A Texas cardiology practice is seeing success with analytics software that unifies clinical and claims data.
Cardiologists are turning to analytics technology to improve care management for their patients, and in doing so are building a better business platform that allows them to stay independent and steer clear of hospital acquisition.
Eduardo Hernandez-Vila, MD, president of the Houston-based Texas Heart Medical Group, oversees 12 cardiologists and three heart surgeons (and growing). The group is owned by the Texas Heart Institute, founded in 1962 by world-renowned cardiovascular surgeon Denton A. Cooley.
In 1984, Cooley established Cardiovascular Care Providers (CVCP), a network of more than 500 cardiovascular physicians, including the Texas Heart Medical Group. CVCP is using software from Cedar Gate Technologies to focus on value-based care reimbursement.
The technology provides a unified view of patient data as gathered from hospitals, other providers, payers, and wearables, in a system purpose-built to facilitate value-based care.
CVCP providers have been able to report very good outcomes to their cardiovascular care and surgeries at a reasonable cost, says Hernandez-Vila, who’s now serving as CVCP’s president. Employers, who contracted with CVCP members, reward the network by referring their patients to these institutions.
Eduardo Hernandez-Vila, president of the Texas Heart Medical Group and Cardiovascular Care Providers. Photo courtesy Texas Heart Medical Group.
"It provided predictability for the [employer] institution," Hernandez-Vila says. "This concept of offering predictability not only included the surgical fee and the hospital fee, it included also the physician cardiologist who's taking care of the patients."
"If you're a really good surgeon, and you have good post-operative care, then risk is going to be mitigated by providing the best and safest care to patients, which I think has led to the success of CVCP," he says.
"One way that we've managed to provide value to that program is by trying to keep almost all of the services are performed in the outpatient setting, with prevention," Hernandez-Vila says.
Under bundled payment arrangements, CVCP operates a risk pool that manages behavior.
"This aligns the interests of the providers with the fixed payments, with a focus on budget," Hernandez-Vila says. "That helps to mitigate the fee-for-service mentality. It teaches the delivery system to look ahead, and to focus on reducing variability in costs."
For example, a patient may say they have a family history of vascular disease.
"We'll do imaging and discover they have a carotid stenosis they would not have known they had had they not had easy access to participate in this program," Hernandez-Vila says. "That patient was just put on lifestyle measures, as well as aggressive medical therapy to control their cholesterol and their blood pressure. That patient will likely never be seen in a hospital setting because they're treated aggressively as an outpatient."
More than 97% of the services provided in the Shell center of excellence are provided in the outpatient setting, he says, freeing up bandwidth in affiliated hospital facilities to take care of sicker patients.
This program boasts one third the inpatient services utilization compared to the industry in general, a low mortality rate, and very low readmission rates, Hernandez-Vila says.
"Almost nil in patients that are enrolled in these programs," he points out. "That's one of the easiest examples of how an integrated system like our works to provide value for the patient, and for the company that participates in this program."
Although the technology is designed to offer complete data on each patient, some challenges persist.
"We're getting regular feeds from our bundled payment claims processing engine, and we're currently in the process of creating ways to modify how our risk pool is distributed to physicians, based on the output of these metrics," Hernandez-Vila says.
If a patient comes in for an open-heart surgery, develops a post-operative infection or septicemia and ends up on a ventilator for a month, the bundle set aside for payment no longer exists because it assumes the patient leaves the hospital in a week, he says.
"The excess care for that patient is going to come from that risk pool," Hernandez-Vila says. "So the entity has every interest in creating every measure that you can think of to limit infection risk and to provide good postoperative management so the patient is in the hospital for the shortest period of time that is safe for them to be there."
Most hospital care is provided by hospital-based physicians or hospitalists, he says.
"One of the great advantages of CVCP is that it has allowed physicians to remain independent from employment models that frequently doctors can get burned out in," he says. "You're being rewarded on production, not necessarily only on quality."
Conversely, that focus on cost control has allowed the Texas Heart Medical Group to grow significantly in the Houston area in the past year, Hernandez-Vila says.
Since 2020, Cedar Gate's technology has used appropriate-care measures algorithms developed by renowned Johns Hopkins University physician Marty Makary, MD.
"A lot of our [CVCP] practices are entering into Epic's EHR system," Hernandez-Vila says. "It's going to make it easier for us to input clinical data into the … platforms to have a comprehensive view not only of claims-based data, but also specific patient-level data that will help them refine their risk and how well doctors are performing in the network."
The Texas Heart Medical Group is also joining the Pinnacle Registry, the American College of Cardiology's outpatient quality improvement registry, capturing data on coronary artery disease, hypertension, heart failure, and atrial fibrillation.
"A lot of value-based care programs can fizzle out if the physicians are not engaged in it, or if the risk and reward is not there," Hernandez-Vila says. "Having an entity that's been present since 1984 in the value-based care space, and is still strong and growing, is something I think that should be replicated outside of the Houston area and outside of the state of Texas."
A new study out of the Regenstrief Institute finds doctors critical of the high volume of texts, while nurses are OK with it.
As hospitals replace pagers with smartphone-based texting platforms connecting doctors, nurses, and other clinicians, some are giving mixed reviews to the technology.
So says a new study from the Regenstrief Institute, published in the journal Applied Clinical Informatics, which found a lack of shared understanding among clinicians on how to use clinical texting.
"Clinical texting is a double-edged sword – it’s easy to contact fellow clinicians, which can be viewed as good or bad," says Joy L. Lee, PhD, M.S., a Regenstrief Institute scientist, assistant professor of medicine at the Indiana University School of Medicine and corresponding author of the study. "Each hospital or hospital system needs to figure out how to use clinical texting to optimize communication, workflow and patient care and develop use guidelines."
In the study, nurses and hospitalists supported clinical texting, noting ease of access and the ability to transmit photos and create a record of communications. But they were also frustrated with implementation challenges and a lack of understanding about the appropriateness of texts.
