OPENACCESS Citation: Wagner KD, Harding RW, Kelley R, Labus B, Verdugo SR, Copulsky E, et al. (2019) Post- overdose interventions triggered by calling 911: Background Centering the perspectives of people who use Opioid overdose deaths have increased exponentially in the United States. Bystander drugs (PWUDs). PLoS ONE 14(10): e0223823. response to opioid overdose ideally involves administering naloxone, providing rescue https://doi.org/10.1371/journal.pone.0223823 breathing, and calling 911 to summon emergency medical assistance. Recently in the US, Editor: Thomas G. Brown, Douglas Mental Health public health and public safety agencies have begun seeking to use 911 calls as a method University Institute, CANADA to identify and deliver post-overdose interventions to opioid overdose patients. Little is Received: June 19, 2019 known about the opinions of PWUDs about the barriers, benefits, or potential harms of post- Accepted: September 30, 2019 overdose interventions linked to the 911 system. We sought to understand the perspectives Published: October 17, 2019 of PWUDs about a method for using 911 data to identify opioid overdose cases and trigger a Copyright:© 2019 Wagner et al. This is an open post-overdose intervention. access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and Methods and findings reproduction in any medium, provided the original We conducted three focus groups with 11 PWUDs in 2018. Results are organized into 4 cat- author and source are credited. egories: willingness to call 911 (benefits and risks of calling), thoughts about a technique to Data Availability Statement: Because of the identify opioid overdoses in 911 data (benefits and concerns), thoughts about the proposed sensitive nature of the information contained in the transcripts (e.g., details about illegal behavior) and post-overdose intervention (benefits and concerns), and recommendations for developing potential for severe ethical, legal, and social an ideal post-overdose intervention. For most participants, calling 911 was synonymous consequences resulting from broken with “calling the police” and law enforcement-related fears were pervasive, limiting willing- confidentiality, full transcripts cannot not be made ness to engage with the 911 system. The technique to identify opioid overdoses and the pro- publicly available, per restrictions imposed by the Reno Research Integrity Office, and IRB. Redacted posed post-overdose intervention were identified as potentially lifesaving, but the benefits excerpts of the qualitative transcripts used in the were balanced by concerns related to law enforcement involvement, intervention timing, current analysis will be made available to qualified and risks to privacy/reputation. Nearly universally, participants wished for a way to summon researchers subject to review and approval by the appropriate Institutional Review Board(s). emergency medical assistance without triggering a law enforcement response. PLOS ONE | https://doi.org/10.1371/journal.pone.0223823 October 17, 2019 1 / 14 PWUDs’ perspectives on post-overdose interventions Requests can be made to the University of Nevada, Conclusions Reno Research Integrity Office by calling +1-775- The fact that the 911 system in the US inextricably links emergency medical assistance with 327-2368 or via email to Reno RIO director, Nancy Mood: firstname.lastname@example.org. law enforcement response inherently problematizes calling 911 for PWUDs, and has implica- tions for surveillance and intervention. It is imperative to center the perspectives of PWUDs Funding: Research reported in this publication was supported by the US National Institutes of Health when designing and implementing interventions that rely on the 911 system for activation. under awards P20GM103440 (KW, BL) and R01DA040648 (PJD, KW). The funder provided support in the form of salaries for authors (KW, BL, PJD) but did not have any additional role in the Introduction study design, data collection and analysis, decision to publish, or preparation of the manuscript. The Opioid overdose deaths have risen exponentially in the United States (US). In 2017, over commercial company FirstWatch, Inc. provided 70,000 people died from drug overdose in the US . This number translates to an age- support in the form of salary for author SRV adjusted rate of 21.7/100,000, representing a 9.6% increase compared to 2016 . Of those beginning in September 2018 and had no role in drug-related deaths, deaths involving synthetic opioids (other than methadone), increased by study design, data collection and analysis, decision to publish, or preparation of the manuscript. The 45% from 2016 to 2017 . Polypharmacy is also a major contributor to overdose deaths, espe- specific roles of these authors are articulated in the cially consumption of opioids in combination with benzodiazepines and, increasingly, stimu- ‘author contributions’ section. The content is solely lant drugs such as methamphetamine and cocaine . the responsibility of the authors and does not While the number and rates of opioid overdose have increased dramatically in recent years, necessarily represent the official views of the drug overdose has been identified as a leading cause of preventable death among people who National Institutes of Health or FirstWatch, Inc. use drugs (PWUDs) for at least two decades , and interventions to train PWUDs in opioid Competing interests: FirstWatch is a company that overdose prevention/response using take-home naloxone have existed in the US since the late creates technological solutions to manage and use 1990’s . While the initial scale up was slow