Cost-Utility Analysis of Geriatric Assessment and Management in Older Adults With Cancer: Economic Evaluation Within 5C Trial

Abstract

PURPOSE

Geriatric assessment (GA) is a guideline-recommended approach to optimize cancer management in older adults. We conducted a cost-utility analysis alongside the 5C randomized controlled trial to compare GA and management (GAM) plus usual care (UC) against UC alone in older adults with cancer.

METHODS

The economic evaluation, conducted from societal and health care payer perspectives, used a 12-month time horizon. The Canadian 5C study randomly assigned patients to receive GAM or UC. Quality-adjusted life-years (QALYs) were measured using the EuroQol five dimension-5L questionnaire and health care utilization using cost diaries and chart reviews. We evaluated the incremental net monetary benefit (INMB) for the full sample and preselected subgroups.

RESULTS

A total of 350 patients were included, of whom 173 received GAM and 177 UC. At 12 months, the average QALYs per patient were 0.728 and 0.751 for GAM and UC, respectively (ΔQALY, −0.023 [95% CI, −0.076 to 0.028]). Considering a societal perspective, the total average costs (in 2021 Canadian dollars) per patient were $46,739 and $45,177 for GAM and UC, respectively (ΔCost, $1,563 [95% CI, −$6,583 to $10,403]). At a cost-effectiveness threshold of $50,000/QALY, GAM was not cost-effective compared with UC (INMB, −$2,713 [95% CI, −$11,767 to $5,801]). The INMB was positive ($2,984 [95% CI, −$7,050 to $14,179]; probability of being cost-effective, 72%) for patients treated with curative intent, but remained negative for patients treated with palliative intent (INMB, −$9,909 [95% CI, −$24,436 to $4,153]). Findings were similar considering a health care payer perspective.

CONCLUSION

To our knowledge, this is the first cost-utility analysis of GAM in cancer. GAM was cost-effective for patients with cancer treated with curative but not with palliative intent. The study provides further considerations for future adoption of GAM in practice.

INTRODUCTION

The health and economic burden of cancer around the world is increasing steadily, primarily because of population aging.¹ In 2021, 48% of all new cancer cases in Canada were detected in those age 70 years and older.² Older individuals have complex health care demands owing to coexisting diseases and impaired functional capacity, which contribute to their need for more services and greater expenditures.³ Previous research has shown that substantial portion of emergency department (ED) and inpatient spending for patients with cancer were avoidable.⁴ Therefore, early identification and management of preventable conditions or events could represent an important method for reducing cancer care costs.

CONTEXT

• Key Objective

• On the basis of the Canadian 5C randomized controlled trial, to our knowledge, this was the first within-trial cost-effectiveness analysis comparing geriatric assessment and management (GAM) versus usual care (UC) in patients with cancer.

• Knowledge Generated

• Under both societal and health care perspectives, GAM was not cost-effective compared with UC in the overall study sample. GAM, however, was more cost-effective than UC for patients treated with curative but not for those with palliative treatment intent. Hospitalization and chemotherapy costs were the main drivers of total costs, while costs of delivering GAM were small.

• Relevance (S.D. Ramsey)

• Programs that comprehensively assess and manage older patients with cancer are recommended by National Comprehensive Cancer Network, American Society of Clinical Oncology, and other guidelines groups, on the basis of evidence that they decrease treatment toxicity and improve treatment completion. This cost-effectiveness study showed that these programs can be highly cost-effective for patients who are treated with curative intent, but have lower value in palliative settings.*

  *Relevance section written by JCO Consulting Editor Scott D. Ramsey, MD, PhD.

