Considering the full care pathway in regional variation in paediatric otitis media treatment in the Netherlands: an observational study

Abstract

Objective Ventilation tube insertion for paediatric otitis media (POM), including acute otitis media (AOM) and otitis media with effusion (OME), has been signalled in the past for potential unwarranted treatment variation. Quality improvement initiatives, like Audit & Feedback (A&F), often ignore the care pathway when identifying such variation, possibly overestimating variation at a specific care step. To gain more insight into the effect of prior care steps, this study examined (1) the degree of regional variation in each step of the care pathway (general practitioner (GP) contacts, referrals and surgeries) and (2) investigated the effect of adjusting for prior care steps. Design Observational study using general practice electronic health record data linked to specialist claims data. Participants 272 790 children ≤12 years with and without POM registered in 320 GP practices between 2017 and 2018. Primary and secondary outcomes Using multilevel logistic regression, the degree of regional variation in each step of the POM care pathway was assessed by calculating the coefficient of variation (CV). The effect of adjusting for prior care steps was determined by estimating correlations between subsequent care steps and analysing the impact on the CV. Results Regional variation in POM treatment was larger in each subsequent step in the care pathway (CV POM GP contacts 0.110; referral 0.179; surgery 0.239). In regions with a higher proportion of children with frequent AOM/ persistent OME, referral rates were higher (POM: OR: 1.06; 95% CI: 1.02 to 1.11) and surgical rates were higher (for OME only: OR: 1.08; 95% CI: 1.02 to 1.15). Regional variation in referrals and surgery decreased after adjusting for the regional frequent AOM/persistent OME rate (CV referrals POM 0.103 vs 0.128; CV surgery OME 0.047 vs 0.059). Conclusions Regional variation is observed in GP contact rates for POM and is larger in referrals and surgeries. Adjusting for the proportion of frequent AOM/ persistent OME significantly reduces regional variation in POM treatment. Future A&F should adjust for prior care processes and develop tailored interventions for quality improvement.