Delay of Detection Of COVID-19 Patients In Bangladesh; An Application To Cox Proportional Hazard Model

Objective: To delineate the survival rate of the patients with coronavirus disease 2019 (COVID-19) who did the diagnostic tests lately after the development of symptoms. The aim is to determine the socio-demographic risk factors associated with the delay of the detection of COVID-19 patients. Methods: For this cross-sectional study, 300 patients were selected who were diagnosed as COVID-19 patients in the Molecular Biology Laboratory of Chittagong Medical College, Chattogram, Bangladesh. Data were collected from May to July 2020. Clinical characteristics were obtained from over phone interviews and laboratory diagnosis by Real-time Reverse Transcriptase Polymerase Chain Reaction (rRT-PCR). Cox proportional hazard model is applied to estimate risk factors affecting the delay of detection of COVID-19 patients. Result: Female mortality rate was 44.9% higher compared to males, graduates died 32% more than undergraduates, unmarried peoples’ death rate were 56% more than married and those who were in traveling irregularly and in contact with symptomatic patients, were 86% more died than non-travelers. Conclusion: Early diagnosis of COVID-19 can save a huge amount of lives and special attention should be emphasized on the signicant explanatory variable. Real-time Reverse transcription-polymerase chain Severe Acute


Introduction
Coronaviruses are signi cant pathogens of humans and animals that can cause diseases ranging from the common cold to fatal respiratory diseases. In the past two decades, two highly pathogenic human coronaviruses have emerged in two separate events. They are the severe acute respiratory syndrome (SARS-CoV) and the Middle East respiratory syndrome (MERS-CoV) [1,2]. They caused lower respiratory tract infection as well as extrapulmonary manifestations, leading to thousands of patients with high mortality rates of up to 50%. In December 2019 a new strain of coronavirus, o cially named severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), was rst isolated from three patients with coronavirus disease 2019 (COVID-19) by the Chinese Center for Disease Control and Prevention [3,4].
Recent epidemiological reports have provided evidence for person to person transmission of the SARS-CoV-2 in family and hospital settings [5,6]. On 30 January 2020, the World Health Organization declared that the outbreak of SARS-Cov-2 constituted a public health emergency of international concern. As of 31 July 2020, the number of patients infected with SARS-CoV-2 has exceeded 2, 54,474 globally, and more than 6002 have now died of COVID-19.
Research illustrates that marked similarities exist between severe acute respiratory syndrome and COVID-19. A recent study reported a 79.5% genome sequence identity between SARS-CoV-2 and SARS-CoV, and SARS-Cov-2 was 96% identical in terms of whole-genome sequence to a bat coronavirus [7]. Clinical and pathological features of patients with COVID-19 have recently been reported, showing that the SARS-CoV-2 infection causes clusters of severe and even fatal pneumonia with clinical presentation greatly Page 3/12 resembling that of SARS-CoV infection, associated with admission to intensive care units and high mortality [8].
To keep going with this economic breakdown, the members of maximum families went outside of their homes to earn money. Regular household necessities were bought by them too from markets in this pandemic situation. Besides, a so long period of home restrictions made them unsteady to breathe fresh air, as a result, affected by COVID-19. But initially, symptoms were mild, and especially the young generation was unwilling to maintain healthy hygiene and to follow restrictions. So, in our study, we tried to identify the hazard of late detection time of COVID-19 of those patients.

