Six-month outcomes after a GP phone call during the first French COVID-19 lockdown (COVIQuest): a cluster randomised trial using medico-administrative databases

Abstract

Objectives: The first COVID-19 lockdown raised concerns about reduced access to primary care, especially for people with chronic diseases particularly at risk in the absence of follow-up. However, the COVIQuest trial, evaluating the impact of a general practitioner (GP) phone call (intervention) to chronic patients with cardiovascular disease (CVD) or mental health disorder (MHD) concluded that the intervention had no effect at 1 month on the rate of self-reported hospitalisations in the CVD subtrial, whereas the intervention group in MHD subtrial might have a higher rate. This second part of the study aimed to describe the 6 month hospitalisation and specialised consultation rates, using the French health data system (Système National des Données de Santé). The secondary objective was to describe these rates during the same period in 2019.

Design: A cluster randomised controlled trial, with clusters being GPs.

Setting: Primary care, 149 GPs from eight French regions.

Participants: Patients ≥70 years old with chronic CVD or ≥18 years old with MHD.

Interventions: A standardised GP-initiated phone call aiming to evaluate patient's need for urgent care (vs usual care for control groups).

Primary and secondary outcome measures: The occurrence of at least one hospitalisation at end point 31 October 2020 (randomisation 30 April 2020), excluding those starting on 30 April 2020, was measured as planned. Another main outcome was the occurrence of at least one specialised consultation during the same period. These 6 month effects were studied, using a logistic regression model within a generalised estimating equation framework, for each subtrial.

Results: 4640 patients were included: 3274 cardiovascular (mean age 79.9±7.0 years; 57.8% male) and 1366 psychiatric (53.2±7.0; 36.5%). For both subtrials, the intervention patients were significantly more hospitalised than the control patients, respectively, 17.3% versus 14.9% of CVD patients (OR=1.26 (1.05 to 1.52)); 14.4% versus 10.7% of MHD patients (OR=1.40 (1.00 to 1.96)). During the same period in 2019, the hospitalisation rates were, respectively, 16.3%, 18.2%, 15.8% and 14.8%. The proportions of patients with at least one specialised consultation were not different between the intervention and control groups, respectively, 24.6% versus 24.3% for CVD patients (OR=1.06 (0.85 to 1.32)); 26.5% versus 24.4% for MHD patients (OR=1.15 (0.84 to 1.57)). During the same period in 2019, these rates were, respectively, 22.7%, 24.6%, 28.0% and 25.5%.

Conclusions: The intervention was associated with higher rates of hospitalisation at 6 months in patients with MHD or CVD. No intervention impact was found in outpatient care. These results are difficult to interpret because of a potential artefact induced by national campaigns promoting medical use during lockdown, overlapping the study inclusion period. This study showed that medico-administrative databases could represent a complementary cost-effective tool to clinical research for long-term and healthcare consumption outcomes.

Trial registration number: NCT04359875.

Keywords: COVID-19; Cardiovascular Disease; Chronic Disease; Hospitalization; MENTAL HEALTH; Primary Health Care.