BIO-CXRNET: a robust multimodal stacking machine learning technique for mortality risk prediction of COVID-19 patients using chest X-ray images and clinical data.

Nowadays, quick, and accurate diagnosis of COVID-19 is a pressing need. This study presents a multimodal system to meet this need. The presented system employs a machine learning module that learns the required knowledge from the datasets collected from 930 COVID-19 patients hospitalized in Italy during the first wave of COVID-19 (March-June 2020). The dataset consists of twenty-five biomarkers from electronic health record and Chest X-ray (CXR) images. It is found that the system can diagnose low- or high-risk patients with an accuracy, sensitivity, and F1-score of 89.03%, 90.44%, and 89.03%, respectively. The system exhibits 6% higher accuracy than the systems that employ either CXR images or biomarker data. In addition, the system can calculate the mortality risk of high-risk patients using multivariate logistic regression-based nomogram scoring technique. Interested physicians can use the presented system to predict the early mortality risks of COVID-19 patients using the web-link: Covid-severity-grading-AI. In this case, a physician needs to input the following information: CXR image file, Lactate Dehydrogenase (LDH), Oxygen Saturation (O2%), White Blood Cells Count, C-reactive protein, and Age. This way, this study contributes to the management of COVID-19 patients by predicting early mortality risk. Supplementary Information: The online version contains supplementary material available at 10.1007/s00521-023-08606-w.

The Effect of Economic Policy Uncertainty on Foreign Direct Investment in the Era of Global Value Chain: Evidence from the Asian Countries

The global value chain has promoted foreign direct investments in emerging markets. Not only resources but also public policies can affect the inflows or outflows of foreign direct investments (FDI). This study investigates the effect of economic policy uncertainty on net foreign direct investment inflows in 48 Asian countries. We use the panel dataset from different sources from 1995 to 2020. Our core dependent variable is net foreign direct investment inflows, and the explanatory variable is economic policy uncertainty. The study's control variables include trade, GDP per capita, GDP growth, population, financial development, inflation, and employment. We use the generalized system method of moment (SYS_GMM). Furthermore, the robustness of our empirical results is checked by using the different proxy variables of policy uncertainty. Our results confirm the negative effect of policy uncertainty on foreign direct investment inflows in 48 Asian countries. Our results show that foreign investment inflows are more sensitive than domestic investment. The influence of domestic and global uncertainty on inward FDI is greater than domestic investment. Furthermore, the interaction effect of financial development (FD) shows that FD does not affect mitigation of the negative impact of global economic policy uncertainty on foreign investment inflow. In contrast, FD mitigates the adverse effects of domestic policy uncertainty on foreign and domestic investment. The findings imply that policies need to be attractive, effective, and transparent to woo FDI to the emerging markets.

What motivated residents of Saudi Arabia to receive the COVID-19 vaccine?

Background: Acceptance of vaccination is a multifactorial issue. The unprecedented speed at which the COVID-19 disease spread globally has meant that people have had to face the idea of receiving novel vaccines for a novel disease. Purpose: Studies conducted earlier in the pandemic had shown high vaccine hesitancy in Saudi Arabia, therefore we wanted to understand the motivating factors for people living in Saudi Arabia with regards to accepting the COVID-19 vaccine, our survey was conducted when the government had already mandated vaccination to enter public spaces. Saudi society is not particularly outspoken and therefore it was of special importance to the authors to explore the motivation behind COVID-19 vaccines. Methods: This is a cross-sectional survey of 802 participants living in Saudi Arabia. The questionnaire was distributed to staff, visitors, and patients in a hospital in Saudi Arabia and via electronic means to the general population. Results: A total of 521 (65%) of the respondents were women, and 281 (35%) were men. A total of 710 (88.5%) were Saudi, and 55 (6.9%) were non-Saudi. The majority of participants (496, 65.7%) stated that they registered for the vaccine as soon as it was available, with 185 (24.5%) stating that they registered when they were mandated to do so and 74 (9.8%) registered only when they felt cases were increasing. Most participants (316, 41%) stated that the main reason for taking the vaccine was one of a self-protective nature, followed by indirect vaccination (240, 31.1%), paternalistic reasons (157, 20.4%) and altruistic reasons (58, 7.5%). Conclusions: With the increased burden on healthcare that is being faced by COVID-19, other resources need to be carefully allocated. This paper may aid the Saudi government in understanding the motivation for the population to take the vaccine and therefore facilitate any future vaccination campaigns to ensure the best utilization of resources.

