narrativemuworkgrouphealthitpolicycommittee051612 niosh by IT4921X


									To:     Health IT Policy Committee
        Meaningful Use Workgroup
        Population and Public Health Subgroup

From: Eileen Storey, MD, MPH
      Chief, Surveillance Branch
      Division of Respiratory Disease Studies
      National Institute for Occupational Safety and Health
      Centers for Disease Control and Prevention

Date:   May 16, 2012

Thank you for this opportunity to speak with you again about incorporating occupational information into
electronic health records and the opportunities that this raises for meaningful use of EHR systems. I am here on
behalf of Dr. John Howard, the Director of NIOSH. I lead the Electronic Health Working Group, a NIOSH-wide
initiative. With me on the phone is Dr. Terri Schnorr, Director of the Division of Surveillance, Hazard Evaluation,
and Field Studies (DSHEFS) and CAPT Margaret Filios, co-lead of the EHR Working Group.

When this committee considered this issue in 2010, you commented on the evident importance of work in
health and you challenged us to demonstrate that the collection of work information was feasible and could be
made useful in the EHR. We have been working hard to meet this challenge and hope that occupational
information will be incorporated as essential data in Stage 3 of Meaningful Use.

Background and Rationale

Population health depends on patient socioeconomic status, environment, medical care, genetics, and behavior.
Occupation has been shown to be a critical socioeconomic factor that impacts health outcomes through
financial security, mental state (e.g., stress), and exposure to health risks. Approximately 2/3 of the civilian,
non-institutionalized U.S. population over the age of 16 participates in the current labor force, and many others
are retired, or perform unpaid work (e.g., students, interns). Thus, much of the patient population is likely to be
or have been a worker or have a close family member who is or has been a worker (family members can be
impacted by work, too—for example, through toxins brought home on clothing). Past work and usual (longest
held) positions are as important to health and care as current positions for some conditions, such as cancer.

Work factors such as occupation (i.e., trade or job) and the industry (i.e., type of business) in which it is
performed are valuable in defining at-risk populations and describing or refining populations that require
different care, and identifying sources of disparities. Occupational health traditionally has defined populations
based on industry and occupation (I/O) and examined health risks associated with those positions; for example,
miners in the coal industry have higher rates of chronic lung disease, including coal worker’s pneumoconiosis
and chronic obstructive pulmonary disease (COPD). A less obvious but also critical application of I/O to
population health outcomes is identifying the impact of work on non-work-related conditions, e.g., closely
monitoring diabetics who may face different challenges in following a treatment regimen because of limitations
imposed by shift work.

Knowledge of work information also can guide direct care, syndromic surveillance, and public health in most
situations and settings. Extensive evidence exists to suggest that work and health cannot be separated for

                                                    Page 1 of 6
working Americans, who spend close to one-half of their waking hours at work, on average. Appreciation of
occupational risks is well known to influence patient health outcomes by contributing to diagnosis and
treatment. Syndromic observations have been used to identify new occupationally-related conditions. The lack
of work information in syndromic data presented challenges in monitoring worker population health during
clean-up of the 2010 Deepwater Horizon oil spill. In addition, there are examples of value in knowing the
patient’s work information such as I/O or, for children, the parents’ I/O, across all public health domains.

Work information also can facilitate the recognition and reduction of health disparities. Jobs with high health
risks are filled disproportionately by immigrants and racial/ethnic minorities, who also may face language
barriers, workplace discrimination, and other factors that can add to their risk of experiencing health disparities.
In 2011, NIOSH co-sponsored the First National Conference on Eliminating Health and Safety Disparities at Work
(see: The workshop highlighted
the contribution of workplace exposures to health disparities and one of the recommendations from the
conference participants was that work information be included in EHRs, “so that work and health care can be
linked and to understand population-based injuries and illnesses.” In March, 2012, NIOSH also provided the
Population Health Subcommittee of the National Committee on Vital Health Statistics with evidence that work is
an important determinant of health and health disparities. The purpose of the hearing was to gather input for
formulating recommendations to the Secretary of Health and Human Services regarding minimum standards for
collecting socioeconomic status in federal surveys; others at the meeting echoed the value of work information
and noted the desirability of harmonizing these standards with data collected in EHRs.