across the US, by 2014 there were 644 sites pro- real-time 9-1-1 data to inform emergency medical response. Prior to working at FirstWatch, SRV was viding take-home naloxone to PWUDs . These programs typically train bystanders to a Project Coordinator on an NIH/NIDA-funded respond to opioid overdose by administering naloxone, providing rescue breathing and/or study related to this report. SRV conducted data CPR, and summoning emergency medical assistance by calling 911. Summoning emergency collection and analysis related the larger study medical services can be an important step for bystanders. Because naloxone is an opioid antag- while she was employed as Project Coordinator for onist with specific affinity for the opioid receptors, a polypharmacy overdose caused by multi- the NIH/NIDA-funded study. Her contributions towards the preparation of the current manuscript ple substances may not be reliably reversed with naloxone administration alone. While rare, occurred while she was employed by FirstWatch, other sequalae (e.g., pulmonary edema subsequent to opioid use or acetaminophen toxicity) Inc. FirstWatch was compensated as a vendor to are possible and many opioid overdose patients could benefit from additional supportive care query 9-1-1 data for the larger NIH-funded studies for optimal recovery . related to this report, but has not contributed in any In the US, 911 is a universal telephone number operating under local governance that way to the development of the current manuscript routes calls to a system of call centers, ultimately resulting in the dispatch of emergency per- other than with salary support paid to Dr. Verdugo as discussed in the ‘Funding’ section. KW and PJD sonnel. Typically, calls are received by a public safety answering point (PSAP), which identifies have testified as unpaid invited experts (PJD) and the nature of the emergency and either dispatches responders immediately or transfers the call public comment (KW) on the topic of opioid to a more specialized secondary PSAP . Medical emergency calls may be handled by the pri- overdose and naloxone availability for the US Food mary PSAPs, which are usually operated by public safety agencies, or may be routed to call and Drug Administration and US National Institutes centers with dedicated medical dispatch. While some variability exists in terms of how 911 of Health and other state and local governmental bodies. KW holds a separate grant from Arnold calls are handled, most dispatch centers use a formalized protocol for asking questions to Ventures to examine the feasibility, acceptability, determine the nature of the emergency and appropriate level of response, dispatch emergency and outcomes of an emergency department-based responders, and record data about the call. post-overdose intervention. These declarations do For many people who do not use drugs, calling 911 for a medical emergency is a relatively not alter our adherence to the PLOS ONE policies uncomplicated act that represents activating the emergency response system and summoning on sharing data and materials. emergency medical responders to provide lifesaving medical assistance. However, for non- majority populations in the US, including people of color, undocumented immigrants, people living in low income neighborhoods, youth, and PWUDs, calling 911 may be fraught with fear that it might act to summon law enforcement or have other risky consequences (e.g., [9, 10]). In fact, global studies among PWUDs have reported rates of calling for medical assistance between 21% and 63% [11–16]. PLOS ONE | https://doi.org/10.1371/journal.pone.0223823 October 17, 2019 2 / 14 PWUDs’ perspectives on post-overdose interventions Multiple factors influence the likelihood of calling 911 for an overdose emergency, includ- ing history of experiencing or witnessing overdose , social norms or social influence , source of information about opioid overdose prevention , and location of drug use/over- dose [13, 17]. In addition to these individual, social, and structural predictors, one of the most critical influences on whether PWUD summon emergency medical assistance is the legal/pol- icy environment . In the US, where drug use is severely criminalized, research has consis- tently identified fear of law enforcement as a significant deterrent to calling 911 [12, 14, 19]. PWUDs fear arrest for drug-related charges, but also for charges related to homicide (if the overdose victim dies and those at the scene were involved in some way with providing the drugs used by the decedent), violation of parole/probation, outstanding warrants, and tres- passing . These fears are not unfounded, since PWUDs do report that law enforcement officers attend medical emergency calls and individuals at the scene can be arrested as a conse- quence. For example, participants in an opioid overdose prevention program in Los Angeles, California, USA reported that law enforcement officers responded to 67% of events, and some- one was arrested in 14% of events . A community-level study in New York City demon- strated a positive association between the rate of misdemeanor arrests (an indicator of policing) and accidental drug overdose mortality . 