Effects of elongation longitudinaux avec decoaption osteo-articulaire and post-facilitation stretching technique on pain and functional disability in mobile users with text neck syndrome during COVID-19 pandemic: A randomized controlled trial.

BACKGROUND: Adolescent smartphone users have developed a variety of musculoskeletal conditions due to their use of smartphones. ELDOA (Elongation longitudinaux avec decoaption osteo-articulaire) is based on maximizing facial and spinal stretching by assuming specific posture for 1 minute and post facilitation stretching is also a type of muscle energy technique that is used for chronically shortened or tight muscles. This aimed to compare the effects of post-facilitation stretching technique and ELDOA on neck pain and functional disability in mobile users experiencing Text neck syndrome during COVID-19. METHODS: This single-blinded randomized control trial with a parallel group design was conducted at the Department of Physiotherapy Safi Hospital (Faisalabad, Pakistan) from September 2021 to April 2022. Forty smartphone users between the ages of 18 and 35 who had a Neck Disability Index score of >10 due to neck pain without unilateral arm symptoms participated in the study. Of the 40 participants, twenty were randomly assigned to the ELDOA group and twenty were assigned to the post facilitation stretching group, and each group received 3 weekly sessions of treatment for 6 weeks. The Numeric Pain Rating Scale (NPRS), Neck Disability Index (NDI), and Smartphone Addiction Scale (SAS) were used to measure pain intensity, functional disability, and smartphone addiction at baseline and after 18 sessions of treatment. SPSS version 22 was used to enter and analyze the data. To find comparisons between groups an independent sample t test was used, and a paired sample t test was used to find the difference within each group. RESULTS: Post-treatment values showed statistically significant difference between groups. ELDOA group showed greater improvement in pain (P < .03) with 95% CI [-1.33, -0.068] and functional disability (P < .05) with 95% CI [-4.44, 0.143] at 6th week. There was no statistically significant difference (P = .35) with 95% CI [-28.6, 10.4] between the two groups regarding smartphone addiction. The NPRS, NDI, SAS scores were significantly different within each group with P < .05. CONCLUSION: The study concluded that ELDOA method and post-facilitation stretching both were effective in treating neck pain and functional disability. However, ELDOA method was superior to post-facilitation stretching effects on neck pain and functional disability among patients with text neck syndrome.

Evaluation of Core Coordination Strategies and Concept of Operations implemented during activation of Public Health Emergency Operations Center for COVID-19 Response in Pakistan (preprint)

Background: Public health emergency management systems encountered difficulties in developing countries, especially in Pakistan. The COVID-19 pandemic was extremely challenging for different agencies, departments, and institutions in Pakistan. Public health emergency management depends on a well-established public health emergency operations center that could generate a coordinated response to escalated incidents. We conducted an assessment of public health emergency response and coordination implemented during the COVID-19 pandemic management at a strategic level in Pakistan. Methods: This was mix-method study including qualitative and quantitative data collection processes implemented in 2022-23. Primary data was collected by using a structured questionnaire, and secondary data was collected by desk review. The agencies engaged in pandemic response at the national level in Pakistan were included in the assessment. Results: The overall score of the emergency response coordination system during COVID-19 was 49% for all agencies. We found that agencies faced challenges in leadership, legislation, and financing issues during the pandemic response (44%). None of the agencies had a fully developed framework for joint planning and response system for health emergencies. Roles and responsibilities attached to designated agencies in response were relatively clear (55%) for most of the agencies. Conclusion: An effective public health emergency response is based on multi-departmental coordination, resource mobilization, and correct information. Pakistan must proactively address these and other relevant challenges for future pandemic response.

