Data source
We conducted a retrospective cohort study using claim data from the National Health Insurance, Late Elders' Health Insurance, and Long-Term Care Insurance in Tsukuba City, Ibaraki, Japan, which is a suburban medium-sized city with a population of 236,842, including 23,994 aged 65–74 years and 21,506 aged ≥ 75 years [13]. Data sets obtained between April 2014 to March 2019 were analyzed.
Japan’s statutory health insurance system consists of two types of mandatory insurance: employment-based plans and residence-based insurance plans [14]. For individuals aged < 75 years, one of these two types of insurance is selected, based on their employment status. Unemployed individuals, excluding the dependents of employed individuals, are covered by residence-based insurance plans, i.e., the National Health Insurance. The proportion of individuals covered by the National Health Insurance among those who were aged 65–74 years in Tsukuba City was 72.9% in 2016 [15]. Individuals aged ≥ 75 years are all automatically covered by residence-based insurance plans, named the Late Elders' Health Insurance. Although recipients of public assistance are excluded from these medical insurance systems, the proportion of recipients of public assistance among the population aged ≥ 65 years is very small. (In Ibaraki Prefecture, the proportion of recipients of public assistance among those aged ≥ 65 years is estimated at about 0.15% [16–18].) The dataset we used covers a considerable proportion of individuals who were aged ≥ 65 years.
The long-term care insurance system in Japan covers persons aged ≥ 65 years as insured persons who could be supported when requiring long-term care or support [19]. Municipalities, who are the insurers, provide long-term care requirement certification or support requirement certification based on voluntary application. Standards for long-term care requirement certification are uniformly and objectively determined nationwide. After receiving initial certification, it has to be updated every 2 years to obtain continued support. Seven levels of long-term care requirement are defined: support levels 1–2 and care need levels 1–5, with care need level 5 being the most severe condition requiring the longest care. The level of long-term care required has been correlated with the Barthel Index, with median Barthel Index scores (interquartile range) by care level (CL) being as follows: CL3, 60 (40–75); CL4, 30 (20–40); and CL5, 20 (5–35) [20].
The study protocol was approved by the Ethics Committee, Faculty of Medicine, University of Tsukuba (approval number: 1445-13, date of approval: 21 September 2023). Data were anonymized prior to being provided to the researchers; thus, the need for obtaining informed consent was waived.
Participants
Patients aged ≥ 65 years who underwent IMV were included. Patients with a claim of medical reimbursement classification code “lifesaving endotracheal intubation” (procedure code J044) and “artificial ventilation” (J045) on the same day or day after intubation was defined as patients who received IMV. “Lifesaving endotracheal intubation” (J044) does not include intubation associated with surgical anesthesia. “Artificial ventilation” (J045) includes not only IMV, but also non-invasive MV in patients with acute respiratory failure with PaO2 / FiO2 ≤ 300 mmHg or PaCO2 ≥ 45 mmHg, as well as intubation associated with surgical anesthesia. Using the combination of these two procedure codes, only patients with IMV intubated outside the operating room could be included. Patients, since the decision-making regarding the initiation of IMV associated with surgical anesthesia differs significantly from decision-making in patients requiring emergency endotracheal intubation were intended to include as objective.
In patients with multiple intubation records, only the data of the first episode was used. Patients who experienced tracheotomy or home MV before the first intubation record, and who had been in the hospital for less than 180 days after intubation at the study cutoff date, March 31, 2019 were all excluded. Patients who died within 3 days after intubation were also excluded, because early deaths are mostly due to the progression of the primary diagnosis with few impacts of complications from intensive care, and at least 3 days are necessary to properly observe the trajectory after MV [21].
Extracted data items
Data were extracted regarding the age, sex, main diagnosis recorded on receipt (multiple diagnoses could exist), past medical history, the level of long-term care required, the duration of MV, tracheotomy followed by MV management after intubation, the days of tracheotomy from intubation, and liberation and discharge status on day 180. The main diagnosis recorded on receipt was categorized using the ICD-10 code, based on previous research [22–25]. Patients who underwent open and closed chest cardiac massage were classified as having a diagnosis of cardiopulmonary arrest. The procedure codes and ICD-10 codes we used to determine each procedure and diagnosis are summarized in Appendix 1. Age was categorized into three categories: 65–74 years, 75–84 years, and ≥ 85 years. The level of long-term care required was extracted from the latest long-term care requirement certification records before intubation, which is updated every 2 years from the initial certification.
The level of long-term care required was divided into two categories based on a previous study, which classified CL based on the degree of bedriddenness and the level of independence of older individuals with dementia [26]. They revealed that among older individuals requiring CL ≥ 3, more than half were bedridden, and more individuals exhibited a functional decline in both level of independence among patients with dementia and degree of bedriddenness compared to other patients. Therefore, patients classified CL < 3 (including no certification) was categorized as mild to moderate, and CL ≥ 3 as severe.
Liberation was defined as liberation from MV without any implication about extubation. If there was an interruption in the MV record of ≤ 3 days during MV management, we considered it as a continuous MV period.
Analyses
A descriptive analysis of the liberation and discharge status was conducted. Data were analyzed using descriptive statistics, including frequency counts, percentages, median, and graphical representations. Liberation and discharge status were summarized as changes over time until day 180. If patients achieved liberation even once, they were treated as liberated patients, even if they underwent MV again 4 or more days after the end of the initial MV.
The stratified results of liberation and discharge status by age categories and CL ≥ 3 or < 3 are also presented. We conducted chi-square test or Fisher’s exact test to assess whether liberation and discharge status differed among age categories or CLs.