Trends in medical care utilization in patients with cancer: An analysis of real‐world data in a tertiary hospital in Korea, 2014–2019

Abstract Background Rising costs of cancer treatments challenge even areas with universal health coverage. There's a need to assess current medical care utilization trends among patients with cancer to guide public health policy, resource allocation, and set informed healthcare goals. Methods We analyzed the latest trends in medical care utilization by cancer patients in four areas—drugs, radiation therapy (RT), surgery, and diagnostic procedures—using clinical databases extracted from electronic medical records of a tertiary hospital in Korea between 2014 and 2019. Compound adjusted growth rates (CAGR) were computed to capture the annual growth over the study period. Results A total of 74,285 cancer patients were identified, with 40.3% (29,962), 14.2% (10,577), 31.1% (23,066), and 92.6% (68,849) of patients having received at least one anticancer agent, RT, surgery, and diagnostic procedure, respectively, over the period. We observed a 1.7‐fold increase in the use of targeted · immune‐oncology agents (from 6.8% to 11.6%) and a 21‐fold increase (from 3.0% in 2014 to 65.7%) in intensity‐modulated RT (IMRT) use over the period. In contrast, we observed a continuous decrease in the proportion of patients who underwent surgical treatment from 12.2% in 2014 to 10.9% in 2019. This decrease was particularly noticeable in patients with colon cancer (from 28.5% to 24.2%) and liver cancer (from 4.1% to 2.9%). Conclusion From 2014 to 2019, there was a significant rise in the use of targeted · immune‐oncology agents and IMRT, alongside a decline in surgeries. While targeted · immune‐oncology agents and IMRT may offer promising outcomes, their financial impact and potential for overuse necessitate careful oversight and long‐term cost‐effectiveness studies.


| INTRODUCTION
As the number of cancer treatment modalities has increased and their types have been diversified over the last few decades, the costs have also dramatically increased. 1or example, the annual price of novel anticancer agents easily exceeds 100,000 US dollars (USD). 2,3 Furthermore, innovative treatments such as CAR T-cell therapy can cost >500,000 USD per year. 46][7] Notably, in South Korea, cancer treatment expenses escalated by approximately 50 billion Korean won (approximately >37,000,000 USD) from 2014 to 2019, with an average annual growth rate of 12.7%. 8hus, many cancer patients may not afford to receive optimal treatments because of their high costs.For example, patients with cancer often take less than the prescribed dose, and avoid filling prescriptions because of the financial burden. 9Furthermore, increased out-of-pocket expenses for cancer treatment inhibit patients from getting access to optimal medical care irrespective of age and insurance status. 1 Universal health coverage, as exemplified by South Korea, aims to ensure all individuals have access to essential healthcare without financial strain. 10However, South Korea's universal health coverage does not cover all medical services. 8,10In South Korea, the expenses for these non-covered medical services have consistently risen, with an average annual growth rate of 7.7% over the past 5 years, from 11.5% in 2015 to 16.6% in 2019. 8,11mong the four major diseases-cancers, cerebrovascular diseases, cardiac diseases, and rare intractable diseases-cancer treatment expenses emerge as the most significant concern, accounting for 11.0% of costs not covered by insurance, surpassing cerebrovascular diseases (5.1%), cardiac diseases (5.2%), and rare intractable diseases (2.7%). 11espite the existence of South Korea's national health insurance, patients still face significant out-of-pocket costs, which affects the quality of cancer care. 12This underscores the urgency of smartly allocating limited resources.Real-world data (RWD) has been essential in determining these priorities, guiding decision-making, resource allocation, and setting healthcare objectives based on current cancer treatment patterns.4][15][16][17][18] Previous studies also have identified age-specific clinical practices for cancer treatments.For example, hormonal therapy was the treatment of choice in elderly breast cancer (BC) patients aged ≥75 years, while chemotherapy was more frequently prescribed in younger patients. 19ndeed, those studies have provided insights into how to better shape public health policy for patients with cancer based on the contemporary treatment trends. 20owever, those previous studies have been limited in that they did not include treatments that were newly developed and/or not covered by insurance.][23][24] Based on this understanding, the objective of this study was to analyze the latest trends in medical care utilization by patients with cancer in the following four areas: drugs, RT, surgery, and diagnostic procedures.To this end, we used two clinical databases extracted from electronic medical records (EMR) of a university hospital for recent 6 years.

| Data sources and ethical statement
We used EMR data extracted, transformed, and loaded (i.e., ETL process) according to the common data model (CDM) specifications by the Observational Medical Outcomes Partnership (version 5.3) and the clinical data warehouse or CDW (SUPREME®) at Seoul National University Hospital (SNUH), a university-affiliated Conclusion: From 2014 to 2019, there was a significant rise in the use of targeted • immune-oncology agents and IMRT, alongside a decline in surgeries.While targeted • immune-oncology agents and IMRT may offer promising outcomes, their financial impact and potential for overuse necessitate careful oversight and longterm cost-effectiveness studies.