Study Protocol for A Statewide Educational Intervention Focused on Reducing COVID-19 Health Disparities Through a Quality Improvement Approach (preprint)

Objective To describe the protocol of a statewide educational program for healthcare professionals (HCP) addressing COVID-19 disparities using quality improvement (QI) approach. Intervention A 19-month program to educate HCP has been developed in the priority areas of health equity, cultural sensitivity, infection prevention and control (IPC), and QI to address COVID-19 disparities. Method This innovative approach combines the Extension for Community Healthcare Outcomes (ECHOTM) model of learning with one-on-one coaching to assist participants with practical application. Participants are a diverse group of HCP and stakeholders working in healthcare and public health organizations. Participants meet twice monthly for 90-minute virtual interactive sessions led by subject matter experts and project consultants. The sessions are divided into didactic presentations, case discussions, COVID-19 updates, and pre/post-session assessments, and are eligible for continuing education credits. Participants are supported to develop a QI project addressing an area for health disparities in their work setting under the guidance of QI and health equity coaches. Using the RE-AIM framework, a mixed-mode approach is used to collect quantitative data on continuing education credit claims, participants’ characteristics, participation, satisfaction, and learning competence. Semi-structured qualitative interviews are used to gain insights into participants’ application of ECHO training and the project's impact on their competence and plan to implement relevant changes using QI approach. Conclusion The ECHO model has significant strengths based on its multidisciplinary approach and case-based learning that help interlinks priority areas to uniquely address COVID-19 disparities. Disseminating protocol for this novel intervention has the potential to generate additional projects addressing health disparities using this educational platform.  

Mutational analysis of the spike protein of SARS-COV-2 isolates revealed atomistic features responsible for higher binding and infectivity

Introduction: The perpetual appearance of Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-COV-2), and its new variants devastated the public health and social fabric around the world. Understanding the genomic patterns and connecting them to phenotypic attributes is of great interest to devise a treatment strategy to control this pandemic. Materials and Methods: In this regard, computational methods to understand the evolution, dynamics and mutational spectrum of SARS-CoV-2 and its new variants are significantly important. Thus, herein, we used computational methods to screen the genomes of SARS-CoV-2 isolated from Pakistan and connect them to the phenotypic attributes of spike protein;we used stability-function correlation methods, protein-protein docking, and molecular dynamics simulation. Results: Using the Global initiative on sharing all influenza data (GISAID) a total of 21 unique mutations were identified, among which five were reported as stabilizing while 16 were destabilizing revealed through mCSM, DynaMut 2.0, and I-Mutant servers. Protein-protein docking with Angiotensin-converting enzyme 2 (ACE2) and monoclonal antibody (4A8) revealed that mutation G446V in the receptor-binding domain;R102S and G181V in the N-terminal domain (NTD) significantly affected the binding and thus increased the infectivity. The interaction pattern also revealed significant variations in the hydrogen bonding, salt bridges and non-bonded contact networks. The structural-dynamic features of these mutations revealed the global dynamic trend and the finding energy calculation further established that the G446V mutation increases the binding affinity towards ACE2 while R102S and G181V help in evading the host immune response. The other mutations reported supplement these processes indirectly. The binding free energy results revealed that wild type-RBD has a TBE of −60.55 kcal/mol while G446V-RBD reported a TBE of −73.49 kcal/mol. On the other hand, wild type-NTD reported −67.77 kcal/mol of TBE, R102S-NTD reported −51.25 kcal/mol of TBE while G181V-NTD reported a TBE of −63.68 kcal/mol. Conclusions: In conclusion, the current findings revealed basis for higher infectivity and immune evasion associated with the aforementioned mutations and structure-based drug discovery against such variants.

An analysis of business students’ stressors and their coping strategies in the post-pandemic era

The COVID-19 pandemic has brought a dramatic change in teaching–learning practices in higher education across the world. The post-pandemic era with a new shift of teaching–learning phase raises the concern of students' psychological well-being, especially the novice, i.e., first-year undergraduate students, as first-year students often struggle to adjust to the new environment of higher institutions, thereby depressive symptoms are more prevalent among them. In this regard, a qualitative study was conducted followed by a dyadic approach to explore the symptoms of students’ psychological stressors, their causes, and provide essential remedies for coping with such behaviors in the post-pandemic era. The findings show that students’ stressors were commonly reported due to the academic workload, institutional regulations, lack of resources, and financial constraints. Regarding indications, unusual behavior, lack of confidence, improper sleep, and lack of motivation were identified as the primary symptoms of students’ stress. However, participants reported a lack of understanding regarding stress-related matters. Yet, students and faculty members try to adopt several strategies for reducing such stressors. By examining the perceived psychological stress, indications, and coping behaviors, this study enriches the literature by bridging pragmatic and theoretical gaps regarding students’ stressors of business management. Furthermore, the study mainly suggests that adequate institutional support should be provided to deal with students’ problems that cause stress. Also, universities should organize workshops or seminars to cultivate psychological health awareness among students and faculty members.

