1/
Feeling overwhelmed by the #COVID19 #SARSCoV2 literature? Me too.
We’re in this together. I’ll show you #HowIReadThisPaper and my appraisal. Add, comment, and correct me where I am wrong.
First up: nejm.org/doi/full/10.10…
(Thread) @BostonChiefs @Curbsiders
Feeling overwhelmed by the #COVID19 #SARSCoV2 literature? Me too.
We’re in this together. I’ll show you #HowIReadThisPaper and my appraisal. Add, comment, and correct me where I am wrong.
First up: nejm.org/doi/full/10.10…
(Thread) @BostonChiefs @Curbsiders
2/
Question: What were the clinical features of #COVID19 among patients hospitalized in China in the first 2 months of the outbreak?
Date published: 2/28/2020
Funding: National Health Commission of China
Question: What were the clinical features of #COVID19 among patients hospitalized in China in the first 2 months of the outbreak?
Date published: 2/28/2020
Funding: National Health Commission of China
3/
Study Design: Case series
Population: 1,099 patients hospitalized in China with RT-PCR confirmed #SARSCoV2 (nasal or pharyngeal swab)
Study Period: 12/11/2019 - 1/29/2020
Study Design: Case series
Population: 1,099 patients hospitalized in China with RT-PCR confirmed #SARSCoV2 (nasal or pharyngeal swab)
Study Period: 12/11/2019 - 1/29/2020
4/
Inclusion criteria: Only patients with a positive #SARSCoV2 PCR were included
Exclusion criteria: None specified
Inclusion criteria: Only patients with a positive #SARSCoV2 PCR were included
Exclusion criteria: None specified
5/
Outcomes reported: Demographic, clinical, lab, radiographic, hospitalization, and treatment data, including % of patients who [required ICU care, intubation, or died] [primary outcome].
Duration of follow-up: up to 2 days after the study period
Loss to follow-up: 0/1,099
Outcomes reported: Demographic, clinical, lab, radiographic, hospitalization, and treatment data, including % of patients who [required ICU care, intubation, or died] [primary outcome].
Duration of follow-up: up to 2 days after the study period
Loss to follow-up: 0/1,099
6/
Results:
Table 1: M >> F, median age=47; fever on admit=44%, fever during LOS=89%, median incubation period=4d [IQR 2-7d]
Sx: cough=68%, fatigue=38%, sputum=34%, SOB=19%, N/V=5%, diarrhea=4%; any comorbidity=24%
Results:
Table 1: M >> F, median age=47; fever on admit=44%, fever during LOS=89%, median incubation period=4d [IQR 2-7d]
Sx: cough=68%, fatigue=38%, sputum=34%, SOB=19%, N/V=5%, diarrhea=4%; any comorbidity=24%
7/
Table 2: Any abnl CXR finding=59%, any abnl CT finding=86%, lymphocytes <1,500=83%
CXR normal in 41%; CT normal in 14%.
Table 3: O2 req=41%, ICU admit=5%, septic shock=1.1%, ARDS=3.4%, intubated=2.3%, death=1.4%, pts who remained hospitalized at study end=94%, median LOS=12d
Table 2: Any abnl CXR finding=59%, any abnl CT finding=86%, lymphocytes <1,500=83%
CXR normal in 41%; CT normal in 14%.
Table 3: O2 req=41%, ICU admit=5%, septic shock=1.1%, ARDS=3.4%, intubated=2.3%, death=1.4%, pts who remained hospitalized at study end=94%, median LOS=12d
8/
Conclusions: During the first 7 weeks of the outbreak, among 1,099 hospitalized patients, 5.0% of patients required ICU care, 2.3% were intubated, and 1.4% died.
Conclusions: During the first 7 weeks of the outbreak, among 1,099 hospitalized patients, 5.0% of patients required ICU care, 2.3% were intubated, and 1.4% died.
9/
My appraisal:
Data collection was NOT systematic → leads to ascertainment bias.
If only the sickest pts get CT scans, this introduces bias since their contribution depends on severity. Look at the denominators to tell (e.g. incubation period estimated in only ~25% of pts).
My appraisal:
Data collection was NOT systematic → leads to ascertainment bias.
If only the sickest pts get CT scans, this introduces bias since their contribution depends on severity. Look at the denominators to tell (e.g. incubation period estimated in only ~25% of pts).
10/
The most common Sx (cough) was only seen in ⅔ of patients = prepare for protean manifestations.
75% of patients had NO underlying comorbidities.
→ Symptom-based case definitions and risk stratification based on comorbidities will be inadequate and will miss cases.
The most common Sx (cough) was only seen in ⅔ of patients = prepare for protean manifestations.
75% of patients had NO underlying comorbidities.
→ Symptom-based case definitions and risk stratification based on comorbidities will be inadequate and will miss cases.
11/
Almost all of the patients (94%) remained hospitalized at the study end, so the duration of follow-up was likely not long enough to capture later complications, including deaths attributable to #COVID19.
Almost all of the patients (94%) remained hospitalized at the study end, so the duration of follow-up was likely not long enough to capture later complications, including deaths attributable to #COVID19.
12/
Estimates of case fatality rate (CFR) biased upwards due to unknown denominator of total cases (ppl w/ mild infection do not seek care) and biased downwards due to delayed reporting of deaths. Net effect unclear, but prepare for CFR to change.
Estimates of case fatality rate (CFR) biased upwards due to unknown denominator of total cases (ppl w/ mild infection do not seek care) and biased downwards due to delayed reporting of deaths. Net effect unclear, but prepare for CFR to change.
13/
Bottom Line:
During the first 7 weeks of #COVID19 in China, among 1,099 hospitalized patients, 5% required ICU care, 2.3% required intubation, and 1.4% died. 94% of pts remained hospitalized at the end of the study period; longer follow-up needed.
(End)
Bottom Line:
During the first 7 weeks of #COVID19 in China, among 1,099 hospitalized patients, 5% required ICU care, 2.3% required intubation, and 1.4% died. 94% of pts remained hospitalized at the end of the study period; longer follow-up needed.
(End)