911 Good Samaritan laws are designed to reduce these concerns by providing protections against some offenses when someone calls 911 in good faith to summon emergency assistance, but research since their passage suggests that the laws often do not provide sufficient protection to mitigate fears . The epidemic scale of opioid overdose deaths in the US has necessitated the rapid develop- ment and scaleup of innovative public health responses to reduce death rates. In many com- munities, public health, public safety, and social service agencies are exploring the possibility of using the 911 emergency response system as a mechanism for monitoring trends in opioid overdoses and initiating intervention efforts. For example, “post-overdose outreach” interven- tions represent novel collaborations between public health and public safety agencies to pro- vide outreach and engagement services to people who use opioids and/or their social networks once they come to the attention of the system via an emergency response . Program mod- els include post-overdose outreach to the overdose victim’s residence (either by police, clini- cians, or a multidisciplinary team), referrals to services for the overdose victim and their social network, or encouragement that the overdose victims visit a fixed community-based site for services. Other law enforcement-based models include referrals initiated by officers at the scene , or via a program that encourages people who use opioids to seek help at the police station . A critical assumption underlying these interventions is that opioid overdose patients will come to the attention of the public health and public safety agencies via a request for emergency medical services (i.e., by calling 911). However, given the substantial literature that describes highly salient and severe risks perceived by PWUDs when considering a call to 911, this assumption requires further interrogation. The current study was initially undertaken to provide pilot data to inform a post-overdose outreach model similar to those mentioned above. We aimed to investigate the feasibility of using a machine-learning algorithm to identify opioid overdoses in data from the 911 emer- gency medical dispatch system, which could then be used to trigger a post-overdose outreach intervention. However, given the well-established findings regarding the deterrents to calling 911, we first sought to investigate the ethics and acceptability of such an initiative with PWUDs. In the current study we report on findings from focus groups with PWUDs designed to elicit their perspectives on post-overdose interventions triggered by 911 calls. We present these findings to advance a patient-centered perspective on the development of interventions that affect the health and wellbeing of PWUDs . PLOS ONE | https://doi.org/10.1371/journal.pone.0223823 October 17, 2019 3 / 14 PWUDs’ perspectives on post-overdose interventions Materials and methods Setting & legal environment The study took place in Reno, Nevada. In 2015, the Nevada Legislature passed State Bill 459, Nevada’s Good Samaritan Drug Overdose Act, which amended multiple sections of Title 40 of the Nevada Revised Statue (Good Samaritan Drug Overdose Act, 2015). This law expanded access to naloxone through multiple mechanisms and enacted liability protections for individ- uals involved in naloxone distribution/prescribing. The law also provided 911 Good Samaritan protections that protect people who call 911 in the event of an overdose or who are the subject of such a call from arrest, charge, prosecution, conviction, or asset forfeiture for use or posses- sion of small amounts of drugs (unless it is for the purpose of sale), possession of drug para- phernalia, or violation of restraining orders or probation or parole. The law explicitly does not prohibit the government from taking action related to the abuse or neglect of a child. Recruitment & data collection The Institutional Review Board (IRB) at the University of Nevada, Reno approved all study activities under protocol #1024876. Respondents provided verbal informed consent using a Consent Information Script, which described the study and the risks/benefits of participation. Because a signature would have been the only identifying information provided by partici- pants, the study was granted a waiver of documentation of consent by the University of Nevada, Reno IRB. We conducted three focus groups over the course of one week in 2018. Participants were recruited through the distribution of flyers at community-based organizations that provide services for PWUDs, including substance use disorder treatment centers and syringe services programs, and via word of mouth, resulting in three groups of participants who came from dif- ferent locations in town and different networks of PWUDs. Criteria for eligibility included being over 18 years old and a current opioid user. Recruitment flyers described the study as seeking to hear the opinions of people who use opioids about a new intervention to reduce opi- oid overdoses. Focus groups were conducted at a university-leased research field site, located in a nondescript building accessible by foot, bike, bus, or car. Participants received $20 for their time and food was provided at the focus group. After the focus groups were complete, participants were offered the opportunity to participate in a brief overdose education session and were provided naloxone by a community-based organization that partnered with the study. Focus groups were facilitated by the first author, a behavioral scientist with 20 years of expe- rience in qualitative data collection, and were attended by 3 additional authors who served as observers, note takers, and co-facilitators as needed. The loosely-structured interview guide began with a scripted description of the proposed intervention: “The idea for the project would be to identify opioid overdoses based on the data that come in on a call to 911, then send a counselor or a peer educator to the scene to provide naloxone, training in overdose risk