Abstract
Background
Community surveillance for acute respiratory illness (ARI) can include unsupervised participant-collected nasal swabs. Little is known about use of self-swabs in low-income populations or among households including extended family members and the validity of self-collected swabs. We assessed the acceptability, feasibility, and validity of unsupervised participant-collected nasal swabs in a low-income, community sample.
Methods
This was a substudy of a larger prospective community-based ARI surveillance study in 405 households in New York City. Participating household members self-collected swabs on the day of a research home visit for an index case, and for 3–6 subsequent days. Demographics associated with agreement to participate and swab collection were assessed, and index case self-collected versus research staff–collected swab results were compared.
Results
Most households (n = 292 [89.6%]) agreed to participate, including 1310 members. Being <18 years old, female, and the household reporter or member of the nuclear family (parents and children) were associated with both agreement to participate and self-swab collection. Being born in the United States or immigrating ≥10 years ago was associated with participation, and being Spanish-speaking and having less than a high school education were associated with swab collection. In all, 84.4% collected at least 1 self-swabbed specimen; self-swabbing rates were highest during the first 4 collection days. Concordance between research staff–collected swabs and self-swabs was 88.4% for negative swabs, 75.0% for influenza, and 69.4% for noninfluenza pathogens.
Conclusions
Self-swabbing was acceptable, feasible, and valid in this low-income, minoritized population. Some differences in participation and swab collection were identified that could be noted by future researchers and modelers.
Keywords: acute respiratory illness, community surveillance, influenza-like illness, pandemic preparedness, self-swab
Daily nasal self-swabbing proved to be an acceptable, feasible, and valid method for detection of respiratory pathogens. Differences in participation and swab collection in terms of key demographics as well as relationship to household reporter were identified.
Acute respiratory illnesses (ARIs) lead to a significant health burden worldwide [1, 2]. In the United States (US), ARIs contribute to more than 25 million primary care and 9 million emergency department visits per year [3, 4]. In addition to direct medical expenses, ARIs also lead to indirect costs and societal burden due to an increase in missed work or school days and help required to care for sick children. Influenza alone has a considerable impact on health in the US. In the 2019–2020 influenza season, the Centers for Disease Control and Prevention (CDC) estimated that influenza resulted in 35.9 million illnesses, 16 million medical visits, 389 000 hospitalizations, and 25 000 deaths [5].
However, medically attended disease surveillance only captures a fraction of ARIs, including influenza, as many infected people never seek care [6]. Community surveillance, “the systematic detection and reporting of events of public health significance within a community by community members” [7], may be complementary to surveillance that focuses on medically attended disease. Unsupervised participant-collected nasal swabs, herein referred to as self-swabs, can be implemented to support community-based surveillance of ARI. This method of swab collection could include at-home testing and resulting without the need to present to medical care [8]. Self-swabs reduce potential exposures to healthcare professionals, provide a fuller ascertainment of overall infection rate, and may also identify ARI pathogens present in the community that might not be captured by surveillance for medically attended ARI, if they lead to illnesses for which people do not generally seek medical care. Additionally, >1 self-swab collected during an ARI episode might improve pathogen detection when compared to collecting a single specimen.
Vulnerable populations of lower socioeconomic status (SES) are disproportionally affected by ARIs [9, 10]. Lower levels of formal education and unemployment have been associated with a greater incidence of ARIs, and children from lower-income families miss more days of school associated with ARIs [9, 10]. As recorded statistics underestimate incidence of ARI [6], it is likely that the effect of ARIs on those of lower SES is also underestimated. Previous studies have employed self-swabs to detect acute respiratory infections in nonmedical settings [11–13]. A study in Seattle, Washington, tested the feasibility of self-swabs among patients of a managed care organization, with a focus on timeliness of specimen collection and appropriate packaging of the specimens [14]. Another study based in North Carolina assessed the feasibility of at-home testing for severe acute respiratory syndrome coronavirus 2 in children [8]. Other studies assessed self-swabs in a community surveillance setting [15–18]. However, to our knowledge, self-swabs collected by members of low-income households from underserved populations have not been studied. In addition, within households, an index case is the first identified case of ARI in a household. To our knowledge, the feasibility of collection of self-swabs among household members of index cases, in particular among extended family members and non–family members who may be less connected to the nuclear family unit, has not been studied. Additionally, while self-swabs to detect pathogens associated with ARI have been used to determine agreement between anterior nasal swabs and saliva specimens [16, 19], self-swabs have not been assessed to determine agreement between different specimen types for different respiratory pathogens. Furthermore, sequential self-swabbing for multiple days in a row has not included data on possible swabbing fatigue [16], meaning a decline in the rate of swab collection over time. Thus, the aims of this study were to assess acceptability and feasibility of collecting self-swabs from all household members during ARI in a low-income, community sample. Additionally, the study aimed to assess validity of collecting self-swabs from index cases to assess for different respiratory pathogens.
