Clinical characteristics, outcomes, and costs of COVID-19 patients in Thai hospitels: a comparative analysis based on chest X-ray findings

Abstract

Background

“Hospitels” are hotels that have been specially converted to healthcare facilities. Their utilization emerged as a resource-optimization strategy during the peak of the COVID-19 pandemic in Thailand. This study evaluated the clinical characteristics, outcomes, and admission costs of asymptomatic and mild-to-moderate COVID-19 patients treated in these facilities.

Methods

We conducted a retrospective analysis of adult COVID-19 patients who were admitted between July and October 2021 to two hospitels in Bangkok that were affiliated with Siriraj Hospital. Patient demographics, clinical data, vaccination status, treatment modalities, outcomes, and total costs were collected and analyzed via univariable and multivariable statistical methods.

Results

Of the 1729 patients, 644 (37.2%) presented with abnormal baseline chest X-rays that could imply to moderate cases. These patients were older (49.2 vs. 42.2 years, P < 0.001), had greater body weights (68.1 vs. 64.7 kg, P < 0.001) and body mass indices (26.3 vs. 24.9 kg/m², P < 0.001), and more frequently presented comorbidities. They demonstrated an increased risk for severe COVID-19 (P < 0.001), lower full vaccination rates (P < 0.001), and higher hospital referral rates (P < 0.001). Symptomatology was largely comparable between patients with normal and abnormal baseline chest X-rays, with significant differences noted only for cough and fatigue. While the admission costs of both groups were similar, the total costs for referred patients were sixfold greater than those for discharged patients (70 545 vs. 12 010 Thai Baht, P < 0.001).

Conclusions

Compared with patients with normal baseline chest X-rays, COVID-19 patients with abnormal chest X-rays admitted to hospitels presented higher comorbidity rates, greater severe disease risk, and increased hospital referral rates. Referrals from the hospitels back to Siriraj Hospital substantially escalated the total costs. Vigilant monitoring of patients with initial abnormal chest X-rays in hospitels is crucial for efficient management and resource allocation.

Introduction

COVID-19, caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), has triggered a global pandemic affecting the respiratory system, with symptoms ranging from asymptomatic to severe pneumonia. Patients who progress to respiratory failure and require mechanical ventilation face significant morbidity and mortality risks [1,2,3]. The rapid surge in COVID-19 cases has strained healthcare systems worldwide, including Thailand, causing substantial economic losses due to disease containment measures and the high demand for healthcare services [4, 5]. Limited healthcare resources, especially in areas with low immunization rates, exacerbate these challenges, highlighting the importance of early testing and prompt treatment to improve patient outcomes [6].

In March 2021, there was an increase in the number of cases of the Alpha strain of SARS-CoV-2 cases in Thailand. By June 2021, the Delta strain of SARS-CoV-2 infection had become predominant. The situation worsened significantly, including limitations in available treatments and hospital beds. The number of new infections exceeded 20 000 cases per day. Home isolation was initiated to address these issues, but it was still inadequate and fell short of providing appropriate care. The primary challenges included a severe patient accumulation, increased vulnerability of high-risk groups, limited resources, and escalated risks of community transmission. The evolving landscape of healthcare interventions during crises such as the COVID-19 pandemic underscores the need for innovative solutions. A notable adaptation was the conversion of hotels into healthcare facilities, termed “hospitels,” to address shortages in healthcare resources and hospital beds. Thailand adopted this strategy to meet the escalating demands on medical infrastructure during the peak of the pandemic [7,8,9].

Patients categorized as mild had symptoms such as fever, myalgia, cough, or upper respiratory tract symptoms without hypoxia, while moderate cases included those with evidence of lower respiratory tract involvement but maintained oxygen saturation levels ≥ 94%. Severe cases were defined by the presence of respiratory distress or an oxygen saturation level below 94%, in accordance with the National Institutes of Health (NIH) guidelines [10]. This study evaluated the use of hospitels in managing asymptomatic, mild, and moderate COVID-19 cases during the pandemic, categorizing them as nonsevere illness. By analyzing clinical characteristics, outcomes, and admission costs, the research highlights the practicality and efficiency of hospitels as an alternative to traditional healthcare settings. These findings offer valuable guidance for optimizing patient care and resource allocation in future pandemics, particularly in settings with limited healthcare capacity.

