Non-attendance in Telephone versus In-person Secondary Care Consultations: Retrospective cohort Study of Patients with Type 2 Diabetes in Northwest London

BackgroundAlthough virtual consultations are increasingly used in healthcare, mode affects attendance patterns remains limited, particularly across demographic groups. Within NHS secondary care, telephone consultations have been the most widely adopted form of telephone care; however, few studies have examined non-attendance (commonly termed Did Not Attend [DNA]) patterns specifically for telephone consultations and fewer still have explored how patient characteristics influence attendance differently across consultation modes. Understanding these patterns is essential for equitable service planning. ObjectiveTo compare non-attendance rates between telephone and in-person secondary care consultations among adults with type 2 diabetes (T2D), and to identify patient characteristics associated with non-attendance under each mode. MethodsData from 853,693 secondary care consultations (January 2020-August 2024) for 45,618 patients with T2D in Northwest London were analysed. Telephone consultations in this dataset consisted exclusively of telephone consultations; we therefore refer to them as telephone consultations throughout. Patient-level consultations were aggregated into patient-mode strata for regression modelling. Zero-inflated Negative Binomial regression assessed factors associated with missed consultation rates by mode (in-person or telephone). Propensity-score balance diagnostics (inverse probability of treatment weighting) were conducted to assess measured confounding by mode assignment. Specialty-stratified non-attendance rates were examined across 34 major specialties. ResultsIn-person consultations had higher unadjusted non-attendance rates than telephone consultations (9.1% vs 7.2%, p<0.001). This pattern was consistent for both first consultations (9.3% vs 6.2%, p<0.001) and follow-up consultations (9.0% vs 7.50%, p<0.001). For in-person consultations, higher non-attendance was associated with younger age (18-39: 12.2%, 40-59: 11.1% vs 60-79: 9.9%, p<0.001), Black or Black British ethnicity (18.9% vs White: 7.6%, p<0.001), and greater deprivation (most deprived IMD1: 10.3% vs least deprived IMD5: 7.0%, p<0.001). For telephone consultations, higher non-attendance was associated with male gender (7.3% vs female: 7.0%, p<0.01), younger age (18-39: 11.3%, 40-59: 9.5% vs 60-79: 6.1%, 80+: 5.6%, p<0.001), and greater socioeconomic deprivation (most deprived: 8.3% vs least deprived: 4.7%, p<0.001). Interaction analyses revealed that demographic disparities were amplified for telephone relative to in-person consultations. Specialty-stratified analysis showed that in-person non-attendance exceeded telephone non-attendance in the majority of high-volume specialties. ConclusionsIn-person consultations had higher non-attendance rates than telephone consultations. Various demographic factors influenced non-attendance rates, with younger age and greater socioeconomic deprivation consistently associated with non-attendance for both in-person and telephone consultations. These findings suggest that a personalised, equity-informed approach to consultation mode selection is needed. Findings apply to telephone consultations and may not generalise to video-based modalities.