[Linking Interventional Cardiology clinical registry data with French hospital administrative data: Development and validation of deterministic record linkage].

BACKGROUND

To recreate the in-hospital healthcare pathway for patients treated with coronary angiography or percutaneous coronary intervention, we linked the interventional cardiology registry (ACIRA) and the pseudonymized French hospital medical information system database (PMSI) in the Aquitaine region. The objective of this study was to develop and validate a deterministic merging algorithm between these exhaustive and complementary databases.

METHODS

After a pre-treatment phase of the databases to standardize the 11 identified linking variables, a deterministic linking algorithm was developed on ACIRA hospital stays between December 2011 and December 2014 in nine interventional cardiology centers as well as the data from the consolidated PMSI databases of the Aquitaine region from 2011 to 2014. Merging was carried out through 12 successive steps, the first consisting in strict linking of the 11 variables. The performance of the algorithm was analyzed in terms of sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV). Strategies complementary to the initial algorithm (change in the order of variables and base preprocessing) were tested. Comparative analysis of merged/unmerged patients explored potential causes of mismatch.

RESULTS

The algorithm found 97.2% of the 31,621 ACIRA stays to have sensitivity of 99.9% (95% CI [99.9; 99.9]), specificity of 97.9% (95% CI [97.7; 98.1]), PPV of 99.9% (95% CI [99.9; 99.9]) and NPV of 96.9% (95% CI [96.7; 97.1]). Complementary strategies did not yield better results. The unmerged patients were older, and hospitalized mostly in 2012 in two interventional cardiology centers.

CONCLUSION

This study underscored the feasibility and validity of an indirect deterministic pairing to routinely link a registry of practices using hospital data to pseudonymized medico-administrative databases. This method, which can be extrapolated to other health events leading to hospitalization, renders it possible to effectively reconstruct patients' hospital healthcare pathway.