Optimal duration of bisphosphonate therapy was unknown until the FLEX study was published in 2006 showing a 5‐year course to be adequate for most women. In 2008 a link between long‐term bisphosphonate and atypical femoral fractures was reported and confirmed in later studies. We hypothesized these landmark observations should have led to a decrease in use of bisphosphonates for >5 or 10 years, from 2010 onward. The Manitoba BMD registry with linkage to provincial pharmacy data was used to determine the percentage of long and very‐long term bisphosphonate users from therapy start. The cohort comprised women age > 50 with BMD between 1995–2018 with oral bisphosphonate first prescribed for >90 days with adherence >75% in the first year. For each calendar year of continued therapy, the percentage of patients and medication possession rate was tabulated. The percentage of users beyond 5 years was compared between patients who started therapy in 1998–2004 (those taking 5 years of therapy still finish prior to 2010) versus 2005–2012 (all new therapy starts overlap 2010 in those taking ≥5 years treatment). The cohort included 2991 women with mean follow up 8.8(1.3) years, 64.9% of whom took continuous oral bisphosphonate for >5 years and 41.9% for >10 years. In the earlier vs later era, there were 74.4% vs 70.2% who completed 5 years. With respect to longer treatment, there were 68.0% and 60.5% of patients treated for 6 or more years (p < 0.0001) and 46.6% vs 33.5% treated for >10 years (p = 0.08). Medication possession rate was >79% in every year of therapy. Landmark studies leading to more limited bisphosphonate courses may have slightly reduced longer‐term treatment, but up to one‐third of adherent patients in the modern era still receive continuous bisphosphonate therapy for >10 years.This article is protected by copyright. All rights reserved.