How a poor evidence base can distort decision-making in cancer

New cancer treatments are typically expensive, prompting committees worldwide to evaluate their value carefully. Approval is often based on clinical trials that show improvements in progression-free survival (PFS) — the time from treatment initiation to when the treatment is deemed to have failed. 

Every year, pharma and biotech companies introduce new cancer treatments with strong evidence of efficacy. Huge progress has been made, but the length of PFS is not enough to justify access. 

The value of new treatments is typically assessed using three variables:

1️⃣ Does the treatment lead to a gain in progression free and/or overall survival?

2️⃣ How does the new treatment affect patients’ health-related quality of life (HRQL)? 

3️⃣ How much does the treatment cost and are there savings elsewhere?

To make informed decisions, we need evidence for all three variables from the start of treatment through to the end of life. But this poses some significant challenges. 

Measuring HRQL in a patient’s final months of life may be considered intrusive, inappropriate, or even unethical. It is often not done. As a result, there is relatively little high-quality data regarding how HRQL changes during this time.

Analysts frequently make two errors:

🚫 The modelling teams make assumptions regarding how HRQL declines as people approach death, often assuming a rapid deterioration in HRQL after treatment failure.

🚫 The modelling teams rely on poor-quality data — often simply because it has been included in previous submissions. 

The reality can be quite different. Patients can maintain good HRQL with effective symptom management by clinicians. 

Two studies we have worked on with colleagues support this suggestion that HRQL can be maintained quite well. 

The first study [1] reanalysed trial data from women with metastatic breast cancer. Trial data were used to describe HRQL before and after women’s cancer had progressed. This study reported that, in both treatment arms, women’s HRQL was maintained quite well following progression. This result contrasts with widely-used data, which suggests quite a sharp decline in HRQL. 

The second study [2] in men with prostate cancer also assessed patients before and after progression of disease in an observational study. The study noted that, following completion of chemotherapy, men reported similar HRQL values to those reported by other men before chemotherapy. 

Important decisions regarding the adoption of new treatments in cancer require more robust HRQL data. And addressing this gap is a worthwhile initiative — one that could benefit from collaboration between industry, advocacy groups and agencies like NICE.