Abstract
Importance
Health care systems should strive to achieve outcomes that matter to patients. Limited data exist on the outcomes achieved by patients with cancer with different treatment outcome preferences.
Objective
To describe treatment outcome preferences among older adults with advanced cancer and explore differences in outcomes between patients prioritizing survival vs quality of life (QoL), testing the hypothesis that patients prioritizing survival would live longer, while those prioritizing QoL would have fewer treatment-related adverse effects (TRAEs) and hospitalizations.
Design, Setting, and Participants
This study was an exploratory secondary analysis of the GAP70+ cluster randomized clinical trial conducted at National Cancer Institute’s Community Oncology Research Program centers. Older adults (≥70 years) with incurable solid tumors or lymphoma and 1 or more impaired geriatric assessment domain(s) who were starting a new systemic cancer treatment as part of the GAP70+ trial were eligible. The primary trial took place between July 29, 2014, and March 13, 2019. All secondary analyses were conducted between January 1, 2025, and May 30, 2025.
Exposure
Patients were divided into 2 cohorts based on their reported preference for prioritizing extending survival vs maintaining QoL.
Main Outcomes and Measures
Treatment outcome preferences, hospitalization, TRAEs, and survival at 6 months and 1 year.
Results
A total of 706 patients were included in the analysis. The mean (SD) patient age was 77.2 (5.4) years; 306 (43.3%) were female, and 400 (56.7%) were male. Gastrointestinal (244 patients [34.6%]), lung (175 patients [24.8%]), and genitourinary (109 patients [15.4%]) cancers were the most common. Only 59 patients (8.4%) preferred to prioritize extending survival over maintaining QoL, while 506 patients (71.7%) preferred to prioritize maintaining QoL. No significant associations were identified between cohorts prioritizing survival vs QoL and treatment modifications (risk ratio, 1.03; 95% CI, 0.84–1.27), grade 3 to 5 TRAEs (hazard ratio [HR], 0.84; 95% CI, 0.57–1.23), hospitalization (HR, 0.74; 95% CI, 0.39–1.41), and survival (at 6 months: HR, 0.72; 95% CI, 0.40–1.29; at 1 year: HR, 1.18; 95% CI, 0.81–1.72).
Conclusions and Relevance
In this secondary analysis of a randomized clinical trial, fewer than 1 in 10 older adults with advanced cancer participating in the trial prioritized extending survival over maintaining QoL. Patient preference for extending survival or maintaining QoL was not associated with up-front treatment modifications or downstream outcomes, suggesting a possible lack of responsiveness of the current oncology care delivery system to patient preference.
ClinicalTrials.gov Identifier
Introduction
Health care systems should achieve outcomes that matter to patients. Preference-concordant care, care aimed at achieving outcomes that matter most to each patient, improves quality of life (QoL) and patient satisfaction, and reduces health care utilization.1,2 To our knowledge, no recent large oncology studies have evaluated the association between the treatment outcome preferences of individual patients and clinical outcomes. In this study of older adults with advanced cancer, we sought to describe patient treatment outcome preferences for prioritizing survival vs maintaining QoL and to evaluate whether these preferences were associated with related clinical outcomes.
Methods
We conducted an exploratory secondary analysis of the GAP70+ trial, a cluster randomized clinical trial (RCT) providing oncology physicians with geriatric assessment (GA) results and GA-guided recommendations. (See Supplement 1 for the trial protocol.) Institutional review board approval and informed consent were obtained for the RCT, including secondary data analyses using all of the collected data. The primary trial took place between July 29, 2014, and March 13, 2019, and the results have previously been published.3,4 Eligible patients were 70 years and older, with an incurable advanced cancer or lymphoma and 1 or more GA domain impairment (excluding polypharmacy), initiating a cancer treatment with high risk of treatment-related adverse effects (TRAEs) within 4 weeks. At enrollment, patients indicated their level of agreement (strongly agree, agree, neutral, disagree, or strongly disagree) with the statement: “Maintaining my quality of life is more important to me than living longer.”5 Participants self-reported their race and ethnicity.
