GPIN Members in the News
Geisinger doctors see 13% boost in all-cause mortality predictions with help from new AI tool
Matt O'Connor | February 08, 2021
A new deep learning tool based on hundreds of thousands of heart ultrasound images can help physicians predict patients’ one-year all-cause mortality, according to research published Monday.
Data scientists and cardiology experts at Geisinger Health System trained their convolutional neural network on echocardiogram videos from upwards of 34,000 patients. The tool beat out a number of cardiologist-based predictions of mortality and a machine learning model constructed from hand-picked variables.
Co-senior author of the paper Chris Haggerty, PhD, an assistant professor in Geisinger’s Department of Translational Data Science and Informatics, said this study is one of the largest medical imaging datasets ever published.
"We were excited to find that machine learning can leverage unstructured datasets such as medical images and videos to improve on a wide range of clinical prediction models," Haggerty said in a statement.
The team explained that imaging has become one of the most “data-rich” components of the electronic health record, with a sole heart ultrasound producing nearly 3,000 individual images. The average cardiologist or radiologist, however, has little time to look over all these pictures and analyze additional clinical data.
With this in mind, the Danville, Pennsylvania-based team trained their model using 812,278 echocardiogram videos gathered at their institution over the past decade.
When put to the test, their model’s predictions outperformed a handful of other metrics, including the Seattle Heart Failure score, widely used pooled cohort equations, and an AI tool combining 58 manually chosen variables from echo images and 100 clinical EHR data points.
And proving the notion that AI can augment human intelligence, cardiologists using the model improved their one-year all-cause mortality predictions by about 13%. The model bolstered their sensitivity while maintaining specificity.
With larger datasets and further testing, the team hopes to bring its machine learning platform to real-world clinical settings.
"Our goal is to develop computer algorithms to improve patient care," Alvaro Ulloa Cerna, PhD, senior data scientist at Geisinger, added Monday. "In this case, we're excited that our algorithm was able to help cardiologists improve their predictions about patients since decisions about treatment and interventions are based on these types of clinical predictions."
Read the entire study published in Nature Biomedical Engineering here.
AdventHealth meets CMS quality metrics, saves money with digital health platform
By Bill Siwicki | January 08, 2021
With the CMS Hospital Readmissions Reduction Program, hospitals can lose up to 3% of their Medicare payments annually due to penalties. But reducing readmissions poses challenges – especially when working with vulnerable populations like Medicare.
AdventHealth is driven by a mission to deliver compassionate care for body, mind and spirit. This includes caring for Medicare populations, who often face unique challenges when it comes to managing their chronic conditions, leading to higher levels of hospitalization and high HRRP penalties for the organization.
Wanting to improve outcomes for these members, AdventHealth needed a way to increase disease control – reducing the need for expensive acute care and readmissions.
Significant correlations have been found between low post-discharge medication adherence and 30-day readmissions, said Joey Pineda, a case manager at AdventHealth and designated project manager for implementing a pilot program surrounding these challenges.
Understanding this, the team decided to pilot a program to reduce readmissions by increasing medication adherence with a digital health platform, Wellth.
"The pilot, which took place at the AdventHealth DeLand, Florida, acute care hospital, engaged Medicare patients with acute myocardial infarction, pneumonia, chronic obstructive pulmonary disease or congestive heart failure who also had recent 30-day readmissions," he explained.
"Participating patients would utilize the Wellth platform daily – completing daily check-ins and submitting photos of them each time they took their medications as prescribed – in exchange for a financial incentive."
The platform itself leverages principles of behavioral economics, as in a study suggesting that, even when educated, people make irrational decisions (i.e. skipping medications), to increase medication adherence.
Patients received up to a set amount of reward money at the start of each month and had money deducted each time a prescribed medication check-in was missed. At the end of each month, the patient was free to spend whatever money they had kept.
MEETING THE CHALLENGE
The Wellth platform was designed to minimize barriers to participation for AdventHealth patients – requiring no additional connected devices – and could be accessed via an app on each individual's existing smartphone.
Through the app, enrolled patients received regular reminders to complete their medication check-ins, submitted photos of themselves taking their medications and received personalized bonus incentives in the form of messages of encouragement from family and friends.
"Each picture submitted by the patient was checked using in-app AI to confirm the medications and amounts taken, and Wellth member-support team members were alerted in the event that a picture submitted was out of the ordinary," Pineda explained.
