This study examined long-term tracheostomy decannulation rates across facilities in Japan using a large health insurance claims database, which included data after discharge from acute care hospitals. Decannulation rates were 40.8% at 3 months, 63.9% at 12 months, and 65.0% at 24 months. Factors associated with prolonged tracheostomy included older age, female sex, cerebrovascular disease, injuries to the head, and other forms of heart disease.
The decannulation rate increased rapidly up to 3 months after tracheostomy, reaching a plateau after 12 months. Decannulation is performed in acute care hospitals and rarely in long-term care hospitals or after being discharged home. A previous study conducted in an acute care hospital in Japan reported a decannulation rate of 31% at 3 months22; another study reported a rate of 59% in a rehabilitation hospital20. Our results are consistent with these reports. The average length of hospital stay in Japan is 30.6 days25, which is the longest among OECD countries26. Although the length of acute care hospital stay is fixed to up to 90 days, critically ill patients tend to remain hospitalized for a much longer period. Many of these patients are transferred to a rehabilitation hospital or long-term care hospital. In the present study, one-third of patients were transferred to a rehabilitation hospital or long-term care hospital. The problem is that, once patients are discharged from acute care hospitals, they may lose the opportunity to have their readiness for decannulation evaluated due to the lack of access to specialists and equipment. Rehabilitation/long-term care hospitals provide rehabilitation or daily care services but not specialized medical care, such as decannulation. In other words, the last opportunity for patients to have their condition and readiness for decannulation evaluated, and then undergo decannulation, is when they are still hospitalized in an acute care hospital.
The time to decannulation in Japan is relatively longer compared to other countries. Previous studies conducted in Canada, the United States, Italy, Spain, and Australia have reported shorter decannulation times and higher decannulation rates, with few adverse events16,17,18,19,27,28. These studies demonstrated that the time to decannulation was shortened by having specialized teams follow tracheostomy patients and developing protocols for decannulation. Patients who survive the initial acute condition but still require other intensive care are said to be chronically critically ill (CCI)29; the recent increase in the number of CCI patients has become a serious problem29,30,31,32. CCI patients are those in devastating situations, and include those with poor long-term survival, severe physical and cognitive disabilities, and significant medical costs29,30,31,32,33. A previous study on CCI found that tracheostomy, one of the clinical conditions of CCI, accounts for more than 20% of all CCI patients in Japan32. Intensivists need to be aware of long-term outcomes such as CCI burden32. Evaluating patient readiness for decannulation prior to discharge from acute care hospitals could reduce the number of CCI patients and thereby reduce the clinical and economic burden, improve patient QOL, and contribute to the establishment of a decannulation protocol.
Older age and female sex were identified as factors associated with prolonged tracheostomy. Population-based studies conducted in the United States have reported that elderly people undergo tracheostomy more often than younger people3,4. According to a Japanese national database, people aged ≥ 65 years account for more than 70% of all tracheostomies performed annually in Japan6 (see Supplementary Figure S2). Elderly people tend to have multiple diseases and are susceptible to worsening conditions, and thus, it is more likely that tracheostomy tubes are kept in place longer. Although the present study did not include patients aged ≥ 75 years, if such patients were included, we would expect a much lower decannulation rate in Japan. While the reason why more females have prolonged tracheostomy compared to males is unclear, there may exist gender disparities in treatment decisions. However, more information will be needed to verify this. Tracheostomy can reduce the QOL of patients as well as their families. When making decisions about tracheostomy, it is important that physicians understand the factors associated with prolonged tracheostomy. Further investigation of decannulation procedures is warranted to establish appropriate protocols34.
Cerebrovascular disease and injuries to the head were also factors associated with prolonged tracheostomy. Tracheostomy is performed in patients with a wide range of diseases, with cerebrovascular disease accounting for more than 20% in the present study. Other diseases accounted for less than 10%. Previous studies have reported that surgical patients are more likely to have early tracheostomy4,11, and that patients with severe brain injury may be more likely to undergo tracheostomy, with a tendency for prolonged tracheostomy. Rehabilitation is important for such patients in terms of achieving improved mobility. Readiness for decannulation should be evaluated at the acute care hospital prior to the transfer to a rehabilitation hospital.
Many patients with other forms of heart disease had cardiac arrest, and the decannulation rate among these patients was extremely low due to high mortality. This finding underscores the importance of end-of-life care in critical care settings. While advances in intensive care have saved the lives of critically ill patients, they have also created a large number of CCI patients30. Healthcare providers in critical care frequently encounter difficult ethical decisions about whether to save lives or deliver end-of-life care35,36,37. In the United States, treatment withdrawal is common among patients with severe brain injury due to stroke, trauma, or cardiac arrest, and those patients do not undergo tracheostomy36. End-of-life guidelines in critical care in Japan recommend discontinuation of life-sustaining treatment for terminally ill patients. However, it is difficult to predict patient outcomes immediately. At present, life-sustaining treatment may be provided without the need to consider these ethically difficult decisions in Japan. Treatment decisions for critically ill patients should be based on patient-centered thinking and an understanding of their values and preferences, while also avoiding overuse or underuse of services29,35,36,37. To achieve optimal end-of-life care in critical care settings, it will be important to have in place a system that makes available communication opportunities with the families of patients35,36,37. Particularly in Japan, where the super-aging of society is progressing, considering end-of-life care in critical care settings is all the more important.
There are several limitations to this study. First, the claims database did not include important factors known to affect decannulation, such as patient socio-economic status, family information, hospital characteristics, and hospital location. Second, we could not obtain certain clinical information, such as the physical condition of patients and their disease severity. Although we identified diseases from the claims data related to the procedure “tracheostomy,” this method has not been validated, and the identified diseases may differ from the actual diseases that led to tracheostomy. We did, however, identify diseases and medical procedures in consultation with researchers with experience in claims database analysis, as well as an infectious disease physician. Third, some data were censored due to patient death or transfer to rehabilitation/long-term care hospitals, from which they were not discharged until after the study period. This may have resulted in an overestimation of decannulation rates. Finally, the database is limited to company employees and their families, and people aged ≥ 75 years were not included. Therefore, our results may not be widely generalizable.