Patient health extends far beyond the healthcare setting itself, and is subject to influence from a number of external factors that can range from everything from medication adherence to diet. Montefiore Health System realized this, and knew it needed a better population health approach that integrated these social determinants in a way that emphasized efficiency and quality of care.
Its solution, at least philosophically, was simple: leverage technology, artificial intelligence in particular, to better engage patients and enable better care coordination.
The health system had some specific goals in mind – to reduce preventable admissions, readmissions and ER utilization; to improve member satisfaction in terms of access, wait times and ease of scheduling; and to implement more preventative care, including wellness visits, breast and colorectal cancer screenings, and behavioral health interventions.
In a HIMSS20 digital presentation, Vanessa Guzman, associate vice president at Montefiore, said the new model was built around the idea of whole-person care.
“In addition to treating a patient’s health needs, their interactions with the health system are governed oftentimes by their life circumstances,” said Guzman, “and their ability to self-manage their health, their quality of immediate care and their overall relationship to the health system.”
Shara Cohen, vice president of customer experience for clinical effectiveness at Wolters Kluwer Health, said the first step in the process was establishing clinical and quality programs initially focused on areas such as cardiac and respiratory care.
“Something that distinguished Montefiore is the lengths they’ve gone to to support this whole-person care model, and recognizing that patient needs extend far beyond the clinical setting and longitudinally over time,” said Cohen.
To that end, the system leveraged social determinants of health assessments to connect people with community resources. It better enabled care management teams to provide support to patients and, importantly, built a technology infrastructure that can support these individualized activities at scale.
The starting point is identifying the patient, said Guzman. Once a patient with specific needs is identified, they can be enrolled in programs that assess those needs. At that point clinicians can personalize care plans.
Predictive analytics are used to segment the patient population. Navigators are placed across the system to see if a high-need patient shows up physically at a hospital or facility, and rising risks are flagged in the EHR. The navigators know who the patients are and what programs they belong to, what services are available to them and who their care teams are. They’ll work with the patient to create a discharge plan and get follow-up appointments scheduled before they’re handed off to the post-discharge phase.
The analytics tools provide care teams with actions they can take. High utilizers get enrolled in a care guidance process, and even those who aren’t in the highest-risk segment receive coaching on how to manage their chronic conditions and understand how their behaviors impact their overall health. Natural language processing is harnessed to call patients automatically and connect them to the appropriate resources in an efficient way. SDOH screenings and community-based organization referrals help to better address patient needs.
NLP processes enable the system to send relevant information back to the care team, which makes tier lives easier in that they can prioritize follow-up for high-need patients. They also benefit from red flags that indicate any changes in acuity.
Since implementation, Montefiore has benefitted from a 28% overall improvement in quality scores, a 6.8% reduction in non-user rates and 27% higher show rates for office visits. The system also projects a 12% positive impact on readmission rates over the next six months.
“It has helped us stay more in touch with what we’re good at,” said Guzman.
Email the writer: firstname.lastname@example.org