key: cord-0988427-p5om2qnl authors: Santos, Ricardo S.; Barros, Danilo S.; Moraes, Thiago M.P.; Hayashi, Cintya Y.; Ralio, Renata B.; Minenelli, Fernanda F.; van Zon, Kees; Ripardo, João P.S. title: Clinical characteristics and outcomes of hospitalized COVID-19 patients in a Brazilian Hospital - A retrospective study comprising first and second waves date: 2022-04-14 journal: IJID Regions DOI: 10.1016/j.ijregi.2022.04.002 sha: 0f78dc21fee614a664a8dd9f88a01946de8b3943 doc_id: 988427 cord_uid: p5om2qnl Objectives : Describe clinical characteristics, hospitalizations flow and outcomes in a cohort of COVID-19 patients in a Brazilian hospital, comprising first and second waves. Methods : Retrospective and observational study, including patients with a confirmed SARS-CoV-2 infection who were evaluated in the ED between March 1, 2020, and June 30, 2021. We used descriptive statistics to report the clinical characteristics, admissions, and outcomes. The comparison between the two waves was inferred through Hypothesis Test techniques. Results : During the period, 7,723 (86.54%) were evaluated in the ED, of which, 1,908 (24.70%) were admitted. Of these, 476 (24.95%) were initially allocated to the ICU and 1,432 (75.05%) to the ward. Of those allocated to the ward, 349 (24.37%) were later transferred to the ICU. 158 patients were intubated (19.15% of ICU admissions) and 110 died (5.77% of admissions). The admission rates decreased in the second wave: in the ICU from 13.84% to 9.56% (p-value < 0.01), and in the ward from 22.41% to 17.16% (p-value < 0.01). The average age decreases in the second wave from 44.06 to 41.87 years old (p-value < 0.01). Patients with severe symptoms, such as dyspnea, decrease from 25.51% to 13.13% (p-value < 0.01). Death rate by admissions saw a reduction of 17.84%; from 6.52 to 5.38 (p-value < 0.01). Conclusion : Results show a higher volume of patients in the second wave, but lower admission and death rates. The mean age of patients in the second wave was lower, and patients arrived at the hospital with less severe symptoms than those in the first wave. The spread of COVID-19 had, up 2021-Jul-30 , affected more than 197 million people in the world and caused at least four million deaths (JHU CSSE COVID-19 Dashboard). This disease is an infection caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) and can vary in severity from asymptomatic to critical. Although the lethality of COVID-19 is around five percent, 15-18% of patients can become severely ill, requiring mechanical ventilation and intensive care unit (ICU) admission (Kummar et al., 2020) . The initial lack of knowledge the disease's behavior raised a concern on how to avoid the collapse of hospital resources. One of the main difficulties was related to patients' characteristics, their clinical evolution and resources consumed during their treatment. This type of comprehension is essential to the planning and management of health resources necessary to fight the disease. For this reason, studies in the past year focused on describing the clinical characteristics and outcomes (Anesi et all., 2020; Grasselli et al., 2020; Holler et al., 2021; Huang et al., 2020; Larsson et al., 2020; Ludwig et al., 2021; Murthy et al., 2021; Olumade et al., 2021; Pouw et al., 2021; Richardson et al., 2020; Sulejmani et al., 2021; Wang et al., 2020; Wen et al., 2021; Zhu et al., 2020) , or developing prediction models (Baas et al., 2021; Kang et al., 2020; López-Cheda et al., 2021; Mahboub et al., 2021; Velasco-Rodríguez et al., 2021; Wang et al., 2021) . ]. However, there is still a need for studies describing differences between the waves of this disease. In Brazil, the first wave increased until reaching its peak around the end of May 2020, then declined from September until the start of the second wave early November 2020. This second wave peaked around the end of March 2021 (da Silva & Pena, 2021). According to Bastos et al. (2021) , the comparison between the first and second wave in Brazil shows an increase in admissions by 59%, with a relative increase of 18% in the proportion of patients younger than 60 years. Despite recent advances in vaccination, new waves of the disease cannot be ruled out due to frequently arising new variants of the virus. Thus, a study that describes the characteristics of patients, hospitalizations, and outcomes, covering more than one wave of the disease can be quite relevant, even after almost two years of pandemic. In this context, the objective of this paper is to describe clinical characteristics, hospitalizations flow and outcomes in a cohort of hospitalized COVID-19 patients in a Brazilian hospital, comprising Brazil's first and second waves. The study was done in a medium size tertiary private hospital in São Paulo, Brazil that comprises 