A Florida-based community hospital is using AI tools to examine patient data and formulate new treatment protocols for deadly illnesses such as sepsis.
Artificial intelligence (AI) is sometimes seen more as hope and hype than as a solution for healthcare that can improve patient outcomes and reduce costs.
Now, a community hospital in Florida is not only proving healthcare AI can generate observable results in the clinical arena, but also showing that initiatives can be launched and sustained without the deep pockets of a large health system.
Flagler has already realized financial gains from new care pathways for pneumonia and sepsis. The pneumonia and sepsis care pathways have generated nearly $850,000 in costs savings in less than a year. Once care pathways are established for about a dozen other high-risk conditions, cost savings could be as high as $20 million over the next three years, says Michael Sanders, MD, CMIO at Flagler.
For health systems and hospitals seeking to capitalize on AI technology, Flagler is a case study on how AI can be used to develop new care pathways that simultaneously cut costs and improve clinical outcomes.
In 2017, Flagler decided to use AI to mine the hospital's clinical information and focus on deadly and costly conditions. The goal was creating new care pathways for conditions such as sepsis to boost clinical outcomes and drive down cost of care.
Early this year, Flagler signed a three-year AI technology contract with Menlo Park, California–based Ayasdi to help with this initiative.
The AI-powered process reviews thousands of patient records from the hospital, then identifies the patient cohort with the best outcomes, such as the lowest direct variable cost, readmission rate, length of stay, and mortality. The commonalities of this patient cohort drive development of new care pathways such as revision of the emergency department order sets with new treatment protocols.
The hospital also has internally published a new care pathway for COPD, and is set to rollout diabetes, total knee replacement, heart attack, and coronary artery bypass graft pathways next.
"Any hospital can do this," Sanders says.
Here are three ways Flagler implemented AI into its organization to improve patient outcomes and reduce total cost of care.
1. Create a Staffing Strategy
To limit costs, Flagler decided to staff its AI capabilities internally, Sanders says.
"We do not have a data scientist at our hospital. All of this work has been done internally, which I think any community-based hospital can do. We have a couple of folks, including myself and a SQL query guy who has great SQL skills," he says.
Most of the work done by humans in a new care pathway effort involve setting up the parameters and queries for data extraction. Once the data has been verified, the AI tools sift through the hospital's data to combine patients into cohorts.
2. Start With a Simple Pilot
In the pilot phase of Flagler's AI initiative, Sanders wanted to pick a condition with relatively few variables to compute.
Picking the right place to start—pneumonia—was essential to the effort, Sanders says. "I wanted something that would be relatively easy and straightforward, so we could get our feet wet with something that wasn't too complicated."
It took nine weeks to complete the pneumonia project—from data extraction to AI analysis, to refinement of new treatment protocols, to internal publication of the new care pathway.
The next care pathway project—sepsis—took two weeks, with learning gains speeding data extraction and the treatment protocol refinement process.
3. Collect and Crunch Patient Data
Flagler took multiple steps to extract and organize the patient data used to form new care pathways:
Patient data is extracted from the hospital's Allscripts EHR, surgical system, enterprise data warehouse, and financial system
Data is loaded to a cloud technology, where it can be stored and manipulated
Three rounds of semantic and syntactic validation are conducted to make sure the data is accurate
AI tools are used to carve out patient cohorts
Once the analytical work has been completed, the AI team members engage the hospital's physicians to craft the new care pathways.
"We sit down with one physician from every physician group in the hospital. We have been having meetings every first and third Wednesday of the month, and we go through the care paths and make changes that are deemed appropriate based on our evidence and experience," Sanders says.
The new care pathways are used to set treatment protocols in the emergency department and inpatient wards, he says. "We publish the care paths and make changes to the ED order sets and admission order sets based on the care paths, [then] we begin monitoring them."
Operationalizing Care Pathways
Care pathways are road maps for treatment, Sanders says.
"The pathway includes everything, from the moment patients enter the ER to when they are admitted and discharged. There are two sets of orders involved—the care that is given in the emergency department and the care that is given after admission," he says.
Positive patient outcome data has fueled physician engagement in the AI-driven care pathways, Sanders says.
"We had 557 patients with septic shock. For all but 19 patients, the sepsis order set was used. When we looked at the data, the mortality in our hospital is below the national average at 9%. If a patient had the sepsis order set used, they had a readmission rate of 6%. If we look at the group where the order set was not used, the readmission rate was 21%."
If you are going to change your discharge model for at-risk geriatric patients, there are four elements you are going to need.
Geriatric inpatient care and the transition of patients to postacute care are among the most daunting challenges for health systems and hospitals because of the possibility of patient functionality losses in the inpatient setting and costly readmissions after discharge.
In one effort to rise to these challenges, Philadelphia-based Penn Medicine has launched Supporting Older Adults at Risk, or SOAR. SOAR which served its first patient in January, is modeled after the "flipped" discharge program crafted at Sheffield Teaching Hospitals in the United Kingdom.
Penn Medicine has modified the U.K. approach to "flipped" discharge, which features in-home patient assessments after hospital discharge more prominently than in-hospital assessments at discharge, says Rebecca Trotta, PhD, RN, director, nursing research and science, at the Hospital of the University of Pennsylvania in Philadelphia.
"We tried to adhere to the central tenets of flipped discharge: maintaining a geriatric focus; having comprehensive services, not just medical services; and offering services early and intensively," she says.
SOAR has served 46 patients so far, and the early data is promising, Trotta says. A primary goal of the initiative is decreasing hospital length of stay, which has dropped about 1.5 days for SOAR patients.
Lowering length of stay reduces risks associated with hospital care, she says. "We are thinking about how we can maximize care in the home setting because we know when patients are in the hospital they are at higher risk for falls, functional decline, and delirium."