We built generalized estimating equation models to assess the associations of treatment outcome preference (survival vs QoL) with treatment modifications. We built Cox shared frailty models to assess the associations of treatment outcome preference with clinician-rated grade 3 to 5 TRAEs; hospitalizations within the first 3 months; and death at 6 and 12 months. We considered death as a competing risk by estimating cause-specific hazards of grade 3 to 5 TRAEs and hospitalizations. We hypothesized that patients prioritizing survival would live longer, while those prioritizing maintaining QoL would have more treatment modifications leading to fewer TRAEs and hospitalizations. More details may be found in the eMethods in Supplement 2. Analyses were performed using SAS statistical software, version 9.4 (SAS Institute Inc), between January 1, 2025, and May 30, 2025. Two-tailed hypothesis testing was conducted at α = .05.
Results
Among 706 older adults with advanced cancer, 506 patients (71.7%) preferred to prioritize maintaining QoL vs 59 (8.4%) extending survival and 141 (20.0%) with no preference/neutral (Table 1). The mean (SD) age of participants was 77.2 (5.4) years; 306 (43.3%) were female, and 400 (56.7%) were male. Fifty participants (7.1%) identified as African American; 620 (87.8%) as non-Hispanic White; and 35 (5.0%) as American Indian or Alaska Native, Asian, or Native Hawaiian or other Pacific Islander.
Table 1.
Baseline Characteristics
| Characteristic | Participants, No. (%) | P valuea | |||
|---|---|---|---|---|---|
| All participants (N = 706) | No preference/neutral (n = 141) | Prioritize extending survival (n = 59) | Prioritize maintaining QoL (n = 506) | ||
| Study arm | |||||
| Control | 365 (51.7) | 72 (51.1) | 26 (44.1) | 267 (52.8) | .21 |
| Intervention | 341 (48.3) | 69 (48.9) | 33 (55.9) | 239 (47.2) | |
| Age, mean (SD; range), y | 77.2 (5.4; 70–96) | 76.3 (5.4; 70–91) | 76.5 (5.3; 70–92) | 77.5 (5.5; 70–96) | .18 |
| Sex | |||||
| Female | 306 (43.3) | 62 (44.0) | 23 (39.0) | 221 (43.7) | .49 |
| Male | 400 (56.7) | 79 (56.0) | 36 (61.0) | 285 (56.3) | |
| Race and ethnicity | |||||
| African American | 50 (7.1) | 7 (5.0) | 11 (18.6) | 32 (6.3) | <.001 |
| Non-Hispanic White | 620 (87.8) | 125 (88.7) | 43 (72.9) | 452 (89.3) | |
| Otherb | 35 (5.0) | 9 (6.4) | 5 (8.5) | 21 (4.2) | |
| Missing | 1 | NA | NA | NA | NA |
| Marital status | |||||
| Single or never married | 17 (2.4) | 5 (3.5) | 2 (3.4) | 10 (2.0) | .76 |
| Married or domestic partner | 442 (62.6) | 87 (61.7) | 36 (61.0) | 319 (63.0) | |
| Separated, widowed, or divorced | 247 (35.0) | 49 (34.8) | 21 (35.6) | 177 (35.0) | |
| Education | |||||
| <High school | 110 (15.6) | 17 (12.1) | 13 (22.0) | 80 (15.8) | .01 |
| High school graduate | 237 (33.6) | 48 (34.0) | 27 (45.8) | 162 (32.0) | |
| ≥Some college | 359 (50.8) | 76 (53.9) | 19 (32.2) | 264 (52.2) | |
| Income, $ | |||||
| ≤50 000 | 364 (51.6) | 71 (50.4) | 34 (57.6) | 259 (51.2) | .64 |
| >50 000 | 189 (26.8) | 37 (26.2) | 14 (23.7) | 138 (27.3) | |
| Declined to answer | 153 (21.7) | 33 (23.4) | 11 (18.6) | 109 (21.5) | |
| Cancer type | |||||
