Safety of Outpatient CAR T-Cell Therapy May Be Improved With Integrated Management Process

The utilization of remote patient monitoring within the first 30 days of outpatient CAR T-cell therapy can promote patient safety, reduce costs, and decrease hospitalization rates by allowing patients with hematologic malignancies to connect with a virtual care platform, according to data presented as during the 2023 ASH Annual Meeting.¹

“We demonstrated that we can safely manage CAR T patients in the outpatient setting using a virtual care [remote patient monitoring] platform and enabling patients to spend more time at home,” Tonya Cox, BSN, RN, OCN, assistant vice president of operations at Sarah Cannon Transplant and Cellular Therapy Network in Nashville, said during the presentation of the data. “We also significantly reduced hospital days and were able to use them when they were most needed.”

Key Takeaways

Investigators reported on the percentage of patients who needed to be hospitalized during the program.

The highest percentage of patients who needed hospitalization was 93% for patients treated with ciltacabtagene autoleucel (Carvykti; median hospital stay, 4 days), and the lowest was for those treated with lisocabtagene maraleucel (Breyanzi; 63%; median hospital stay, 2 days).

Rates of ICU admissions ranged from 0% in those from the tisagenlecleucel group (median hospital stay, 4 days) to 44% in those treated with brexucabtagene autoleucel (Tecartus; median hospital stay, 6 days). The rates of ICU admission with ciltacabtagene autoleucel, lisocabtagene maraleucel, and axicabtagene ciloleucel (Yescarta) was 28%, 17%, and 32%, respectively.

Notably, the average planned length of stay for patients receiving inpatient CAR T-cell therapy recipients between January and June of 2023 was 16 days. Through the outpatient program, the median number of hospital days was 4 days.

Of the 100 patients in this study, 70% developed cytokine release syndrome. Most of the patients who developed cytokine release syndrome were considered grade 1 or 2 (30% an 36%, respectively). Among those who never developed cytokine release syndrome, half still experienced a hospitalization for reasons like tachycardia and hypoxia.

In patients that were hospitalized, the days from infusion to hospitalization ranged from 1 to 6 depending on the treatment they received.

Researchers extracted data from 100 patients who participated in the virtual care platform, which included phone and video call logs with virtual nurses and clinical data. The median age was 61.2 (± 13.3 years) and 62% were male.. The data were obtained between February 20, 2023, and June 15, 2023. Summaries were prepared on patient adherence, inpatient admission, calls, alarms, and clinical metrics.

Researchers derived information on patient adherence from tasks/wear time completed vs prescribed.

Patients who completed the outpatient CAR-T program had diagnoses including non-Hodgkin lymphoma (60%), acute lymphoblastic leukemia (11%), and plasma cell disorder (29%). All evaluated patients had received CAR T-cell therapy, including lisocabtagene maraleucel (24%), axicabtagene ciloleucel (25%), brexucabtagene autoleucel (18%), ciltacabtagene autoleucel (29%), and tisagenlecleucel (Kymriah; 4%).

Patients were monitored on the program for a median of 17.2 days (range, 11.6-29.9). Overall wearable adherence was 80.7% (range, 71.4%-89.5%); the wearable adherence rate was 66.4% in the day (range, 48.0%-83.3%) and 89.7% at night (range, 79.3%-94.0%).

Cox noted that patients were asked to remove their wearable devices when they were in the clinic, which affected the daytime adherence rates.

The median survey completion was 58.4% (range, 33.3%-88.8%) and compliance to blood pressure measurements was 98.8% (range, 82.2%-100%). Of note, out of 2,040 patient days, 100% compliance to blood pressure measurements (with 3 readings a day) was obtained in 63.2% of days.

“If patients don’t measure their blood pressure 3 times a day, they are prompted by the virtual nurses to do so,” Cox explained.

The Learning Curve

During the presentation, Cox noted that there was a learning curve for clinical teams to recognize the difference between intermittent and continuous monitoring.

