetime risk of occupational fatal injury. Am J Ind Med 1997;31:459-67.
Gallagher RP, Threlfall WJ, Band PR, Spinelli JJ. Occupational Mortality in British Columbia 1950-1984. Vancouver, BC: Cancer Control Agency of British Columbia, 1989.
Gomez Marin O, Fleming LE, Lee DJ, LeBlanc W, Zheng D, Ma F, Jane D, Pitman T, Caban A. Acute and Chronic Disability among US Farmers and Pesticide Applicators: The National Health Interview Survey (NHIS). Journal Agricultural Safety Health in press.
Guralnick L. Mortality by industry and cause of death among men 20-64 years of Age: US, 1950. Vol 53(4). Washington DC: NCHS, 1963b.
Guralnick L. Mortality by occupation and causes of death among men 20-64 years of Age: US, 1950. Vol 53(3). Washington DC: NCHS, 1963a.
Guralnick L. Mortality by occupation and industry among men 20-64 years of Age: US, 1950. Vol 53(2). Washington DC: NCHS, 1962.
Hurwitz EL, Morgenstern H. Correlates of back problems and back-related disability in the US. J Clin Epid 1997;50:669-81.
Idler EL, Benyamini Y. Self-rated health and mortality: a review of twenty-seven community studies. Journal of Health & Social Behavior 1997;38(1):21-37.
Kaminski R, Spirtas R. Industrial Characteristics of Persons Reporting Morbidity during the Health Interview Surveys Conducted in 1969-1974. Cincinnati, OH: NIOSH, 1980.
Kircher T, Nelson J, Burdo H. The autopsy as a measure of accuracy of the death certificate. NEJM 1985;313:1263-1269.
La Rosa JH. Women, Work and Health: Employment as a risk factor for coronary heart disease. Am J Ob Gyn 1988;158:15597-1602.
Lee DJ, LeBlanc W, Fleming LE, Gomez-Marin O, Pitman T. Trends in US Smoking Rates in Occupational Groups: The National Health Interview Survey 1987-1994. JOEM 2004;46:538-548.
Liao Y, Cooper RS, Cao G et al. Mortality patterns among adult Hispanics: findings from the NHIS, 1986-1990. Am J Pub Health 1998;88:227-32.
Loomis D, Richardson D. Race and the risk of fatal injury at work. Am J Pub Health 1998;88:40-4.
Ma F, Lee DJ, Fleming LE, Mustafa D. Race specific cancer mortality in US Firefighters: 1984-1993. J Occ Env Med 1998;40:1134-8.
Markides KS, Lee DJ. Predictors of health status in middle-aged and older Mexican Americans. Journal of Gerontology 1991;46(5):S243-9.
Massey JT, Morre TF, Parson VL, Tadros W. Design and Estimation for the NHIS 1985-94. Washington DC: NCHS, Vital & Health Statist Series 2#110 DHHS Pub PHS89-1384, 1989.
Milham S. Occupational Mortality in Washington State 1950-71. Cincinnati, OH: NIOSH, Dept of HEW #76-175-A, 1976.
Milham S. Occupational Mortality in Washington State 1950-79. Cincinnati, OH: NIOSH, Dept of HEW #83-116, 1983.
Monson RR. Occupational Epidemiology. Boca Raton: CRC Press, 1990, second edition.
Murphy PL, Sorock GS, Courtney TK, Webster BS, Leamon TB. Injury and illness in the American workplace: a comparison of data sources. Am J Ind Med 1996;30:130-141.
National Center for Health Statistics (NCHS). Documentation for the NHIS Multiple Cause of Death Public Use Data File 1986-94 Survey Years; Dates of Death 1986-97. Washington, DC: US Dept of Health and Human Services, NCHS, 2000.
Nelson DE, Emont SL, Brackbill RM, Cameron LL, Peddicord J, Fiore MC. Cigarette smoking prevalence by occupation in the US. JOEM 1994;36:516-25.
NIOSH. Fatal Injuries to Workers n the US, 1980-89. Cincinnati, OH: NIOSH #93-108, 1993.
