Using Hospital Inpatient Discharge Data to Supplement Active Surveillance for Invasive Pneumococcal Disease: Is the Extract Worth the Exertion?

OBJECTIVE

Invasive pneumococcal disease (IPD) surveillance systems monitor morbidity, mortality, and vaccine impact; accurate surveillance is important to detect changes in epidemiology. We evaluated completeness of IPD reporting in New Mexico by comparing data from the Hospital Inpatient Discharge Database (HIDD) and the New Mexico Active Bacterial Core Surveillance (ABCs) program.

METHODS

We linked data from the HIDD and the ABCs program. We defined cases of IPD in the HIDD among New Mexico residents with hospitalizations during 2007-2009 as specific (320.1 or 038.2) or nonspecific (481, 320.2, or 041.2) using International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes. To validate if HIDD records that could not be matched to ABCs data were true IPD cases, we reviewed laboratory data and determined if Streptococcus pneumoniae (S. pneumoniae) had been isolated from a sterile body site.

RESULTS

We examined 732 HIDD records for cases that were not matched in the ABCs database; of such records, S. pneumoniae was isolated from a sterile body site in 10 HIDD records.

CONCLUSION

ABCs data detected the majority of IPD cases in New Mexico. Laboratory and medical record review is essential when using HIDD data because ICD-9-CM coding alone does not ensure data accuracy. The addition of IPD cases to the ABCs program from the HIDD was minimally beneficial to active surveillance and reporting completeness in New Mexico. States that rely exclusively on passive reporting and that have access to HIDD data might use linkages of pneumococcal and IPD-specific ICD-9-CM-coded HIDD data to improve IPD surveillance and case ascertainment.