| Gastrointestinal | 244 (34.6) | 44 (31.2) | 24 (40.7) | 176 (34.8) | .74 |
| Genitourinary | 109 (15.4) | 22 (15.6) | 10 (16.9) | 77 (15.2) | |
| Lung | 175 (24.8) | 39 (27.7) | 12 (20.3) | 124 (24.5) | |
| Otherc | 178 (25.2) | 36 (25.5) | 13 (22.0) | 129 (25.5) | |
| Cancer stage | |||||
| III | 76 (10.8) | 11 (7.8) | 4 (6.8) | 61 (12.1) | .31 |
| IV | 617 (87.4) | 127 (90.1) | 53 (89.8) | 437 (86.4) | |
| Other | 13 (1.8) | 3 (2.1) | 2 (3.4) | 8 (1.6) | |
| Cancer treatment type | |||||
| Chemotherapy alone (single agent) | 146 (20.7) | 27 (19.1) | 10 (16.9) | 109 (21.5) | .78 |
| Chemotherapy alone (multiagent) | 325 (46.0) | 70 (49.6) | 28 (47.5) | 227 (44.9) | |
| Chemotherapy plus other agents | 150 (21.2) | 29 (20.6) | 12 (20.3) | 109 (21.5) | |
| Nonchemotherapy (monoclonal antibody, hormone, immunotherapy) | 85 (12.0) | 15 (10.6) | 9 (15.3) | 61 (12.1) | |
| No. of impaired geriatric assessment domains, mean (SD; range) | 4.5 (1.6; 1–8) | 4.5 (1.6; 1–8) | 4.6 (1.6; 1–8) | 4.5 (1.6; 1–8) | .52 |
| Impaired geriatric assessment domains | |||||
| Physical performance | 657 (93.1) | 128 (90.8) | 52 (88.1) | 477 (94.3) | .09 |
| Polypharmacy | 573 (81.2) | 112 (79.4) | 48 (81.4) | 413 (81.6) | .96 |
| Comorbidity | 474 (67.1) | 99 (70.2) | 45 (76.3) | 330 (65.2) | .09 |
| Functional status | 406 (57.5) | 86 (61.0) | 27 (45.8) | 293 (57.9) | .07 |
| Nutrition | 427 (60.5) | 93 (66.0) | 35 (59.3) | 299 (59.1) | .97 |
| Cognition | 256 (36.3) | 45 (31.9) | 29 (49.2) | 182 (36.0) | .048 |
| Social support | 191 (27.1) | 37 (26.2) | 18 (30.5) | 136 (26.9) | .55 |
| Psychological status | 198 (28.0) | 39 (27.7) | 19 (32.2) | 140 (27.7) | .46 |
| Treatment modification at the start of therapy | |||||
| No | 408 (57.8) | 84 (59.6) | 31 (52.5) | 293 (57.9) | .42 |
| Yes | 296 (41.9) | 56 (39.7) | 28 (47.5) | 212 (41.9) | |
| Missing | 2 | NA | NA | NA | NA |
Abbreviation: NA, not applicable; QoL, quality of life.
P value was based on the comparison between participants who prioritize extending survival vs adults who prioritize maintaining QoL. Post hoc comparison after a significant P value: patients who were White (vs African American), had some college education or above (vs high school graduate), or without cognitive impairment (vs with) were more likely to prioritize QoL over extending survival.
Race and ethnicity were self-reported. The other category included individuals who identified as American Indian or Alaska Native, Asian, and Native Hawaiian or other Pacific Islander.
Other cancer types include breast cancer (n = 56), gynecological cancer (n = 41), lymphoma (n = 46), and other types of cancer (n = 35).
No difference in preference was found based on age, sex, income, or marital status. Patients completing some college or above were more likely to prioritize QoL vs those with high school degrees alone (odds ratio [OR], 2.3; 95% CI, 1.2–4.3). No difference in preference was observed based on cancer type, stage, or impaired domains of GA, except for cognition. Patients without cognitive impairment, vs those with cognitive impairment, were more likely to prioritize QoL (OR, 1.7; 95% CI, 1.0–3.0).