Research Triangle Institute (RTI). Software for Survey Data Analysis (SUDAAN), Version 8.0.0. Research Triangle Park, NC: RTI, 2001.
Reviere R, Schneider S, Woolbright K. Associations between disease and occupation: hypotheses generated from the National Mortality Followback Survey. Am J Ind Med 1995;27:195-205.
Schade WJ, Swanson GM. Comparison of death certificate occupation and industry data with lifetime occupational histories obtained by interview: variations in the accuracy of death certificate entries. AM J Ind Med 1988;14:121-136.
Schumacher MC. Comparison of occupation and industry information from death certificates and interviews. AmJPH 1986;76(6): 635-637.
Sterling T, Weinkam J. The confounding of occupation and smoking and its consequences. Soc Sci Med 1990;30:457-67.
Sterling TD, Weinkam J. Comparison of smoking related factors among black and white males. Am J Ind Med 1989;15:319-33.
Stout NA, Jenkins EL, Pizatella TJ. Occupational injury mortality rates in the US: Changes from 1980-89. Am J Public Health 1996;86:73-7.
Swanson GM, Schwartz G, Burrows R. An assessment of occupation and industry data from death certificates and hospital medical records for population based cancer surveillance. AmJPH 1984;74:464-467.
Tanaka S, Wild DK, Seligman PJ, Halperin WE, Behrens V, Putz-Anderson V. Prevalence and work-relatedness of self-reported carpal tunnel syndrome among US workers: Analysis of the occupational health supplement data of 1988 National Health Interview Survey. Am J Ind Med 1995;27:451-70.
Thornberry O. An experimental comparison of telephone and personal health surveys. Vital Health Stat 2. 1987;106:1-186.
Toscano G, Jack T. Occupational injury fatalities -- 1994. Statistical Bulletin – Metropolitan Insurance Company 1996;77:12-22.
Wagener DK, Walstedt J, Jenkins L, Burnett C, Lalich N, Fingerhut M. Women: work and health. Vital Health Stats 1997;31:1-91.
Wagener DK, Winn DW. Injuries in working populations: black-white differences. Am J Public Health 1991;81:1408-14.
Zwerling C, Whitten PS. Davis CS, Sprince NL. Occupational injuries among workers with disabilities: the NHIS 1985-94. JAMA 1997;278:2163-6.
Zwerling C, Whitten PS. Davis CS, Sprince NL. Occupational injuries among older workers with visual, auditory, and other impairments. A validation study. J Occ Env Med 1998;40:720-3.
Table 1. Occupational Groups, US Census Occupational Codes, Sample Numbers, and Population Numbers by 13, 41, and 206 Occupations (Download Table 1 in Excel Format)
Table 2. Monograph NHIS Definitions of Disability
Table 3. Overall Prevalence of Health Indicators by 13, 41, and 206 Occupations (Download Table 3 in Excel Format)
Table 4. Average Number of Restricted Days (≥1) in the past 2 weeks by 13, 41, and 206 Occupations (Download Table 4 in Excel Format)
Table 5. Average Number of Bed Days (≥ 1) in the past 2 weeks by 13, 41, and 206 Occupations (Download Table 5 in Excel Format)
Table 6. Average Number of Lost Work Days (≥ 1) in the past 2 weeks by 13, 41, and 206 Occupations (Download Table 6 in Excel Format)
Table 7. Average Number of Doctor Visits (≥ 4) in the past 12 months by 13, 41, and 206 Occupations (Download Table 7 in Excel Format)
Table 8. Average Number of Days in Hospitalizations (≥1) in the past 12 months by 13, 41, and 206 Occupations (Download Table 8 in Excel Format)
Table 9. Prevalence of Self Reported Health by 13, 41, and 206 Occupations (Download Table 9 in Excel Format)
Table
10. Average
Number of Self Reported Health Conditions (≥ 1) by 13, 41, and 206 Occupations (Download Table 10 in
Excel Format)
If you
experience any problems with this site, please contact the webmaster
© University of Miami 2004.
Page last updated September 4, 2004.