Given the complexity of cancer management in older patients, the use of comprehensive geriatric assessment (GA) has been recommended by many organizations.^5-7 To enable individualized treatment approaches, GA evaluates comorbidities, cognitive and mental health, functional and nutritional status, social support, and assesses tolerance to chemotherapy to avoid overtreatment or undertreatment of cancer.^3,5 Recent randomized controlled trials (RCTs) and a systematic review concluded that GA had a favorable impact on outcomes in patients with cancer, namely decreased treatment toxicity,^8,9 improved treatment completion,¹⁰ quality of life (QoL),¹¹ and improved patient-provider communication.¹² A single, model-based economic evaluation reported that GA can be marginally cost-effective depending on setting.¹³ GA can be resource-intensive, often requiring a multidisciplinary team and the presence of geriatric oncology-trained specialists. Although GA has been shown to be cost-effective in hip fracture¹⁴ and ambulatory geriatric care,¹⁵ the cost-effectiveness of GA has not been established in cancer.

The 5C was a single-blind, RCT (ClinicalTrials.gov identifier: NCT03154671) conducted in Canada between March 2018 and March 2020 and evaluated the effectiveness of a GA and management (GAM) on QoL in older patients with cancer.¹⁶ The study did not find a statistical difference in either the primary outcome of QoL at 6 months (difference in global QoL of 0.01 points [95% CI, −0.46 to 0.48]) or in secondary outcomes of survival, functional status, number of ED visits and unplanned hospitalizations, treatment continuation, and toxicity.¹⁶ Cost-effectiveness was a prespecified secondary outcome considering the resource intensity of the GAM.¹⁶ Therefore, the aim of this study was to conduct an economic evaluation alongside the 5C RCT to compare GAM plus usual care (UC) against UC alone in older adults with cancer.

METHODS

Study Design and Population

This was a within-trial economic evaluation to assess cost utility of GAM relative to UC in older adults with cancer, conducted from societal and health care perspectives. Since no differences in clinical outcomes were observed in the 5C RCT, we had no basis to anticipate any impact of GAM beyond the trial duration. Thus, we used a 12-month time horizon reflecting the longest follow-up in the trial.¹⁷ No discounting of costs or benefits was applied. Research ethics boards of participating centers approved the study protocol, and informed consent was obtained from each participant. We followed The Consolidated Health Economic Evaluation Reporting Standards for reporting of this study (Data Supplement, Table S1 [online only]).¹⁸

A detailed trial protocol has been reported elsewhere.¹⁹ Briefly, 350 patients from eight Canadian cancer centers were randomly assigned to receive GAM for 6 months in addition to UC or UC alone and were followed for 12 months or until death, whichever occurred first. Patients were eligible if they were age 70 years and older, had an estimated life expectancy of more than 6 months, and had an Eastern Collaborative Oncology Group Performance Score of 0-2.

Intervention and Comparator

The GAM intervention aligned with clinical guidelines and recommendations included functional status, cognitive, nutritional, and psychosocial evaluation, mobility and falls risk assessment, medication review and polypharmacy management, care planning and coordination, and geriatric syndrome management.^7,19 GAM group participants received UC from their oncologist and GAM from a geriatric team (a registered nurse and a geriatrician) at enrollment and repeated assessments thereafter if necessary. The GA results and any ongoing recommendations were communicated to the oncologists to optimize the care plan. After the initial visit, a nurse followed up with patients at least monthly for 6 months to identify any new or ongoing issues or needs.

Participants allocated to the UC group received care from their oncologist.

Effectiveness Outcomes

The effectiveness outcome of this economic evaluation was quality-adjusted life-years (QALYs). Health utilities were estimated using the EuroQol five dimension (EQ-5D)-5L questionnaire²⁰ completed at baseline, monthly for the first 6 months, and then at 9 and 12 months either in person, by telephone, or by email surveys. We converted health states into health utilities using the Canadian time trade-off scoring algorithm.²¹ QALYs were calculated as a product of health utility and time period.