reduction techniques, and/or connection to services like methadone, buprenorphine, or syringe services.” The interventionist was subsequently described as a peer recovery support specialist (i.e., someone with lived experience of substance use who is trained in recovery sup- port). Then, participants were asked what they think of the idea and what benefits/harms might result from such a program. Throughout the focus groups, the technique and interven- tion were described in increasingly more detail, with additional examples of how they could be implemented and the potential benefits and harms. For example, we told participants that the machine-learning technique would use data from the 911 call to identify overdose cases, even PLOS ONE | https://doi.org/10.1371/journal.pone.0223823 October 17, 2019 4 / 14 PWUDs’ perspectives on post-overdose interventions if the caller did not explicitly state that the call was for an overdose (e.g., by using other contex- tual information from the call). We also asked participants about how knowledge of such a program would affect their willingness to call 911. The focus groups were digitally recorded and facilitators took notes to record impressions. Audio recordings were transcribed verbatim, and reviewed to ensure any identifiable information was redacted prior to analysis. Analysis The first author conducted the qualitative analysis using a thematic approach that relied on a priori categories and emergent themes [27, 28]. First, transcripts were loaded into the Atlas.ti software program for management. Transcripts were read in their entirety and memos were used to document initial impressions. Codes were developed iteratively, based on the initial reading of the transcripts, and were grouped into higher and lower order concepts, based on our a priori questions about participants’ impressions of the proposed technique and interven- tion. Codes were applied systematically to the transcripts for each focus group. Then, the coded segments of the transcripts were output and read a final time, during which additional axial codes were applied. Findings were shared with authors 2, 5, 6, 7, 8, and 9 and interpreta- tions were solidified through reflection, discussion, and revision until consensus was achieved. Results Three focus groups were attended by 11 participants (Focus group [FG] 1: 2 women; FG2: 1 woman, 5 men; FG3: 1 woman, 2 men) over the course of one week in November, 2018. Focus groups lasted between 40 and 55 minutes in duration. We present findings related to four a priori categories that formed the basis for our inquiry: (1) willingness to call 911 and the fears and meanings associated with that act, (2) thoughts about the development of a machine- learning technique to identify opioid overdoses in 911 data, (3) thoughts about the subsequent intervention that could be deployed, and (4) recommendations for an ideal post-overdose intervention. Within each category, we present emergent themes and supporting narratives from the focus group transcripts. The categories and themes are summarized in Table 1. Willingness to call 911 Nearly universally, respondents described hesitance to call 911 in the event of an overdose. For example, one respondent said, I feel threatened every time if I have to call for assistance, because of cops. . ..Now, if a friend is overdosing, “Oh shit. I’ve got to call 911. Cops are going to come too. Oh fuck. They’re going to start harassing me again.” The exchange below between the interviewer (I) and two female respondents (R1 and R2) describes the multifaceted nature of people’s concerns, and highlights the additional fears faced by people with children: I: So, the question is, ‘How worried are you right now, about. . .right now how worried would you be about calling 9-1-1 for somebody that was overdosing’? R2: As worried as a person could be. I would do everything in my power not to call 9-1-1. R1: And I’m opposite. I’ve always called because, like I said, if it comes down to it and I have to go to prison, I trust in God that . . . at least I know that I saved someone’s life. Like, I will take the repercussions, I don’t hesitate. PLOS ONE | https://doi.org/10.1371/journal.pone.0223823 October 17, 2019 5 / 14 PWUDs’ perspectives on post-overdose interventions Table 1. A priori categories and emergent themes related to PWUDs’ perspectives on a post-overdose outreach intervention triggered by calling 911. Category 1: Willingness to call 911 Benefits of calling: Risks of calling: - Save a life - Fear of CPS/losing children - Fear of police - Fear of arrest and/or incarceration - Impact on privacy/reputation Category 2: Thoughts about technique to identify overdoses in 911 data Benefits: Concerns: - Shorten time to appropriate care - Inaccuracy, leading to misdiagnosis or delayed treatment - Save a life - Privacy, violation of choice not to disclose Category 3: Thoughts about a post-overdose intervention Benefits: Concerns: - Peer support specialist could empathize with - Privacy, violation of confidentiality patient - Timing of intervention is suboptimal because of precipitated withdrawal - Disincentivize transport to hospital Category 4: Recommendations for an ideal post-overdose intervention - Active follow-up (not just referrals) - Flexible intervention, allow people to talk about their “real problems” not just substance use disorder treatment - Create alternative number that can summon emergency medical assistance without linking to law enforcement https://doi.org/10.1371/journal.pone.0223823.t001 I: And you both talked about this a little bit, but the question says, ‘What are you worried