With so many great submissions, it is a real challenge for Healthcare Innovation’s editors to pick only a few projects as the top innovations for the year. It is like picking one movie to win the Oscar in a year when there are several great movies. So we wanted to offer readers an idea of some of the other impressive work being done across the country with the following summaries of project descriptions submitted by our seven runner-up organizations, presented in alphabetical order.
AdventHealth’s Virtual MRI Operations Center
AdventHealth is the largest healthcare organization in Central Florida with multiple campuses that spread across several counties. For decades, its quaternary flagship campus, AdventHealth Orlando, served as the major hub for all complex imaging studies because that is where its highly skilled technologists who could perform complicated exams (ex: cardiac MRI) were located. Its imaging volumes have increased exponentially over the years, which ultimately resulted in transporting seven to 10 inpatients each week to AdventHealth Orlando for advanced MRI exams. Outpatients requiring a similar advanced study had to wait up to three weeks to be seen at the Orlando campus due to bottleneck issues from high demand. Across six campuses, the Virtual MRI Operations Center provides comprehensive scanning assistance to staff and physicians–regardless of their physical location. It enables real-time knowledge sharing and coaching across teams/campuses. Using live audio, chat, and video functions, one experienced technologist can collaborate with up to three scanning workplaces simultaneously. Patients do not need to travel to a specific campus.
This innovation has reduced inpatient MRI transfers by 94 percent across the division AdventHealth has grown cardiac MRI services by 38 percent across its local markets since launching this program, and it saves $150,000 annually in length-of-stay costs.
The new system also has helped with team engagement: Before, the only option was to train staff at Orlando and incur additional labor costs to backfill staff. There was no medium or method for assisting team members during complicated studies, and no backup for call-outs or staffing shortages resulting in lost exam volume. Now, expert MRI technologist are available whenever needed for coaching and guidance on complex exams. They have provided coaching and training over 414 times since launching Virtual MRI. The Virtual MRI technologist can pilot up to three scanners at once and can cover call-outs and staffing gaps in the event of an emergency reducing the need for overtime.
Plans are in the works for geographic expansion across Central Florida. The current concept is limited to Siemens MRI scanners, but the future state will open all vendor scanners for virtual scanning. AdventHealth also plans to develop a virtual staffing concept for CT that allows for quicker table turnover without sacrificing quality.
Kaiser Permanente’s Risk Quantification Program
Risk quantification is an evolution in risk management that characterizes risks in terms of probability and dollars of loss expectancy. Optimizing technology and cyber risk management in healthcare is a paramount concern because of the potential for spend to divert funds from patient care priorities such as care quality and access. Decision makers need to know how much technology risk Kaiser Permanente is carrying and if investments are effective in reducing risk in the face of uncertainty.
Legacy qualitative risk evaluation methods (e.g., rating risk as “medium” or “very high”) are unable to provide objective results. Quantitative risk evaluation addresses the problem, but it requires highly specialized capabilities that are not broadly available in industry.
To deploy and scale quantitative risk evaluation capabilities, Kaiser Permanente invested in a broad program using the FAIR standard in a centralized enterprise service model to supply quantified risk results to decision makers. This provides validated, standardized results that are coherent across the enterprise. Since inception, decision makers at Kaiser Permanente have access to templated risk models, using the results to facilitate decision making across various functions, including:
• Information Technology – incremental risk of new technology deployment
• Cyber Security – cyber insurance coverage, prioritization of cyber security projects
• Compliance – control prioritization, prioritization of audit remediation
• Business Operations – power resiliency investment, supply chain management (including Personal Protective Equipment supply and location)
Under the leadership of Kaiser Permanente’s Chief Technology Risk Officer George DeCesare, the team engaged engaged with outside departments to apply quantification in decision support.
As implemented, the model allows Kaiser to scale adoption of risk quantification for business decision making in the form of a service model based on FAIR. Thus far, this approach has brought over $20 million in value to the organization attributed to quantified calculations optimizing risk-based decisions.
Mayo Clinic’s Automated Chemotherapy Dose Rounding Rules
Mayo Clinic has worked to decrease cancer drug cost and waste through the implementation of automated chemotherapy dose rounding rules in the electronic health record system.