Materials and methods

Study design and participants

This retrospective study included all eligible patients with nonsevere COVID-19 infection admitted to Siriraj Hospital’s two affiliated hospitels in Bangkok between July 1 and October 13, 2021. The inclusion criteria were COVID-19 confirmed by real-time reverse transcription-polymerase chain reaction, age ≥ 18 years, and an initial chest X-ray (CXR) performed before hospitel admission. Patients admitted for step-down therapy, pregnant individuals, and those with incomplete data were excluded. Demographic data, clinical characteristics, and treatment-related information were collected from medical records. The study protocol was authorized by the Scientific Ethics Committee of the Siriraj Institutional Review Board (approval number Si 732/2021; March 14, 2023), Faculty of Medicine Siriraj Hospital, Mahidol University, Bangkok, Thailand.

Definition and COVID-19 management at the hospitel

In mid-April 2021, the Faculty of Medicine at Siriraj Hospital, Mahidol University, established “hospitels” in collaboration with partner hotels in response to an emergency policy promulgated by Thailand’s Ministry of Public Health. A hospitel was a repurposed hotel that admitted COVID-19 patients to single or double rooms. Physicians and other healthcare professionals stationed at each hospitel managed these patients via telemedicine [9].

Patient eligibility and criteria

Patients eligible for hospitel management were categorized as asymptomatic or as having mild to moderate COVID-19 symptoms, regardless of risk factors for severe disease. The hospitel model combined remote monitoring and telehealth consultations to safely manage COVID-19 patients outside of hospitals, preventing resource strain. Risk factors for severe disease were not the sole criterion for hospitel exclusion, as the model ensured adequate medical oversight for all patients, regardless of comorbidities. This approach aligned with Thailand’s guidelines, which prioritized reserving hospital beds for severe cases while managing asymptomatic and mild-to-moderate patients in alternative settings like hospitels or home isolation.

In accordance with the Thai guidelines established during the study period, the treatment of COVID-19 patients was categorized into the following four groups on the basis of disease symptoms and the risk factors for severe disease [11]:

• Group 1: asymptomatic COVID-19 patients.

• Group 2: non-severe symptomatic COVID-19 patients without pneumonia and no risk factors for severe disease.

• Group 3: non-severe symptomatic COVID-19 patients with risk factors for severe disease or pneumonia without hypoxia.

• Group 4: pneumonia with hypoxia (resting SpO₂ < 96% or exercise-induced hypoxemia).

At Siriraj Hospital, we primarily used Thai guidelines for considering hospitel admission, which included groups 1,2, and 3. Patients with SpO₂ ≥ 96% and no decrease in SpO₂ ≥ 3% after exercise were classified as having nonsevere illness [11]. Risk factors for severe disease were age > 60 years, chronic lung disease, chronic kidney disease, cardiovascular disease, congenital heart disease, uncontrolled diabetes mellitus, cerebrovascular disease, active malignancy, cirrhosis, immunocompromised status, body weight > 90 kg, or body mass index > 30 kg/m² [11].

Diagnostic and admission process

Upon receiving a positive SARS-CoV-2 real-time reverse transcription-polymerase chain reaction results, patients were instructed to visit the Acute Respiratory Infection Clinic at Siriraj Hospital for CXR evaluation. Radiologists or infectious disease specialists assessed the CXR findings. Abnormal CXR findings were defined as any pulmonary opacity, including unilateral, bilateral infiltration. These could be consolidations or ground glass opacities with peripheral and basal predominance [12]. Patients with pulmonary infiltration visible on CXR could still be managed in hospitels as long as they did not show severe symptoms or hypoxia. Following this assessment, eligible patients were transported to a hospitel by a designated van.

Admission and monitoring

Upon admission to a hospitel, patients received a thermometer, a pulse oximeter, and supportive medications such as paracetamol, antihistamines, and cough suppressants. For those requiring antiviral therapy, favipiravir was administered for 5 to 10 days, as it was the only available and promising oral antiviral medication in Thailand at the time [8, 13]. Treatment with favipiravir was provided according to infectious disease specialists’ judgments which were mostly based on the following criteria:

• those diagnosed with pneumonia either initially or during their stay.