Doctors in particular criticized the high volume of texts, though nurses did not.
With clinical texting guidelines and etiquette, Lee says, one size may not fit all.
The study’s authors created a checklist of considerations to take full advantage of the benefits of clinical texting, diminish user dissatisfaction, and create shared understanding of use. They recommend creating team buy-in and developing user consensus on aspects of appropriateness, such as when it would or would not be appropriate to text a non-urgent message.
The study, “Qualitative Analysis of Team Communication with a Clinical Texting System at a Midwestern Academic Hospital,” is a collaborative effort of health services researchers, clinicians and hospital administrators.
In addition to Lee, it was conducted by Monica Huffman, BS, Marianne Matthias, PhD, April Savoy, PhD, and Michael Weiner, MD, MPH, all of the Regenstrief Institute; and Areeba Kara, MD, Bethany Radecki, BSN, MSN, and Jason T. Schaffer, MD, all of IU Health.
A Maryland physician is using speech dictation software to translate notes from a medical encounter into the patient record, reducing his time on a computer and enriching the visit for both him and his patient.
A family physician in Maryland is improving his documentation (and spending less time doing it) with technology that combats physician burnout.
"The most important thing is you can dictate anywhere you want," such as via a mobile phone, he says.
Suki is one of several technology solutions designed to improve clinician workflows by translating notes into the EHR. He says he’s used other platforms in the past, including Dragon with Epic, and likes Suki for its ability to help providers who use English as a second language.
In the year that he has used Suki with athenahealth's EHR, he says, he’s been able to facilitate 80 patient encounters per week.
"I dictate for at least 95% of my patients," he says.
Anuj Bhatnagar, MD. Photo courtesy Frederick Primary Care Associates.
Lately, the industry has begun paying attention to newer conversational AI technology, such as that made by Nuance, which also makes the previous generation of Dragon dictation software. In the newer version, software captures the entire conversation between physician and patient, then generates SOAP notes that are reviewed by both Nuance and the physicians themselves.
These ambient technologies enable clinicians to make more eye contact between with patients, rather than focusing on their computers, thus enriching the experience for both provider and patient. DAX queues up orders it generates for physician approval after the encounter.
He notes that with the approval of most of his patients, he uses patient encounter time to enter those orders, including e-prescriptions and lab tests, with mouse clicks.
"When I'm talking to a patient, I'm ordering as I'm talking," he says. "I never tend to dictate when I'm with a patient. It just takes that attention away."
"It's a mixed bag,” he adds, “but I believe in giving the control to the patient, saying, ‘Hey, would you mind if I just have my ears and not my eyes towards you?’"
Some patients prefer the eye contact instead, he notes.
He also likes the ability to invoke Suki from within different parts of the medical record, rather than relying upon the software to make sense of where in the medical record parts of the conversation are intended to go.
He says he has also trained Suki to recognize athenahealth smart texts or text macros, pre-populated portions of notes, to expedite note building in his EHR.
"Even with Suki, it took some time," he says. "I think it was learning my accent. But in the last eight months, I realized [Suki] is just catching it pretty easy."
Some macros he built himself, while others are provided by athenanet, an online service for which athenahealth requires a subscription.
The cost of Suki, he says, is more than made up for in the addition of seeing between one and two more additional patients per day.
He notes that dictation works best immediately after the encounter.
"If we saw a patient and we dictate right away, the juiciness of the note is way more than after two or three days," he says. "After the patient leaves, I stay back in the patient room, take my two minutes or so to dictate, [then] come out and see the next patient.”
"Honestly speaking, I never bring work home," he adds. "I finish all my full 20 patients a day [on] the same day."
By having those records completed on the same day, the patient’s medical record is more quickly up to date, and can be access by another physician if that patient sees someone else.
"My colleague knows what my thought process was," he says. "I think that seals the deal for us in this case."
AAFP has used Suki technology in its Innovation Laboratory to study the adoption, use and impact of an AI assistant for documentation by primary care physicians. According to an AAFP study, participants in family medicine and other primary care specialties saw a 72% reduction in their median documentation time per note. Lab participants reported a calculated time savings of 3.3 hours per week per clinician and improved satisfaction with their workload and in their practice. Full results and analysis are available in a white paper on the pilot.
Phase 2 of the lab studied the adoption and impact of an AI assistant used for visit note completion for 30 days by over 132 family physicians and primary care clinicians. Four out of ten lab participants who tried the AI assistant for documentation adopted the solution, which was assessed based on the number of participants agreeing to purchase the technology, and the impact realized during the lab trial. The impact was assessed through a quantitative assessment of documentation time.
Vital signs sticker aims to detect early signs of trouble, including infection, in bone marrow transplant recipients.
A Colorado health system is testing the feasibility of remote patient monitoring for the early detection of febrile neutropenia, a common, life-threatening complication of cancer therapy that is typically treated as an oncologic emergency.
Ten bone-marrow transplant recipients will initially participate in the trial at the University of Colorado Anschutz Medical Campus, also known as CU Anschutz. The phased approach, through a series of studies, will scale the trial up over time to 100 participants, overseen by an institutional review board, and will include the use of predictive analytics, telemedicine, portable imaging, and supportive therapies such as antibiotics and hydration via IV.
The trial, overseen by CU oncologists, will include round-the-clock oversight from a team of oncology-certified advanced practice providers via a virtual care center operated by Reimagine Care, a company specialized in providing virtual care to cancer patients.
Febrile neutropenia leads to major complications in 25% to 30% of patients, and may lead to death in approximately 10% of patients. Time to antibiotic administration has been independently associated with mortality, and each delay in starting antibiotics can increase the risk of 28-day mortality by 18%.
