Physical health-related quality of life trajectories over two years following breast cancer diagnosis in older women: a secondary analysis

The number of women diagnosed with breast cancer aged ≥ 65 years in the U.S. is rising substantially, largely due to the aging population [1]. By 2030, the estimated number of new cases of invasive breast cancer diagnosed in women aged ≥ 65 years in the U.S. is projected to be 179,000, representing a 57% increase from 2010 [2]. Given this demographic increase, greater attention has been directed at the impact of cancer on the health-related quality of life (HRQoL) in older cancer survivors [3‐5].

HRQoL refers to the impact of health on the physical, emotional, and social domains of life [6]. In general, older age is associated with poorer physical HRQoL [7, 8], which may be attributed to a greater number of medical comorbidities that can impact health, physical functioning, and survival [9]. Older cancer survivors are even more vulnerable to poorer physical HRQoL compared to age-matched adults without a history of cancer [10, 11]. Population-based case–control studies suggest that women with breast cancer aged ≥ 65 years report significantly poorer physical HRQoL compared to their matched counterparts without a history of cancer, particularly in the early years following breast cancer treatment [11‐13].

Despite a growing body of literature on the HRQoL of older women with breast cancer [12, 14‐24], the majority of these studies examine older patients as one age group. However, individuals aged ≥ 65 years comprise a highly heterogeneous group, varying in both physical health status and functioning [25‐27]. Examination of this heterogeneity may help to identify characteristics of survivors who have different patterns in physical HRQoL trajectories and, consequently, identify those patients most at risk. To date, four studies have examined the physical HRQoL trajectories of breast cancer survivors [28‐31]. None of these studies, however, focused on women age ≥ 65.

The present study extends the literature by exploring physical HRQoL trajectories with a specific focus on older women diagnosed with breast cancer. We include variables that others have found to be related to heterogeneity in HRQoL such as breast cancer treatment, medical comorbidities, and psychosocial variables [28, 30, 32]. Our objectives were: (1) to determine physical HRQoL trajectories among breast cancer survivors aged ≥ 65 years; and (2) to identify sociodemographic, cancer-related, health-related, and psychosocial factors associated with trajectory group membership.

This study examined physical HRQoL trajectories in women aged ≥ 65 years diagnosed with breast cancer. We found two distinct PCS trajectories in our sample. One group, constituting 58% of our sample, had high PCS following diagnosis that improved slightly over time, and the second group (42%) had consistently low PCS that did not improve over time. In a multivariable model, older age, difficulty paying for basics, higher BMI, and greater number of medical comorbidities were predictive of low PCS group membership. Cancer-related and treatment-related factors were not significant predictors, nor were any of the psychosocial variables still significantly associated with trajectory group in the multivariable setting.

In the first 2 years after a breast cancer diagnosis, the majority of survivors in our sample had similar or better physical HRQoL than population reference norms of women ≥ age 65 without a history of cancer [7, 8]. This is an important finding, as cross-sectional and longitudinal studies that do not consider heterogeneity among older breast cancer survivors women suggest that older survivors in general, report significantly lower physical HRQoL compared to age-matched adults without a history of cancer [11‐13].

Nonetheless, 42% of women were in a low PCS group that did not improve over time. In fact, survivors in this group scored six to nine points lower on PCS compared to population reference norms of older women without a history of cancer [7, 8], exceeding the minimally important difference of two to three points on the PCS [11, 46]. In contrast to other studies of HRQoL trajectories [29, 31, 32], we did not find a group that declined over time. It is possible that a longer follow-up time would show greater heterogeneity and a declining group [31].

Consistent with previous research [15, 29, 30, 32, 47], we found several sociodemographic, health-related, and psychosocial variables that significantly distinguished between trajectory groups in bivariate analyses. Cancer-related variables (e.g. cancer stage, treatment) did not significantly predict group membership [29, 31]. In the multivariable analysis, older age, difficulty paying for basics, higher BMI, and greater number of comorbidities were significantly associated with the low PCS trajectory group [15, 29, 32, 47].

It is also worth noting that although a small percentage of survivors reported difficulty paying for basics, women in the low PCS group had six times the odds of having a somewhat/very hard ability to pay for basics compared to the high PCS group. This finding is consistent with the literature on associations between low socioeconomic status and poor physical HRQoL in breast cancer survivors [48‐51]. Incorporating screenings such as the Comprehensive Score for Financial Toxicity-Functional Assessment of Chronic Illness Therapy [52] may be an efficient way to identify vulnerable patients at risk for financial toxicity [53]. This may result in earlier referrals for social workers, financial counselors, transportation vouchers, and co-pay assistance resources to decrease financial toxicity [54].

