Abstract
Objectives:
Heroin-related deaths have increased substantially in the past 10 years in the United States, particularly in Florida. Our objectives were to measure heroin-related morbidity and mortality rates in Orange County, Florida, and to assess trends in those rates during 2010-2014.
Methods:
We used 3 heroin surveillance methods, based on data from the Florida Medical Examiner, the Florida Agency for Health Care Administration (AHCA), and the Electronic Surveillance System for the Early Notification of Community-Based Epidemics–Florida (ESSENCE-FL). We conducted descriptive and geographic spatial analyses of all 3 data sets, determined heroin-related mortality and morbidity (emergency department [ED] visit) rates, and compared the timeliness of data availability from the 3 data sources.
Results:
Heroin-related deaths in Orange County increased by 590%, from 10 in 2010 to 69 in 2014. Heroin-related ED visits during the same period increased 12-fold (from 13 to 154) and 6-fold (from 49 to 307) when based on AHCA and ESSENCE-FL data, respectively. ESSENCE-FL identified 140% more heroin-related visits than did AHCA. Spatial analysis found geographic clustering of heroin-related morbidity and mortality. Hospitals facing the greatest burden of heroin-related ED visits were close to communities with the highest crude heroin-related ED visit rates. Of the 3 data sources, ESSENCE-FL provided the timeliest data availability.
Conclusions:
These 3 data sources can be considered acceptable surveillance systems for monitoring heroin-related events in Orange County. The timely availability of data from ESSENCE-FL makes it the most useful source for obtaining near–real-time data about the heroin epidemic, potentially leading to improved identification of populations most in need of interventions to reduce morbidity and mortality.
Keywords: heroin, surveillance, morbidity, mortality
Communities across the United States have seen increases in deaths related to opioid overdoses, to the extent that opioid overdoses are now a rapidly growing local and national public health problem.1,2 From 1999 to 2013, the prescription opioid overdose mortality rate in the United States nearly quadrupled, from 1.4 to 5.1 deaths per 100 000 population.3 Opioids include heroin as well as prescription drugs, such as hydrocodone, oxycodone, fentanyl, and codeine.
Florida is certainly not immune to this epidemic. From 2003 to 2009, the number of drug overdose deaths in Florida increased 61%, from 1804 to 2905, with especially large increases noted in the numbers of deaths caused by oxycodone overdose.4 In 2010, Florida had the highest rate of prescription painkiller sales per person—12.6 sales per kilograms per 10 000 population (extrapolated data5)—and it was home to 98 of the 100 highest–oxycodone-prescribing physicians in the United States.4,5 Statewide efforts from 2010 to 2011 to reduce the sales of painkillers included legislation, law enforcement actions, and the creation of the prescription drug monitoring program. The temporal association between these efforts and the subsequent >20% decrease in prescription painkiller rates (eg, oxycodone prescriptions dropping from 26 049 per 100 000 population in 2010 to 19 790 per 100 000 population in 2012) and in deaths rates from prescription drug overdoses (from 14.5 per 100 000 population in 2010 to 11.1 per 100 000 population in 2012) suggests that these initiatives in Florida were successful.4
In addition to deaths involving prescription opioids, the number of heroin-related drug-poisoning deaths has also increased in recent years.2 A recent study by Rudd et al reported that from 2010 to 2012, heroin overdose deaths in 28 states doubled from 1779 to 3635.6 In 2013, >8200 deaths in the United States resulted from heroin-related overdoses; the annual age-adjusted rate was 2.7 heroin-related deaths per 100 000 population, almost quadruple the rate observed in 2000 of 0.7 deaths per 100 000 population.2,7 The potential reasons for this rise in heroin use have included the emergence of chemical tolerance toward prescription opioids, the increasing difficulty in obtaining prescription opioids illegally,8 and the fact that heroin may now be cheaper and easier to obtain than prescription opioids.9–11
Florida has had similar trends. Heroin was responsible for more deaths in the state in 2014 than in the prior 15 years. The drug was involved in 447 deaths in Florida in 2014, a 125% increase from the 199 heroin-related deaths in 2013. In 2014, Orange County in particular had one of the highest heroin-related mortality rates in Florida, second only to Manatee County.12 We reasoned that an investigation into the heroin-related morbidity and mortality in Orange County might provide a better understanding of heroin use and its consequences, potentially aiding in identifying high-risk populations that might benefit from interventions to prevent heroin-related morbidity and mortality.