METHODS
Study Population
This substudy was part of a larger prospective 5-year community-based study of ARI surveillance consisting of 405 households, the Mobile Surveillance for ARI and Influenza-like Illness in the Community (MoSAIC) study [20, 21], which took place among a low-income, primarily immigrant Latino population in northern Manhattan in New York City; 20% of households in the northern Manhattan neighborhoods have incomes below the poverty level, as defined by the US Census Bureau [22]. In brief, eligible households for MoSAIC were limited to those with ≥3 persons per household with at least 1 member <18 years of age, were Spanish or English speaking, and had a household reporter with a cellphone with text-messaging capabilities. Household reporters were self-selected and, after reviewing the study procedures, were sent twice-weekly text messages and asked if any household members had ARI symptoms: “Reply with 1 or 2. Does anyone in the household have runny nose, congestion, sore throat, cough, body aches, or fever, or feels [sic] hot? 1: yes; 2: no.” Partway into the study, “allergies” was added because households were not reporting qualifying symptoms that they believed were due to allergies as illness. If a household member developed symptoms in between the 2 text-message time points, they were asked to call research staff to report those symptoms. Research staff determined the index case and confirmed by phone calls if ARI criteria were met. If met, research staff conducted a home visit generally within 2 days of ARI confirmation to collect a nasal swab from the index case, herein referred to as the research swab. ARI criteria consisted of having ≥2 of the following symptoms: fever (feverish), runny nose and/or congestion, sore throat, cough, and/or myalgias. For infants <1 year old, congestion alone qualified.
Between December 2014 and March 2015 and February 2015 through May 2016, research staff consented households to participate in a self-swab substudy, provided self-swab kits for the ill household member (index case) and other household members, demonstrated swab collection, and provided written swabbing instructions with pictures. Following determination of an index case, all household members were asked to collect a nasal swab (or parent-collected swab if the parent of a child aged <18 years deemed it more suitable to collect a swab from their child) later that day and for the 6 subsequent days if the index case had influenza and the 3 subsequent days if they had a noninfluenza pathogen, for a total of 7 or 4 self-swabs, respectively. Households were sent a daily text message as a reminder to collect swabs in the evening. Families were asked to store the kits and the collected specimens in the refrigerator, and the research staff collected them at the end of the swabbing period. Households were compensated with $20 for their time if at least 1 person in the household collected swabs for at least 3 days. The institutional review board of the Columbia University Irving Medical Center approved this study, and the CDC relied upon this determination.
Laboratory Analysis
All research swabs and the first self-swab for the index case (day 0) were analyzed by multiplex reverse-transcription polymerase chain reaction (RT-PCR), using a commercially available assay (FilmArray respiratory panel, BioFire Diagnostics) for 20 respiratory pathogens including adenovirus, seasonal human coronaviruses (hCOVs; HKU1, NL63, 229E, OC43), metapneumovirus, rhinovirus/enterovirus, influenza (A, A/H1, A/H3, A/H1-2009, B), parainfluenza (types 1–4), respiratory syncytial virus (RSV), Bordetella pertussis, Chlamydia pneumoniae, and Mycoplasma pneumoniae [23]. All research swabs and self-swabs were analyzed at the same laboratory using the same equipment and procedures.
Statistical Analysis
We assessed the acceptability of self-swabbing by determining the proportion of participants who agreed to participate versus those who did not among those who were approached, as well as the proportion who collected at least 1 self-swab versus no swabs among those who agreed to participate. We then used χ2 tests to assess the relationship, key demographic characteristics (age group, sex, primary language spoken [adults only], birth in US [US born; not US born and in US ≥10 years; not US born and in US <10 years], education level, and the participant's relationship to the household reporter [nuclear family member; other family member; unrelated to household reporter]), and participation and self-swab collection, separately.
We assessed the feasibility of swab collection by calculating the proportion of swabs collected for each of the days expected from each participant. We also assessed the relationship between whether a participant was the index case with an ARI that met criteria versus a potentially exposed household member and the number of swabs collected using χ2 analyses.
Referencing research swabs as the gold standard, we conducted validity analysis of self-swabbing by determining concordance between the day 0 self-swab in the index case and the research swab taken the same day, for research swabs that tested positive for a pathogen and negative for a pathogen. Concordance was assessed for the respiratory pathogens individually and broadly categorized as influenza positive, noninfluenza positive, and negative. In the event a research swab tested positive for 2 pathogens, we separated the co-detections by duplicating the individual-level data for each pathogen, and assessed for concordance for each pathogen individually. We also assessed differences in concordance for each pathogen between children and adults. Sensitivity was defined as concordance between self-swab and research swab results when the research swab tested positive for any pathogen. Specificity was defined as concordance between self-swab and research swab results when the research swab tested negative for any pathogen. SPSS version 26.0 software was used for all analyses (IBM Corporation, Armonk, New York).
