Risk of evaluation bias joins list of Covid-19 impacts on clinical trials

By on October 13, 2022 0

It has been well reported that the Covid-19 pandemic has had a direct impact on the execution of clinical trials, such as the inability of patients to travel to clinical sites or procedures affected by the lack of staff. Now questions arise about how best to analyze missing data due to perturbed assessments and retain statistical power without introducing bias.

Clinical trials that used time-to-event parameters are particularly prone to a disconnect between the scientific question the trial intends to answer and the statistical analysis. For example, progression-free survival (PFS), one of the most frequently used endpoints in oncology trials, is a major concern, say the authors of a recently published research paper.

Although PFS can be measured using clinical signs, symptoms, or tumor marker levels, radiographic assessments are the preferred method. In the scenario of a pandemic-caused global shutdown and extended periods between assessments, PFS is “vulnerable to lack of data,” the authors state.

Approaches to estimate PFS

To combat the inevitable loss of data and biased PFS data in the context of a pandemic, the authors of the article explore an estimation-based framework, a concept that was introduced recently in the addendum to the E9 guideline. (R1) of the International Conference on Harmonization. By accounting for intercurrent events that preclude outcome measures, this framework allows for a more precise definition of the estimation target and ensures that the statistical analysis is better aligned with the scientific question. Using a simulated six-month discontinuation as an intercurrent event, the authors investigate two approaches to dealing with missing data: a treatment policy strategy and a hypothetical strategy.

The treatment policy approach ignores intercurrent events and discloses disease progression with delayed assessments and other pandemic-related issues. The hypothetical strategy censors patients whose progression or death occurred upon discontinuation at the last assessment.

Censorship means patients will not be included in the final analysis due to factors such as loss to follow-up or toxicity, says Rachel Woodford, immuno-oncology researcher at Melanoma Institute Australia.

Although censoring patients allows researchers to complete the analysis in the presence of missing data, it also creates bias, adds Woodford. Disregarding patients who drop out of study drug due to toxicity could lead to approval or trial progression of a drug that may be unnecessarily toxic to other patients, says- she.

The authors found that although the treatment policy approach slightly overestimated the median PFS, which was expected since the assessments were conducted after the quit period, the statistical power was not affected. In contrast, the hypothetical strategy approach further overestimated median PFS times, and the smaller number of events decreased statistical power. In a case against censorship, the authors conclude that treatment policy strategy should remain the primary method of evaluation, with Woodford calling it a logical conclusion.

Endpoints affected by the pandemic

The sensitivity of the time to event endpoints to disruption poses a scenario with potentially misleading data from clinical trials that have been conducted during the pandemic. According to GlobalData’s Clinical Trials Database, 863 industry-sponsored clinical trials with PFS as the primary endpoint were initiated between June 1, 2019 and December 31, 2021. AstraZeneca topped the list of 10 leading sponsors measuring PFS throughout the pandemic, with 47 clinical trials. GlobalData is the parent company of Clinical Trials Arena.

The PFS is not the only endpoint that has been affected by the Covid-19 pandemic. Woodford explains that any event delay endpoint that required measurements that it could not be attended to could be error-prone, such as time to second progression (PFS-2) or time to second subsequent therapy (TSST).

Overall survival (OS) is easier to verify than PFS. However, in the context of the pandemic, patients might have died from SARS-CoV-2 infection and not from their disease, which could also affect the reliability of the data, she says.

“It is hoped that subsequent presentations and data releases will mention whether there have been any significant effects caused by the pandemic and missed visits,” notes Woodford.