We conducted heroin-related morbidity and mortality surveillance in Orange County, Florida, using 3 data sources to estimate heroin-related events involving Orange County residents or occurring in Orange County hospitals. To assess heroin-related morbidity, we tracked trends in heroin-related emergency department (ED) visits in Orange County. Although rates of heroin-related mortality have been documented by others,6,12 we also tracked trends in heroin-related deaths in Orange County. Our objectives were to (1) measure local heroin-related morbidity and mortality rates and (2) assess trends in those rates from 2010 to 2014.
Methods
We identified heroin-related deaths in Orange County and heroin-related ED visits made by Orange County residents and made to Orange County hospitals between January 1, 2010, and December 31, 2014. Because this study only included de-identified data and no human subjects, institutional review board review was not required.
Mortality Surveillance
We obtained an unpublished Florida Medical Examiner data set, which identified heroin-related deaths as those in which heroin was listed as a cause of death or heroin was identified in the bloodstream of the deceased at the time of death. We conducted a descriptive analysis of the demographic characteristics (eg, sex, race/ethnicity, and age) of those who died to better understand the affected population. We obtained denominator data from the US Census Bureau, 2010 Demographic Profile,13 and calculated crude heroin-related mortality rates per 10 000 population by ZIP code.
Morbidity Surveillance
To analyze heroin-related morbidity in Orange County, we measured heroin-related ED visits using 2 unpublished surveillance system databases: the Florida Agency for Health Care Administration (AHCA) and the Electronic Surveillance System for the Early Notification of Community-Based Epidemics–Florida (ESSENCE-FL). Neither database contained unique patient identifiers (eg, Social Security numbers), so we were unable to de-duplicate the data that we obtained from either system. We conducted a descriptive analysis of the demographic characteristics of the AHCA population, including sex, race/ethnicity, and age; however, ESSENCE-FL allowed us to capture only sex and age characteristics. Using the ED visit and US Census Bureau 2010 Demographic Profile data, we calculated crude heroin-related morbidity rates per 10 000 population by ZIP code. Using the ED visit data, we determined the total number of heroin-related ED visits, by patient residence and hospital location, in Orange County from 2010 to 2014.
AHCA
AHCA, an agency that was statutorily created by chapter 20, Florida Statutes, is the principal health policy and planning entity in Florida. AHCA is responsible for the state’s Medicaid program, the licensure of its health care facilities, and the sharing of health care data.14 It maintains records, based on required reporting by licensed hospitals, of inpatient, ambulatory care, and ED visits from across the state. Each row in the AHCA ED database contained details pertaining to one patient’s ED visit and included, among other things, demographic characteristics (including residence location), chief complaint codes, diagnosis codes, and the name and location of the treating hospital.
We used Centers for Disease Control and Prevention presentation and guide resources to validate the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes used for this analysis.15,16 We used the ICD-9-CM code 965.01 (poisoning by heroin) to identify heroin-related visits, searching for it in the chief complaint field, discharge diagnosis field, and up to 9 other diagnosis fields in the ED visit records of the AHCA database. We used the external cause of injury code (E-code) E850.0 (accidental poisoning by heroin) to identify heroin-related visits, searching in the 3 E-code fields in the ED visit records. We then filtered this subset of heroin poisoning visits by the patient’s residence and by the treating hospital’s location. We did not use International Classification of Diseases, Tenth Revision codes, because we did not analyze data after the third quarter of 2015, and these new codes were not implemented until October 1, 2015.
For our analysis, we defined a heroin-related ED visit as either a visit made to any ED in any Florida hospital by an Orange County resident or a visit made by any person to an Orange County hospital, between January 1, 2010, and December 31, 2014, and in which the visit record included a chief complaint, discharge or other diagnosis, or E-code field with a code for heroin poisoning.
ESSENCE-FL
ESSENCE-FL is a biosurveillance system that collects ED chief complaint and discharge diagnosis data from participating hospitals and urgent care centers in Florida, call data from the Florida Poison Information Center Network, mortality data from the Florida Office of Vital Statistics, and reportable disease information from the state’s data repository, Merlin.17 We used the ESSENCE-FL database to identify heroin-related ED visits, both by a patient’s residence and by the treating hospital’s location.
