Page 170 - Risk quantification and modification in older patients with colorectal cancer

Chapter 8
Our observed improvements in multiple domains of HRQoL including symptom scales after surgery could be related to the reduction of symptoms after surgery (i.e. therapeutic effect). An alternative explanation might be the ‘response shift phenomenon’, i.e. the reframing of the perception of their health over time, resulting in the reporting of improved HRQoL.42,43 Last, the majority of our patients were treated in hospitals were oncogeriatric care is extendedly integrated into the pre-, peri-, and postoperative care processes and this may influence the effect of surgery on physical functioning and quality of life.44
It is unlikely that the postoperative improvement in HRQoL was caused by loss of follow-up as the majority of these patients were not found to be the ones with the lowest scores at baseline. That improvements in the majority of the HRQoL scales were not significant at T12 for dependent patients, may have been caused by the lower number of respondents at this time point or that patients return to their original level of HRQoL indicating only a temporarily improvement.
Strengths of our study are the longitudinal follow-up of a cohort of older patients, with baseline assessment before surgery and a high response rate at follow-up (78%-87%). We reported on multiple time points showing a trajectory of HRQoL and using linear mixed-models we corrected for possible confounders resulting in more robust findings.
There are some limitations in our study that need to be addressed. First, we chose functional dependency as an indicator of possible frailty. However, frailty encompasses multiple domains, including cognitive and social functioning. The impairments in other geriatric domains may have further influenced HRQoL. Second, in our study patients had a Barthel Index score of 15 or higher, hence, patients that were highly dependent on care (score < 10) were not included, and neither were patients with acute surgery. Third, we cannot exclude the possibility that patients who did not respond or were not included in our study exhibited better or worse HRQoL, which would limit the generalizability of our results. Lastly, the interval between the HRQoL questionnaires was 3 months, and may not have captured the nadir of postoperative decline in HRQoL as was seen in earlier trials.14,45,46
168