Hospitalization rates from radiotherapy complications in the United States

Hospitalizations due to radiotherapy (RT) complications result in significant healthcare expenditures and adversely affect the quality of life of cancer patients. Using a nationally representative dataset, the objective of this study is to identify trends in the incidence of these hospitalizations, their causes, and the resulting financial burden. Data from the National Inpatient Sample was retrospectively analyzed from 2005 to 2016. RT complications were identified using ICD-9 and ICD-10 external cause-of-injury codes. The hospitalization rate was the primary endpoint, with cost and in-hospital death as secondary outcomes. 443,222,223 weighted hospitalizations occurred during the study period, of which 482,525 (0.11%) were attributed to RT. The 3 most common reasons for RT-related hospitalization were cystitis (4.8%, standard error [SE] = 0.09), gastroenteritis/colitis (3.7%, SE = 0.07), and esophagitis (3.5%, SE = 0.07). Aspiration pneumonitis (1.4-fold) and mucositis (1.3-fold) had the highest relative increases among these hospitalizations from 2005 to 2016, while esophagitis (0.58-fold) and disorders of the rectum and anus were the lowest (0.67-fold). The median length of stay of patient for hospitalization for RT complications was 4.1 (IQR, 2.2–7.5) days and the median charge per patient was $10,097 (IQR, 5755–18,891) and the total cost during the study period was $4.9 billion. Hospitalization for RT-related complications is relatively rare, but those that are admitted incur a substantial cost. Use of advanced RT techniques should be employed whenever possible to mitigate the risk of severe toxicity and therefore reduce the need to admit patients.

Study design. Cancer patients were identified in the NIS using Classifications Software (CCS) codes 11-45 14 . Cancers were divided into two categories: solid tumors and hematologic malignancies (see Table 1). External cause-of-injury codes (E-codes) from ICD-9-CM and ICD-10-CM were used to identify RT complications. The primary reason for hospitalization was the first reported non-cancer disease identified using an ICD-9-CM or ICD-10-CM diagnosis code (Supplemental Table 1). The CONSORT diagram depicts the technique for selecting the patient cohort (Supplemental Fig. 1). The study examined demographics such as age, gender, and race; socioeconomic factors included median household income by zip code, insurance type; and hospital characteristics included region, bed size, and teaching status. The key end measures were the hospitalization rate and total cost, while the length of stay in the hospital and in-hospital mortality for RT complications were secondary outcomes. Our findings were reported using the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) guidelines. The methods for this study were conducted in accordance with relevant guidelines and regulations. The study was reviewed by the Miami Cancer Institute's Institutional Review Board, which exempted the study from institutional review board approval and waived the requirement for informed consent because it uses previously collected deidentified data stored in NIS.  15 . The NIS was updated in 2012 to be more nationally representative, as previously noted. We used modified discharge weights for the years 2005-2011 to account for these changes in data collection 16 . Temporal factors, cancer types, demographics, socioeconomic factors, and hospital characteristics were all evaluated using descriptive statistics. Frequencies and percentages were used to describe categorical variables and median and interquartile ranges were used for continuous variables. The total number of hospitalizations due to RT complications was divided by the total number of hospitalizations to calculate the hospitalization rates. Total hospital charges and cost-to-charge ratios were multiplied to calculate the costs of individual inpatient stays. The expenditures for each year were adjusted based on 2016 inflation levels using the US Consumer Price Index. By subtracting the admission date from the discharge date, the length of stay in the hospital was calculated. The number of patients admitted for RT complications who died in the hospital was divided by the total number of patients hospitalized for RT complications to calculate the In-hospital mortality. Weighted frequencies were used in all analyses to calculate national estimates using discharge weight, or "DISCWT." Ethics approval statement. Since the study uses previously gathered deidentified data maintained in NIS, the study was excused from the institutional review board approval. All of the methodologies followed the appropriate ethical requirements for handling data.
Informed consent statement. Since this study involves an administrative database that does not contain any personal information that may be connected to an individual participant, informed consent was not essential.