For our analysis, we identified heroin-related ED visits by querying the ESSENCE-FL system for the location of patient residence in which the region under the geography system heading was set to “Orange, FL.” We set the time resolution for this query as monthly, with start and end dates of January 1, 2010, and December 31, 2014, respectively. To search for heroin-related ED visits made by Orange County residents, we queried the chief complaint and discharge diagnosis field in ED visit records. This field allows for free-text entries and uses the caret symbol (⁁) for wildcards and multiple entries. During a previous search for heroin-related ED visits in ESSENCE-FL, we noticed various misspellings for “heroin” in the chief complaint field. Because of the various spellings and our desire to capture all heroin-related visits, we queried the ESSENCE-FL database for not only the term “heroin” but also for its related misspellings: ⁁herion⁁ or ⁁heroine⁁ or ⁁herione⁁ or ⁁heriun⁁ or ⁁heruin⁁.
To determine the number of heroin-related ED visits made to Orange County hospitals, we queried the ESSENCE-FL database with the same location parameters that were used for patients, except we searched by hospital location instead of the patient’s residence. For our search in ESSENCE-FL, we defined a heroin-related ED visit as either a visit made by any Orange County resident to any Florida hospital or a visit made by any person to an Orange County hospital between January 1, 2010, and December 31, 2014, and in which “heroin” (or a related misspelling) was mentioned in the chief complaint and discharge diagnosis free-text field. We were unable to ascertain race and insurance coverage information because these variables were not completely captured in ESSENCE-FL.
We also determined the relative timeliness of data availability from each data source by continuing to capture data on heroin-related ED visits and deaths after the end of our initial analysis, until the third quarter of 2015, at which time we recorded the date of the most recent ED visit or death from each source.
Statistical Methods
We analyzed data from both morbidity databases (AHCA and ESSENCE-FL) using SAS version 9.4.18 The Medical Examiner data set contained the deceased’s address, not the location where the injury occurred. We linked the Medical Examiner data set by each heroin-related death to a supplemental data set obtained from the Florida Department of Health, Bureau of Emergency Medical Oversight, Health Information and Policy Analysis Section, which contained the location where the heroin-related injury was reported. We used ArcGIS 10.3.1 and data from each data source about the locations where heroin-related overdose deaths or ED visits occurred to conduct geographic spatial analyses.19 We created a hotspot map that represented the relative density of where heroin-related injuries (which resulted in death) occurred. In addition, we created 2 choropleth maps, in which particular areas within Orange County were shaded in proportion to ED visits per 10 000 population.
Results
Mortality Surveillance
A total of 145 heroin-related deaths occurred in Orange County from 2010 to 2014, increasing 590% (from 10 in 2010 to 69 in 2014). White men aged 30-34 accounted for 36 (25%) of the deaths and were the predominant group (by race, sex, and age) that died from heroin-related causes (Table 1). The crude mortality rate for 2010-2014 for the entire county was 1.0 heroin-related death per 10 000 population. Among all 72 ZIP codes in Orange County, the top 10 highest heroin-related crude mortality rates ranged from 1.3 per 10 000 population in the 32807 ZIP code to 3.3 per 10 000 population in the 32833 ZIP code. Spatial analysis showed the locations within Orange County that were most heavily affected by heroin-related deaths (Figure 1).
Table 1.