Flipping the discharge model at Penn Medicine means equipping its SOAR program with the following four elements:
1. Geriatric nurse consultants
Four nurses who serve as geriatric nurse consultants in the inpatient setting are a building block of the SOAR initiative. They are staffed through the department of nursing.
"Their full-time job is to identify older adults upon admission who could benefit from a comprehensive geriatric assessment, to share recommendations and findings with the interprofessional team, to follow up on those recommendations, and to collaborate with caregivers," Trotta says.
The geriatric nurse consultants establish working relationships with patients that underpin the SOAR program, says David Resnick, MEd, MPH, innovation manager for the Acceleration Lab at the Penn Medicine Center for Health Care Innovation in Philadelphia.
"The geriatric nurse consultants spend a lot of time with patients in the hospital conducting assessments and getting to know them and their caregivers, which builds trust. So, we have seen refusals of home care at the door—which happen about 15% of the time with traditional home care—fall to zero with SOAR," he says.
2. Home assessments
SOAR provides home health services through Penn Care at Home, a division of Penn Medicine. When patients return home, a Penn Care at Home staff member conducts an assessment.
The home assessments confirm or revise discharge assessments conducted in the hospital including physical therapy and occupational therapy evaluations, she says. "SOAR verifies what is needed. We often find patients need more or different things than we thought in the hospital."
Home assessments help recovering patients to live at home, Trotta says.
"The goal for our older patients is to maximize their ability to take care of themselves, which includes their daily living and functioning. They need to do things like get to the bathroom, prepare food, do laundry, and keep their house clean. Seeing how that unfolds in their real environment lets us see where they might need help versus seeing it in the hospital," she says.
3. Rigorous handoff
"We do a handoff call with the geriatric nursing consultants that includes highlighting key things they have learned about the family and caregivers. They also work on medication reconciliation jointly with the home care team," Resnick says.
Hospital staff and the home care team work together closely for the first 48 hours after a patient has been discharged, he says.
"For the first two days that a patient is home, the home care team is tethered back to the hospital. The home care team can either call or message the acute care provider and talk with the geriatric nurse consultants for issues that arise in the home such as medication discrepancies and other concerns. They are not on their own," he says.
4. Intensive services
SOAR provides a level of service that is more extensive than traditional home healthcare, Trotta says.
"It is organized differently than standard home care. Typically, across the country the average time to the first home visit is at least two to three days. For an older person leaving a hospital without a connection to their next provider for two or three days, there is a risk of something going wrong in that time," she says.
With SOAR, patients get a same-day visit after discharge. They leave the hospital in the morning, then see a home care nurse that afternoon. They also see that same nurse the next day.
Same-day and next-day visits following hospital discharge allow home care team members to address immediate needs such as questions about medications and usage of durable medical equipment, Trotta says. "There is more immediate attention as patients transition from the hospital to the home."
SOAR provides a level of services, which are reimbursable by Medicare and commercial insurance, that is often not included in traditional home care, she says.
"Our patients are defaulted to receive physical therapy, occupational therapy, and social work. In traditional home care, those services can be delayed or not recognized as needed at all," Trotta says.
Collection of patient-generated data and embracing patient-centered communication are called crucial factors in avoiding harm from diagnostic errors.
Drawing information from patients can help boost understanding of why diagnostic errors happen and reduce the risk of future errors, research published this week says.
Diagnostic errors are a serious patient safety problem, impacting about 12 million adult outpatients each year and causing as many as 17% of adverse events for hospitalized patients.
"Health systems should develop and implement formal programs to collect patients' experiences with the diagnostic process and use these data to promote an organizational culture that strives to reduce harm from diagnostic error," researchers wrote in an article published today in the journal Health Affairs.
The research features an examination of 184 narratives from patients or family members about diagnostic errors collected in a new database maintained by the Empowered Patient Coalition.
The data provide unique and valuable insight into diagnostic errors, the researchers wrote.
"Patients' reports of their experiences of diagnostic errors can provide information that traditional measurement mechanisms often fail to capture. Given the absence of diagnosis-specific experiences in most surveys and patient-reported outcomes, the only current way to capture patients' experiences of diagnostic error is via patient complaints. However, complaints are often viewed as satisfaction matters rather than safety signals," the researchers wrote.
Pain points
The Empowered Patient Coalition narratives identified four areas where poor clinician-patient relations contributed to diagnostic errors.
Patient knowledge was ignored in 92 of the narratives. Patients or family members said that clinicians ignored or disregarded reports of clinical indications such as symptoms and changes in patient status.
Disrespect of patients was considered a possible contributing factor in several diagnostic errors. Clinician disrespect of patients was reported in several forms such as belittling, mocking, and stereotyping.
Failure to communicate was another theme in the narratives, with clinician failings ranging from ineffective communication styles to refusal to talk with patients and family members. Examples of poor communication included unanswered phone calls and unresponsiveness to questions.
Manipulation or deception was reported in 15 of the narratives. This behavior fell into two categories: Clinicians using fear to influence care decisions or patients who were misled or misinformed.
Addressing the problem
To help reduce diagnostic errors, the Health Affairs researchers propose five methods to collect patient experience data and encourage better communication between clinicians and patients.
Creating new requirements for clinicians to conduct lifelong communication training. These requirements could include training to manage patient expectations through discourse.
Including communication skills, professionalism, and safety knowledge in certification and continuing medical education programs.
Health systems and providers should encourage patient engagement in safety through active and systematic collection of patient observations of clinician behaviors. These patient engagement efforts should be incorporated in mechanisms that are designed to change clinician behaviors.
Patient reports identifying clinician behaviors that pose a risk of diagnostic errors should result in interventions to foster patient-centered communication. These reports should be corroborated through the medical record or some other form of independent analysis.