At the core of the trial is an RPM system consisting of an FDA-cleared BioSticker made by BioIntelliSense. The sticker, which adheres to the patient's chest, collects a number of vital signs, including skin temperature, and is able through algorithms to monitor core body temperature, according to Aaron Gerber, MD, chief executive officer and co-founder of Reimagine Care.
Continuous monitoring of body temperature can detect weak signals or trends in ways that traditional temperature-taking, even in hospitals, can find challenging, according to Clay Smith, MD, director of the Blood Disorders and Cell Therapies Center and medical director of UCHealth and CU Innovation Centers.
The protocol also acknowledges that typical body temperature can vary from patient to patient.
Clay Smith, MD, director of the Blood Disorders and Cell Therapies Center and medical director at UCHealth and CU Innovation Centers. Photo courtesy of UCHealth.
"If a patient is above a certain number relative to their baseline, we have a rules engine we set up working with Dr. Smith, and it triggers an alert," Gerber says. "Then our nurse practitioners can figure out how we want to proceed from there."
Historically, oncologists encouraged patients to buy thermometers and call their oncologist's office if they feel bad, or they notice their temperature going up, Smith says.
"It's so very much time for improvement," Smith says. "One thing we've learned is that temperature, like virtually everything, is very different from one person to the next. The notion that there is a line you can draw on a thermometer, and that works for everybody, we've known for some time is just wrong. Everybody's different. Temperature can also change if you just drink a cold drink or have a hat on, or if your temperature tends to run low or high, or you're on some medicate that changes that."
The new system represents a different, more personalized approach to detecting early infection, Smith says.
"What really excites me is bringing all of the different data from these remote patient monitors together, and developing new machine learning and other AI based algorithms to say, well, it's not just temperature, it's your heart rate's going up, your breathing is getting a bit faster, you're not as steady on your feet," he says.
The new system is also expected to do a better job of detecting dehydration, or are not eating well, either of which can weaken patients, Smith says.
With this enhanced detection, the hope is to "nip [infections] in the bud before something really bad has happened," he says.
The initial study will determine if the data is accurate and reproducible, Smith says. "Then what we hope to do is incrementally, in a fairly rapid and intuitive series of studies, to build upon that, and start to add more and more questions and tests of functionality," he says.
Over time, the hope is to change the historical model to move much of the evaluation and care of patients from the emergency room to the home as much as possible, Smith says.
"Our mantra here is patients come first," Smith says. "I also think it can be a great approach to improving access, improving quality of life, improving healthcare disparities. We do a lot of very high-end complicated types of treatments like bone marrow transplants, where people have to travel from sometimes a thousand miles away, because we're the only academic medical center for the Rocky Mountains."
The initial clinical trial will require participants to reside in the Denver metro area for the duration of the study, within 45 minutes of the academic medical center. But the intent is to be able ultimately to have participants throughout the region, Smith says.
Participants are asked to have reliable telephone and home internet service and a stable wireless network. The BioIntelliSense sticker communicates with an app on patients' mobile phones via Bluetooth, Gerber says. As a backup medium, the sticker is also capable of communicating with clinicians via an associated cellular-based BioHub, which will also be provided to patients, Gerber says.
The Ohio State University Wexner Medical Center hasn’t seen an uptick in cyberattacks as a result of the Russian invasion of Ukraine just yet, but that and other threats are keeping security experts on their toes these days
The Russian invasion of Ukraine has heightened concerns about cyberattacks worldwide, including in the US, where the Cybersecurity & Infrastructure Security Agency issued a Shields Up warning to the private sector, including healthcare organizations.
In this conversation with HealthLeaders, Phyllis Teater, chief information officer at The Ohio State University Wexner Medical Center, discusses best practices and steps that healthcare organizations can take to strengthen their cybersecurity strategies. This interview has been edited lightly for clarity and brevity.
HealthLeaders: What's your assessment of the mood of cybersecurity officers such as yourself, given that we're more than two months into the war and everyone remains on edge about maybe there's a bigger cyber threat just around the corner?
Phyllis Teater: I think that's certainly something that's concerning. We watch. We're fortunate. I'm the CIO here, not the CISO, but our IT security group watches a lot of markers to try to understand if we're seeing more threats, such as, do we see a significant uptake in the number of spam or denial of service attacks? Are we seeing phishing? It's [things] like that, that we are currently blocking. But of course, things can get through. And we have not seen a significant difference in the number of things that are trying to get in. But we are trying to be diligent.
Phyllis Teater, chief information officer at The Ohio State University Wexner Medical Center. Photo courtesy The Ohio State University Wexner Medical Center.
Healthcare has been such a big target through the pandemic. We made a lot of investments to try to really continue to address our risks, knowing that there is no perfect protection. I feel like this is just another thing that is continuing to increase the risk. We've had so many scares and media attention about the attacks on healthcare, particularly because we have been pretty busy the last two years with our business. And folks are distracted and tired, and so their proclivity to fall for an attack of some kind through our user base has certainly been higher. We've continued to be diligent and know that many of the changes in the world mean that the attacks could get more and more, but we have not seen any indication of a major uptick in the threats, in particular, from anywhere.
HL: To what extent do you think the government ought to strengthen security requirements for critical infrastructure such as hospitals and healthcare?
Teater: My opinion of this is probably biased as a CIO who has to oversee all of the investments. I think that the limitation is not the requirements or the regulations we should be following. It is all about resources, both people and monetary resources, to invest in all of the things that we should be doing. I mean, I could have a security department that [comprises] 100 people. I don't. You can invest a limitless amount of money, and you still may have a breach, because there is no perfect protection. I don't know that strengthening the requirements will do anything but feel like an unfunded mandate for any of the critical infrastructure industries.
HealthLeaders: Should the federal government give more cybersecurity support to enterprises such as yourself? For things like training? Should there be a government tax credit?
