Profile of the Public Health Workforce: Registered TRAIN Learners in the United States

Abstract

Objectives: We analyzed data from the TrainingFinder Real-time Affiliate Integrated Network (TRAIN), the most widely used public health workforce training system in the United States, to describe the public health workforce and characteristics of individual public health workers.

Methods: We extracted self-reported demographic data of 405 095 learners registered in the TRAIN online system in 2012.

Results: Mirroring the results of other public health workforce studies, TRAIN learners are disproportionately women, college educated, and White compared with the populations they serve. TRAIN learners live in every state and half of all zip codes, with a concentration in states whose public health departments are TRAIN affiliates. TRAIN learners’ median age is 46 years, and one third of TRAIN learners will reach retirement age in the next 10 years.

Conclusions: TRAIN data provide a limited but useful profile of public health workers and highlight the utility and limitations of using TRAIN for future research.

To paraphrase a common dictum in management, one can only manage what one can measure, and in the arena of public health, data on workers are surprisingly sparse. In evaluating the efficacy of public health systems, researchers currently rely on 3 levels of data: community, agency, and individual public health worker. Vital statistics, the US Census Bureau, and disease registries provide considerable data on the demographics, socioeconomic characteristics, and health measures of communities and populations at the state, county, and municipal level. Data gathered by public health professional organizations such as the National Association of County and City Health Officials (NACCHO) and the Association of State and Territorial Health Officials (ASTHO) profile individual state and local health departments at the agency level. Data in the third area, individual health workers, remain relatively rare and challenging to collect. In this area, the TrainingFinder Real-time Affiliate Integrated Network (TRAIN) offers some limited but unique insights.

Structurally, governmental public health workers are divided among federal, state, and local agencies with large variations in the services offered, training provided, and jurisdictions served by these health workers. No single system currently provides the number of public health employees working in the United States or provides information on the composition and training levels of these employees. For more than a decade, many articles have described the needs, frustrations, and challenges involved in enumerating and describing the public health workforce.^1–4

HISTORY OF ENUMERATION

At least 2 comprehensive reviews of the relatively long history of public health workforce enumeration in the United States have been published. A 2003 review⁵ identified a national census of public health workers as early as 1908, and the federal government conducted regular enumerations until 1964. This review also identified a decline in the number of public health workers, from 220 per 100 000 population in 1980 to 158 per 100 000 population in 2000. A subsequent review⁶ of 126 scholarly works published from 1985 to 2010 on the public health workforce found these works to be primarily descriptive in nature, with 30 articles focusing primarily on the size and composition of the workforce.

Studies such as a seminal 2000 study⁷ commissioned by the Health Resources and Services Administration have aggregated various data sets to describe the public health workforce. This study pointed out the critical need for workforce data to make effective planning and policy decisions. Examining public health workers at the combined federal, state, and local levels, the study’s authors counted 448 254 workers divided among federal (19%), state (33%), local (34%), and other (14%) agencies. This other category included individuals working in schools of public health and other settings. Physicians accounted for only 1.3% of workers. Public health nurses, however, made up the largest professional job category among public health workers and accounted for almost 11% of all workers. Another 2.8 million individuals volunteered for public health initiatives.

In the aftermath of the 9/11 terrorist attacks, the US government invested heavily in expanding emergency preparedness training and public health capacity surge capabilities. These funds supplemented hiring in public health and allowed for an expansion of computer equipment and networking.⁸ Around the same time, the Public Health Foundation developed TRAIN, which was initially a mechanism to quickly provide and track emergency responder training via the Internet. Today, agencies use TRAIN to offer a variety of training sessions, including new employee orientation and sessions on continuing education topics spanning public health practice.

RECENT ENUMERATION DATA

At the agency level, the preeminent data sources are NACCHO and ASTHO, who conduct regular censuses of local and state health departments.⁹ NACCHO’s 2013 profile report¹⁰ focused exclusively on local health departments. The 2013 NACCHO census found that local health departments employ about 162 000 people, equating to 146 000 full-time-equivalent positions, which is a decline from 190 000 local public health employees in 2008. As was found in the 2000 Health Resources and Services Administration study, the most common professional position was nurse.

