Method for establishing a covid-19 triage system based on diagnosis timeliness, the system and the triage method

Abstract

The application discloses a method for establishing a triage system of COVID-19 based on diagnosis timeliness, a computer system established by the method and a related COVID-19 triage method. The computer system comprises a data receiving module for receiving information including the age, chronic disease history, symptom onset date and first visit date of a patient; a data processing module for comparing the received information with preset priority visit conditions; and an output module for highlighting the patient meeting the preset priority visit conditions and prompting priority visit; wherein the preset priority visit conditions are selected from one or more of the following conditions: (a) having a chronic disease and being older than 61 years old, (b) having a chronic disease and being more than 4 days from the symptom onset date to the first visit date, and (c) not having a chronic disease but being older than 51 years old.

Description

[0001] This application is a divisional application of the invention patent application with application number 202110432543.0, application date April 21, 2021, entitled "Method for establishing a COVD-19 triage system based on diagnostic timeliness, the system and the triage method". Technical Field

[0002] This invention relates to disease triage, and more specifically, to a method for establishing a COVID-19 triage system based on diagnostic timeliness, the COVID-19 triage system established by the method based on diagnostic timeliness, and related COVID-19 triage methods based on diagnostic timeliness. Background Technology

[0003] From an imaging perspective, several studies have documented that COVID-19 patients who presented to the clinic or underwent imaging within 6–14 days of symptom onset were more likely to have more severe abnormal lung CT findings and lesions. To date, the relationship between COVID-19 prognosis and time-efficiency in diagnosis, and the importance of COVID-19 diagnostic time-efficiency in different populations, require further in-depth analysis.

[0004] Improving the timeliness of COVID-19 diagnosis means shortening the time from symptom onset to diagnosis. This includes both patient-led time from symptom onset to first visit and institution-led time from first visit to diagnosis. On one hand, we need patients to identify the earliest symptoms and seek medical attention as soon as possible, but currently this time interval varies from person to person. On the other hand, hospitals should improve diagnostic efficiency. Currently, the time required to diagnose COVID-19 in patients varies, as does the number of reverse transcription polymerase chain reaction (RT-PCR) nucleic acid tests. Some patients require more than two nucleic acid tests for diagnosis, while most patients are diagnosed within two days, with some taking up to seven days.

[0005] Efficient triage of suspected COVID-19 patients is needed to prevent delays in diagnosis and treatment for high-risk patients with poor prognoses. Furthermore, methods and systems applicable to different population samples are required to establish efficient triage systems tailored to the specific circumstances of each country or region. Summary of the Invention

[0006] This invention analyzes the distribution and cutoff values ​​of diagnostic timeliness-related durations, performs survival analysis and comparisons on different diagnostic timeliness groups to compensate for aspects neglected in other studies and better validate the role of diagnostic timeliness. This invention constructs a model based on diagnostic timeliness and meaningful clinical characteristics in the analysis to identify high-risk populations requiring higher diagnostic timeliness. The selection of indicators for the data receiving module and the node selection for the data processing module of this invention's system are derived from the triage system establishment method described below.

[0007] Based on the above research, this application proposes a COVID-19 triage system based on diagnostic timeliness, and a method and system for establishing such a triage system for novel population samples. Accordingly, in one aspect, the present invention also provides a method for establishing a COVID-19 triage system based on diagnostic timeliness, the method comprising: (a) collecting information including the age, chronic disease history, symptom onset date, first visit date, diagnosis date, hospitalization date, adverse outcome, and earliest date of adverse outcome occurrence of confirmed COVID-19 cases; (b) calculating the diagnostic timeliness-related duration for each confirmed case, in days, including the duration from symptom onset date to diagnosis date, the duration from symptom onset date to first visit date, and the duration from first visit date to diagnosis date; calculating the duration from hospitalization date to the earliest date of adverse outcome occurrence, and directly excluding cases if the duration is negative or greater than 31 days; (c) determining the 95th percentile of the diagnostic timeliness-related duration, and excluding data anomalies with values ​​greater than the 95th percentile; (d) using the diagnostic timeliness-related duration as the independent variable, and survival data such as whether an adverse outcome occurred and the duration from hospitalization date to the earliest date of adverse outcome occurrence as the dependent variable, and using the Maxstat maximum statistic method (maximally selected rank) to calculate the duration. (e) Using statistical analysis to determine the optimal cutoff value for converting the diagnostic time-related duration into a binary variable, and directly replacing the original continuous variable of diagnostic time-related duration with a binary variable, in order to reduce the impact of extreme or abnormal time-related duration on the stability of the triage model and strengthen the variable effect, thereby enhancing the applicability and generalizability of the triage model; (e) using, but not limited to, the diagnostic time-related duration after binary conversion, age, gender, smoking history, and chronic disease history as independent variables, and survival data such as whether adverse outcomes occurred and the time from the date of admission to the earliest date of occurrence of adverse outcomes as dependent variables, a univariate log-rank test was performed. (f) Using the diagnostic time-related duration after binary classification transformation, variables with log-rank test p-values ​​below a set threshold, age, gender, smoking history, and chronic disease history as independent variables, and survival data on whether adverse outcomes occurred and the duration from the date of admission to the earliest date of adverse outcome as dependent variables, multivariate Cox proportional hazards regression survival analysis was performed to obtain the Cox regression p-values ​​for each independent variable; (g) Using the diagnostic time-related duration after binary classification transformation and variables with Cox regression p-values ​​below a set threshold as independent variables, and survival data on whether adverse outcomes occurred and the duration from the date of admission to the earliest date of adverse outcome as dependent variables, conditional reasoning tree analysis was performed to obtain a decision tree model; (h) A triage system for COVID-19 was established based on the obtained decision tree model.

[0008] In one specific implementation, for the Maxstat maximum statistic method in step (d), the minprop (the minimal proportion of observations per group) parameter value is set to 0.15. In another specific implementation, for the Maxstat maximum statistic method in step (d), the break.time.by parameter value is 5.

[0009] In one specific implementation, the Maxstat maximum statistic method in step (d) is calculated using the surv_cutpoint() function in R programming. The minprop parameter is set to 0.15, the progressbar parameter is set to TRUE, the variables parameter is named "timeStoD", the data parameter is used to call the data frame structure dataset containing survival data (time, event) and continuous variables to be segmented, the time parameter calls the survival time index, which is set to the time required from admission to the occurrence of the adverse outcome, and the event parameter calls the survival outcome index, which is set to the status of whether the adverse outcome has occurred, with the event value set to 0 or 1.

[0010] In one specific implementation, for the Maxstat maximum statistic method in step (d), the surv_categorize() function is used to segment each variable value according to the cutpoint returned by the surv_cutpoint() function above, and the variable value of "timeStoD" is updated. Then, the output cutpoint value, i.e. the optimal cutoff value, is obtained through the summary() function.

