Good morning, everyone, and welcome to ACOM's webinar presentation, Advancing Occupational Health Through Intolerable and Computable Data, Best Practices in Real-World Implementation of the ODH Framework in EHR. My name is Nikki Hoffman. I'm an ACOM staff member, and I will be your staff liaison today. Before we get started, I'd like to go over just a few housekeeping tips. There are two features available to communicate with the panelists and other attendees. You may post general messages in the chat feature. Messages can be shared with either of the panelists or all the participants. Use the drop-down box to select who you want to share those messages with, and go ahead and give it a try by introducing yourself to all panelists and attendees, and let us know who you are, what's your role, and where you're from, if you would like. Questions on the other hand should be submitted in the Q&A box. Panelists are monitoring this for the questions, so please be sure to post all your questions here and not into the chat box, and we'll be taking Q&A at the end of today's presentation. If you're not familiar with ACOM, we are a membership organization that promotes the health and safety of workers, workplaces, and environments through education, research, development of public policy, and advancing the field of occupational health. Before we get started, I'd just like to set a reminder that there will be a recording of today's session, and that recording will be made available for purchase in our LMS following sometime later this week. We are delighted to have Drs. Uli, Dr. Kowalski-McGraw, and Dr. Chung Bridges with us as faculty today. I'll do some brief introductions of our faculty. Dr. Catherine Chung Bridges is the Chief Community Research Officer at the Health Choice Network, focused on increasing research capacity at the federal qualified health centers. She's a board-certified family physician and health disparities researcher who earned her undergraduate degree from Harvard, her medical degree from UMDNJ New Jersey Medical School, and completed her family medicine residency at Montefiore's Residency Program in Social Medicine. Dr. Chung Bridges earned a master's in public health and epidemiology from UMDNJ New Jersey School of Public Health while completing a National Research Service Award Primary Care Research Fellowship. She has held faculty appointments at UMDNJ SPH, University of Miami, Florida International University, and has worked as a family physician as one of HCN's founding FQHSs, Jesse Trice Community Health System. Dr. Chung Bridges is HNC site PI for the Advanced Clinical Research Network, as well as PI for NIH-funded community-led health equity structural intervention, who focuses on food insecurity and foundation-funded studying, assessing, and the effectiveness of social determinants of health intervention at FQHS. Dr. Raymond Yui is a physician informationist at the National Association of Community Health Centers. Dr. Yui serves as a clinical informatics SME and a lead of multiple projects funded by the CDC and others on nationwide health information, technology policies, informatics, health data standards, data quality, clinical decision support, clinical data workflow, and machine learning. Dr. Yui supports intertribality and QI initiatives between electronic health records, electronic quality measures, and value-based care through the curation of more than 150-plus VSACs published value sets, grouping together healthcare terminology standards intended for the health data extraction, segmentation, security, and analytic for reporting to the CDC and other health agencies for syndrome surveillance. Dr. Michelle Kowalski McGraw has extensive experience in occupational medicine and multiple health systems and is duly certified in family medicine and occupational medicine. Through her practice in occupational medicine, she studies how environmental factors shape health and customizes solutions to promote wellness. As a lead physician in occupational medicine and the community care, she is working at Bridge Occupational Medicine and Primary Care. Her most recent project is development of the CDC's Occupational Date for Health, which will facilitate communication about work between patients and their providers to be used for improved individual patient care and population health. She's also a graduate at UCSD Stanford Compassionate Communication Academy Fellowship and affiliate faculty leadership since 2003. She is a fellow and an active member of the American College of Occupational and Environmental Medicine and has spoken on topics such as health informatics, fitness for duty, and with respect to use of opioid and benzomedicine medication use, prevention of bloodborne pathogen exposure, and stress management. She serves as the chair of ACOM Council on External Relationships and Communication. She pursues her passion for prevention through research, program development, education of patients, students, coworkers, and friends. And with that, I would like to welcome our faculty. Well, thanks, Nikki. And for everyone joining us today, don't forget to put your questions on the Zoom Q&A section. And welcome again. My name is Dr. Raymond Uy. There are four learning objectives today. For everyone to learn and differentiate between the CDC NIOSH's ODH framework, or Occupational Data for Health, compared to traditional employer data capture methods, which is what is happening currently right now in your EHRs, in your private practices, or whatever software you use to record your information. Next learning objective is to construct or think about different clinical scenarios where you can apply ODH data capture or ODH data elements for public health interventions, thinking about at least two data elements from ODH, which we will present to you later in this session. Next, to think about and design maybe a step-by-step workflow for integrating the capture of Occupational Data for Health data elements in your existing workflow and EHR, and thinking about considering about the potential implications and challenges and their solutions in your organizations. Last but not least, think about and evaluate the impact of current and upcoming health IT policies from the federal and state level on occupational data capture in your practices, and think about two relevant policy changes and implications. We will begin today by a great introduction from Dr. Kowalski-McGraw on ODH and the journey to USCDI. Next, I will be showing you some data standards, terminology, and case exemplars, and USCDI discussion as well. And we will wrap things off with a real-world experience and implementation in a federally qualified health centers by Dr. Chong Bridges from Health Choice Network. And I would like to welcome Dr. Kowalski-McGraw. Dr. McGraw, thank you for joining us today and telling us about ODH. Well, thank you, Ray, and thanks to the folks at NICCI and everybody at ACOM who's enabled us to bring this webinar to you today. I'm really excited to be here. When we were rehearsing just a few days ago, I was telling NICCI how much I love this group and I love this project. It's been wonderful to work for the past couple of years on this Occupational Data for Health project with Dr. Dewey and Dr. Chong Bridges. We've taken a multidisciplinary approach to this, informatics, patient care, public health. And today, we're going to share that with all of you. That is the work that we've done in optimizing how to use work information to deliver better health. So, as a physician who's practiced both family medicine and occupational medicine, it's very clear to me that we cannot effectively treat our patients without