#### Abstract *Objective* To estimate the incremental public healthcare utilisation and costs, incidence of comorbid conditions, and socioeconomic outcomes following newly diagnosed osteoarthritis in primary care in New Zealand. *Design* Matched longitudinal cohort study using routinely collected administrative data from the New Zealand Integrated Data Infrastructure. Patients with a first diagnosis of hip or knee osteoarthritis between 2009 and 2010 were identified from primary care electronic health records and matched with patients without osteoarthritis diagnosis across a large set of pre-diagnosis variables. Extensive diagnostic checking was undertaken to establish a very close balance of covariates between matched osteoarthritis and non-osteoarthritis cohorts. Healthcare utilisation, healthcare costs, incidence of comorbid health conditions, income and employment, receipt of social welfare benefits, and mortality were measured in administrative record data over 10-year follow-up and compared between the osteoarthritis and matched non-osteoarthritis cohorts by linear regression. *Results* We successfully matched 792 patients with newly diagnosed osteoarthritis with 35 904 controls without osteoarthritis diagnosis. The matched case and control cohorts were very well balanced across all baseline covariates. Compared with matched controls, osteoarthritis patients had significantly higher healthcare utilisation and healthcare costs over the 10 years following diagnosis, with total incremental public healthcare costs of \$10 771 (95\% CI: 3 047, 18 494) per patient. There was no consistent evidence of differences in socioeconomic outcomes, mortality, or incidence of comorbid conditions. *Conclusions* People with osteoarthritis incur substantial healthcare costs in the decade following primary care diagnosis.