[0011] In one specific implementation, the effectiveness of the cutoff in step (d) is verified using the Maxstat maximum statistic method. A survival function is constructed using the Surv() function, with the time, event, and data settings unchanged from previous steps. The "~predictors" parameter uses the "timeStoD" indicator generated in the previous step and is set to "~timeStoD". The ggsurvplot() function is used to plot survival curves for the above function, where the data settings remain unchanged, the pval parameter is set to TRUE, the risk.table parameter is set to TRUE, the conf.int parameter is set to TRUE, the surv.median.line parameter is set to "hv", and the linetype parameter is set to "strata". The break.time.by parameter is set to 5 for optimal plotting. The parameters for style themes, palettes, or color panels are not limited. If the two sets of survival curves do not intersect, then the binary cutoff value of the diagnostic timeliness-related indicators is valid.

[0012] In one implementation, for the univariate log-rank test in step (e), a survival function is constructed using the Surv() function, with the settings for time, event, and data being the same as in the previous steps; each variable is entered into the "~predictors" parameter one by one; the Survdiff() function is used to incorporate each variable into the Surv() survival function, and a log-rank test is performed on each variable to obtain the p-value; wherein the rho parameter value is set to 0, and the timefix parameter value is set to TRUE.

[0013] In one implementation, for the multivariate Cox proportional hazards regression survival analysis in step (f), the Surv() function is used to further construct the multivariate survival function, with the time, event, and data settings the same as in the previous steps; all independent variables are entered into the model; the coxph() function is used to incorporate the above Surv() survival function, and multivariate Cox proportional hazards regression survival analysis is performed to obtain the output value Pr(>|z|), which is the Cox regression p value of each independent variable.

[0014] In one implementation, for the multivariate Cox proportional hazards regression survival analysis in step (f), the singular.ok parameter is set to TRUE, the model parameter to FALSE, the x parameter to FALSE, the y parameter to TRUE, and the method parameter to ties. Among the three values ​​of "efron", "breslow", and "exact", ties is set to "efron". Efron is used as the default value because it is more accurate and computationally efficient in handling bound outcome times.

[0015] In one implementation, for the conditional reasoning tree analysis in step (g), the Surv() function is used to further construct a multivariate survival function, with the settings of time, event, and data being the same as in the previous steps; independent variables that meet the conditions are put into the model; the ctree() function is used to incorporate the above Surv() survival function, and continuous, truncated, ordered, nominal variables and multivariate response variables are recursively partitioned in the conditional reasoning framework to establish a decision tree model.

[0016] In one implementation, for the conditional inference tree analysis in step (g), the subset parameter value is set to NULL, the weights parameter value is set to NULL, the xtrafo parameter value is set to ptrafo, the ytrafo parameter value is set to ptrafo, and the scores parameter value is set to NULL.

[0017] In some embodiments, the threshold values ​​for the log-rank test p-value and the Cox regression p-value are set to 0.05 to 0.1. In some embodiments, the abnormal cases are cases lacking one or more of the following information: age, chronic disease history, symptom onset date, first visit date, diagnosis date, hospitalization date, adverse outcome, and earliest date of occurrence of the adverse outcome. In some embodiments, the adverse outcome includes one or more of the following: ICU admission, severe pneumonia, invasive ventilation, and death.

[0018] This invention assesses the risk of adverse outcomes in COVID-19 patients by stratifying diagnostic timeliness. Severe cases of pneumonia or admission to the ICU / invasive ventilation / death were analyzed. Cutoff values ​​for diagnostic timeliness (defined as the time from symptom onset to diagnosis, including the time from symptom onset to first visit and the time from first visit to diagnosis) were studied using the maximum selection rank statistic. Survival analyses (log-rank test, Cox regression, conditional inference tree) and comparisons between different timeliness groups were performed.

[0019] Through the above analysis, the inventors found that the median time from symptom onset to diagnosis, from symptom onset to first visit, and from first visit to diagnosis were 6 days, 3 days, and 2 days, respectively, with cutoff values ​​of 5 days, 4 days, and 3 days, respectively. After adjusting for age, sex, smoking status, and comorbidities, age (hazard ratio [HR: 1.03; 95% CI: 1.01–1.04]), comorbidities (HR: 1.84; 95% CI: 1.23–2.73), and time from symptom onset to diagnosis >5 days (HR: 1.69; 95% CI: 1.10–2.60); and age (hazard ratio [HR: 1.03; 95% CI: 1.01–1.04]), comorbidities (HR: 1.77; 95% CI: 1.18–2.68), and time from symptom onset to first medical visit >4 days (HR: 1.56; 95% CI: 1.07–2.26) were independent predictors of COVID-19 prognosis. This indicates that the longer the time from symptom onset to diagnosis, the worse the prognosis of COVID-19, especially with delayed medical visits. The CTREE model shows that time to diagnosis, age, and comorbidities are important milestones. Delayed medical visits were defined as a time interval of more than 4 days from symptom onset to the first visit. Differential analysis showed that patients with characteristics such as male, advanced age, dry cough, expectorant cough, shortness of breath, and COPD were more common in the delayed medical visit group.

[0020] Another aspect of the present invention provides a computer system for COVID-19 triage based on diagnostic timeliness, comprising: a data receiving module for receiving information including the patient's age, chronic disease history, symptom onset date, and first visit date; a data processing module for comparing the received information with preset priority treatment conditions, wherein the priority treatment includes one or more of priority prescription, priority sampling, priority sample delivery, and priority testing; and an output module for highlighting patients who meet the preset priority treatment conditions and prompting them to seek priority treatment; wherein the preset priority treatment conditions are selected from one or more of the following conditions: (a) having a chronic disease and being older than 61 years, (b) having a chronic disease and having a symptom onset date to the first visit date being more than 4 days, and (c) not having a chronic disease but being older than 51 years.

[0021] In some implementations, the output module is configured to prioritize patients according to the following rules: (1) patients with chronic diseases and older than 61 years are given first priority; (2) patients with chronic diseases and whose symptom onset date is more than 4 days prior to their first visit are given second priority; (3) patients without chronic diseases but older than 55 years are given third priority; and (4) patients without chronic diseases but older than 51 years are given fourth priority. In some implementations, the output module is further configured to prompt the patient when they are older than 51 years without chronic diseases, suggesting that a COVID-19 nucleic acid test result be provided within 5 days from the onset of symptoms. In some implementations, the output module is further configured to prompt the patient when they are older than 55 years without chronic diseases, suggesting that a COVID-19 nucleic acid test result be provided within 3 days from the first visit.

[0022] In some embodiments, the chronic disease history includes one or more of COPD, diabetes, hypertension, coronary heart disease, cerebrovascular disease, hepatitis B, tumors, chronic kidney disease, and immunodeficiency. The selection of these chronic diseases in this invention is based on calculating the frequency and proportion of comorbid chronic disease characteristics in the patient's electronic medical record discharge diagnosis, obtaining 15 high-frequency and high-proportion chronic comorbidities. These comorbidities are selected from cases with adverse outcomes, including one or more of ICU admission, severe pneumonia, invasive ventilation, and death.

[0023] In some embodiments, the symptoms include one or more of the following: dry cough, sore throat, conjunctival congestion, nasal congestion, headache, expectoration, fatigue, hemoptysis, shortness of breath, nausea, vomiting, diarrhea, joint and muscle pain, chills, and fever. The selection of these symptoms in this invention is based on symptom records obtained from the patient's chief complaint in the admission record and retrospective information collection. Frequency and percentage calculations are performed, and the top 15 most frequent, high-percentage, and consistently recurring symptoms are selected to determine the above symptom definitions.

[0024] Analysis of the triage system established in this invention revealed that the median times from symptom onset to diagnosis, from symptom onset to first visit, and from first visit to diagnosis were 6 days, 3 days, and 2 days, respectively, with cutoff values ​​of 5 days, 4 days, and 3 days. The CTREE model used in the established method showed that diagnostic timeliness, age, and comorbidities were important factors.

[0025] Another aspect of the present invention provides a COVID-19 triage method based on diagnostic timeliness, comprising: (S1) receiving information including the patient's age, chronic disease history, symptom onset date, and first visit date; (S2) comparing the received information with preset priority treatment conditions, wherein the priority treatment includes one or more of priority prescription, priority sampling, priority sample delivery, and priority testing; (S3) highlighting patients who meet the preset priority treatment conditions and prompting them to seek priority treatment; wherein the preset priority treatment conditions are selected from one or more of the following conditions: (a) having a chronic disease and being older than 61 years, (b) having a chronic disease and having a symptom onset date to the first visit date being more than 4 days, and (c) not having a chronic disease but being older than 51 years.

[0026] In some implementations, step S3 sets priorities according to the following rules: (1) patients with chronic diseases and older than 61 years are given the first priority; (2) patients with chronic diseases and whose symptom onset date is more than 4 days before the first visit are given the second priority; (3) patients without chronic diseases but older than 55 years are given the third priority; and (4) patients without chronic diseases but older than 51 years are given the fourth priority.

[0027] In some embodiments, step (S3) further includes prompting, when the patient has no chronic disease but is older than 51 years, that it is recommended to provide a COVID-19 nucleic acid test result no more than 5 days from the onset of symptoms. In some embodiments, step (S3) further includes prompting, when the patient has no chronic disease but is older than 55 years, that it is recommended to provide a COVID-19 nucleic acid test result no more than 3 days from the first visit.

[0028] In some embodiments, the chronic disease history includes one or more of COPD, diabetes, hypertension, coronary heart disease, cerebrovascular disease, hepatitis B, tumors, chronic kidney disease, and immunodeficiency. The selection of the above-mentioned chronic diseases in this invention is based on the calculation of the frequency and proportion of comorbid chronic disease characteristics in the patient's electronic medical record discharge diagnosis, obtaining 15 high-frequency and high-proportion chronic comorbidities. These chronic comorbidities are selected from cases with adverse outcomes, wherein the adverse outcomes include one or more of ICU admission, severe pneumonia, invasive ventilation, and death.

[0029] In some embodiments, the symptoms include one or more of the following: dry cough, sore throat, conjunctival congestion, nasal congestion, headache, expectoration, fatigue, hemoptysis, shortness of breath, nausea, vomiting, diarrhea, joint and muscle pain, chills, and fever. The selection of these symptoms in this invention is based on symptom records obtained from the patient's chief complaint in the admission record and retrospective information collection. Frequency and percentage calculations are performed, and the top 15 most frequent, high-percentage, and consistently recurring symptoms are selected to determine the above symptom definitions.

[0030] Analysis of the triage system established in this invention revealed that the median times from symptom onset to diagnosis, from symptom onset to first visit, and from first visit to diagnosis were 6 days, 3 days, and 2 days, respectively, with cutoff values ​​of 5 days, 4 days, and 3 days. The CTREE model used in the established method showed that diagnostic timeliness, age, and comorbidities were important factors.

[0031] Based on the analysis of a large number of hospital admission records, this invention proposes the aforementioned triage method and system. It extracts key node information and enables efficient triage of patients based on different diagnostic timeliness needs, avoiding diagnostic delays, improving patient prognosis, and rationally allocating medical resources, thus fully realizing the value of diagnostic timeliness research. Furthermore, by summarizing the logic and methods used in the above research, the inventors have also identified a method for establishing an efficient triage system. This method allows for the rapid development of triage systems suitable for the specific circumstances of each country using data from various countries. Attached Figure Description

[0032] Figure 1 Frequency density plot of diagnostic timeliness and duration of results.

[0033] Figure 2 Maximum selection log-rank statistic for the cutoff point of time from symptom onset to diagnosis. a) Patients are divided into high (right-hand) and low (left-hand) groups based on the time from symptom onset to diagnosis; the cutoff point (dashed line, showing the highest point) is defined by the maximum selection rank statistic; b) the time-dependent risk of reaching severity between patients with "high" time from symptom onset to diagnosis (lower curve) and patients with "low" time from symptom onset to diagnosis (upper curve); the transparent portion represents the 95% confidence interval (95% CI). The maximum selection rank statistic allows for the assessment of the cutoff point, which divides observations into two groups by continuous or ordered predictor variables. The calculation of the exact distribution of the maximum selection rank statistic is discussed, and a new lower bound for the distribution is derived based on an extension of the algorithm for the exact distribution of the linear rank statistic.

[0034] Figure 3Maximum choice log-rank statistics for cutoff points of time from symptom onset to first visit and from first visit to diagnosis. a) Patients were divided into high (right-side) and low (left-side) groups based on the time from symptom onset to first visit; the cutoff point (the intersection of the two lines and the point indicated by the dashed line) was defined by the maximum choice rank statistic; b) Time-dependent risk of reaching severity between patients with "high" time from symptom onset to first visit (lower curve) and patients with "low" time from symptom onset to first visit (upper curve); the transparent portion represents the 95% confidence interval (95% CI); c) Patients were divided into high (right-side) and low (left-side) groups based on the time from first visit to diagnosis. The cutoff point (the intersection of the red and blue lines and the dashed line) is defined by the maximum selection rank statistic; d) the time-dependent risk of reaching severity between patients with "high" time to diagnosis (lower curve) and "low" time to diagnosis (upper curve); the transparent portion represents the 95% confidence interval (95% CI). The maximum selection rank statistic allows for the assessment of the cutoff point, which divides observations into two groups by continuous or ordered predictor variables. The calculation of the exact distribution of the maximum selection rank statistic is discussed, and a new lower bound for the distribution is derived based on an extension of the algorithm for the exact distribution of the linear rank statistic.