about exactly’? R2: Um, my children [laughter]. I: Mhm. R1: And [I] think me, that’s why I don’t hesitate so much anymore, ‘cause I’ve already lost my children. I don’t want to lose my dog, but. . .you know, she’s still just a dog. It’s a lot dif- ferent from losing a child so. . . R2: So yeah, that’s my number one concern. Also my, my reputation, I don’t want anyone to know anything about me, so I could. . . there’s a good possibility I could know a first responder or know somebody from, you know. I: Mmmm. The person who shows up? The people who show up are people that you might know? R2: Sure. I: And then they would know things about you? R2: Yeah, or going, or ending up in the hospital or something, that I could know an RN or I could know the social worker or any number of people. The passage above highlights one respondent’s fear that a call to 911 could involve her with Child Protective Services (CPS) and lead to the loss of her children. Later, she goes on to say, “I fear CPS so much, because my children are my world, you know? And. . .that makes me want to at least steer clear of law enforcement regardless of. . .you know. . .it’s a really, really, really, really big deal.” The second respondent in the passage above reveals that she has already lost PLOS ONE | https://doi.org/10.1371/journal.pone.0223823 October 17, 2019 6 / 14 PWUDs’ perspectives on post-overdose interventions children, though the circumstances leading to that loss are not discussed. The ultimate conse- quence for her would be going to prison, though she says she would be willing to face that con- sequence in the interest of saving a life. Later in the conversation, this respondent talked about fear of facing murder charges if she were at the scene of a fatal overdose. The first respondent also highlights that a call to 911 could threaten one’s privacy and impact one’s reputation. This fear was echoed by a male respondent in another focus group, who described an experience in which he had encountered individuals through both the emergency medical and law enforce- ment systems (e.g., a law enforcement officer who was also a volunteer paramedic), and believed that information about his drug use learned through the clinical encounter was subse- quently used against him during a law enforcement investigation. For many, “calling 911” was synonymous with “calling the police”. For example, when asked by the interviewer (I) about what would make him more or less willing to call 911 in the event of an opioid overdose, one participant (R) explained through a hypothetical example that calling 911 would be a last resort only if the administration naloxone did not work, R: Personally. . .this is make-believe, pretend. . .this would only be if the naloxone that I already administered myself wasn’t working, that I would be having to call the police. [emphasis added] I: You do what you can first, and then. . .? R: “Hell yeah. All things. I don’t call the police ever, for any reason, unless I have exhausted every means.” [emphasis added] In this passage we see that even though the interviewer was asking about willingness to call 911, the respondent answered by talking about “calling the police”. Another respondent described the events following his own opioid overdose, including his observation of the arrests of his friends and his perceived risk of arrest himself, Yeah. . .I overdosed, they call[ed] the cops for me. [emphasis added] Yet, when I came to after they did Narcan on me, the cops were arresting my friend because they were high too, and they had dope on them, but they went to jail. I was in the hospital and I was next to go to jail, but I snuck out. And they were waiting. They were. I seen them and I overheard them and everything. The semantic substitution of “police” and “cops” for “911” in these quotes reveals how deeply connected the 911 and law enforcement systems are in the minds of these respondents. To many, there is no difference between calling 911 to summon emergency medical assistance and calling 911 to summon law enforcement. In fact, this observation may be accurate in many communities, where the primary PSAP call center may be operated by a public safety agency . Thoughts about a technique to identify opioid overdoses in 911 data We asked participants to share their opinions about our idea to develop a machine-learning technique that would use existing data in the 911 dispatch system to identify opioid overdoses. If it worked, this technique would have the potential to use data from the call to identify opioid overdose cases even if the caller did not explicitly state that the medical emergency was for an overdose. Respondents perceived both benefits and concerns related to such a technique. The primary benefit was that respondents believed it could help shorten the time to appropriate care for someone dying of an opioid overdose. This assessment was tied to their understanding PLOS ONE | https://doi.org/10.1371/journal.pone.0223823 October 17, 2019 7 / 14 PWUDs’ perspectives on post-overdose interventions of the mechanisms of opioid overdose–specifically, that the respiratory depression associated with opioid overdose must be reversed quickly: R: If it can save somebody faster, I mean, that’s good, right? I mean. . .sometimes it’s just minutes without oxygen can cause brain damage. R: All I can really think of is that it can improve. . . is that you’re able to realize more quickly that it is an overdose and treat it appropriately. Concerns related to the technique focused on two domains: inaccuracy and privacy. Some respondents expressed concerns that if the technique were