Distribution of tobacco and nicotine use indicators from the Periodic Health Assessment and medical diagnostic codes...

Military service members remain a priority population for assessing the prevalence, patterns, and long-term consequences of tobacco and nicotine use. The limitations inherent to documenting use among military service members, however, complicate the design of exposure assessment. This study combined 2 data sources-by aggregating self-reported Periodic Health Assessment (PHA) survey data with International Classification of Diseases, 9th and 10th revisions, Clinical Modification (ICD-9-CM/ICD-10-CM) medical diagnostic codes-to classify nicotine and tobacco use as exposures delineated by recent use or history of any use. The study population included a total of 921,394 U.S. active component service members who completed a PHA in 2023. PHA classification for 'recent use' was defined by self-reported use of any tobacco or nicotine product within the past 30 days, whereas 'history of any use' included recent users in addition to those who reported cessation of use. The full roster of service members who completed the PHA in 2023 was matched to ambulatory and inpatient medical records within 30 days, before or after, the PHA sample period (December 1, 2022-January 31, 2024) to identify selected ICD-10-CM codes for recent use. Selected diagnostic codes for a 'history of any use' were queried for a period of 20 years preceding and 30 days following (January 1, 2004-January 31, 2024) the PHA sample period. Among PHA respondents, 22.0% (n=203,156) self-reported recent nicotine or tobacco use. When aggregating PHA data with recent exposure classified from diagnostic codes, the resulting assessment of recent nicotine or tobacco use increased to 28.7% (n=264,194). Critically, this aggregation identified 61,038 U.S. service members with no evidence of recent use on the PHA but with a concurrent clinical record during the specified matching period. Aggregating data sources for a history of any use only nominally improved the estimate, increasing it from 41.1% (PHA alone) to 43.1%. Agreement between sources was fair for both recent use (κ=0.28) and historical use (κ=0.36). The results of this study indicate that neither self-reported PHA data nor medical diagnostic codes alone provide a complete picture of tobacco and nicotine use among U.S. active component service members. The combination of medical diagnostic codes with self-reported PHA survey responses increases exposure estimates of recent tobacco or nicotine use among U.S. active component service members to 28.7%, in comparison to 22.0% if exclusively assessing recent use from the PHA. The integration of multiple data sources may provide a more comprehensive assessment of recent nicotine and tobacco exposure among service members, directly supporting enhanced public health surveillance.