ASTHO’s 2010 survey¹¹ identified 106 815 state health agency workers, equating to 102 760 full-time-equivalent positions. The ASTHO survey asked some demographic questions about the composition of states’ public health workforce. It also found that the median age of state public health employees was 49 years and the average number of years of service was 12. At the federal level, a study published in 2013¹² calculated that the Centers for Disease Control and Prevention employed 11 223 public health workers in 2012. Of these workers, 61% were women, and 30% of workers will be eligible for retirement by 2017.

Combining data from the NACCHO, ASTHO, and CDC studies provides a figure of 280 038 public health workers employed in state and local health departments and this one federal agency. This figure certainly does not include all public health workers (especially at the federal level, in schools of public health, or in nongovernmental agencies), but it does provide an aggregation of existing data.

Several studies have also looked at specific professions within public health, including epidemiology^13,14 and environmental health.¹⁵ Other studies have focused on specific geographic areas such as Nebraska,¹⁶ Switzerland,¹⁷ or rural areas.¹⁸ The Nebraska study found significant worker shortages in the areas of environmental health, chronic disease, health disparities, sexually transmitted infections, and mental health.¹⁶ A 2014 chapter¹⁸ focusing on the rural public health workforce described rural health workers as more likely to work part-time and to work in an agency led by a nurse rather than a physician. Such rural, nurse-led, local health agencies tend to offer more direct patient care but fewer environmental and regulatory oversight services. Especially in states with a decentralized public health system, declining rural populations result in a declining tax base. The result is often an aging population with increased needs served by a public health workforce undergoing budget-driven staffing cuts. These studies share a common theme: the lack of consistently collected and accurate data on public health workers.

ROLE OF TRAIN

A few studies described the use of TRAIN in the critical education of public health workers during disasters,¹⁹ in emergency preparedness training,²⁰ and as an effective health informatics tool for online continuing education.²¹ Indeed, the Public Health Foundation created TRAIN in response to the 9/11 terrorist attacks and the need to quickly provide updated training to public health workers and emergency responders on bioterrorism and other health threats. As such, it was not created as an enumeration tool. TRAIN’s success in enrolling hundreds of thousands of public health workers, however, creates a natural interest in analyzing the data collected through direct access to public health workers otherwise fractured into multiple state and agency systems.

In 2012, 21 states plus the CDC, the Medical Reserve Corps, and the CDC’s Division of Global Migration and Quarantine were TRAIN affiliates. Affiliates provide financial support for the system, which uses a network model. The national TRAIN hub allows affiliates’ state-level TRAIN systems to share courses nationally. Instructors design and upload course materials, which learners can then access. Courses may include evaluation tools such as quizzes and exams.

As of 2014, more than 800 000 learners have registered in TRAIN. In some cases, individual learners may register because of personal interest in continuing education. In Kentucky, all governmental public health workers must register in TRAIN for mandatory employee orientation and training. As such, Kentucky’s TRAIN data include almost all of its governmental public health workers. TRAIN’s initial registration process asks learners to self-report their demographic information. Other TRAIN data include the courses, competency and capability covered by the content, and the number of learners who have taken a particular course.

TRAIN thus provides the largest existing data set of individual public health workers in the United States. These data are especially important in light of the scarcity of data on the number and demographic characteristics of individual health workers.

METHODS

After approval by the University of Kentucky institutional review board, we extracted TRAIN learners’ demographic data from the TRAIN system, using its administrator report functions. We then redacted personal identifiers from the data set.

We filtered out any user who lived in a country other than the United States. The resulting analytic data set consisted of 405 095 learners. The data set included several thousand duplicate learner accounts, which were not removed. We conducted univariate and bivariate analyses of aggregate demographic data at the national, state, and zip-code levels using IBM SPSS Statistics version 20 (IBM Corporation, Armonk, NY). We then imported these statistical tables into ArcMap 10.1 (ESRI, Redlands, CA) for spatial analysis and mapping.