inaccurate, it could contribute to misdiagnosis and/or delayed treatment. For example, one respondent said, R: Well I guess if they’re focused specifically, if they go into it specifically focused on, with it in their head that this is an overdose, then it might delay them from taking the measures that they need if it was something else. I mean they only have seconds to save lives or save a life, you know? I: Oh, like if it’s wrong—Then the paramedics go in with an idea about what it is— R: Exactly and it may delay them a minute or two [cross talk]. So that could be one down- fall, but then the upside is that it could also save a life if, you know, people are being honest enough to say, you know, they go in thinking it’s an overdose and it is, they are able to immediately. . .they go less time without the lack of oxygen, which can cause brain damage. So, it can work both ways, you know. Privacy concerns were also raised. Some participants believed that some callers have rea- sons for not disclosing the nature of the call, and that a technique like this would override that choice. This could be particularly problematic for people on parole/probation or other com- munity supervision, if a call to 911 signaled that there was drug use occurring at their house. However, most participants balanced that concern with the potential benefits, as illustrated in the quote above. Some respondents also identified a sense of inevitability regarding the discov- ery of the true nature of the call, As soon as they get there, they’re [going] to know anyway. It doesn’t matter what you say. . .If I say it’s an overdose [or] you get the idea based on this [technique] that it’s an overdose, the end result is the same. It’s an overdose. So, they say it’s an overdose and based on that, you’re going to send somebody to help or whatever, or you get the idea it’s an over- dose [based on the technique], you’re going to send somebody. The end result is the same. Thoughts about a post-intervention intervention initiated by a 911 call The machine-learning technique was proposed as a way to identify opioid overdoses in the 911 data, which would allow for the deployment of a peer recovery support specialist or other interventionist to the scene. Respondents perceived both benefits and concerns related to this idea. In general, they thought it was a good idea to send a peer recovery support specialist (i.e., someone with lived experience of substance use who is trained in recovery support), who would know what the patient was going through. This would generate empathy and connec- tion between the patient and the interventionist. Some had seen other similar models, such as mobile crisis intervention teams, and perceived that this kind of intervention might work. PLOS ONE | https://doi.org/10.1371/journal.pone.0223823 October 17, 2019 8 / 14 PWUDs’ perspectives on post-overdose interventions Two primary concerns about the intervention were raised. First, similar to their concerns about the technique to identify opioid overdoses in the 911 dispatch data, participants were concerned about risks to privacy. For some, having an interventionist come to their house would feel invasive, while for others they worried that if the interventionist were known as someone who comes for opioid overdoses, this could compromise confidentiality. Second, some respondents thought that having someone arrive immediately after the overdose was treated was not optimal timing, I think it’s a good idea that they would have the education to know what type of people they are dealing with and to be more prepared, but I think trying to give a person who is just waking up from an overdose a bunch of information as soon as they wake up is pretty much senseless. The concern about timing was mostly related to the anticipation that the patient would be experiencing withdrawal symptoms from the opioid overdose reversal, and therefore would not be able to retain or process the information they were given. Others suggested that the ben- efit of the intervention might be more for the friends or family members or other bystanders, since the patient would likely be transported to the hospital and wouldn’t be able to receive the intervention. One respondent thought that a post-overdose intervention provided on-scene might disincentivize a patient from going to the hospital, which could compromise their health. Recommendations for developing a patient-centered intervention In terms of intervention content, respondents stressed that simply providing referrals is not enough. For some, this sentiment was motivated by experiences in which they hit dead ends when trying to pursue referrals they had been given, leading to the suggestion that there also be active follow-up with patients after the initial intervention and referrals. Others emphasized that the intervention should be flexible enough to allow people to talk about their “real prob- lems”. That is, the intervention should be responsive to and centered on patients’ needs and not exclusively focused on substance abuse treatment. In all three focus groups, respondents spontaneously generated the same solution to their concerns regarding the risks of calling 911 and the perceived risks of the proposed interven- tion: Create an alternative number that could be used to summon emergency medical assis- tance without being connected to the law enforcement dispatch system. For some, this idea was derived from past experiences. For example, one respondent recalled a time when a bystander called directly to the hospital to summon an ambulance for a medical emergency. For others, the idea was