• those with mild symptoms and any risk factor for severe disease.

• those with moderate symptoms.

Favipiravir (200 mg/tablet) was administered as follows: 1800 mg (9 tablets) twice daily on day 1, followed by 800 mg (4 tablets) twice daily for 5 days. For patients whose symptoms did not improve by day 5 or who had some degree of pneumonia, the treatment duration was extended to 10 days. Patients who weighed more than 90 kg received 2400 mg (12 tablets) twice daily on day 1 and 1000 mg (5 tablets) twice daily on subsequent days [10, 14].

Patient communication and daily assessments

Medical histories were collected over the phone, eliminating the need for direct physical examinations. Patients were instructed to monitor and record their vital signs at least once daily and to report these signs to healthcare staff via a secure online application. Additionally, patients had 24-hour access to medical personnel through direct phone calls if necessary. Physicians and staff from various departments of Siriraj Hospital volunteered to work at each hospitel under the close supervision of infectious disease specialists. These volunteers initiated daily phone calls to assess patients’ clinical symptoms and addressed any concerns. However, intravenous medication and blood tests were not performed at the two hospitels affiliated with Siriraj Hospital.

Management of severe symptoms

Patients with SpO₂ < 96%, who were diagnosed to have some degree of COVID-19 pneumonia, were offered dexamethasone treatment. Infectious disease specialists provided guidelines for the on-site doctors to determine the initiation of dexamethasone. Oral dexamethasone was initially administered at a dose of 6 mg once daily, with an increase in dosage if necessary [15]. Patients who potentially needed hospital referrals were evaluated by the on-site doctors in consultation with an infectious disease specialist. Patients who developed severe symptoms or who did not improve after receiving dexamethasone were referred to Siriraj Hospital for further management and admitted to a cohort ward, an airborne infection isolation room, or the intensive care unit.

Discharge protocol

According to the guidelines, all patients were discharged by day 14 of illness or earlier, but those discharged earlier were required to be isolated at home until day 14 [11].

Data collection

We collected data from Siriraj Hospital’s electronic database, including baseline characteristics, vaccination status, underlying diseases, COVID-19-related symptoms, COVID-19 testing, and treatment outcomes. Additionally, we examined the direct medical costs of patient admission. The “total cost of COVID-19 care” was calculated as the cost per patient from a hospital perspective, which included all direct medical costs accrued during a hospitel stay plus any subsequent admissions. The direct medical costs of concern included costs related to drugs, medical procedures, service fees (including room and food), and others.

Statistical analysis

We classified patient data into two groups on the basis of normal and abnormal CXR findings and compared the characteristics of each group. Demographic and clinical information are presented via descriptive statistics, with categorical variables shown as frequencies and percentages. Continuous variables with a normal distribution are represented by their means and standard deviations, whereas nonnormally distributed variables are described using medians and interquartile ranges. We utilized Fisher’s exact test to compare categorical data and employed Student’s t-test and the Mann–Whitney U test for normally and nonnormally distributed continuous data, respectively.

A multivariable logistic regression model was used to identify factors associated with referral status. The results are presented as adjusted odds ratios (aORs) with 95% confidence intervals (CIs). The results with a P value less than 0.05 were considered statistically significant. Statistical analyses were conducted using IBM SPSS Statistics, version 28 (IBM Corp, Armonk, NY, USA) and Stata statistical software, release 17 (StataCorp LLC, College Station, TX, USA).

Results

Patient enrollment and exclusion criteria

Between July 1 and October 13, 2021, during the Delta wave of the pandemic, 2253 patients with confirmed SARS-CoV-2 infection were admitted to the two hospitels affiliated with Siriraj Hospital. Of these, 524 patients were excluded from our investigation because of step-down therapy indication from hospitalization (464, 20.6%), lack of CXR at admission (32, 14.2%), pregnancy (15, 0.7%), age under 18 years (9, 0.4%), or incomplete data (4, 0.2%). The final analysis included 1729 patients, as illustrated in Fig. 1.