considering where they live and where they work. It's what drew me to occupational medicine in the first place and why I'm passionate about what we're talking about today, which is how we use this information to make better health, make better medicine. So, this project, this webinar came about as a result of two things. The first is this ACOM guidance paper. And the second is the experience on how the recommendations of this paper are put into action, which you will hear about shortly. So, you can see, first, the paper offers guidance on something very important that's coming up, and that is U.S. core data for interoperability. Work information is included as the core data elements, as you can see, industry and occupation. So, in this webinar, we're going to talk about how we get this information into the record in a way that makes sense and that we can actually use it for what we're trying to achieve. You can see highlighted here, that is to be able to use work information every time a patient hits the medical system to make better health, because what people do at work matters to all of their health. So as you know, social determinants of health have become increasingly recognized as something we need to consider in treating patients and populations. And until now, work has been a social determinant of health that has not been completely integrated into patient care. So the challenge is how do we integrate so many into our medical encounters? There are so many social determinants of health, which this slide illustrates, race, gender, ethnicity. But work stands out among these as one that we're under recognizing and under utilizing. I love being an OCDOC, but one of my frustrations has always been that we sometimes learn of problems only after many injuries or illnesses occur. We can intervene then, and we do, but there exists the capacity to move upstream and identify groups of workers that might be at risk for particular conditions, and apply that information to the person sitting in front of you before any injury or illness occurs. Many work injuries are seen by primary care before they even get to us, as you know. So we need to do better at leveraging what we know in occupational medicine for all medical encounters. How do we do that? Well, the EHR. And I think I can speak freely in saying that there's not a doctor that says, wow, my HR really does the trick, I just love it. No, we complain it takes time away from our patients, that it's cumbersome. But the whole beauty of this occupational data for health project that we worked on takes that into account. And the whole idea is that the information is there by the time you see the patient, it's in that record in an interoperable way. And so once it's there, we can use it for public health to identify risks, track patterns of disease transmission, track trends of illness or injury. We can provide opportunities for intervention, as I just mentioned, improving clinical care for the person sitting in front of you, identifying sentinel cases, or even protecting coworkers. How did this all come about? Well, in 2011, the NIOSH sponsored an Institute of Medicine letter report. The Institute of Medicine is now the National Academies of Science, Engineering and Medicine. And basically they looked at, what are the benefits of incorporating occupational information into the EHR? And they found many, as you can see the top five here, improve quality, safety, efficiency of care, reduce health disparities, engage patients and families, improve care coordination, improve population and public health. So they came out with 10 recommendations to advance the likelihood of incorporating occupational data for health into the general EHR while protecting privacy. And so with this, we began demonstration projects, we started developing meaningful use metrics and performance measures. We want to assess the impact of incorporating occupational information to electronic health records. And so when they were developing how to do this, we looked at three different categories, patient care, population health and public health. We really think, how's this information going to be used effectively? So patient care, very exciting. That's what drew me to this in the first place. For instance, you see a person with a rash. If you knew where they work, perhaps you're considering Lyme disease in certain areas where there are agricultural workers, prescribing a medicine. You want to know, do they drive for a living? We want to make sure those medicines are not sedating. Someone that's pregnant, we need to know where they work so we can help them to continue a safe pregnancy and their work as long as possible. We want to assist patients in successful return to work after illness or injury. And we can't do that if we don't know what work they're going back to. It gets really exciting with patient care and clinical decision support that can be developed. And these three categories are three that ACOM has worked upon previously and are available now as decision support tools. They're waiting to be integrated into the EHR with things that are coming up, which you're going to hear about. But basically the idea is asthma, something we see often in occupational medicine. How about when that person hits the emergency room or the urgent care, if there's some trigger that says, oh, this asthma began in the past two years, there's been emergency or unscheduled visits. So then you will be triggered to ask about the temporal relationship between asthma symptoms and work. Someone comes in with uncontrolled diabetes or serious hypoglycemia. You want to know, what's their work? Maybe their shift change, i.e. during the COVID pandemic with healthcare workers, you know, that threw off their diabetes management. Return to work, low back pain. You know, certainly we want to get them active, but we need to know what kind of job they're going back to. So once we know this job information, we can leverage clinical decision support. As far as population health, you want to look at groups of workers and what they might be encountering. So for example, you know, many patient information things are in Spanish, but there was a project in which they found that the housekeepers were actually Portuguese. So they tailored the communication to basically educate them on how to reduce injuries in their native language. Lead is a big one for us. Certainly we want to know, you know, groups of workers that we might need to be testing for lead or at least educating on lead. And then COVID-19 and the vaccinations. That was a lot of what you'll hear about in our project that we started was identifying those essential workers. In my state of California right now, silicosis is a big deal. So we want to know, you know, what are those groups of workers that are at risk so that when they hit the healthcare system, we can screen them for silicosis. And the last one I highlighted because you'll hear more about it later in this webinar, but that is looking at groups of workers in ICD codes. A lot of information there, you know, that we have just barely begun to tap. So that's really exciting for me. You know, we can find out so much once this information is in the record and usable. Public health. Certainly we use work information now for public health. We need to provide usual industry and occupation to cancer registries, provide current industry occupation, employer and employer address in electronic case reports. We need to identify patients with reportable conditions such as pesticide poisoning in agricultural workers. And some of this we do now, but it's not optimal. As I'm sure anyone in public health will tell you. So how do