How to Read: Clinical Characteristics of Coronavirus Disease 2019 in China [Reading Guide]

This guide will help you efficiently read "Clinical Characteristics of Coronavirus Disease 2019 in China" published in the New England Journal of Medicine in 2020, using the three-pass approach.

First pass (5–10 min)

What to skim in this paper:

• Title: "Clinical Characteristics of Coronavirus Disease 2019 in China" – Immediately tells you the subject and geographic focus.
• Abstract (Page 1): Read this carefully. It provides a concise summary of the background, methods, key results (patient demographics, primary endpoint, common symptoms, incubation period, radiologic findings, lab findings), and conclusions.
• Introduction (BACKGROUND, Page 1-2): Skim the first few paragraphs to understand why the study was conducted – the emergence and rapid spread of Covid-19 and the need for data on clinical characteristics.
• Conclusion (Page 1 and DISCUSSION, Page 5): Read the final paragraph of the Abstract and the "CONCLUSIONS" section on page 1. Then, quickly read the first and last paragraphs of the "DISCUSSION" section on page 5 to grasp the main takeaways and the authors' interpretation.

What to notice:

• Problem: The paper addresses the urgent need for data on the clinical characteristics of Covid-19 patients during the initial outbreak in China.
• Setting: Data was collected from 1099 laboratory-confirmed patients across 552 hospitals in 30 provinces of mainland China, covering December 2019 through January 2020 – a crucial early snapshot of the pandemic.
• Contributions: This is one of the earliest large-scale studies providing comprehensive clinical details, including demographics, symptoms, radiologic findings, laboratory results, and outcomes for Covid-19 patients. Key findings include the median age (47 years), the primary composite end point (ICU admission, mechanical ventilation, or death) occurring in 6.1% of patients, and the observation that patients often presented without fever or abnormal radiologic findings.

Second pass (up to 1 hour)

Using SpatialRead in the Second Pass

• Use the Screenshot tool to capture figures, tables, diagrams, and formulae from the PDF and use AI for simple explanations.
• Ask: "Explain Figure X at a high level", "Compare Table 2 vs 3", or "What does the y-axis represent here?"
• Download SpatialRead →

What sections/figures to focus on:

• METHODS (Page 2-3):
  • Study Oversight: Note the funding and ethical approval.
  • Data Sources: Understand the patient cohort (1099 laboratory-confirmed cases) and the time frame of data collection.
  • Study Outcomes: Clearly identify the primary composite end point (admission to ICU, mechanical ventilation, or death).
  • Study Definitions: Pay attention to how key terms like "fever," "lymphocytopenia," and "severe vs. nonsevere disease" were defined. This is critical for interpreting results.
  • Laboratory Confirmation: Note that diagnosis was based on RT-PCR.
• RESULTS (Page 3-5):
  • DEMOGRAPHIC AND CLINICAL CHARACTERISTICS: Read this section to understand the patient population and initial presentation.
  • Figure 1 (Page 4): Examine the map showing the distribution of patients across mainland China. This visualizes the rapid spread and concentration of cases.
  • Table 1 (Page 6-7): This is a core table. Focus on the "All Patients" column for overall characteristics. Then, compare characteristics between "Nonsevere" and "Severe" disease, and between "Yes" and "No" for the primary composite end point. Notice differences in age, sex, exposure history, symptoms (e.g., fever on admission vs. during hospitalization), and coexisting disorders.
  • RADIOLOGIC AND LABORATORY FINDINGS: Read this section for objective measures of disease.
  • Table 2 (Page 8-9): Another crucial table. Analyze the radiologic findings (e.g., ground-glass opacity on CT) and laboratory abnormalities (e.g., lymphocytopenia, D-dimer levels). Compare these findings across disease severity and outcomes.
  • CLINICAL OUTCOMES and TREATMENT AND COMPLICATIONS: Read these sections to understand the disease progression and management.
  • Table 3 (Page 10-11): Review the complications (e.g., ARDS, septic shock), treatments (e.g., antibiotics, mechanical ventilation), and clinical outcomes (e.g., length of hospital stay, death rates) presented here. Again, compare across severity and outcome groups.

Key terms/questions to note:

• What was the median age and gender distribution of the patients? (Table 1)
• What were the most common symptoms on admission and during hospitalization? (Table 1)
• What was the median incubation period? (Table 1)
• How frequently did patients present without fever or radiologic abnormalities? (Abstract, RESULTS, DISCUSSION)
• What were the key differences in demographics, symptoms, and lab/radiologic findings between severe and nonsevere cases? (Tables 1 and 2)
• What was the overall case fatality rate in this cohort? (Table 3)
• What were the most common treatments administered? (Table 3)

Third pass (1–5 hours)

Using SpatialRead in the Third Pass

• Highlight complex keywords, dense paragraphs, equations, or proofs and ask for a plain-English explanation that preserves notation.
• Use AI to explain the concepts in a way that is easy to understand.
• Ask for alternative phrasings or analogies to confirm understanding without changing technical meaning.
• Connect concepts with each other using AI to build a comprehensive understanding of the paper.
• Download SpatialRead →

Deep reading guidance:

• Methods (Page 2-3): Critically evaluate the study design. Given the early stage of the pandemic, consider the challenges of data collection across 552 hospitals. Pay attention to the definitions of "severe" and "nonsevere" disease (American Thoracic Society guidelines for community-acquired pneumonia) and the primary composite end point. Understand that the statistical analysis is descriptive, meaning findings are specific to this cohort and not necessarily generalizable without further studies.
• Results (Page 3-5) and Tables 1, 2, 3: Go through each table row by row, focusing on the percentages and interquartile ranges (IQR). Look for statistically significant differences (though the paper states all statistics are descriptive, the magnitude of differences can still be informative). For example, in Table 1, note the higher median age and prevalence of coexisting illnesses in patients with severe disease or those who met the primary composite end point. In Table 2, observe the higher rates of bilateral patchy shadowing and lymphocytopenia in severe cases. In Table 3, analyze the higher incidence of ARDS and the greater use of mechanical ventilation in severe cases.
• Discussion (Page 5): Read this section thoroughly. The authors interpret their findings, compare them to other coronaviruses (SARS-CoV, MERS-CoV), and discuss the implications for diagnosis and management.
• Limitations (Page 11): This is a critical section. The authors acknowledge several limitations, including incomplete documentation, recall bias for incubation period, censoring of outcomes for patients still hospitalized, and potential underestimation of mild/asymptomatic cases. These limitations are crucial for understanding the scope and generalizability of the study's findings. For instance, the study cohort likely represents the more severe end of the Covid-19 spectrum, meaning the reported case fatality rate might be higher than the true population rate.

Replication pointers if applicable:

• This is a retrospective observational study, so direct "replication" in the experimental sense isn't applicable. However, the methodology for data collection and patient classification is well-described, allowing for similar cohort studies to be conducted in other regions or at different stages of the pandemic.
• The clear definitions for "laboratory-confirmed Covid-19" (RT-PCR) and "severe disease" (American Thoracic Society guidelines) are key for ensuring consistency if one were to conduct a similar study.
• The paper highlights the challenges of data collection during an urgent outbreak, including varying electronic database structures and incomplete documentation. Future studies aiming for similar large-scale data collection should consider these practical difficulties and implement robust data standardization protocols.