[0035] Figure 4 Overall survival curve over 31 days.

[0036] Figure 5 : A conditional inference tree model for COVID-19 prognosis with diagnostic timeliness and pre-hospital factors. a) Based on significantly separated nodes in the model tree, the time-dependent risk of reaching severity is divided into 5 parts; the p-value is calculated using the corresponding time series test (log-rank test); the model includes the duration from symptom onset to diagnosis, age, and chronic comorbidity status; b) Based on significantly separated nodes in the model tree, the time-dependent risk of reaching severity is divided into 5 parts; the p-value is calculated using the corresponding time series test (log-rank test); the model includes the duration from symptom onset to first visit, the duration from first visit to diagnosis, age, and chronic comorbidity status. Detailed Implementation

[0037] definition

[0038] Throughout the specification and claims, unless the context otherwise requires, the word “comprising” and variations such as “including” and “containing” will be understood to include the said integers, steps, or ingredients, but not exclude any other integers, steps, or ingredients. When used herein, the term “comprising” may be replaced by the terms “including” or “containing”, or sometimes by the term “having”.

[0039] As used herein, the term "triage" refers to the process of determining the severity and urgency of a patient's condition and its corresponding specialty based on their main symptoms, signs, and medical history; quickly assessing and classifying the type and severity of the condition; and ensuring that patients receive appropriate diagnosis, treatment, and care at the appropriate time and in the appropriate treatment area for the appropriate reason, according to a triage priority order. In this invention, triage identifies patients who require priority treatment. Priority treatment typically includes priority in issuing medical orders, priority in sample collection, priority in sample delivery, and priority in testing.

[0040] As used herein, the term "diagnostic timeliness" refers to the time efficiency of a patient's diagnosis process at key time points or events. Typically, these key time points or events include symptom onset, initial consultation, and diagnosis (i.e., obtaining a diagnosis). Accordingly, the term "diagnostic timeliness-related duration" can include the time from symptom onset to diagnosis, the time from symptom onset to initial consultation, and the time from initial consultation to diagnosis. In this invention, the duration is specifically measured in "days".

[0041] As used in this article, the term "prognosis" refers to the predicted course of a disease based on experience. Disease prognosis involves understanding a disease, including its clinical manifestations, laboratory and imaging findings, etiology, pathology, and disease progression. Crucially, it involves assessing the short-term and long-term efficacy, recovery, or progression of the disease based on the timing and method of treatment, combined with any new findings during the treatment process. Disease prognosis is related to many factors, including the timing of treatment, the severity of the disease, medical expertise, co-existing diseases, the physician's skill, the patient's physical condition, age, whether the patient faces the disease squarely or has a clear understanding of it, and whether treatment is continued. Even with the same treatment, prognoses can vary significantly. Some of these factors are uncontrollable and unchangeable, such as age and underlying health conditions; others require early intervention to improve prognosis, such as early detection, facing the disease squarely, and early intervention and treatment. All of these factors contribute to a better prognosis.

[0042] As used herein, the term "disease outcome" refers to the outcome of a disease progression to a certain extent or stage. In this invention, an "adverse outcome" can refer to the progression of COVID-19 to a severe level. This severity can include severe pneumonia or admission to an intensive care unit / invasive ventilation / death.

[0043] As used in this article, the criteria for defining severe pneumonia cases (which should also be transferred to the ICU for close observation and active treatment) are as follows: meeting one of the primary criteria or more than (or equal to) three secondary criteria, where the primary criteria are: ① receiving invasive ventilation; ② septic shock, requiring vasoactive drugs even after fluid resuscitation; and the secondary criteria are: ① increased respiratory rate (≥30 breaths / min); ② arterial oxygen partial pressure (PaO2) / inspirated oxygen partial pressure (FiO2) ≤250 mmHg (1 mmHg = 0.133 kPa); ③ lung imaging showing multilobar infiltration; ④ altered consciousness and / or disorientation; ⑤ blood urea nitrogen ≥7 mmol / L; and ⑥ hypotension requiring fluid resuscitation.

[0044] This invention uses the terms "maximum selection rank statistic" and "cutoff value." In an ROC curve, the dependent variable must be a binary variable. However, sometimes we may face situations where the dependent variable is survival data, quantitative data, or other types of data, in which case the ROC curve is ineffective. In such cases, the maximum selection rank statistic can be used to find the critical value. Specifically, assume there is a dependent variable y (y can be categorical, continuous, or survival data) and an independent variable x. The maximum selection rank statistic is equivalent to partitioning the x variable for each value, dividing the data into two groups at each step and calculating a standardized statistic. This standardized statistic varies depending on the type of dependent variable, but generally reflects the difference between the two groups after partitioning by a certain value. After all partitions, multiple standardized statistics can be obtained. The largest one is the optimal cutoff value (essentially still a binary variable), which is the "cutoff value" used in this paper.

[0045] Specifically, in the technical solution of this invention, the continuous variables of three durations (time from symptom onset to diagnosis, time from symptom onset to first visit, and time from first visit to diagnosis) can be transformed into binary variables using the maximum selection rank statistic. Because a difference of one day is not significant and there is considerable confounding, converting them into binary variables makes the analysis results more stable. The cutoff value for this continuous variable is the cutoff value mentioned here. Then, the optimal cutoff value is found using the above method and incorporated into the subsequent model calculations.

[0046] Example

[0047] Design, data source, and data extraction

[0048] According to interim guidance from the WHO, the cases involved in this invention were diagnosed through real-time RT-PCR analysis or high-throughput sequencing of throat and nasal swab specimens. Clinical data were examined and extracted into a computer database by experienced respiratory clinicians and validated through dual input prior to analysis.

[0049] Data collection

[0050] Demographic, clinical, and prognostic characteristics (i.e., sex, age, smoking status, primary chronic disease, and symptoms) were collected, along with initial examination and chest X-ray results upon admission. Regarding chronic comorbidities, chronic obstructive pulmonary disease (COPD), diabetes, hypertension, coronary artery disease, cerebrovascular disease, hepatitis B, malignant tumors, chronic kidney disease, and immunodeficiency were recorded.

[0051] The analysis also recorded the symptom onset date, first visit date, and diagnosis date, and calculated the time between the symptom onset date and the first visit date, as well as the time between the first visit date and the diagnosis date. If any number was less than 1.0 day, it was removed from the analysis.

