Characteristics of Life-Sustaining Treatment Decisions: National Data Analysis in South Korea.

This study analyzed the national data on life-sustaining treatment decisions from 2018 to 2020 to find out the characteristics of South Korea's end-of-life procedure according to the decision-making approach and process. We collected the data of 84,422 patients registered with the National Agency for Management of Life-sustaining Treatment. We divided the patients into four groups (G1, G2, G3, and G4) according to the decision-making approach. A descriptive analysis of each group was conducted using indicators such as the patient's age, status, diagnosis, and content of forgoing life-sustaining treatment. Additionally, logistic regression analysis was performed by dividing the patients into self-determining (G1, G2) and non-self-determining patients (G3, G4). Cancer was the most common diagnosis for each group. The period from life-sustaining treatment decision to implementation was 10.76, 1.01, 0.86, and 1.19 days for G1, G2, G3, and G4, respectively. In the logistic regression analysis, the self-determination ratio was higher for 40-49 years old and lower for cardiovascular disease and gastrointestinal disease. Age was has a major impact on life-sustaining treatment decisions (LSTD), and with increase in age, the family, and not the patient, made the LSTD. The LSTD method also differed depending on the disease. The self-determination rates of patients with circulatory or digestive diseases were somewhat lower than that of those with neoplastic diseases. The period from decision-making to implementation is short for end-of-life care.


Introduction
In 2018, the Act on Decisions on Life-sustaining Treatment for Patients in Hospice and Palliative Care or at the End of Life (LSTD Act) was enacted in South Korea to determine hospice-related medical treatment. This is the rst law on end-of-life care respecting the patient's right to self-determination and ensuring their best interests regarding a digni ed death (Article 1). [1]: First, the LSTD Act documents the contents of advance care planning, which is based on patients' intention. The documents are categorized based on who makes the decision: the patient (Advance Directive [AD]) or the doctor (Physician Orders for Life-Sustaining Treatment [POLST]). Treatment preferences recorded in these forms go through veri cation procedures during the end-of-life, which is documented and reported to the national agency in South Korea. Second, only families of patients prescribed by civil law may be involved in treatment decisions. If the patient cannot directly con rm treatment preference, the family will state the patient's intention or obtain consent of the whole family.
The decisions are determined on the patient's behalf by the authority granted by law. Third, the lifesustaining treatment decision (LSTD) process is divided into terminal-term and end-of-life. If a doctor diagnoses an end-of-life patient, he/she should write and register legal documents. Fourth, the LSTD system has a national supervisory and managing agency that establishes a database to manage legal forms, including POLST and AD, and guides and supervises matters related to LSTD. It also calculates monthly statistics on LSTD status.
As of January 2020, 85,076 cases have been implemented, [2] of which the number of decisions made through family statements or agreements (55,775) is nearly twice as high as that made using POLST or AD (29,301). The LSTD system is thus based on family decisions rather than on patients' preferences or wishes, [3][4][5] and a more accurate understanding of the decisions is needed. However, the national agency only provides simple gures of the cumulative number of implementation patients, and limitations exist in understanding the system's trends and characteristics. An international study used national death statistics on age of death, cause, and contents of forgoing treatment to identify the proportion and characteristics of patients forgoing treatment.[6-8] A South Korean study identi ed the types and related factors of life-sustaining treatment (LST) for geriatric patients, attitudes and characteristics of elderly patients to LST, and whether they wrote POLST. [9][10][11] However, these studies mostly considered data shared by a single or several medical institutions; as no data analysis was conducted using National Institute of Statistics data, there was limited information on the system's characteristics. Therefore, this study aims to analyze the national agency's data for two years after the LSTD system's implementation to identify the LSTD's status and characteristics according to patients' decision-making method.

Design
This retrospective study analyzes data on legal forms prepared in South Korea's LSTD system and the characteristics of patients who implemented LSTD. The purpose is to analyze LSTD records collected and kept by a national agency and how much patient preferences are re ected in the decision-making and factors involved.
According to patients' LSTD forms, four types of decision-making groups exist ( Figure 1) depending on the decision, decision maker, patient's condition, and timing of decision-making.
Group 1 (G1) involves determination by POLST, wherein a terminally ill or end-of-life patient asks the doctor for future medical decisions, and which documents the patient's condition and treatment options, hospice choices, type of forgoing treatment, and LSTD (Articles 10, 17.1).
Group 2 (G2) involves determination by AD, which lists LST preferences previously prepared by patients, and the doctor in charge recon rms them with the patient to complete the patient's LSTD. If the patient does not have su cient competency to verify the AD, the patient's decision can be respected by checking this document with a specialist (Article 17.2).
Group 3 (G3) involves determination to con rm the patient's intention based on a family statement. This occurs when the patient cannot implement the above two decision-making methods and is not capable of decision-making. The presumed intent is recognized by the family's testimony that the patient did not want meaningless LST. The family statement is made by at least two members and must be consistent Group 4 (G4) involves decision-making by every patient's family member. This decision is made if the patient does not have a document or witness for the LSTD. The family then ensures the decision is in the patient's best interest. A proxy decision is made even if the patient is underage, when the legal representative is limited to the patient's parental authority who may express the intention to make a lifelong medical decision (Article 18). G1 and G2 involve patients making decisions based on explicit preference expression, and they are classi ed as "self-determining patients." Conversely, G3 and G4 include "family-determining patients" for whom proxy decisions are made based on family statement or agreement.