Resource Utilization and Costs

Self-reported health care utilization was evaluated using cost diaries collected at baseline and then prospectively every 3 months. Diaries included questions on hospital and ED admissions, same-day surgeries and procedures, visits to health care specialists including phone consultations, laboratory tests, time and out-of-pocket (OOP) costs spent on receiving care (eg, transportation, private hospital room), over-the-counter medications, medical aids or equipment, accompanying family member (their sex and age), paid and unpaid assistance with household chores, and productivity loss (loss of paid employment both for patients and caregivers). Diaries were reviewed by research personnel and, if needed, were supplemented by review of medical records for validation. Data on primary causes of hospitalization, chemotherapy regimens, and radiation therapy were retrieved from the medical records as well.

For the intervention arm, the number and duration of in-person and phone follow-up visits with the geriatrician and nurses were recorded. Unit costs for geriatrician (Data Supplement, Table S2) and other physician consultations, laboratory tests, and procedures were assigned from the Ontario Schedule of Benefits for Physician Services,²² while hourly labor costs for nurses and allied health specialists were obtained from the hospital financial records.²³ Unit costs for ED and hospital admissions were obtained using the 2019-2020 Patient Cost Estimator by the Canadian Institute for Health Information and selecting case mix groups for Ontario elderly population and respective cause of hospitalization.²⁴ Unit costs of chemotherapy medications were obtained from multiple sources.^23,25,26 The costs for radiation therapy incorporated fixed costs for supporting infrastructure (eg, computed tomography simulation, dosimetry planning, and quality assurance by physicists) and per-fraction cost for treatment delivery.²³

Costs of productivity loss were operationalized using the human capital method.²⁷ For each patient, we estimated total time spent on receiving care. We assumed that persons (caregivers) who typically accompany patients to health care visits would spend time equivalent to that of patients. Next, to monetize productivity loss for working patients and caregivers, we multiplied hours by mean age- and sex-specific hourly wages in Canada, which were inflated by 11% to cover potential benefits (eg, days off work, vacation etc), as per Canadian economic evaluation guideline.²⁷ We also monetized the amount of unpaid help around household chores, by assuming minimal wage of $15 (in 2021 Canadian dollars) per hour of help. All costs were converted to 2021 Canadian dollars using the Canadian Consumer Price Index for health and personal care.²⁸

Statistical and Economic Analysis

All statistical analyses were conducted in R Statistical Software.²⁹ Baseline characteristics, QALYs, costs per resource utilization category, and cumulative costs were reported using descriptive statistics.

Considering intermittent missing patterns of cost and QALY data at study visits (Data Supplement, Fig S1), multiple imputation by chained equation technique was chosen to account for missing data.^30,31 The imputation model included age, sex, treatment intent, frailty status, baseline utility values, and all available outcome measures. We imputed total costs rather than costs by resource category and utility score rather than individual domains of EQ-5D-5L. We created 10 complete data sets.

We calculated incremental QALYs and costs by determining the mean difference of the corresponding outcomes between the GAM and UC groups. The 95% CIs for the incremental QALYs and costs were computed using bootstrapping with 2,000 replications drawn from each imputed data set.³²

Since we anticipated a small difference in QALYs, we used incremental net monetary benefit (INMB) as a primary measure to overcome limitations of incremental cost-effectiveness ratio (ICER).¹⁷ For each bootstrap sample, INMB was determined using a cost-effectiveness threshold (λ) of $50,000 and $100,000 per QALY as follows:

$$\text{INMB}=\lambda \times \Delta \text{QALY}-\Delta \text{Costs}$$

GAM was considered cost-effective if the INMB was positive, that is, meaning the benefits outweigh the costs, at the defined cost-effectiveness threshold. The probability of cost-effectiveness was calculated as the proportion of the 2,000 bootstrap samples with a positive INMB in each imputed data set, and then averaged across all data sets.

One-way sensitivity and threshold analyses were performed to evaluate the impact of uncertainty in the key parameters on the results. Since the treatment intent (curative v palliative) was considered as a stratification factor in the original paper, a preplanned subgroup cost-effectiveness analysis by treatment intent was conducted. We further explored the results by frailty status measured using the Geriatric 8, whereby a score of ≤14 indicated a higher risk of impairment warranting a GA.^33,34 Additionally, we considered a sensitivity analysis, which included adjustment for the baseline utility score.