[0052] The study analyzed patients who reached severity (as an endpoint), including those with severe pneumonia or admitted to the intensive care unit (ICU) / invasive ventilation / death. The first day of reaching one of these criteria determined the time from admission to reaching severity.

[0053] Prognostic outcomes and duration from admission to endpoint were recorded, but both were removed when the duration was negative. Patients with outcome-related durations exceeding 31 days were excluded from prognostic correlation analyses along with patients with missing outcomes.

[0054] Statistical analysis

[0055] This invention summarizes demographic information for all samples and compares data from severe and non-severe cases. Descriptive analyses were performed on diagnostic timeliness (time from symptom onset to diagnosis, time from symptom onset to first visit, and time from first visit to diagnosis) and survival time (time from hospital admission to reaching severity), and frequency density plots were generated. Finally, demographic and clinical data were compared between high and low diagnostic timeliness. Continuous data are presented as means and ranges. The Wilcoxon rank-sum test was used to compare nonparametric values. Categorical data are presented as counts and percentages and compared using the chi-square test and Fisher's exact test.

[0056] To transform continuous variables into categorical variables, reduce the disproportionate impact of extreme values, and stabilize the computational model (using "days" as a unit of measurement is too precise for assessing prognostic outcomes), maximum selection rank statistic (Maxstat) analysis was performed to determine the appropriate cutoff point for the diagnostic time-to-determinacy (DTD). The Kaplan-Meier (KM) method was used to plot survival curves for different DTD groups. In the Maxstat analysis, the `surv_cutpoint()` function was used in R programming, with the `minprop` parameter set to 0.15, the `progressbar` parameter set to TRUE, and the `variables` parameter named "timeStoD". The `data` parameter calls a data frame structure containing survival data (time, event) and the continuous variables to be segmented. The `time` parameter calls the survival time index, set to the time required from admission to the occurrence of the adverse outcome. The `event` parameter calls the survival outcome index, set to the status of whether the adverse outcome has occurred, with the event value set to 0 or 1. The `surv_categorize()` function is used to segment each variable value based on the cutpoint returned by the `surv_cutpoint()` function above, and the value of the variable "timeStoD" is updated. Then, the `summary()` function is used to obtain the output cutpoint value, which is the optimal cutoff value. Verify the effectiveness of the cutoff: Construct a survival function using the Surv() function, with the time, event, and data settings the same as in the previous steps; the "~predictors" parameter takes the "timeStoD" indicator generated in the previous step and sets it to "~timeStoD"; use the ggsurvplot() function to plot the survival curves for the above function, where the data settings remain unchanged, the pval parameter is set to TRUE, the risk.table parameter is set to TRUE, the conf.int parameter is set to TRUE, the surv.median.line parameter is set to "hv", and the linetype parameter is set to "strata"; set the break.time.by parameter to 5 for optimal plotting effect; the parameters of style themes or color panels such as ggtheme and palette are not limited; if the two sets of survival curves do not intersect, then the binary cutoff value of the diagnostic timeliness-related indicators is valid.

[0057] Overall survival curves were plotted for patients with varying degrees of severity. Without violating the proportional hazards assumption, the prognostic significance of all pre-admission factors (duration related to time to diagnosis, sex, age, smoking status, and presence of pre-existing chronic diseases) was analyzed using log-rank tests and multivariate Cox proportional hazards regression. HRs and 95% confidence intervals (95% CI) are described. For the univariate log-rank test, a survival function was constructed using the Surv() function, with the time, event, and data parameters set as before. Each variable was entered into the "~predictors" parameter. The Survdiff() function was then used to incorporate each variable into the Surv() survival function, and a log-rank test was performed to obtain p-values. The rho parameter was set to 0, and the timefix parameter was set to TRUE. For the multivariate Cox proportional hazards regression survival analysis, the Surv() function is used to further construct the multivariate survival function, with the time, event, and data settings the same as in the previous steps. All independent variables are entered into the model. The coxph() function is then used to incorporate the Surv() survival function to perform the multivariate Cox proportional hazards regression survival analysis, obtaining the output value Pr(>|z|), which is the Cox regression p-value for each independent variable. Specifically, the singular.ok parameter is set to TRUE, the model parameter to FALSE, the x parameter to FALSE, the y parameter to TRUE, and the method parameter to ties. Among the three values ​​of "efron", "breslow", and "exact", ties is set to "efron," using Efron as the approximation as the default value because Efron is more accurate and computationally efficient in handling bound outcome times.

[0058] Conditional inference tree analysis (CTREE) is a machine learning method performed as a supplementary analysis to significant factors (probability values ​​<0.1 in multivariate Cox proportional hazards regression). It typically utilizes multiple significance tests and information measures (e.g., Gini coefficient) on the feature permutations of tree nodes to segment the most relevant predictors of the outcome and recursively splits the tree to divide patients into subsamples with varying degrees of severity risk. CTREE recursively performs univariate splitting of the dependent variable based on values ​​on a set of covariates. CTREE tends to select variables with many possible splits or many missing values, using significance testing procedures to select variables rather than those that maximize information measures (e.g., Gini coefficient). Specifically, the Surv() function is used to further construct the multivariate survival function, with the time, event, and data settings identical to the previous steps; eligible independent variables are added to the model; the ctree() function is used to incorporate the Surv() survival function, recursively splitting continuous, truncated, ordinal, nominal variables and multivariate response variables within the conditional inference framework to build a decision tree model. Specifically, the subset parameter is set to NULL, the weights parameter is set to NULL, the xtrafo parameter is set to ptrafo, the ytrafo parameter is set to ptrafo, and the scores parameter is set to NULL.

[0059] This invention also compared the prognostic characteristics and initial examination results of patients who presented within 4 days of symptom onset with those of other patients. A p-value < 0.05 was considered statistically significant, and statistical analysis was performed using R software (R version 4.0.0 https: / / www.r-project.org / ).

[0060] result

[0061] Demographic and clinical characteristics

[0062] A summary of demographic information for the entire cohort is shown in Table 1, and data on severity and non-severity are compared.

[0063] Table 1. Demographic and clinical characteristics of patients stratified by severity.

[0064]

[0065]

[0066] Data are expressed as mean ± standard deviation, or median with range, n(%), where n is the number of patients in the sample and % is the proportion of available data. COPD = Chronic Obstructive Lung Disease.

[0067] Of the 1590 cases, the median age was 48.0 years, and only 647 patients (40.6%) were female. Besides fever during or after hospitalization (88.0%), the most common symptom was dry cough (70.2%). 399 patients (25.1%) had at least one pre-existing chronic disease. Over 85% of patients had at least one abnormal chest CT or X-ray. Significant differences were also observed in diagnostic time and outcome-related time.

