At the 26th Annual Conference of the International Society for Quality of Life Research, 2 groups of US Food and Drug Administration (FDA)-affiliated researchers presented abstracts for studies aiming to test the limits of patient-reported outcomes (PROs) as a useful metric in the assessment of the efficacy of certain treatment approaches in cancer. One study investigated whether patient awareness of treatment administration, or unblinding, impacted reporting of PROs,1 and another questioned whether PROs are truly a stable metric for a patient’s deteriorating condition.2
That second study took a closer look at how PROs are currently used (and could be used in the future) in cancer drug development. While these data could be harnessed in the pursuit of regulatory approval for a drug, or as surrogate end points in the evaluation of the effect of a treatment — both seemingly positive uses of the information — the use of PROs as a metric to gauge quality of life may also be susceptible to gaming.
The evaluation of PROs, then, must be standardized — especially because PRO data are increasingly finding their way into pharmaceutical companies’ drug trials and across presentations of pivotal cancer trial results.
“These days, an industry sponsor is putting itself at a disadvantage if it does not have expertise on PROs or include PROs in development programs,” said Ethan Basch, MD, MSc, director of cancer outcomes research at the University of North Carolina-Chapel Hill’s Lineberger Cancer Center, who wasn’t involved in either study. “This is a fast-growing field of increasing emphasis … Many patients have baseline impairments that should be measured by PROs so as not to be misattributed to products.”
The first of the 2 aforementioned abstracts presented at the conference aimed to suss out whether bias is detectable in PROs across open-label trials in oncology compared with other trial designs.1
Part of the impetus for the research was that, as the abstract noted, open-label cancer trials — where researchers and participants are both aware of the parameters of the trial, and treatment assignment is unblinded — are used more often in cancer to bolster research than in other clinical settings. If bias were detectable, it could have a marked impact on health outcomes for cancer patients specifically.
The study enrolled patients with multiple myeloma and sorted them into 2 groups. One group received treatment by way of an open-label design, while the other received the same treatment through a double-blind trial. PROs were assessed twice: at baseline, or before patients were randomly assigned to 1 of these 2 treatment groups, and again after treatment was assigned (but before treatment actually began).
“PRO Domains of Interest”
Patients were asked to report at both data collection times any emotional, physical, and social symptoms they were experiencing, as well as their general quality of life. All of the questions were framed around symptoms that would be relevant to their forthcoming treatment, but patients were unaware of this when they filled out the survey for the first time. What researchers were looking for in both groups was any change in PROs between the first and second time they answered the questions.
Ultimately, when posttreatment answers from patients in the open-label trial were compared with those in the double-blind trial, the investigators determined that there were no significant differences in reporting between the arms. Patient knowledge of the treatment that was administered did not appear to bias PROs.1
This is good news for open-label cancer drug trials that include PROs as domains of interest, and Dr Basch said that he believes the FDA ultimately values these patient-reported details.
“Some have objected to [the FDA’s] rejection of quality of life as a construct in drug approvals, which is inconsistent with the European Medicines Agency,” he pointed out. “But overall, I feel the FDA has done much to advance the field of PROs in oncology regulatory science.”
Not everything is smooth sailing, though. In the second study presented at the same conference, the researchers (some of whom also participated in the first PROs study discussed in this overview), showed evidence that using PROs to monitor patient deterioration during trials is not always reliable.2
While it could be logical to assume that patients reporting deterioration in physical function would die during the trial, the researchers showed that this is not actually always the case: some patients are capable of later returning to a state of non-deterioration. They cautioned that a reversal or pause in deterioration is not necessarily an indicator that the drug is suddenly working again. Instead, the improvement in physical function could potentially be attributed to “treatment interruption, dose modification, or other clinical interventions” that are not adequately disclosed along with the PROs.
When questioned about what type of systemic change would be required to make PROs more useful in drug trials, Dr Basch said that “input from regulators would be optimal, to minimize missing data and to maximize meaningfulness.”
“If PROs were systematically collected during routine cancer care, that would both improve quality of care and provide a trove of data for analyses,” he added.
- Roydhouse J, Mishra-Kalyani P, Bhatnagar V, King-Kallimanis B, Kluetz P. Presented at: 26th Annual Conference of the International Society for Quality of Life Research; October 20-23, 2019: San Diego, CA. Does knowledge of treatment assignment affect patient-reported outcomes? An evaluation of open-label bias in multiple myeloma. Qual Life Res. 2019;28(Suppl 1):1-190. Abstract B204.5. doi: 10.1007/s11136-019-02257-y
- Roydhouse J, Lee H, Cheng J, Bloomquist E, Bhatnagar V, King-Kallimanis B, Kluetz P. Presented at: 26th Annual Conference of the International Society for Quality of Life Research; October 20-23, 2019: San Diego, CA. 26th Annual Conference of the International Society for Quality of Life Research. Evaluating time to physical function deterioration in multiple myeloma. Qual Life Res. 2019;28(Suppl 1):1-190. Abstract 303.4. doi: 10.1007/s11136-019-02257-y
This article originally appeared on Cancer Therapy Advisor