RESULTS

Baseline Characteristics

The RCT sample included 350 patients. Baseline characteristics were comparable between the arms (Table 1). The mean age was 76 years, 40% were female, and two-thirds were considered frail. More than half (54%) of the participants were treated with curative intent and 46% with palliative intent. In treatment intent subgroups, meaningful differences were noted in the type of cancer and baseline utility score between the study arms, with patients treated with palliative intent showing higher baseline utility scores in the UC arm.

TABLE 1.

Baseline Characteristics of Study Population

During the 12-month follow-up period, 20 (11.6%) and 25 (14.1%) participants withdrew, and 39 (22.5%) and 42 (23.7%) participants died in the intervention and UC arms, respectively. The Data Supplement (Table S3) reports a similar distribution of missing information between the arms in EQ-5D-5L surveys and cost diaries at each data point.

Effectiveness

Perceived problems rated in all five dimensions of the EQ-5D-5L questionnaire (Data Supplement, Fig S2), the utility scores, and the number of survived days were comparable between arms across all study time points, but overall were favorable among patients treated with curative versus palliative intent (Data Supplement, Table S4, Figs S3 and S4). We found no difference in QALYs between the arms. After imputation, the mean (standard deviation [SD]) QALYs were 0.728 (0.263) in GAM and 0.751 (0.230) in UC groups (ΔQALY, −0.023 [95% CI, −0.076 to 0.028]; ΔQALY adjusted for baseline utility, −0.016 [95% CI, −0.069 to 0.034]). In patients treated with palliative intent, the mean (SD) QALYs were 0.638 (0.294) in GAM and 0.706 (0.254) in UC groups (ΔQALY, −0.068 [95% CI, −0.153 to 0.015]; ΔQALY adjusted, −0.049 [95% CI, −0.136 to 0.034]). In patients treated with curative intent, the mean QALYs (SD) were 0.802 (0.207) in GAM and 0.790 (0.201) in UC groups (ΔQALY, 0.012 [95% CI, −0.045 to 0.072]; ΔQALY adjusted, 0.010 [95% CI, −0.046 to 0.070]).

Resource Utilization and Costs

Resource utilization and associated costs were comparable between the arms (Data Supplement, Table S5, Table 2). The average total costs from health care payer and societal perspectives were $39,812 and $46,739 in the GAM group versus $37,450 and $45,177 in the UC group, translating into an incremental cost of $2,362 (95% CI, −$5,483 to $10,495) and $1,563 (95% CI, −$6,583 to $10,403) per patient, respectively. This difference, although not significant, was driven by higher chemotherapy ($1,438 [95% CI, −$5,830 to $9,442]) and radiation therapy ($1,239 [95% CI, −$191 to $2,682]) costs in the GAM arm (Fig 1A). The intervention costs accounted only for a small portion of total costs ($329 per patient [$214 + $41 + $74, Table 2], and $353 after accounting for adherence, Data Supplement, Table S2). Furthermore, patients in the intervention arm had slightly lower costs related to hospitalizations (−$280 [95% CI, −$3,020 to $2,541]) and lower out-of-pocket expenses for medical equipment (−$479 [95% CI, −$1,102 to −$16]).

TABLE 2.

Mean Costs by Resource Categories

FIG 1.

Mean incremental costs by resource category. Incremental costs, defined as the differences in cost between GAM and usual care groups, are displayed by (A) cost category for the full sample and (B) stratified by treatment intent. Costs were rounded to the nearest dollar and expressed in 2021 Canadian dollars. Procedures included day surgeries and ambulatory visits. Out-of-pocket expenses were related to transportation, parking, and other expenses related to visits to health care practitioners and facilities, as well as expenses for medication and medical equipment paid by a patient. Productivity loss included both patients' and caregivers' productivity loss. Note: Error bars were intentionally omitted from the plot to enhance visualization. Including error bars would have widened the y-axis range, making it difficult to visualize the differences in incremental costs between the categories. The 95% CIs of plotted incremental costs are available in Table 2 and the Data Supplement (Tables S6 and S7). ED, emergency department; GAM, geriatric assessment and management; HCP, health care provider; OOP, out-of-pocket.