Heroin-related morbidity (ED visits) and mortality (deaths), by data source and demographic characteristics, Orange County, Florida, 2010-2014
| Data Source, No. (%) | |||
|---|---|---|---|
| Demographic Characteristics | Florida Medical Examiner (Deaths) | AHCA (ED Visits) | ESSENCE- FL (ED Visits) |
| Total heroin-related deaths or ED visits | 145 (100) | 268 (100 | 643 (100) |
| Sex | |||
| Male | 128 (88) | 175 (65) | 447 (70) |
| Female | 17 (12) | 93 (35) | 196 (30) |
| Age, y | |||
| <18 | 0 (0) | 2 (1) | 3 (<1) |
| 18-24 | 17 (12) | 75 (28) | 130 (20) |
| 25-29 | 20 (14) | 59 (22) | 133 (21) |
| 30-34 | 36 (25) | 50 (19) | 160 (25) |
| 35-39 | 25 (17) | 20 (7) | 77 (12) |
| 40-44 | 15 (10) | 22 (8) | 46 (7) |
| 45-49 | 5 (3) | 14 (5) | 33 (5) |
| ≥50 | 27 (19) | 26 (10) | 61 (9) |
| Racea | |||
| White | 117 (81) | 212 (79) | NA |
| Black | 7 (5) | 13 (5) | NA |
| Asian | 2 (1) | 1 (<1) | NA |
| Other | NA | 36 (13) | NA |
| Unknown | NA | 6 (2) | NA |
| Ethnicityb | |||
| Hispanic | 2 (1) | 72 (27) | NA |
| Non-Hispanic | 17 (12) | 178 (66) | NA |
| Unknown | NA | 18 (7) | NA |
| Health insurance coverage | |||
| None | NA | 166 (62) | NA |
| Medicaid | NA | 44 (16) | NA |
| Medicare | NA | 2 (1) | NA |
| Private/other | NA | 56 (21) | NA |
Abbreviations: AHCA, Agency for Health Care Administration; ED, emergency department; ESSENCE-FL, Electronic Surveillance System for the Early Notification of Community-Based Epidemics–Florida; NA, not available (from data source).
aRace not fully captured for all deaths in Florida Medical Examiner data.
bEthnicity included in race category in Florida Medical Examiner data and not fully captured for all deaths; ethnicity was a separate demographic variable from race in AHCA data.
Figure 1.
Geographic spatial analysis of heroin-related mortality, Orange County, Florida, 2010-2014. Areas in Orange County are shaded in proportion to heroin-related deaths per 10 000 population; darker shading indicates higher mortality rates. Data source: Florida Department of Health, Bureau of Emergency Medical Oversight, Health Information and Policy Analyses Section. The hotspots on the map are based on supplemented medical oversight data (location of where injury was reported), although the final ruling on the cause of death of these injuries was based on Florida Medical Examiner data.
Morbidity Surveillance
AHCA
According to the AHCA database, 268 heroin-related ED visits were made by Orange County residents from 2010 to 2014 (Table 1). The number of heroin-related ED visits increased nearly 12-fold from 2010 (n = 13) to 2014 (n = 154). People who were non-Hispanic (n = 178, 66%), white (n = 212, 79%), male (n = 175, 65%), and aged 18-24 (n = 75, 28%) visited the ED for a heroin-related event more than any other demographic group. More than half (n = 166, 62%) of the heroin-related ED visits were made by uninsured Orange County residents (Table 2). The top 10 crude heroin-related ED visit rates per 10 000 population from 2010 to 2014 among all Orange County ZIP codes ranged from 4.3 in the 32833 ZIP code to 6.3 in the 32827 ZIP code; the overall crude heroin-related ED visit rate for the county was 2.3 per 10 000 population. Spatial analysis showed that hospitals with the greatest burden of heroin-related ED visits were close to communities with the highest crude heroin-related ED visit rates (Figure 2).
Table 2.
