Post COVID-19 condition diagnosis among older people: findings from Swedish national register data (Preprint)

BACKGROUND Post COVID-19 condition (PCC) denotes the persistence of symptoms following Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) infection. PCC poses large disease burden worldwide. OBJECTIVE We aimed to explore the epidemiological characteristics of PCC diagnosis among older people in Sweden and to establish a machine learning tool to predict PCC risk. METHODS This nationwide study employed a cross-sectional design. Prevalence of PCC among older people in Sweden was calculated. Logistic regression models were employed to assess the associations between various factors and PCC. Prediction model based on extreme gradient boosting was constructed based on selected features. RESULTS Of 236,943 older people diagnosed with COVID-19, 2.5% of these were diagnosed with PCC. The odds of PCC diagnosis in women were lower (OR = 0.78, 95% CI: 0.73-0.82). The odds decreased for each additional year of age (OR = 0.98, 95% CI: 0.98-0.98). Higher odds of PCC were found in people with higher Charlson Comorbidity Index and who used more types of medication. The odds of PCC were the lowest for infections during Omicron-predominant period (OR = 0.09, 95% CI: 0.08-0.10). The predicted PCC risk was generally higher during early variant-dominant periods. CONCLUSIONS In summary, the prevalence of PCC was relatively low among older people in Sweden. A higher burden of comorbidities and more types of medication use were associated with increased odds of PCC. Register-based predictive models may help identify older people at higher risk of PCC.