Fig. 1

Study flow chart

Baseline characteristics and clinical features

The baseline characteristics and clinical features of the COVID-19 patients were compared on the basis of their initial CXR findings (Table 1). Among the 1729 patients, 644 (37.2%) had an abnormal baseline CXR. These patients were older (49.2 vs. 42.2 years, P < 0.001), had greater body weights (68.1 vs. 64.7 kg, P < 0.001), and had higher body mass indices (26.3 vs. 24.9 kg/m², P < 0.001). They were also less likely to be fully vaccinated (P < 0.001). No significant difference in sex distribution was noted, with a greater proportion of females than males. The patients’ characteristics based on referral status were shown in Table S1 and Figure S1 as supplemental information.

Table 1 Baseline characteristics of patients (n = 1729)

Underlying diseases and risk factors

Patients with abnormal CXRs were more likely to have underlying diseases such as diabetes (13.4% vs. 8.5%, P = 0.002), hypertension (23.9% vs. 15.3%, P < 0.001), obesity (22.8% vs. 15.1%, P < 0.001), chronic lung disease (0.8% vs. 0.1%, P = 0.029), and autoimmune disease (1.6% vs. 0.5%, P = 0.028). They also had a greater risk of severe COVID-19 (53.6% vs. 32.2%, P < 0.001). The initial SpO₂ level among patients with abnormal CXRs was lower than that of the normal CXR group (98.0% vs. 98.2%, P < 0.001). The initial SpO₂ level among patients referred back was also lower than that the group discharged to home (97.1% vs. 98.2%, P < 0.001).

COVID-19-related symptoms

There were significant differences in the incidence of some COVID-19-related symptoms including cough (67.7% vs. 55.0%, P < 0.001) and fatigue (13.4% vs. 8.7%, P = 0.003) among patients with abnormal CXRs. Other symptoms were similar between the two groups.

Treatment, hospital referral status, and associated factors

According to the univariable analyses in Table 2, patients with abnormal CXRs were more likely to receive dexamethasone for treating COVID-19 infection (26.4% vs. 7.8%, P < 0.001) and were more frequently referred back to Siriraj Hospital (7.3% vs. 1.8%, P < 0.001). Figure 2 shows the adjusted odds ratio (aOR) of referral status. The adjustments were made for variables including CXR results, age, body weight, comorbidities, having high-risk factors for COVID-19, related symptoms, duration from symptoms onset to test date, initial SpO₂, and dexamethasone use. A higher initial SpO₂ level was significantly associated with discharge status (aOR 0.54, 95% CI 0.42–0.70). Conversely, the use of dexamethasone was significantly associated with patients being referred back to Siriraj Hospital (aOR 7.63, 95% CI 4.20–13.83). The data between patients discharged to home and referred back group are shown in table S1 and figure S1.

Table 2 Comparison of outcomes between groups regarding baseline chest X-ray (n = 1729)

Fig. 2

Adjusted odds ratio of referral status

Economic aspects and costs

There was no significant difference in the total cost of COVID-19 care between patients with normal and abnormal CXR findings. However, the total cost of COVID-19 care was significantly greater for patients referred back to Siriraj Hospital than for those who were discharged (70 545 vs. 12 010 Thai Baht, P < 0.001; these data are displayed in the table S1; in 2021, 1 USD = 31.81 Thai Baht (THB) [16]).

Discussion

This study focused on a population with asymptomatic and mild to moderate COVID-19 symptoms who received care through the hospitel system managed by Siriraj Hospital during the severe outbreak of the Delta variant, when there were insufficient hospital beds available. Baseline differences were observed between patients with normal and abnormal initial CXR findings, which were associated with varying disease outcomes and higher total care costs, particularly in patients with lower initial SpO₂ levels.

Patients with abnormal initial CXR findings were older, had higher body weights and BMI, and were more likely to have underlying conditions, indicating a greater risk of severe disease. They also experienced more symptoms, including cough and fatigue, had a longer time from symptom onset to diagnosis, and showed lower SpO₂ levels. These findings align with previous studies comparing mild and moderate COVID-19 patients, showing worse baseline characteristics in the latter group [17,18,19]. However, comparisons with patients having severe symptoms or pneumonia were not made in this study.