we get better at this? Well, this is really exciting. This is what's coming up, the US Core Data for Interoperability. And so the idea is all these data classes and data elements will be a requirement to have in every record so that wherever the patient goes, you'll have this information. And you can see occupation and industry are in there. So we're very excited about that. That's huge in terms of prevention, you know, so we'll hear more shortly. So with Dr. Yerwee and Dr. Chung-Bridges, I had the opportunity to be part of a pilot on this. And so for the past several years, as I've told you, we've worked through National Association of Community Healthcare Centers to see how can we use, get this information into the record and then use it effectively. And so basically the big questions that were asked were, you know, what is the patient's occupation? What is their industry? What's their employer address? And with that information, we can do a lot. So without further ado, I'm going to transfer this to my colleague, Dr. Yerwee, who will tell you the informatics aspects of all of this. Thanks so much, Michelle, for that great intro and really setting the stage for why we're talking about ODH. So let's get started. So let's talk a little bit about ODH. Let's get ahead, shall we? We're going to, hi, my name's again, Dr. Ray Bandua. You can call me Ray. We're going to add some more context on everything that Michelle just explained to everyone. This is a publication in the American Journal of Industrial Medicine earlier this year, February, 2024, which examined different data systems in occupational medicine and health. And one of the results, and you can see the highlight in the bottom left here, is that only 12 of 39 data systems were identified as collecting workload information. And even within those 12 of those 39, the information is very minimal, like only employment status, which is mainly probably an administrative information for insurance purposes. And this information is only getting gathered periodically through annual surveys or a service every two, every three years as part of the Bureau of Labor Statistics or statistical offices or oversight. So historically, it's really this inconsistent collection of work-related information, and not just in public health data systems, but also in the systems that you are using right now. It doesn't mean that you are not collecting it. It means it might be going on a note, in a SOAP note or a summary note, but as everybody knows, free text is not computable. This next diagram shows you and really drives the point that Dr. McGraw was showing us earlier, which is why work is such a social determinant or driver of health. We know that housing, education, all play a critical role in healthcare access, but there is little focus on employment status and the actual occupations that people work here in America, 30, 40 hours, even 80 hours, especially in medicine, for example. And it's these working conditions that affect not just again, access to healthcare, but also the burden of health as we now see the paradigm right now of mental health issues, chronic burnout across healthcare professionals as well. This is a paper published in the Journal of Allergy and Clinical Immunology back in 2021, which is titled Occupational Health Disparities During the Pandemic, which demonstrated a data-driven evidence that demonstrated how occupation is a prism that focuses and exacerbates occupational and non-occupational social determinants of health, both in improving or decreasing workplace conditions or safety or increasing or decreasing the incidence and prevalence of different occupational medicine risk and outcomes. We saw an increase, especially during the COVID pandemic. Thankfully, with the work we've done with NIOSH and with all our partners, including Health Choice Network, we've created different publications, an implementation guide, an action guide, and also an online course. So please reach out if there's any interest for those physicians here who are more tech savvy or more into the data, please contact us and let us know what we want to get from this presentation is really action, not action from you guys, or possibly you can identify who the data people in our organizations are, informaticists or data analysts. For those who are more into the weeds in the informatics aspect of ODH, there's a publication in the CDC website called the Guide to Collecting ODH for Health IT Systems Developers. For those who know HL7, which is a standard development organization that facilitates the secure and standard transfer of data across different electronic health record systems, health IT, HINs, or health information exchange networks, there are implementation guides which show you the technical specifications for implementing this. Don't feel overwhelmed or feel scared looking at all of these. These are mostly on the purview of your EHR vendor, but knowing that these exist and there are standards that exist to define ODH or occupational health is in your hands to really put your vendors, to make sure that their capability and they're covering this standards of exchanging occupational data. So ODH is generally divided into seven. You're seeing it here. One, employment status. All two, retirement dates. Three, jobs, and this was divided into past jobs or current jobs. Your longest held or your usual work, volunteer work, combat zone periods, and work of household members, including minors. If you check the guide, you will see different on the very first column here are the data elements. These are each specific concepts. Think about this as something that you will see in a form that we hope is in a structured form or in a field. For example, in the top left here, employment status, data type, it should be in a code, and the formal description on semantically what this data element really means or what is the operational definition. On the top right, so that's employment status. On the top right here, you're seeing occupation, description, industry, industry description. On the bottom left, it's basically what is your occupation? What is the description of your occupation? What industry are you in this occupation in? And what is the description? And you're seeing that industry and occupation here has a code data type. You're going to see occupation and industry across different general categories, including in longest held work, voluntary work, work of household members. Basically, this data element exists across all of the major categories. And we'll see more of it later. So US CDI, why do we keep talking about this? There really is no incentive or really, really low incentive for vendors to really interoperate and share their data with other vendors due to data being, as everybody says in our conferences, data is the new oil. And these data can be trained to create new AI models, new machine learning models. So if you're a vendor and you have all of this data that you have access to, you prize that and make sure that your competitors don't have it. But that's the bottom line. But the bottom line that we have as clinicians, as healthcare professionals, is to help our patients and to make sure that we practice medicine in a way that we use technology ethically and use it to its fullest potential. And that's what US CDI is about. This is from the 21st century, Kirshad. This is from a federal standpoint. It's really a common data model, a combination of data elements and data classes required for all certified EHR technologies in the United States. And you're seeing version three here because version three is the one that is going to be required. So again, the core principles