“With intermittent monitoring in a clinical setting, there’s a fair amount of control,” she said. “For example, patients are usually stationary. With continuous monitoring, the patient could be going up and down the stairs, they could be playing with their grandkids. It represents real life. So, the parameters being monitored need to be considered over a period of time, usually 30 or 60 minutes, and they need to be coupled with other parameters to ‘reduce the noise’ for the virtual RN.”

Certain alarms were in place to let nurses know when the patient’s condition had changed. For instance, if the patient’s median pulse rate was less than or equal to 45 for at least 30 minutes, they were flagged for bradycardia.

For many of these “alarms,” the median number of flags that were reported per patient-day, were low. For bradycardia, tachycardia, hypoxia, pyrexia, hypertension, and hypotension, the median was 0. However, there were some patients who had a significant number of alarms; on one day, for example, 8.60 alarms came through for bradycardia. According to Cox, this shows that the “alarms” need to be correspond to a patients baseline status to reduce the amount of “noise” that the RN has to navigate.

The alarm with the most value, Cox said, was the pyrexia alarm, as that led to the greatest number of calls to providers after hours.

Overall, however, there was an average of 11 calls per month going to providers.

“This was not adding a burden to providers to use this technology,” Cox said.

Cox highlighted some of the future opportunities she and her team plan on focusing on regarding this project, including reducing the number of daily clinic visits, investigating product-specific patterns in the dataset, refining the alarm triggers, and individualizing alarms based on a given patient’s baseline.

Giving Patients Control and Independence

Cox and her team had sought to establish an outpatient CAR T-cell therapy administration because their experience has demonstrated that patients have a better experience when they have more control and independence over their care.

Outpatient CAR T-cell therapy administration is also desirable because it offers patients greater mobility and lowers the risk of hospital acquired infections. Further, as CAR T-cell therapy increases in popularity as a treatment modality, certain limitations in institutional capacity may emerge. The inpatient team may need to focus their attention on the patients who experience treatment-related complications, such as cytokine release syndrome.

“We had concerns about our capacity with the growing number of products in clinical trials and with therapies offered [in earlier line settings],” Cox said. “We know that we need to focus those inpatient days when they are most needed [because] it also lowers costs and resource utilization.”

Patients were eligible for outpatient CAR T-cell therapy if they were 18 years or older, lived between 30 and 60 minutes away from a treating hospital, always had a caregiver with them, and if their disease did not have any central nervous system involvement.

Once cleared and following their CAR T-cell therapy administration, the patients proceeded to have in-person clinic visits during the high-risk period (days 1 through 14) and engaged with their nursing team virtually for the duration of the program (30 days).

The remote monitoring kit included a wearable device that transmitted vital signs (skin temperature, O2 saturation, pulse, and respiratory rate) to their health care team, an axillary temperature patch, a tablet, and a blood pressure cuff. Patients were also provided with an FDA-approved oral thermometer.

Patients were asked to always wear devices outside of clinic visits, complete regular surveys on their tablet, and take blood pressure readings 3 times a day. During this time, patients were able to connect with their health care providers via video or chat if they needed.

Cox highlighted that part of this remote patient monitoring platform and kit included a hub with cellular data capability.

“The patient does not need WiFi,” she explained. “This is really important from an equity standpoint; we don’t have to worry about what the internet capabilities are where the patient is staying.”

Patients were given the remote patient monitoring kits 3 days before undergoing the CAR T infusion, “which enabled them a few days to get comfortable with the technology, to work with the technical support team if they needed to so that by infusion day, they were being continuously monitored,” Cox said. All patients were monitored for a minimum of 14 days and for up to 30 days. The physician determined the total monitoring time.

Clinical pathways for remote monitoring were established with a multidisciplinary taskforce, who developed parameters for virtual nurse check-ins, alarms, and escalation of care to the ER or clinic. Virtual nurses were available to respond to patient concerns, contact them based on vital sign trends, and triage consistent with clinical pathways. The nurses were able to escalate care, as appropriate.

Reference

Cox T, Zahradka N, Billups RL, et al. Standardization of Outpatient Care after CAR-T Therapy across a Large Cell Therapy Network- through Technology and Decentralized Virtual Nurses: Preliminary Results. Blood. 2023;142(1):254. doi:10.1182/blood-2023-187253