RESULTS

In our analysis of the 405 095 learners in the analytic data set, TRAIN learners live in all 50 states but are more common in TRAIN affiliate states (Figure 1). We found that 3 of 4 learners responded to the optional demographic questions; 77% self-reported demographics for gender, 74% for race, 73% for ethnicity, and 76% for educational attainment. Among the 77% of learners reporting their gender, two thirds (66%) were women. Women make up the majority of public health workers nationally, and this pattern also extended to TRAIN learners, except for those in 1 state: The majority of TRAIN learners in Utah were men. Learners’ median age was 46 years, with a mean age of 47. By 2025, 30% of learners will have reached age 65. Six percent of learners have already reached age 65.

FIGURE 1—

TRAIN learners per 100,000 population: Public Health Foundation, 2012.

Note. TRAIN = TrainingFinder Real-time Affiliate Integrated Network. Asterisks indicate states that were TRAIN affiliate states in 2012.

In terms of racial composition, the learners were predominately White (88%), followed by Black or African American (7%), Asian (2%), multiracial (2%), American Indian (0.6%), and Native Hawaiian/Pacific Islander (0.4%). White learners constituted the majority in 49 states and the District of Columbia (Figure 2). Among Hawaiian learners, however, Asian, Native Hawaiian, Black, and other racial minorities outnumbered Whites. Hawaii was also home to the largest percentage of Asian public health learners. Louisiana and the District of Columbia had the largest percentage of Black learners, and American Indians made up significant portions of learners in various Western states (e.g., Oklahoma, Arizona, New Mexico, South Dakota). Alaska was, not surprisingly, home to the largest percentage of Alaska Native public health workers registered in TRAIN.

FIGURE 2—

Race/ethnicity of 2012 TRAIN learners: Public Health Foundation, 2012.

Note. TRAIN = TrainingFinder Real-time Affiliate Integrated Network.

Regarding ethnicity, Latino workers made up 6% of TRAIN learners. In most states, the majority of learners were non-Hispanic. New Mexico had the largest percentage of Hispanic TRAIN learners, mirroring the state’s population, which is largely Hispanic. States with a large Latino population, such as Texas, California, Arizona, and Florida, also had a sizable percentage of Hispanic TRAIN learners. TRAIN learners also reported using dozens of primary languages at home. After English and Spanish, Tagalog, Armenian, Chinese, and Hindi were the most commonly spoken primary languages.

Education levels among the states’ TRAIN learners varied considerably. Among the 76% of learners who reported their educational attainment, 56% reported having a 4-year college degree. Twenty-eight percent reported also having completed a graduate or professional degree (e.g., MPH, MD, JD, PhD). One percent of learners reported not having completed high school. In 16 states and the District of Columbia, 3 of every 4 learners had at least a bachelor’s degree. In 5 states—Utah, Kansas, Arkansas, Wisconsin, and Virginia—the majority of registered TRAIN learners report having less than a 4-year college degree.

In 2012, only 88 of the 405 095 registered TRAIN learners in the United States were not associated with a particular state or the District of Columbia. The largest percentage of learners lived in Kentucky, where 17.8% of all US TRAIN learners were registered. This is not by chance. Kentucky uses TRAIN to orient and train all new governmental public health workers. Twenty-one states and the CDC, Medical Reserve Corps, and CDC’s Division of Global Migration and Quarantine were TRAIN affiliates in 2012. TRAIN affiliate states are home to a larger percentage of TRAIN learners than nonaffiliate states.

Ninety-four percent of learners reported their zip code. Registered TRAIN learners lived or worked in 15 466 of 30 306 zip codes (Figure 3). Although TRAIN learners are concentrated in the TRAIN affiliate states, 51% of zip codes across the United States have TRAIN learners.

FIGURE 3—

TRAIN learners by zip code: Public Health Foundation, 2012.

Note. TRAIN = TrainingFinder Real-time Affiliate Integrated Network.

One workforce question regards whether the public health workforce reflects the demographics of the population it serves. In only 2 states—Louisiana and North Dakota—did the percentage of minority TRAIN learners exceed the percentage of minorities in those states’ total population (Figure 4). In 22 states and the District of Columbia, White TRAIN learners exceeded the percentage of Whites in the general population by 11% or more. In Nevada and the District of Columbia, they exceeded the percentage of Whites in the general population by more than 20%.

