Roifman Idan, Qiu Feng, Connelly Kim A, Wright Graham A, Farkouh Michael, Jimenez-Juan Laura, Wijeysundera Harindra C
Department of Medicine, Sunnybrook Health Sciences Centre, Toronto, Ontario, Canada.
Department of Medicine, Institute of Health Policy Management and Evaluation, Toronto, Ontario, Canada.
BMJ Open. 2018 Oct 8;8(10):e021370. doi: 10.1136/bmjopen-2017-021370.
Cardiovascular magnetic resonance (CMR) imaging is the gold-standard test for the assessment of heart function. Despite its importance, many jurisdictions lack specific billing codes that can be used to identify patient receipt of CMR in administrative databases, limiting the ability to perform 'big data' CMR studies. Our objective was to identify the optimal billing code combination to identify patients who underwent CMR using administrative data in Ontario.
Retrospective cohort study.
Quaternary care academic referral centre in Ontario, Canada.
We tested all billing code combinations in order to identify the optimal one to determine receipt of CMR. The reference gold standard was a list of all cardiothoracic magnetic resonance scans performed at Sunnybrook Health Sciences Centre between 1 January 2014 and 31 December 2016, verified by chart audit. We assessed the diagnostic performance (accuracy, sensitivity, specificity, positive predictive value and negative predictive value) for all code combinations.
Our gold-standard cohort consisted of 2339 thoracic MRIs that were performed at Sunnybrook Health Sciences Centre from 1 January 2014 to 31 December 2016. Of these, 2139 (91.5%) were CMRs and 200 (8.5%) were chest MRIs. We identified the most accurate billing combination for the determination of patient receipt of CMR. This combination resulted in an accuracy of 95.3% (95% CI 94.4% to 96.2%), sensitivity of 97.4% (95% CI 96.6% to 98.1%), specificity of 86.4% (95% CI 83.1% to 89.6%), positive predictive value of 96.9% (95% CI 96.1% to 97.6%) and negative predictive value of 88.4% (95% CI 85.4% to 91.5%).
Our study is the first to verify the ability to accurately identify patient receipt of CMR using administrative data, facilitating more robust population-based CMR studies in the future.
心血管磁共振成像（CMR）是评估心脏功能的金标准检测方法。尽管其很重要，但许多司法管辖区缺乏可用于在行政数据库中识别患者接受CMR检查情况的特定计费代码，这限制了开展“大数据”CMR研究的能力。我们的目标是确定最佳计费代码组合，以便利用安大略省的行政数据识别接受CMR检查的患者。
回顾性队列研究。
加拿大安大略省的四级医疗学术转诊中心。
我们测试了所有计费代码组合，以确定用于判定是否接受CMR检查的最佳组合。参考金标准是2014年1月1日至2016年12月31日在阳光布鲁克健康科学中心进行的所有心胸磁共振扫描列表，并经病历审核验证。我们评估了所有代码组合的诊断性能（准确性、敏感性、特异性、阳性预测值和阴性预测值）。
我们的金标准队列包括2014年1月1日至2016年12月31日在阳光布鲁克健康科学中心进行的2339例胸部MRI检查。其中，2139例（91.5%）为CMR检查，200例（8.5%）为胸部MRI检查。我们确定了用于判定患者接受CMR检查情况的最准确计费组合。该组合的准确性为95.3%（95%CI 94.4%至96.2%），敏感性为97.4%（95%CI 96.6%至98.1%），特异性为86.4%（95%CI 83.1%至89.6%），阳性预测值为96.9%（95%CI 96.1%至97.6%），阴性预测值为88.4%（95%CI 85.4%至91.5%）。
我们的研究首次验证了利用行政数据准确识别患者接受CMR检查情况的能力，为未来开展更强大的基于人群的CMR研究提供了便利。