[0068] Characteristics and cutoff points of diagnostic timeliness or outcome-related duration

[0069] This study further analyzed and described the diagnostic timeliness and outcome-related duration (Table 2), and plotted a frequency density diagram. Figure 1 Of these 1,590 cases, the median durations from symptom onset to diagnosis, from symptom onset to first visit, from first visit to diagnosis, and from hospitalization to reaching severity were 6, 3, 2, and 8 days, respectively.

[0070] Table 2 Diagnostic time and result duration

[0071]

[0072] *In this cohort, there were a total of 1,590 cases, of which 1,501 had symptom onset dates, 1,512 had first visit dates, 935 had diagnosis dates, 1,246 had hospitalization dates, and 234 progressed to a severe stage with a definite date.

[0073] *Based on the logit statistical method, considering age, pre-existing conditions, and endpoint factors, the above four durations are completely randomized missing (MCAR).

[0074] To determine the cutoff value for duration, 350 observations were removed due to missing durations (missing admission dates or dates when the condition progressed to severity, or negative calculations). Furthermore, 11 cases (5%, 11 / 237) with days from admission to reaching severity were significantly outliers (far exceeding 1.5 times the difference between the third quartile plus the first and third quartiles, making their validity untraceable) and excluded. Therefore, a maximum of 1229 cases were retained in the Maxstat analysis to determine the optimal threshold.

[0075] Based on the best log-rank test p-value ( Figure 2 , Figure 3 The cutoff values ​​for the time between symptom onset and diagnosis, between symptom onset and first visit, and between first visit and diagnosis were 5, 4, and 3 days, respectively (Table 3).

[0076] Table 3 Cutoff values ​​and log-rank test results for diagnostic timeliness

[0077]

[0078] *<0.0005,**<0.0001

[0079] Prognostic analysis

[0080] Further prognostic analysis was performed on a total of 1229 cases with complete prognostic outcome parameters (admission date, endpoint date, endpoint outcome). Overall survival curves were plotted. Figure 4 Univariate log-rank analysis showed that age, sex, comorbidities, and time to diagnosis were suspected factors influencing COVID-19 prognosis (P<0.05, Table 4). After adjusting for age, sex, smoking status, and comorbidities, multivariate Cox proportional hazards regression analysis showed that age (hazard ratio [HR]: 1.03; 95% CI: 1.01–1.04), comorbidities (HR: 1.84; 95% CI: 1.23–2.73), and time from symptom onset to diagnosis (HR: 1.69; 95% CI: 1.10–2.60) were strong independent predictors of COVID-19 severity. When considering the time from symptom onset to first visit and from first visit to diagnosis, age (HR: 1.03; 95% CI: 1.01–1.04), comorbidity status (HR: 1.77; 95% CI: 1.18–2.68), and time from symptom onset to first visit (HR: 1.56; 95% CI: 1.07–2.26) were significantly associated with COVID-19 severity (Table 3).

[0081] Table 4 Univariate log-rank analysis and multivariate Cox proportional hazard regression analysis

[0082]

[0083]

[0084] ***=p<0.001, **=p<0.01,*=p<0.05, #=p<0.1

[0085] Further analysis using conditional inference trees revealed the relationship between prognostic outcomes and significant factors in multivariate Cox regression (P<0.1), identifying the risk threshold and its relationship with overall survival. A CTREE model considering age, chronic comorbidities, and time from symptom onset to diagnosis indicated that patients over 61 years of age with a history of chronic disease were more likely to reach severity. For patients without chronic comorbidities, those diagnosed >5 days after symptom onset and over 51 years of age also had a similar risk of reaching severity. Figure 5 a).

[0086] Another CTREE model, considering age, chronic comorbidities, time from symptom onset to first visit, and time from first visit to diagnosis, showed that patients without chronic comorbidities, diagnosed >3 days after first visit, and over 55 years of age had the worst prognosis. Patients with chronic comorbidities who sought medical attention >4 days after symptom onset ranked second. Figure 5 b). Both CTREE prognostic models showed that diagnosis-related time to diagnosis plays an important role in disease progression, especially in older individuals and those with more comorbidities.

[0087] Furthermore, by comparing the prognostic characteristics of patients who sought medical attention within 4 days of symptom onset with those of other patients, as well as the initial examination results upon admission, it was found that the incidence of ARDS was higher in patients who sought medical attention more than 4 days after symptom onset. Within the same group, there were more patients characterized by male sex, advanced age, dry cough, sputum production, shortness of breath, and COPD (P<0.05) (Table 5).

[0088] Table 5. Comparison of prognostic characteristics and initial examination results of patients who sought medical attention within 4 days of symptom onset with other patients.

[0089]

[0090]

[0091] Data are presented as mean ± standard deviation, or median with range, n (%), where n is the number of patients in the sample and % is the proportion of available data. ARDS = Acute Respiratory Distress Syndrome. DIC = Disseminated Intravascular Coagulation. COPD = Chronic Obstructive Pulmonary Disease.

[0092] This invention is the first in-depth study to explore the impact of diagnostic timeliness on prognosis and clinical characteristics in COVID-19 patients. In a comprehensive risk assessment and prognostic survival model, pre-hospital factors of COVID-19 patients, such as comorbidities, age, sex, and smoking status, were considered in conjunction with the time-related factors associated with diagnostic timeliness. This invention demonstrates that diagnostic timeliness plays a crucial role in the prognosis of COVID-19, especially in elderly patients and those with chronic comorbidities; patients with chronic obstructive pulmonary disease (COPD) are more likely to delay their first visit. According to this invention, reducing the total time between symptom onset and diagnosis to within 5 days is beneficial. Patients with chronic comorbidities are advised to seek medical attention within 4 days of symptom onset. Patients over 55 years of age should be diagnosed within 3 days of their first visit. Therefore, elderly patients and those with chronic comorbidities are more sensitive to diagnostic timeliness. Notably, COPD patients were significantly more numerous in the group with longer time from symptom onset to first visit, and COPD patients often present with symptoms such as dry cough, sputum production, and shortness of breath. This invention also identified more severe and worse prognoses or ARDS cases in the delayed visit group. Among laboratory-confirmed COVID-19 cases, patients with longer time intervals between symptom onset and diagnosis have poorer prognoses, especially those who delay seeking medical attention. Diagnostic timeliness is particularly important for older patients; delayed diagnosis (e.g., more than 5 days after symptom onset) may worsen their prognosis. Furthermore, for patients over 55 years of age, a diagnosis should be made as quickly as possible within 3 days. Patients with chronic comorbidities should seek medical attention within 4 days of symptom onset. Accurate classification should be determined by taking a more thorough medical history to establish risk stratification, identifying patients with delayed diagnosis and those more likely to have poor outcomes, especially elderly patients with comorbidities. Patients with COPD or any other respiratory illness should be more aware of changes in their symptoms, seek medical attention promptly, and receive better guidance.