Cancer medicine spend has reached $185 billion worldwide, including the spending of $75 billion just in the United States. In 2021 alone, there were 30 oncology novel active substances brought into the market. Many of these are packaged in single-dose vials (SDVs) and have a cost of greater than $100,000 annually. Traditionally, the doses for cancer drugs are calculated based on patient’s weight or body surface area (BSA) resulting in a partial vial with drug left over. Since these drugs are packaged as SDVs, and due to regulatory constraints, it creates a system that leads to waste and increases financial burden. According to the findings by Peter Bach, the proportion of waste varies from 1 percent to 33 percent. There is a mismatch between the required doses and the doses available based on the SDV. The cost of wasted drug is usually passed on to patients and payers. These rising and unnecessary costs pose a risk of “financial toxicity” for cancer patients. It is clear that more measures need to be taken by hospitals to reduce the waste and burden on our healthcare system.
To counter this issue, a team at Mayo Clinic implemented a system of automated dose rounding rules in their EHR. This system specifically aims to prevent waste and reduce cost by not opening another SDV when the dose is close to the vial size. Dose rounding is an approach where the weight-based or BSA-based dose is rounded to the nearest SDV size to minimize waste. At Mayo Clinic, this system was implemented to round the calculated dose to the nearest SDV size if the vial size is within 10 percent of the original calculated dose. Rounding of doses to the nearest 10 percent of the SDV is supported by a position statement from the Hematology/Oncology Pharmacy Association.
To implement this system, Mayo Clinic created a workgroup of pharmacists, pharmacist technicians, and pharmacy leaders. This workgroup was tasked with creating a list of all injectable cancer drugs and develop a set of rounding rules. The rounding rules were approved by multiple committees that included oncology providers, finance, nursing, and pharmacy across all sites. The final step was an approval by the Pharmacy and Therapeutics Committee. Following the implementation, Mayo Clinic collected data for doses administered in both the inpatient and outpatient settings during the first six months of 2019. In six months, this implementation generated a total cost savings of over $7 million and saved a total of 9,814 SDVs from being wasted.
Northwell Health Physician Partners’ EHR Optimization
New York-based Northwell Health Physician Partners (NHPP) had received physician feedback that revealed ongoing dissatisfaction with ease and efficiency within the Ambulatory Electronic Health Record (AEHR).
In response, a multidisciplinary collaborative launched in 2021 and engaged a third-party consultant, EHR Concepts, to perform an AEHR learning needs assessment. They conducted 1:1 interviews and practice site interviews, held service-line focus groups, captured feedback through surveys, and end-user forums. They engaged over 600 end-users to provide the collaborative with recommendations on creating an action plan to improve training, optimization, and stakeholder engagement.
To help physicians best navigate the AEHR, a campaign called “Home for Dinner” launched, offering new training and proficiency services. This goal of Home for Dinner is to offer a personalized learning and support experience designed to help physicians best navigate the AEHR by delivering personalized services, tools, and training to get more time back for their patients, their families and themselves.
One goal was to achieve measurable improvement in overall satisfaction with AEHR by adopting evidence-based interventions to improve proficiency and reduce time spent in AEHR after work hours
Northwell developed a program that focuses on end-user efficiency and adoption with specialty-specific learning plans. They launched an online site designed to provide AEHR users with personalized and on-demand resources to help them best navigate the AEHR. They created an exclusive group of AEHR experts who help providers and team members elevate their skills and knowledge within the application The organization also created a network to communicate new resources available due to new best practices or changes within the AEHR. To measure results, they identify KPI metrics that demonstrate the value-add of the Home for Dinner workstreams using concrete data points.
A cross-functional workgroup was created to support the implementation of the new program deliverables. Due to the magnitude of each deliverable, Northwell arranged them into workstreams with designated management leads. Each workstream was first charged with understanding and adapting the recommendations by their consultants, determining resource requirements, and creating specific goals and deliverables to ensure completion of all program components. Since launching in October 2022, the new Ambulatory EHR Learning Center has received more than 2,000 views. This comprehensive online learning center is comprised of videos, job aids, and FAQs designed to enhance users' understanding of various AEHR features. The site has received over 14,000 site views since launch. Fourteen physician specialty-specific advisory groups have been created, with over 50 participating physicians, to assist in the design of specialty-specific learning plans. The AEHR Super User Community launched, with over 65 Super Users since the October 2022 launch and a 94 percent average meeting attendance rate.
