Preliminary Exploration of the Initial Diagnostic Prediction Model of Moderate Coronavirus Disease 2019 (2019-nCoV) Based on Clinical Data

Objective: To explore those differences and relationships of the initial diagnostic clinical data between confirmed cases of 2019-nCoV and suspected cases of COVID-19, and then to establish prediction models for predicting the probability of the first diagnosis of 2019-nCoV. Methods: A total of 81 suspected cases and 87 confirmed cases of moderate 2019-nCoV diagnosed initially in the isolation wards of the First People’s Hospital of Wuhu and the People’s Hospital of Wuwei and Wuhan Caidian Module Hospital with the help of our hospital doctors were gathered, and retrospectively analyzed. Results: The most common symptoms were fever (76.79%) and cough (64.29%) in the total of 168 cases. The median age was 45 (35 - 56) years old in the 87 confirmed cases of moderate 2019-nCoV, older than the median age 36 (29 50) in the 81 suspected cases. There were significant more in the former than in the latter in the incidence of myalgia, ground-glass opacity (GGO), invasions of lesion in the peripheral lobes, vascular thickening and bronchial wall thickening, interlobular septal thicking, and small pulmonary nodules. On the contrary, there were less in the former than in the latter in the total number of leukocytes and neutrophils in blood routine examination and the levels of procalcitonin (PCT). Two groups predicting model for predicting the probability of the first diagnosis of 2019-nCoV is: P = e x /(1 + e x ), x = −6.226 + (0.071 × ages) + (1.720 × fever) + (2.858 × myalgia) + (2.131 × GGO) + (3.000 × vascular thickening and bron-chial wall thickening) + (3.438 × invasions of lesion in the peripheral lobes) + (−0.304 × the number of leukocytes) + (−1.478 × cough) + (−1.830 × pharyngalgia) + (−2.413 × headache), where e is a natural logarithm. The area under the ROC curve (AUC) of this model was calculated to be 0.945 (0.915 -0.976). The sensitivity is 0.920 and the specificity is 0.827 when the appropriate critical point is 0.360. Conclusions: A mathematical equation prediction model for predicting the probability of the first diagnosis of 2019-nCoV can be established based on the initial diagnostic clinical data of moderate 2019-nCoV. The prediction model is a good assistant diagnostic method for its high accurateness.


Introduction
During this current public health emergency of international concern, screening and diagnosing patients quickly to aid containment is a priority and these limitations make RT-PCR unsuitable for use in the field. Consequently, new tools are in great demand. Unfortunately, the sensitivity of the RNA test in the real world is not satisfactory. Clinical sensitivity of PCR decreases with days post symptom onset, Serological assay sensitivity increases with days post symptom onset [1]. Scholars ultimately found that computed tomography (CT) was associated with a higher rate of diagnostic accuracy than a real-time quantitative polymerase chain reaction (qPCR)-based approach. Even so, it is important that clinicians utilize a combination of laboratory and radiological testing where possible in order to ensure that this virus is reliably and quickly detected such that it may be treated and patients may be isolated in a timely fashion, thereby effectively curbing the further progression of this pandemic [2].
The population is generally susceptible to the 2019-nCoV. In order to prevent missed diagnosis, the diagnosis and treatment guidelines of the National Health Commission [3] point out in particular that suspected cases of 2019-nCoV can be screened first based on clinical manifestations, and then confirmed based on the new coronavirus nucleic acid detection, but the early clinical characteristics of suspected cases and confirmed cases have no significant difference [4], the detection rate of new coronavirus nucleic acid is very low [5], it is easy to delay diagnosis and treatment, and it is urgent to continue to develop new clinical diagnosis study of pneumonia caused by coronavirus infection. This study analyzes the clinical manifestations of common 2019-nCoV, screens out common 2019-nCoV risk factors, and initially explores a risk prediction model based on

Statistical Methods
Using SPSS 19.0 statistical software, some patients' peripheral blood C-reactive protein (CRP) and procalcitonin (PCT) vacancies were assigned missing values for statistical analysis. All observations were tested for normal distribution. Except for the age variable of the confirmed group which conformed to the normal distribution, the other variables did not conform to the normal distribution.
Therefore, all measurement data are represented by the median and the 25th and

General Information and Clinical Symptoms
Among all 168 patients, 93 were male (55.36%), and the median age was 39 (31 -54) years old. Fever (76.79%) and cough (64.29%) were the most common symptoms. Between the two groups, the difference in patient age was statistically significant (P = 0.013). Among them, the minimum age of the 81 suspected cases was 5 years, the maximum age was 86 years, and the median age was 36 (29 -50) soreness, which is higher than 4 in the suspected group. (4.94%) cases, the difference was significant (P = 0.003); there were no significant differences in the number of cases of fever, cough, sore throat, anorexia, diarrhea, headache, dizziness, chest tightness, and chest pain in the remaining clinical manifestations between the two groups (P > 0.005). See Table 1 for details.