Teater: I think that that would certainly help. The most important defense is training and helping [employees]. At OSU we call it the human firewall. To be able to make sure that they understand the risks, the linkages between the things that they do, and keeping our organization safe. We think about the hierarchy of technology to be protect -- keeping the bad actors out, of course, detect -- that's making sure that you have mechanisms to understand if they did get a foothold somewhere, contain -- that's making sure if a bad actor should get in that you have protections for them being able to spread and then restore -- when you do have something that's been compromised, how fast can you turn around getting that back in a secure way?
We look at that four-part hierarchy in terms of our security posture and the initiatives that we have. Protect is so important for our humans that work here to help us with, to make sure that they are cognizant about the risks that can approach them, and how their actions can affect the institution. So I do think investing in training and being able to do some of those things and be able to have some help for organizations to build strong training for programs is really important. But of course, a tax credit doesn't help us at all, because we're not for profit.
HL: The other issue you keep hearing about is these breaches where security updates were not applied in a timely fashion. Do you think that that remains one of the big vulnerabilities, just unpatched systems?
Teater: I think so. And again, that comes back to the resources issue and having the resources to do that. [With] most IT departments, and I'll just talk about ours at the medical center, our primary focus is supporting the business of our organization. And when we think about the literally thousands of patch notices and updates, everything from critical to low, that we get every day, it's a pretty formidable thing for any organization of any size. But especially frankly, [with] large organizations, we have thousands and thousands of servers, as does every organization of any size. All different operating systems, different levels, different manufacturers, and frankly, for any organization, to feel like they're keeping up super well is a huge challenge, because of the resources it takes.
And when you think about the work that we need to do, to make our organization secure, and you think about zero trust, those kinds of things, we have to balance our ability to do business and our ability for us to take care of patients in a timely fashion. We could do all sorts of things to make our environment more secure. That would make it difficult for us to take care of patients or to do the business side of our work. There is always a balance between making your environment more secure to reduce risk, and impacting your users in such a way that it makes their job harder. Those almost never go in the same direction.
Now there are some technology advancements that have kind of made it there. Two-factor authentication, which is such a huge piece of any defense against cyberattacks, is now pretty standard. And pretty darn easy to use. When I log into our, say, HR system, every time I log in, I can have my notices go right to my Apple Watch, right? And it buzzes on my wrist. And I say approve and I'm in. And that is a big piece of our protection. So I do think there's opportunity for technology advancements that can help our security posture and not be too intrusive with our users. But that is not everything.
And we have a lot of security mechanisms at the medical center that they do not require at the university here at OSU. And that is because of our patient information and the HIPAA requirements that we have. And we hear from users regularly, why is it so hard over here at the medical center? And because we have much stronger controls that we've had to put in place.
HL: How important are the principles of zero trust, to improving enterprise security and healthcare?
Teater: It provides a very valuable framework to be thinking about, not a utopia, but a vision for how we could operate, knowing again that absolute zero trust would mean, nobody's logging onto your system. So you're sure they're fine. So it comes back to the impact to the users for zero trust environment. Has the technology advanced enough, that we can really do a full zero trust without driving our users crazy?
HL: How is the cyber insurance industry doing? Some say it's just facilitating ransomware payouts.
Teater: I think it's a relative mystery. I'm not sure that the people that need to buy the insurance or the people that are offering insurance have really figured it out. That's my sort of personal opinion. It is helpful to have the insurance if you have a major breach. Not every breach is ransomware. There are other types of breaches that still happen all the time. But I think the ransomware world is very unique, and of course, very prevalent. Many organizations have significant deductibles on their cyber insurance. Thinking about how that plays into it also within the insurance is in many situations, but not all.
HL: The security climate also makes IT leaders consider cloud alternatives to on-premises data centers more than you may have in the past.
Teater: The way I think about it, and encourage our organization to think about it, is simply a different set of security issues, some better and some not better. If you have an application that's in the cloud, that provider can have a security issue or a breach. And if they do, the capability of the organization that is using that service, to impact the restore process is very limited, and you are dependent on somebody else's best practices. You try to do due diligence, of course, when you are obtaining that cloud service and understand the process, but it is just a different set of security issues. And I think it's still quite an impact to any organization, because it's about the business continuity. It's about not having access to that system. And it can happen with a cloud provider the same way it can happen with something on prem.
HL: Are there other best practices that have come to the forefront this year, as you further try to respond to this threat?
Teater: Continuing to up your game with training, which we already talked about, and really having a program to train your organization, a regular ongoing program, about how you do self-phishing to try to identify those users that are perhaps high risk and give them additional training. And so lots of innovation, I think in that space, although that's not really technological.
It's more about social engineering innovation to be able to really hone your practices for the areas that are the highest risks. Many organizations are investing in EDRs - endpoint detection and remediation. And, and I do think that that endpoint protection and detection is so important, because that is often the way that many, many vectors can get in. They're starting to really, I think, advance and mature and be able to help organizations really contain as well as protect their assets.
HL: Any other concerns?
Teater: It is a huge challenge for all industries in their technology space to have the talent, the amount of talent we really need working on these problems. And so, hoping for advancements in degrees that are offered in cybersecurity and protections and in training programs, to be able to really build that talent base, is going to be so important to our future.
In the second of a two-part interview, Micky Tripathi points to resources to optimize best practices, but says the ONC could do more.
Micky Tripathi, PhD, MPP, took the helm of the Department of Health and Human Services Office of the National Coordinator for Health Information Technology on the first day of the Biden administration in 2021. In a sequel to his two-part conversation with HealthLeaders a year ago, Tripathi addresses ONC's top priorities. Part 1, published yesterday, explored top ONC priorities including data governance, data sharing, security, and privacy. This interview that has been lightly edited for brevity and clarity.
HealthLeaders: Some healthcare IT leaders continue to express concern about the total cost of ownership of digital health technology going on back a decade. It's as if the government helped fund the purchase of cars, but did not budget for funding for the fuel for the cars. Enterprises say they don't have the resources necessary to necessarily drive home the potential of this technology. What can the government do to help?