Northwell says the effort shows the power of collaboration in convening ambulatory operations, clinical informatics, the AEHR Solutions Group, end-users, Physician Partners, project management and physicians. Home for Dinner will offer ongoing support and opportunities for physicians to regularly share ideas to make things easier for everyone. It will also help inform how the organization approaches learning for future programs and platforms across Northwell Health Physician Partners.
Penn Medicine’s Cancer Registry Case Finder
Although Penn Medicine uses a commercial third-party vendor tool to collect and store its cancer registry data, it still needed to determine how to cohesively identify, manage, and report on all the potential cancer cases from its large academic medical center. Penn Medicine’s Abramson Cancer Center consists of six hospitals, many clinical departments, and divisions and has approximately 75,000 to 100,000 related encounters as well as 30,000 to 35,000 unique patients related to the treatment or diagnosis of cancer each month.
Penn Medicine built an in-house application to automate, organize, and manage this process, empowering the organization to identify which patients qualify for inclusion in the registry.
In the previous process, the cancer registry management team received an Excel report from its electronic medical record (EMR) system on a monthly basis, containing all potential encounters. The management team normalized these data prior to dividing and distributing the large spreadsheet across the Tumor Registry-Registrar team. Registrars worked on their individual spreadsheets each month and then reported their progress to the managers. Any outstanding potential encounters were added to the next month’s list, creating a perpetual backlog. This manual process required numerous Excel spreadsheets and e-mail exchanges between approximately 25 to 30 employees.
The application Penn Medicine built to solve this business problem is called the Cancer Registry Case Finder (CRCF), a web-based application that was built in PHP using the Symfony framework and supported with an Oracle database back end. This application was fully designed and built by the Penn Medicine Software Development - Database and Applications Group (DAG). All software used to build and maintain the CRCF is open-source, can be used with any major database, and uses the BSD-style and MIT licenses.
The business problem was addressed by implementing the following improved process: This application has been consistently used in the Cancer Services Line’s Cancer Registry, which is an American College of Surgeons (ACOS) Commission on Cancer-accredited program. The CRCF application was integrated with the EMR for source data, and the cancer registry tool allowed Penn Medicine to import and export to streamline the full abstraction process. The application immediately began to save approximately 3 to 5 person-hours per month just in the management of the current workflow (eliminating numerous Excel files and hundreds of emails).
Additional benefits are still being realized. The application provides: • Real-time data reporting, which decreases data delays • Analytics to assess appropriate staffing • Improved data quality • Data for research
Although unexpected, the CRCF application also successfully identifies previously missed cases. Over five months of working within the application, an estimated 386 previously missed cases are now identified with a year of first contact ranging from 2021 cases back to 2000. Since the application is showing all encounters for each facility up through the latest import of data (currently set to load new encounters once a month), there are significantly fewer chances of cases being missed across organizational facilities moving forward.
Michael Restuccia, senior vice president, and chief information officer at Penn Medicine, remarks, “The development of the Cancer Registry Case Finder application is the result of the trustful collaborative efforts between the Corporate IS team and the Cancer Center representatives to produce a solution that improves operational efficiency, enhances patient care and has the ability to benefit not only Penn Medicine but also other leading institutions throughout the country. Such a unique and far-reaching initiative is only possible through leveraging a deep partnership with the Cancer Center subject matter experts and applying the best technical approach to generate such a successful outcome.”
UnityPoint Health’s Use of Machine Learning to Predict Census
The COVID-19 pandemic forced unprecedented challenges on healthcare providers spanning clinical, financial, and operations domains. Perhaps chief among them was balancing a precarious and volatile surge in patient volumes unmatched by prior experience along with an equally challenging reduction in nurse staffing due to COVID infection and staff resignations.
These factors forced nursing leaders onto a tightrope of staffing to adequate levels compared to volume, retaining existing workforce, and minimizing the use of expensive traveler nurses. For many healthcare providers, patient volumes in the last two years do not follow any trend experienced prior to 2020, and predictions about where they are headed have fallen short due to numerous factors including the unpredictable emergence of COVID variants and their varied impact on hospital admissions along with changing behavior of non-COVID patients in their interactions with healthcare. The current environment has stressed traditional nurse staffing models that look at prior year trends to develop staffing plans, putting healthcare systems at risk of understaffing to the detriment of care quality as well as overstaffing to great financial cost.