Population and data collection
The study subjects were patients registered in the LSTD system from March 2018 to January 2020. Data from the database of national institutions were used, which included the total number of registrations in the six main legal forms of the LSTD process for the presented period-POLST, AD recon rm, witness by the family, consent of the family, end-of-life judgement, and implementation. Duplicate data were removed and as a result of merging based on the author of the implementation form, 84,422 patients were selected for analysis. Among them, 2,179 patients who were only registered with an implementation without end-of-life or decision-making documents (AD, POLST, witness by the family, and consent of the family) were excluded. Moreover, 579 patients without a registered decision-making document and 1,451 patients without end-of-life judgments were excluded. Finally, 80,213 subjects were extracted, who had a written end-of-life judgment, written statement of implementation, and at least one decision-making document. Among them, 79,863 were selected for the nal analysis, excluding 350 with unclear decisionmaking methods because of two or more decision-making documents.

Ethics
In this retrospective study, data analysis was conducted after approval of deliberation exemption from the Institutional Ethics Committee of the researcher's hospital (No: Y-2020-0008).

Data analysis
The patients were divided into four groups according to the decision-making method ( Figure 1) and subjected to descriptive statistical analysis. Moreover, logistic regression analysis was performed by dividing them into self-determining and family-determining patients according to the decision-making method. For analysis, SAS ver. 9.4 was used. The data are described in Table 1. and implementation on the same day. Moreover, 29.1% of patients wrote the POLST rst and then later completed the end-of-life judgment and implementation on the same day.

Characteristics of patients determined by AD (G2)
G2 included patients who prepared AD, rea rmed the contents through the con rmation form, and completed the end-of-life judgment and implementation. They included 60.56% men and 39.55% women. Their average age was 74.13 years, and they were mainly in their 70s (37.23%) and 80s or older (34.50%).

Characteristics of self-decision-making
This study also conducted a regression analysis to determine the characteristics of patients' selfdetermination of withdrawal and withholding of LST. Signi cant differences existed between selfdetermined and family-determined according to age, type of medical institution, region, whether AD was prepared, and refusal of some medical treatment (see Table S2).
Self-determination rate of patients in their 40s was 5.685 times higher than that of patients in their 80s. The self-determination rate for patients with circulatory disorders was 0.246 times lower than that of patients with neoplasms; this rate for patients with gastrointestinal disorders was 0.552 times lower than that of patients with neoplasms.
The rate of self-determination of general hospital inpatients was 0.021 times lower than that of those in tertiary hospitals. Regarding subjects of specialists, the self-determination rate of patients with end-of-life judgments by family medicine specialists was 3.460 times higher than that of patients with judgments by internal medicine specialists.
Moreover, the self-determination rate in Gwang-ju was 0.082 times lower than that in Seoul, and that in Gyeongbuk was 0.084 times lower than that in Seoul. Compared with Seoul, the rate was 1.934 times higher in Ulsan, 2.471 times higher in Gyeong-gi, 1.597 times higher in Chung-buk, and 1.987 times higher in Je-ju.
The self-determination rate of patients who had previously prepared AD was 117.475 times higher than that of those who did not.
This rate for patients who decided to withdraw or withhold a ventilator was 5.300 times higher than that of patients who did not, and the rate for patients who decided to withdraw or withhold hemodialysis was 2.120 times higher than that of patients who did not.

Discussion
This study found differences in the LSTD method according to patients' age and speci c disease when the LSTD was made at an impending time of death for the patient.
Difference in LSTD method according to age As age increases, the family, not the patient, makes the LSTD. Considering that the average age of G1 is 66.86, G3 is 75.24, and G4 is 70.46, [1] age has a major impact on LSTD. Patients in their 80s or older are dominant in G3 and G4, compared with G1, which has the largest proportion of patients in their 60s and 70s, and G2, which has the largest proportion in their 70s.
These results are con rmed by the logistic regression analysis, according to which the self-determination rate of patients in their 40s was 5 times higher than that of patients in their 80s. Thus, elderly patients' families generally make LSTDs as major proxy decision-makers. Family-centered decision-making is common in Asian countries as a characteristic of Asian cultures. [12,13] This is in line with previous research that interviewed doctors in intensive-care units to nd that family-led decisions to withdraw and withhold patient treatment were more prevalent than patients' own decisions. [14] Particularly, for elderly patients, the family did not deliberately inform the patient of the medical condition, fearing patients' lack of understanding and weakening of willingness to ght the disease. [15,16] Families play a major role in determining patient treatment as they have key information about patient preferences and values. However, elderly cancer patients showed the attitude of preparing for and accepting death in physical, psychological, social, and spiritual aspects if they had knowledge of LST. [17,18] This means that even if the patient is older, has less understanding, and has a protective family, it is necessary to seriously reconsider the family's decision to exclude the patient. The LSTD Act was implemented to protect patients' right to self-determination in treatment choices. To realize this, it is fundamental to give patients the opportunity to decide.