RESULTS
Overall, 326 households, including 1658 household members, were eligible for this substudy. Of the 326 households, 292 (89.6%) agreed to participate, including 1486 household members. Within the households who agreed, 179 household members were not in the household during this substudy due to vacation or other reasons. Therefore, the study comprised 1307 household members. Age, sex, primary language spoken, being born or having spent more time in the US, and relationship to the household reporter were associated with participation (Table 1). Household reporters and members of the nuclear family (parents and their children) were more likely to participate (84.7% and 79.7%, respectively); in comparison, 54.2% of extended family members and 46.7% of nonrelated household members agreed to participate.
Table 1.
Characteristics of the Study Population and Relationship Between Demographic Characteristics and Feasibility of Self-Swabbing
| Characteristics | Eligible, No. |
Agreed to Participate (n = 1658) |
Agreed to Participate, No. |
Self-Swabs Collected of Those Participating (n = 1307) |
||||
|---|---|---|---|---|---|---|---|---|
| Yes | No | P Value | Yes | No | P Value | |||
| Total | 1658 | 1307 (78.8) | 351 (21.2) | 1307 | 1103 (84.4) | 204 (16.6) | ||
| Age, y | ||||||||
| <18 | 761 | 637 (83.7) | 124 (16.3) | <.001 | 637 | 565 (88.7) | 72 (11.3) | <.001 |
| ≥18 | 897 | 670 (74.7) | 227 (25.3) | 670 | 538 (80.3) | 132 (19.7) | ||
| Sex | ||||||||
| Female | 957 | 776 (81.1) | 181 (18.9) | .009 | 776 | 683 (88.0) | 93 (12.0) | <.001 |
| Male | 701 | 531 (75.7) | 170 (24.3) | 531 | 420 (79.1) | 111 (20.1) | ||
| Primary language spoken (adults) | ||||||||
| Spanish | 654 | 489 (74.8) | 165 (25.2) | .29 | 489 | 415 (84.9) | 74 (15.1) | <.001 |
| English | 224 | 164 (73.2) | 60 (26.8) | 164 | 109 (66.5) | 55 (33.5) | ||
| Both equally | 19 | 17 (89.5) | 2 (10.5) | 17 | 14 (82.4) | 3 (17.6) | ||
| Born in the US | ||||||||
| Yes | 768 | 643 (83.7) | 125 (16.3) | <.001 | 643 | 557 (86.6) | 86 (13.4) | .098 |
| No, in US ≥10 y | 509 | 407 (80.0) | 102 (20.0) | 407 | 335 (82.3) | 72 (17.7) | ||
| No, in US <10 y | 378 | 254 (67.2) | 124 (32.8) | 254 | 209 (82.3) | 45 (17.7) | ||
| Education (adults only) | ||||||||
| Less than high school | 372 | 283 (76.1) | 89 (23.9) | .60 | 283 | 238 (84.1) | 45 (15.9) | .032 |
| High school | 198 | 143 (72.2) | 55 (27.8) | 143 | 105 (73.4) | 38 (26.6) | ||
| Some college or college graduate | 320 | 238 (74.4) | 82 (25.6) | 238 | 191 (80.3) | 47 (19.7) | ||
| Relationship to household reporter | ||||||||
| Household reporter | 326 | 276 (84.7) | 50 (15.3) | <.001 | 276 | 267 (96.7) | 9 (3.3) | <.001 |
| Nuclear family member | 1219 | 972 (79.7) | 247 (20.3) | 972 | 791 (81.4) | 181 (18.6) | ||
| Other family member | 83 | 45 (54.2) | 38 (45.8) | 45 | 31 (68.9) | 14 (31.1) | ||
| Unrelated | 30 | 14 (46.7) | 16 (53.3) | 14 | 14 (100.0) | … | ||
Data are presented as No. (%) unless otherwise indicated.
Abbreviation: US, United States.
Of the 1307 members who participated, 1103 (84.4%) collected at least 1 of the 7 self-swabs to the research team. Age, sex, language, education level, and relationship to the household reporter were associated with taking of self-swabs among those who agreed to participate (Table 1). A high proportion of extended family (n = 31 [68.9%]) and unrelated household members (n = 14 [100.0%]) also collected at least 1 swab. The proportion of self-swabs collected was highest on days 0–3 with >80% of swabs collected, decreasing to <70% for days 4–6 (Figure 1A). We observed statistically significant differences in test collection between days 2 and 3, 3 and 4, and 4 and 5 (McNemar test, P = .025, .002, and .023, respectively). There was no significant difference between days 0 and 1, days 1 and 2, or days 5 and 6. On regression analysis, there was not a significant difference in return rates comparing day 1, 2, or 3 to day 0; however, days 4, 5, and 6 had decreased odds of return compared to day 0 (odds ratios, 0.37 [95% confidence interval {CI}, .28–.50]; 0.36 [95% CI, .27–.49]; and 0.36 [95% CI, .26–.48], respectively). Participants were significantly more likely to collect swabs on all 7 days if they were an index case than if they were a potentially exposed household member (P < .001; Figure 1B). Exposed household members who became symptomatic had similar swabbing rates to index cases (0.4%–2% absolute difference).
