The slow de‐implementation of non‐evidence‐based treatments in low back pain hospital care—Trends in treatments using Dutch hospital register data from 1991 to 2018

BACKGROUND: Low back pain (LBP) is the leading cause of disability worldwide and has an excessive societal burden. Accumulating evidence has shown that some medical approaches such as imaging in absence of clear indications, medication and some invasive treatments may contribute to the problem rather than alleviating it. OBJECTIVES: To determine the extent of de‐implementation of non‐evidence‐based hospital treatments for LBP care in the Netherlands in the last three decades. METHODS: Using a register‐based population‐level observational study with Dutch hospital data, providing a nearly complete coverage of hospital admissions in the Netherlands in 1991–2018, we assessed five frequently applied non‐evidence‐based hospital treatments for LBP. Time trends in treatment use (absolute and per 100,000 inhabitants) were plotted and analysed using Poisson regression. RESULTS: The use of bed rest for non‐specific LBP and hernia nuclei pulposi, and discectomy for spinal stenosis decreased 91%, 81% and 86% since the availability of evidence/guidelines, respectively. De‐implementation, beyond 84%, was reached after 18 and 17 years for bed rest for non‐specific LBP and discectomy respectively, while it was not reached after 28 years for bed rest for hernia nuclei pulposi. For spinal fusion and invasive pain treatment, there was an initial increase followed by a reduction. Overall, these treatments reduced by 85% and 75%, respectively. CONCLUSIONS: In the Netherlands, de‐implementation of five non‐recommended hospital LBP treatments, if at all, took several decades. Although de‐implementation was substantial, slow de‐implementation has likely resulted in considerable waste of resources and avoidable harm to many patients in Dutch hospitals. SIGNIFICANCE: Medically intensive approaches to low‐back pain care contribute to the high societal burden of this disease. There have been calls to avoid such care. Using Dutch hospital data, we showed that de‐implementation of five non‐recommended hospital low‐back pain treatments, if at all, took several decades (i.e. ≥17 years) after availability of evidence and guidelines. Slow de‐implementation has likely resulted in considerable waste of resources and avoidable harm to hospital patients; better ways for de‐implementation of non‐evidence‐based care are needed.