here are minimum data elements and classes that are going to be, again, required, mandatory for all certified EHRs, Epic, Cerner, Athena, you name it, they're all certified EHR products in states to make these data elements interoperable, which means they are adhering to a standard, just like how your cell phone, you charge it with your standard USB type C or a standard plug in your electronic devices at home. In the same way, we want our health data to be interoperable and EHR agnostic as well. The 21st Century Cures Act, in a nutshell, really was supporting all of these things, but it was not mandatory in the beginning. They were saying it encourages, again, the number one here, encourages adoption of standards. Noting here that it just encourages it, and there's an attestation to support the standards, but it's not mandatory. Number two, again, EHRs meeting specific standards to support interoperability, but also it's not as explicit in doing and adhering to these standards. Similarly, now that we have patient portals, patients are accessing their information using APIs or application programming interfaces, but not right now. This is all open and available. As in the regulatory standpoint, ER, EHRs are supposed to have what's called the FHIR API, F-H-I-R, API open, about a year and a half ago. And now I present to you the HDI-1 final rule. The purpose of this is to expand the 21st Century Cures Act to make and help data to be more interoperable. And this HDI-1 final rule, again, this is already in effect, also ensures different things, such as EHR reporting to public health agencies, ensuring racial equity, supporting underserved communities through the federal government, through, again, interoperable data and using data standards for representing race and ethnicity. And also, this is one of the first things that they did when artificial intelligence or AI machine learning and generative AI has been the buzzword everywhere. So this is one of the first steps to ensure that. It says on the final rule, again, it's the final rule, it's now in effect, is to adopt USCDI version three as the baseline, set baseline across all certified products. And I will show you what has happened these past few years. We started with USCDI version one at top left. It doesn't have occupation, it doesn't have a lot of data elements. This is currently happening now, but it was not mandatory for version one. Version two is not mandatory either, but you're seeing I highlighted a few things here. We added social drivers of health, goals, assessments, interventions, all that stuff. On USCDI version three, you're seeing on your screen on the bottom left here, occupation and occupation industry is part of USCDI version three, which means before January 1st, 2026, you all certified health IT technologies in the United States need to represent occupation and occupation industry using a standard. What is that standard exactly? If you open the guide and the version three, you will see that ODH or Occupational Data for Health is the mandatory and required standard to represent occupation and occupation industry. What's in occupation anyway? So if you look at the submission on USCDI, you will see that on the top right here highlighted in red, in the red box, representing industry requires a standard coding system called NAICS, N-A-I-C-S, or the North American Industry Classification System. This is being maintained by the Bureau of Labor Statistics and that these are the value sets or groups of codes required to represent industry, which means if you're in the healthcare industry, there's one and the one single only NAICS code or N-A-I-C-S code to represent that. Similarly, within industries, again, industry and occupations are not mutually exclusive. These are pairs. So occupation and occupation industry are paired together. So when you are representing occupation, you're not going to be using the NAICS codes. You will be using what's called SOC codes or SOC codes. This is being also being stewarded by ONET, which means if you are an occupational therapist, there is one and one single SOC code to represent you being an occupational therapist. And here in the bottom, next, employment status also has a specific code. In this case, the question of asking employment status or the concept of employment status is also represented by one and one single code only, a line code, L-O-I-N-C code 74165-2, which means all EHR products in the United States, if they are capturing employment status, they should be in the back end, the things you're not seeing that's behind the screen, that information is mapped or linked to this one line code so that any researcher, any occupational medicine researcher in the United States can do a SQL query or a question or a database extract using 74165-2. And the idea is, using that one single code, they will be able to get all of the employment status in their databases. And again, if adhering to these rules, you can reuse those SQL queries across multiple databases, so you don't have to look at and figure out where the employment status question is. And this is the same concept as you see in industry and occupation, you're seeing the line code on occupation and industry, the same rules apply. This is a good slide to show you what's contained in these value sets or how the coding happens in the back end. Employment status in this case, for example, the code is employed, not in labor force, unemployed, again, very straightforward, they're not coded. And the bottom left here are different work classifications. Again, these are not currently standard, but this is part of the ODH standard for paid work, but in the armed forces would be PWAF. And that's what we're trying to do, we're standardizing the coding in the back end. On the right hand side, you're seeing an example of SOC codes, again, SOC codes for occupation, SO, the O is in the middle, occupation. And you're seeing these digits on the right hand side, again, from top to bottom, .NET developer, .NET programmer, all the way down to Abalone fishermen have one and single codes there. And you can go as granular as you can, there's some granularity in SOC and NAICS codes, so I welcome you guys to explore and check those out. Some examples of what this has been used, if you recall, during the pandemic, there were essential workers, part of our project was to use NAICS and SOC codes and use those codes to identify essential workers, because the Department of Homeland Security released a cross map showing which SOC codes are those occupations that should be identified as essential. And we can do this, again, on hopefully not the next pandemic, we will be able to use these crosswalks to immediately identify those essential workers for those events. This is a screenshot of Epic showing you a drop down specifically for occupation, this was an example with an organization called Ochin, this is Ochin Epic, and in this case, there's a number here in the back end, this is mapped to an SOC code. So if they select here, travel agent or tire builder or therapist, there will be a code attached to it on DidExtract. The challenges as we've seen, I'll be quick here, trust. When you ask someone what's your job and they're your patient, they'll be, why are you asking me this question? This is none of your business, or they don't want to share it with you due to different reasons. And that's one of the barriers that we saw. Privacy concerns, again, people generally don't want to tell them their occupations, there are migrant communities also that are wary about law enforcement getting into trouble with those things. There's also language barriers, literacy issues, data collection, as