300 beds in the general ward and has 50 ICU beds. The hospital assists adult and pediatric patients, and there is no the need of referral from a primary care unity. Since the beginning of the pandemic, the hospital has defined an overall workflow for COVID-19 patients (Figure 1 ), which has not undergone major changes during the pandemic. Some operational and structural adjustments were made, but the workflow activities were always maintained. Workflow begins in the emergency department (ED) where patients go to registration and nursing triage, after which they proceed to medical evaluation. In this assessment, the doctor decides whether to request the RT-PCR test, as well as prescribe other tests and medications. According to clinical status and results of tests, the doctor decides whether the patient should be released to home or admitted to the ward or the ICU. The clinical condition of patients initially admitted to the ward can worsen which may lead to transfer to the ICU. In the ICU, patients receive respiratory support according to their severity: non-invasive ventilation (NIV) through a high-flow nasal catheter, or mechanical ventilation through endotracheal intubation (ETI). Patients who are discharged from the ICU are transferred to the ward, where they remain until full recovery before being discharged home. The workflow also includes hospital readmissions, which, in the context of this study, correspond to patients who are released to home at their first medical evaluation and, after a few days, return to the hospital. This retrospective observational study includes all patients with a confirmed SARS-CoV-2 infection (by RT-PCR test) who were evaluated in the ED between March 1, 2020, and June 30, 2021, and those who were discharged from the hospital at the time of database closure. All analyzed data were from the electronic health record (EHR) where, as part of clinical routine, they were collected from arrival at the ED to discharge or death. We developed database stored procedures for automatic data collecting from the EHR, including clinical characteristics, demographics, comorbidities, patient flow tracking and procedures executed during hospitalization. We used descriptive statistics to report the clinical characteristics, admissions, and outcomes. Discrete variables are expressed as frequency (percentage) and continuous variables as mean with standard deviation (SD) for normally distributed data. The comparison between the two waves were inferred through Hypothesis Test techniques, with the first wave covering March-2020 to October-2020 and the second one November-2020 to June-2021. We opted for the two-tail Z-test for continuous variables and the two-tail T-test for discrete ones. The null hypothesis takes the format: H 0 : x (p1) = x (p2) ; where p 1 is the parameter evaluated for the first wave and p 2 is the same parameter evaluated for the second. We adopted a significance level of 5%. Analyses of medications, procedures, or therapeutic approaches were outside the scope of this study; our focus was on analyzing the differences between the first and second waves of the pandemic, considering the characteristics of patients, hospitalizations, and outcomes. Basically, we were interested in analyzing whether these characteristics were maintained in the two waves. This study was approved by the local ethics committee (code: 4.718.282). Figure 1 shows the general statistics for COVID-19 patient flow. Between March-2020 and June-2021, the hospital detected 8924 COVID-19 cases of which 7723 (86.54%) patients were evaluated by a physician in the ED. The remaining 13.46% only went to the hospital to do a PCR test. From the patients evaluated by a physician in the ED, 6302 (81.60%) were released to home on the same day; however, 528 (8.38%) returned to the hospital several days after that discharge. The average time for this readmission was 5.76 days. It was not possible to track patients who were readmitted to another hospital; however, we believe that the probability of this happening is small, since there is no other referral hospital in the neighborhood. From the evaluated patients, 1,908 (24.70%) were admitted, of which, 476 (24.95%) were initially allocated to the ICU and 1,432 (75.05%) to the ward. Of those initially allocated to the ward, 349 patients (24.37%) had a deterioration in their clinical condition and were later transferred to the ICU. Of the 825 patients who went through the ICU, 158 (19.15%) were intubated. Table 1 shows patient characteristics, such as demographics, symptoms, comorbidities, as well as information regarding admissions and lethality, stratified according to the first and the second wave. The main results are presented in the sections below. In general, there were no major changes in the