Figure 1.
Feasibility of swabbing. A, Proportion of self-swabs collected by day by child vs adult participants. B, Proportion of self-swabs collected by day for index case versus other household members. *P < .01.
There were 358 sets of research swabs from those who met ARI criteria for which a self-swab was collected. Most (n = 347 [96.9%]) ARIs occurred among members of the nuclear family with the remaining 3.1% (n = 11) among extended family members or nonrelated household members. More than half (n = 197 [55%]) occurred in children. Of those, 188 (52.5%) tested positive for 1 respiratory pathogen, 15 (4.2%) tested positive for 2 respiratory pathogens, and 155 (43.3%) tested negative. There were 19 instances in which a research swab and/or self-swab detected a co-detection, and concordance of each pathogen between the self-swab and the research swab was assessed. There were 27 instances in which a valid laboratory result could not be obtained from a self-swab. Thus, to assess for validity of self-swabbing, we assessed for concordance for 204 pathogens from positive swabs and 146 negative swabs for a total of 350 pairs.
Overall, 144 (70.6%) self-swabs from those with ARIs were concordant with research swabs that were positive for a pathogen, and 129 (88.4%) of self-swabs were concordant with research swabs that were negative for a pathogen. Concordance levels did not vary based on demographic characteristics (Table 2).
Table 2.
Characteristics of Index Cases and Relationship Between Demographic Characteristics and Validity of Self-Swabbing
| Characteristic | Self-Swabs With Positive Research Swabs, No. | Self-Swabs Concordant With Positive Research Swabsa (n = 204) | Self-Swabs With Negative Research Swabs, No. | Self-Swabs Concordant With Negative Research Swabsa (n = 146) | ||||
|---|---|---|---|---|---|---|---|---|
| Yes | No | P Value | Yes | No | P Value | |||
| Total | 204 | 146 | ||||||
| Age, y | ||||||||
| <18 | 123 | 85 (69.1) | 38 (30.9) | .57 | 73 | 61 (83.6) | 12 (16.4) | .07 |
| ≥18 | 81 | 59 (72.8) | 22 (27.2) | 73 | 68 (93.2) | 5 (6.8) | ||
| Sex | ||||||||
| Female | 129 | 91 (70.5) | 38 (29.5) | .99 | 105 | 92 (87.6) | 13 (12.4) | .66 |
| Male | 75 | 53 (70.7) | 22 (29.3) | 41 | 37 (90.2) | 4 (9.8) | ||
| Primary language spoken (adults only) | ||||||||
| Spanish | 63 | 46 (73.0) | 17 (27.0) | .64 | 59 | 56 (94.9) | 3 (5.1) | .26 |
| English | 16 | 11 (68.8) | 5 (31.2) | 11 | 9 (81.8) | 2 (18.2) | ||
| Other | 2 | 2 (100.0) | … | 3 | 3 (100.0) | … | ||
| Born in the US | ||||||||
| Yes | 115 | 80 (69.6) | 35 (30.4) | .87 | 74 | 63 (85.1) | 11 (14.9) | .41 |
| No, in US >10 y | 54 | 38 (70.4) | 16 (29.6) | 44 | 41 (93.2) | 3 (6.8) | ||
| No, in US <10 y | 35 | 26 (74.3) | 9 (25.7) | 28 | 25 (89.3) | 3 (10.7) | ||
| Education | ||||||||
| Less than high school | 28 | 22 (78.6) | 6 (21.4) | .52 | 31 | 29 (93.5) | 2 (6.5) | .55 |
| High school | 16 | 10 (62.5) | 6 (37.5) | 11 | 11 (100.0) | … | ||
| Some college or college graduate | 36 | 26 (72.2) | 10 (27.8) | 31 | 28 (90.3) | 3 (9.7) | ||
| Relationship to household reporter | ||||||||
| Household reporter | 52 | 42 (80.8) | 10 (19.2) | .23 | 53 | 49 (92.5) | 4 (7.5) | .42 |
| Nuclear family member | 149 | 100 (67.1) | 49 (32.9) | 86 | 73 (84.9) | 13 (15.1) | ||
| Other family member | 2 | 1 (50.0) | 1 (50.0) | 5 | 5 (100.0) | … | ||
| Unrelated | 1 | 1 (100.0) | … | 2 | 2 (100.0) | … | ||
Data are presented as No. (%) unless otherwise indicated.
Abbreviation: US, United States.