In 2011, NIOSH received a report from an Institute of Medicine committee appointed, at NIOSH’s request, to
examine the rationale and feasibility of incorporating occupational information in patients’ EHRs (“Incorporating
Occupational Information in Electronic Health Records: Letter Report,”
Letter-Report.aspx). The committee concluded that work information, “could contribute to fully realizing the
meaningful use of EHRs in improving individual and population health care.” It affirmed the importance of
including occupation and industry as core data in the EHR. In addition, the committee found this to be feasible
and necessary to set the platform upon which innovative work could be done to enhance information associated
with these data. For example, methods for patients to provide input into their own EHR through a Personal
Health Record (PHR) or a patient portal could expand upon specific work risks.

1. What are you working on that can help inform stage 3?

Participating in the development of standards related to work information-- NIOSH is working closely with
others to:
     Ensure that work information coding systems are accurately and fully reflected in PHIN VADS, the public
        health vocabulary service provided by CDC;
     Contribute to comments on the HL7 EHR Functional Model Release 2 and incorporate the occupational
        health domain in the HL7 Public Health Functional Profile Release 1.1;
     Include I/O in public health case reporting standards by participating in: the CSTE Case Report
        Standardization Work Group, which provided the information leading to the current HL7 Case Reporting
        Implementation Guide; the S&I Framework Public Health Case Reporting Initiative; and the development
        of CDA templates for public health case reporting as led by the Public Health Data Standards Consortium
     Update relevant Integrating the Healthcare Enterprise (IHE) profiles with work information; and
     Model the relationship between I/O (current and usual) and the risk factors relevant to other public
        health domains.

                                                    Page 2 of 6
Demonstrating how to collect and structure work information (I/O) in EHRs-- NIOSH has been collecting, coding,
and evaluating work information in health surveys and other data sources for decades. NIOSH is currently
working to counter the perception that I/O are too cumbersome to collect and incorporate in EHRs.
    NIOSH has demonstrated that self-reported occupational survey data is codable: trained coders were
       able to assign census occupation codes to 99% of the records from the NIH Multi-Ethnic Study of
       Atherosclerosis (MESA).
    NIOSH is developing the NIOSH Industry and Occupation Computer Coding System (NIOCCS): the beta-
       version of this web-based application will be available as of fall, 2012 and the first public- use version
       will be available in December, 2012. An early version of NIOCCS autocoded industry and occupation for
       87% of approximately 15,000 death certificate records with 74% accuracy and occupation for
       approximately 73% of 8200 MESA survey responses with 67% accuracy. The NIOCCS algorithms are
       being further updated to improve performance.
    NIOSH is expanding its autocoding application in two additional projects funded by a CDC Foundation
       Innovation Award: testing integration of NIOCCS with EHR software and developing and testing a
       prototype autocoding module that would be incorporated directly into EHR software, allowing EHR
       users to select standard I/O values via intelligent drop down lists.
    NIOSH has partnered with colleagues from the Massachusetts Department of Health and a primary care
       network to evaluate the capture of work information by several ambulatory care sites, including
       community health centers. NIOCCS successfully coded 44% of 27,300 records without human assistance
       or data cleaning, which is particularly remarkable given the absence of industry information which can
       inform coding of occupation. (Preliminary examples of potential impact are in #4, below.)
    NIOSH is currently undertaking two activities to more broadly ascertain the state of work information in
       EHRs: a Request for Information is being developed for publication in the Federal Register to query EHR
       users and vendors, and the Association of Occupational and Environmental Clinics (AOEC) is discussing
       the use of EHRs with its members and reporting the findings to NIOSH.
    A team has applied for intramural NIOSH funds to support pilot activities for collecting, recording, and
       viewing I/O history in EHRs.
    NIOSH is developing an information model for structuring and using work history information in EHRs.
       The model will provide guidance for areas such as time-in-job and retaining text I/O entries; we
       anticipate that it will be valuable to keep and manage I/O text fields for validation as new occupations
       appear and the way people describe their also work changes.

Developing tools to support the value of documenting work information--
    Occupational health researchers have historically developed job exposure matrices to link occupation
       and industry information to job exposures primarily for use in epidemiological studies. Utility of
       matrices are especially well researched for asthma/COPD and cancer outcomes. Example: case control
       study by NIOSH and Kaiser showing jobs that have higher exposures to diesel exhaust, irritant gases and
       vapors, mineral dust, and metal dust were associated with clinical COPD. NIOSH is developing a similar
       matrix that can be applied to a broad range of occupations in a community sample of close to 4,000.
    O*NET is a system maintained by the Department of Labor and assigns a wide range of job
       characteristics by occupation codes, including work-related physical activity levels, psychological work
       demands, and noise exposures (among many others). Recent studies show that O*NET measures can be
       used to predict health outcomes in community based samples of workers- e.g., the relationship
       between psychosocial work environment and hypertension and noise exposure and hearing loss. NIOSH
       is working to convert O*NET data into exposure indicators that could then be used to trigger clinical
       decision support messages.