Micky Tripathi: I'm not sure I agree with the analogy. Every industry has to move forward, and I think there's a tremendous amount of value that's been gotten already from these systems as well. With anything, and particularly with technology, the costs are very, very clear and discrete -- what you pay eClinicalWorks or Epic or Cerner every month -- but the benefits are very diffuse. So your CFO doesn't see on their ledger, here is the return on investment that I got from that, even though there is a return in terms of quality, in terms of safety, and better documentation. So I disagree with the premise that that analogy is right. I don't think that there's a choice in that matter.
There are a variety of federal programs that provide additional benefits. For example, participation in value-based purchasing, where there's upside potential for organizations that participate in those. You can't successfully participate in those if you don't have an electronic health record. And there's no way that you could do the kind of quality measurement, the kind of care coordination, the kind of care management that you're required to do to participate in those programs without the benefit of those systems. So that's where there's a lot of benefit. And I think we're going to start to see more and more.
Micky Tripathi, head of the Health and Human Services Department's Office of the National Coordinator for Health Information Technology. Photo Courtesy ONC.
We talked about algorithms. I have a little bit of a black box, and we need to worry about, are there unintended consequences for what's going on in there? There is a huge upside potential to what algorithms offer as well, when you think about the burden of too many clicks. Algorithms can work in the background, to try to make better sense of non-structured data to allow physicians to dictate in the ways that they used to be able to dictate into systems through natural language processing, so that they're not having to type all that information in. Those are all things that are tremendous benefits on the algorithm side, but also are part of the benefit side of the ledger, as we think about what electronic health records and health IT in general provides to people.
HL: Although to be fair, if you're adding elements like conversational natural language processing, that also adds to the cost side of the equation as well. It doesn't come with the EHR.
Tripathi: No, it doesn't. As data varies by the system, how much of that is an extra feature? Like any other investment decision, every organization has got to make the decision of, do I want to hire a scribe, who will do that kind of data entry for me? Or do I want to invest in that technology? Or do I want to place that responsibility on the provider? And these are all sort of the regular automation versus labor intensity kinds of decisions that every organization has to make.
HL: The ONC and the government are best known for writing and implementing regulations. How can the government move beyond this, to offer more in the way of workflows, best practices, recipes for success, educational curriculum, and sample policies. Some say adding these would improve clinical workflow in ways that similar measures do for aviation by going beyond regulations.
Tripathi: Yeah, it's really an interesting perspective. We would love more feedback on where we can do that more. We try to do some of that more, like we have the Health IT Playbook, for example. It's on our website, but that kind of walks through different types of capabilities and gets a little bit into workflow sorts of related things. We do a ton in the way of learning collaboratives and webinars and all of that, but if we're not doing enough, or if there are areas where we can do more, we are absolutely happy to talk to anyone who wants to help us with that.
We work with the AMA and others to develop joint educational materials and resources as well. The challenge with things like workflows is that once you get to a certain level of detail, it really varies by what's your specialty, and what EHR are you on. It's really hard to be generic, or to the extent that you are generic, people feel like that's just motherhood and apple pie, and doesn't apply to what they're trying to do.
One area that I think is an area that we're starting to look harder at, though, and since you mentioned aviation, is safety. Are there things where we can drill into more with respect to surfacing safety issues that might be common across EHRs, for example, and be able to surface those and make those available to others for them to learn and be able to make sure that they design certain things out of their systems? If that has popped up in one system, it seems like it could be an issue in other systems as well. And that's very much in line with the way that the aviation industry handles safety issues. Regardless of whether you're on Boeing or Airbus, or regardless of which avionics system you're using, being able to surface those issues so that others can learn from them, that's an area that we're starting to look at for sure.
In a two-part interview, Micky Tripathi, the head of the Health and Human Services Department's Office of the National Coordinator for Health Information Technology, talks about his department's top priorities and other pressing issues.
Micky Tripathi, PhD, MPP, took the helm of the Department of Health and Human Services’ Office of the National Coordinator for Health Information Technology on the first day of the Biden administration in 2021. In a sequel to his two-part conversation with HealthLeaders a year ago, he addresses the ONC's top priorities, as well as health IT issues such as total cost of ownership and going beyond rule-making to help healthcare systems optimize their investment.
This interview has been lightly edited for brevity and clarity.
HealthLeaders: What you see right now as the top technology challenges facing U.S. healthcare?
Micky Tripathi: Certainly security is going to be top of mind for everyone. It's a huge and ongoing challenge. I don't know that that's any different than it should be at any given time. But as we have more and more electronic health record penetration and more interoperability, certainly the security issues related to that rise equally. So that's one very big challenge that we have from a technical perspective.
The other is having appropriate governance around data sharing. I know that's not necessarily a technology issue, but the minute you start to share information, you start to have issues related to data quality, data governance, data use, all of those kinds of things. And then last is privacy. Privacy is a very big deal. And certainly, as we think about the intersection of privacy and technology in healthcare, you start to confront the issues related to information crossing the HIPAA boundary into the hands of individuals and the lack of general privacy law protecting that information once it's in their hands.
HL: How is the US government addressing these challenges?
Tripathi: With security, it's a kind of a multi-pronged thing. From an ONC perspective, we have our certification role, and certification requirements that relate to the basic security processes and technologies that are required for certified health IT systems. And that is very much coupled with the HIPAA security rule, which has a whole bunch of requirements related to what HIPAA-regulated entities are required to put in place, things like encryption. So that works very much hand in glove. And other parts of the department, like the CIO's office, also have been active in trying to inform the healthcare delivery system and healthcare providers about the need for security protections in their environments.
At the end of the day, every place that is managing IT, the security is only as good as the policies that they implement, and their diligence around those policies and technologies. That's the bigger challenge: That we live in a very fragmented healthcare delivery system, and there's nothing that the federal government can do to say, let's just turn this crank, or flip this switch, and everyone will be secure. It's really about constant diligence, constant awareness, and making sure that there's alignment and awareness of security issues.
