Department for Healthcare of Older People, Queens Medical Centre, Nottingham University Hospital NHS Trust, Nottingham, UK.
Faculty of Medicine, University of Malaya, 50603, Kuala Lumpur, Malaysia.
Osteoporos Int. 2021 May;32(5):921-926. doi: 10.1007/s00198-020-05710-8. Epub 2020 Nov 10.
Integration of a vertebral fracture identification service into a Fracture Liaison Service is possible. Almost one-fifth of computerised tomography scans performed identified an individual with a fracture. This increase in workload needs to be considered by any FLS that wants to utilise such a service.
This service improvement project aimed to improve detection of incidental vertebral fractures on routine imaging. It embedded a vertebral fracture identification service (Optasia Medical, OM) on routine computerised tomography (CT) scans performed in this hospital as part of its Fracture Liaison Service (FLS).
The service was integrated into the hospital's CT workstream. Scans of patients aged ≥ 50 years for 3 months were prospectively retrieved, alongside their clinical history and the CT report. Fractures were identified via OM's machine learning algorithm and cross-checked by the OM radiologist. Fractures identified were then added as an addendum to the original CT report and the hospital FLS informed. The FLS made recommendations based on an agreed algorithm.
In total, 4461 patients with CT scans were retrieved over the 3-month period of which 850 patients had vertebra fractures identified (19.1%). Only 49% had the fractures described on hospital radiology report. On average, 61 patients were identified each week with a median of two fractures. Thirty-six percent were identified by the FLS for further action and recommendations were made to either primary care or the community osteoporosis team within 3 months of fracture detection. Of the 64% not identified for further action, almost half was because the CT was part of cancer assessment or treatment. The remaining were due to a combination of only ≤ 2 mild fractures; already known to a bone health specialist; in the terminal stages of any chronic illness; significant dependency for activities of daily living; or a life expectancy of less than 12 months CONCLUSION: It was feasible to integrate a commercial vertebral fracture identification service into the daily working of a FLS. There was a significant increase in workload which needs to be considered by any future FLS planning to incorporate such a service into their clinical practice.
本服务改进项目旨在提高对常规影像学检查中偶然发现的脊柱骨折的检出率。它将一个脊柱骨折识别服务（Optasia Medical，OM）嵌入到医院的骨折联络服务（FLS）中，作为其一部分，对常规计算机断层扫描（CT）进行检查。
该服务整合到医院的 CT 工作流程中。前瞻性地检索了 3 个月内年龄≥50 岁的患者的 CT 扫描及其临床病史和 CT 报告。通过 OM 的机器学习算法识别骨折，并由 OM 放射科医师进行交叉检查。然后，将识别出的骨折作为原始 CT 报告的附录添加，并通知医院 FLS。FLS 根据商定的算法提出建议。
在 3 个月的时间内，共检索了 4461 例 CT 扫描患者，其中 850 例患者发现有脊柱骨折（19.1%）。仅有 49%的骨折在医院放射学报告中描述。平均每周发现 61 例患者，中位数为 2 例骨折。36%的患者由 FLS 确定需要进一步采取行动，并在骨折检测后 3 个月内向初级保健或社区骨质疏松团队提出建议。对于不需要进一步采取行动的 64%的患者，几乎一半是因为 CT 是癌症评估或治疗的一部分。其余的是由于≤2 处轻度骨折；已经被骨健康专家知晓；处于任何慢性疾病的晚期；日常生活活动严重依赖；或预期寿命不足 12 个月。
将商业性脊柱骨折识别服务整合到 FLS 的日常工作中是可行的。这会显著增加工作量，任何未来计划将此类服务纳入其临床实践的 FLS 都需要考虑这一点。