                                                   Page 3 of 6
Engaging in ongoing dialogue with multiple stakeholders who have an expressed interest in and are working to
ensure inclusion of occupational information in electronic health records
            o Council of State and Territorial Epidemiologists
            o American Public Health Association
            o American College of Occupational and Environmental Medicine
            o Association of Occupational and Environmental Clinics
            o Labor Unions, AFL-CIO
            o Occupational Safety and Health Administration (OSHA)
            o Mine Safety and Health Administration (MSHA) in collaboration with the Health Resources and
                Services Administration (HRSA) and the Black Lung Clinics and rural health clinics that HRSA
            o National Center for Environmental Health (NCEH) and Agency for Toxic Substances and Disease
                Registry (ATSDR)
            o Other government agencies (e.g., VHA, NASA)

2. What barriers have you faced?

Establishing work history information as a core component of an EHR is challenging. Work information does or
should contribute to clinical decisions, syndromic surveillance, public health case reporting and surveillance, and
population health evaluations. Current public health HIT activities are structured along traditional
programmatic lines and NIOSH must both represent occupational health as a domain and serve as subject
matter experts for the inclusion of work information by every domain. S&I framework activities also are
structured such that work information subject matter experts are needed in too many places to manage.

NIOSH is working to counter perceptions that it is ‘too hard’ to structure work information in the healthcare
setting, especially I/O, and that the return-on-investment would be too low. Structuring data greatly facilitates a
population- based approach to healthcare; standardized data can be readily aggregated, compiled, and
analyzed. Structured work information will facilitate both population healthcare, by allowing this important
socioeconomic factor to be taken into consideration in monitoring and evaluating health outcomes, and direct
care, by providing the terms to link to diagnostic tools, care and treatment considerations, and return-to-work
information. While it is not expected that all care providers will become occupational health specialists, it is
beneficial to most patients, since they work, for their care providers to take into account their work information
in relation to diagnosis, treatment, and return-to-work considerations, and to know when to send them to an
occupational health specialist.

Incorporating work information in electronic health records raises concerns regarding ethics, privacy and
security. The potential for personal health information to inadvertently become a part of occupational health
records or (in some states) public workers’ compensation records and/or to be inappropriately accessed during
work-related care in settings that provide both personal and work-related care are of concern. In addition, job
security is a concern for workers who fear reprisals if they report work-related illness or injury. Finally, patients
and care providers are reluctant to engage in time-consuming struggles between personal health insurance and
workers’ compensation insurance to determine who should pay for the patient’s care. Until these concerns are
addressed, widespread and standardized documentation of work and potential work-relatedness of conditions
in EHRs is unlikely to be broadly supported. NIOSH will hold a workshop on these issues in the fall of 2012,
engaging stakeholders in a detailed discussion of the challenges and potential solutions.

                                                     Page 4 of 6
3. What infrastructure (policies, tools, training, communication) is needed to make this successful? What
strategies would you recommend to get there?

NIOSH is advocating for standards and measures that reflect the incorporation and preservation of Industry and
Occupation in EHRs in the future rules issued by CMS and ONC for Stage 3 Meaningful Use and EHR certification,
respectively. We believe that these are the best tools for ensuring that a history of work information is
incorporated and managed appropriately in EHRs such that it becomes a part of the framework of data used to
improve healthcare. We enthusiastically support the inclusion of usual, or longest-held, occupation and industry
as structured data in the draft cancer registry reporting implementation guide for eligible providers proposed in
Stage 2. And while we are working to incorporate work information, particularly I/O (current and/or usual), in
other data transport standards as well, we believe that the direct and comprehensive approach afforded by the
future CMS and ONC rules would be beneficial in assuring the presence of the data in a structured format where
it is most needed—in the EHR—to link to clinical decision support and facilitate population health activities that
fall outside of the public health arena, e.g., clinician management of at-risk diabetic populations as described

To further prepare for the incorporation of work history information in EHRs, NIOSH is drafting an information
model for the data and recognizes that there will be a need for quality measures that reflect incorporation
and/or use of the information. In preparing these items, NIOSH is making the assumption in preparing for the
incorporation of work information in EHRs that population health activities will be performed at the local level,
within the care setting or care organization, as well at the public health agency level (federal, state, territorial,
tribal, local government area)