Micky Tripathi, head of the Health and Human Services Department's Office of the National Coordinator for Health Information Technology. Photo Courtesy ONC.
With respect to data use, we [have] things like TEFCA, the Trust Exchange Framework and Common Agreement. The hope for that is that you start to have greater commonality and transparency around the rules for sharing information, and a common understanding of a basic set of ground rules for the sharing of information. And that helps to lead to more responsible and more appropriate use. We're also starting to look at issues like algorithms, and what role they play as we think about data use, and certainly everything we're doing with respect to health equity, related to being able to have the capture of that data in better ways, and how do we think about appropriate and responsible use as well.
So [those are] areas that we need to explore and get more into, but we're certainly looking at those as well. TEFCA allows individuals to access information through apps, and those apps are required, contractually. If they're going to want to provide services on TEFCA to individuals, they're required to essentially meet the requirements of the HIPAA privacy and security rule, where as a general matter, they're not required to, because they aren't regulated entities. What we've done is tried to say, well, that if TEFCA is offering a value that they want to participate in, then contractually, they would be required to meet those requirements as well. So in the absence of regulations and policies over which we have no direct authority, we're trying to say, let's do everything we can, to get more of those kinds of protections in the interests of individuals.
HL: Is there a bigger policy role that the federal government should play in looking at equity through the lens of, we have this tremendous investment in AI, through all these private actors, but fundamentally, it seems like a very much succession of black boxes? We don't know how these APIs work. The algorithms are proprietary. So we end up with equity questions that we don't have easy answers to, regarding, is this algorithm endangering a particular population?
Tripathi: Yeah, that's some of the things that we're looking at. There is an interagency group that's been launched to look specifically at AI and equity that's across the US government. And that's just getting started. But the idea there is to start to catalog and capture what's going on across the different federal agencies. What are the different authorities that those agencies have, and how do we want to think about this from a federal government perspective?
Different agencies, the NIH and FDA, have published different sets of guidelines, rules, and principles related to ethical uses of AI. But to your point, a lot of it is still proprietary. There's limits to what you can do there. But we are taking a look at that for sure, across federal agencies and within the federal government. Within ONC, we are looking at it through the lens of electronic health records, which are increasingly the source of data, and increasingly, the ways that AI results get injected into decision making. So we're certainly taking a look at that.
Editor's Note: Part 2 of the HealthLeaders interviewwith ONC Chief Micky Tripathi can be accessed here.
EHR data reveals that 1 in 5 Indiana county pregnancies are not screened for syphilis
The analysis of electronic health records is one of public health's newest tools to reduce preventable diseases.
A recent study from the Regenstrief Institute and Indiana University, analyzing EHR data, found that one in five women in an Indiana county do not receive a test for syphilis during pregnancy. The Centers for Disease Control and Prevention (CDC) recommends screening all pregnant women for syphilis during their first prenatal visit.
The study was conducted by a research team looking at live births in Marion County, Indiana, from 2014 through 2016. Researchers were able to link birth certificates from the health department with maternal health data from EHRs in the Indiana Network for Patient Care, created by the Regenstrief Institute and managed by the Indiana Health Information Exchange.
During National Public Health Week last week, HealthLeaders spoke to study co-author Brian Dixon, PhD, MPA, about the study and the state of public health in 2022. Dixon is the director of public health informatics at the Regenstrief Institute and the Indiana University Fairbanks School of Public Health. This interview has been edited for brevity and clarity.
HealthLeaders: Syphilis hasn't been the most talked-about disease for some time.
Brian Dixon: We saw STD rates go up during the pandemic, and as we come out of the pandemic. During the pandemic, we put a lot of focus on COVID And we took people out of STD divisions and other divisions of public health in order to help with the response to COVID. So we now have to focus again on things like diabetes and syphilis.
Syphilis is a is a challenge to the infant, but also could be problematic for the mother. We have a real problem of maternal morbidity in this country. This being Public Health Week, we want to encourage people to think beyond COVID, and think about their health and doing things that they need to do to screen for early detection. We want to identify syphilis early in the pregnancy. We have effective treatments for it. But we've got to detect it in order to treat it.
HL: Will you continue to study it and look at years subsequent to 2016 and broader cohorts?
Dixon: Yes, this is not the end of our research. We are continuing to work with CDC on this type of work. We are currently looking at treatment of gonorrhea and chlamydia. This time, we're going to try to link to some datasets that are at the state levels, not just Marion County, but look at statewide data. As you can imagine, we have to jump through a few hoops to get that state data in order to do our linkage. But we're in the process of doing that now. The new project will be used looking at data all the way through 2021, to look at not just diagnosis of it, but also the treatment.
HL: Is this kind of study common in public health? Or do you feel it represents a kind of a new wave of population health analysis?
Dixon: The research question we posed is in line with traditional public health research. Are our clinicians following guidelines? Are women or any other population getting preventative health screenings at recommended rates? However, I do think that our research does represent a growing trend that is still relatively new: taking advantage of electronic health records and multiple electronic data sources in a way where they are integrated.
Brian Dixon, director of public health informatics, Regenstrief Institute and Indiana University Fairbanks School of Public Health. Photo courtesy Regienstrief Institute and IUFSPH.
We still largely in this country have a problem of siloed systems. We still have an HIV system and other STDs separated. We still have public health largely separated from clinical data sets. So in that regard, this is a new type of research, and one that I think we will increasingly see as, especially post-pandemic, a real focus now in the public health world on connecting clinical data sources and public health data sources. We need to have things at the health department like a master person index. We need to be able to link across datasets in a way that we've never been able to do before. I still think today it's pretty rare.
HL: What percentage of birth certificates, even nationwide, is electronic or digital today? Does anybody know?
Dixon: That's a great question. I'm not sure anybody knows that answer.