Case report: A brief description of unseen facts in the comorbid COVID-19 patients (preprint)

Data on two deceased individuals with COVID-19 and comorbidities such as hepatitis C, chronic kidney disease, diabetes mellitus type 2 and hypertension are discussed. Changes in laboratory signatures with impact on COVID-19 severity in both cases indicate the need for extensive monitoring of comorbid individuals to reduce morbidity and mortality.

BIO-CXRNET: A Robust Multimodal Stacking Machine Learning Technique for Mortality Risk Prediction of COVID-19 Patients using Chest X-Ray Images and Clinical Data (preprint)

Fast and accurate detection of the disease can significantly help in reducing the strain on the healthcare facility of any country to reduce the mortality during any pandemic. The goal of this work is to create a multimodal system using a novel machine learning framework that uses both Chest X-ray (CXR) images and clinical data to predict severity in COVID-19 patients. In addition, the study presents a nomogram-based scoring technique for predicting the likelihood of death in high-risk patients. This study uses 25 biomarkers and CXR images in predicting the risk in 930 COVID-19 patients admitted during the first wave of COVID-19 (March-June 2020) in Italy. The proposed multimodal stacking technique produced the precision, sensitivity, and F1-score, of 89.03%, 90.44%, and 89.03%, respectively to identify low or high-risk patients. This multimodal approach improved the accuracy by 6% in comparison to the CXR image or clinical data alone. Finally, nomogram scoring system using multivariate logistic regression -- was used to stratify the mortality risk among the high-risk patients identified in the first stage. Lactate Dehydrogenase (LDH), O2 percentage, White Blood Cells (WBC) Count, Age, and C-reactive protein (CRP) were identified as useful predictor using random forest feature selection model. Five predictors parameters and a CXR image based nomogram score was developed for quantifying the probability of death and categorizing them into two risk groups: survived (<50%), and death (>=50%), respectively. The multi-modal technique was able to predict the death probability of high-risk patients with an F1 score of 92.88 %. The area under the curves for the development and validation cohorts are 0.981 and 0.939, respectively.

Bifurcation analysis of a discrete-time compartmental model for hypertensive or diabetic patients exposed to COVID-19.

In this article, a mathematical model for hypertensive or diabetic patients open to COVID-19 is considered along with a set of first-order nonlinear differential equations. Moreover, the method of piecewise arguments is used to discretize the continuous system. The mathematical system is said to reveal six equilibria, namely, extinction equilibrium, boundary equilibrium, quarantined-free equilibrium, exposure-free equilibrium, endemic equilibrium, and the equilibrium free from susceptible population. Local stability conditions are developed for our discrete-time mathematical system about each of its equilibrium point. The existence of period-doubling bifurcation and chaos is studied in the absence of isolated population. It is shown that our system will become unstable and experiences the chaos when the quarantined compartment is empty, which is true in biological meanings. The existence of Neimark-Sacker bifurcation is studied for the endemic equilibrium point. Moreover, it is shown numerically that our discrete-time mathematical system experiences the period-doubling bifurcation about its endemic equilibrium. To control the period-doubling bifurcation, Neimark-Sacker bifurcation, a generalized hybrid control methodology is used. Moreover, this model is analyzed along with generalized hybrid control in order to eliminate chaos and oscillation epidemiologically presenting the significance of quarantine in the COVID-19 environment.

COVID-19 outcomes associated with clinical and demographic characteristics in patients hospitalized with severe and critical disease in Peshawar (preprint)