Costs by treatment intent subgroup are summarized in Figure 1B and the Data Supplement (Tables S6 and S7). Although cost differences between the GAM and UC arms were nonsignificant for most categories, the mean total costs were higher in the intervention group for patients treated with palliative intent (incremental costs, $6,509 [95% CI, −$6,763 to $21,070]), but were lower for patients treated with curative intent (−$2,384 [95% CI, −$12,432 to $7,846]). Patients treated with palliative intent in the GAM arm reported higher costs for hospitalization ($746) and chemotherapy ($4,485) relative to UC. By contrast, patients treated with curative intent in the GAM arm had lower costs for the same resource categories (hospitalization, −$1,059 and chemotherapy, −$795) relative to UC.

Cost Utility and Incremental Net Benefit Analysis

Considering a societal perspective, on average, GAM was associated with decreased QALYs (−0.023 [95% CI, −0.076 to 0.028]) and increased costs ($1,563 [95% CI, −$6,583 to $10,403]) per patient (Table 3). Bootstrap replications representing pairs of incremental cost and utilities are plotted on a cost-effectiveness plane (Fig 2A). At a cost-effectiveness threshold of $50,000/QALY, the INMB was −$2,713 (95% CI of INMB, −$11,767 to $5,801; ICER, −$67,957/QALY), meaning that GAM was not cost-effective compared with UC (the probability of GAM being cost-effective was 25%, Fig 2A, Data Supplement, Fig S5).

TABLE 3.

Cost-Effectiveness Analysis for Geriatric Assessment and Management Relative to Usual Care (imputed data set)

FIG 2.

Cost-effectiveness plane for the (A) full sample and by (B) treatment intent. The graph shows difference in costs (y-axis) and difference in QALYs (x-axis) between GAM versus usual care for 20,000 bootstrap samples. Teal, blue, and red small dots represent bootstrap values for the full sample, for palliative treatment intent, and for curative treatment intent, respectively, while the bold circles of the same colors represent corresponding mean estimates. The blue dashed line represents the $50,000 per QALY threshold. Interventions are deemed cost-effective if they are below the cost-effectiveness thresholds. For example, for the full sample at $50,000 threshold, GAM is considered cost-effective in 25% of simulations compared with usual care (since 25% of bootstrapped samples [teal dots] are located below the $50,000 threshold line). GAM, geriatric assessment and management; ΔCosts, difference in costs; ΔQALYs, difference in quality-adjusted life-years.

For participants treated with curative intent, GAM was cost-effective in 72% of simulations (Fig 2B) with the INMB of $2,984 (95% CI of INMB, −$7,050 to $14,179; ICER, −$198,667/QALY). On the contrary, for patients treated with palliative intent, GAM was cost-effective in 8% of simulations with the INMB of −$9,909 (95% CI of INMB, −$24,436 to $4,153; ICER, −$95,720/QALY; Fig 2B and Table 3). Similar findings were found after adjusting for the baseline utility score, using a threshold $100,000/QALY and for the analysis conducted from a health care payer perspective (Table 3, Data Supplement, Table S8).

In the one-way sensitivity analysis, the chemotherapy and hospitalization costs had the biggest impact on the INMB both in the full sample and in treatment intent subgroups (Fig 3, Data Supplement, Fig S6). In the full sample, at least approximately 10% reduction in chemotherapy (<$18,000) or 30% reduction in hospitalization (<$6,200) costs would be required for GAM to become cost-effective at $50,000 per QALY threshold. In a subgroup of patients treated with curative intent, close to a 10-fold increase in intervention cost (>$3,000) would be required for GAM to become not cost-effective (Data Supplement, Fig S6).