"These support-team members could then reach out to individual patients to see what the issue was, and refer those patients back to us – their hospital care manager – at DeLand for appropriate follow-up."
Having the AI to identify and send alerts for missed check-ins and incorrect dosages allowed the Wellth support team to quickly step in, assess and engage DeLand staff early on, before acute care was needed, he added.
The DeLand team was also able to set specific parameters for alerts and notifications, reducing alert fatigue and ensuring that the data sent was that most valuable and relevant for its needs, he said.
Prior to the Wellth program, the average 30-day readmission rate for AdventHealth DeLand was 19% for patients with the four targeted disease states.
Within the first 30 days of the program, however, medication adherence for enrolled patients improved significantly, reaching an average of 91% adherence. This not only allowed DeLand to achieve the 80% CMS quality metric for medication adherence, but improved the individual patient outcomes, as well.
"With the pilot, our DeLand location saw a 57% reduction in readmissions for these previously high-risk patients, taking their readmissions from 19% to 8%," Pineda said. "As readmissions dropped, so did the cost of care for these patients, ultimately saving the organization $183,000 in care costs during the length of the pilot."
ADVICE FOR OTHERS
"First, while the program was helpful in achieving adherence across the board, it was most impactful for patients who had a history of poorly or moderately controlled conditions," Pineda advised. "In treating the mind, body and spirit of your patient populations, start with those populations that are struggling the most rather than trying to find a one-size-fits-all solution."
By investing resources into these higher-risk patients with solutions designed specifically for them, AdventHealth saw better health outcomes and an increased return on investment as a whole, he noted.
"Second, the beauty of this technology is that it worked with what the patients already have – smartphones," he said. "This made implementation quicker on our end, with minimal disruption to our workflow, and enrollment easier for our patients. Connected devices, though they have their place, can be expensive and don't have to be your only option for adopting digital health."
Finally, the goal has to be patient-centric, he said.
"Our pilot, while cost-effective, was ultimately about creating healthier patients through fewer readmissions," said Pineda. "When we focused on improving the health of these individuals, the readmissions, quality metrics and cost savings all just fell into place."
SSM Health Medical Minute – Kangaroo Care and massage therapy is proven to help NICU preemies
by: Dan Gray | Posted: Jan 6, 2021
Two proven methods for preemies inside the NICU, Neonatal Intensive Care Unit at SSM Health Cardinal Glennon Children’s Hospital are being used every day to help their patients thrive and achieve positive outcomes.
The first is Kangaroo Care, a method that encourages either parent to hold the baby with skin-to-skin contact. The baby is typically naked except for its diaper and is place in an upright position against the chest of the parent for a few hours.
Dr. Amit Mathur, a SLUCare Neonatologist at SSM Health Cardinal Glennon Children’s Hospital, says that Kangaroo Care has proven many benefits including developing a close bond with the parents, but it is also medically beneficial. Babies held close to their mother’s body not only have a higher survival rate, they thrive and achieve development outcomes faster.
Kangaroo Care is encouraged at SSM Health hospitals with NICU babies and healthy newborns because the outcomes include:
- Stabilizing baby’s heart rate
- Improve baby’s breathing pattern and making breathing more regular
- Improving oxygen saturation levels
- Gaining in sleep time
- Experiencing more rapid weight gain
- Decreasing crying
- Having more successful breastfeeding episodes
- Having earlier discharge from the hospital
While the baby benefits, so does the mother. Kangaroo Care improves bonding with your baby in addition to:
- Increasing your breast milk supply
- Increasing your confidence in the ability to care for your new baby
- Increasing your confidence that your baby is well cared for
- Increasing your sense of control
Massage therapy is also being used in the NICU to help babies with physical touch. Often, preemies are attached to life saving equipment and often spend many months outside from a natural environment. These preemies often associate touch with negative experiences as they need to be hooked up, poked or prodded to help them medically.
Massage therapy provides positive long-term development of the infant and gives families the positive opportunity to touch and feel their baby to bond while helping their baby maximize sleep, comfort and neurodevelopmental outcomes. Dr. Mathur says that touch is one of the most developed senses at birth. Touch provides support and pleasure to the baby to provide stress relief and parent-infant bonding.
SSM Health Cardinal Glennon Children’s Hospital NICU is nationally ranked for care as one of the “Best Children’s Hospitals” ranked by U.S. News & World Report for 2020-2021.