One promising avenue for incorporating work history in EHRs that will be explored is to gather the information
via personal health records, patient portals, and/or other patient-entered data mechanisms. Occupational
medicine practices have long collected this information on self-administered questionnaires brought to the
clinical visit by the patient. It has also been shown that coding of patient-entered I/O text can be accomplished
by trained coders. NIOSH is working to provide autocoding software as a service and/or as a module that can be
incorporated into an EHR. NIOSH recognizes that it will be important also to enhance the knowledge of
providers and clinical staff engaged in the change to incorporate and use work information in EHRs. NIOSH has a
long history of providing training and educational materials for capturing and using work information in the
context of cancer registries, vital statistics, and population surveys. The project in Massachusetts suggests that
codable occupational information can be captured by clinical registration personnel with limited training.

There are a number of laws that are intended to protect patients’ rights to healthcare data privacy, including the
Health Insurance Portability and Accountability Act (HIPPA), the Genetic Information Nondiscrimination Act
(GINA), the Americans with Disabilities Act (ADA), and federal and state workers’ compensation rules and
statutes. These laws need to be carefully examined for guidance and enforcement in the new era of
electronically stored and shared health information. We anticipate that meeting the intention of these laws in
the implementation of EHRs depends more on improving guidance, policies, governance, training,
implementation, and enforcement and less on technological breakthroughs. A clearer picture of the privacy and
ethical considerations for patients that work and their family members, particularly pertaining to documentation
of work-relatedness of an illness or injury, will be available after the fall, 2012 workshop mentioned above.

Clinical decision support (CDS) tools are expected to substantially increase the value of work information in
EHRs. NIOSH plans to work with vendors and sponsors of existing data sources to link structured work
information in EHRs to resources for work exposure and risk information. We also foresee development of new

                                                     Page 5 of 6
CDS tools that build on the extensive knowledgebase of NIOSH subject matter experts and others, particularly in
support of diagnostic and care management.

4. What is the impact of these activities and the cost/economic savings?

The Bureau of Labor Statistics estimates that in 2010, over 3.9 million workers experienced a nonfatal
occupational injury or illness. Occupational injuries claimed 4,547 workers in 2010 and approximately 49,000
workers die from occupational illnesses annually. Recent calculations estimate that the annual economic
burden of occupational illness and injury is approximately $250 billion. And yet there is good reason to consider
the estimates above as gross underestimates of the numbers of work-related cases of injury and chronic disease,
although they may be more reliable for fatal injuries than for other conditions.

Incorporation of patients’ work information into electronic health records provides valuable information that
can be used to both inform patient care and provide appropriate follow-up resources to patients. During 2010,
the Massachusetts Department of Public Health (MDPH) worked with a primary care network in Cambridge, MA
to include occupation information in their EHR and train staff to collect this information from patients.
Demonstrating that work information can be collected through patient registration, current or former
occupation for more than 27,000 patients was recorded. In 2011, NIOSH, MDPH and the primary care network
entered into an agreement to evaluate the data collection, test automated coding of the occupation entries, and
assess the clinical utility of the data. NIOSH fed the occupation-only data (NIOCCS was being developed to code
linked industry and occupation data) into a beta version of NIOCCS, NIOSH’s automated industry and occupation
coding system. An early test of the health network’s data coded 44% of occupation entries automatically
(without human assistance or data cleaning). Coded data is therefore available to be queried and analyzed and
potentially linked to resources such as CDS.

Valuable information can be gained from even this preliminary analysis of occupation data from an EHR. Among
primary care network patients for whom coded occupation data is available, the most frequent job is cleaning
maid or housekeeper. Maids and housekeepers are exposed to a variety of factors that can affect their health,
among them irritants among commonly used cleaning products, which can trigger asthma and skin problems,
and ergonomic stressors. A recent NIOSH analysis found that housekeeping is one of the occupations with an
injury rate that doubles the national average. Combined with demographic information, occupation information
points out specific groups at higher risk for health effects. Maid/housekeeper occupations are especially
common among female Portuguese speakers (55%), the largest foreign language group represented in the
dataset (4496 patients). The most frequent occupation for Portuguese-speaking male workers is construction
and maintenance painters (22%). Construction and maintenance painters can be provided with information
about exposure to lead-based paint. The clinicians that initiated this project do now conduct screening for lead
exposure for these patients. Linked demographic and occupation information is enabling this primary care
network to tailor resources and information to its patient population.

                                                  Page 6 of 6

To top