Identifying potential post-COVID-19 condition among people experiencing homelessness using longitudinal symptom patterns: A prospective cohort study
[Line breaks added]
Objectives
People experiencing homelessness have high SARS-CoV-2 infection and re-infection burden, potentially leading to higher prevalence of post-COVID-19 condition (PCC). However, high baseline symptom rates may make identification of PCC difficult or impossible in this population. This study evaluates symptom patterns over time to assess their ability to identify potential PCC among individuals experiencing homelessness.
Study design
Prospective cohort study
Methods
We prospectively followed a large (n = 736), representative cohort of people experiencing homelessness recruited at random from 62 sites in Toronto, Canada between June and September 2021. Participants were interviewed up to five times over approximately 12 months.
Longitudinal patterns of twelve self-reported symptoms were assessed through latent transition analysis, and generalized estimating equations with logit link were applied to determine their association with known risk of potential PCC.
Results
Among 736 participants, three latent statuses were identified: (1) ‘No/Few Symptoms’ (≥70 %), (2) ‘Non-Specific Symptoms’ (15–23 %), and (3) ‘Infection-Related Symptoms’ (≤5 %). Statuses 2 and 3 were associated with being at risk of PCC following symptomatic infection (aOR 3.41 [95 % CI 2.3–5.0] and 3.18 [95 % CI 1.6–6.4]) but not with being at risk of PCC overall.
Transition probabilities suggested PCC would mostly occur among individuals with symptoms at baseline. However, the clustered area under the curve was modest (0.70 [95 % CI 0.65–0.75]), indicating symptom-based approaches are suboptimal for identification of potential PCC.
Conclusions
Self-reported symptoms do not reliably identify potential PCC among people experiencing homelessness, due to high rates of underlying symptoms and asymptomatic infections. Alternative, strengths-based approaches are recommended to more equitably identify post-COVID condition in this population.
Web | DOI | PDF | Public Health | Open Access
Richard, Lucie; Nisenbaum, Rosane; Dyer, Allison; Mergarten, Daniela; Brown, Michael; Stewart, Suzanne; Hwang, Stephen W.
[Line breaks added]
Objectives
People experiencing homelessness have high SARS-CoV-2 infection and re-infection burden, potentially leading to higher prevalence of post-COVID-19 condition (PCC). However, high baseline symptom rates may make identification of PCC difficult or impossible in this population. This study evaluates symptom patterns over time to assess their ability to identify potential PCC among individuals experiencing homelessness.
Study design
Prospective cohort study
Methods
We prospectively followed a large (n = 736), representative cohort of people experiencing homelessness recruited at random from 62 sites in Toronto, Canada between June and September 2021. Participants were interviewed up to five times over approximately 12 months.
Longitudinal patterns of twelve self-reported symptoms were assessed through latent transition analysis, and generalized estimating equations with logit link were applied to determine their association with known risk of potential PCC.
Results
Among 736 participants, three latent statuses were identified: (1) ‘No/Few Symptoms’ (≥70 %), (2) ‘Non-Specific Symptoms’ (15–23 %), and (3) ‘Infection-Related Symptoms’ (≤5 %). Statuses 2 and 3 were associated with being at risk of PCC following symptomatic infection (aOR 3.41 [95 % CI 2.3–5.0] and 3.18 [95 % CI 1.6–6.4]) but not with being at risk of PCC overall.
Transition probabilities suggested PCC would mostly occur among individuals with symptoms at baseline. However, the clustered area under the curve was modest (0.70 [95 % CI 0.65–0.75]), indicating symptom-based approaches are suboptimal for identification of potential PCC.
Conclusions
Self-reported symptoms do not reliably identify potential PCC among people experiencing homelessness, due to high rates of underlying symptoms and asymptomatic infections. Alternative, strengths-based approaches are recommended to more equitably identify post-COVID condition in this population.
Web | DOI | PDF | Public Health | Open Access