K E Y W O R D S
cancer, immune-oncology agent, IMRT, medical care utilization, targeted agent tertiary-care hospital located in Seoul, South Korea. 25,26rom the SNUH CDM, we collected data on drug prescriptions, surgical treatments, and diagnostic procedures for each patient.Additionally, we extracted data on RT procedures from SUPREME®.
This study was reviewed and approved by the SNUH Institutional Review Board (IRB).The IRB waived the requirement for obtaining informed consent from the participants (IRB No: C-2105-213-1226).Information that may identify patients was anonymized.

| Study population
Eligible patients were those who had been diagnosed with BC, colon cancer (CC), liver cancer (hepatocellular cancer, HC), lung cancer (LC), or prostate cancer (PC), and visited SNUH at least once between January 2014 and December 2019.We selected these cancer types based on their high incidence, prevalence, and the growing financial burdens of their treatments in South Korea. 27Patients with cancer were identified using the following diagnosis terms: "cancer," "carcinoma," "tumor," and "malignant".Any patients diagnosed with benign, borderline, and/or in situ cancer were excluded from the analysis.

| Medical care utilization
Medical care utilization by patients with cancer in anticancer agents, RT, surgery, and diagnostic procedures was investigated.We selected anticancer agents from the list of all drugs prescribed to patients with cancer identified in the SNUH CDM.Selected anticancer agents were classified into four groups: cytotoxic agents, hormonal agents, targeted • immune-oncology agents (e.g., nivolumab and pembrolizumab), and immunomodulators (i.e., immunosuppressants and immune enhancers such as granulocytecolony stimulating factor [G-CSF]).Anticancer agents, which were not listed in the SNUH CDM at the time of conducting this study (e.g., talazoparib and ponatinib), were excluded from the analysis.
RT was classified into five types: photon beam RT (e.g., 10MV X-ray 2 ports), proton beam RT, three-dimensional conformal RT (3D-RT), stereotactic RT (i.e., stereotactic ablative RT [SABR] and fractionated stereotactic RT [FSRT]), and IMRT.We selected the most performed surgical treatment in each cancer type (Table S1) from the list of all surgical treatment operated to patients with cancer identified in the SNUH CDM.Surgeries operated on <0.5% of participants and not directly pertinent to cancer treatment were excluded from the analysis.Likewise, diagnostic procedures were categorized into six types: biopsy, computed tomography (CT), magnetic resonance imaging (MRI), positron emission tomography (PET), ultrasonography and X-Ray.

| Statistical analyses
We identified the number of patients with cancer who received ≥1 anticancer agent, RT, surgery, and diagnostic procedure, respectively, by year and/or type.We examined the prescription proportions of drug class and the proportions of patients who underwent each RT type, surgical treatment, and diagnostic procedure, respectively, by year.We summarized trends by calculating compound adjusted growth rates (CAGR) over the period.CAGR offers a representation of the mean annual growth rate over a specified duration, accounting for compounding across multiple time periods 28 : Data management, statistical analyses and data visualizations were performed with PostgreSQL (PostgreSQL Global Development Group) with DBeaver (ver 21.0.0) and R software (ver 4.1.3;R Foundation, Vienna, Austria).

| Radiotherapy (Figure 4)
From 2014 to 2019, the use of proton beam RT and photon beam RT dramatically decreased, that is, from 22 Note: Values are presented as frequency (%) and age was presented as mean ± standard deviation (years).
3.5 | Diagnostic procedure (Figure 6) X-ray was the most performed diagnostic procedure while the proportion of patients undergoing X-ray kept decreasing over the period from 75.8% in 2014 to 69.8% in 2019 (CAGR, −1.6%).In contrast, the proportions of patients undergoing ultrasonography, CT, and MRI increased over the period (CAGR, 2.7%, 1.9%, and 1.3%, respectively).All in all, PET, biopsy and MRI were used least among the diagnostic procedures.