Concordance was analyzed as the percentage of self-swabs that, in diagnostic testing for respiratory pathogens, yielded the same result as its corresponding research swab.
Of the 350 case pairs, 44 research swabs were positive for influenza, 160 were positive for another pathogen, and 146 were negative (Table 3). Of the additional 27 research swabs for which the self-swab in the index case did not have a valid result, 5 were positive for influenza, 2 for a seasonal hCOV, and 11 for rhinovirus/enterovirus, and 9 were negative. The overall concordance was 78.0% (McNemar P = .070, κ = 0.7, p < 0.001) and was 74.5% (McNemar P = .29, κ =0.7, p < .001) for children and 82.5% (McNemar P = .052, κ = 0.7, p < .001) for adults. Overall, the concordance was 75.0% (n = 33) for influenza, 69.4% (n = 111) for noninfluenza pathogens, and 88.4% (n = 129) for negative samples. For influenza, there were 11 instances in which the research swab identified influenza but the self-swab did not, and 3 instances in which the self-swab identified influenza and the research swab did not. For noninfluenza pathogens, there were 49 cases in which the research swab identified a pathogen but the self-swab did not, and 14 cases in which the self-swab identified a noninfluenza pathogen and the research swab did not (seasonal hCOV, n = 3; rhinovirus/enterovirus, n = 9; RSV, n = 1; and M pneumoniae, n = 1).
Table 3.
Concordance of Research Swabs and Self-Swabs
| Pathogen | Results of Testing | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Research Swabs Testable for Pathogen (n = 350) |
Concordant Self-Swab | Sensitivity | Specificity | Child Research Swab Testable for Pathogen (n = 196) |
Child Concordant Self-Swab | Sensitivity | Specificity | Adult Research Swab Testable for Pathogen (n = 154) |
Adults Concordant Self-Swab | Sensitivity | Specificity | |
| Influenza | 44 | 33 | 75.0% | 98.4% | 22 | 18 | 81.8% | 97.7% | 22 | 15 | 68.2% | 99.2% |
| Influenza A | 38 | 28 | 73.7% | 98.4% | 19 | 16 | 84.2% | 97.7% | 19 | 12 | 63.2% | 99.3% |
| Influenza B | 6 | 5 | 83.3% | 100% | 3 | 2 | 66.7% | 100% | 3 | 3 | 100.0% | 100.0% |
| Noninfluenza | 160 | 111 | 69.4% | 87.9% | 101 | 67 | 66.3% | 81.1% | 59 | 44 | 74.6% | 93.7% |
| Seasonal hCOVs | 51 | 40 | 78.4% | 98.7% | 23 | 19 | 82.6% | 97.7% | 28 | 21 | 75.0% | 100.0% |
| RSV | 11 | 8 | 72.7% | 99.4% | 5 | 3 | 60.0% | 99.0% | 6 | 5 | 83.3% | 99.3% |
| hMPV | 16 | 9 | 56.3% | 99.7% | 14 | 8 | 57.1% | 99.7% | 2 | 1 | 50.0% | 100% |
| Rhinovirus/enterovirus | 70 | 44 | 62.9% | 95.0% | 49 | 29 | 59.2% | 93.9% | 21 | 15 | 71.4% | 97.1% |
| Adenovirus | 8 | 7 | 87.5% | 99.7% | 7 | 6 | 85.7% | 99.5% | 1 | 1 | 100.0% | 100.0% |
| Chlamydia pneumoniae | 1 | 1 | 100.0% | 100% | 1 | 1 | 100.0% | 100.0% | 0 | … | … | … |
| Parainfluenza viruses | 2 | 2 | 100.0% | 100% | 1 | 1 | 100.0% | 100.0% | 1 | 1 | 100.0% | 100.0% |
| Mycoplasma pneumoniae | 1 | 0 | … | 99.7% | 1 | … | … | 99.7% | 0 | … | … | … |
| Negative | 146 | 129 | 88.4% | 76.0% | 73 | 61 | 83.6% | 76.4% | 73 | 68 | 93.2% | 75.3% |
Abbreviations: hCOVs, human coronaviruses; hMPV, human metapneumovirus; RSV, respiratory syncytial virus.
Although the sample size was too small to assess for significance, concordance was highest for research swabs that were positive for C pneumoniae (n = 1 [100.0%]), parainfluenza viruses (n = 2 [100.0%]), adenovirus (n = 7 [87.5%]), seasonal hCOVs (n = 40 [78.4%]), and RSV (n = 8 [72.7%]), as well as swabs that tested negative for a pathogen (n = 129 [88.4%]) (Table 3). In comparison to swabs from children, swabs from adults had higher concordance for influenza B (100.0% vs 66.7%), RSV (83.3% vs 60.0%), adenovirus (100.0% vs 85.7%), and rhinovirus/enterovirus (71.4% vs 59.2%), and swabs that tested negative (93.2% vs 83.6%). Conversely, swabs from children had higher concordance for influenza A (84.2% vs 63.2%), seasonal hCOVs (82.6% vs 75.0%), and human metapneumovirus (57.1% vs 50.0%).