Hospitals and health systems should include patient reports of diagnostic errors into training and patient safety programs.
A multi-pronged approach is needed to address aberrant clinician behaviors that lead to diagnostic errors, Traber Giardina, PhD, lead author of the Health Affairs research, told HealthLeaders today.
"We recommend health systems use a systematic method to collect patient reports of these types of behaviors. This would allow for these behaviors to be identified and monitored. A safety culture that encourages not just patients but also clinicians and staff to report these behaviors is needed. Additionally, we suggest reforms in medical education that highlight patient safety," she said.
These efforts require walking a fine, said Giardina, a patient safety researcher at the Michael E. DeBakey VA Medical Center and assistant professor of medicine at Baylor College of Medicine, both in Houston.
"Fostering clinician accountability for the unprofessional behaviors experienced by the patients who reported diagnostic errors is sure to be challenging and will need to be balanced by the need to address pressures on clinicians that lead to burnout, which may even contribute to these behaviors. These at-risk behaviors that compromise patient safety must be addressed though. More policy priority to nurture the patient-physician relationship is long overdue."
The new rules, which focus on the 2019 and 2020 calendar years, update payment regulations, support expansion of remote patient monitoring, and seek to ease regulatory burdens.
The Centers for Medicare & Medicaid Services (CMS) are updating payment and innovation rules for home health.
As home health agencies expand and health systems establish more robust capabilities for post-acute care including home health services, a CMS final rule announced last weekfinalizes calendar year 2019 and 2020 payment and policy changes, including regulations to promote remote patient monitoring and ease regulatory burdens.
"[The final] rule overhauls how Medicare pays for home health, refocusing on the needs of patients, promoting innovation, and reducing burden for physicians and home health providers," CMS Administrator Seema Verma said in a prepared statement.
Medicare payment
Medicare payments to home health agencies in calendar year 2019 are estimated to increase 2.2%, or $420 million, based on the agency's finalized policies, according to a CMS fact sheet.
A rule change that eases operations for home health agencies as well as health systems and hospitals that offer home health services speeds Medicare payments.
The rule change under the Home Health Prospective Payment System alters the unit of payment under HHPPS from 60-day episodes of care to 30-day episodes of care. The change is set to be implemented on Jan. 1, 2020.
Another new payment regulation slated to start on Jan. 1, 2020, is designed for Medicare to pay for value rather than volume of services. Under this change, Medicare will discontinue the practice of determining home health payments based on the number of visits provided. Instead, a patient's medical condition and care needs will be the determinative factors.
"Therapy thresholds encourage volume over value and do not acknowledge that all patients are not the same, with some patients having complex needs that do not involve a lot of therapy," the CMS fact sheet says.
Remote patient monitoring
The final rule sets a new definition for remote patient monitoring and makes it easier to receive Medicare payment for the service.
The final rule states remote patient monitoring can be a beneficial service in the home-health setting.
"Fluctuating or abnormal vital signs could be monitored between visits, potentially leading to quicker interventions and updates to the treatment plan. Additionally, … remote patient monitoring may improve patients' ability to maintain independence, improving their quality of life," the final rule says.
Some patients garner major health benefits from remote monitoring, the final rule says. "Particularly for patients with chronic obstructive pulmonary disease and congestive heart failure, research indicates that remote patient monitoring has been successful in reducing readmissions and long-term acute care utilization. Other benefits included fewer complications and decreased costs."
The final rule's new definition of remote patient monitoring is relatively succinct: "The collection of physiologic data—for example, ECG, blood pressure, glucose monitoring—digitally stored and/or transmitted by the patient or caregiver or both to the home health agency."
Under the final rule, remote patient monitoring is billable to Medicare as an administrative cost.
Regulatory burden reduction
Reducing bureaucratic costs is a central theme of the final rule.
For example, CMS is ending a requirement that certifying physicians estimate how long home-health services are needed. "This policy is responsive to industry concerns about regulatory burden reduction and could reduce claims denials that solely result from an estimation missing from the recertification statement," the CMS fact sheet says.
For calendar year 2019, CMS estimates this rule change will generate $14.2 million in cost savings for certifying physicians.
The best way physicians can adapt to patient feedback online is to actively participate in the online review process, the chief medical officer of Press Ganey says.
Whether physicians like it or not, online reviews and transparency are an expanding feature of the healthcare landscape.
That's the conclusion of Thomas Lee, MD, chief medical officer of South Bend, Indiana-based Press Ganey Associates and a practicing primary care physician.
Prior to joining the Press Ganey staff in 2013, Lee was network president for Boston-based Partners Healthcare System and CEO for Partners Community HealthCare Inc.
In addition to his role at Press Ganey, Lee serves on the board of directors at Geisinger Health System in Danville, Pennsylvania, and holds professorships at Harvard Medical School and the Harvard School of Public Health.
HealthLeaders spoke recently with Lee. Following is a lightly edited transcript of that conversation.
HL: What are the primary online review concerns for physicians?
Lee: Press Ganey has done a lot of quantitative analysis of what really matters to patients—what drives their likelihood to recommend.
The analyses of millions of patient surveys indicate it is more about the "how" of what is delivered rather than the "what." Was the care coordinated? Was the care empathic? Was there good communication?
Our data overwhelmingly indicate that if we get the "how" right, then patients cut us slack on the "what." We rarely find that waiting time is a driving factor of likelihood to recommend. Patients don't like to wait, but the statistical drivers are about the "how." Was the teamwork good? Was there empathy? Was there good communication?
This is good news. These factors are within my control as a physician. I can control whether I am empathic. I can control whether I make an effort to have the care coordinated. I can control whether I communicate well.
HL: Why should physicians participate in online reviews?
Lee: They should participate in the process and drive it for multiple reasons.