FIG 3.

One-way sensitivity analysis. Parameters are listed in a descending order of their impact on the incremental INMB. The range of input values for each parameter is in parentheses and is based on 95% CIs. The black dashed vertical line corresponds to INMB of zero, while the teal axis represents the INMB from the base-case scenario (−$2,712). The bars and corresponding data labels indicate the INMB associated with maximum (red bars) and minimum (blue bars) inputs for each parameter (eg, as the cost of chemotherapy decreases for GAM arm, the INMB becomes more positive [the first bar] meaning that GAM is becoming a more attractive strategy compared with usual care). The values of the parameters when the bars are crossing zero (ie, when INMB becomes positive) are the thresholds at which the cost-effectiveness results are switching (eg, cost of hospitalization in GAM group needs to be $6,200 and less for GAM to become cost-effective, ie, for INMB to become >0). GAM, geriatric assessment and management; INMB, incremental net monetary benefit; K, thousands; UC, usual care.

For patients who were frail, the results were uncertain (INMB, $398 [95% CI, −$11,372 to $10,966]; probability of being cost-effective, 50%), and for not frail individuals, they were not cost-effective (INMB, −$7,623 [95% CI, −$23,078 to $7,362]; probability of being cost-effective, 15%), considering $50,000 per QALY threshold.

DISCUSSION

To our knowledge, we conducted the first economic evaluation alongside a trial comparing GAM with UC in older patients with cancer. Under the societal perspective, GAM was not cost-effective compared with UC in the full sample, with probability of being cost-effective of 25% at a $50,000/QALY threshold. The intervention costs accounted for only a small portion of total costs. The uncertainty decreased after evaluating by treatment intent with GAM being cost-effective for patients treated with curative intent and not cost-effective for patients treated with palliative intent.

Similar to other RCTs, we found no significant differences in survival or in QoL between the groups.^8-10,35 In the full sample, the average health care cost was higher in the GAM group (mean difference, $2,362 [95% CI, −$5,483 to $10,495]), which was primarily attributable to higher chemotherapy and radiotherapy costs. In our study, the rates (and subsequently costs) of ED visits and hospitalizations were similar, consistent with GAIN and GAP70+ trials.^8,9 The difference in total costs decreased after considering patient and caregiver costs using the societal perspective.

For patients treated with curative intent, GAM appeared to be cost-effective in >70% of simulations, supporting its added value in practice. Examining categories of costs, main factors contributing to cost-reductions appear to be reduced hospitalization-related costs likely because of shorter length of stay in the GAM group (Data Supplement, Table S5), as well as decreased chemotherapy-related costs resulting from higher rates of any chemotherapy dose reduction in GAM versus UC (50% v 45%, respectively, Table 1). The sensitivity analysis confirmed that the results were most sensitive to these two cost categories.

By contrast, GAM was not cost-effective for patients treated with palliative intent in >90% of simulations. Although not statistically significant, patients in the UC group appeared to have a better survival compared with the GAM group (Data Supplement, Fig S4B). This result may be due to unequal distribution of cancer types with poorer prognosis between the treatment groups (eg, a higher prevalence of thoracic and gastrointestinal cancer in the GAM group). Patients tend to incur the highest cost in the months preceding death.³⁶ Thus, having a higher proportion of patients nearing the end of life in the GAM group could explain the relatively higher costs there, for example, hospitalization costs, which could be less amenable to GAM interventions, particularly if the GA is not conducted before treatment initiation.

This study has a number of strengths. First, this economic analysis was based on patient-level data collected alongside a trial. Second, we adopted both a health care payer perspective and a societal perspective, which is sparse in cancer research. This broader perspective allowed to fully estimate the impact of intervention not only on the health care system but also on time lost from work by both patients and caregivers and OOP expenses.

