Patient-friendly summary
If you read nothing else
Bottom line
A national education-and-feedback program for GPs was associated with a statistically significant reduction in low-value CT scans for low back pain and was cost saving for the health system, though as an observational study it cannot prove the program directly caused the change.
Moderate evidencePublished
Evidence hierarchy
Study participants
Australian general practitioners and GP registrars referring patients for Medicare-subsidised lower-back imaging, with non-GP health professionals as a control
Study Summary
This Australian study evaluated a 2013 national education program (NPS MedicineWise) that aimed to reduce unnecessary lower-back imaging for non-specific acute low back pain by general practitioners (GPs). The program — a personalised imaging-referral feedback report, a patient self-management pad, and an online decision-support tool — was sent to 19,997 GPs (about 60% of Australian GPs). Using a Bayesian structural time-series analysis of national Medicare claims with non-GP referrals as a control, the program was associated with a statistically significant 10.85% relative reduction in GP-referred CT scans of the lumbosacral region, but no measurable change in X-ray volume. The reduction translated into an estimated AUD$11.6 million in saved Medicare reimbursements against AUD$141,154 in program costs, so each dollar spent was estimated to save about AUD$82.
Key Findings
| Finding | Detail | Impact |
|---|---|---|
| CT scans of the lower back dropped by an estimated 10.85% after the program | Following the June 2013 intervention, observed GP-referred CT scans of the lumbosacral region were 10.85% lower than the modelled counterfactual, a statistically significant reduction. This equated to an estimated 50,186 fewer GP-referred CT scans between July 2013 and February 2015 (95% posterior interval 3,919 to 96,476). | High |
| The program was cost saving from the government funder's perspective | The CT reduction corresponded to an estimated AUD$11,600,898 cut in Medicare reimbursement (95% posterior interval $1,093,011 to $22,100,313) against program costs of AUD$141,154, giving a cost-benefit ratio of about 1:82 and a net benefit of roughly AUD$11.4 million. | High |
| Low cost per CT scan averted, robust to sensitivity analysis | The incremental cost-effectiveness ratio was AUD$2.82 per lumbosacral CT scan averted versus no program, far below the AUD$231 average Medicare cost of such a scan. The program had an ICER below AUD$231 in 98.3% of probabilistic-analysis iterations. | Medium |
| No effect on X-ray imaging volume | No association was observed between the program and GP-referred X-ray items, and there was no evidence that averted CT scans were being replaced by additional X-rays. The Bayesian model could not be applied to X-rays due to lack of stationarity, and a conventional time-series model showed no impact. | Medium |
Following the June 2013 intervention, observed GP-referred CT scans of the lumbosacral region were 10.85% lower than the modelled counterfactual, a statistically significant reduction. This equated to an estimated 50,186 fewer GP-referred CT scans between July 2013 and February 2015 (95% posterior interval 3,919 to 96,476).
The CT reduction corresponded to an estimated AUD$11,600,898 cut in Medicare reimbursement (95% posterior interval $1,093,011 to $22,100,313) against program costs of AUD$141,154, giving a cost-benefit ratio of about 1:82 and a net benefit of roughly AUD$11.4 million.
The incremental cost-effectiveness ratio was AUD$2.82 per lumbosacral CT scan averted versus no program, far below the AUD$231 average Medicare cost of such a scan. The program had an ICER below AUD$231 in 98.3% of probabilistic-analysis iterations.
No association was observed between the program and GP-referred X-ray items, and there was no evidence that averted CT scans were being replaced by additional X-rays. The Bayesian model could not be applied to X-rays due to lack of stationarity, and a conventional time-series model showed no impact.
Strengths
- National implementation and evaluation, providing a census of all GP-referred, Medicare-funded CT scans and X-rays rather than a sampled subset
- Use of administrative claims data avoided self-report bias
- No selection bias from selective recruitment, and the study was well powered to detect change
- Used non-GP referrals as an observational control to account for unmeasured factors influencing imaging volume
- Cost-saving conclusion was robust across best/worst-case and probabilistic sensitivity analyses
Limitations
- Observational design: causation cannot be inferred, though no other competing national program or guideline change was identified during the period
- Program costs were estimated retrospectively rather than measured prospectively
- Medicare data do not include reasons for imaging, so the appropriateness of the reduced referrals (e.g., any missed or delayed diagnoses) could not be assessed
- Possible substitution of imaging to medical specialists could not be excluded
- Downstream benefits (reduced radiation/cancer risk) and downstream costs (specialist referrals, delayed diagnosis) were outside the study scope
- The sustainability of the effect over longer periods was not evaluated
Key Takeaways for Patients
What This Means for You
- 01For most people with new low back pain and no warning signs, scans like CT or X-ray are not needed and do not improve outcomes — Australian guidelines have advised against routine imaging for over a decade.
- 02Avoiding an unnecessary CT scan also avoids unnecessary radiation exposure, which carries a small cancer risk, especially relevant over a lifetime of tests.
- 03This study looked at whether doctors changed their ordering habits, not at individual patients, so it does not mean any specific person should skip imaging — your doctor still checks for 'red flag' symptoms that may require a scan.
- 04Giving GPs feedback on their own ordering patterns, plus decision tools and patient handouts, was linked to fewer unnecessary CT scans and substantial savings to the health system.
Read the Full Paper
Access the complete peer-reviewed study from BMC Health Services Research
View Full Study