HL: Are EHRs essentially a proxy for digital birth certificates, or is that stretching it?
Dixon: That's stretching it. We still largely in the US rely on birth registrars that are assigned at each hospital to record data in a separate system. Birth registrars do use the electronic medical record system to get information that they need to fill out the birth certificate. But in some other research that we've done, we've found that those individuals have to access three or four different sorts of health IT systems within the hospital to get all the information they need. So it's not all in one place, even within the hospital.
Birth registrars do the hard work of extracting the information from multiple places and putting it into the birth certificate system. A high proportion of birth certificates these days are being entered into an electronic system that is maintained at the health department. But it's still largely a manual process. Then you have the challenge, of course, of getting data from non-hospital settings into that birth registry. But there has been a really strong, consistent effort by the CDC and state and local health departments over the last decade to move to electronic birth records. And so most places have that habit.
HL: You mentioned that there's several sources around the hospital that these birth registrars have to refer to. You mentioned EHRs. What are the others?
Dixon: Typically, the OB practices are private, so they're affiliated with the hospital, but the EHR in their clinic is totally separate from the hospital. And they typically do not talk to one another. There's very little interoperability between OB EHRs and hospital EHRs. That's why when a woman shows up in a hospital, she's often retested for things because they don't have access to the OB's EHR electronically. And if the woman came into the emergency department first and then transfered to the OB or the maternity ward, then [birth registrars] may have to access the ED information system as well. And then also, if there are complications with their pregnancy, then they may have to go into a separate imaging system or PACS system or something to look at data as well.
HL: That has implications also for other public health priorities such as vaccination registries, correct?
Dixon: Yes, it does. Vaccine registries are their own silo. That's a big challenge in public health, and each state has its own siloed system. Let's say you give birth to your child in Minnesota, and then you move a few years later to Wisconsin. Your child gets the next series of vaccines there. Those are in two separate immunization registries. In clinics, when a pediatrician gives a vaccine, they do enter the information into the state's immunization registry electronically. That is pretty well established these days. But the immunization registry, for example, is not connected at all to the birth certificate data typically in a state.
HL: Post-pandemic, are we seeing a renewed commitment to funding for public health? Or is it too early to say?
Dixon: There is a renewed commitment to it at the federal level. I'm not sure that it's going to persist at the state level. That's part of the challenge. There is a bill working its way through Congress now for additional public health funding. Now we need to put money into recovery, because it's going to take some communities a very long time to recover. Technically, states are responsible for the health of their citizens, and so we need state legislatures to provide sustainable funding for state and local public health. That's going to be more of a challenge, because states coming out of the pandemic are not necessarily putting more money into public health.
People now are way more aware of what public health does than they did before the pandemic. Before, public health was kind of invisible to many people. Now we realize the important work that they do in communities. There's this commitment at the federal level. I'd like to see a similar commitment at the state level, so that when we get to 2023, we don't go back to pre-pandemic funding levels.
Hospitals can now use magnetic technology to improve the accuracy and speed of the procedures.
Patients living with breast cancer often face long, gut-wrenching wait times for surgeries (during the pandemic that wait could last 22 weeks), not to mention invasive and not-always-accurate testing before the procedure to determine where to operate and how much breast tissue to remove. New technology is now helping to reduce those delays, while at the same time improving the accuracy and clinical outcomes.
"With the advent of more and more detailed mammography, we're finding lesions that are super, super small – cancers that are really tiny that we cannot feel," says Courtney Fulton, MD, a breast surgical oncologist and general surgeon at Community Hospital in Grand Junction, Colorado.
Such super-small cancers are typically found using localization, a procedure used to mark the breast before surgery to guide the surgeon in locating the abnormality, Fulton says.
"The traditional way to do that, the day of surgery, is to stick a tiny stainless steel wire, with a hook on the end of it, into the patient's breast, around the lesion and sticking out of their skin," she says.
This method has "a whole lot of problems," Fulton says. "Mostly that the patients are then sitting waiting for the surgery with a wire sticking out of them." There’s also a problem with accuracy, she says, leading to unnecessary surgeries, removal of too much healthy breast tissue, and the possibility of missing cancer cells.
Courtney Fulton, MD, breast surgical oncologist and general surgeon at Community Hospital. Photo courtesy Community Hospital.
Add to that the dangers caused by waiting. According to a 2020 BMJ study, delays in breast cancer treatments could lead to 5,100 excess deaths in the U.S., fueling the push for innovative ways to reduce those delays.
A new generation of localization devices is improving accuracy and providing more convenience to patients, Fulton says. They’re wire-free and can be implanted in patients up to 30 days in advance.
An example of this new technology is the wire-free localization technology platform developed by MOLLI Surgical and approved by the US Food & Drug Administration earlier this year. The non-radioactive and solid magnet technology measures 3.2 mm in size, and is designed to allow both radiologists and surgeons to detect the smallest lesions. The accompanying wand allows Fulton to get into small spaces and take only the tissue she needs to remove.
"It's super small," she says. "We're removing lesions that are a few millimeters in size sometime. You don't want a localization device that's 10 times the size of the thing that you're removing."
"It's about advancing the science of getting breast cancers out of patients quicker and easier for the patients and the surgeons," she adds.
Some localization alternatives rely upon radioactive seeds instead of magnetic ones, which have their own problems.
"You have to be very diligent about tracking radioactive substances around the hospital, and it's kind of a pain in terms of paperwork," Fulton says.
Magnetic devices like MOLLI are sensitive to any other surgical instruments that themselves are magnetic.
"We had some issues with our sterilizer that was making our instruments too magnetic, so we revamped our demagnetizing process," Fulton says. "Now I test all my instruments at the beginning of the case, to make sure they won't interfere, and then I'm good to go. We've got it pretty much down to a good science now."
The only patients for whom this technology wouldn’t be used are patients with a severe allergy to nickel. The device also isn’t needed if the lesion is palpable, though it can also help surgeons find other sites, such as lymph nodes.
