Using Remote Monitoring to Reduce Hospital Visits for Cancer Patients
Until the 1990’s, patients with cancer requiring chemotherapy typically had to be admitted to the hospital to receive treatment. This inpatient care was stressful for patients and expensive. Since then, thanks to improvements in chemotherapy administration and symptom management, most treatment has moved from the inpatient setting to outpatient clinics, which prove more patient-centered and able to provide high quality care at lower cost. Today, aided by ongoing advances in treatment and monitoring technologies, cancer care is poised to make another transition: from the clinic into patients’ homes.
In October 2018, Memorial Sloan Kettering Cancer Center (MSK) launched a pilot program called InSight Care which enables seamless connection with its patients wherever they are, whenever they are in need. This pilot has had promising early clinical results and, just as important, has been well received by patients. Said one of the patients under its care, “The best part is that I don’t feel alone in this. They have been a lifesaver to me.”
Before we describe how the MSK program works, let’s look at the forces that are driving the move toward connected cancer care. Most current oncology-care models can’t effectively manage patients at home. As a result, those receiving chemotherapy in the hospital followed by stretches at home between treatments may have side effects or other issues that send them back to the hospital for acute visits that could have been avoided with proactive monitoring and management; on average, they have one unplanned hospital admission and two emergency department visits a year. In addition to reducing patients’ quality of life and delaying treatment, these acute admissions are expensive, accounting for 48% of total cancer care expenditures.
Patients receiving traditional oncology care wind up in the hospital for two main reasons: Traditional approaches often fail to identify those at high risk for hospitalization, and they are not designed to manage these patients once they’ve left the clinic. Thus the InSight Care program seeks to identify high-risk patients and provide digitally enabled, proactive, coordinated care before they need hospitalization. The digital connections allow us to keep patients in our line of sight beyond our walls and lowers barriers for patient communication with team members. To this end, it leverages three interlocking elements:
1. A novel risk prediction model
InSight’s predictive analytics framework was built from 10,000 observations of patients starting chemotherapy and has been refined to predict risk of acute hospitalization based on 270 patient characteristics spanning sociodemographics, the nature of the malignancy and treatment, lab results, medical and social history, medications and prior ED visits and hospitalizations. This model runs nightly on patients scheduled to begin chemotherapy and generates an email to the clinical team with each patient’s risk assessment from high to low — for example “This patient is deemed high risk (40.8% probability) for coming to urgent care for potentially preventable symptoms in the next 6 months.” The risk score is also integrated into the InSight Care digital platform via a web application, RiskExplorer, that displays the top ten characteristics that contribute to an individual patient’s risk. Patients who fall into the top risk quartile account for more than 50% of preventable inpatient bed days and would potentially benefit from more intensive symptom management.
2. Digital monitoring
Patients enrolled in the program receive a daily symptom survey through a patient portal based on our analysis of common symptoms leading to acute-care visits. These include pain, fatigue or reduced activity, difficulty eating or drinking, nausea, vomiting, constipation, diarrhea, dehydration, anxiety, and depression. The system alerts the care team when a patient reports mild/moderate symptoms (a “yellow alert”) or severe symptoms (a “red alert”), with red alerts requiring an immediate response. Prior work at MSK found that managing symptoms proactively in patients receiving chemotherapy leads to a 30% enhancement in quality of life, 7% fewer ED visits, and, most significantly, a 5-month improvement in overall survival. The survival benefit is thought to flow from patients’ reduced symptoms; with those better controlled, patients can spend more time on treatment. There’s also a potential psychosocial component; when patients are activated and educated about their symptoms, they are more engaged in treatment and better able to cope with their symptoms, which leads to better outcomes.
3. Digital team-based care
A dedicated, centralized cohort-management team consisting of oncology registered nurses and nurse practitioners acts as an extension of the primary oncology team with whom they collaborate. Unlike the primary team, these clinicians engage with patients exclusively through the digital platform which includes a dashboard for monitoring high-risk patients and allows the nurses to track symptoms. Using an internal web-based application, the Darwin Symptom Tracker, the team can monitor symptom trends and how these relate to chemotherapy. They can then connect with patients 24/7 as needed by phone, through the patient portal or by tele-visit. If a patient needs to be physically seen, the team can determine the right setting for evaluation.
Promising Results
As of August, 106 patients have participated in the program since the pilot launched. On average, patients complete more than 50% of the daily symptom assessments. Over the course of the pilot, the most common symptom generating a red alert and requiring immediate follow-up was pain (64% of participating patients reported severe pain during their enrollment). Patient interviews show that they have valued the speedy response to problematic symptoms, the 24/7 access to clinicians, and the convenience of avoiding in-person visits.
The pilot was not designed to study the program’s impact on acute hospital visits, but the data suggest a positive effect: Just 14 of the 106 high-risk patients (13%) in the pilot cohort went to MSK for acute care compared to 32% of those high-risk patients not enrolled in the pilot.
Going forward we plan to enroll more patients in the program to confirm these findings and to improve the model. We’ll be focusing on several areas including fine-tuning the identification of patients who could benefit most from the Insight Care model; determining the optimal cadence for patient-symptom assessments; and optimizing the interventions based on these assessments. More broadly, we are working to make Insight Care a learning model that encourages iterative patient and provider feedback and improves as a result. This includes finding the best ways to communicate analysis of the symptom data we collect to both patients and providers. Patients can cope better with symptoms when they understand them, and providers can integrate symptom data with other clinical reports to improve their treatment plans.
A more proactive, connected cancer care system will ultimately benefit patients, providers and society at large as cancer-care quality and patient experience improve and costs fall. Our Insight Care program is a promising initiative on that path which we hope other providers can learn from. Everyone wins when patients with cancer can avoid the emergency room and do better as a result.
The authors would like to acknowledge the contributions to this manuscript of Diane Reidy Lagunes, MD; Brett A. Simon, MD, PhD; Rori Salvaggio, RN, NEA-BC; Gilad Kuperman, MD, PhD; Ophelia Chiu, MHA; Isaac Wagner, MS; Margarita Rozenshteyn; Melissa Zablocki, MPH and the entire InSight Care team.