Direct medical costs of young-onset colorectal cancer: a worldwide systematic review

We conducted a literature search of Ovid MEDLINE, Ovid Embase and Science Citation Index and Social Sciences Citation Index via Web of Science from inception to July 15, 2021, and then updated the search on May 30, 2022. To ensure comprehensive capture of articles that may assess yCRC as a subgroup of CRC, we employed a broad search strategy to identify articles on CRC across all ages, from which data pertaining to yCRC could be extracted. Our search strategies used a combination of subject headings (e.g. Medical Subject Headings in Medline) and keywords to locate studies. Search terms related to economics/costs were adapted from a search filter developed by Canadian Agency for Drugs and Technologies in Health [7] (Tables S1-3).

We reviewed titles and abstracts to identify published studies that met our systematic review inclusion criteria of: 1) an original study; 2) published in a peer-reviewed journal as a full-length article; 3) patient population with CRC or yCRC; 4) published in English; and 5) reported the direct medical cost of yCRC, defined as costs directly related to provision of medical services for patient care (e.g., surgery, prescription medications, inpatient care, outpatient visits, chemotherapy therapy and radiation therapy). Given potentially different age cut-offs for defining ‘young-onset’, for the purposes of our systematic review, we broadly considered an age cut-off of up to 65 years old in their cost estimation and reported as yCRC. We excluded non-original literature (i.e., reviews and editorials), economic evaluations (i.e., cost-effectiveness of interventions or programs), studies which estimated costs associated with co-morbidities related to CRC or side effects of cancer treatment, studies referring to the cost of CRC screening, studies comparing costs associated with different cancer treatment protocols, and conference proceedings. While the aim of our study was to extract the direct medical costs associated with CRC treatment, we also accepted studies that included cost of illness prior to diagnosis in their definition for CRC treatment, as many countries lack standard screening protocols, resulting in increased spending prior to pathological confirmation of diagnosis. Three authors (RG, VC, VV) independently reviewed the titles and abstracts of articles identified from literature search and resolved discrepancies by consensus. Abstracts that met our inclusion criteria were forwarded for full-text review. The same three authors independently assessed articles eligible for a full-text review based on the inclusion criteria. To assess included studies for quality, we used a modified version of the Consolidated Health Economic Evaluation Reporting Standards (CHEERS) checklist [8], which includes elements that assess for both study and reporting quality and has been used in prior cost of illness systematic reviews [9] (Table S4). We defined studies that scored 14 or greater points on the modified CHEERS checklist as studies of ‘good quality’.

To characterize the included studies, we extracted information on country, data source, length of follow-up, cancer site (i.e., colon, rectum), sex, sample size, and age-cut off used for estimating and reporting costs. Of particular interest in our systematic review, we extracted detailed information on: 1) costing approach, such as source of payment (i.e., public health spending, private health insurance, out-of-pocket costs); 2) payer perspective (i.e., societal, healthcare provider); 3) whether costs were attributable to (i.e., the mean difference in cost of care between individuals with cancer and without, also referred to as net costs) or associated (i.e., all-cause costs incurred after a CRC diagnosis, which may include the cost of CRC treatment and co-existing conditions) with CRC; and 4) cost components (e.g., chemotherapy, radiation, outpatient visits, inpatient care, and prescription medications). We then extracted the reported per-capita direct medical costs for both prevalent (i.e., existing and newly diagnosed patients) and incident cases of yCRC, and where relevant, aCRC. Also, where relevant, we extracted information on cost comparisons between yCRC and aCRC cases, specifically those based on the use of multivariable approaches. For brevity in our reporting of results, we use the term ‘costs’ to refer to direct medical costs. To facilitate comparisons, all extracted costs were inflation-adjusted to 2020 USA dollars (USD) using the Consumer Price Index unless otherwise specified. Costs in the original currency are provided in Table S5.

Our search strategy resulted in 17,764 articles on July 15, 2021 and 1,584 articles in the May 30, 2022 update (Fig. 1). Article assessment led to exclusions for the following reasons: did not report direct medical costs; did not include patients under the age of 65; and did not report costs stratified by age. Overall, 14 studies met all eligibility criteria and were included in the systematic review – 10 from the original search and 4 from the update.

