Measuring Mortality Information in Clinical Data Warehouses

Area of Science:

• Clinical research methodology
• Health informatics
• Data quality assessment

Background:

• Accurate long-term outcome tracking, particularly mortality, is vital for robust clinical research.
• Existing clinical research databases may have variable quality in mortality data.

Purpose of the Study:

• To present a framework for evaluating the quality of mortality information within clinical research databases.
• To assess the reliability of mortality data using a specific institutional case study.

Main Methods:

• Utilized Columbia University Medical Center's clinical data warehouse (CDW) for a case study.
• Compared vital status between the institution's patient registration system and the U.S. Social Security Administration's Death Master File (DMF).
• Quantified discrepancies including missing deaths, post-mortem encounters, and centenarian/supercentenarian overcounting.

Main Results:

• 39.5% of recorded deaths were absent from the DMF.
• 2.1% of patients listed as deceased in the DMF had subsequent clinical visits.
• 43.6% and 43.1% of patients in the CDW were erroneously marked as living over 100 and 120 years, respectively.

Conclusions:

• The developed framework provides valuable metrics for assessing mortality data quality in clinical research.
• Significant data quality issues were identified, emphasizing the need for data validation and improvement strategies.
• These quality measures can aid researchers in enhancing the reliability of long-term outcome reporting.