We faced some difficulties when collecting data using patient diaries. The 5C study was halfway through when COVID-19 pandemic hit. This might have affected intervention delivery, data collection, and study outcomes, given associated mortality with the COVID-19 pandemic. Nevertheless, the patterns and rate of missing data were comparable between arms. We also applied multiple imputation to address this issue.^37,38 Although the randomization in 5C RCT was stratified by treatment intent, there were several cancer types (with variable sample sizes) not equally distributed within treatment arms in subgroups. Small sample size within subgroups limited our ability to conduct analysis adjusted for cancer type.

Ultimately, the decision to adopt GAM should consider scientific evidence, cost-effectiveness, and the overall health care structure. So far, RCTs have reported the largest effects for patients with advanced disease⁷ or with frailty.¹⁰ Expected survival alone, however, should not dictate whether to perform a GA. In the 5C study, 66% were frail and 46% had palliative treatment intent but health outcomes in the full sample or by frailty status were not different between GAM and UC groups.¹⁶ On the basis of cost-effectiveness analysis, GAM may not be a cost-effective approach for all patients with cancer, regardless of frailty status. However, it may be considered for patients treated with curative intent. Conversely, GAM is not likely to be cost-effective for patients with palliative treatment intent if it is provided after treatment initiation, which was how it was delivered in 98% of patients in the 5C trial. Considering the collective evidence, GAM may work best when performed before treatment initiation, regardless of treatment intent, to allow institution of supportive care measures, dose reduction, or other treatment modifications. This aligns also with the only published cost-effectiveness study of GA in cancer population to date, where the authors explored various GA delivery approaches, from least to most resource-intensive, and suggested that GA might become cost-effective if administered before chemotherapy with minimal resources.¹³ In addition, a recent single-center costing study of GA clinics demonstrated that GA was associated with cost-savings primarily because of decreased treatment intensity.²³ Further research in select populations (ie, by cancer type and treatment intent) is warranted to better understand for whom GAM may work.

To inform policy decisions, the economic evidence needs to be context-specific.³⁹ The 5C enrolled patients across Canada; hence, generalizability of findings in other countries and health care systems may be limited. To optimize resource allocation, for example, it is important to consider which components of GAM may be insured so that local practice guidelines can be modified. These components could include an assessment by a specialized, multidisciplinary team, by a dedicated geriatrician, by a trained oncologist, by a nurse, or by simply using electronic GA tools.⁴⁰ Although intervention costs are small and GAM may contribute to holistic management of patients, the resulting chemotherapy intensity reduction may translate into substantial cost-savings, particularly in countries with similar reimbursement structures and treatment practices.

This cost-effectiveness analysis provides additional considerations for future practice guidelines and efforts toward a wider adoption of GAM into care. As the field of geriatric oncology evolves and clinical and economic evidence accumulate, guidelines and coverage policies should be revisited and updated accordingly.

AUTHOR CONTRIBUTIONS

Conception and design: Yeva Sahakyan, Shabbir M.H. Alibhai, Martine Puts, Fay Strohschein, George Tomlinson, Lusine Abrahamyan

Financial support: Shabbir M.H. Alibhai

Collection and assembly of data: Shabbir M.H. Alibhai, Martine Puts, Sarah Brennenstuhl, Bianca McLean, Fay Strohschein, Aria Wills, Lusine Abrahamyan

Data analysis and interpretation: Yeva Sahakyan, Qixuan Li, Shabbir M.H. Alibhai, Martine Puts, Shant T. Yeretzian, Mohammed R. Anwar, George Tomlinson, Lusine Abrahamyan

Manuscript writing: All authors

Final approval of manuscript: All authors

Accountable for all aspects of the work: All authors

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

Cost-Utility Analysis of Geriatric Assessment and Management in Older Adults With Cancer: Economic Evaluation Within 5C Trial

The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated unless otherwise noted. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/jco/authors/author-center.

Open Payments is a public database containing information reported by companies about payments made to US-licensed physicians (Open Payments).