Claims

1. A method for establishing a COVID-19 triage system based on diagnostic timeliness, the method comprising: (a) Collect information including age, chronic disease history, date of symptom onset, date of first visit, date of diagnosis, date of hospitalization, adverse outcomes, and earliest date of adverse outcome in individuals diagnosed with COVID-19; (b) Calculate the diagnostic time-related duration for each confirmed case in days, including the time from the date of symptom onset to the date of diagnosis, the time from the date of symptom onset to the date of first visit, and the time from the date of first visit to the date of diagnosis; calculate the time from the date of admission to the earliest date of adverse outcome, and exclude the case if the time is negative or greater than 31 days; (c) Determine the 95th percentile of the time relevant to the diagnostic time-out and exclude cases with abnormal data values ​​greater than the 95th percentile; (d) Using the duration of diagnosis-related time as the independent variable and survival data on whether adverse outcomes occurred and the duration from the date of admission to the earliest date of the adverse outcome as the dependent variable, the Maxstat maximum statistic method was used to determine the optimal cutoff value for converting the duration of diagnosis-related time into a binary variable, and the original continuous variable of the duration of diagnosis-related time was directly replaced with a binary variable. (e) Using the diagnostic time-related duration after binary classification conversion, age, gender, smoking history, and chronic disease history as independent variables, and the survival data of whether adverse outcomes occurred and the duration from the date of admission to the earliest date of occurrence of adverse outcomes as dependent variables, a univariate log-rank test was performed to obtain the log-rank test p-values ​​of the independent variables respectively. (f) Using the diagnostic time-related duration after binary classification conversion, variables with a p-value of log-rank test below the set threshold, age, gender, smoking history, and chronic disease history as independent variables, and survival data on whether adverse outcomes occurred and the duration from the date of admission to the earliest date of adverse outcome as dependent variables, multivariate Cox proportional hazards regression survival analysis was performed to obtain the Cox regression p-values ​​of each independent variable. (g) Using variables including the diagnostic time-related duration after binary classification conversion and Cox regression p-values ​​below a set threshold as independent variables, and survival data such as whether an adverse outcome occurred and the duration from admission date to the earliest occurrence date of the adverse outcome as dependent variables, conditional inference tree analysis was performed to obtain a decision tree model; and (h) Establish a triage system for COVID-19 based on the obtained decision tree model; In step (d), the Maxstat maximum statistic method is calculated using the surv_cutpoint() function in R programming. The minprop parameter is set to 0.15, the progressbar parameter is set to TRUE, the variables parameter is named "timeStoD", the data parameter is used to call the data frame structure dataset containing survival data and continuous variables to be segmented, the time parameter calls the survival time index, which is set to the time required from admission to the occurrence of adverse outcome, and the event parameter calls the survival outcome index, which is set to the status of whether the adverse outcome has occurred, with the event value set to 0 or 1.

2. The method according to claim 1, wherein for the Maxstat maximum statistic method in step (d), the surv_categorize() function is used to segment each variable value according to the cut point returned by the surv_cutpoint() function, the variable value of "timeStoD" is updated, and then the output cutpoint value, i.e. the optimal cut-off value, is obtained through the summary() function.

3. The method according to claim 2, wherein the validity of the optimal cutoff value is verified for the Maxstat maximum statistic method in step (d): a survival function is constructed using the Surv() function, with the settings of time, event, and data being the same as those in the surv_cutpoint() function; the "~predictors" parameter is taken from the "timeStoD" index generated by the surv_categorize() function and set to "~timeStoD"; the ggsurvplot() function is used to plot the survival curve for the surv() function, wherein the settings of the data function are not changed, the pval parameter is set to TRUE, the risk.table parameter is set to TRUE, the conf.int parameter is set to TRUE, the surv.median.line parameter is set to "hv", and the linetype parameter is set to "strata"; the break.time.by parameter is set to 5 to achieve the best plotting effect; the parameters of style themes or color panels such as ggtheme and palette are not limited; if the two sets of survival curves do not have an intersecting trend, then the binary cutoff point value of the diagnostic timeliness-related index is valid.

4. The method according to any one of claims 1 to 3, wherein the adverse outcome includes one or more of ICU, severe pneumonia, invasive ventilation, and death.

5. The method according to any one of claims 1 to 3, wherein the threshold values ​​for the log-rank test p-value and the Cox regression p-value are both 0.05 to 0.

1.

6. The method according to any one of claims 1 to 3, wherein the abnormal case is a case lacking one or more of the following information: age, history of chronic disease, date of symptom onset, date of first visit, date of diagnosis, date of admission, adverse outcome, and earliest date of occurrence of adverse outcome.

7. A computer system for COVID-19 triage based on diagnostic timeliness, characterized in that, include: The data receiving module is used to receive information including the patient's age, chronic disease history, symptom onset date, and first visit date. The data processing module is used to compare the received information with preset priority medical treatment conditions, which include one or more of the following: priority medical treatment, priority prescription, priority sampling, priority sample delivery, and priority testing. as well as The output module is used to highlight patients who meet the preset priority treatment conditions and prompt them to seek priority treatment. The preset priority medical treatment conditions are determined by the method according to any one of claims 1 to 6, and are selected from one or more of the following conditions: (a) Having a chronic disease and being over 61 years of age (b) Having a chronic illness and having a symptom onset date more than 4 days prior to the first visit, and (c) No chronic disease but older than 51 years.

8. The triage computer system according to claim 7, characterized in that, The output module is further configured to set priorities according to the following rules: (1) Prioritize patients with chronic diseases who are over 61 years of age; (2) Patients with chronic diseases whose symptom onset date is more than 4 days prior to their first visit are given the second priority. (3) Patients without chronic diseases but older than 55 years of age are given the third priority; and (4) Patients with no chronic diseases but older than 51 years of age are given the fourth priority.

9. The triage computer system according to claim 7, characterized in that, The output module is further configured to: when the patient has no chronic disease but is older than 51 years old, suggest providing the COVID-19 nucleic acid test result within 5 days from the onset of symptoms; or when the patient has no chronic disease but is older than 55 years old, suggest providing the COVID-19 nucleic acid test result within 3 days from the date of the first visit.