patient characteristics from the first to the second wave. We observed a slight reduction in the average age of the infected patients, from 44.06 years old in the first wave to 41.87 in the second. This reduction is also perceived for different age groups, through a slight increase in the percentage of people between 40 and 60 years old (from 41.48% to 43.98%) and, especially, for a reduction in the percentage of people over 60 years old (from 17.49% to 12.56%). Regarding the most frequent symptoms, we noticed that patients from the first wave arrived at the hospital more symptomatic. The major change was for dyspnea, which had a significant decrease in the second wave. In the first wave, 25.51% of the patients presented this symptom in their first assessment, while in the second wave only 13.13% did. The only two symptoms that increased their occurrence in the second wave were sore throat, from 35.28% to 42.81%, and headache from 55.54% to 60.11%. Although in the second wave we had an increase in the percentage of admitted patients without comorbidities (18.05% to 21.15%), its p-value does not allow us to consider it as statistically significant. In addition, we also saw a slight decrease in the average number of comorbidities per inpatient (3.10 to 2.63). The list of main comorbidities, as well as the percentage of patients for each one of them, remained practically unchanged for the two waves. In the second wave, we had a significant reduction in admission rates. ICU admission rates decreased from 13.84% to 9.56%, and the admission rates in the ward reduced from 22.41% to 17.16%. Figure 2 shows the monthly evolution of admission rates, where we can see the peaks of admissions for the two waves, respectively, in May 2020 (118 admissions) and March 2021 (335). However, we observed that admission rates in the first wave were much higher than in the second one, especially in the first month of the pandemic when 65.71% of patients who visited the ED were admitted. The mean length of stay (LOS) in the ICU showed a slight decrease in the second wave (from 13.47 to 12.12 days). However, this reduction is not statistically significant (p-value = 0.3), and we can observe a very high standard deviation (greater than the mean). In addition, we noticed numerous cases with very high LOS, considered as outliers (Figure 3 ). Considering the values provided by Boxplot graph (Figure 3 ) and compiled it in Table 2, we can see that 50.30% of ICU admissions has a LOS <= 7 days, however, 7.27% of ICU admissions exceed 37 days. The maximum ICU LOS was 155 days in the first wave and 102 in the second. Regarding ward admissions, we also observed a reduction in the LOS (from 7.6 to 5.8 days), however, unlike the ICU, this decrease is statistically significant (p-value < 0.01). Another point to note is that the standard deviation for the ward' LOS in the first wave was much higher than in the second one: 15.7 and 7.8, respectively. Regarding the mean age of patients admitted to the ICU, we observed a reduction of almost 6 years in the second wave (from 62.23 to 56.26). Interestingly, in the ward admissions we had a slight increase (from 52.56 to 53.56 years old). This increase was not characterized as statistically significant (p-value = 0.36). Like hospitalization rate, we can also observe a statistically significant reduction (p-value < 0.01) in the death rate in the second wave. If we consider the death rate by admissions, we had a reduction of 17.84% (6.52 to 5.38). Considering the death rate by infected cases (patients evaluated at the ED) we had an even greater reduction, namely 42.51% (2.07 to 1.19). Figure 4 shows the monthly evolution of the death rate, computed from the admission date. That is, of the 69 patients admitted in March 2020, 11 had died (in March or in subsequent months). We can also observe a similarity between the graphs of Figures 4(b) and 2(b) ; the highest death rates coincide with the highest admission rates, particularly in the peak months and in the first month of the pandemic. Like ICU admissions, we observed a reduction of approximately 6 years in the mean age of patients who died in the second wave (from 77.71 to 71.31 years old). Analyzing deaths in different age groups, we noticed an expressive increase in the second wave of deaths of people between 40 and 60 years old (from 4.76% to 23.53%), as well as a significant reduction in deaths of people over 60 years old (from 92.86% to 75.00%). The results presented allow us to evidence four main differences between the first and second waves: 1) Reduction in the admissions rates, despite the expressive increase in the number of infections. 