A Quality Improvement Analysis of Death Ascertainment Among Utah HIV/AIDS Cases

BACKGROUND:  In 2011, 2,614 Utahns were reported to be living with HIV/AIDS.  Accurate surveillance and death ascertainment are essential to evaluate the impact and burden of HIV.  The objective of this analysis was to assess two techniques of linking death certificate records and HIV surveillance data in order to determine best practices for death ascertainment in Utah.

METHODS:  Electronic record linkage was utilized to match HIV cases reported in the Utah Enhanced HIV/AIDS Reporting System (eHARS) with vital statistics death certificate records from 1985 to 2011.  The variables used for record linkage were first and last name, gender, date of birth, and Social Security Number.  Cause of death information was available from 2000 to 2011.  The death certificate records analyzed included only those cases with underlying or contributing ICD-10 codes directly pertaining to HIV infection (B20-B24).  A subsequent manual search matched these data to eHARS records in order to ascertain additional deaths and HIV cases not previously identified within this population.

RESULTS:  Using electronic record linking, 38 deaths that had not been previously identified in eHARS were identified.  Manual search of death certificate records identified 260 individuals with HIV-associated underlying or contributing codes.  Among these, the proportion previously documented in eHARS as HIV cases and as deceased HIV cases was 94% (n=245) and 89% (n=220), respectively.  Six additional deaths among HIV cases were identified through manual search that had not been detected by electronic linkage. Fifteen (6%) individuals with death certificate codes associated with HIV were newly identified as cases not previously reported to the state health department.  

CONCLUSIONS:  The results of this study suggest that electronic record linking may identify deaths among HIV/AIDS cases not previously identified in eHARS through routine surveillance activities.  However, electronic linking may not be sufficient to ascertain all deaths among HIV/AIDS cases.  In areas with lower HIV prevalence and incidence, manually searching those with HIV-associated underlying and contributing codes may be feasible to help detect additional deaths and cases not identified by electronic record linking alone.  The combined methodology utilized in this analysis may also be applicable to the surveillance of similar diseases or conditions.  Additionally, linking to death records may be a way to ascertain some HIV/AIDS cases that have been previously missed through routine case surveillance activities.