A not-so-lofty reason is that it is going to happen to you anyway. HealthGrades, Vitals, and other websites are collecting information and posting it. They won't have very many patients, and the ones who log on will be disproportionately negative. So, you should take control and post reviews on your own website because it's going to happen anyway, and it's going to be done in a way that is not going to make you look great.
A little loftier, posting your reviews online is good marketing. Google and other search engines prioritize the websites that have more data. So, if you take control of online reviews, not only do you have a better snapshot of how your patient population feels—which tends to be positive—your website gets pulled to the top.
The best reason to post reviews online is it makes physicians better. It's a vivid reminder that every interaction with a patient is a high-stakes interaction. All you have to do is be the kind of clinician who patients are hoping for—and the kind of clinician you want to be.
Transparency puts the focus on the future. It nudges physicians to be reliably at their best.
HL: How can physicians promote generation of positive reviews?
Lee: The way to get positive reviews is to be the way you want to think of yourself reliably. It's not hard. If you sent a relative to me, my goal would be for that relative to say to you, "Thanks for sending me to Tom Lee. He was fantastic." Virtually all physicians know how to make that happen.
The real question is, are we going to be reliable in treating our patients. That is all that it takes to get patients to write positive reviews about you.
HL: How can physicians react constructively to negative reviews?
Lee: First, if there are no negative reviews posted, it just doesn't look credible. So, you should not weed out negative reviews.
My last negative comment was from February 2018 about the completeness of my patient notes. When I first saw that review, I felt not all of the patient's comments are important and I'm not going to spend my time typing. Then I calmed down and behaved more maturely. In this era of open notes, I realized the notes I take are not just for me. My notes are a way of communicating with my colleagues as well as the patient, and I needed to improve. I learned from that feedback.
HL: What is the next frontier for online reviews and transparency?
Lee: The broad theme in transparency is collecting information from patients.
For a long time, there has been an information imbalance in medicine, where physicians and other healthcare professionals have known more than patients about medicine. Now that baby boomers have reached Medicare age and are more frequently patients, what is becoming clear is that there is a second information imbalance—patients know more than physicians do about what is important to them.
So, the collection of information from patients to correct that second information imbalance is going to be one of the major themes in the evolution of healthcare over the next 10 to 20 years.
There are unique challenges to teaching knowledge and skills in the ICU setting such as patient urgency.
Part of the challenge of teaching in the ICU setting is the wide array of team members, from a variety of professionals such as pharmacists and nurses to various learner levels such as students and fellows.
In this diverse and multidisciplinary environment, education is fostered with structured bedside teaching techniques, learning end-of-life communication skills, and practicing medical procedures.
There are three strategies to teach diverse ICU teams, according to a research article published recently in CHEST:
Bedside teaching should include assigning roles, responsibilities, and expectations for all team members
Open communication should be encouraged, and educators should create a safe learning environment with critical thinking
Interprofessional education can take place outside the hospital, including simulations that can train teams
Leadership is an important skill to learn in the ICU, Lekshmi Santhosh, MD, lead author of the CHEST article, told HealthLeaders last week.
"The ICU is particularly important to teach leadership skills, not only in acute situations like Code Blues, but also in the day-to-day rounds situation," she said.
Santhosh said there are several methods to teach leadership skills in an ICU: role modeling, entrustable professional activities that target leadership skills, leadership simulation activities, and small-group workshops and didactics.
End-of-life communication skills
End-of-life conversations with families are frequent in the ICU setting and require advanced communication skills.
"Teaching about communication, specifically in the context of family meetings in the ICU, is another invaluable skill for trainees and a teaching opportunity for faculty and fellows," the CHEST researchers wrote.
Earlier research provides several tips on holding difficult conversations with family members in an educational setting:
Before a family meeting, the learner's baseline communication skills and understanding of the clinical situation should be assessed
The learner's educational needs should be matched with one of the assigned roles in the conversation
Before the meeting, the learner should review the meeting agenda, the instructor should discuss the learner's goals, and pertinent communications skills should be highlighted
Once the meeting starts, the patient and the family should be informed that a trainee will be leading the conversation
The instructor should observe and take notes
After the meeting, the learner should be debriefed with opportunities for reflection, corrective feedback, and planning for future encounters.
"In the meeting itself, the trainee should lead the conversation. The observer should take notes and avoid interruptions—only stepping in if truly needed," Santhosh said.
Procedure training
The ICU can be a daunting setting to learn medical procedures, Santhosh said.
"Procedural teaching has added challenges in the ICU environment, in that procedures often have to be done rapidly, with urgency, and with actively decompensating patients," she said.
Procedural basics such as tray preparation, sterilization of the surgical field, and communication cannot be compromised despite the urgency of the situation, Santhosh said. "Maintaining a calm presence under pressure is of utmost importance."
When performing medical procedures, there is a fine line between teaching and treating, she said.
"The balance between autonomy and supervision in the critical care setting is also crucial, since the procedures are high-stakes. Thus, simulation can be extremely helpful—trainees can practice concepts of 'motor memory' and master the various steps, so that they can more readily free up their mental bandwidth when the procedure needs to be done urgently in the ICU."
Online communities for patients with cancer and other serious illnesses provide emotional support and vital information.
Online cancer communities allow groups of people with expressed interest in cancer to communicate using a website, instant messaging, or email.
Forming online communities generates several patient benefits, according to a recent research article in Journal of Oncology Practice.
"Online communities can provide a measure of emotional support that may be comforting to patients and caregivers. Connecting with someone who has undergone a similar experience can be helpful by providing a personal perspective in addition to valuable information," the researchers wrote.
They found additional benefits included connecting individuals with rare diseases to patients who share the diagnosis, low or no cost to patients for the community platform, and privacy and anonymity protection.
Building and maintaining online communities includes planning, platform selection, proactive management of user posts, and caregiver engagement.