The analytics team at Iowa-based UnityPoint Health developed a real-time analytics solution that uses machine learning to predict the maximum census for each of the next nine nursing shifts on each nursing unit across the health system. This solution allows decision-makers to strategically shift department staffing to meet the expected demands. Nursing units that are expected to increase relative to the previous shift are colored shades of red by magnitude (“heating up”), and units that are expected to decrease are colored shades of blue (“cooling down”). Units that are expected to maintain their current census are colored grey.
The predictive models are developed for every unit of every hospital in the health system and are retrained each run based on most recent data trends, resulting in a solution that is continuously learning and adapting to emerging data trends. This approach has made the solutions robust enough to be generalizable to each hospital floor throughout the system. The approach shows an improvement over baseline by 10 to 40 percent depending on the floor being analyzed. In general, the models perform better on Med/Surg floors and are less accurate in departments such as the ED or OB due to the inherent volatility of these floors (though still improved over baseline). Most importantly, nursing leaders are now able to confidently make staffing decisions that better support the nursing teams, who in turn, ensure the best care is provided to the patients they serve.
Vanderbilt University Medical Center’s MyHealth Bundles Program
The complex healthcare marketplace is stressful for consumers searching for the level of care they need—and figuring out how to pay for it. Even if health insurance is part of their compensation package, more than 80 percent of employees are confused by the terms and conditions of those benefits. At the same time, 80 percent of employers say their employees don’t open or read their benefits information. That’s often because much of the communication from providers and payers lacks the personalized touch that helps patients understand details about their health and know how to navigate their care journey—gaps that lead to poor health outcomes and higher costs.
Vanderbilt Health Employer Solutions (VHES) addresses these critical gaps with the MyHealth Bundles program—an innovative, value-based approach to managing some of the most common and costly health conditions, such as pregnancy, musculoskeletal pain, hearing loss, and medical and surgical weight loss. When patients feel that their care team listens to them, they’re more likely to be highly involved and proactive with their care. That’s why Vanderbilt Health’s marketing communications and clinical teams worked cross-functionally to identify pivotal moments and pain points in a patient’s care path when digital communication could help alleviate confusion and stress. From there, an informed digital journey was created for each bundle to address specific health needs—automating many aspects of preventive care.
For example, the MyMaternityHealth bundle delivers precise communication at key points of the pregnancy journey. When pregnant women are at a gestational age of 8-10 weeks, the program sends information-packed emails to proactively address common early symptoms and concerns, as well as develop a patient navigator connection to make it easier to ask questions. At a gestational age of 17-19 weeks, the patient receives communication tailored to address certain pregnancy-related conditions that can occur during this period, such as hypertension and gestational diabetes. The communication also includes content tailored to the patient’s support person—often a spouse or close family member—that serves to create an even larger team caring for the patient. This educational messaging, which features at-home care instructions and hints for self-care, ensures the patient is better able to manage their condition. These digital touch points are written from the voice of a bundles patient navigator who facilitates all non-clinical needs, such as booking appointments, navigating the health system, providing helpful resources and answering questions. VHES and Vanderbilt marketing teams work together to help employers communicate these benefits to employees in a personalized, non-jargon way.
The cross-departmental Employer Success Program features a library of assets detailing the bundles’ features and promoting their advantages to both an employee’s health and an employer’s bottom line.
This human-technology hybrid approach is working. In its first year using the MyMaternityHealth bundle, Metro Nashville Public Schools saw employee engagement with the program and communication increase. The average open email rate for MyMaternityHealth users was 68 percent with a click-through rate of 42 percent —significantly higher than the industry benchmarks where the open email rate for health care communications is only 22 percent and the click-through rate is 3 percent. In addition, maternity bundle patients gave the program rave reviews, as evidenced by an impressive Net Promoter Score of 95. Most important, these patients are experiencing dramatically better health outcomes, such as 25 percent fewer C-sections and 16 percent lower NICU utilization. The successful MyMaternityHealth model has expanded, with nine more bundles being added to the portfolio in the past three years: MyHearingHealth for cochlear implant surgery, MySpineHealth for spine surgery, three MyOrthoHealth bundles for osteoarthritis and knee and hip pain; two MyWeightLossHealth bundles for surgical and medical weight loss; MyUrologyHealth for kidney stone treatment; and MyRecoveryHealth for substance use disorders.