you know, we are not in the health profession to collect data, we're here to give care. And they saw that the more stuff we ask in the front or more administrative data we ask, we know that it is data intensive, but all of this can be automated. On the right hand side, training, registration, lack of multilingual occupation list, this actually exists in Spanish, other questions at least. There's some staff or front desk confusion, transitions in EHRs during the project, we had partners who were transitioning from one EHR to another, which delayed some implementation. Some things we learned during our implementation was, we really need to enhance how we ask this question to patients to make sure we're up front, saying that this information is private, and we're using this for your care, nothing else. Streamlining the data collection, you don't have to ask the patient what their occupation industry is, you can ask them to fill this up themselves while they're in the waiting room or in the patient portal. And that's what's actually recommended in the guide. And last but not least, training the staff to make sure they understand why we're collecting this information and to what end. Leveraging all these technologies, enhancing privacy measures, and being culturally sensitive to patient populations who may be wary about sharing this information. Some lessons learned again, from again our project with NIOSH and our partners, optimizing EHRs, accessibility in telling the folks what their occupations are, and doing Kaizen or continuous improvement. We're celebrating our success stories in these implementations. There has been increased improvement and consistent data and collection engagement across the three years with our partners. And you will see one or two of those with our next speaker. And really the point here is to do real world data, which means this is not a research project. This is not a once in a year survey. This is using the information you are capturing from your patients, their occupations, and making sure we can use again clinical decision support algorithms to deliver care. Or especially now during climate, you know, with inclement climate or heat related illness, and all of these illnesses that we know are from occupations or environments, we can reach out to make sure we reach those folks during those times of crisis. We're truly learning from the past here, if you remember. And during the 1700s, chimney sweepers related to scrotal cancer, coal workers and pneumoclonicosis, radium dial painters and radiation poisoning in the early 20th century, textiles and vesonosis, asbestosis, rubber workers and bladder cancer. In the 60s, we know about DDT and farm workers, radiation related cancers in those working in the nuclear industry. And nowadays in the late 20th century, firefighters and healthcare workers, especially during the pandemic. It's just a few more papers on all of these, and these are quite recent. We know that while OSHA, for example, wants to collect information on work-related injury, these are really exclusive to work-related incidents and fatalities and not necessarily things that we care for, like cardiovascular disease and all that. And again, highlighting the fact that work is really infrequently queried or documented due to how it's being captured. And this is known throughout all of these publications, and they're all saying the same thing, that adoption of clinical decision support tools requires standardization of occupational data, and that's what we have achieved. Thankfully, here we have 2017, this publication in the Journal of American Medical Association, representation of occupational information across resources and validating ODH. And it says almost 10 years ago, they were hoping that the growing adoption of EHRs, that there would be a possibility of EHR certification criterias, including occupational information. And here we are, 2026, January 1st, the deadline, we got occupation and occupation industry, at least in U.S. CDI version 3 through HDI 1. So great success. What's the future here? There's the HDI 2 proposed rule from the ONC, which increases the baseline version of U.S. CDI to version, I believe, 4 here. Unfortunately, employment status has not reached acceptance in U.S. CDI, but we are pushing every year to get employment status, and then eventually, hopefully, the rest of ODH. But right now, at least we have occupation using SOC codes and industry using NAICS codes, N-A-I-C-S codes. Thank you. And now you will listen to Dr. Katherine Chung Bridges for her work with her health center using ODH. Hello, greetings, and thank you for this opportunity to speak with you today about this wonderful work that we've been engaged in for several years. This is the real-world experience implementation and use in a federally qualified health center. So I know we've, the other speakers have already talked to you, but I sort of want to set the stage. Why is this important? Occupation is a key social determinant of health, corresponds to several different dynamics of socioeconomic status, especially according to a classification by Weber. It corresponds to class, which is the economic component, status, which is the social prestige component, and power, which is the ability to influence others. The example I like to give is looking at a school teacher and somebody who collects a garbage collector, for example, in a particular city, you might find that the teacher makes less than the garbage collector. So the garbage collector would have greater level in the class spectrum, but in the status spectrum, we'd expect the teacher would, you know, might have greater status. Power can be dynamic and change depending on the situation. Most times we're able to collect information on class through income and factors such as that, but the occupation sort of is a package that gives us a lot more information about an individual and several different dynamics of their socioeconomic status. It determines their income, the neighborhood they live in, and therefore the environmental exposures that they themselves and their family are exposed to, health insurance access to care, physical and psychological exposures to environment over time. We know how much time we all spend on our jobs. So any exposure, whether physical or psychological, is going to have a significant impact on our health and the health of our family, and as I've been mentioning, it really affects the entire family. Really good conceptual framework that sort of puts it all together, all the different components, the retirement benefits, the income, sick leave, vacation, health insurance, community resources, all of these are afforded due to the type of work that an individual does. I wanted to take a minute to set the stage by providing some evidence on Latinx communities and the impact of work and the ways that we see some of the occupational disparities, especially in our Latinx communities. So this graph shows you that Latinx workers are underrepresented in managerial and professional jobs. When we look at all U.S. workers compared to Latinx workers, we see 42.4% represented in managerial positions versus 24.5% of Latinx workers. And the converse, that Latino men are overrepresented in construction and maintenance occupations, and again, underrepresented in the managerial and professional roles. So as you can see, construction and maintenance occupations, there's about 27.9% representation versus 16.3% in all U.S. men. Latina women are more likely to be employed in service occupations than other women workers in the United States. So looking at this orange, darker orange, 29.7% representation among