Evaluating the effect of data standardization and validation on patient matching accuracy

Area of Science:

• Health Informatics
• Data Science
• Patient Data Management

Background:

• Accurate patient matching is vital for healthcare quality and safety.
• Existing datasets often suffer from inconsistencies hindering effective patient identification.
• Standardization of demographic data is proposed as a solution to improve matching.

Purpose of the Study:

• To evaluate the impact of demographic data standardization on patient matching accuracy.
• To assess the effectiveness of standardizing specific fields like address and last name.
• To determine the influence of standardization across different real-world datasets.

Main Methods:

• Utilized four manually reviewed datasets: HIE, public health registry, Social Security Death Master File, and newborn screening records.
• Standardized key demographic fields: last name, telephone number, social security number, date of birth, and address.
• Evaluated matching performance using sensitivity, specificity, positive predictive value, and accuracy metrics.

Main Results:

• Address standardization independently improved matching sensitivity in public health and HIE datasets (0.6%–4.5%), though overall accuracy was unchanged due to reduced specificity.
• Last name standardization improved HIE dataset sensitivity by 0.6%, with unchanged overall accuracy.
• Combined standardization of address and last name significantly boosted HIE dataset sensitivity from 81.3% to 91.6%.

Conclusions:

• Data standardization, especially for address and last name, can substantially improve patient match rates.
• Enhanced match rates lead to better data for clinical decision-making, improving care quality and safety.
• Targeted data standardization offers a pathway to more reliable patient identification in healthcare systems.