Dataset Information

Establishing Classifiers With Clinical Laboratory Indicators to Distinguish COVID-19 From Community-Acquired Pneumonia: Retrospective Cohort Study.

ABSTRACT:

Background

The initial symptoms of patients with COVID-19 are very much like those of patients with community-acquired pneumonia (CAP); it is difficult to distinguish COVID-19 from CAP with clinical symptoms and imaging examination.

Objective

The objective of our study was to construct an effective model for the early identification of COVID-19 that would also distinguish it from CAP.

Methods

The clinical laboratory indicators (CLIs) of 61 COVID-19 patients and 60 CAP patients were analyzed retrospectively. Random combinations of various CLIs (ie, CLI combinations) were utilized to establish COVID-19 versus CAP classifiers with machine learning algorithms, including random forest classifier (RFC), logistic regression classifier, and gradient boosting classifier (GBC). The performance of the classifiers was assessed by calculating the area under the receiver operating characteristic curve (AUROC) and recall rate in COVID-19 prediction using the test data set.

Results

The classifiers that were constructed with three algorithms from 43 CLI combinations showed high performance (recall rate >0.9 and AUROC >0.85) in COVID-19 prediction for the test data set. Among the high-performance classifiers, several CLIs showed a high usage rate; these included procalcitonin (PCT), mean corpuscular hemoglobin concentration (MCHC), uric acid, albumin, albumin to globulin ratio (AGR), neutrophil count, red blood cell (RBC) count, monocyte count, basophil count, and white blood cell (WBC) count. They also had high feature importance except for basophil count. The feature combination (FC) of PCT, AGR, uric acid, WBC count, neutrophil count, basophil count, RBC count, and MCHC was the representative one among the nine FCs used to construct the classifiers with an AUROC equal to 1.0 when using the RFC or GBC algorithms. Replacing any CLI in these FCs would lead to a significant reduction in the performance of the classifiers that were built with them.

Conclusions

The classifiers constructed with only a few specific CLIs could efficiently distinguish COVID-19 from CAP, which could help clinicians perform early isolation and centralized management of COVID-19 patients.

SUBMITTER: Dai W

PROVIDER: S-EPMC7901596 | biostudies-literature | 2021 Feb

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Establishing Classifiers With Clinical Laboratory Indicators to Distinguish COVID-19 From Community-Acquired Pneumonia: Retrospective Cohort Study.

Dai Wanfa W Ke Pei-Feng PF Li Zhen-Zhen ZZ Zhuang Qi-Zhen QZ Huang Wei W Wang Yi Y Xiong Yujuan Y Huang Xian-Zhang XZ

Journal of medical Internet research 20210222 2

<h4>Background</h4>The initial symptoms of patients with COVID-19 are very much like those of patients with community-acquired pneumonia (CAP); it is difficult to distinguish COVID-19 from CAP with clinical symptoms and imaging examination.<h4>Objective</h4>The objective of our study was to construct an effective model for the early identification of COVID-19 that would also distinguish it from CAP.<h4>Methods</h4>The clinical laboratory indicators (CLIs) of 61 COVID-19 patients and 60 CAP patie ...[more]

