Chris Wallis: Hello, and thank you for joining us for this UroToday Journal Club discussion. Today, we're talking about a recent publication entitled The Effect of Electronic Symptom Monitoring on Patient-Reported Outcomes Among Patients With Metastatic Cancer: A Randomized Clinical Trial. I'm Chris Wallis, an assistant professor in the division of urology at the University of Toronto. With me today is Zach Klaassen, assistant professor in the division of urology at the Medical College of Georgia. You can see here the citation for this recent publication led by Dr. Basch and published in the Journal of the American Medical Association.

In terms of advanced cancer treatment, we focus heavily on treatments that improve survival outcomes, but symptom management is a large portion of the care of these patients. Both the disease itself, as well as the treatments we administer, may contribute to a symptom burden. However, these are often unrecognized and untreated.

Symptom monitoring allows for the detection and treatment of symptoms, which may improve outcomes for cancer patients. Use of an electronic system to facilitate patient-reported outcome surveys can identify symptoms earlier that are treatable, allowing for early intervention. This may lead to improvements in physical function, symptom control, health-related quality of life, and potentially even hospitalizations and survival.

This study was designed as a multicenter cluster randomized trial among community oncology practices within the context of the Alliance for Clinical Trials in Oncology. Each site was asked to approach up to 50 adults with metastatic cancer of any type who were receiving treatment with chemotherapy, targeted oral therapy, or immunotherapy. Eligible patients must understand English, Spanish, or Mandarin, and the authors excluded patients with indolent lymphoma or acute leukemia, as well as those receiving hormonal monotherapies.

The participating practices were randomized in a one-to-one fashion to the use of this electronic symptom monitoring. Participating practices were stratified by their rural or urban designation according to the US census, and this randomization was performed using permuted blocks or block sizes of two or four. A cluster randomized approach was used to avoid influencing symptom management procedures among those people randomized to usual care.

The intervention comprised a patient-reported outcome metric. All practices received online access to standardized educational materials, whereas those randomized to receive the electronic symptom monitoring with patient-reported outcomes received access to an electronic survey system based at the University of North Carolina's PRO-Core facility. This included an assessment of PRO-CTCAEs. The symptoms included pain, nausea, vomiting, constipation, diarrhea, dyspnea, insomnia, depression, oral intake, performance status, falls, and financial challenges. Patients in the intervention group were asked to complete surveys weekly for one year and when they hadn't completed their regular survey, they received an email or an automated call that prompted them to do so.

Patients with PRO scores that reached the threshold of absolute magnitude or had a substantial worsening from prior received a link with patient level education materials, and here is an example of one for sleep-related problems. [inaudible 00:03:25] patients was also sent to their clinical team. In addition to these changes in PRO scores, ECOG performance status greater then two and ECOG worsening by two or more, falls, or financial distress over two also triggered alerts that were sent to the clinical team.

When the study was initially designed, there was two co-primary endpoints, including physical function at three months and overall survival. As of December 2020, the protocol was revised such that only overall survival was a primary endpoint and physical function moved to a key secondary endpoint, joining symptom control and health-related quality of life. There are additional exploratory endpoints, including ER visits and the duration of chemotherapy.

Assessments were performed for physical function, symptom control, and health-related quality of life at baseline and then at months 1, 3, 6, 9, and 12. These were performed using the EORTC's QLQ-C30 measure, and each was assessed on a 100-point scale with higher scores indicating better function.

In terms of statistical analysis, the authors initially specified 1,000 patients to be included. They subsequently increased this to 1,200 due to a increased available funding. In terms of overall survival, this sample size allows a 90% power to detect a hazard ratio of 0.76 with a two-sided alpha 0.05 using 576 observed events. When physical function was changed to the key secondary endpoint, 1,200 patients across 50 to 55 sites provides more than 90% power to detect a difference of 0.37 standard deviations between these groups, assuming an intracluster correlation coefficient of 0.055 and an 85% response rate at the three-month interval. The authors assessed the mean change in baseline function in physical functions, symptom control, and health-related quality of life at each visit using a general linear mixed model. In this model, study group, the time point, cancer type, and group-by-visit interaction were considered as fixed effects, whereas the practice was a random effect. The compound symmetric correlation structure was used to account for repeated observations.

They further performed a responder analysis among patients responding to the QLQ-C30 at baseline and at each time point, categorizing them as improved if their scores were at least five points higher than baseline, worse if their scores were at least five points lower than baseline, and stable otherwise. In a sensitivity analysis, they repeated this by using a 10-point change for the thresholds for worsening and improvement. The proportion of patients in each group were compared using cumulative logistic regression models with fixed effects of the group and cancer type and a random practice intercept. When interpreting these results, it's notable that there is no minimal clinically important difference that's been defined for response rates. While the primary endpoint was to find at 3 months, the authors also assessed differences between groups at 1, 6, 9, and 12 months. They performed a number of preplanned subgroup analyses.

Now I'm going to hand it over to Zach to walk us through the results of this trial.

Zach Klaassen: Thanks so much, Chris, for that great introduction. This is the participant flow in the PRO-TECT trial. You can see here that 52 practices were randomized, including 26 practices to the PRO intervention, of which 597 patients were included in that arm of the study, as well as 26 practices randomized to usual care, of which 600 patients were included in that arm of the study. Ultimately, you can see at the bottom here that 593 patients in the PRO intervention were included in the analysis, as well as 598 patients included in the analysis for the usual standard of care.

This is the participant characteristics. Essentially table one, it's a large table, so I've broken it down into two slides. We can see the characteristics on the left of this table, followed by to the right of that with the PRO patients into the far right, the usual care patients. You can see here that median age was roughly early to mid-60s. About 60% of patients were female, the majority of which patients were white at roughly 80%, but also did include roughly 16% of Black or African American patients.

Moving down to the middle of the table, interestingly, we can see that 63.7% of patients in the PRO group used the internet for their assessment and 36.3% used the telephone prompt. Education-wise, the most common level of education was roughly 35% of patients having some college, an associates degree, or other certification, whereas roughly 30% of patients were high school graduates or had a GED. With regards to employment status, roughly three-quarters of patients were not currently working at the time of the study. It of rural practice location, just over one-quarter of patients were categorized as rural. About two-thirds of patients were married or had a partner. And interestingly, roughly 20% of patients had never used email. Roughly 15 to 20% of patients never used the internet. And in terms of difficulty paying monthly bills, a wide range here. Not at all, so no difficulty paying bills, around 40% and around 10% of patients having very or extreme difficulty paying their bills.

This is the second part of the table, one which looks at cancer type. The most common cancers included in this study included colorectal, thoracic, and breast, and we can see here that prostate patients were included, 5.6% of patients in the intervention arm and 3% of patients in the usual care arm.

The next several slides will look similar to this one. This looks at the score distribution and model-based mean change from baseline at each assessment time point. And we can see here that the PRO intervention is in orange, the usual care is in blue. And the first assessment is for physical function. We can see at the top here, at baseline, there's no difference between these two arms. Again, at month 1, no difference, but we see an improvement in physical function, statistically significant improvement, at 3, 6, and 9 months, whereas at 12 months, there was no difference in physical function between the two groups.

A similar slide looking at symptom control, and symptom control, we do see a difference at one, three, six, and nine months between these two groups, favoring the intervention arm. Again, at 12 months, there is no difference in the intervention and the control arm. Finally, looking at health-related quality of life, similar to the symptom control, we see differences at 1, 3, 6, and 9 months, favoring the PRO intervention, and again at 12 months, we see no difference in the groups with regards to health-related quality of life.

This figure looks at the proportion of patients completing the surveys, as well as the patients remaining in the study, and completed surveys at each week of participation. This gray line at the top suggests a very good survey completion rate, and we can see that on the right here among the PRO group patients, 91.5% completed the expected weekly PRO surveys with no substantial reduction in completion rate over time. With regards to who completed the surveys, 84.4% were completed by the patients, 3.8% by caregivers, and 2.7% with the assistance of the staff from the study.

Several important discussion points for the PRO-TECT trial. Among patients receiving treatment at US community oncology practices, electronic symptom monitoring with PROs improved patient physical function, symptom control, and health-related quality of life at three months. This is important as identifying symptoms early via PROs and alerting clinicians to their presence facilitated interventions to prevent subsequent symptom worsening or complications. Secondly, this trial also demonstrated that high rates of patient survey completion can be attained in routine clinical practice, even when patients are ill.

The following should be considered within the context of the findings of this study. First, the effect size was smaller than demonstrated in a prior single center study. Secondly, the effect size diminished over time, as we saw that none of the comparisons were significant at 12 months, which may have reflected reduced benefits after one year of participation in the study. And finally, dropout rates were higher in the intervention group at 6.9% versus 1.2% in the control group, suggesting that some participants may have found the intervention burdensome.

So, in conclusion, in this report of secondary outcomes from a RCT of adults receiving cancer treatment, use of weekly electronic PRO surveys to monitor symptoms versus usual care resulted in significantly improvements in physical function, symptom control, and health-related quality of life at three months. Finally, these findings should be interpreted provisionally pending results of the primary outcome of overall survival.

We thank you very much for your attention and we hope you enjoyed this Uro Today Journal Club discussion of the recently published PRO-TECT trial in JAMA.