No association was identified between treatment outcome preference (survival vs QoL) and treatment modifications (risk ratio, 1.03; 95% CI, 0.84–1.27) or any clinical outcomes, including grade 3 to 5 TRAEs (hazard ratio [HR], 0.84; 95% CI, 0.57–1.23), hospitalization (HR, 0.74; 95% CI, 0.39–1.41), or survival at 6 months or 1 year (6 months: HR, 0.72; 95% CI, 0.40–1.29; 1 year: HR, 1.18; 95% CI, 0.81–1.72) (Figure). No evidence of interaction between study arm and treatment preference was observed; no significant differences in clinical outcomes by patient preference were seen in models stratified by study arm (Table 2).
Figure. Forest Plot of the Association Between Patient Preference for Survival or Quality of Life and Clinical Outcomes.
A hazard ratio (HR) of 1.0 represents no association between outcomes and preference. The dark green box is the point estimate. Error bars represent 95% CI. Included participants were older adults 70 years and older with an incurable advanced solid tumor or lymphoma, at least 1 geriatric assessment domain impairment (excluding polypharmacy), and scheduled to initiate a cancer treatment with a high risk of treatment-related adverse effects (TRAEs) within 4 weeks.
aThe measure for treatment modifications is risk ratio.
Table 2.
Association of Patient Preference With Clinical Outcomes
| Clinical outcome | Events, No./total (%) | Prioritizing extending survival vs prioritizing maintaining QoL (reference) | |||||
|---|---|---|---|---|---|---|---|
| Neutral | Prioritizing survival | Prioritizing QoL | Primary analysis | Sensitivity analysis: adding treatment modification at the start of therapy as a covariate | |||
| Uncorrected | Bonferroni correction | Uncorrected | Bonferroni correction | ||||
| Generalized estimating equations | |||||||
| Treatment modification | 56/140 (40.0) | 28/59 (47.5) | 212/505 (42.0) | RR = 1.03 (0.84–1.27) | RR = 1.03 (0.81–1.31) | NA | NA |
| Cox shared frailty models (death as a competing risk) | |||||||
| Any hospitalization within 3 mo | 35/141 (24.8) | 12/59 (20.3) | 131/506 (25.9) | HR = 0.74 (0.39–1.41) | HR = 0.74 (0.36–1.54) | HR = 0.75 (0.40–1.41) | HR = 0.75 (0.36–1.55) |
| Any grade 3–5 TRAEs within 3 mo | 87/141 (61.7) | 31/59 (52.5) | 318/506 (62.8) | HR = 0.84 (0.57–1.23) | HR = 0.84 (0.54–1.30) | HR = 0.84 (0.57–1.23) | HR = 0.84 (0.54–1.30) |
| Cox shared frailty models | |||||||
| Died at 6 mo | 33/141 (23.4) | 13/59 (22.0) | 140/506 (27.7) | HR = 0.72 (0.40–1.29) | HR = 0.72 (0.37–1.40) | HR = 0.71 (0.40–1.28) | HR = 0.71 (0.37–1.39) |
| Died at 1 y | 62/141 (44.0) | 33/59 (55.9) | 239/506 (47.2) | HR = 1.18 (0.81–1.72) | HR = 1.18 (0.77–1.81) | HR = 1.20 (0.82–1.74) | HR = 1.20 (0.78–1.84) |
Abbreviations: HR, hazard ratio; NA, not applicable; QoL, quality of life; RR, risk ratio; TRAE, treatment-related adverse effects.
