Chapter 5 112 comparing COVID-19 patients to the patients with respiratory infections, we believe other factors during the pandemic than type of infection alone play a role in this increase. However, early reports of the risk during a COVID-19 infection on severe symptoms necessitating long intensive care admissions10 11 might have led to more restrained physicians in COVID-19 infections. Other possible explanations are increased awareness in patients and physicians to the harms of intubation and ICU admission along with raised attention to ICU shortages.7–9 Our study was not designed to differentiate between these explanations. One of the major strengths of this study is the unique comparison between code status documentation of patients admitted with COVID-19 and patients before the COVID-19 pandemic. To our knowledge, only Coleman et al analysed this before.7 Another strength is the few missing values (all <1% except for the MEWS scores, in which it was 12%), improving the accuracy and reliability of our results. There are some limitations to our study, the primary being that we cannot distillate what caused the differences we found: the type of infection (COVID-19), factors associated with being in a worldwide pandemic (shortage of care, awareness in physicians, awareness in patients) or differences in the patients. We chose to use two existing databases, to be able to have results as early as possible to guide practice in the developing pandemic. Our goal was to describe code status documentation during COVID-19, rather than calculate an effect size. Because we compared two existing cohorts that were essentially different, we used descriptive statistics instead of performing statistical tests for significance. Another potential limitation is that we could not assess the quality of the code status. In our opinion, discussing the code status with the patient is of utmost importance for its quality; this was done equally in the cohorts. Code status in COVID-19 patients contained more often limitations, what could suggest code status is considered more thoughtful (one could say it is easier to check the box ‘full code’ than a treatment limitation). However, measuring the actual quality of the code status (discussion) is difficult and was not possible with our data. Next to this, we did not know if patients had former documented code status before admission, which could influence code status documentation.29 However, this effect applied to both cohorts and we regarded an important difference in predocumented code status between both periods unlikely.
RkJQdWJsaXNoZXIy MTk4NDMw