Background As a novel disease, understanding the relationship between the clinical and demographic characteristics of coronavirus disease 2019 (COVID-19) patients and their outcome is critical. We investigated this relationship in hospitalized patients in a tertiary healthcare setting. Aims/objectives To study COVID-19 severity and outcomes in relation to clinical and demographic characteristics of in admitted patients Methodology In this cross-sectional study, medical records for 1087 COVID-19 patients were reviewed to extract symptoms, comorbidities, demographic characteristics, and outcomes data. Statistical analyses included the post-stratification chi-square test, independent sample t-test, multivariate logistic regression, and time-to-event analysis. Results The majority of the study participants were >50 years old (67%) and male (59%) and had the following symptoms: fever (96%), cough (95%), shortness of breath (73%), loss of taste (77%), and loss of smell (77%). Regarding worst outcome, multivariate regression analysis showed that these characteristics were statistically significant: shortness of breath (adjusted odds ratio [aOR] 31.3; 95% CI, 11.87-82.53; p < 0.001), intensive care unit (ICU) admission (aOR 28.3; 95% CI,9.0-89.6; p < 0.001), diabetes mellitus (aOR 5.1; 95% CI;3.2-8.2; p < 0.001), ischemic heart disease (aOR 3.4; 95% CI,1.6-7; p = 0.001), nausea and vomiting (aOR 3.3; 95% CI, 1.7-6.6; p = 0.001), and prolonged hospital stay (aOR 1.04; 95% CI, 1.02-1.08; p = 0.001), while patients with rhinorrhea were significantly protected (aOR 0.3; 95% CI, 0.2-0.5; p < 0.001). A Kaplan-Meier curve showed that the symptoms of shortness of breath, ICU admission, fever, nausea and vomiting, and diarrhea increased the risk of mortality. Conclusion Increasing age, certain comorbidities and symptoms, and direct admission to the ICU increased the risk of worse outcomes. Further research is needed to determine risk factors that may increase disease severity and devise a proper risk-scoring system to initiate timely management.

Genomic surveillance of SARS-CoV-2 reveals emergence of Omicron BA.2 in Islamabad, Pakistan (preprint)

The Omicron variant of SARS-CoV-2 has rapidly replace previous variants of SARS CoV2 around the globe and is now a major variant of concern. The genomic surveillance of Omicron variant also reveals spread of its subvariant BA.2 which has differing transmissibility in comparison to its other subvariant BA.1. BA.1 and BA.2 harbors different mutational profile. One of the important change in both the subvariants is the presence of 69-70 deletion in BA.1 and absence of this deletion in BA.2. This deletion can be used as tool for the detection of omicron sub variants using real time PCR. In the current study we have used the TaqPath COVID-19 PCR kit for the detection of 69-70 deletion followed by genotyping using SNPsig(R) SARS-CoV-2 (EscapePLEX) kit (PrimerDesign, UK) that target K417N, E484K, and P681R mutations. The samples with the amplification of spike gene and K417N were termed as probable BA.2 cases. A subset of samples (n=13) were further subjected to whole genome sequencing. The results showed all the 13 samples were of BA.2. Hence, this assay can be used as a cost effective method for the detection of omicron BA.2 variant using real time PCR in resource limiting settings. Moreover, the detection of BA.2 with highly transmissible mutations in Islamabad, Pakistan may potentially increase the number of positive cases. In that scenario, there has to be stringent genomic surveillance to understand the circulating lineages in the country.

Genomic Characterization of Sars-Cov-2 from Islamabad Pakistan by Rapid Nanopore Sequencing (preprint)

Since the start of COVID-19 pandemic, Pakistan has experienced four waves of pandemic. The fourth wave ended in October, 2021 while the fifth wave of pandemic starts in January, 2022. The data regarding the circulating strains after the fourth wave of pandemic from Pakistan is not available. The current study explore the genomic diversity of SARS-CoV-2 after fourth wave and before fifth wave of pandemic through whole genome sequencing. The results showed the circulation of different strains of SARS-CoV-2 during November-December, 2021. We have Omicron BA.1 (n=4), Lineage A (n=2) and delta AY.27 (n=1) variants of SARS-CoV-2 in the population of Islamabad. All the isolates harbors characteristics mutations of omicron and delta variant in the genome. The lineage A isolate harbors a nine amino acid (68-76) and a ten amino acid (679-688) deletion in the genome. The circulation of omicron in the population before the fifth wave of pandemic and subsequent upsurges of COVID-19 positive cases in Pakistan highlights the importance of genomic surveillance.

Detection and upsurge of SARS-CoV-2 Omicron variant in Islamabad Pakistan (preprint)

The SARS-CoV-2 omicron variant was first detected in South Africa in November, 2021 and has rapidly spread to more than 90 countries. The emergence of Omicron variant demands for enhanced genomic surveillance to track the mutation profile and spread of virus. In the current study, we have sequenced 15 whole-genome sequences of SARS-CoV-2 Omicron variant from Islamabad region of Pakistan. Among the 15 isolates, 66% were from Islamabad whereas 33% of cases had international travel history of United Kingdom, Maldives, South Africa, and Oman. The detection of Omicron in local community and in travelers highlights the need for rigorous screening at national level and at entry points in order to contain the spread of variant.