Heroin-related emergency department visits by Orange County residents, by insurance payer and year, based on Agency for Health Care Administration data set, Orange County, Florida, 2010-2014
| Visits, No. (%) | ||||||
|---|---|---|---|---|---|---|
| Insurance Payer | Total | 2010 | 2011 | 2012 | 2013 | 2014 |
| Total | 268 (100) | 13 | 12 | 33 | 56 | 154 |
| Self-pay (no insurance) | 166 (62) | 9 | 7 | 20 | 39 | 91 |
| Commercial health insurance | 44 (16) | 1 | 2 | 6 | 7 | 28 |
| Medicaid | 15 (6) | 1 | 0 | 4 | 2 | 8 |
| Medicaid Managed Care (patients covered by Medicaid HMOs, Medicaid provider-sponsored networks, or other Medicaid-funded plans licensed in the state of Florida) | 15 (6) | 1 | 2 | 1 | 2 | 9 |
| Other state/local government insurance | 9 (3) | 0 | 0 | 1 | 1 | 7 |
| Medicare Managed Care (patients covered by Medicare Advantage plans, Medicare HMO, Medicare PPO, Medicare private fee for service, or any other type of Medicare plan in which CMS is not a direct payer) | 8 (3) | 0 | 0 | 0 | 1 | 7 |
| Nonpayment (includes charity, professional courtesy, no charge, research/clinical trial, refuse to pay/bad debt, Hill-Burton free care, research/donor) | 4 (1) | 1 | 0 | 0 | 2 | 1 |
| Medicare | 2 (1) | 0 | 0 | 0 | 0 | 2 |
| Veterans Health Administration | 2 (1) | 0 | 1 | 0 | 1 | 0 |
| TriCare or other federal government insurance | 1 (<1) | 0 | 0 | 0 | 1 | 0 |
| Kidcare (includes Healthy Kids, Medikids, and Children’s Medical Services) | 1 (<1) | 0 | 0 | 1 | 0 | 0 |
| Commercial liability coverage (patients covered under a liability policy; eg, automobile, homeowners, or general business) | 1 (<1) | 0 | 0 | 0 | 0 | 1 |
Abbreviations: CMS, Centers for Medicare & Medicaid Services; HMO, health maintenance organization; PPO, preferred provider organization.
Figure 2.
Geographic spatial analysis of crude heroin-related ED visit rates per 10 000 population, by ZIP code and total number of heroin-related ED visits by hospital, based on AHCA and ESSENCE-FL data sets, Orange County, Florida, 2010-2014. Data sources: AHCA, US Census Bureau, 2010 demographic profile data, and Florida Department of Health in Orange County Epidemiology Program. Abbreviations: AHCA, Agency for Health Care Administration; ED, emergency department; ESSENCE-FL, Electronic Surveillance System for the Early Notification of Community-Based Epidemics–Florida.
ESSENCE-FL
Using the ESSENCE-FL database, we found that all heroin-related visits from 2010 to 2014 were made to hospital EDs and none were made to urgent care centers. We identified 643 heroin-related ED visits made by Orange County residents from 2010 to 2014 (Table 1). The number of heroin-related ED visits increased 6-fold, from 49 in 2010 to 307 in 2014. Males made the most heroin-related ED visits (n = 447, 70%). Adults aged 30-34 made 160 (25%) heroin-related ED visits.
The top 10 crude heroin-related ED visit rates per 10 000 population from 2010 to 2014 for all Orange County ZIP codes ranged from 8.5 to 14.5; the overall crude heroin-related ED visit rate for the entire county was 5.6 per 10 000 population. The ZIP code 32817 had the highest crude heroin-related ED visit rate in Orange County (14.5 per 10 000 population). Spatial analysis showed that hospitals with the greatest burden of heroin-related ED visits were close to communities with the highest crude heroin-related ED visit rates (Figure 2).
In comparing data from the 3 sources, we found that heroin-related morbidity and mortality trended similarly over time and that each source provided relatively timely data (Figure 3). However, at the time of our later analysis, during the third quarter of 2015, heroin-related morbidity data were available from ESSENCE-FL through the third quarter of 2015 and from AHCA through the first quarter of 2015; heroin-related mortality data were available from the Florida Medical Examiner only through the fourth quarter of 2014.
Figure 3.
Trends in heroin-related morbidity (ED visits based on ESSENCE-FL and AHCA data sets) and mortality (deaths, based on Florida Medical Examiner data set), Orange County, Florida, 2010-2015. Analysis of timeliness was conducted in the third quarter of 2015. At the time of this analysis, ESSENCE-FL data were available through the third quarter of 2015, from AHCA through the first quarter of 2015, and from the Florida Medical Examiner through the fourth quarter of 2014. Abbreviations: AHCA, Agency for Health Care Administration; ED, emergency department; ESSENCE-FL, Electronic Surveillance System for the Early Notification of Community-Based Epidemics–Florida.
Discussion
We demonstrated a novel use of existing data sources in Florida to monitor an emerging health issue, heroin poisoning. All 3 data systems showed an increase in heroin-related events from 2010 to 2014; most events occurred in 2014.