FIGURE 4—

Race/ethnicity of 2012 TRAIN learners by state compared to the racial/ethnic composition of each state: Public Health Foundation, 2012.

Note. TRAIN = TrainingFinder Real-time Affiliate Integrated Network.

Limitations

These data have several limitations and potential sources of bias. TRAIN is a training management system, not a research or enumeration tool. Most public health workers in most states are not registered in TRAIN and are therefore not included in this analysis. Although TRAIN courses span the breadth of public health topics, the Public Health Foundation originally developed this system after the 9/11 attacks to train emergency responders. Thus, learners may be drawn more from particular professions within public health (e.g., environmental health) than from areas such as health education. In turn, learners from particular public health subfields who use this training system may have different educational, racial, or age profiles than workers from other public health subfields potentially not well represented in TRAIN. Outside of Kentucky, where TRAIN is used to orient all government public health workers, TRAIN’s current data pose significant issues of potential bias in terms of generalizability to the national public health workforce.

The data also have other limitations. TRAIN learners have the option of updating their demographic data, but the data set may contain outdated, duplicative, or missing information for an unknown number of learners. In most states, no routine mechanism exists for removing former health workers who have retired, died, or moved to other jobs. The Public Health Foundation specifically designed TRAIN for customization by affiliates. The data set thus lacks standardization in many elements, including whether learners list their home or work zip codes. Other than for Kentucky, in which almost all state and local public health workers use TRAIN for employment orientation, the percentage of the state public health workforce represented among registered TRAIN learners is unknown. This limitation restricts analysis to descriptions specific to TRAIN learners and limits valid extrapolation to other states or the national US workforce. TRAIN also does not record learners’ motivations for registering with the system. Thus, researchers cannot compare different populations (e.g., physicians vs lower-educated workers) on the basis of their reason for using the system for job training.

Conclusions

The 2012 TRAIN data add to our understanding of the public health workforce and the characteristics of individual health workers. Unlike data gathered at the agency level, TRAIN data are reported directly by workers. These data reinforce demographic patterns found in other studies: US public health learners are disproportionately women and White compared with the general population of the areas they serve. The median age of TRAIN-registered public health workers is 46, and one third of learners will have reached retirement age within the next 10 years.

In all but 5 states—Utah, Kansas, Arkansas, Wisconsin, and Virginia—most TRAIN learners have a college degree. These 5 states have a long history of using TRAIN, and their learners’ lower educational attainment may reflect differences between individuals who have signed up through personal interest (higher education levels) and individuals who have signed up for mandatory training (lower education levels). These educational attainment percentages, however, must be evaluated carefully in light of the low number of registered TRAIN learners in some states. For example, 1 possible bias exists. In 7 states—Washington, Nevada, North Dakota, South Dakota, Mississippi, Iowa, and New Jersey—physicians accounted for 15% to 24% of registered TRAIN users; this percentage is unusually high. In the benchmark state of Kentucky, in which almost all governmental public health workers are registered in TRAIN, the percentage of physicians is less than 5%, and the percentage of Kentucky learners who are college educated is 54%.

Converting the number of registered TRAIN learners to the number of learners per 100 000 population provides a better picture of the existing and future leveraging potential of TRAIN learners on public health. Nationally, the median is 27 TRAIN learners for every 100 000 people. In Colorado, Kansas, and Kentucky, there is 1 TRAIN learner for at least every 200 state residents. In fact, Kentucky has 1 TRAIN learner for every 80 Kentuckians. Considering that almost every governmental public health worker (current or former) in Kentucky is a registered TRAIN user, the 1:80 ratio is a useful if limited indicator of the number of public health workers per population served for this state. The limitations of the TRAIN data make the 27:100 000 national estimate unlikely to be correct. The Kentucky estimate of 1:80 is also likely inflated by the inclusion of former public health employees.

The utility of this study nevertheless rests on providing a snapshot of data from a very large number of individual public health workers. This study also demonstrates how TRAIN or similar systems can serve as enumeration and capacity tools when implemented comprehensively and even more so if supported by procedures to regularly update users’ data.