Characteristics of studies included in the systematic review are shown in Table 1. We grouped studies according to those of prevalent (n = 5) [10‐14] or incident disease (n = 9) [15‐23]. Studies varied in countries where they were conducted, which included the USA [11, 17, 19], England [15, 18], France [16], China [12, 20], Korea [10], Vietnam [13], Italy [14], Australia [21], Canada [22] and Japan [23]. Five studies utilized claims data from a health insurance database [10, 16, 17, 19, 23], three used data abstracted from hospital medical records [12, 13, 20], four used administrative health data [14, 15, 18, 22], one study used data from a nationally representative medical expenditure survey [11], lastly another study linked self-reported survey data with an administrative health database [21]. It must also be noted that studies which relied on either survey data [11, 21] or hospital medical records [12, 13, 20], may omit data from non-responders and those receiving care from other hospital(s), respectively, which may impact the generalizability of their cost estimates. Most studies estimated the cost of CRC, except for three studies which reported cost estimates by cancer site. Specifically, studies by Gigli and Taplin et al., focused on rectal [14] and colon [17] cancer, respectively, while Goldsbury et al., stratified cost estimates by age and cancer site (i.e., colon and rectum) [21]. CRC was identified using International Disease Classification codes in 10 studies [10, 13, 14, 16‐19, 21‐23], two studies used hospital medical records [15, 20], and two studies relied on self-reported diagnosis [11, 12]. With respect to cancer stage at diagnosis, 10 studies included CRC patients with stage I to IV disease [12‐18, 20, 22, 23], two studies did not report cancer stage [10, 11], one study included CRC patients with a stage IV de novo (no prior CRC diagnosis) diagnosis or a recurrent (previously diagnosed with stage I-III CRC) cancer diagnosis [19], and one study reported cancer severity by the extent of cancer metastasis (i.e., localized, regional, distant metastases) [21]. Overall, nine studies [12, 14, 15, 18‐23] scored well against the modified CHEERS checklist with scores ≥ 14 and thus being classified as studies of ‘good quality’ (Table S4). Items not consistently reported by studies that scored below 14 points on the modified CHEERS checklist [10, 11, 13, 16, 17] included, specification of their study perspective (i.e., authors failed to specify their costing perspective), detailed methods for how they estimated resource utilization and costs components, the price date and conversion rate used, and provision of study parameters (i.e., values, references and ranges for input values used to estimate cost).

Table 2 highlights aspects of the costing methodology employed by each included study. Given our particular interest in those with yCRC, we synthesized age cut-offs applied, which varied from < 50 years in five studies [10, 13, 14, 22, 23], < 55 years in three studies [12, 20, 21], and < 65 years in six studies [11, 15‐19]. While all studies reported costs of CRC treatment (i.e., chemotherapy and radiation therapy) and in-patient visits (i.e., hospitalization), only seven of the included studies considered costs related to ambulatory care, including both outpatient medical visits and prescription medications [10‐12, 14, 16, 17, 23]. Additionally, studies varied in terms of the selected costing perspective, with five studies adopting a societal perspective (i.e., considering both patient out-of-pocket costs and health provider spending) [10‐13, 20] and nine studies adopting a health provider perspective, of which six estimated public payer spending [14, 15, 18, 21‐23] and three utilized data from private payers, including insurance companies [16, 17, 19].