Fifteen research swabs identified co-detections and of those, 3 (20.0%) detected infection with influenza virus and a noninfluenza pathogen, and 12 (80.0%) detected infection with 2 noninfluenza pathogens. Among the 3 influenza/noninfluenza co-detections, no self-swabs detected both pathogens, 2 swabs detected influenza virus only, and 1 swab detected a noninfluenza pathogen only. Among the 12 research swabs with 2 noninfluenza co-detections, 6 self-swabs detected both pathogens, 4 detected 1 pathogen, and 2 detected neither pathogen. An additional 4 self-swabs detected coinfections, none of which included influenza. Of those, 3 had a corresponding research swab that detected only 1 pathogen, and 1 swab detected neither pathogen.
DISCUSSION
In this prospective study comparing research staff–collected swabs and self-swabs, we found that self-collected nasal swabs are suitable for the detection of respiratory pathogens by RT-PCR in ARIs among low-income households. This lends support to the idea that self-swabbing could be a safe and effective surveillance tool to support community-based surveillance that can reduce delays in time to diagnosis while also helping to minimize participant contact with the healthcare system, and could be incorporated into protocols for pandemic preparedness of ARIs [13, 24].
Most participants agreed to collect self-swabs, indicating that self-swabbing could be helpful in further understanding household ARI transmission. This high participation rate is in agreement with other studies assessing the feasibility of nasal self-swabbing for population-based studies, suggesting a likelihood of compliance in future large-scale prospective studies [11, 25, 26]. However, there are some important potential differences in feasibility for use of self-swabbing for household transmission studies. In this population, age, sex, whether the participant was born in the US, and their relationship to the household reporter did affect who agreed to participate. It might, therefore, be important to consider the potential need to account for possible biases in participation when surveillance studies are used to extrapolate to population-level estimates, depending on the study. There were also differences in collecting self-swabs that could affect the study procedures for future investigations. Household reporters and members of the nuclear family (parents and their children) were more likely to take swabs, but a high proportion of unrelated household members also collected swabs, suggesting that once they agreed to participate, they were likely to follow through with study-related procedures. Household self-swabbing studies that recruit extended family members and nonrelatives should make independent contact with those adults to improve participation. Other household ascertainment studies do not remark on the likelihood of self-swab collection among nuclear family members versus other household members [16, 25].
Those who had less than a high school education and/or were Spanish speakers were more likely to collect swabs, which could alleviate concerns in future studies about potential disparities in collection of self-swabs. Swabs from children were slightly more likely to be collected than those from adults. A German study found self-swabbing to be well accepted among adults and children alike [27]. In our study, female participants were more likely to collect swabs than males. This may be because most of our household reporters were female, and these contacts were more likely to swab than other family members.
Participants had the highest self-swab rates in the first 4 days of ARIs, suggesting swabbing fatigue over the course of the week. This new information is important for future studies, as incentives to continue collecting self-swabs are likely needed for studies with sequential swabs. An alternative is to have participants not collect swabs daily, with notifications to remind participants what day to swab. Individuals were also more likely to swab if they were the index case.
In general, concordance was high between index case research swabs and self-swabs. Negative concordance was higher among adults than children, which may suggest that parents were better than research staff at collecting swabs from their children. However, a previous pilot study did not find differences in concordance based on demographic variables, which may have reflected the relatively small sample size [28]. Concordance was higher among research swabs and self-swabs when research swabs were negative for a pathogen than when they were positive. There were instances where the self-swab detected a pathogen that was not detected on the research swab, including for influenza virus. In those cases, it is possible the participants swabbed more diligently than the research staff. Concordance also varied by pathogen; it was slightly better for influenza A than influenza B virus among children and, among the noninfluenza pathogens, slightly better for seasonal hCOVs, RSV, and adenovirus. Concordance varied among swabs that resulted in coinfections. We were not able to obtain valid results on a relatively small number of self-swabs, suggesting that additional support from research staff may be required to conduct community surveillance or that swabbing more than one time may be considered. Despite these limitations in concordance, the generally high concordance between index case research swabs and self-swabs make self-swabbing an important method for studies to consider.
Limitations
We used the FilmArray Respiratory panel BioFire diagnostic test, for which sensitivity may be lower if the specimen was not of peak quality, and sensitivity may not be generalizable across different diagnostic tests. Specimen quality and storage conditions may have led to false negatives in some samples, but assessing the impact of this possibility was, in part, the purpose of this substudy. Families received $20 to collect self-swabs, which may have provided added motivation to collect swabs. Compensation, or lack thereof, may pose a potential barrier that would need to be navigated if self-swabbing were to be implemented on a wider scale. The study was conducted in northern Manhattan and participants consisted of a low-income, primarily immigrant Latino population. In addition, participants were already part of a larger study, which could have impacted their willingness to both participate and collect self-swabs. Self-swabbing studies should be performed in populations from other geographical areas to increase the generalizability of these findings.