BMI was another significant predictor. Women in the low PCS group were significantly more likely to be overweight compared to the high PCS group, which averaged in the normal BMI range. This finding is consistent with the literature on the relationship between obesity and physical health [55]. Obesity is a risk factor for chronic health conditions including breast cancer, and is associated with greater treatment complications, less effective treatment results, likelihood for recurrence and increased mortality [55, 56], all of which may contribute to poorer physical HRQoL.

The finding that number of comorbidities was a highly significant predictor of PCS is consistent with other studies [31, 32]. More than half (57%) of cancer survivors in the low PCS group reported having at least three medical comorbidities, with two-thirds of women in this group having arthritis or hypertension. In fact, survivors in the low PCS group were more than twice as likely to have arthritis, cardiovascular, gastrointestinal, or respiratory disease, all of which negatively impact physical HRQoL [57‐59]. Given that presence of comorbidities may limit treatment options, increase risk of toxicity, and negatively impact survivorship [60], identification of specific comorbidities and/or clusters may help predict poorer PCS trajectories in older women with breast cancer. The National Comprehensive Cancer Network recommends use of comorbidity assessment tools to older patients prior to cancer treatment [61]. However, a systematic review on the utilization of comorbidity tools suggest limited use in clinical practice [62]. Barriers to implementation may be due to time constraints and lack of perceived value or benefit from clinicians [62]. Nonetheless, our finding that number of medical comorbidities was a primary variable associated with poorer PCS trajectories suggests the importance of these in geriatric oncology care. Assessing the presence or absence of comorbidities using tools such as the Charlson Comorbidity Index (CCI) or Cumulative Illness Rating Scale-Geriatric (CIRS-G) may facilitate treatment decision making, determine toxicity risk, identify vulnerable patients, and enhance treatment success. Potential interventions may include optimizing specific health conditions prior to cancer treatment and coordinating a multidisciplinary treatment team consisting of oncologists, primary care physicians, psychologists, and other specialists to optimize health and well-being [61, 63].

The incorporation of routine comprehensive geriatric assessments (CGA) that screen for the above factors may help providers identify older women at risk for poorer physical HRQoL post breast cancer treatment. The International Society of Geriatric Oncology (SIOG) guidelines recommend clinicians to routinely administer CGA as an effective evidence-based care model for older adults with cancer to assist with assessment and discussion of treatment options [64, 65]. These assessments typically include the following domains: functionality, nutrition, cognition, psychological state, social support, comorbidities, medications, and geriatric syndromes [65]. A recent systematic review of utilization of CGA in older women with early stage non-metastatic breast cancer suggested that results of the CGA can be successfully used to assess QOL and predict treatment outcomes (e.g. survival and mortality rate) [66]. As such, CGA may assist with identifying varying levels of fitness of patients based on an algorithm of factors to determine the most optimal treatment options that increase chances of survival, minimize treatment adverse effects, and decrease risk for poorer physical HrQOL outcomes [66]. Although there is no current gold standard for CGA, our results suggest that age, number of comorbidities, BMI, and financial hardship may be important domains to consider. Future research is encouraged to explore the domains of the CGA that are most applicable for older women diagnosed with breast cancer to assist with cancer treatment options and treatment decision making to optimize survival and health outcomes.

There are several limitations to this study. A larger sample or longer follow-up may have identified a greater number of trajectories, and/or allowed for a more thorough examination of the impact of specific medical comorbidities on PCS group membership. Baseline assessment occurred between 3 days and 8 months following diagnosis, and thus our reported mean values of the psychosocial covariates at 4 months post-diagnosis are estimates, specifically least-squares estimates derived from repeated measures models. Age and gender-based population reference norms were used to compare the PCS scores to older women without a history of cancer. Our sample was racially and socioeconomically fairly homogeneous, which limits generalizability. Finally, we recognize it is unlikely that the MCAR assumption underlying trajectory analysis hold fully in these or any observational data; this is an additional reason for caution in generalizing our findings.

This paper contributes to the geriatric oncology literature by examining the PCS trajectories among older cancer survivors in the early years post-breast cancer diagnosis. Although the majority of survivors reported high PCS, a large subgroup reported consistently low PCS that did not improve over time. Older age, BMI, ability to pay for basics, and number of medical comorbidities were strongly associated with low PCS group membership. Incorporating comprehensive geriatric assessments that assess for these characteristics may help identify individuals at risk for poorer physical HRQoL post breast cancer treatment.