Fostering online communities
There are essential ingredients to grow an online community for patients, Lidia Schapira, MD, the corresponding author for the Journal of Oncology Practice research, told HealthLeaders this week.
"Important steps in launching an online community include defining its goals and deciding which platform on which to host it. Since the initial weeks and months help establish the culture and norms for the community, proactive management during this time is especially important. It is also important to involve and acknowledge needs of caregivers and to anticipate the need for revisions as the community evolves," she said.
With a cultural foundation in place, an online community can offer different levels of communication:
Individual users post questions, answer questions, and create content in a shared space
Users have the option to lurk—read postings in the community without communicating
Multiple users can influence content
There are several approaches to maintaining an online community, Schapira said.
"Important ingredients for maintaining an online community include: community managers, a regular influx of new members, and mechanisms for members to provide feedback on how to improve the community. What we learned from our interviews and research is that, in addition to motivation, it takes a concerted and disciplined effort to maintain a community," she said.
Online community pitfalls
There are several risks associated with online communities.
"Oncologists' most pressing concern is the exposure to misinformation. Scientific information may be misunderstood, and patients often report that comments they read on the Web are not only unhelpful but also scientifically flawed," Schapira and her coauthor wrote.
Online communities can pose social risks, they wrote.
"Other risks include the possibility of becoming so addicted to or reliant on Internet-based relationships as to become more socially isolated. A possible risk may also come from oversharing or sharing personal details that may compromise the patient’s own privacy and that of family members and members of the care team."
The helper therapy principle also bears risk in online communities. The principle applies when patients provide support to others who are facing similar circumstances.
"Although some have suggested that serving this role can be therapeutic, there is concern that occupying the role of the helper could also limit patients from being able to express their needs openly, thereby preventing them from attaining full benefit of the support group," the researchers wrote.
Clinical effect unclear
The impact of online communities on clinical outcomes is unknown, Schapira said.
"Our discovery work taught us that we need novel research methods to analyze how and if participating in an online community impacts measurable health outcomes—these could be measures of self-efficacy, coping, managing symptoms of illness or side-effects of cancer therapy," she said.
Caregivers are another area requiring more online community research, Schapira said.
"There is also a need to study the impact on family caregivers' ability to handle their increasingly important roles in providing support and advice and managing complex decision-making."
Strategies for overcoming ICU instruction obstacles include condensing educational interactions and conducting bedside teaching sessions.
The ICU is a unique care environment that includes teaching of a diverse set of caregivers such as nurse practitioners, physician assistants, medical students, residents, and fellows.
Teaching in an ICU involves several obstacles, according to a recent article published in CHEST.
"Teaching in the ICU comes with unique challenges given the medical complexity of the patients, the time pressure, the diverse levels and professions of learners, and the challenges of communication at the end of life," the researchers wrote.
Overcoming these hurdles includes assigning capable instructors, condensing interaction time with students, and teaching at the bedside.
Instructor qualities
The researchers surveyed internal medicine resident physicians to find out which qualities an excellent ICU teaching physician should possess. Enjoyment of teaching was the most coveted quality:
Enjoyment of teaching residents
Demonstrating empathy and compassion to patients and families
Explaining clinical reasoning and differential diagnosis in the critically ill patient
Treating non-MD staff members respectfully
Showing enthusiasm for topics discussed on rounds
"These teaching tactics can be refined and continuously improved, making them ideal for faculty development. Beyond behaviors that educators can learn, they also need to know the targets for board certification to prepare their learners," the researchers wrote.
Time pressure
There are ways for ICU instructors to overcome busy schedules, Lekshmi Santhosh, lead author of the CHEST article, told HealthLeaders this week.
"We recognize that the ICU clinical environment is extremely fast-pace and that balancing high clinical workload with teaching is always a challenge. Fortunately, there is robust medical educational literature from the outpatient, ED, and inpatient settings discussing how to approach teaching in a time crunch," she said.
The "1-minute preceptor model" and mini-chalk talks are examples of condensing interactions with trainees.
Mini-chalk talks can be banked for future use and should be honed in advance of presentation. Simple symbols are effective such as Venn diagrams and flow charts.
Bedside instruction
The bedside is a prime location for instructing residents and other trainees in the ICU, Santhosh said.
"Learners crave clinically applicable, relevant teaching points delivered just-in-time to reinforce learning: bedside teaching thus engages learners immediately and they crave good bedside teaching. Moreover, it is patient and family centered as it brings patients and families into the educational process in the role of educators. Lastly, bedside teaching is often efficient for teaching faculty as it combines direct patient care with education."
Bedside instruction must also be constrained to the ICU environment, the researchers wrote.
"ICU bedside teaching has to be deliberately and carefully refined. Educators should thoughtfully consider the limited scope of a bedside teaching session and resist the temptation to over-teach. Instead, giving a brief 5-min to 15-min talk at the bedside that is relevant to current or recently admitted ICU patients may be more helpful," they wrote.
Some doctors reject a provocative proposal to delay antibiotics for sepsis patients without shock. But proponents argue antibiotics pose significant risks, too.
Drawing the line on antibiotics for sepsis patients is drawing scrutiny, pitting an antibiotic stewardship perspective against the established reliance on antibiotics.
A recent editorial in Journal of the American Medical Association calls for a more conservative approach for treating sepsis—delaying antibiotics treatment in some cases of sepsis without shock.
With its deadly mortality rate, an aggressive approach to treating sepsis patients with antibiotics is common practice.
Treating patients with possible sepsis or septic shock requires a nuanced approach, the editorial co-authors wrote.
"The need to treat patients rapidly and aggressively ought to reflect on the severity of illness and certainty of diagnosis rather than applied uniformly to all patients. If a patient clearly has a bacterial infection, prompt treatment is indicated. If there is diagnostic uncertainty, however, clinicians should calibrate their response to severity of illness and probability of infection," they wrote.
