EPH34 long COVID symptoms and diagnosis in primary care: a cohort study using the thin database including unstructured text

Background Long COVID is a widely recognised consequence of COVID-19 infection, but there is currently little information about the symptoms that patients present with in primary care, as these are typically recorded only in free text clinical notes. Objectives To compare the occurrence of symptoms in patients with and without a history of COVID-19 infection. Methods We used data from the The Health Improvement Network (THIN, a Cegedim Database). We included patients aged 18 or over registered with participating practices in England, Scotland or Wales. We compared COVID-19 cases (defined by Read-coded diagnoses) with unexposed controls. We extracted baseline and symptom information from both the Read codes and free text, which was analysed by natural language processing. We calculated hazard ratios (adjusted for age, sex, baseline medical conditions and prior record of symptoms) for each of 89 symptoms from 12 weeks after the COVID-19 diagnosis. Results We compared 10,229 patients with confirmed COVID-19 and 22,654 unexposed controls. Patients had a mean age of 52.4 years and 62.5% were female. Around 80% of the symptom mentions in the primary care were only in the free text. A wide range of symptoms were associated with previous COVID-19 infection, including shortness of breath (hazard ratio (HR) 3.2, 95% confidence interval (CI) 2.8, 3.5), chest pain (HR 2.35, 95% CI 2.03, 2.72) and fatigue (HR 3.4, 95% CI 3.0, 3.8). There were 603 free text entries of ‘Long Covid’, but none recorded using Read codes. Conclusions These preliminary results show that there are numerous symptoms which are more commonly recorded after COVID-19 infection. There is a lack of structured recording of symptoms and Long COVID diagnoses, showing the importance of free text in health records for studying these topics