Auxiliary Inspection
The peripheral blood median white blood cell count and neutrophil count  In chest CT imaging omics, ground-glass shadows and lesions infiltrating extrapulmonary bands were the most common in the two groups of patients. The suspected group was 36 (44.44%) and 42 (51.85%), which was less than 64 (73.56%) in the confirmed group.) and 82 (94.25%), the difference is significant (P < 0.001). In addition, the median number of bronchial vascular bundle thickening, interlobular thickening, and disease infiltrating lung lobes in the suspected group were less than those in the confirmed group, and the former were 6 (7.41%), 5 (6.17%), and 2 (1, 3), the latter are 50 (57.47%), 33 (37.93%) and 3 (2,4), the difference is statistically significant (P < 0.001). The median of small nodules in the suspected group was 15 (18.52%), and the median of small nodules in the confirmed group was 34 (39.08%), the difference was significant (P = 0.003). However, there was no significant difference in pleural effusion and air bronchial signs between the two groups. See Table 2 for details.

Factors Related to the First Diagnosis of Clinical Manifestations of Ordinary 2019-nCoV
Taking the 2019-nCoV as the dependent variable and the possible influencing factors in clinical manifestations as the independent variables, the unconditional binary logistic regression was used to perform multivariate regression analysis. The independent variable elimination level was 0.05, and there were missing data in the first diagnosis. PCT has also been eliminated, and the hazard ratio (HR) and its 95% CI estimate the intensity of the influence of related factors. The regression analysis results (see Table 3) showed that the general data of the patient, fever and muscle aches in the clinical symptoms, ground glass shadow, bronchial blood vessel thickening, and lesion infiltration outside the lung are common 2019-nCoV in the imaging omics. Independent risk factors; WBC count in blood routine, cough, sore throat and headache in clinical symptoms are independent negatively correlated factors of common 2019-nCoV.
Among them, e is the logarithm of the natural number, the age unit is one year old, the white blood cell count is derived from the peripheral blood routine (×10 9 /L), fever, muscle aches, cough, sore throat, headache, ground glass shadow, bronchial vascular bundle Thickening and lesions infiltrating the lung extracorporeal zone were calculated by two classifications (0 = none, 1 = yes).
Calculate the predictive value of ordinary 2019-nCoV for the above-mentioned

Discussion
This study compared the general data of 168 patients with common suspected cases of 2019-nCoV and confirmed cases, and found that the median age of common confirmed cases of 2019-nCoV was 45 (35 -56) years old, which is consistent with literature reports [6], and more than the normal suspected case of 2019-nCoV.
In this study, the clinical symptoms were more common with fever and In this study, after comparing the chest CT imaging of 168 patients with COVID-19 common suspected cases and confirmed cases, it was found that only pleural effusion and air bronchial signs were not significantly different between the two groups. The number of lesions infiltrating the outer pulmonary band, the thickening of the bronchial vascular bundle, the thickening of the interlobular septum, the number of infiltrating lung lobes and the shadow of small nodules were all higher than those in the suspected group, and the difference was significant ((P < 0.001 for the first 5 items, P < 0.001 for small nodules) = 0.004), the former are 64 (73.56%), 82 (94.25%), 50 (57.47%), 33 (37.93%), 3 (2, 4) and 34 (39.08%), the latter are 36 (44.44%), 41 (50.62%), 6 (7.41%), 5 (6.17%), 2 (1, 3) and 15 (18.52%). It has been reported in the literature that ground-glass density shadows, lesions along the bronchial vascular bundles, and thickening of the internal lobules help to distinguish between 2019-nCoV patients and suspected patients [9] [10]; due to the rich blood flow in the subpleural pulmonary lobules, the virus is easier Invasion, the GGO of patients with 2019-nCoV is mostly distributed along the subpleural [10]; the first diagnosis of 2019-nCoV lung CT is characterized by ground glass and nodular shadows located under the pleura [11]; The number of lesions of chest CT of pneumonia infected by 2019-nCoV is higher than other pneumonia infections [12]. Obviously, the chest CT imaging omics results of this study are consistent with the above literature reports.
After fully introducing various factors of clinical manifestations and performing binary logistic unconditional regression analysis, this study established a general 2019-nCoV first diagnosis prediction mathematical model, and the area under the ROC curve (AUC) reached 0.945 (0.915 -0.976). With the best critical value of 36%, the sensitivity of the prediction model is 0.920 and the specificity is 0.827. The accuracy of the diagnosis is good, and it has good predictive significance.
This study is a retrospective case-control study. The clinical data are enrolled in their first diagnosis. The number and source of samples have specific regional limitations. These limitations need to be improved and supplemented in future studies.