The proposed approach challenges the standard of care, which calls for rapid administration of antibiotics to all sepsis patients whether they are in shock or not.
Aggressive track
Administering antibiotics to all sepsis patients should remain the standard of care, James O'Brien, MD, director of quality and patient safety at OhioHealth in Columbus, told HealthLeaders this week.
"The data supports that the sicker a patient is—and septic shock is probably the best example—the more likely they are to benefit from antibiotics. But there is evidence of patients with sepsis without shock who also benefit," he said.
O'Brien said a family medical case exemplifies the advisability of administering antibiotics to sepsis patients without shock.
His mother had melanoma removed from her back, then she got confused at home. She went to the hospital with no signs of shock but was treated for sepsis.
"As a result of the care that the hospital provided—getting rapid antibiotics and opening the wound to drain—she recovered. But if we had waited for tests to come back, it's very possible she could have progressed to develop shock. Then you're dealing with somebody whose mortality jumps up significantly," O'Brien said.
"So, I have hard time from the patient's standpoint waiting until they have shock, then all of a sudden we'll jump onboard," he said.
Adverse impacts
Patients face significant risks from antibiotics, Michael Klompas, MD, the lead author of the JAMA editorial, told HealthLeaders this week.
"The risks of antibiotics extend beyond Clostridium difficile alone. They can cause organ damage, interact with other medications, and promote colonization and infection with drug resistant pathogens that then pose downstream treatment problems," said Klompas, an infectious disease physician at Brigham & Women's Hospital in Boston.
The JAMA editorial lists several other adverse impacts from antibiotics: kidney injury, hepatitis, cytopenias, sever rash, mitochondrial toxicity, and alteration of the microbiome.
There are adverse effects associated with antibiotics, Klompas said.
"Studies estimate that a remarkable 20%-25% of hospitalized patients exposed to antibiotics develop some sort of adverse effect. This flies in the face of the commonly held perception amongst both doctors and the general public that antibiotics are "free"—they are very safe and there's little downside."
Treatment for sepsis
Treatment of sepsis requires the exercise of good medical judgment, O'Brien said.
"It's more art than science now. For me, it winds up at the intersection between patient risk, the severity of illness, and the likelihood of alternative diagnoses. I factor all of those to determine the relative risk of treatment for sepsis versus the relative risk for not treating for sepsis," he said.
Sepsis patients stretch across a continuum, O'Brien said.
"If I have a patient who is elderly, is currently being treated for leukemia, and was recently in the hospital, these are all risk factors for sepsis. If they come in with shock, and I don't have an alternative diagnosis or reason for them to be in shock, I'm going to need to have significant evidence that the patient does not have sepsis."
"Alternatively, I could have a healthy 22-year-old who has no medical problems, has not been in contact with healthcare, and comes in with unclear complaints. There's no signs of shock and the patient is talking. That's someone I am going to be more delayed with."
Physicians are needed most keenly to treat more complicated sepsis cases, O'Brien said.
"There are a whole bunch of patients who end up being in that spectrum—between the polar opposites. In between is where you need a physician to integrate the data and make a decision when there is uncertainty."
Under pressure to deliver value, increase efficiency, and lower costs, health systems and hospitals are seeking ways to advance the capabilities of their physicians.
At least half of the nation's physicians are sick of their jobs at a time when demands on their skills are more critical than ever. In the past, hospital leaders could reliably throw money at the problem. Not so much anymore.
A tried-and-true way to "engage" physicians historically has been through their paychecks to prod along everything from quality incentives to EHR implementation. The kitty of cash that leadership teams had on hand to make physicians happy is almost empty these days, operating margins being what they are. Without income in hand, hospital leaders must rely on some thrifty and tested "satisfiers" that work in any good business: create a reliable team around them, involve them in the business, and add technology that—for one—makes their job easier.
Leaders may recognize an opportunity: that the underlying fundamentals of physician labor are changing. The benefits of high reliability, team-based healthcare takes the physician from being the only voice in the huddle to a leading voice in a coordinated care plan. Bringing physicians into decisions on supply chain creates the beginnings of a business partnership.
Under pressure to deliver value, increase efficiency, and lower costs, three leading health systems are advancing the capabilities of their physicians through a supportive and productive environment.
A 2017 study published by the National Academy of Medicine found that more than half of surveyed physicians were exhibiting substantial signs of burnout.
No health system can address burnout without first understanding one of the root causes: quality of care. At Cleveland Clinic, efforts to become a high-reliability organization since 2013 have achieved a trifecta: improved patient outcomes, boosted physician satisfaction, and reduced physician burnout.
"There is a big problem in healthcare with burnout, which is complex and involves lack of job enjoyment, feeling stressed, and work-life balance challenges. All of those things improve when you have a team working together smoothly to get the best outcomes for patients," says Edmund Sabanegh, MD, main campus hospital president.
The correlation between team-based care and physician satisfaction is direct, says Sabanegh.
"Things that help us successfully treat patients—team approaches, checklists, and spreading of responsibility—improve our engagement and satisfaction with our career field," he says.
1. Happiness Begins With High-Reliability
Cleveland Clinic's high-reliability initiative has revolved around basic team building, policy standardization, real-time operational management, creating a culture of safety, and sustaining redundancy in the clinical setting.
"There has been a revolution at Cleveland Clinic over the past several years to emphasize a culture of high reliability and safety, as well as to emphasize a team approach to everything we do," Sabanegh says.
But inconsistency in administrative and operational policies is a major challenge for health systems seeking to attain high-reliability, says Sabanegh.
"One of the challenges for any large healthcare system is there are many sites for delivery of care. A pitfall that you can have is failing to recognize the nooks and crannies of the system, then having different policies and standard operating procedures for different areas," he says.
