| Sumario: | Comorbidities are common in older patients with cancer. The Cumulative Illness Rating Scale for Geriatrics (CIRS-G, developed in 1992) provides a comprehensive assessment of comorbidities and is independently associated with the mortality rate, the hospital admission rate and functional limitations. Although the literature data indicate that the CIRS-G's reliability is good to very good, most of the studies were small and did not include cancer patients. Here, we evaluated the CIRS-G's reliability in a large, prospective cohort of cancer patients aged 70 or over. A stratified, random sample of 150 patients was used to assess inter-rater reliability, and a random sample of 30 patients was evaluated by two investigators to assess intra-rater reliability. For inter-rater reliability, the observed intraclass correlation coefficient (ICC) [95% confidence interval] of 0.53 [0.12–0.74] was substantially lower than the literature values. For intra-rater reliability, the ICC was 0.66 [0.40–0.82] for investigator 1 and 0.77 [0.54–0.89) for investigator 2. These findings suggest that applying the CIRS-G consistently across raters and centres may be challenging in real-world oncology settings for older patients with cancer. Revising the scoring guidelines could help improve its reliability.
|
|---|