| DISCUSSION
We observed a 1.7-fold increase in the use of targeted • immune-oncology agents (from 6.8% to 11.6%, Figure 3), and a 21-fold increase (from 3.0% in 2014 to 65.7%, Figure 4) in IMRT use from 2014 to 2019.These findings were consistent with the results in previous studies, reporting a 1.7-and 1.6-fold increase in the use of antineoplastic monoclonal antibodies and protein kinase inhibitors, respectively, from 2010 to 2016, 29,30 as well as an 18-fold increase in IMRT utilization from 2011 to 2018 in Korea. 17The increased use of targeted • immune-oncology agents and IMRT in patients with cancer is most likely attributed to the fact that their clinical utility has been repeatedly shown.][36][37] In contrast, we observed a continuous decrease in the proportion of patients who underwent surgical treatment, from 12.2% in 2014 to 10.9% in 2019 (CAGR, −2.2%, Figure 2).This decrease was particularly noticeable in patients with CC (from 28.5% to 24.2% [CAGR, −3.2%]) and HC (from 4.1% to 2.9% [CAGR, −7.1%], Figure 5).The decline in surgical treatment may have been caused by increased adoption of non-surgical treatments, possibly driven by advancements and wider use of targeted • immune-oncology agents. 38,392][43][44] We observed that such programs have indeed amplified the usage of the targeted • immune-oncology agent (Figure 3). 29,30Despite these efforts in South Korea, many individuals have not benefited from these programs due to the complexities and limitations of insurance coverage for such treatments. 8,11,126][47][48] Thus, we highlight that further studies are necessary to evaluate the cost-effectiveness of targeted • immune-oncology agents to ensure long-term affordable and sustainable patient access.
Over the past few years, the adoption of IMRT has raised globally. 17,495][36] In South Korea, the expansion of health insurance coverage for IMRT since 2015 may have led to overutilization of IMRT. 17Indeed, we observed a sharp increase of IMRT use since 2015 (Figure 4).We emphasize the necessity to implement quality control measures to ensure appropriate use and prevent the misuse and abuse of IMRT.
This study had three major limitations.First, this study's results may not be generalizable to other hospital settings because the analysis was based on data from a single institution.However, the consistency of our findings with those from other studies may support that the results are practically extrapolated to other situations.Second, some anticancer agents and procedures were not included in the analysis because they were not available in the SNUH CDM at the time of the study.This may have led to an underestimation of the actual number of patients who received each treatment or diagnostic procedure.Third, we were unable to segregate the patient cohort into more granular subcategories and perform advanced statistical analysis such as time series analysis with diverse parameters.This limitation is inherent to the retrospective observational study and the available data variables.For example, patient phenotyping was problematic, as we could not discern between those newly diagnosed and returning patients; the original diagnosis location for patients at SNUH, especially if they had prior diagnoses elsewhere, was unclear.
Despite these limitations, to the best of our knowledge, our study is the most up-to-date and exhaustive analysis in South Korea which have included treatments and diagnostic procedures that are newly developed and not yet covered by insurance.We successfully achieved our primary objective of illustrating the current trends.However, more intricate analyses, such as detailed stratification of patient with richer datasets and enhanced time series analysis, could potentially reveal trends and patterns not identified in our study.
In conclusion, the use of targeted • immune-oncology agents and IMRT had continuously increased over the period of 2014-2019.Therefore, healthcare policymakers and practitioners should focus on the evaluation of longterm cost-effectiveness of targeted • immune-oncology

F I G U R E 3 F I G U R E 4
Annual trends in the use of anticancer agents by patients with breast, colon, liver, lung, or prostate cancer between 2014 and 2019: cytotoxic agents (blue), hormonal agents (red), targeted • immune-oncology agents (green), and immunomodulators (orange).(A) Number of patients prescribed ≥1 anticancer agent by drug class.(B) Number of anticancer agent prescriptions by drug class.(C) Prescription proportions of each drug class.Annual trends in the utilization of RT by patients with breast, colon, liver, lung, or prostate cancer between 2014 and 2019: photon beam RT (red), proton beam RT (orange), 3D-RT (green), stereotactic RT (blue), and IMRT (purple).(A) Number of patients received ≥1 RT by RT type.(B) Number of treatments performed by RT type.(C) Proportions of RT type.RT, radiotherapy; three-dimensional conformal RT, 3D-RT; intensity-modulated RT, IMRT.

F I G U R E 5
Annual trends in the utilization of surgical procedures by patients with breast, colon, liver, lung, or prostate cancer between 2014 and 2019.(A) Proportions of patients who received surgical treatments by year and cancer types: breast cancer (blue), colon cancer (orange), liver cancer (gray), lung cancer (yellow), and prostate cancer (green).(B) Proportions of patients who received each surgical procedure by year in colon cancer: excision of mucosa of colon (blue), rectal resection (orange), partial of total removal of colon (gray), polypectomy (yellow).(C) Proportions of patients who received each surgical procedure by year in liver cancer: hepatectomy (blue) and liver transplantation (orange).

F I G U R E 6
Annual trends in the utilization of diagnostic procedures by patients with breast, colon, liver, lung, or prostate cancer between 2014 and 2019.(A) Number of diagnostic procedures performed by modality and year.(B) Proportions of patients who received each diagnostic procedure by year.CT, computed tomography; MRI, magnetic resonance imaging; PET, positron emission tomography.
T A B L E 1