CONCLUSIONS
Daily nasal self-swabbing proved to be an acceptable, feasible, and valid method to assess pathogens associated with ARIs among this low-income community sample. This approach could be useful for the surveillance of other seasonal and pandemic pathogens. Differences in participation and collection of swabs in terms of age, sex, primary language spoken, whether the participant was born in the US, education, and relationship to household reporter were identified that could be noted by future researchers and modelers.
Contributor Information
Priyam Thind, Division of Child and Adolescent Health, Department of Pediatrics, Columbia University Vagelos College of Physicians and Surgeons, New York, New York, USA.
Celibell Y Vargas, Division of Child and Adolescent Health, Department of Pediatrics, Columbia University Vagelos College of Physicians and Surgeons, New York, New York, USA.
Carrie Reed, Influenza Division, Centers for Disease Control and Prevention, Atlanta, Georgia, USA.
Liqun Wang, Division of Pediatric Infectious Diseases, Department of Pediatrics, Columbia University Vagelos College of Physicians and Surgeons, New York, New York, USA.
Luis R Alba, Division of Pediatric Infectious Diseases, Department of Pediatrics, Columbia University Vagelos College of Physicians and Surgeons, New York, New York, USA.
Elaine L Larson, School of Nursing, Columbia University Irving Medical Center, New York, New York, USA; Department of Epidemiology, Mailman School of Public Health, Columbia University Irving Medical Center, New York, New York, USA.
Lisa Saiman, Division of Pediatric Infectious Diseases, Department of Pediatrics, Columbia University Vagelos College of Physicians and Surgeons, New York, New York, USA; Department of Infection Prevention and Control, New York–Presbyterian Hospital, New York, New York, USA.
Melissa S Stockwell, Division of Child and Adolescent Health, Department of Pediatrics, Columbia University Vagelos College of Physicians and Surgeons, New York, New York, USA; Department of Population and Family Health, Mailman School of Public Health, Columbia University Irving Medical Center, New York, New York, USA.
Notes
Author contributions. M. S. S., L. S., E. L. L., and C. R. conceptualized and designed the study, designed the data collection instruments, and reviewed and revised the manuscript. P. T. analyzed the data and drafted and revised the manuscript. C. Y. V. aided in the conduct of the study, contributed to data collection, and reviewed and revised the manuscript. L. W. conducted the laboratory analyses and reviewed and revised the manuscript. L. R. A. contributed to data collection and reviewed and revised the manuscript.
Acknowledgments. The authors thank the other members of the MoSAIC study team for their assistance: Hilbania Diaz, Yaritza Castellanos de Belliard, Maria Morban, Julia Nunez, Asha Liverpool, Othanya Garcia, Raul Silverio, and Manuel Cifuentes.
Patient consent. The authors confirm that the patients’ consent was obtained.
Disclaimer. The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the CDC.
Financial support. This work was supported by the Centers for Disease Control and Prevention (grant number U01IP000618 to M. S. S.). BioFire Diagnostics, Inc, loaned machines for this study and provided kits at a reduced rate for research.
References
- 1. Liu L, Oza S, Hogan D, et al. Global, regional, and national causes of under-5 mortality in 2000–15: an updated systematic analysis with implications for the Sustainable Development Goals. Lancet 2016; 388:3027–35. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Vos T, Allen C, Arora M, et al. Global, regional, and national incidence, prevalence, and years lived with disability for 310 diseases and injuries, 1990–2015: a systematic analysis for the Global Burden of Disease Study 2015. Lancet 2016; 388:1545–602. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Cairns C, Kang K. . National hospital ambulatory medical care survey: 2019 emergency department summary tables. National Center for Health Statistics. 2022. . Available at: https://www.cdc.gov/nchs/data/nhamcs/web_tables/2019-nhamcs-ed-web-tables-508.pdf. Accessed 7 May 2022.
- 4. Santo L, Okeyode T. National ambulatory medical care survey: 2018 national summary tables.2021. Available at: https://www.cdc.gov/nchs/data/ahcd/namcs_summary/2018-namcs-web-tables-508.pdf. Accessed 7 May 2022.
- 5. Centers for Disease Control and Prevention . Estimated influenza illnesses, medical visits, hospitalizations, and deaths in the United States—2019–2020 influenza season.2022. Available at: https://www.cdc.gov/flu/about/burden/2019-2020.html. Accessed 7 May 2022.