We found 140% more heroin-related ED visits in the ESSENCE-FL database than in the AHCA database. This difference was likely due to how each system captured heroin-related ED visits. In AHCA, ED visits specific to heroin poisoning were identified by ICD-9-CM code entries in the chief complaint and discharge (and other) diagnosis fields; in ESSENCE-FL, the visits were identified by free-text entries in a single chief complaint and discharge diagnosis field. Some health care providers and medical coders may classify heroin more generally as an opioid. Thus, some patients seeking care related to heroin use at an ED could have had chief complaints or discharge diagnoses that were coded as “opioid poisoning” rather than “heroin poisoning.” Because the AHCA database contained only coded data, patients whose records were coded for “opioid poisoning” would not have been captured in our analysis with the AHCA database and therefore would have resulted in an underestimation of heroin-related morbidity. Conversely, the ESSENCE-FL database primarily contained free-text data. Therefore, if patients came to an ED with any chief complaint related to heroin or if heroin was mentioned in any way in the patient’s discharge diagnosis (eg, heroin poisoning, patient requesting detox from heroin, heroin withdrawal), these patients would have been captured in our analysis with the ESSENCE-FL database. Additionally, records with free-text phrases such as “no longer uses heroin” or “quit using heroin” may have been captured as heroin-related ED visits. Thus, it is likely that the use of free-text entries in the ESSENCE-FL database led to more observed heroin-related ED visits compared with the AHCA database and resulted in some overestimation of ED visits.
We observed that the highest proportion of heroin-related deaths and ED visits occurred among males who were white, non-Hispanic, and aged 18-34. This finding parallels the findings of other reports that examined heroin morbidity and mortality trends in the United States.2,6 A survey conducted by Cicero et al found a marked shift in the demographic characteristics of heroin users during the past several decades. The survey found that the heroin problem was not only an inner-city issue primarily affecting racial/ethnic minority populations; rather, it extended to white middle-class men and women living outside of large urban areas.11
The survey by Cicero et al also found that many recent heroin users had transitioned from prescription opioids and were using heroin because it was cheaper and more accessible than prescription opioids.11 Indeed, decreased accessibility to prescription opioids has been noted in Florida. This decreased accessibility may be the result of new legislation, increased law enforcement activity, and introduction of the prescription drug monitoring program in 2010-2011. Although these efforts may have reduced the number of opioid poisonings and deaths in the state, they have coincided with a resurgence in heroin use and deaths, a trend that has also been observed nationally.4,7
Another possible cause of the increase in heroin-related deaths identified in our analysis is the emergence of the abuse of fentanyl and its analog acetyl-fentanyl. Fentanyl, a synthetic and short-acting opioid analgesic, is 50 to 100 times more potent than morphine and is approved for managing acute or chronic pain associated with advanced cancer.20 According to the US Drug Enforcement Administration, beginning in late 2013 and extending throughout 2014, several states, including Florida, reported spikes in overdose deaths due to fentanyl and acetyl-fentanyl.21 Further study is warranted to determine if any of the increases in heroin-related deaths found in our analysis were attributable to fentanyl.
Using the AHCA data, we observed that most heroin-related ED visits were made by uninsured patients. This finding is consistent with another report, which showed that the people most at risk for heroin abuse or dependence were those who were uninsured.22
Using spatial analysis, we found that clusters of heroin-related deaths correlated with the ZIP codes of Orange County residents who most frequently made ED visits for heroin-related events. Using both the ESSENCE-FL and AHCA data sources, we also found that the hospitals with the most heroin-related ED visits were located in close proximity to the communities with the highest heroin-related morbidity rates. In light of these results from geographic information system mapping, targeted interventions to reduce the rates of heroin-related morbidity and mortality in these areas and hospitals may be warranted.
We completed our data timeliness analysis in the third quarter of 2015 and found that heroin-related morbidity data were available in ESSENCE-FL through the third quarter of 2015, whereas the AHCA morbidity data and the Florida Medical Examiner mortality data were both available in a less timely fashion. People involved in heroin surveillance should keep in mind the differences in timeliness of capturing heroin-related events among these 3 data sources. The rapid data availability from ESSENCE-FL may be particularly valuable in understanding a population health problem such as heroin poisoning in near-real time.