Cleveland Clinic, which features 19 acute-care hospitals, has made policy standardization a priority, he says. "We have worked hard to standardize our policies to make sure that a nurse who works in one ward, then works in another location in our system has a similar expectation and similar understanding of processes," he says.
One of Cleveland Clinic's high-reliability cultural initiatives has upended decades of tradition in the health system's operating rooms. As opposed to the top surgeon dominating discussions and decision-making in the OR, the health system has adopted a team-oriented approach including operating room pauses, he says.
"If we have a surgery and anyone in the room is unsure of equipment status or a missing supply like a sponge, there can be a pause. Any member of the team can say, 'I want to look at where we are before we proceed any further with this procedure.' It could be the most junior member of the surgical team or it could be the most senior member."
To achieve real-time operational management, Cleveland Clinic adopted a reporting system based on tiered huddles this year.
"Every morning, on every nursing unit, there is a huddle of the team. They discuss what has gone right, opportunities, and concerns for the day ahead," Sabanegh says.
The discussions at the ward level are reported to the hospital leadership level, including the president, chief nursing officer, chief medical officer, and chief quality officer.
The hospital leadership's huddle is reported to the health system leadership. Information gathered through the tiered reporting allows senior leadership to take action quickly at any location in the organization,
he says.
"As the hospitals' president, I am hearing every day from every hospital in our system about their challenges and opportunities for the day ahead. What is our workload and how can we balance it? What kind of infrastructure support do we need? What kinds of repairs are needed?" Sabanegh says.
Gathering timely information from throughout the health system is invaluable from both management and labor perspectives, Sabanegh says.
"Real-time operational management gives us both an early warning system for problems and challenges for the entire enterprise, and a great venue to communicate up and down the organization. Everybody is hearing about challenges at other places and how those challenges are being solved," he says.
Culture is essential to creating a high-reliability organization, he says. "We are working very hard to create a culture of safety and high-reliability. Every time the leaders of the organization speak, they talk about
this theme."
Staff members are encouraged to identify quality concerns with public recognitions such as awards. "We don't want to be in a reactive mode. Our system fails when we have a serious safety event. What we want to identify is the near miss or something that could turn into a serious safety event down the road,"
Sabanegh says.
Although redundancy is often equated with waste, Cleveland Clinic sees value in redundancy in the clinical setting, he says.
"We still believe some redundancy is necessary. We are leveraging technology to assist in catching things; but, in this generation, technology will not replace the need to have multiple sets of eyes looking at a challenge," he says.
Education and communication have been key elements of engaging physicians in high-reliability efforts at Cleveland Clinic, Sabanegh says.
Educational programs that support the health system's high-reliability efforts include Solutions for Value Enhancement (SolVE).
"Physician leaders learn about high-reliability, performance improvement, and tackling processes with risk and opportunity while avoiding risk. We have trained thousands of people in our organization in these areas,"
he says.
Cleveland Clinic also communicates with clinicians about the benefits associated with high-reliability organizations, he says.
Engaged clinicians have helped Cleveland Clinic achieve significant high-reliability gains.
The average 30-day readmission rate has fallen from 14% to 12.65%, which represents 2,100 patients per year who did not require a readmission.
Outpatient hypertension control has increased from 66% to 76%, with 15,000 more patients at prescribed goals. Cleveland Clinic estimates improved hypertension care has saved about 100 lives.
"We have seen a steady improvement in our quality outcomes, a reduction in serious safety events, and improvements in our readmissions—all things that are important to our patients and improve when our care team makes sure we are highly reliable," Sabanegh says.
2. Save Physicians and Patients With Predictive Modeling
Predictive modeling is not brand-new technology, but its utility has advanced into care scenarios that have the potential to improve patient care and give physicians a critical tool.
NorthShore University HealthSystem is using prediction models to give physicians important information about their patients. The health system has about 20 prediction models to target high-risk patients for factors such as cardiac arrest and readmission.
Prediction models help physicians decide whether their patients need interventions, says Nirav S. Shah, MD, an infectious disease specialist at NorthShore.
"The prediction modeling is refining the patient population, so that when you perform an intervention you can find the highest-risk patients. Instead of performing interventions on an entire population, we can limit the intervention to a small subset of the patients," he says.
In NorthShore's cardiac arrest prediction model, a patient's risk level for cardiac arrest helps determine whether an intensive care consultation is held. The model has resulted in fewer patients having cardiac arrest and a trend toward decreased mortality, Shah says.
NorthShore is preparing to take a leap forward in its prediction modeling efforts with a cutting-edge technology.
"The most exciting thing we are doing is embarking on a journey to integrate all of our prediction models into a single engine," he says.
Combining prediction models will expand the utility of the data for physicians and population health initiatives.
"Most institutions are getting more and more into these prediction models, but they are each in their own silo. Each prediction model has its own lexicon of risk. What we are doing is trying to combine every prediction model into a large engine, so we can subdivide patients," Shah says.
Subdividing a patient population will help NorthShore manage high-risk patients with multiple morbidities, he says. "We may have an intervention for high-risk readmission patients that could be an intervention for patients who are also high-risk heart failure."
One of the top goals of NorthShore's Clinical Analytics Prediction Engine (CAPE) is establishing a powerful data-driven learning capability, Shah says.
"We are finding a way to sub-segment our entire patient population using analytics and prediction models, so we can target specific patient populations. Then we will use this engine to quickly learn whether interventions on patients or subgroups of patients make sense."
Running randomized trials can take months to a year to complete, but CAPE will run randomized trials much faster, he says.
"On the academic side, you use randomized controlled trials, where you take populations, get consent from them to be part of the study, use resources such as the people who design these studies, then it takes six months to a year to enroll patients and analyze the data. … We will be able to speed up deciding whether an intervention makes sense from six months to a year, to a couple of weeks," Shah says.
