Thirty-Day Mortality Following Systemic Anticancer Therapy: Evaluating Risk Factors Without Selection Bias in a Real-World, Population-Based Cohort From 2009 to 2019.
Level 3 - non-randomized controlled study
Retrospective population-based cohort study without experimental intervention
PubMed 35400599 · doi:10.1016/j.clon.2022.03.015
What was done
The authors modified an existing quality indicator for 30-day mortality following systemic anticancer therapy (SACT) to remove selection bias caused by conditioning on future events. They evaluated the adapted indicator alongside the original Wallington et al. indicator in a population-based cohort comprising 10,213 patients treated with 16,622 SACT courses for common malignancies between 2009 and 2019 in the North Denmark Region.
What was found
Associations between clinical variables and 30-day mortality were similar between the original and adapted indicators, except in patients aged 75 years and older. The original indicator displayed shifting absolute risks when measured over quarterly versus yearly intervals. For the adapted versus original indicator, 30-day mortality was 1.0% versus 1.1% for curative SACT and 9.1% versus 11.7% for palliative SACT. Under the adapted indicator for palliative treatment, 30-day mortality exceeded 10% for gastroesophageal, pancreatic, and lung cancers, but remained below 4% for prostate cancer. Mortality values for the adapted indicator were significantly lower in the later study years compared to the 2009–2011 baseline period.
Why it matters
Adjusting quality indicators to avoid conditioning on future events provides clinicians with a mathematically unbiased assessment of short-term mortality risks when deciding on palliative or curative systemic therapies.
Limits
The abstract does not report exact effect sizes, odds ratios, confidence intervals, or specific p-values for the temporal decline and subgroup differences. The analysis is limited to registry data from a single Danish region, which may not generalize to health systems with different treatment selection practices or palliative care pathways. Individual causes of death and functional performance status were not reported in the abstract.
Cited by
- contradicts Nearly 50% of people treated for metastatic cancer are killed by the treatments rather than the cancer itself.