Discussion
This secondary analysis of a cluster RCT provides insights into the preferences and care of older adults with advanced cancer. Most older adults with advanced cancer prioritize QoL over extending survival; less than 10% of patients prioritized survival over QoL. Although some associations between preference and baseline factors were discovered, patient preferences were not reliably associated with any specific demographic or clinical factor, including frailty. This finding reinforces the necessity of eliciting individual patient preferences. Some have argued that combining preference elicitation with GA at time of treatment may be the best strategy to inform treatment decisions for older adults.6
This large analysis investigated the association between patient preference, treatment choice, and clinical outcomes. Contrary to our hypothesis, we did not find significant differences in up-front treatment modifications between patients who prioritized survival and those who prioritized QoL, suggesting that treatment modifications may not have been responsive to patient preference. Further, we did not find any statistically significant differences in downstream clinical outcomes (grade 3–5 TRAEs, hospitalizations, or death) between groups. Taken together, these data suggest a possible lack of responsiveness of the current oncology care delivery system to patient preferences.
Determining why patient preferences were not associated with clinical outcomes was not possible in this study. Unfortunately, structural incentives to enhance productivity/reimbursement increasingly promote adherence to protocolized regimens (e.g., pathways) and disincentivize clinicians from tailoring treatments to individual patient preferences. Data from 2 studies suggest that only approximately one-third of the time oncologists ask patients what matters most to them before determining therapy.7,8 Alternatively, was the primary desired outcome for most patients, improved QoL, simply too vague to operationalize for clinicians or too difficult to achieve? Only 3% of all RCTs in the last 20 years in oncology have focused on improving QoL for patients, suggesting a gap between what matters to older adults and what is prioritized in trials.9 Trials to identify evidence-based interventions to improve or maintain QoL for older adults with cancer, who have been historically underrepresented in clinical trials,10 are especially needed in oncology and may appeal to patients even more than trials evaluating end points they find less important.
Limitations
This study has limitations. The small sample size of those who prioritized survival (n = 59) resulted in a lack of precision around estimates. While no significant associations between preferences and outcomes were found in this study, these data alone are not precise enough to exclude the possibility of identifying meaningful associations in a different sample. We followed prior work11,12 in using a 1-question measure to dichotomize preferences between maintaining QoL and extending survival. Including more attributes distinguishing aspects of QoL or using a more robust preference elicitation method, such as best-worst scaling, may have provided a more complete understanding of patient preferences. Capturing preferences and QoL longitudinally would be valuable to better understand the relationship between these variables. Patient preferences from minoritized racial and ethnic groups are not well characterized in this study, limiting generalizability of the findings and potentially obscuring important differences in preferences.
Although eliciting patient preferences is essential to high-quality care, oncology clinical teams do not routinely ask patients about their preferences or integrate preferences into care plans.7 A 2024 rule from the Centers for Medicare & Medicaid Services now requires documentation of how preferences of older adults are integrated into individual care plans.13 By financially incentivizing preference-concordant care, this rule may be a catalyst to improve cancer care for older adults.
Conclusions
This secondary analysis of an RCT suggests a possible lack of responsiveness of the current oncology care delivery system to patient preferences. Further work is necessary to develop evidence-based strategies for preference elicitation and aligning individual treatment plans to what matters most to each patient.
Supplementary Material
Key Points.
Question
Do outcomes differ between older adults with advanced cancer who prioritize maintaining quality of life vs extending survival?
Findings
In this secondary analysis of a large cluster randomized clinical trial including 706 patients, 506 patients (71.7%) prioritized maintaining quality of life over survival. No significant differences were identified in survival, hospitalization, or treatment-related adverse effects between groups.
Meaning
The current oncology care delivery system may not be responsive to the preferences of older adults with advanced cancer.
Funding/Support:
The primary cluster randomized clinical trial received funding from the National Cancer Institute (R01CA177592, U01CA233167 SGM, and UG1CA189961)
Role of the Funder/Sponsor:
The funder had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Footnotes
Conflict of Interest Disclosures: Dr Richardson reported grants from the National Institutes of Health (NIH)/National Cancer Institute to develop values elicitation tools for older adults with cancer. Dr Flannery reported grants from NIH during the conduct of the study. Dr Magnuson reported grants from NIH and FAIR Health and personal fees from the American Society of Clinical Oncology outside the submitted work. Dr Wells reported grants from NIH during the conduct of the study; stock ownership in Agenus and MiNK Therapeutics outside the submitted work. Dr Mohile reported grants from NIH during the conduct of the study and grants from Gilead outside the submitted work. No other disclosures were reported.