Table 2 summarized key results, namely reported costs for yCRC and where relevant, aCRC. Five studies of prevalent disease reported the annualized per-capita cost for existing as well as new cases of CRC [10‐14]. Four out of the five aforementioned studies estimated the annualized per-capita cost associated with CRC (i.e., all-cause costs incurred after a CRC diagnosis, which may include the cost of CRC treatment and co-existing conditions) [10‐13]. Specifically, the annualized per-capita cost associated with yCRC ranged from $2,263 to $16,801 (inflation-adjusted to 2020 USD) and $1,412 to $14,997 for aCRC (costs in original currency are provided in Table S5). Among studies estimating the prevalence based cost of CRC, Byun et al. used administrative health data in Korea to account for the cost of treatment (i.e., chemotherapy, radiation and outpatient/inpatient visits) and prescription medications related to a CRC diagnosis (i.e., costs for prescription medications unrelated to CRC treatment were excluded), with estimated per-capita costs associated with yCRC ranging from $9,778 to $16,078 and $13,658 to $16,801 among males and females, respectively. [10]. Inspection of costs stratified by sex and age categories (i.e., 10–19, 20–29, 30–39, 40–49 years) suggest higher costs among females and younger age groups. Next, Zheng et al., used the Medical Expenditure Panel Survey in the US and estimated costs associated with yCRC to be $13,837 and the net cost of illness (or attributable costs, defined as costs of CRC cases minus cancer-free controls) to be $9,745 [11]. Here, net costs accounted for 70% of the total cost of illness incurred by individuals with yCRC. This contrasts to those with aCRC where net costs ($5,537) accounted for 37% of the total cost of illness ($14,997) [11]. Authors also stratified the overall reported cost by cost components (i.e., outpatient/inpatient visits and prescription medications), with outpatient visits accounting for 57% ($5,597) of the net cost for yCRC, in comparison to 30% ($1,668) of the net cost for aCRC patients [11]. Additionally, it must be noted that the yCRC age cut-off applied by Zheng et al., is 65 years [11], while remaining studies employed age cut-offs of < 50 [10, 13, 14] or 55 years old [12]. Using hospital medical records and patient-reported survey data, Huang et al., used generalized linear models to compare the cost of yCRC and aCRC [12]. Specifically, they demonstrated that costs did not significantly differ for those diagnosed at less than 45 (p = 0.419), 45–54 (p = 0.131) and 55–64 (p = 0.522) years old compared to those diagnosed at 65 (reference) years or older [12]. Among prevalence-based, cost estimates by Tran et al., in Vietnam, were the lowest – ranging from $2,263 to $2,517 and $1,412 to $2,191 for yCRC and aCRC, respectively. Low estimation of costs associated with CRC may be attributed to the use of medical expenditure data from a single hospital site. Additionally, authors did not specify whether they included costs for radiation and chemotherapy in their cost estimates, as they simply stated that ‘cancer treatment’ were included in their calculations [13].

Among more recent studies, Gigli et al., used administrative health data to estimate costs directly related to rectal cancer treatment (i.e., cost calculation only considered the cost of procedures and medications used for rectal cancer treatment) by phases of care [14]. Specifically, among young-onset rectal cancer patients costs were estimated to be $27,692, $3,709 and $31,359 during the initial (first 12 months of treatment), continuing (the months between the continuing and terminal phase) and terminal (last 12 months before death) phases of care, respectively [14]. Here, authors found the cost of hospitalizations to be the main driver of the total cost estimate for each phase of care. Cost estimates followed a decreasing trend with age, with average-age onset rectal cancer treatment costs ranging from $18,214 to $24,083, $1,620 to $3,290, and $7,311 to $25,450 during the initial, continuing and terminal phases of care, respectively [14].

Nine studies estimated the per-capita cost of incident CRC cases, with differences between the selected time horizon, point of care (i.e., treatment phase, continuing phase) and cancer site (i.e., colon or rectum). Of note, in contrast to prevalence-based costing studies that reported age cut-offs consistent with current definitions of yCRC (i.e., cut-offs at 45, 50 years), the majority (n = 5) of incidence-based costing studies used an age cut-off of less than 65 years old [15‐19].

