Syndromic Surveillance as an Early Indicator of the

Syndromic Surveillance as an Early Indicator of the Effectiveness of Overdose Prevention Legislation in Utah, 2012 -2015 Anne Burke, MS, Anna Fondario, MPH, Meghan Balough, MPH, Gregg Reed, MPH, David Jackson, MPH, Allyn K. Nakashima, MD Utah Department of Health, Salt Lake City, UT Results Introduction The objective of this analysis was to evaluate the impact of this legislation on emergency department (ED) visits for overdose in Utah using the National Syndromic Surveillance Program (NSSP). Methods • Line-level syndromic surveillance data for ED visits from June 2012 -June 2015 was exported from National Syndromic Surveillance System (NSSS). • Visits binned to the “poisoning by medicines” syndrome were identified and manually reviewed. • Gender of patient and intention of overdose were characterized and tabulated. • Monthly counts of overdoses were determined 24 months pre-intervention and 12 months post-intervention. • The Loess procedure for smoothed non-parametric fit was utilized to examine trends in overdose. • A segmented regression analysis of interrupted time series data was utilized to determine the impact of legislation on overdose ED visits overall and stratified by gender and intent. • Regression coefficients were compared using Student’s t-tests. “Poisoning by Medicines” Syndrome Definition Bio. Sense Revised Inclusion Criteria Revised Exclusion Criteria Chief Complaint Terms “bebio”, “chlorine”, “cloro”, “consumio”, “dosis”, “envenenameinto”, “intoxicacion”, “o d”, “over dose”, “overdose”, “pastillas para dormer”, “poisioning”, “poisoning”, “sleeping pills”, “sobredosis”, “swallowed”, “tomo”, “took pills”, “toxicity”, “toxico”, “trago” “od”, “SI/OD”, “ingestion”, “took too much”, “took extra”, “took wrong”, “accidentally took”, “accidentaly took”, “took meds”, “o. d”, “ingestion”, “injestion”, “wrong drug”, “ingeston”, “ingestions”, “drugged”, “poison”, “ingession”, “ingest”, “took multiple”, “overmedicated”, “over medicated”, “took too many”, “injested”, “ingested”, “poisen” “foreign body”, “food poisoning”, “alcohol”, “etoh”, “fb”, “foreign object”, “oak”, “carbon monoxide”, “CO 2”, “food poison”, “ivy”, “sumac”, “carbonmonoxide”, “carbonmonxide”, “f/b”, “f/o”, “object”, “staple”, “battery”, “lego”, “quarter”, “coin”, “penny”, “magnet”, “bobby pin”, “earring”, “marble”, “nickel”, “toy”, “swallowed foreign obj”, “marble”, “nail”, “push pin”, “dime”, “ball”, “button”, “tooth” Table 1: Characteristics of patients and visits for overdose in Utah, 2012 -2015 30 20 10 0 6. 15. 2012 300 1 b 250 200 150 100 50 0 6. 15. 2012 Accidental 6. 15. 2013 Intentional 6. 15. 2014 Male 6. 15. 2013 Characteristic Gender Male Female Unknown N (%) 5, 144 (44) 6, 545 (56) 78 (1) Intent of Overdose Intentional Accidental Undetermined 703 (8) 762 (6) 10, 302 (88) Table 2: Trends in Overdose by intention of overdose and gender, Linear Regression Analysis Female 6. 15. 2014 Figure 2: Segmented regression analysis of trends in overdose in Utah pre- and post-legislation, 2012 -2015 600 Variable and Stratification Date by Gender (Male) Date by Gender (Female) Co-efficient (SE) Date by Overdose (Accidental) Date by Overdose (Intentional) 0. 0007 (0. 1788) -0. 0270 (0. 0572) Comparison t -statistic 0. 1705 p-value 0. 0952 0. 9250 0. 8662 -0. 0142 (0. 0486) -0. 0014 (0. 0128) Table 3: Trends in overdose in Utah pre- and postintervention, Segmented Regression Analysis Time Period Pre-Legislation Post-Legislation 500 Variable Intercept Date Estimate (SE) -7954. 77 (1793. 83) 0. 42 (0. 09) Intercept Date 4185. 57 (1535. 17) -0. 19 (0. 07) p-value 0. 0002 0. 0285 Figure 3: Non-parametric estimation of a smooth regression for trends in overdose in Utah, 2012 -2015 600 400 500 • Overall, 11, 767 visits for “poisoning by medicines” were identified out of 2, 629, 950 ED visits during the 3 -year time period (0. 45% of visits). • The results of this analysis suggest that there was a significant reduction in ED visits for overdose following the legislation. • The non-parametric smooth supported the utilization the legislation implementation (beginning March 2014) as the point of change, with the steadily increasing rate of overdose visits decreasing at that time. • The segmented regression analysis further supported this conclusion, with the rate of overdoses increasing prior to March 2014 and decreasing following March 2014. • The comparison of overdose visits for men and women and accidental and intentional overdoses suggested similar decreases for these four groups. • The results of this analysis suggest that the 2014 legislation has positively impacted overdoses in Utah. Limitations • It is possible that other factors impact the rate of overdoses in the state as observed in this ecological analysis. • If individuals were less likely to seek emergency care after administering naloxone, this may have contributed to the rate decrease. • Beginning in 2007, Utah women, but not men, experienced a decrease in the rate of overdose deaths which persisted despite discontinued funding for the Pain Medication Management and Education Program in 2010. It is possible that this difference impacted the decreases observed in this analysis, however a comparison of post-legislation in men and women suggests a similar impact. • Utah’s “Use as Directed” initiative which encouraged safe use and disposal of prescription medications began around the same time and may have impacted rates. • While this analysis stratified by intent, chief complaints were often nonspecific. Although some overdose chief complaints include descriptions of intention, the majority (88%) do not and chief complaints of “OD” and “overdose” alone are common. • While syndromic surveillance provides near real-time reporting of ED visits, it is limited by data quality and completeness and it is currently necessary to evaluate syndromes to determine their sensitivity and specificity to capture trends in data. Acknowledgements 300 Visit Count Purpose Figure 1 a and 1 b: Linear regressions of overdose trends by intention of overdose (1 a) and gender (1 b) in Utah, 2012 -2015 50 1 a 40 Visit Count Drug overdoses are an epidemic in Utah, with an average of 49 Utahns dying as a result of drug poisoning every month. Having experienced a 315% increase in deaths related to misuse and abuse of prescription drugs since 2000, Utah ranked fifth in the country for prescription drug overdose deaths in 2013. In Spring 2014, legislation was signed into effect with the potential to impact drug overdose rates in the state. The Good Samaritan (effective March 2014) and Naxolone administration (effective May 2014) legislation provided limited criminal immunity to people who notify the authorities in the event of an overdose and expanded access to naloxone for opioid overdose, respectively. While similar legislation has been evaluated using mortality data, syndromic surveillance provides a timely mechanism by which to evaluate its impact on overdose cases. Discussion 200 400 Funding for this conference was made possible (in part) by the Centers for Disease Control and Prevention. The views expressed in written conference materials or publication and by speakers or moderators do not necessarily reflect the official policies of the Department of Health and Human Services, nor does the mention of trade names, commercial practices, or organizations imply endorsement by the US Government. 300 Contact 200 100 100 1/2012 7/2012 1/2013 7/2013 1/2014 7/2014 1/2015 7/2015 Date Anne Burke, MS Utah Department of Health Bureau of Epidemiology aburke@utah. gov Anna Fondario, MPH Utah Department of Health Bureau of Epidemiology afondario@utah. gov