- 6. Reed C, Chaves SS, Daily Kirley P, et al. Estimating influenza disease burden from population-based surveillance data in the United States. PLoS One 2015; 10:e0118369. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Technical Contributors to the June WHO Meeting . A definition for community-based surveillance and a way forward: results of the WHO global technical meeting, France, 26 to 28 June 2018. Euro Surveill 2019; 24:1800681. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Ahmed A, Rossman W, Lu LC, et al. Feasibility of at-home virological and serological testing for SARS-CoV-2 in children. Open Forum Infect Dis 2022; 9:ofac459. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Sims DG, Downham MA, McQuillin J, Gardner PS. Respiratory syncytial virus infection in north-east England. Br Med J 1976; 2:1095–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Egbuonu L, Starfield B. Child health and social status. Pediatrics 1982; 69:550–7. [PubMed] [Google Scholar]
- 11. Akmatov MK, Krebs S, Preusse M, et al. E-mail–based symptomatic surveillance combined with self-collection of nasal swabs: a new tool for acute respiratory infection epidemiology. Int J Infect Dis 2011; 15:e799–803. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Larios OE, Coleman BL, Drews SJ, et al. Self-collected mid-turbinate swabs for the detection of respiratory viruses in adults with acute respiratory illnesses. PLoS One 2011; 6:e21335. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Akmatov MK, Pessler F. Self-collected nasal swabs to detect infection and colonization: a useful tool for population-based epidemiological studies? Int J Infect Dis 2011; 15:e589–93. [DOI] [PubMed] [Google Scholar]
- 14. Jackson ML, Nguyen M, Kirlin B, Madziwa L. Self-collected nasal swabs for respiratory virus surveillance. Open Forum Infect Dis 2015; 2:ofv152. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Emanuels A, Heimonen J, O’Hanlon J, et al. Remote household observation for noninfluenza respiratory viral illness. Clin Infect Dis 2021; 73:e4411–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Grijalva CG, Rolfes M, Zhu Y, et al. Performance of self-collected anterior nasal swabs and saliva specimens for detection of SARS-CoV-2 during symptomatic and asymptomatic periods. Open Forum Infect Dis 2021; 8:ofab484. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Dawood FS. Clinical characteristics, and risk factors of SARS-CoV-2 infection among pregnant individuals in the United States. Clin Infect Dis 2021; 19:ciab713. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Dawood FS, Porucznik CA, Veguilla V, et al. Incidence rates, household infection risk, and clinical characteristics of SARS-CoV-2 infection among children and adults in Utah and New York City, New York. JAMA Pediatr 2022; 176:59–67. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Teo AKJ, Choudhury Y, Tan IB, et al. Saliva is more sensitive than nasopharyngeal or nasal swabs for diagnosis of asymptomatic and mild COVID-19 infection. Sci Rep 2021; 11:3134. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Stockwell MS, Reed C, Vargas CY, et al. MoSAIC: mobile surveillance for acute respiratory infections and influenza-like illness in the community. Am J Epidemiol 2014; 180:1196–201. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Stockwell MS, Reed C, Vargas CY, et al. Five-year community surveillance study for acute respiratory infections using text messaging: findings from the MoSAIC study. Clin Infect Dis 2022; 75:987–95. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Hinterland K, Naidoo M, King L, et al. Community health profiles 2018, Manhattan Community District 12: Washington Heights and Inwood. 2018. Available at: https://www1.nyc.gov/assets/doh/downloads/pdf/data/2018chp-mn12.pdf. Accessed 10 February 2023.
- 23. BioFire Diagnostics . FilmArray respiratory panels. Available at: https://www.biofiredx.com/products/the-filmarray-panels/filmarrayrp/. Accessed 19 October 2021.
- 24. Emerson J, Cochrane E, McNamara S, Kuypers J, Gibson RL, Campbell AP. Home self-collection of nasal swabs for diagnosis of acute respiratory virus infections in children with cystic fibrosis. J Pediatric Infect Dis Soc 2013; 2:345–51. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Malosh RE, Petrie JG, Callear AP, Monto AS, Martin ET. Home collection of nasal swabs for detection of influenza in the household influenza vaccine evaluation study. Influenza Other Respir Viruses 2021; 15:227–34. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Kojima N, Turner F, Slepnev V, et al. Self-collected oral fluid and nasal swab specimens demonstrate comparable sensitivity to clinician-collected nasopharyngeal swab specimens for the detection of SARS-CoV-2. Clin Infect Dis 2021; 73:e3106–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Haussig JM, Targosz A, Engelhart S, et al. Feasibility study for the use of self-collected nasal swabs to identify pathogens among participants of a population-based surveillance system for acute respiratory infections (GrippeWeb-Plus)—Germany 2016. Influenza Other Respir Viruses 2019; 13:319–30. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Vargas CY, Wang L, Castellanos de Belliard Y, et al. Pilot study of participant-collected nasal swabs for acute respiratory infections in a low-income, urban population. Clin Epidemiol 2016; 8:1–5. [DOI] [PMC free article] [PubMed] [Google Scholar]