2) Reduction in the death rate. 3) Reduction in the mean age of the patient (both in terms of infections, admissions, and deaths). Reduction in the admissions rates, despite the expressive increase in the number of cases is a phenomenon that has also been reported in other national (Bastos et al., 2021) and international (Wolfisberg et al., 2021) studies. In the context of this study, the last two items contribute, albeit partially, to the first two occurrences. Since the second wave affected younger patients, with fewer comorbidities, and those patients arrived at the hospital with less severe symptoms, the need for hospitalization tended to be lower. However, there are other factors that may have influenced this reduction. One of them, and perhaps the most important, is the assertiveness in the criteria for patient's admission. Basically, this was due to improved knowledge of the management of this disease, given that in the first wave, especially in the first months, the disease was totally unknown. This fact is also reported by Jain et al. (2021) . Another factor that also may have contributed to the reduction in the admission rates in the second wave is the underreporting of positive cases in the first wave. Due to the scarcity of resources at the beginning of the pandemic, especially of RT-PCR tests, which were intended only for patients with severe symptoms, the number of confirmed diagnoses may have been underreported, and, consequently, the calculation of the admission rate has been affected. Da Silva & Pena (2021) also quote this underreported diagnosis, thus corroborating our statement. Finally, the recommendation of the health authorities at the beginning of the pandemic that patients should go to the hospital only if they had severe symptoms, to avoid the collapse of the health system, meant that most patients arriving at the hospital needed to be hospitalized. This recommendation of the health authorities may also partially explain item four, especially the considerable reduction in the percentage of patients with dyspnea at the first assessment in the second wave. Another factor that contributed to the reduction in the percentage of patients with severe symptoms in their first assessment was the expansion of the health system. Many hospitals in the country, including the one where this study was done, considerably expanded their care resources since the pandemic began. Thus, in the second wave, more patients went to a hospital at the first onset of symptoms, which may also explain the increase in the "sore throat" symptom seen in the second wave. Patients with less intense symptoms at their first clinical evaluation in the second wave were also observed in other studies (Jain et al., 2021; Soriano et al., 2021) . The reduction in the mean age of hospitalized patients was also observed in other national studies. Bastos et al. (2021) found numbers very close to those we measured in our study. They observed a reduction in the average age of patients from 63 to 59 years old, while we measured a reduction from 62 to 56. We believe that this reduction is basically due to two factors. The first is the characteristic of the variant prevalent in the second wave in Brazil, the E484K mutation, which was much more transmissible, thus affecting younger people more than in the first wave (Wise, 2021) . Second, and probably the most relevant, is vaccination, which started in February 2021 for elderly people. People under 60 years of age only began to get vaccinated in June 2021. Thus, at the peak of the second wave, a large percentage of the elderly population was already vaccinated, while younger people were not. Regarding the reduction in the mortality rate in the second wave, we believe that the reason for that is the combination of all the factors mentioned above. Younger patients infected tended to have fewer comorbidities; the clinical team had more knowledge of the disease and was better able to manage it; hospital expanded their infrastructure, and patients went to a hospital earlier, as soon as symptoms appeared. Probably, the decrease in the age of patients affected in the second wave was the most relevant factor for the reduction in the mortality rate, as age is the greatest predictor of survival in COVID-19. Unfortunately, not all Brazilian hospitals experienced this mortality rate reduction in the second wave. On the contrary, in the national context, the second wave was more lethal than the first. Bastos et al. (2021) , who worked with data at the national level, report an increase of 18.47% in in-hospital mortality, although they recommend analyzing this number with caution. Some factors may explain the lower mortality rate measured in our study compared to national rates. First, the national rates measured by Bastos et al. (2021) also included the public hospitals, which had higher mortality rates than private ones. Da Silva & Pena (2021) report the collapse of the