Data Sharing Statement:
See Supplement 3.
References
- 1.Stacey D, Lewis KB, Smith M, et al. Decision aids for people facing health treatment or screening decisions. Cochrane Database Syst Rev. 2024;1(1):CD001431. doi: 10.1002/14651858.CD001431.pub6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Witteman HO, Ndjaboue R, Vaisson G, et al. Clarifying values: an updated and expanded systematic review and meta-analysis. Med Decis Making. 2021;41(7):801–820. doi: 10.1177/0272989X211037946 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Mohile SG, Epstein RM, Hurria A, et al. Communication with older patients with cancer using geriatric assessment: a cluster-randomized clinical trial from the National Cancer Institute Community Oncology Research Program. JAMA Oncol. 2020;6(2):196–204. doi: 10.1001/jamaoncol.2019.4728 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Mohile SG, Mohamed MR, Xu H, et al. Evaluation of geriatric assessment and management on the toxic effects of cancer treatment (GAP70+): a cluster-randomised study. Lancet. 2021;398(10314):1894–1904. doi: 10.1016/S0140-6736(21)01789-X [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Loh KP, Mohile SG, Lund JL, et al. Beliefs about advanced cancer curability in older patients, their caregivers, and oncologists. Oncologist. 2019;24(6):e292–e302. doi: 10.1634/theoncologist.2018-0890 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.DuMontier C, Loh KP, Soto-Perez-de-Celis E, Dale W. Decision making in older adults with cancer. J Clin Oncol. 2021;39(19):2164–2174. doi: 10.1200/JCO.21.00165 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Williams CP, Miller-Sonet E, Nipp RD, Kamal AH, Love S, Rocque GB. Importance of quality-of-life priorities and preferences surrounding treatment decision making in patients with cancer and oncology clinicians. Cancer. 2020;126(15):3534–3541. doi: 10.1002/cncr.32961 [DOI] [PubMed] [Google Scholar]
- 8.Pieterse AH, Kunneman M, Engelhardt EG, et al. Oncologist, patient, and companion questions during pretreatment consultations about adjuvant cancer treatment: a shared decision-making perspective. Psychooncology. 2017;26(7):943–950. doi: 10.1002/pon.4241 [DOI] [PubMed] [Google Scholar]
- 9.Sherry AD, Miller AM, Parlapalli JP, et al. Overall survival and quality-of-life superiority in modern phase 3 oncology trials: a meta-epidemiological analysis. JAMA Oncol. 2025;11(7):718–724. doi: 10.1001/jamaoncol.2025.1002 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Sedrak MS, Freedman RA, Cohen HJ, et al. ; Cancer and Aging Research Group (CARG). Older adult participation in cancer clinical trials: a systematic review of barriers and interventions. CA Cancer J Clin. 2021;71(1):78–92. doi: 10.3322/caac.21638 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Weeks JC, Cook EF, O’Day SJ, et al. Relationship between cancer patients’ predictions of prognosis and their treatment preferences. JAMA. 1998;279(21):1709–1714. doi: 10.1001/jama.279.21.1709 [DOI] [PubMed] [Google Scholar]
- 12.Wright AA, Zhang B, Ray A, et al. Associations between end-of-life discussions, patient mental health, medical care near death, and caregiver bereavement adjustment. JAMA. 2008;300(14):1665–1673. doi: 10.1001/jama.300.14.1665 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Medicare and Medicaid programs and the Children's Health Insurance Program; hospital inpatient prospective payment systems for acute care hospitals and the long-term care hospital prospective payment system and policy changes and fiscal year 2025 rates; quality programs requirements; and other policy changes. Fed Regist. 2024;89(167):68986–70046. [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
See Supplement 3.