Difference in LSTD method according to disease
According to the results of this study, we found that the LSTD method differs depending on the disease. In G1, 86.36% of patients had neoplasms, and G2 patients also showed the highest proportion of neoplasms. Comparing the groups, the proportion of neoplastic diseases decreased in the order of G1-G2-G3-G4, and the proportion of respiratory and circulatory diseases increased in reverse order. This can be explained in the same context that the self-determination rates of patients with circulatory or digestive diseases were somewhat lower than that of those with neoplastic diseases.
Neoplastic diseases' progression is relatively predictable, [19] while in chronic diseases, it is generally di cult to distinguish between terminal and end-of-life stages. For cardiopulmonary diseases, it is generally di cult to determine the treatment direction because the patient repeatedly improves and worsens. These characteristics are con rmed by existing studies. [20,21] In G1, which accounts for a large proportion of neoplastic patients, LSTD was mainly made in the terminal stage. This means that changes in the disease's progress are related to LSTD's timing. However, the more di culty chronic patients experience in their daily lives, the more they hope to receive terminal diagnosis and information from their doctors. [22] Therefore, information and opportunities for LSTD must be given equally to all patients regardless of the type of disease and change in disease status.
Period from decision-making to implementation for LST POLST can be created in the terminal stage, and additional forms are required at the end-of-life and transition periods. However, in only 29.1% of cases was POLST prepared in the terminal stage and then judged and implemented at the end-of-life. More than half (58.2%) of patients who made LSTD by POLST simultaneously received end-of-life judgment and implementation on the same day.
This was similar for patients who wrote AD (G2). The doctors recon rmed the AD content, judged the endof-life, and implemented it in one day (79.7%). Even if the patient's family made a proxy decision (G3 and G4), witnessed by the family or with consent of the family, end-of-life judgment and implementation all occurred in one day. In previous studies, most patients were found to withdraw or withhold treatment just before their death.
[16] In conclusion, the LSTDs in this study seem to have been made in a hurry during the end-of-life period, and the patient was not given enough time to prepare for death.
Knowing the treatment, patients generally wish to reduce the burden on surrogate decision-makers. [23] In particular, advanced care planning, which is a way for patients to identify their preferences in advance and document them, is a major method of end-of-life treatment. Advance care planning is not just an expression of the patient's intention, but a result of active communication between patients and medical staff, and decisions tend to reduce the burden on the patient and agent. [24,25] Summary, contributions, and limitations Based on the research results and discussion, this study argues the following. First, in making LSTDs, patients must be the subjects of their treatment decisions. Particularly, even if patients are old or have poor medical understanding, they should be given the opportunity to make their own decisions. Second, patients' LSTD consultation should begin earlier than it is now. If it is determined that the patient has reached the terminal stage, the medical staff should be able to think about the quality of death the patient will face and discuss it with the patient and family.
This empirical study examined the actual condition and characteristics of the LSTD system by analyzing data of patients who have used the system since its implementation. It is a signi cant research that identi es the characteristics of patients' end-of-life care decision-making process and objectively shows the current state of self-determination. Unlike previous studies that analyzed a single form, this is the rst study to objectively examine the trends of patients' life-sustaining treatment decisions by identifying their characteristics throughout the LSTD process.
However, some data had missing values and this study only con rmed the system's trends and characteristics, but could not identify speci c problems related to the phenomenon. In order to suggest effective improvement plans, further empirical studies will be needed to identify the cause of the signi cant difference in variables.
[1] For G4, whose LSTD was made through patient family agreement, the average age was somewhat lower than that of other groups, likely because 474 underage patients were included.

Conclusion
In South Korea, decisions are made about the discontinuation of treatment for patients at the end-of-life through the LSTD system. Despite the introduction of this system to protect the dignity of the dead, LSTDs for elderly and chronically ill patients are made by their families rather than by self-determination. Moreover, the time for decision-making to stop LST is tight. This does not satisfy the purpose of the law, which con rms the patient's right to self-determination and best interest. Thus, efforts to protect this law are needed.