![[ Visit Client Website ]](images/banner.png)
BACKGROUND: Causes of American Indian and Alaska Native (AI/AN) hospitalizations in the Pacific Northwest are insufficiently understood, partly due to the lack of an Indian Health Service (IHS) hospital in the Portland IHS Area, as well as high rates of AI/AN racial misclassification in state public health databases. The Northwest Tribal Registry (NTR), maintained by the Improving Data and Enhancing Access - Northwest (IDEA-NW) Project, is a demographic dataset of known AI/AN, composed from patient rosters from the Portland IHS, three urban Indian health clinics, and Northwest Tribes’ enrollment lists. Using this data source, IDEA-NW links with a variety of state public health databases to quantify prevalence of AI/AN misclassification, analyze race-corrected data, and communicate findings to Tribes. This was the first linkage study with Oregon hospitalization data.
METHODS: Using probabilistic record linkage methods, we corrected for AI/AN misclassification in Oregon inpatient hospital discharge records for 2010-2011 by comparing the NTR to the Oregon database. A case of AI/AN racial misclassification was defined as a state record that matched an NTR record and was either coded as a race other than AI/AN or missing race data in the state database; results were reported to the state. We analyzed de-identified race-corrected data to determine the major underlying causes of AI/AN hospitalizations in Oregon, using ICD-9 principal diagnosis (2013 HCUP Clinical Classifications Software).
RESULTS: From 2010-2011, 55.4% of records that matched to the NTR were misclassified as non-AI/AN in the state dataset. Of those misclassified, 66.5% were coded as white, and 11.9% were missing race information. Our linkage increased ascertainment of AI/AN in the Oregon database by 31.8%. AI/AN hospitalizations tended to be younger (mean age = 39.1 years vs. 51.4 years for non-Hispanic Whites (NHW)), and were more likely to have Medicaid and governmental sources (both IHS and non-IHS) as primary payer. The most frequent principal diagnoses among AI/AN hospitalizations were complications of pregnancy and childbirth (17.5%), conditions originating in the perinatal period (12.7%), and diseases of the digestive system (9.8%).
CONCLUSIONS: To our knowledge, this is the first report based on linked data between the Oregon hospital discharge dataset and Indian health databases. This project demonstrated how data linkage can increase the quality of hospitalization data for AI/AN populations, and emphasizes the need for high quality data to better understand causes and trends of hospitalization, and ultimately improve outcomes for AI/AN patients.