Five out of the nine included studies captured costs incurred 12 months following a CRC diagnosis, with reported per-capita costs ranging from $23,368 to $89,945 and $19,929 to $67,195 among yCRC and aCRC, respectively [15, 16, 18, 19, 22]. Using data from a health insurance database in France, Clerc et al., stratified their cost estimates by the different healthcare components and reported that 54% ($24,988) and 52% ($22,489) of overall costs incurred during the first 12 months following a CRC diagnosis to be attributed to inpatient hospital visits, among yCRC and aCRC respectively [16]. Additionally, the study employed a multivariate linear regression model to evaluate the impact of covariates on the cost associated with CRC and found no significant association with age at diagnosis [16]. Next, using electronic health records from a cancer registry in the US, Ritzwoller et al., estimated the cost associated with de novo metastatic (stage IV) cancer to be $89,945 and $67,195 among yCRC and aCRC, respectively [19]. Paszat et al., conducted a cohort study in Ontario, Canada using administrative health data and estimated the cost CRC among individuals with a hereditary CRC syndrome to be $44,291 and $41,063 among yCRC and aCRC patients, respectively [22]. Lastly, studies by Hall et al. and Laudicella et al., used an extended time horizon and captured costs beyond 12 months after diagnosis [15, 18]. Specifically, with the use of administrative health data in England, Hall et al., estimated the cost associated with yCRC to be $19,097, $23,368 and $25,319 at 6, 12 and 15 months post-diagnosis, respectively [15]. Authors also used multivariate linear regression models, which demonstrated that age at diagnosis did not significantly impact cost associated with CRC care [15]. Similar results are seen by Laudicella et al., who used administrative health data in England and stratified costs by stage at diagnosis. Here authors extended the time horizon to 9 years since diagnosis, with the highest cost being incurred during the first 12 months after diagnosis [18]. Cost estimates stratified by stage of diagnosis, indicate a greater difference in the cost of illness between early and late stage diagnosis among yCRC (stage I/II $27,360; stage III/IV $35,206; difference -$7,846), as this trend was less pronounced among aCRC patients (stage I/II $26,048; stage III/IV 28,277; difference -$2,229) [18].

Shi et al., used hospital medical records in China to examine trends in the cost of CRC treatment over time (time horizon unspecified) [20]. Here, authors found the costs associated with CRC become increasingly more expensive over the years. Particularly among individuals with yCRC, as the per-capita cost between 2009 and 2011 was estimated to be $9,846 compared to $7,085 between 2002 and 2011 among yCRC patients less than 45 years old. However, this trend was less pronounced among aCRC patients greater or equal to 65 years old (2002–2011, $6,933; 2009–2011, $8,354) [20].

Utsumi et al., used national health insurance claims data in Japan and estimated costs associated with CRC stratified by age and treatment type [23]. Specifically, among yCRC patients the mean cost of care ranged from $5,717 to $8,733 for those who received endoscopic treatment, $22,755 to $26,386 for those who received surgery, and $55,713 to $72,016 for those who received palliative care (i.e., consider non-curable CRC) [23]. In contrast to the aforementioned studies, cost estimates followed an increasing trend with age at diagnosis, as the mean cost of care ranged from $8,184 to $12,811, $26,507 to $33,424 and $70,936 to $80,516 among aCRC patients who received endoscopic, surgical and palliative care, respectively. The low cost of care for patients who received endoscopic treatment, which relates to a lower stage at diagnosis relative to the surgical and palliative treatment groups (indicating late stage diagnosis) among both yCRC and aCRC patients, emphasizes the cost-benefits of early stage diagnosis and treatment.

Lastly, two out of the nine included studies stratified cost estimates by age and cancer site [17, 21]. Taplin et al., used electronic health records from a cancer registry in the USA to estimate the cost of colon cancer at different stages of care among young-onset patients. Specifically, costs were estimated to be $28,013 $30,046 and $2,594 during the initial (first 6 months of treatment), terminal (last 6 months before death) and continuing (the months between the continuing and terminal phase) phases of care, respectively [17]. Meanwhile, Goldsbury et al., linked survey data to an administrative health database and estimated the cost associated with CRC, stratified by both age and cancer site [21]. Specifically, they estimated the cost of colon cancer to be $36,064 and $23,251 to $31,869 among young-onset and average-age onset patients, respectively [21]. In comparison to colon cancer, there was a greater cost associated with rectal cancer, estimated to be $40,720 and ranged from $30,754 to 36,488 among young-onset and average-age onset patients, respectively. Additionally, use of multivariable gamma regression models demonstrated costs estimates to be greater among those diagnosed with colon cancer between the age of 45 and 54 years (effect size 1.10, 95% confidence interval [CI] 1.00–1.12), compared to those diagnosed at the age of 65–74 (reference) [21]. Among rectal cancer patients, a greater estimated cost of care was observed among those diagnosed between the ages of 45–54 (effect size 1.15, 95% CI 1.03–1.28) and 55–65 (Effect size 1.09, 95% CI 1.00–1.18), compared to those diagnosed at the age of 65–74 (reference) [21].