Point prevalence of SARS-CoV-2 infection in Sweden at six time points during 2020 | BMC Infectious Diseases
Survey design and participants
Between March and December 2020, six population-based cross-sectional surveys (surveys 1–6) have been conducted in Sweden.
Survey 1 was conducted in the Stockholm region from March 26 to April 3. Surveys 2 to 6 were conducted at the national level during the following time periods: 21–24 April, 25–28 May, 24–28 August, 21–25 September and 30 November–4 December.
Participants in surveys 1 to 5 were recruited from an already existing web panel that is regularly used for health-related questionnaires at the Public Health Agency. The web panel was built in 2015 from a simple random sample of 35,000 individuals from the general Swedish population and in 2016 (n = 15,000) and 2018 (n = 25,000) top samples were taken according to age, gender and education [7]. At the time of the first survey, the web panel included 4,500 individuals.
Survey 6 was designed as an age-stratified random sample from the general population aged 16 and over. The estimated sample sizes for the age groups 16–29 years, 30–59 years, and 60 years and older were 2486, 1257, and 1700, respectively, assuming a point prevalence of 2.0%, 1.0%, and 0.4% and a precision of 0.6%, 0.6% and 0.3% respectively at 5% significance level [8]. From past experience, we assumed response rates of 30%, 40%, and 40%, respectively, resulting in a total of 15,701 people being invited. Statistics Sweden sent invitations to individuals in survey 6 and provided population figures for 2020 [9].
Ethical approval and consent to participate
The evaluation has been done as part of the Public Health Authority’s mission to monitor infectious diseases and evaluate infection prevention measures in accordance with §§ 18 of the regulation (2021:248) from the Riksdag. For this reason, specific ethical approval was not required for the investigations described in this manuscript. All methods were performed in accordance with relevant guidelines and informed consent was obtained from all participants and/or their guardians.
Procedures
Invitations to potential participants were sent 3–5 weeks before the start of each survey and included instructions on how to perform self-sampling at home, information on how to collect the sample and when to complete the questionnaire. Participants accessed a telephone helpline 13 hours a day and enrolled in the study by completing an online form and providing consent; for persons under the age of 16, guardians have registered. Participation was voluntary and could be withdrawn at any time.
In survey 1, material for self-sampling was delivered to the participant’s home by the Swedish Armed Forces, while material for the other surveys was sent by regular mail. For all examinations, the Swedish Armed Forces have coordinated and collected the samples at the participants’ homes. The samples were taken on the same day as the self-sampling or the day after. Participants were asked to keep the samples refrigerated until collection.
For examinations 1 to 5, the self-sampling material consisted of sterile cotton swabs and a test tube containing 1 ml of phosphate-buffered saline. Participants were asked to use a cotton swab to perform a throat swab by scraping the posterior pharyngeal wall for 10–20 s and then swirling the cotton swab in the buffer. A second cotton swab was used to sample the distal nasal cavity through both nostrils, followed by swirling the cotton swab in the buffer. In studies 1 and 2, saliva was collected in a separate collection tube into which participants had spit 3–4 times, while in the following studies participants had swirled a third cotton swab in the saliva before the swab was swirled in the buffer.
For Study 6, participants received a swab and a test tube containing 0.5 mL of sample storage reagent consisting of 0.9% normal saline. A swab was used to sample the posterior pharyngeal wall, distal nasal cavity and saliva.
In all examinations, participants younger than 16 years had their sampling performed by a healthcare provider.
The samples were analyzed using real-time RT-PCR assays routinely used to diagnose COVID-19. In total, three different laboratories performed assays with different targets for real-time RT–PCR for SARS-CoV-2 [10,11,12,13] targeting ORF1ab and the N gene with minor modification of primers. The housekeeping genes hBeta-actin (Thermo Fisher Scientific art. no. 4333762F) or hRNAse P (IDT, RNase P, cat. no. 10006603) were amplified in parallel with SARS-CoV-2-specific RT-PCR as an internal control for the nucleic acid extraction procedure and to confirm that a sample taken correctly in terms of swabbing against mucosal surfaces in the nose or throat. Samples were considered invalid when hBeta-actin or hRNAse P signals were indeterminate or detected as negative, indicating a lack of human cell material.
Data on the participant’s age, sex, and area of residence were collected at the time the sample was collected. In survey 6, the participants’ level of education was obtained from administrative records. On the day of self-sampling, participants completed an online questionnaire about specific symptoms experienced in the past 24 hours and 2 weeks prior to self-sampling. An additional symptom questionnaire was distributed seven days after self-sampling to those who tested positive for SARS-CoV-2.
The number of reported covid-19 cases and the number of individuals tested for covid-19 have been obtained from the Public Health Authority.
Statistical analysis
The point prevalence was calculated as the weighted proportion of individuals who tested positive for SARS-CoV-2 among those with valid test results. In surveys 1 to 5, the weighting included sample weights from the web panel cohort, survey dropouts, and population size by age, sex, and region. In Study 6, weighting was based on participants’ age, sex, region, and education level to account for sampling error, dropout rate, and population size. Confidence intervals (CI) were calculated using the Clopper-Pearson exact method. Estimates were reported by gender and age group (0–15 years, 16–29 years, 30–59 years, 60 years and older) for Sweden and for the Stockholm region. We calculated the proportion of each symptom among those who tested positive for SARS-CoV-2; the proportion and its 95% CI were not weighted. Differences in prevalence between sexes were tested with univariate weighted logistic regression. Analyzes were performed in R v.3.6.2 (R Core Team 2019) [14] and survey package v.4.0 (Lumley, 2020) [15].