CAPE's capabilities will deepen NorthShore's understanding of its patients.
"We are essentially using this new engine to create a learning health system, which will allow us to continuously learn from the patient population. Instead of creating randomized controlled trials that are separate from patient care, we are building the ability to conduct these trials into our system of care," he says.
Physician engagement and winning over skeptics is a crucial element of launching prediction modeling initiatives, Shah says.
"We have a track record of implementing predictive models. Initially, it can grate some physicians when they feel their autonomy is being taken over by algorithms. You have to show you are improving patient care. In the end, all providers want to improve patient care," he says.
Demonstrating positive outcomes for patients is powerfully persuasive for even the most doubtful physicians, he says. "Even if it's an algorithm that physicians feel is supplanting some of their autonomy, if they see that a prediction model is resulting in better outcomes, that will click a lightbulb immediately and they will buy into it."
Rather than diminishing physician autonomy, prediction modeling is already helping many physicians make wiser decisions, Shah says.
"Physicians use a lot of algorithms already. The Model for End-Stage Liver Disease (MELD) score shows whether someone is at high risk for death if they have liver cirrhosis. There are all kinds of tools that already exist that providers do not complain about," he says.
Shah predicts CAPE—which is being built in the health system's Epic electronic medical record, will be appealing to physicians who view EMR data entry as a waste of time.
"You are essentially putting in a lot of data, and it's not giving you valuable information in return. You can show physicians that prediction models are providing them with key insight for outcomes that are important to them and that the data they are entering has a return on investment," Shah says.
In addition to using prediction models to supplement physician decision-making, NorthShore has been improving Epic's user experience for years, Shah says.
"We were the first to adopt Epic on the inpatient and outpatient sides. So, there are many things that we have done from an optimization standpoint—providing physicians with dashboards and other capabilities to help them make the best decisions for their patients," he says.
An Epic improvement adopted this summer helps inpatient clinicians pick the right antibiotic for patients before culture data is available, he says.
"When someone comes in for treatment, we often do not know which antibiotics are going to work for the patient. We want to find the perfect antibiotic for any given patient, so we can reduce the risk of the patient getting worse and reduce the risk of resistance," Shah says.
The new antibiotics capability in Epic includes data from What's Going Around—a graphical representation of five illnesses in Chicago's northern suburbs such as strep throat and flu-like illness.
The key was creating a program that draws several patient variables directly from Epic, Shah says.
"It uses infectious disease guidelines, it uses prediction modeling, and it uses the What's Going Around epidemiological tool. All of that information is integrated into a single decision-making tool that allows a provider to determine the best antibiotic for their patients before there is culture data to guide therapy," he says.
3. Physicians Empowered to Help With Supply Chain
Kettering Health Network is tapping the clinical expertise of physicians to improve clinically integrated supply chains, which support value-based care with physician engagement, data analytics, point-of-use management, and strategic contracting.
Trisha Gillum, director of supply chain management at Kettering, says physicians can play diverse roles in a clinically integrated supply chain.
"It can be as small as a physician champion on a single project, to being a physician champion for a service line, to being on the payroll for supply chain," she says.
Gillum says the best physician champions for medical supply changes are personally engaged in the effort. "They are willing to understand both the financial and the clinical nuances to a project. They are also willing to speak with their peers—to be a cheerleader or champion for a project."
When Kettering identifies engaged and respected physicians who are interested in serving in the champion role, the doctors receive training from The Advisory Board Company. Physician champion programs at this Washington, D.C.–based consultancy range from individual sessions to a physician leadership track that has sessions held over several months.
Two primary elements of the education programs are business instruction and learning about the nuances of the changing healthcare industry.
"Many physicians are not in tune with everything that is going on in the hospital environment," Gillum says. "They don't understand when we say we need to save money. So, there is education about financial pressures and clinical pressures."
Another educational goal is giving physicians leadership skills, she says.
"We are asking them to step out of their traditional roles and communicate with their peers at an advocacy level. To do that, we not only need to provide the data to support product conversations, but also give them the tools necessary to have those conversations."
In a clinically integrated supply chain, physician champions play a potentially decisive role in proposed supply changes, Gillum says.
"If you really want a physician to get engaged, they will bring their own mindset about what the answers should be. You cannot expect a physician to come onboard and rubberstamp the process," she says.
Supply chain managers and other leaders should be open to opposing views from physician champions, she says. "They are going to want to engage in the process. They are going to want to modify it. So, you may end up in an entirely different place than you expected."
When physician champions object to proposed supply changes, open communication is essential, Gillum says.
"You have to be transparent. You can't ask a physician to own something like a cost-savings initiative unless you are willing to say how much we are going to make on a procedure. You have to be willing to share all of the data and to give physician champions all the facts to make intelligent decisions," she says.
Supply chain managers should treat physician champions as valued teammates, Gillum says. "You have to realize that you have asked physicians to play a supply chain role and to provide information. If you disregard what they are saying, you are going to lose partners."
The best-case scenario for physician champions is when they take ownership of a supply change project, she says.
"I had a physician who went out and talked with every one of his peers who performed a particular procedure. He convinced every one of them that we needed to make a change. He was able to accomplish more in those conversations than I could have accomplished in months of conversations with the same group of physicians," she says.
Gillum says the two primary benefits for clinicians who assume physician champion roles are gaining experience that helps them compete for hospital administration jobs and helping to decide supply changes that could impact their patients.
"I had one physician say [that] he was passionate about the supplies he used on his patients. The best way physicians can control the supplies that they get is to be part of the conversation and part of the decision," she says.
Photo credit: Nirav S. Shah, MD, infectious disease specialist, NorthShore University HealthSystem (Jean-Marc Giboux/Getty Images)