We presented the results of our analysis to the Orange County Heroin Task Force, which comprises key community leaders in law enforcement, health care, education, and prevention. We also shared our findings with the mayor of Orange County for use in strategic action planning.23 Ultimately, we hope these results will allow for more informed policy decision making and for the allocation of more resources for heroin-related preventive interventions. Nevertheless, information is lacking about the circumstances leading to heroin poisonings and to heroin-related ED visits in Orange County. We speculate that interviews with heroin poisoning survivors might provide additional insight into the pathways leading to abuse, the methods used to obtain heroin, and the combination of drugs that result in overdoses and other heroin-related events.
Limitations
This analysis had several limitations. First, our use of ICD-9-CM codes alone to search for heroin-related ED visits in the AHCA database may not have captured all heroin-related ED visits; therefore, our AHCA-based results may have underestimated the true burden of heroin-related morbidity in Orange County from 2010 to 2014. Second, medical chart review was not available to verify heroin-related chief complaints, discharge diagnoses, other diagnoses, and external injury diagnoses in the data from the AHCA and ESSENCE-FL systems. Also, hospital discharge data are collected for billing purposes and tend to reflect care for which the payer is billed, but they may not always accurately reflect the primary reasons for a visit. A medical chart review would have been necessary to determine with more certainty that an ED visit was correctly attributed to heroin, which would have allowed us to determine the sensitivity and specificity of the codes and keywords that were used. Third, we were unable to capture information about race/ethnicity and insurance payer from the ESSENCE-FL database; as such, we were unable to compare data with the AHCA data for these variables. We have recommended that ESSENCE-FL data managers include race/ethnicity and payer information in ESSENCE-FL in the future because these data are important for targeted heroin interventions. Finally, this analysis was specific to Orange County; as such, the results may not be generalizable to the rest of Florida or the United States.
Conclusions
Our analysis of 3 data sources—Florida Medical Examiner, AHCA, and ESSENCE-FL—showed an observable increase in and similar trends for heroin-related morbidity and mortality in Orange County from 2010 to 2014. The 3 sources are acceptable surveillance systems that can be used to monitor heroin-related events in Orange County. Although data from ESSENCE-FL may overestimate the true burden of heroin-related morbidity and mortality, the timely availability of data from ESSENCE-FL may make it the most useful data source for obtaining near–real-time understanding of this population health problem, until data from potentially more complete sources (eg, ACHA and Florida Medical Examiner) become available for analysis. Ultimately, a better understanding of heroin-related events may lead to the improved identification of high-risk populations, which are potentially most in need of interventions, to reduce heroin-related morbidity and mortality.
Acknowledgments
We thank Sarah Geiger, PhD, Northern Illinois University, and Steven Hale, MD, Florida Department of Health in Orange County, for their review and suggestions, which improved the quality of this article.
Footnotes
Declaration of Conflicting Interests: The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding: The authors received no financial support for the research, authorship, and/or publication of this article.
References
- 1. Slavova S, Bunn TL, Talbert J. Drug overdose surveillance using hospital discharge data. Public Health Rep. 2014;129(5):437–445. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Hedegaard H, Chen LH, Warner M. Drug-Poisoning Deaths Involving Heroin: United States, 2000-2013. Hyattsville, MD: National Center for Health Statistics; 2015. NCHS data brief 190 http://www.cdc.gov/nchs/data/databriefs/db190.pdf. Accessed April 3, 2016. [PubMed] [Google Scholar]
- 3. Chen LH, Hedegaard H, Warner M. QuickStat: rates of death from drug poisoning and drug poisoning involving opioid analgesics—United States, 1999-2013. MMWR Morb Mortal Wkly Rep. 2015;64(01):32. [Google Scholar]
- 4. Johnson H, Paulozzi L, Porucznik C, Mack K, Herter B. Decline in drug overdose deaths after state policy changes—Florida, 2010-2012. MMWR Morb Mortal Wkly Rep. 2014;63(26):569–574. [PMC free article] [PubMed] [Google Scholar]
- 5. Centers for Disease Control and Prevention. Prescription painkiller overdoses in the US. https://www.cdc.gov/vitalsigns/painkilleroverdoses. Accessed April 3, 2016.