PMID: 33534722

Similar Datasets

Project description:BackgroundThis retrospective cohort study is the first in North America to examine population-level appropriate antibiotic use for community-acquired pneumonia (CAP) in older adults, by agent, dose and duration. With the highest rates of CAP reported in the elderly populations, appropriate antibiotic use is essential to improve clinical outcomes. Given the ongoing crisis of antimicrobial resistance, understanding inappropriate antibiotic prescribing is integral to direct community stewardship efforts.MethodsAll outpatient primary care visits for CAP (aged ≥65 years) were identified using physician billing codes between January 1 2014 to December 31 2018 in British Columbia (BC) and Ontario (ON). Categories of prescribing were derived from existing literature, and constructed for clinical relevance using Canadian and international guidelines available during the study period. Categories were mutually exclusive and included: guideline adherent (first-line agent, adherent dose/duration), clinically appropriate (non-first line agent, presence of comorbidities), effective but unnecessary (first-line agent, excess dose/duration), undertreatment (first-line agent, subtherapeutic dose/duration), and not recommended (non-first line agent, absence of comorbidities). Proportions of prescribing were examined by category. Temporal trends in prescribing were examined using Poisson regression.ResultsA total of 436,441 episodes of CAP were identified, with 46% prescribed an antibiotic in BC, and 52% in Ontario. Guideline adherent prescribing was minimal for both provinces (BC: 2%; ON: 1%) however the largest magnitude of increase was reported in this category by the final study year (BC-Rate Ratio [RR]: 3.4, 95% Confidence Interval [CI]: 2.7-4.3; ON-RR: 4.62, 95% CI: 3.4-6.5). Clinically appropriate prescribing accounted for the most antibiotics issued, across all study years (BC: 61%; ON: 74%) (BC-RR: 0.8, 95% CI: 0.8-0.8; ON-RR: 0.9, 95% CI: 0.8-0.9). Excess duration of therapy was the hallmark characteristic for effective but unnecessary prescribing (BC: 92%; ON: 99%). The most common duration prescribed was 7 days, followed by 10. Not recommended prescribing was minimal in both provinces (BC: 4%; ON: 7%) and remained stable by the final study year (BC-RR: 1.1, 95% CI: 0.9-1.2; ON-RR: 0.9, 95% CI: 0.9-1.1).ConclusionThree quarters of antibiotic prescribing for CAP was appropriate in Ontario, but only two thirds in BC. Shortening durations-in line with evidence for 3 to 5-day treatment presents a focused target for stewardship efforts.

Project description:BackgroundFrailty is associated with poor prognosis in a wide range of illnesses. However, its prognostic implications for older patients with community-acquired pneumonia (CAP) are not adequately addressed.MethodsIn this study, patients were classified into 3 groups according to the frailty index based on standard laboratory tests (FI-Lab) score: robust (FI-Lab < 0.2), pre-frail (FI-Lab 0.2-0.35), and frail (FI-Lab ≥ 0.35). The relationships between frailty and all-cause mortality and short-term clinical outcomes (length of stay, duration of antibiotic therapy, in-hospital mortality) were examined.ResultsFinally, 1164 patients were included, the median age was 75 years (interquartile range: 69, 82), and 438 patients (37.6%) were women. According to FI-Lab, 261(22.4%), 395(33.9%), and 508(43.6%) were robust, pre-frail, and frail. After adjustment for confounding variables, frailty was independently associated with prolonged antibiotic treatment (p = 0.037); pre-frailty and frailty were independently associated with longer inpatient days (p < 0.05 for both). The risk of in-hospital mortality was independently increased in frail patients (HR = 5.01, 95% CI = 1.51-16.57, p = 0.008) but not pre-frail patients (HR = 2.87, 95% CI = 0.86-9.63, p = 0.088) compared to robust patients. During a median follow-up of 33.9 months (interquartile range: 32.8 to 35.1 months), 408 (35.1%) patients died, of whom 29 (7.1%) were robust, 112 (27.5%) were pre-frail, and 267 (65.9%) were frail. Compared to robust patients, frail and pre-frail were significantly associated with increased risk for all-cause death (HR = 4.29, 95%CI: 1.78-10.35 and HR = 2.42 95%CI: 1.01-5.82, respectively).ConclusionsFrailty is common among older patients with CAP and is strongly associated with increased mortality, longer length of stay, and duration of antibiotics. A routine frail assessment at the admission of elderly patients with CAP is necessary as the first step for appropriate multidisciplinary interventions.

Dataset Information

Establishing Classifiers With Clinical Laboratory Indicators to Distinguish COVID-19 From Community-Acquired Pneumonia: Retrospective Cohort Study.

Background

Objective

Methods

Results

Conclusions

Publications

Establishing Classifiers With Clinical Laboratory Indicators to Distinguish COVID-19 From Community-Acquired Pneumonia: Retrospective Cohort Study.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets