Lessons Learned from Matching Infectious Disease Surveillance Data at the New York City Department of Health and Mental Hygiene

BACKGROUND: Sharing data within local health departments is important to identify co-infections and syndemics, identify communities at risk, and target programs to prevent disease transmission. In 2010, the New York City Department of Health and Mental Hygiene (NYC DOHMH) implemented CDC’s Program Collaboration and Service Integration (PCSI) initiative to increase collaboration across infectious disease programs including HIV, sexually transmitted diseases (STD), tuberculosis (TB), hepatitis B (HBV) and hepatitis C (HCV).  Guided by an advisory committee comprised of staff from all infectious disease programs, we matched infectious disease surveillance data, mortality data, and Hemoglobin A1C surveillance data. 

METHODS: Using a deterministic matching algorithm, we matched incident cases of TB, chlamydia, gonorrhea, and syphilis reported between 1/1/00 and 12/31/10; prevalent and incident cases of HIV, hepatitis B, and hepatitis C, reported before 12/31/10 and excluding those known to have died before 2000.  We next matched vital statistics data of NYC deaths occurring between 1/1/00 and 12/31/11.  Finally, Hemoglobin A1C registry data between 1/1/06 – 12/31/12 of persons with at least one A1C test of 6.5% or greater was matched. The final dataset was a relational, de-identified, line-listed dataset of all records, including those that matched and did not match.   

RESULTS: The final dataset included 840,248 people with one or more infectious disease; 13% had two or more infectious diseases. People with syphilis had the greatest proportion of matches with other diseases (64%), with 49% matching to HIV and 21% matching with gonorrhea.  52% of people with gonorrhea matched to another disease with 46% matching to chlamydia and 7% matching to HIV.  31% of people with HIV matched to another disease with 16% matching to HIV.  21% of people with TB matched to another disease with 14% matching to HIV.  16% of people with chlamydia matched to another disease and 11% of people with hepatitis B matched to at least one additional disease.  The prevalence of diabetes was highest among those with HCV (11%), followed by TB (9%), HIV (6%), and HBV (5%).  Across all diseases, a greater proportion of persons with >1 disease died between 2000 and 2011 as compared to those with only one disease.

CONCLUSIONS: Matching surveillance data can help us understand the prevalence of infectious disease syndemics and to appropriately target services to address them. Findings can also be used to inform programmatic activities and educate community providers.  Matching also promotes collaboration between staff in disease-specific programs who do not routinely work closely.