Geara Sammy-Jo, Nygren Clara, Borgfeldt Christer, Bergman Ida
Division of Obstetrics and Gynecology at Södersjukhuset, Stockholm, Sweden.
Statistical Department GynOp Register, Norrlands University Hospital, Umeå, Sweden.
Acta Obstet Gynecol Scand. 2025 Jul;104(7):1390-1398. doi: 10.1111/aogs.15136. Epub 2025 Apr 27.
The Swedish National Quality Register in Gynecological Surgery (GynOp) is widely used in gynecological and obstetric departments throughout Sweden. With a coverage rate of 91%-98%, it collects a substantial amount of data, fulfilling a crucial role in quality assurance and research. This study aimed to assess the internal validity of key variables within the GynOp register.
A multicenter, single-blinded study was conducted across 10 healthcare providers. Data were collected for 150 patients at each center, with 25 patients selected from each sub-register, totaling 1500 patients. Source data were extracted from medical records and compared to the data in the GynOp register for validation. Agreement was evaluated by calculating the exact agreement and the within-interval agreement using Pearson correlation (0-1) for numerical variables and Cohen's kappa (0-1) for ordinal variables.
The primary finding was that the overall validity of key variables in the GynOp register was high when compared with source data. For example, the correlation between source data and GynOp register data was 1.00 for uterine weight and 0.97 for date of delivery, and exact agreement was 97% for grade of perineal rupture and 100% for pathological-anatomical diagnosis. Some variables such as ASA-classification, genital prolapse stage, and devices for intrauterine surgery showed moderate agreement.
The comparison between the GynOp register and source data demonstrates strong internal validity. This study has identified several areas for improvement, leading to the initiation of an update in the registration module of the sub-register for intrauterine surgery. Additionally, certain categorical variables, such as ASA classification and prolapse stage, should be grouped and presented as dichotomized data to enhance data quality. In the future, integrating extraction of data from the Swedish Perioperative Register (SPOR) and providing education to urogynecologists on the POP-Q system could further contribute to the improvement of data quality.
瑞典妇科手术国家质量登记册（GynOp）在瑞典各地的妇产科广泛使用。其覆盖率为91%-98%，收集了大量数据，在质量保证和研究中发挥着关键作用。本研究旨在评估GynOp登记册中关键变量的内部有效性。
在10家医疗服务机构开展了一项多中心、单盲研究。每个中心收集150例患者的数据，从每个子登记册中选取25例患者，共计1500例患者。从医疗记录中提取源数据，并与GynOp登记册中的数据进行比较以进行验证。对于数值变量，使用Pearson相关性（0-1）计算精确一致性和区间内一致性，对于有序变量，使用Cohen's kappa（0-1）评估一致性。
主要发现是，与源数据相比，GynOp登记册中关键变量的整体有效性较高。例如，源数据与GynOp登记册数据之间子宫重量的相关性为1.00，分娩日期的相关性为0.97，会阴裂伤分级的精确一致性为97%，病理解剖诊断的精确一致性为100%。一些变量，如ASA分级、生殖器脱垂阶段和宫内手术器械，显示出中等一致性。
GynOp登记册与源数据之间的比较显示出很强的内部有效性。本研究确定了几个需要改进的领域，从而启动了宫内手术子登记册登记模块的更新。此外，某些分类变量，如ASA分级和脱垂阶段，应进行分组并以二分数据形式呈现，以提高数据质量。未来，整合瑞典围手术期登记册（SPOR）的数据提取，并为泌尿妇科医生提供关于POP-Q系统的培训，可能会进一步有助于提高数据质量。