Rapid Sepsis Testing Generates Value-Based Care Gains at AHN
BY CHRISTOPHER CHENEY | JANUARY 29, 2020
Rapid sepsis testing at Allegheny Health Network has boosted care quality with improved clinical outcomes and reduced costs.
Sepsis is the body's extreme reaction to an infection, which can result in life-threatening symptoms such as multiple organ failure. Annually, more than 1.7 million people get sepsis in the United States, with about 270,000 fatalities, according to Centers for Disease Control and Prevention statistics. One-third of patient deaths in hospitals involve sepsis, the CDC says.
Time to effective treatment is a critical factor for patients infected with sepsis bacteria, says Thomas Walsh, MD, medical director of the Antimicrobial Stewardship Program at Pittsburgh-based AHN. "Every hour delay in antibiotic administration is associated with decreased survival."
For the past eight months, AHN has been using the Accelerate Pheno test system to analyze bloodstream infections and determine the best antibiotic therapy for sepsis patients.
Before adopting the relatively new technology, Walsh says it could take AHN two to five days to detect sepsis and match the strain of sepsis bacteria to a narrow-spectrum antibiotic. Now, that process has been shortened to seven hours.
In addition to cutting time to treatment, which improves clinical outcomes, reducing the use of broad-spectrum antibiotics has significant benefits, he says. "We can avoid the downstream effects of unnecessarily broad antibiotic use such as propagating antimicrobial resistance and higher rates of Clostridium difficile."
Antimicrobial resistance to antibiotics is one of the most daunting public health problems of this generation, the CDC says.
GENERATING POSITIVE RESULTS
For sepsis patients who were not treated in an ICU, AHN has posted several clinical gains, Walsh says.
"We dropped the time that we were able to identify bacteria from 39 hours to 90 minutes. We were able to decrease the time to knowing which antibiotics would be optimal from 46 hours to 7 hours," he says.
Walsh continues, "For patients who were initially on an inadequate antibiotic, we were able to reduce the time to get them on effective antibiotics from 51 hours to 11 hours. We reduced our length of stay from 8 days to 5.5 days. Our total duration for antibiotics went from 14 days to about 9.5 days."
Similar results have been achieved for sepsis patients treated in an ICU, he says. "For patients who were critically ill who were on inadequate antibiotics initially, we dropped the time to effective antibiotics from 43 hours to 12 hours. That led to a two-day drop in length of stay for those patients. For duration of antibiotics use in the ICU, we went from 15 days to 10 days."
The Accelerate Pheno testing has reduced cost of care, Walsh says. "For these kinds of rapid tests, to run one of the tests is usually between $150 and $200. The cost of being in the hospital is usually between $600 and $1,000 per day. If a patient is in an intensive care unit, the cost is usually between $1,000 and $2,000 per day. So, if you can use this new technology and get patients home two days quicker, you are saving about $1,000–$2,000 per day."
The rapid testing also has reduced medication costs, he says. "We are using less broad-spectrum antibiotics, which tend to be more expensive than narrow-spectrum antibiotics."
INCORPORATING RAPID TESTING INTO THE SEPSIS CARE PATHWAY
The rapid testing technology must be combined with efficient workflows, Walsh says. "For us, the critical part was tying this testing to our antimicrobial stewardship team, which is a team of infectious disease doctors who help our bedside physicians use the appropriate antibiotics to maximize our clinical benefit while minimizing the collateral damage of broad-spectrum antibiotic use."
He says there are three primary steps in the care pathway associated with the rapid testing:
- Once a blood culture flags positive for possible sepsis bacteria, microbiology technicians start the Accelerate Pheno testing and call nurses on the floor to alert them that bacteria is growing in the patient's blood and test results will be available within seven hours. The technicians also page the antimicrobial stewardship team, so they are aware as well.
- The patient is given a broad-spectrum antibiotic as soon as possible.
- When the testing results are available, the technicians call the nurses on the floor, who relay the message to the patient's attending physician that sepsis bacteria have been matched to effective antibiotics. The stewardship team is also alerted, and an infectious disease clinician and a pharmacist review the test results and the patient's medical record. Then the stewardship team members call the patient's care team to make recommendations for antibiotic administration.
"The antimicrobial stewardship team plays a key role. It acts as an intermediary between the technology being performed in the lab and how we act on that information at the bedside," Walsh says.