Latin, Latina women versus 19.6% in all U.S. women. This is a nice nuance to look at, something that we, you know, might not think about, having advanced notice of your work schedule and how that would impact on your health and your ability to seek care for any illness or even prevention. And we see a lot more Latinx workers having less than one week notice about their work schedules relative to non-Latinx workers. This is a very sobering slide showing you that Latinx workers have the highest rates of occupational fatalities in the United States compared to other racial and ethnic groups and are overrepresented in some of the occupations with the highest rates of workplace illnesses and injuries. So some of these occupations, construction and laborers, maintenance and repair workers, there's greater representation there. And also industries that have the highest rates of minimum wage violations, we see greater representation or a significant representation among our Latinx workers in private households, food and service industries, and agriculture. So now switching to our particular project, which was Labor of Love, where we were able to implement occupation and industry data collection at a particular Federally Qualified Health Center, a particular FQHC, a Federally Qualified Health Center. Federally Qualified Health Centers serve one in four racial ethnic minorities, one in six Medicaid beneficiaries, one in five uninsured, and one in three persons living in poverty. This slide is to demonstrate that this comes from NAC. That health centers care for a population that is sicker than the general population. A greater proportion of health center patients rate their health as fair or poor, have diabetes, asthma, high cholesterol, and hypertension. The focus of our intervention or implementation was the amazing La Casa Family Health Center in Portales, New Mexico. This is an amazing health center that does pretty significant work in the community that they serve. In 2023, 76% of their population identified as a member of a racial ethnic minority group. They were 71% Hispanic or Latinx, 24% non-Hispanic white, only 3% Black or African American, and 31% preferring a language other than English. They had 68% income with income below the poverty level in 2023, and 92% with incomes below two times the poverty level. 19% were uninsured and 50% Medicaid or CHIP beneficiaries. So again, this project was with support from a National Association of Community Health Centers, NAC, as well as the CDC. And our goals were to implement the collection of occupational data for health at a particular HCN member, FQHC. And as mentioned, we worked with La Casa Family Health Center and had a great partnership with their chief medical officer, Dr. Gidel Tom. So what were the steps that we took? We started off by reviewing the previous methods of capturing occupation-related data, developed a plan of EHR optimization and implementation. We worked internally with HCN staff on EHR optimization, trained the La Casa staff on the new workflow, went live, did a continuous process of reviewing and analyzing data, doing feedback and improving what we were doing. Lots of conversations, lots of conversations with staff about why we were doing what we were doing, the staff having lots of conversations with patients about why the information was being collected. Data analysis was reviewed with La Casa staff, and the health center took action steps to address disparities that were identified. So when we were establishing the main data elements that needed to be captured, the questions that were looked at were, are you currently working? And what is your employment company name or industry? And what is your job, your occupation? As mentioned, there was constant review of the process. The system optimization was done, and an action plan was put into place. I should mention that this particular health center during the course of the project changed electronic health records, which was also a challenge, but it meant that this process kind of happened twice. But it was a very interesting process that we learned from each time we went through it. The health center used a train-the-trainer approach where we provided training to the health center staff on how to collect the data, the occupation and industry data, and then the staff were trained by the leaders whom we trained. So we provided live training and materials. And this was a continuous process throughout the project. So there was a main training that happened in the beginning, but then throughout the project, we, you know, were in constant communication with the staff and needing to provide more information as we went along. This just shows you sort of the cycle of feasibility, optimization, implementation, quality checks, improvement opportunities, and then starting that cycle all over again. In the beginning, the EHR was not collecting the data. So Dr. Oye talked to you about the SOC and the NAICS coding, which were what were implemented. In one particular system, we had to have a higher level of these codes due to limitations of the EHR. And in the newer EHR, we were able to drill down to a smaller level. There was a new employment status that was added not in the labor force to be distinguished from unemployed. Employer information was enabled. A new page was added to capture the industry and occupation. And as mentioned, the classification was a little broader in the beginning. And the new EHR allowed for searchable data. FQHC main challenge was really everybody understanding the why. I think that in any of this work that we do around increasing health equity, eliminating health disparities, it's so critical and important for the staff, you know, who are the main ones having these conversations with the patients, to understand the reason for collecting this data. It's another thing we're asking them to do in the long list of things that they're already doing. So the why is very important and communicating similar to the way we've been communicating to you, the importance of occupation as a social determinant of health, how important it is as a clinician, how important it is for clinicians to know occupation and industry of the patient that they're taking care of. All of those messages were provided over and over again with our staff and also shared with patients as well when they asked the question, why do you want to know this, you know, why are you asking me this question? We found that the enrollment specialists have a trusting staff-patient relationship, but we always want to take advantage of those existing trusting relationships that patients have within the health center. FQHCs tend to have really great relationships with the patients that they serve, and we use that in this project to help us to, you know, gain the trust, provide the information and the education and increase the collection of the data that we needed. In one situation, there was a cheat sheet that was developed by the health center of common local occupations, industries and employers that tried to help speed up the process, sort of, you know, hey, we spent the time to figure out this particular occupation and industry. It's a common one in our area. So that sheet sort of referenced those categories for others to use and to facilitate their collection of data. Challenges, patients may be hesitant or unwilling to provide information due to privacy or lack of understanding. So we talked a little bit about gaining their trust and engaging and educating. Privacy is a factor, so it's going to take some extra time to spend to educate people. It also means that while self-collection of the data is ultimately the goal, that