Study population:
The current study draws on a unique hospital information system (HIS) established in Chiang Mai, located in Northern Thailand, with a population of 1.6 million. Adults with laboratory confirmed COVID-19 between 1 February – 31 July 2022 were included in the study. To understand the epidemiology of circulating SARS-CoV-2 variants, the ministry of health performs molecular testing on a random sample of COVID-19 positive cases every month. Test results from northern Thailand revealed that BA.1 omicron sub-lineage accounted for >90% of the cases in February, with cases of BA.2 appearing in March and BA.4/BA.5 appearing from mid-June onwards. Subsequently, BA.4/BA.5 spread rapidly, dominating through July to >90% of cases by mid-August [2].
Non-Thai residents and migrants were excluded from the study as the vaccination data and outcome capture for this group may be incomplete. Cases with missing age were also excluded. The patient selection flow is presented under Supplementary Figure 1.
Data Sources:
We have previously published the details on creating and implementing the information systems used in this study [21]. In brief, all COVID-19 cases detected in Chiang Mai province are reported into the web based HIS of Chiang Mai Provincial Health Office (CMC-19 HIS). During the study period, reporting of all COVID-19 cases was mandatory under the Communicable Disease Control Act. When a COVID-19 case is detected, either at screening centers, hospitals or out-patient clinics, the healthcare staff enter the patient details, including laboratory results into the CMC-19 HIS under a unique ID. Data on severity and progression of the disease including requirement of ventilatory support and treatments are recorded in each hospital’s information system, which is linked with CMC-19 HIS. Deaths which occur within the province are reported to Chiang Mai Provincial Health Office and are routinely updated in CMC-19 HIS.
All national vaccination records are centrally captured in the Ministry of Public Health Immunization Center (MOPH IC) database maintained by the Ministry of Public Health, Thailand.
Ethical Approval Statement:
The study was conducted on routine data collected as part of the national COVID-19 response under the Communicable Disease ACT (B.E. 2558) and was exempted from ethics review.
Study Design:
We conducted a retrospective cohort study on Thai residents aged 18 years or older, with a laboratory confirmed SARS-CoV-2 infection during 1 February – 31 July 2022 period. Date of first positive SARS-CoV-2 test served as the index date. Reinfections, defined as a positive SARS-CoV-2 test at least 90 days prior, accounted for <0.5% of this cohort.
Baseline clinical characteristics and SARS-CoV-2 test details were extracted from the CMC-19 HIS. The types of COVID-19 vaccines, and dates of vaccinations were extracted from MOPH-IC immunization database.
Severe COVID-19 outcome was defined as requiring Invasive Mechanical Ventilation (IMV) during hospital admission or death within 30 days of positive SARS-CoV-2 test. Records of all included subjects were followed till death, or up to 30 days from first positive test, whichever was earlier. The severe outcome capture for the study population is near complete as the clinical information of all hospitalised COVID-19 cases of the 26 public and 8 private hospitals in Chiang Mai province, including the only two tertiary care referral hospitals providing IMV support in Chiang Mai, are entered into a single CMC-19 HIS platform.
Statistical Analysis:
Descriptive statistics are reported separately for the subjects with and without severe COVID-19 to understand the differences in baseline characteristics between the groups. Continuous variables are summarized as mean and standard deviation (SD) for normally distributed data or median and interquartile range (IQR) for skewed data. Categorical variables are summarized as frequency and percentages. Between group comparisons were done using Mann–Whitney U test or t-test for continuous variables and Chi-squared test or Fisher’s exact test for categorical variables, as appropriate.
Cox proportional hazards regression was used to estimate hazard ratios (HRs) for severe COVID-19 and mortality outcomes. Follow up period was taken from the first positive SARS-CoV-2 test date and censored at the earliest of: date of first starting IMV, date of death or 30 days from first positive test date. If the outcome occurred on the first positive SARS-CoV-2 test date, the follow-up time was taken to be 0.5 days. Age, gender, calendar day of test (in weekly units), vaccination status, and time since last vaccination were added as factors in the regression model to estimate adjusted HRs (95% CI) for severe COVID-19 and mortality outcomes. The vaccination status was categorised as partially vaccinated (received only one dose), primary series (received two recommended doses), three-dose (received primary series and one booster dose), and four-dose or more (received primary series and two or more booster doses) with the unvaccinated group serving as reference group. To examine waning of vaccine response, subjects who were vaccinated were sub-grouped using time specific indicators defined by two-month intervals of time since vaccination (≤14 days, 14 to 60 days, >60 to 120 days, >120 to 180 days and >180 days) for each vaccine status. Due to limited number of events, subjects who received four-dose or more were combined with the three-dose group during analysis (three-dose or more), and a separate sensitivity analysis was done after excluding subjects who received four-doses or more. Separate analyses were done to examine if the waning vaccine response differed by age groups and booster vaccine types.
All statistical analyses were be conducted using stata (version 15.0 SE, College station, TX:StataCorp LP). Significance tests are 2 sided and a p-value <0.05 was considered statistically significant.