generated spontaneously after consideration of the potential risks and benefits of the technique that was the subject of discussion. However, it was not clear that respondents thought their suggestions would be heard. In discussing how law enforcement officers may arrive on scene to support the emergency medical response team, one respondent opined that the decisions about overdose response are made outside the drug using community, The community does not dictate what the first responders do. . .. It’s beyond us. It’s beyond the patient. It’s the cops, the paramedic, they have it all worked out. Whatever they do, I don’t think any victim or any person that needs an ambulance is going to dictate who shows up and who doesn’t. I want the cops. I don’t want the cops. It’s like, it’s—It’s what they do. They do it for some reasons. PLOS ONE | https://doi.org/10.1371/journal.pone.0223823 October 17, 2019 9 / 14 PWUDs’ perspectives on post-overdose interventions Discussion We conducted three focus groups to elicit the perspectives of PWUDs about a machine-learn- ing technique to identify opioid overdose cases in 911 data and a post-overdose intervention that would be deployed to provide naloxone, training in overdose risk reduction techniques, and/or connection to services like methadone, buprenorphine, or syringe services. This study was motivated by a concern that a well-meaning public health-minded intervention could also produce significant harm to the people it is supposed to benefit. Many technologies are inher- ently “dual use.” That is, there is a risk that they can be used to design or produce something that causes harm, when causing that harm is not their primary purpose . In this case, a technique to detect opioid overdose emergencies in 911 dispatch data was conceptualized as a way to increase the speed and accuracy with which an intervention could be deployed to the scene. Upon further deliberation, we also wondered whether it could cause harm by putting PWUDs at greater risk for the criminal justice-based consequences they fear. Our findings suggest that participants’ concerns about the technique to detect opioid overdoses and inter- vention were significant, yet often balanced by the perceived benefits, but only because they already believe that a law enforcement response will be mobilized, no matter what they say or do. Fears about and the inevitability of a law enforcement response to a 911 call was pervasive in the discussions. So much so, that for many respondents, “calling 911” was synonymous with “calling the cops.” We cannot overstate the significance of this finding. Unlike many members of the general public, who do not often question that a call to 911 will result in a helpful emer- gency medical response, PWUDs are among those who weigh the risks associated with a call for help and perceive that harmful actions could result [9, 10, 30]. Respondents in this study expressed a desire to seek medical assistance to save the life of someone experiencing an opioid overdose and thought the proposed technique and interven- tion could help in that effort, which is consistent with a large body of research that demon- strates that PWUD are willing and able to take action in the event of witnessing opioid overdoses [e.g., 14, 31, 32]. Our findings demonstrate that the desire to seek that help is tem- pered by fears of a potential law enforcement response, which could result in CPS involve- ment, arrest, incarceration, and threats to personal reputation and privacy. Some respondents mentioned the threat of being charged with murder if they are at the scene of an overdose–a phenomenon that has become more common across the US during this opioid epidemic , and is in direct conflict with efforts to engage PWUDs in potentially life-saving intervention efforts. Other concerns related to the machine-learning technique included a loss of autonomy since the caller’s choice not to disclose could be overridden, and a fear that an inaccurate algo- rithm could result in misdiagnosis and delayed care. In terms of the intervention model, respondents generally perceived benefit in the deploy- ment of an intervention, especially if the interventionist was a peer, or person with lived expe- rience of substance use, who could be empathetic and provide relevant information and resources. Concerns about the intervention were related to violation of privacy (similar to con- cerns about identifying opioid overdoses in the 911 data), and a sense that the moments imme- diately post-overdose are suboptimal for intervention delivery due to the experience of naloxone-precipitated withdrawal symptoms. Referrals alone were seen as insufficient, and more active follow-up and assistance were recommended. Participants emphasized that an intervention should be responsive to their “real problems”, rather than focusing exclusively on substance use disorder treatment linkage. Across all three focus groups, participants expressed a desire for an alternative number that could summon emergency medical assistance without triggering a law enforcement response. However, fundamentally, participants in the current study expressed a sense of helplessness, stemming from the sense that their ideas for how PLOS ONE | https://doi.org/10.1371/journal.pone.0223823 October 17, 2019 10 / 14 PWUDs’ perspectives on post-overdose interventions interventions and emergency response systems should work would not be heeded, and that as with most drug-related polices in the US, any future interventions would be both designed and implemented from above without their input. A patient centered perspective on the