Materials And Methods
A cross-sectional study was conducted from May 2020 to July 2020 in the molecular biology laboratory Patients with a SARS-CoV-2 infection con rmed by an rRT-PCR of a nasopharyngeal and oropharyngeal combined sample within 3 days of onset of symptoms were included. Presence of symptoms, days of illness, contact with symptomatic patient /health personnel, types of PPE used, travelling history, supplemental oxygen therapy requirement, hospital admission history, ICU support requirement, comorbid conditions, baseline investigations, and total treatment sheet were collected through interviewing over the phone.
The WHO website has provided several rRT-PCR protocols for detecting SARS-CoV-2 in different countries [9]. Here an automated system was repeated the ampli cation process for about 45 cycles until the viral cDNA could be detected, usually by a uorescent signal from different channels such as FAM, ROX, and CY5. We considered the SARS-CoV-2 open reading frame-ORF 1ab and the speci c conserved sequence of coding nucleocapsid protein N gene as the target regions which were designed for the conserved sequence of the double-target genes, to achieve detection of sample RNA through FAM and ROX signal changes respectively. Internal standard gene fragments (Rnase P) were detected by the CY5 signal. Here a single tube containing the necessary primers, probes, dNTPs, MgCl2, Rnasin, PCR buffer, RT enzyme, and Taq enzyme to run the entire RT-PCR reaction (24 tests/kit) was used. Positive internal control was xed to monitor the presence of PCR inhibitors in test specimens and normal saline as a negative control.
After getting the rRT-PCR results, we sorted 300 SARS-CoV-2 positive patients for our study purpose; from those we got proper history and relevant data completely. Cox proportional hazard model (1972) is well admired model used in survival analysis that can be used to assess the importance of various covariates in the survival time of individuals or objects through the hazard function. It makes us capable of estimating the relationship between the hazard rate and Page 4/12 explanatory variables without having to make assumptions on the shape of the baseline hazard function.
In our study, the cox proportion hazard model was used to analyze the factor affecting COVID-19 patients' mortality due to late diagnosis by considering any extra heterogeneity present in the data. To evaluate the equality of the hazard function we used a log rank test (does not make any assumptions regarding the distribution of the data set) to test the null hypothesis: all the survival curves were the same. To explore the relationship among factors having multiple-categories the Bonferroni test was applied.

Results
This study described 300 patients with con rmed COVID-19 from May 09, 2020 to July 12, 2020 in Molecular Biology Laboratory, Chittagong Medical College, Chattogram. In the above table, censored were those who were alive during the study period whether he/she was in isolation at home or the hospital. Total 1%female and 9% male were not alive at p = 0.006 level of signi cance. 0.7% of health professional as well as service holder died and in the case of the student, it is only 1% but the people related to the 'others' profession 7.7% died which is much more. Rural patients died 7%, where urban people were only 3% (p = 0.024). The death of married persons was 10%. Middleage adults had a 4.7% death rate; on the contrary, only 0.3% of children died. Besides, 7% of death was from those groups of patients whose family members were not affected and it was signi cantly higher (p = 0.001) than those who had affected family members. Surprisingly 5.6% of patients were died (p = 0.000) those who had no travelling history. For testing the proportional hazard assumption, we checked whether the independent variables meet the proportional assumption by using a log-minus-log (LML) plot. The LML plot is a graph constructed by applying the log-log transformation of the survival function, in other words, it is a graph of the logarithm of time against the logarithm of the negative logarithm of the estimated survival function. Figure 1 we have seen that curves of each risk factor are not crossing but they are parallel. That means all the risk factors are well de ned and satisfy the proportional hazard (PH) assumptions.
From Patients whose family members were not affected were .715 times [CI (.563, .908)] less in risk than those whose family members were affected. Besides, the patients who were in contact with symptomatic patients were 86% more in hazard [(p = 0.001) CI (28%, 71%)] than those who had no travelling history during this pandemic.

Discussion
Delay of detection of COVID-19 plays a signi cant role in this survival analysis regarding the patient's status: he/she survived or not. By the cox proportional hazard model, we explore the hazard rate of several risk factors and explained with respect to the odds ratio. From this study, it has obvious that female (1.499 times) participants and graduates (1.316) are in more risky positions compare to other participants. The Students (1.353) are in more vulnerable positions, however, the service holders and the health and personnel are at low risk. The urban (1.283 times higher) affected, and the unmarried (1.559 times higher) affected, should be more careful as they both at high risk to befall in COVID-19. We should avoid traveling because the people who travel regularly are 1.807 times at high risk. Besides, without health personnel, Contact with symptomatic patients should be kept off. As the people who were in contact with symptomatic patients are at 1.86 times more risk. This research has some limitations. First, information regarding SARS CoV-2 is limited. Second, the information provided here is based on current evidence but may be modi ed as more information becomes available. Third, the sample size was small.

Conclusion
Late detection of COVID-19 can cause serious health hazards. Besides, its outcome is signi cantly worse than those who have comorbidities which concomitantly exacerbates the death of a patient. So, early earnestness to identify the SARS-CoV-2 virus and taking prompt treatment can save many lives.