10. A COVID-19 triage method based on diagnostic timeliness, comprising: (S1) Receive information including the patient’s age, history of chronic diseases, date of symptom onset, and date of first visit; (S2) The received information is compared with the preset priority medical treatment conditions, which include one or more of the following: priority medical treatment, ... (S3) Highlight patients who meet the preset priority treatment conditions and prompt them to seek priority treatment; The preset priority medical treatment conditions are determined by the method according to any one of claims 1 to 6, and are selected from one or more of the following conditions: (a) Having a chronic disease and being over 61 years of age (b) Having a chronic illness and having a symptom onset date more than 4 days prior to the first visit, and (c) No chronic disease but older than 51 years.

11. A method for establishing a COVID-19 triage system based on diagnostic timeliness, the method comprising: (a) Collect information including age, chronic disease history, date of symptom onset, date of first visit, date of diagnosis, date of hospitalization, adverse outcomes, and earliest date of adverse outcome in individuals diagnosed with COVID-19; (b) Calculate the diagnostic time-related duration for each confirmed case in days, including the time from the date of symptom onset to the date of diagnosis, the time from the date of symptom onset to the date of first visit, and the time from the date of first visit to the date of diagnosis; calculate the time from the date of admission to the earliest date of adverse outcome, and exclude the case if the time is negative or greater than 31 days; (c) Determine the 95th percentile of the time relevant to the diagnostic time-out and exclude cases with abnormal data values ​​greater than the 95th percentile; (d) Using the duration of diagnosis-related time as the independent variable and survival data on whether adverse outcomes occurred and the duration from the date of admission to the earliest date of the adverse outcome as the dependent variable, the Maxstat maximum statistic method was used to determine the optimal cutoff value for converting the duration of diagnosis-related time into a binary variable, and the original continuous variable of the duration of diagnosis-related time was directly replaced with a binary variable. (e) Using the diagnostic time-related duration after binary classification conversion, age, gender, smoking history, and chronic disease history as independent variables, and the survival data of whether adverse outcomes occurred and the duration from the date of admission to the earliest date of occurrence of adverse outcomes as dependent variables, a univariate log-rank test was performed to obtain the log-rank test p-values ​​of the independent variables respectively. (f) Using the diagnostic time-related duration after binary classification conversion, variables with a p-value of log-rank test below the set threshold, age, gender, smoking history, and chronic disease history as independent variables, and survival data on whether adverse outcomes occurred and the duration from the date of admission to the earliest date of adverse outcome as dependent variables, multivariate Cox proportional hazards regression survival analysis was performed to obtain the Cox regression p-values ​​of each independent variable. (g) Using variables including the diagnostic time-related duration after binary classification conversion and Cox regression p-values ​​below a set threshold as independent variables, and survival data such as whether an adverse outcome occurred and the duration from admission date to the earliest occurrence date of the adverse outcome as dependent variables, conditional inference tree analysis was performed to obtain a decision tree model; and (h) Establish a triage system for COVID-19 based on the obtained decision tree model; For the univariate log-rank test in step (e), a survival function is constructed using the Surv() function. The data parameter is used to call a data frame structure dataset containing survival data and continuous variables to be segmented. The time parameter calls the survival time index, set as the time required from admission to the occurrence of an adverse outcome. The event parameter calls the survival outcome index, set as the status of whether an adverse outcome has occurred, with the event value set to 0 or 1. Each variable is entered into the "~predictors" parameter one by one. The Survdiff() function is used to incorporate each variable into the Surv() function, and a log-rank test is performed on each variable to obtain the p-value. The rho parameter is set to 0, and the timefix parameter is set to TRUE.

12. The method of claim 11, wherein the adverse outcome includes one or more of ICU, severe pneumonia, invasive ventilation, and death.

13. The method according to claim 11, wherein the threshold values ​​for the log-rank test p-value and the Cox regression p-value are both 0.05 to 0.

1.

14. The method of claim 11, wherein the abnormal case is a case lacking one or more of the following information: age, history of chronic disease, date of symptom onset, date of first visit, date of diagnosis, date of admission, adverse outcome, and earliest date of occurrence of adverse outcome.

15. A computer system for COVID-19 triage based on diagnostic timeliness, characterized in that, include: The data receiving module is used to receive information including the patient's age, chronic disease history, symptom onset date, and first visit date. The data processing module is used to compare the received information with preset priority medical treatment conditions, which include one or more of the following: priority medical treatment, priority prescription, priority sampling, priority sample delivery, and priority testing. as well as An output module is used to highlight patients who meet the preset priority treatment criteria and prompt them to seek priority treatment; wherein the preset priority treatment criteria are determined by the method according to any one of claims 11 to 14, and are selected from one or more of the following conditions: (a) Having a chronic disease and being over 61 years of age (b) Having a chronic illness and having a symptom onset date more than 4 days prior to the first visit, and (c) No chronic disease but older than 51 years.

16. The triage computer system according to claim 15, characterized in that, The output module is further configured to set priorities according to the following rules: (1) Prioritize patients with chronic diseases who are over 61 years of age; (2) Patients with chronic diseases whose symptom onset date is more than 4 days prior to their first visit are given the second priority. (3) Patients without chronic diseases but older than 55 years of age are given the third priority; and (4) Patients with no chronic diseases but older than 51 years of age are given the fourth priority.

17. The triage computer system according to claim 15, characterized in that, The output module is further configured to: when the patient has no chronic disease but is older than 51 years old, suggest providing the COVID-19 nucleic acid test result within 5 days from the onset of symptoms; or when the patient has no chronic disease but is older than 55 years old, suggest providing the COVID-19 nucleic acid test result within 3 days from the date of the first visit.

18. The triage computer system according to claim 15, characterized in that, The chronic disease history includes one or more of the following: COPD, diabetes, hypertension, coronary heart disease, cerebrovascular disease, hepatitis B, tumors, chronic kidney disease, and immunodeficiency.

19. The triage computer system according to claim 15, characterized in that, The symptoms include one or more of the following: dry cough, sore throat, conjunctival congestion, nasal congestion, headache, sputum production, fatigue, hemoptysis, shortness of breath, nausea / vomiting, diarrhea, joint and muscle pain, chills, and fever.

20. A COVID-19 triage method based on diagnostic timeliness, comprising: (S1) Receive information including the patient’s age, history of chronic diseases, date of symptom onset, and date of first visit; (S2) The received information is compared with the preset priority medical treatment conditions, which include one or more of the following: priority medical treatment, ... (S3) Highlight patients who meet the preset priority treatment conditions and prompt them to seek priority treatment; The preset priority medical treatment conditions are determined by the method according to any one of claims 11 to 14, and are selected from one or more of the following conditions: (a) Having a chronic disease and being over 61 years of age (b) Having a chronic illness and having a symptom onset date more than 4 days prior to the first visit, and (c) No chronic disease but older than 51 years.

Images