- 6. Rudd RA, Paulozzi LJ, Bauer MJ, et al. Increases in heroin overdose deaths—28 states, 2010 to 2012. MMWR Morb Mortal Wkly Rep. 2014;63(39):849–854. [PMC free article] [PubMed] [Google Scholar]
- 7. Centers for Disease Control and Prevention. Today’s heroin epidemic. http://www.cdc.gov/vitalsigns/heroin. Accessed April 5, 2016.
- 8. Slevin KA, Ashburn MA. Primary care physician opinion survey on FDA opioid risk evaluation and mitigation strategies. J Opioid Manag. 2011;7(2):109–115. [DOI] [PubMed] [Google Scholar]
- 9. National Institute on Drug Abuse. Epidemiologic Trends in Drug Abuse: Highlights and Executive Summary. Bethesda, MD: National Institutes of Health; 2006. Proceedings of the Community Epidemiology Work Group; vol 1 https://archives.drugabuse.gov/pdf/cewg/Vol1_605.pdf. Accessed March 12, 2016. [Google Scholar]
- 10. Cicero TJ, Ellis MS, Surratt HL. Effect of abuse-deterrent formulation of OxyContin. N Engl J Med. 2012;367(2):187–189. [DOI] [PubMed] [Google Scholar]
- 11. Cicero TJ, Ellis MS, Surratt HL, Kurtz SP. The changing face of heroin use in the United States: a retrospective analysis of the past 50 years. JAMA Psychiatry. 2014;71(7):821–826. [DOI] [PubMed] [Google Scholar]
- 12. Florida Department of Law Enforcement. Drugs identified in deceased persons by Florida medical examiners: 2014 annual report. https://www.fdle.state.fl.us/cms/MEC/Publications-and-Forms/Documents/Drugs-in-Deceased-Persons/2014-Annual-Drug-Report-FINAL.aspx. Published 2015. Accessed April 17, 2016.
- 13. US Census Bureau. 2010. population finder. https://www.census.gov/popfinder. Published 2013. Accessed November 2015.
- 14. Fla Stat Title 4, Ch 20.42.
- 15. Centers for Disease Control and Prevention. Prescription drug overdose data & statistics: guide to ICD-9-CM and ICD-10 codes related to poisoning and pain. https://www.cdc.gov/drugoverdose/pdf/pdo_guide_to_icd-9-cm_and_icd-10_codes-a.pdf. Published 2012. Revised August 2013. Accessed November 21, 2016.
- 16. Paulozzi L. The other epidemic: prescription drug abuse in the US: ten prevention ideas from the states. Presented at: the National Conference of State Legislatures “Injury Prevention Through the Life Cycle” Meeting; May 14, 2009; Atlanta, GA http://www.ncsl.org/portals/1/documents/health/Paulozzi509.pdf. Accessed October 4, 2015.
- 17. Florida Department of Health. Surveillance Systems—ESSENCE. http://www.floridahealth.gov/diseases-and-conditions/disease-reporting-and-management/disease-reporting-and-surveillance/surveillance-systems.html. Accessed October 6, 2015.
- 18. SAS Institute Inc. SAS Version 9.4. Cary, NC: SAS Institute Inc; 2014. [Google Scholar]
- 19. ESRI. ArcGIS Desktop: Release 10.3.1. Redlands, CA: ESRI; 2015. [Google Scholar]
- 20. Algren DA, Monteilh CP, Punja M, et al. Fentanyl-associated fatalities among illicit drug users in Wayne County, Michigan (July 2005-May 2006). J Med Toxicol. 2013;9(1):106–115. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. US Drug Enforcement Administration Strategic Intelligence. National heroin threat assessment summary. http://www.dea.gov/divisions/hq/2015/hq052215_National_Heroin_Threat_Assessment_Summary.pdf. Published 2015. Accessed January 7, 2016.
- 22. Centers for Disease Control and Prevention. New research reveals the trends and risk factors behind America’s growing heroin epidemic. https://www.cdc.gov/media/releases/2015/p0707-heroin-epidemic.html. Updated July 7, 2015. Accessed February 3, 2016.
- 23. Orange County Heroin Task Force. Orange County Heroin Task Force recommendations. http://orangecountyfl.net/Portals/0/Library/Board%20Of%20County%20Commissioners/docs/Orange%20County%20Heroin%20Task%20Force%20%20Recommendations.pdf. Accessed November 19, 2016.