Discussion
To our knowledge, this is the first study to use a nationally representative dataset to report hospitalizations for RT complications. There was an overall decrease in hospitalizations between 2005 and 2016, but hospitalizations related to complications from RT remained stable over the same time period. A vast majority of these hospitalizations occurred in adults > 60 years of age. In terms of insurance providers and hospital locations associated with admission, Medicare and urban teaching centers were dominant. The complications associated with the highest frequency of hospitalization were radiation cystitis, radiation-induced gastroenteritis and colitis, and esophagitis. The steady rate of hospitalizations for complications of RT, in spite of an overall increase in the number of patients receiving RT, is likely explained by the overall decrease in adverse events associated with RT over time because of technical improvements and better management of toxicities in the outpatient setting. Factors such as improved image guidance during treatment, advanced treatment planning systems and highly conformal and accurate delivery techniques of radiotherapy such as intensity-modulated radiation therapy (IMRT), imageguided radiation therapy (IGRT), volumetric-modulated arc therapy (VMAT), stereotactic radiotherapy and particle therapy (proton beam therapy) have been increasingly adopted in the last few years, and have reduced the dose delivered to normal structures thereby decreasing the intensity of radiation associated side effects 17 .
In this analysis, we noted that the most patients were admitted to large urban teaching hospitals. One possible explanation for this could be the fact that RT facility involves high investment and is usually only available at larger hospitals, and patients are prone to seeking care at hospitals where they received their RT 18 . By 2030, it is projected that 74.1 million (20.6%) Americans will be 65 years or older, and cancer incidence is expected to increase by 67% between 2010 and 2030 19 . A significant proportion of these patients will receive RT as part of their treatment 20 . In our study, we discovered that more than half of the patients who required hospitalization after RT were 65 years and above. Understanding the unique differences that older cancer patients have as compared to their younger counterparts, both in terms of tumor biology, overall health and functional status is critical in approaching the management in this patient population 21,22 . The treating physicians may find it beneficial to have a working understanding of the available methods for completing a geriatric assessment that may affect treatment outcomes in older adults in order to provide comprehensive care 23 .
Radiation therapy is a localized treatment, therefore the complications that may arise are limited to the organs that are in close proximity to the tumor 24 . In our study, esophagitis was common among lung and breast cancers since esophagus is the one of the organs at risk (OAR) in these tumors. Symptoms such as throat pain, dysphagia, and the feeling that food is stuck in the chest normally appear 2 to 3 weeks after treatment 25 . The more the esophagus is included in the radiation field (the region of the body receiving radiation), the higher the risk of esophagitis. Furthermore, with the introduction of hypofractionation, the patient is more likely to develop esophagitis 26 . Similarly, intestinal obstruction was also found to be common among cancers of the bladder, cervix, colon, rectum, and uterus, owing to the close proximity of the bowels, which serve as the OAR for these tumors. This condition develops as a result of intestine discomfort following radiation therapy and can be alleviated by reducing the bowel radiation dose 27 .
In our study, aspiration pneumonitis, intestinal obstruction without a hernia, and mucositis were all linked to a longer duration of stay and higher mortality rates in patients undergoing radiotherapy. Given that RT disrupts mucosal barriers, resulting in mucosal alterations and muscle fibrosis, this is understandable. Aspiration www.nature.com/scientificreports/ pneumonitis is caused by a variety of factors, including acute and chronic radiation-induced mucosal changes, muscle fibrosis, and xerostomia. These factors often result in swallowing difficulties, increasing the risk of aspiration and aspiration pneumonia, as well as the duration of stay in the hospital [28][29][30] . Similarly, Berger et al. looked at the burden of oral mucositis across 65 studies, showed 30% patients required hospitalization due to mucositis 31 . We also found that managing the complications of RT from these hospitalizations involved significant expenditures over a 10-year time period-$4.9 billion-leading to an annual estimated health-care burden of $408 million. This poses an enormous burden on healthcare resources. Surprisingly, there are very few studies that report the cost associated with treating the complications of RT at the national level 32 . The high hospitalization rate identified in our study could be used to identify patients whose treatment could be managed in an outpatient setting. Considering these findings, future studies should focus on estimating the cost associated with these complications across different healthcare settings, as well as different RT technology platforms, such as 2D, 3D, IMRT, proton, brachytherapy, etc. A better understanding of variables that predispose patients to toxicities leading to hospitalization could result in prospective interventional approaches. Recommendations for screening patients at high risk of complications in the outpatient setting and development of standardized management plans may have real world potential benefits to improve care and reduce hospitalizations.

Limitations
E-codes were utilized to identify patients who had RT-related complications resulting in hospitalization. As NIS is an administrative database, E-code documentation which is not an obligatory requirement for reimbursement could result in underestimation of the actual hospitalization burden related to RT complications. Furthermore, the NIS does not contain data on tumor staging, concomitant medications, or the type and dose of RT administered, making it difficult to perform more sophisticated analysis. Furthermore, because NIS removes all personal identifiers for anonymity, successive readmissions of the same patient would be viewed as separate new admissions, leading to some overestimation. Despite these drawbacks, we chose NIS since other data sources for constructing a comprehensive picture of this subject are even more limited and less helpful in terms of national estimates. NIS is the largest all payer in-patient national database in the United States and its extensive reach and enormous size regarding data capture makes it more nationally relevant in comparison to other sources.

Conclusions
RT-related hospitalizations are a small percentage of all hospitalizations. During 2005-2016, hospitalization rates for complications of RT remained almost same and did not show significant reductions; as a rough estimate, one in every thousand admissions to US hospitals occurs because of a RT-related complication, and this rate has remained unaltered between 2005 and 2016. The most common RT complication that required hospitalization were radiation cystitis, radiation-induced gastroenteritis and colitis, and esophagitis. When extrapolated to the national level for all admissions, the overall financial burden of managing RT related complications from these hospitalizations over a 10-year time period was $4.9 billion, leading to an annual estimated health-care burden of $408 million. Improved outpatient and emergency-room strategies for detecting and managing these complications could dramatically reduce the need for these hospitalizations. Identifying patients at greatest risk could also be useful as they could potentially be treated with more sophisticated techniques that reduce radiation dose to critical organs-at-risk, thereby likely diminishing toxicities and hospitalizations.

Data availability
The data that support the findings of this study are available from the corresponding author upon reasonable request.