Community Acquired Pneumonia
Description
This data set is from two public (one each secondary and tertiary) and two private (one secondary and one tertiary) health care facilities at Delhi, Kolkata, Pune and Chennai. All patients aged ≥60 years admitted in medicine, pulmonary and geriatric wards of the selected facilities between 4th December 2018 to 20th March 2020 were screened using an operational definition modified from 2009 update of British Thoracic Society (BTS) guidelines , described as new onset cough within last 10 days along with at least one lower respiratory tract symptom (dyspnea, chest pain), and at least one systemic feature (sweating, fevers, shivers, aches and pains and/or temperature of 38° C or more) and tachypnea with a respiratory rate> 20/min or a physician diagnosed pneumonia. Patients who had been hospitalized for > 48 hours in the hospital or any other facility for the same episode of illness were not enrolled to exclude possible nosocomial illnesses. At the time of enrollment, trained project nurses collected data about age, sex, influenza vaccination in past one year, history of current smoking and use of alcohol and pre-existing chronic morbidity by interviewing participants. Details of clinical assessment, diagnostic tests results of hemogram, blood urea, blood glucose, X-ray and use of medications were gathered by reviewing medical records using a standard data collection tool on open data kit (ODK) modules on handheld tablets. The enrolled participants' nasal and oropharyngeal specimens were collected within 48 hours of admission using the standard protocol. These combined swabs were placed immediately into viral transport media (VTM) on ice or ice pack, triple-sealed for transportation. All the specimens were transported to the respective virology laboratory of the institute on the same day. Viral RNA was extracted using the QiAmp Viral RNA kit (Qiagen, Germany) as per the manufacturer’s protocols. We used U.S. Centers for Disease Control and Prevention (CDC) approved one-step Real Time-PCR protocol and primers and probes for influenza and RSV virus to detect the influenza (A and B) and RSV virus. All samples were also tested for human metapneumovirus (hMPV), parainfluenza viruses (PIV), rhinoviruses and adenoviruses using a multiplex kit developed by National institute of Virology .The study team visited the hospitals every alternate day to enroll new patients and evaluate the existing ones till discharge. Any change in status of the patient or admission to the ICU was noted. After discharge, participants were followed up telephonically on day 7 and day 30 from the date of discharge to collect data on their outcome status. During telephonic follow-up, if a participant could not be contacted on first attempt, two more attempts were made to contact within next 7 days, before categorizing a participant as lost to follow-up. The outcomes were categorized as alive and at home, alive and readmitted, dead, or lost to follow-up.
Files
Steps to reproduce
Participants needing ICU care were those who were either admitted to ICU or referred to other facility for ICU admission. Mortality was recorded as deaths during hospitalization or within 30 days after hospital discharge. The proportion of participants with adverse clinical outcome were compared using chi-squared test by age groups, sex, viral RNA detection, self-reported co-morbid conditions, type of hospital (e.g., private vs. public), clinical and laboratory parameters at the time of admission. We categorized the clinical and laboratory parameters as described in pneumonia severity index, CURB-65 scoring (CURB-65 score evaluate the clinical severity of CAP and the acronym of "CURB-65" stands for C: Confusion or mental status altered; U: increased Urea (Blood Urea Nitrogen) level; R: high Respiratory Rate; B: low Blood Pressure, 65: age ≥65), and WHO classification of anaemia [19–21]. To identify factors associated with ICU admission we estimated adjusted odds ratio (adj OR) and their 95% confidence intervals using multivariable logistic regression with a hierarchical framework. We also applied the same hierarchical framework to identify factors associated with mortality by estimating hazard ratio (HR) using multivariable Cox proportional hazard regression models. For the hierarchical framework we categorized the independent variables into four levels (Table 4). The most distal level in hierarchal framework or level-1 (L1) included baseline characteristics of participants like age group, sex, smoking status, comorbid conditions (chronic respiratory disease, cardiovascular disease including hypertension, diabetes). Level 2 (L2) variables included etiological agents (influenza or RSV). Level 3 (L3) included clinical and laboratory condition at the time of admission including mental status (altered sensorium vs conscious & oriented), high respiratory rate (>=30 per minute), low blood pressure (Systolic blood pressure <90mmHg or Diastolic BP<=60mm Hg), low oxygen saturation (<90% without oxygen), anemia (hemoglobin level <12gm/dl for females and <13gm/dl for males), total leucocyte count(<5000/dl or >20,000/dl), elevated blood urea nitrogen (>19mg/dl). The most proximal or Level 4 (L4) variables included type of care received including ICU admission, mechanical ventilation, oseltamivir therapy and type of hospital (i.e., private vs. public). Regression was conducted using variables at each level (model 1-4) separately to identify variables with p value<0.1. Subsequently, regression was conducted with selected L1, L2 variables (p value <0.1) (model 5); L1, L2, L3 variable (model 6); and L1, L2, L3, L4 variable (model 7).