there might be some patients for whom we still need to sit down with them to collect the data. So it's not always a, you know, one size fits all solution for every patient. Accurately collecting the information takes a long time. It takes resources of staff and time, and so that has to be dedicated in order to achieve the goal that we wanted to achieve. And addressing these challenges just requires continued communication, privacy protections, streamlined data collection processes and making sure that we're always culturally sensitive as we're collecting information. So more about education, which we've kind of talked through that had to be done continuously. Quickly wanted to share with you some work that we did around dairy and cattle farm workers at LA CASA. We had some initial conversations with the staff and the leadership there to try to understand what some of the concerns were from an occupational health standpoint, and they brought up the dairy and cattle farm workers as a category that there was concern about respiratory illnesses and challenges among that population. Sure enough, when we looked at the data, we found among 63 patients that 49% had a diagnosis of respiratory tract and or asthma issues. And so LA CASA did onsite COVID-19 vaccination clinic for the dairy cattle farm workers. Vaccines were offered between work shifts and family members were invited to the health center for vaccines. But the fact that we were able to use this occupation industry category to identify a health need, it was there, and that the health center was able to take some actions to address the health was a great example of the power of this type of data for the individual delivery of care, but for population-based health care, which we like to provide. Really quickly, the last year that we were participating in this project, we had 94% of patients reporting their employment status, of whom 36% were currently employed. And of those currently employed individuals, 81% reported their occupation and 74% reported their industry, which we think are really great numbers. 78% reported their employer's name and company. So just powerful data that's there. Dr. Oye and Dr. Kowalski-McGraw talked to you about how all of this data is going to be collected similarly from other sites. So imagine the power of combining this data to be able to understand what's going on with people at a population level. That's the power of the data. Finally, I'd like to acknowledge the LA CASA Family Health Center team, which were amazing. Dr. Gidel Tom, the Chief Medical Officer, Yvonne Armijo, Director of Operations, Amy Ruiz has been amazing, Tammy Jones, the Practice Manager. At HCN, we could not have done this work without Daniel Paras, our Research Data Scientist, and Lizzie Endemano, our Research Operations Manager. And of course, NAC and CDC, we are indebted to you because this has been a labor of love. We've enjoyed every minute of this work, and it's been a joy. So thank you. And with that, I'm going to turn it back to Ms. Hoffman. Thank you. So at this time, we're going to go ahead and take questions. If anybody has any questions, you can submit those in the Q&A box. I believe we do have a few that have come in. Dr. Uy, do you want to take those, or do you want me to help moderate those for you? Yes, you can help moderate. So our first question, let's see here. Does need... Is there a code for stay-at-home mothers and fathers? There is, in fact, a code for that. It's not... I have to check the SOC code, but I believe it's in the ODH data model. It is going to be under employment status, which is homemaker. So not necessarily employed, but that's a great question. Very, very relevant. Thank you. So I can take over, Nikki. There's another question here, Michelle and Catherine, with large language models in AI. Do we need to spend time coding this data now? And in five years, will we be able to extract this from free text and other unstructured data? Arguably, this is possible now. I have an answer to this. I'll quickly answer this. Oh, gosh. We want to be optimistic that having a natural language processing application running live in the back end of an EHR is the dream that we've been dreaming of in the past decade or two decades. I fear that continuing to collect this in free text may not be a solution. As you know, you can do this with other text, but the sensitivity of healthcare data makes it difficult to maybe do the coding or to extract. There are actually AI tools by NIOSH that does AI coding using the free text. It's called the NIOCS Autocoder. N-I-O-C-C-S, please Google that, and it will be done. There is always no need to not capture this in a structured form or field because that seems to be, how do you say this, easier. So, Occam's razor. I wanted to just comment on that, too. I would totally echo what Dr. Uy said about, yes, that is the hope and the dream to make these things work to our advantage, but we need to tell these things that we're creating how to use the data, too. So, that's not understood right now by these AI models. So, that's part of the Occupational Data for Health Project, is to really look at all the ramifications of this. Where do we put it? Once the computer does recognize it, where is it going to store it? How is it going to connect with other things? But to answer your question, yes, as Dr. Uy said, the NOCS coding is working on that. We're looking at all this, but we can't just rely on this to happen. We have to continue to actively make it happen the way we want it to happen. I agree. And Catherine, I have a two-part question directed to you. I think you're the best person to answer this. Dr. Chung Bridges, isn't a cheat sheet of job titles and codes risk coding bias when you were implementing this? How do we ensure the data quality? And on top of that, Dr. Chung, were healthcare workers and providers included in the collected data and results presented in the graphic? Beautiful presentation, by the way. Thank you. So those two questions. Catherine, what do you think? Yep. So I don't think that the cheat sheet caused bias because, you know, the point of the cheat sheet was that if a person fit those particular categories, then those categories would be used. There were particular employers that were very large in the area, particular dairy farms, particular, you know, other industries that were, you know, had big companies that employed a lot of people. And so those sheets, you know, the cheat sheet was just used for those types of situations where somebody worked at the same company, maybe in the same position. That's when that was used. I do think that validation is important. The exercise of going back and, you know, talking to individual patients who have a classification that's been put. And, you know, maybe later on coming back and seeing if that really measures up to what they truly reflect as their occupation and industry. So validation is always very important. And the second part of the question was, I think, whether or not the staff were involved. And absolutely, they were involved in that process of reviewing the data. They were the ones who gave us the information that sort of led us down that road of looking at the dairy farmers and trying to see the risk of respiratory illnesses among those patients. And that information was fed back to the staff, reviewed with the staff. And for that reason, the intervention that they created around a clinic for vaccinations and all of that was developed by the staff there at the health center. So I hope I answered all the questions. Yes. Thank you, Dr. Trombridges. There are some questions here I've tried to answer live on the application. Fire accelerator funding is an issue. NIOSH is really working on it. I know there are fire accelerators for social drivers of health, including the Gravity Project. And they're looking at unemployment and work instability as well. So I'm involved in the working group with Gravity. So yes, we hope there's a fire accelerator that can support this. The other one is the tenure. Tenure is actually part of the full ODH framework. So if you look at the developer guide, you will see it there. Is there any ability to tie occupational data with industry hygiene measure data? Example, pulling noise or air quality samples from employment job site. I want to answer this live because we have two prospective grants in the Gulf states, NAC in particular, where we will be crossing GIS air quality index measures with health center data and looking at respiratory illness related to environmental factors such as possibly flooding, forest fires, extreme heat, for example. So yes, 100%, this can be done. But for this to be done, using the standard codes is absolutely necessary. Again, standards sound boring, but standards drive all of this work, fortunately or unfortunately. In occupational data collect, is there any idea on the hazard involved in the occupation of that patient so we can connect the disease of our patient to the hazard exposed? Michelle, any insight on this? This is a good one. Yes, so in occupational data for health, Dr. Uri had mentioned the SOC and the NAICS codes. Well, it's actually more than that. There's ONET tied in with the SOC. And so within, if you look at SOC and ONET together, then you can get an idea of which occupations have which hazards. So the idea, again, how we use this data that I brought up earlier is very, very important. Not only just basically what they do, but we want to tie it to exactly what you described. What are the hazards? What are the things we need to look for? Even potentially the employer data, which I think was brought up in the previous question. But like Dr. Uri said in his presentation, the employers aren't just going to give this data. They don't know what we're going to do with it, that we're going to use it somehow in a non-deidentified way. So that's, again, where the informatics becomes key, because we want to get it in there in a safe way that can help populations and not cause any problems. Yes. Everyone, thank you for sticking with us. I know we are five minutes over, but we're happy to keep on, happy to stay. But for those who need to go to their next meeting, please go ahead. I want to answer this one question really quickly. There was a question on, will there be a code for retired? So the occupation prior to retirement will be on the records. There is, in fact, a retired concept. And not just retired. This is retired with the concept of employed post-retirement. As you know, doctors don't like retiring up until 75, 80. They want to do locum or something else. So there is retired with post-retirement employment, which is retired from a usual occupation and not working, retired but working for pay, retired but volunteering, or completely retired. So there is, in fact, value sets and data elements for that specific one. So thank you for asking that. Oh, GEMS. Yes. Thank you, Anil, for bringing this up. The GEMS for occupation. And thank you for linking that publication. We really appreciate that. And I've read that. So with another answer for the hazard that Dr. Kowalski-McGraw actually answered, there is the whole concept of having a domain-specific common data model drives these hazards. And I've seen these hazards in SNOMED CT. So for those who don't know, SNOMED CT is an ontology, not like ICD-10, not like the other systems that we use for billing. SNOMED CT is actually an ontology that shows relationships. And there is actually in that SNOMED CT ontology, again, this is free in the United States. This is a standard. There is an ontology there that shows relationships, different hazards that affect different occupations. So there is a data model. And if you guys are interested, we can work together and create this common data model for hazards surrounding and linking it to NAICS and SOC codes. That would be absolutely important. And I can see that being used a lot in clinical decision support. So if you have any residents in research and occupational medicine, this can be their dissertation or this can be their final project. Who knows? But that's such a great idea. So final remarks, Catherine, Michelle. Oh, do you talk to patients about how their own medical issues may affect their job ability to work safely? Don't we do that all the time, Michelle? Yeah, absolutely. You know, again, this is where you would have the database to back you up on this. So in occupational medicine, we do that all the time. But as I mentioned, a lot of these things come to primary care and they don't always have the resources to be able to do that. So that's key to this project is get those resources to them so that it's available. Thank you. I can add really quickly, a lot of the work of the clinician, as we know, is connecting the dots for patients, right? So, you know, we know clearly the connection between occupation and health, kind of bi-directional impact there. But being able to point that out to patients and help them to realize how their health impacts their ability to work and their work impacts on their health, that's a lot of the job of the clinician and the staff to educate on that connection. Yes, absolutely. I see a lot of people who would like to collaborate. Thank you. We're all in a data-driven world. And this is how we get occupational medicine, OEM data out there to all of the great researchers and, you know, drive the policy and drive really automating all of the stuff that we can automate. And that's what I think Anil mentioned on the chat with using an auto-coder, automated coding for job descriptions. Again, this is a really cool tool. I suggest looking at it. But what we would like is for our actual vendors to look at the papers and publish it and use it. That's what we want. That's why we're spreading awareness. And I think the biggest takeaway for everyone, and I think Michelle and Catherine will agree with me on this, is if this is not your cup of tea, if you're not an informaticist, find out who's doing this in your organization and start the conversation because it's these conversations that drive innovation and drive actual change. Again, it's going to be mandatory before January 1st, but we don't know if they're actually going to be effective or actually use those mandatory codes. But it's up to you guys to really drive this innovation so that we don't have to do surveys and don't have to do, we can use real-world evidence in the EHRs to drive OEM research. And that is my dream, and I hope everybody shares it. So, Nikki? Thank you. Thank you all for, a big, huge thank you to all of our faculty for today's presentation with us and ACOM. To our attendees, thank you for joining us. As a reminder, there will be an evaluation, and that will be required in order to claim your CME. So that should pop up in another window upon closing the webinar. If not, we will be having a follow-up email this afternoon with those instructions. So, again, thank you so much for joining us all today, and I hope you all have a great day. Thank you, everyone. Happy Monday. Thanks, Michelle. Thanks, Catherine. Thank you.

Occupational Health

ODH Framework

Electronic Health Records

EHR Integration

Social Determinants of Health

US Core Data for Interoperability

NAICS

SOC Codes

Public Health Interventions

Clinical Decision Support

Healthcare Standardization