Our analysis of the NIS data from 2016 to 2020 indicated that weekend admissions for myxedema coma were associated with higher mortality rates when compared to weekday admissions. However, weekend admissions were not related to a statistically significant difference in hospital LOS, resource utilization, AKI requiring dialysis, ARF requiring intubation, circulatory shock requiring vasopressors, use of TPN and palliative care use. Although overall and weekday mortality rates showed an upward trend from 2016, there was no significant difference in mortality trend for weekend admissions.
In our study, we found that patients hospitalized with myxedema coma had a mortality rate of 11.60%. However, another retrospective study utilizing the national inpatient database from Japan between July 2010 and March 2013 showed a considerably higher inpatient mortality rate for myxedema coma at 29.5%. The mean age calculated in that study was 77 years, surpassing that of our study population, which may explain the cause for the increased mortality rate. Nevertheless, we observed a similar demographic distribution to this study by Ono et al, where two-thirds of the patients were female [5].
A previous study conducted from 1999 to 2006 at a single center in South Asia, involving 23 patients, reported a mortality rate of 53%. A higher mortality rate may have been observed due to limited access to diagnostic and treatment facilities in developing countries. Comparatively, our study has a lower probability of a beta error due to a larger size as it was conducted using a multicenter, and nationally representative database (NIS) [11].
Our study revealed a 91% increase in the odds of mortality for weekend admissions as compared to admissions done during the weekday. These findings are consistent with those of other studies investigating the effects of weekend admissions on various other medical conditions like stroke [12], acute myocardial infarction [13], and pulmonary embolism [14].
Several factors may contribute to the increase in mortality observed during weekend admissions including lower staffing [15] and limited resource levels. The quality of care may also vary depending on the day of the week as different providers may have different levels of expertise. Additionally, delays in seeking medical attention on weekends or differences in the severity of the disease among patients admitted on weekends may also contribute to elevated mortality rates [16].
We also examined the mortality trends for myxedema coma, which revealed an upward trend from 2016 to 2020. This increase in mortality trends is concerning, but due to the rarity of the condition, there are insufficient large-scale studies available to compare mortality trends over the years. In 2020, the overall mortality rate spiked to 13.36%, which coincided with the onset of the COVID-19 pandemic. This increase in mortality during 2020 may be related to some patients with a concurrent COVID-19 infection, however evaluating this impact was outside the scope of our study. Nevertheless, many studies have demonstrated an increase in mortality for non-COVID-19 related hospitalizations which occurred during the pandemic [17, 18].
According to a recent study [17] that examined the total number of admissions from the COVID-19 research database [19] from January 2019 to December 2020, the odds of in-hospital death among non-COVID-19 patients during the pandemic were 1.2 times higher than before the pandemic. This increased risk of death may be attributed to disruptions in healthcare continuity, inpatient services and resource availability resulting from the pandemic.
Our research is subject to certain limitations. To begin with, it is a retrospective study that utilized administrative and claims-based datasets, both of which are prone to misclassification, missing codes, and inaccurate coding. These factors may have an impact on the precision and reliability of our results. Second, the retrospective nature of the study meant that we could not fully randomize the exposure. To address this issue, we used multivariate regression models that considered multiple patient and hospital-level characteristics as well as co-morbidities to adjust for any potential confounders. Despite these adjustments, a chance of residual confounding remains, although the likelihood of this is low.
We were unable to categorize the severity of myxedema coma because the HCUP data does not include information on laboratory values. Nevertheless, we utilized the Charlson comorbidity index, a well-established and validated prognostic tool, to adjust for the comorbidity burden. Lastly, we analyzed the all-cause inpatient mortality rate for patients admitted with a primary diagnosis of myxedema coma, as the database did not contain enough data to determine the exact cause of death. In our opinion, additional randomized trials are essential to address the limitations encountered in this study.
Despite the limitations outlined earlier, our study has several strengths. To the best of our knowledge, it is the first nationwide investigation that examines the impact of the day of admission on hospitalization outcomes for patients with myxedema coma.
The NIS database overcomes a common limitation of single-center studies by offering a large sample size which is made possible by the fact that the NIS is the largest publicly available all-payer database that contains data on the inpatient population at a national level. This enhances the study’s statistical power, which, in turn, decreases the risk of type II errors in the analysis. Also, the characteristic variables found in the NIS database provided an opportunity to investigate variables that are not usually available in single-center studies, such as household income estimates, hospitalization costs, and hospital-related factors. Our study’s results are representative of the patient population admitted to hospitals across the country, making them highly generalizable and externally valid. The use of NIS helps to eliminate bias associated with practice patterns that arise in single or multi-center studies.