Patients in hospitels underwent initial CXR to determine whether they could be safely managed remotely, with daily self-reporting of vital signs. Although the mortality rate could not be determined due to the retrospective nature of the study, patients with abnormal CXR findings used dexamethasone more frequently and had a higher referral rate to hospitals. These patients tended to have lower initial SpO₂ levels and were more likely to experience disease progression or complications that led to hospital referral. These findings suggest the need for closer monitoring of patients with abnormal CXR findings. After adjusting for variables, CXR findings were not associated with hospital readmission despite our initial patient categorization on the basis of CXR results. Analysis of referral status revealed higher initial SpO₂ levels in the discharged patients, whereas those referred back to Siriraj Hospital reported increased dexamethasone use. Future criteria for safe hospitel retention should consider factors such as dexamethasone use and SpO₂ levels. Our findings on patients requiring dexamethasone, which indicates a risk for severe disease or worsening symptoms, align with those of prior studies [20,21,22].

The study adds valuable insights into COVID-19 management in hospitels, comparing patients based on their initial CXR findings, which provides a clearer risk stratification. This approach contrasts with previous studies that primarily categorized patients by symptom severity, allowing for more targeted interventions. The use of telemonitoring and radiographic tools also showed promise in improving clinical outcomes, including reducing hospital transfers and mortality [7, 8].

This research offers insights into hospitel management and associated costs in Thailand’s healthcare system, which comprises three main schemes: the Civil Servant Medical Benefit Scheme, the Social Security Scheme, and the Universal Coverage Scheme. During the pandemic, the government mandated free COVID-19 treatment at any hospital, with hospital expenses later recovered from the Ministry of Public Health.

Treatment costs varied substantially on the basis of symptom severity, resources, equipment, services, and miscellaneous expenses. Each hospital overseeing a hospitel initially covered these expenses in full, with the entire outlay later reimbursed by the Ministry of Public Health. Cost analysis revealed that patients with abnormal CXRs, or those requiring hospital referral, incurred higher costs due to increased medication and monitoring needs. While home isolation was more cost-effective in some cases, hospitels provided an alternative when home isolation was not feasible. In that study the cost per admission for home isolation was approximately 15 492 THB. However, the cost of hospitel was approximately 27 165 THB per admission. Therefore, if patients can self-monitor their symptoms at home, home isolation may be more suitable. Compared with full hospitalization, this approach significantly reduced unnecessary expenses for patients with mild to moderate COVID-19, particularly for those lacking space for self-isolation. Despite the comparable length of stay in hospitel of almost 8 days, similar patient characteristics, and same perspective of analysis, our cost for hospitel was lower than that reported in the previous study. The study also highlighted variations in care costs across university hospitals, with hospitel costs comparable to home isolation in previous studies, suggesting potential for more cost-efficient management strategies [23].

This study’s strengths include comprehensive data collection from all eligible hospitel patients, yielding a large sample with detailed demographic and clinical information. This study uniquely analyzed patient characteristics and outcomes based on CXR findings, providing insights into the utility of radiographic assessments in resource-limited settings. Additionally, we assessed hospitel care costs, addressing a gap in the literature.

The limitations of this study include potential data incompleteness and recall bias inherent to retrospective studies. CXR evaluations, which were predominantly performed by infectious disease specialists due to workforce constraints, may have affected the accuracy of the clinical assessments. Patient self-reports of symptoms during hospitel stays could also have introduced errors. Insufficient follow-up data on postdischarge outcomes, including potential ongoing symptoms affecting quality of life, limited our analysis. Finally, a lack of mortality data prevented a full evaluation of the effectiveness of the hospitels.

Despite these limitations, the low hospital referral rate (< 8%), even in those with nonsevere COVID-19, suggests that the utilization of hospitels may be an effective pandemic strategy for conserving healthcare resources for patients with greater needs.

Conclusions

Patients with abnormal CXRs had higher risks for severe disease and hospital referral, highlighting the importance of close monitoring in hospitels. This group was also incurred higher healthcare costs than patients discharged to home did. The hospitel model may prove valuable for managing future high-transmission illnesses. However, further prospective studies are needed to validate these findings.