Genomic surveillance reveals the emergence of SARS-CoV-2 Lineage A from Islamabad Pakistan (preprint)

The lineage A of SARS-CoV-2 has been around the world since the start of the pandemic. In Pakistan the last case of lineage A was reported in April, 2021 since then no case has been reported. In November, 2021 during routine genomic surveillance at National Institute of Health we have found 07 cases of lineage A from Islamabad, Pakistan. The study reports two novel deletions in the spike glycoprotein. One 09 amino acid deletion (68-76 a.a) is found in the S1 subunit while another 10 amino acid deletion (679-688 a.a) observed at the junction of S1/S2 referred as furin cleavage site. The removal of furin cleavage site may result in impaired virus replication thus decreasing its pathogenesis. The actual impact of these two deletions on the virus replication and disease dynamics needs to be studied in detail. Moreover, the enhanced genomic surveillance will be required to track the spread of this lineage in other parts of the country.

A Computational Dissection of Spike protein of SARS-CoV-2 Omicron Variant (preprint)

The emergence of SARS-CoV-2 omicron variant in late November, 2021 and its rapid spread to different countries, warns the health authorities to take initiative to work on containing its spread. The omicron SARS-CoV-2 variant is unusual from the other variants of concerns reported earlier as it harbors many novel mutations in its genome particularly with >30 mutations in the spike glycoprotein alone. The current study investigated the variation in binding mechanism which it carries compared to the wild type. The study also explored the interaction profile of spike-omicron with human ACE2 receptor. The structure of omicron spike glycoprotein was determined though homology modeling. The interaction analysis was performed through docking using HADDOCK followed by binding affinity calculation. Finally, the comparison of interactions were performed among spike-ACE2 complex of wild type, delta and omicron variants. The interaction analysis has revealed the involvement of highly charged and polar residues (H505, Arg498, Ser446, Arg493, and Tyr501) in the interactions. The important novel interactions in the spike-ACE2-omicron complex was observed as S494:H34, S496:D38, R498:Y41, Y501:K353, and H505:R393 and R493:D38. Moreover, the binding affinity of spike-ACE2-omicron complex (−17.6Kcal/mol) is much higher than wild type-ACE2 (−13.2Kcal/mol) and delta-ACE2 complex (−13.3Kcal/mol). These results indicate that the involvement of polar and charged residues in the interactions with ACE2 may have an impact on increased transmissibility of omicron variant.

Co-infections and antimicrobial use among hospitalized COVID-19 patients in Punjab, Pakistan: findings from a multicenter, point prevalence survey.

There are reports of high rates of antibiotic prescribing among hospitalized patients with COVID-19 around the world. To date, however, there are few reports of prescribing in relation to COVID-19 in Pakistan. Herein, we describe a point prevalence survey of antibiotic prescribing amongst patients hospitalized with suspected or proven COVID-19 in Pakistan. A Point Prevalence Survey (PPS) was undertaken in seven tertiary care health facilities in Punjab Provence, Pakistan. Baseline information about antimicrobial use according to the World Health Organization (WHO) standardized methodology was collected on a single day between 5th and 30 April 2021. A total of 617 patients' records were reviewed and 578 (97.3%) were documented to be receiving an antibiotic on the day of the survey. The majority (84.9%) were COVID-19 PCR positive, 61.1% were male and 34.9% were age 36 to 44 years. One quarter presented with severe disease, and cardiovascular disease was the major comorbidity in 13%. Secondary bacterial infection or co-infection (bacterial infection concurrent with COVID-19) was identified in only 1.4%. On the day of the survey, a mean of 1.7 antibiotics was prescribed per patient and 85.4% antibiotics were recorded as being prescribed for 'prophylaxis'. The most frequently prescribed antibiotics were azithromycin (35.6%), ceftriaxone (32.9%) and meropenem (7.6%). The majority (96.3%) of the antibiotics were empirical and all were from WHO Watch or Reserve categories. Overall, a very high consumption of antibiotics in patients hospitalized with suspected or proven COVID-19 was observed in Pakistan and this is concerning in view of already high rates of antimicrobial resistance in the region. Antimicrobial stewardship programs need to urgently address unnecessary prescribing in the context of COVID-19 infection.