public health system during the second wave and also presented a higher mortality rate than we measured in our study. Second, the hospital where this study was done is in an upper-middle-class neighborhood and, consequently, the population served by it tends to have a better general state of health than the population from poor neighborhoods. Another reason is that the hospital considerably expanded their care resources since the pandemic began. In this way, it was not as pressured as hospitals that were unable to expand their resources at the same speed and quality. This study provides insights into characteristics, hospitalizations flow and outcomes in a cohort of hospitalized COVID-19 patients in a medium size tertiary private Brazilian hospital, comprising first and second waves. Our results suggest that the hospital resources were more pressured in the second wave due to a higher volume of patients, however, admission and death rates were lower than in the first wave. The mean age of patients in the second wave was lower than in the first wave, and these patients arrived at the hospital with less severe symptoms than those patients in the first wave. However, the average length of stay in the ICU did not change. We believe that this study has produced valuable information that can help managers to plan the bed logistics and hospital resources needed for the treatment of COVID-19. Mortality rates refer to patients who died in hospital; we cannot track patients after hospital discharge. Failures (absence) in filling in the patients' initial symptoms and comorbidities can have occurred since some professionals recorded them as free text. We did not analyze the different clinical protocols adopted during the pandemic. It is possible that in the second wave better protocols were adopted than in the first wave. The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. This study was approved by Internal Ethics Committee from Hospital Samaritano. This research did not receive any specific grant from funding agencies. It was developed by Philips researchers in collaboration with clinical staff from Hospital Samaritano, covered by a research agreement between the two institutions. Covid-19 related hospital admissions in the United States: Needs and outcomes Real-time forecasting of COVID-19 bed occupancy in wards and Intensive Care Units COVID-19 hospital admissions: Brazil's first and second waves compared. The Lancet Respiratory 2021;2600 Collapse of the public health system and the emergence of new variants during the second wave of the COVID-19 pandemic in Brazil Baseline Characteristics and Outcomes of 1591 Patients Infected with SARS Admitted to ICUs of the Lombardy Region, Italy First wave of COVID-19 hospital admissions in Denmark: a Nationwide populationbased cohort study Clinical features of patients infected with 2019 novel coronavirus in Wuhan Differences between First wave and Second wave of COVID-19 in India Jiming-yang M. Machine learning predictive model for severe COVID-19. Infection Characteristics and outcomes of patients with COVID-19 admitted to ICU in a tertiary hospital in Estimating lengths-of-stay of hospitalized COVID-19 patients using a non-parametric model: A case study in Galicia (Spain) Clinical outcomes and characteristics of patients hospitalized for Influenza or COVID-19 in Germany Prediction of COVID-19 Hospital Length of Stay and Risk of Death Using Artificial Intelligence-Based Modeling Characteristics and outcomes of patients with COVID-19 admitted to hospital and intensive care in the first phase of the pandemic in Canada: a national cohort study Clinical characteristics of 4499 COVID-19 patients in Africa: A meta-analysis Clinical characteristics and outcomes of 952 hospitalized COVID-19 patients in the Netherlands: A retrospective cohort study Presenting Characteristics, Comorbidities, and Outcomes among 5700 Patients Hospitalized with COVID-19 in the Main differences between the first and second waves of COVID-19 in Madrid Baseline characteristics of COVID-19 Italian patients admitted to Desio Hospital, Lombardy: a retrospective study Development and validation of a predictive model of in-hospital mortality in COVID-19 patients Study on the prognosis predictive model of COVID-19 patients based on CT radiomics Hospitalized Patients with 2019 Novel Coronavirus-Infected Pneumonia in Wuhan, China Clinical Characteristics and Outcomes of COVID-19 in West Virginia Covid-19: The E484K mutation and the risks it poses Comparison of characteristics, predictors and outcomes between the first and second COVID-19 waves in a tertiary care centre in Switzerland: an observational analysis The authors would like to thank Philips Research North America and Hospital Samaritano Higienópolis who made this research possible.