Fulton has treated 30 patients with the technology in less than a year. Because it can be implanted up to 30 days ahead of time, it can be inserted when patients come into the clinic for preoperative bloodwork, not necessarily requiring an extra visit.
While it’s hard to measure how this technology improves accuracy, Fulton says she feels better about her ability to get all the cancer. Still, "even with all the technology and everything, you still have microscopic cancer at the edge of what you take out," she adds.
About half of Fulton's practice is cancerous and benign breast surgery. Melanoma skin cancers are another focus of hers, as well as regular general surgery including hernias and gall bladder surgeries.
"In the breast community, everyone's very excited about this kind of technology," she says. "There's still tons and tons of hospitals using wire localization, the old-fashioned way."
As the pandemic wanes and curbs on in-person treatment end, Fulton is getting busier. She’s seeing a backlog not only because there’s a limited number of healthcare providers like her able to do these procedures, but also because people are reluctant to get mammograms.
"This month, I was very, very busy with breast cancers," she says. "I was seeing multiple breast cancers a day in [the] clinic and we were trying to get them all into surgery really fast, because these patients have been waiting. So we're trying to get everyone scheduled quickly. [The new technology] definitely helps with that."
In this HealthLeaders Q&A, MediReport co-founder and CEO Serge Makowski, MD, an interventional cardiologist, explains how unified cardiovascular information systems (CVIS) are ushering in a new era in cardiology reporting. This interview has been lightly edited for clarity and brevity.
HealthLeaders: How does this guidance conform to best practices in clinical documentation overall? EHRs, and the standards defined by meaningful use, have specified discrete data input fields that feed into quality measures being used to implement value-based care. Do structured narratives conform to these meaningful use guidelines?
Serge Makowski: The adoption of structured reporting in the cardiovascular space is being recommended with the goal of improving the quality of cardiovascular reporting and patient care. The use of dictation and free text makes it difficult for physicians to search for and locate valuable information and procedure findings retroactively. In this regard, structured reporting can bolster improved quality of care, as it ensures that clinical data such as patient history, past treatments, findings, and procedure results are standardized and easily retrievable for any caregiver that will treat the patient throughout the entire care continuum.
Embedded in the EHR as a single platform or standalone, these platforms allow the clinician to access data captured directly in the EHR as well as have access to unstructured data like diagnoses, summaries, notes, etc. By using consistent vocabulary and cardiology specialty-specific report templates, this also ensures that the communication between different patient care teams is complete and clear without any ambiguity.
To comply with Meaningful Use, various specialties, including cardiovascular, are adopting structured reporting with such benefits as reduced errors for future diagnostic or care decisions, lower semantic errors and increase in care quality and consistency while abiding to the latest clinical guidance.
Serge Makowski, MD, co-founder and CEO of MediReport. Photo courtesy MediReport.
Directly acquiring data from the clinicians (unstructured data) may create obstacles in fully leveraging the EHR’s potential, including the secondary use of data. But modern and flexible structured reporting systems employ discrete data fields to collect structured and standardized data from all physicians. This allows the facility to reuse the discrete data in various instances, like clinical research or registries.
The technology intelligently guides the clinician on the required and important procedural data and generates near real-time reports, which are automatically incorporated into the facility’s existing electronic health record (EHR).
The solution integrates discrete data fields in a precise order to collect structured pertinent data, relying on a structured database that incorporates requested criteria from registries (for instance, NCDR and SCAI) or international classifications like SNOMED.
HL: Of the 10% using structured reporting today, how is this being accomplished -- through which platforms, practices, or workflows?
Makowski: Different facilities are at different levels in their structured reporting initiative. The basic stage is the use of procedure documentation to record all relevant procedure and patient information as structured data.
Other facilities go even further by integrating this documentation system via HL7 to allow communication and with other departments and the hospital EHR. A third, more advanced level in the structured reporting process is embedding additional value-based functionalities via customization of reports per physician or facility specifications.
The most advanced stage in structured reporting implementation is maximum interoperability with different systems or platforms, such as integration with industry-specific registries. We have seen more modern and sophisticated CVIS systems offering all these functionalities and other business-oriented features as real time analytics and inventory management.
All of this results in greater physician satisfaction, as it allows them to save a great amount of time and resources when performing their clinical reporting and administrative tasks with better outcomes. This, in turn, has a strong downstream effect on improving billing accuracy and reducing costs.
HL: These findings come at a time when health IT systems are augmenting reporting systems with querying of "live" data. Does structured reporting facilitate this?
Makowski: Absolutely. Modern structured reporting systems facilitate the acquisition of live data by seamlessly interfacing with the EHR, the HIS, and modalities (like the PACS, Hemodynamic system, X-ray station) in real-time. This allows the system to automatically, and effortlessly, capture pertinent live data, before (i.e. patient history, consent), during (i.e findings, hemodynamic values, dosimetry, measurements) and after (i.e. patient recovery, adverse event) the procedure.
This data then automatically populates the structured report, and the doctor only has to review and sign the final report before it is automatically sent to the referring doctor and the EHR.
HL: What industry standards support structured reporting? To what degree do such standards now need to be defined to assure compatibility between different structured reporting implementations?
Makowski: It is essential that standards are defined. Systems that follow standards work together better, they are easier to implement, and they improve efficiency, resulting in better patient care. Effective platforms are fully connected to the hospital system be adhering to the Integrating the Healthcare Enterprise (IHE) Profiles. They use the IHE as a baseline, and may also follow main standards such as HL7, FHIR, and DICOM. They also have the ability to support proprietary standards from other vendors; for example, in electrophysiology they might communicate with pacemakers and defibrillators via the Implantable Device Cardiac Observation (IDCO) profile, while In the catheterization lab they use HL7 and DICOM to communicate with Hemodynamic and x-ray systems.
