Effective referral from general practitioners to orthopaedic services and timely decisions about surgery are important for health service productivity and patient care quality. The efficiency of the referral-to-decision pathway, including variation between GPs, resource use, costs and conversion to surgery, is poorly characterised. This study evaluated GP referrals to a public orthopaedic service up to the point of surgical determination, aiming to estimate GP-level variability in referral volumes and types, orthopaedic resource usage, surgical conversion rates and the time and costs of care up to surgical decision or discharge.
The investigation was a retrospective observational cohort with a cost and outcomes analysis covering a two-year period. The study included all GP orthopaedic referrals to a large regional Australian public health service. Care provision was orthopaedic-led, and the analysis captured health service activity and costs incurred until surgical determination or discharge.
Time-to-event analyses used the Kaplan–Meier estimator to accommodate right-censored data. Cost estimates were adjusted for inflation and handled using inverse probability censoring to address censoring in cost accumulation. Costs are reported in Australian dollars from the health service perspective. The methods reported are those used to derive time, conversion and cost outcomes in the cohort.
A total of 18,901 GP orthopaedic referrals were identified in the two-year cohort. Referral volume per GP showed marked heterogeneity: individual GPs generated between 1 and 238 referrals during the study period. The study highlights substantial GP-level variability in referral behavior and volume within a single regional public service.
Time from referral to a surgical decision was frequently prolonged and highly variable. The mean time to surgical decision was reported as 520 days. Observed times spanned a wide interval, with some cases reaching a decision in approximately 1 week while others extended to around 4 years. The authors note prolonged non-surgical care in many pathways and high variability in the interval to surgical determination.
Mean pre-surgical health service cost per referral was AUD $785 (standard error 11.89). Costs were influenced by the duration of the pathway: longer time intervals and cases that progressed to surgical listing were associated with greater pre-surgical costs. All costs reported were calculated from the health service perspective and adjusted for inflation as described in the methods.
The surgical conversion rate in the cohort was low at 25%, meaning one quarter of referrals proceeded to surgery. The study also reports high drop-out rates, indicating that a substantial proportion of referred patients did not complete the pathway to a surgical decision or to surgery. Longer pathways and listings were linked to increased costs, while overall conversion remained modest.
This single-centre evaluation found pronounced variability in GP referral volumes and extended periods of non-surgical orthopaedic care before a surgical decision. The cohort demonstrated a low surgical conversion rate (25%), long mean time to decision (520 days), and mean pre-surgical cost of AUD $785 per referral. The authors conclude that further research is required to understand the drivers of variation in GP referral parameters, the reasons for prolonged non-surgical management and drop-out, and to explore strategies to better align referral objectives between primary care and orthopaedics and to improve service efficiency.
All data and outcomes derive from a single regional Australian public health service over a two-year window and reflect costs from the health service perspective. The study design was observational and retrospective; the abstract reports the core findings, methods and numerical results noted above. Details beyond what is contained in the abstract (for example, granular referral types, patient-level characteristics or specific reasons for drop-out) were not reported in the source abstract and would require access to the full text for further detail.