development of post-overdose interventions could be helpful to inform a more effective model. Patient centered care (PCC) is a perspective that pri- oritizes attention to the patient’s experience of illness and health care in the design and delivery of high quality care . PCC is guided by 7 dimensions: 1) respect for the patient’s values, preferences, and expressed needs; 2) coordination and integration of care; 3) information, communication, and education; 4) physical comfort; 5) emotional support and alleviation of fear and anxiety; 6) involvement of family and friends; and 7) transition and continuity . In the current study, participants provided perspectives on an ideal intervention that addressed nearly all of these 7 dimensions. Participants expressed preferences and needs related to how an intervention should be delivered (by people with lived experience, at a time that the patient is receptive, in a manner that addresses the patients’ “real needs” and that pro- tects patient privacy and safety). They discussed barriers to access to care for opioid overdose (risks associated with calling 911) and proposed a solution (an alternative number to summon emergency medical assistance), highlighted the value of peer support specialists as empathetic interventionists who can provide support, and described how the intervention could benefit patients and their families and friends. The recommendation to provide active follow-up for patients (rather than just referrals) could serve to bolster continuity of care, secure transitions to services, and coordinate care across sectors. Finally, participants discussed timing the inter- vention to ensure that patients were not physically uncomfortable due to precipitated with- drawal. The only dimension of PCC that was not explicitly discussed by participants in these focus groups was “information and education”, which Gerteis  explains is related to patients’ fear that information is being withheld from them during their care. In our study, respondents demonstrated a high level of understanding regarding the mechanisms underly- ing opioid overdose and the recommended treatment, which affirms previous research dem- onstrating high levels of knowledge related to opioid overdose among PWUDs [36, 37]. Limitations This study should be considered in light of its limitations. Attendance at the focus groups was low, resulting in 3 “mini” focus groups (FG1 n = 2, FG2 n = 6, FG3 n = 3), which is fewer than the recommended 8–10 participants per group. This may have heightened the risk that responses were subject to social desirability bias, especially because focus group-based inquiry is expected to uncover normative beliefs about the topic of study  and there could have been greater normative pressure in smaller groups. However, we did observe disagreement among some respondents, suggesting that at least some of them felt comfortable expressing non-nor- mative opinions. Respondents for this study live in a small city in the Western US, therefore their attitudes and beliefs may not be reflective of PWUDs in other communities. Their reports may also reflect a diversity of historical experiences and are not necessarily accounts of recent experiences or experiences in their current city. However, our findings related to concerns about calling 911 are consistent with other studies that used different methods (some conducted in larger urban areas; e.g., [14, 22]), lending support for our conclusions. This study was under- taken to elicit opinions about a proposed intervention that had not yet been implemented. Therefore, the participants had not yet had the opportunity to experience the program, and their opinions may change after exposure to the program. Future research should examine the experiences of and attitudes towards post-overdose intervention programs in the early and later stages of implementation to determine their acceptability among PWUDs. PLOS ONE | https://doi.org/10.1371/journal.pone.0223823 October 17, 2019 11 / 14 PWUDs’ perspectives on post-overdose interventions Conclusion The current study builds upon the existing literature by encouraging us to reflect critically on the use of 911 data for the purposes of surveillance and intervention deployment. Interventions that are predicated on engagement with the 911 system for their delivery will only reach a small proportion of the population: those who overdose in the presence of people willing to call 911. For others, including people who use drugs alone and people who use drugs in the presence of individuals who fear calling 911, access to interventions designed to link opioid overdose patients with follow up care and resources will be severely constrained. Relatedly, surveillance efforts that seek to enumerate opioid overdose cases or other types of events based on 911 calls should recognize the inherent limitations of 911 data sets, which consist only of those cases for which a 911 call is made. This may create selection bias by which particularly marginalized or vulnerable individuals are underrepresented, and majority populations are overrepresented. Public health systems must account for the fact that a large share of overdoses will never come to the attention of the emergency response system until fears are mitigated and trust is restored. In the context of the current North American opioid epidemic, innovative intervention strategies to reduce opioid overdose deaths are urgently needed. To maximize effectiveness, it is imperative to center the perspectives of PWUDs when designing and implementing inter- ventions. In exploring PWUDs’ opinions about a 911-based post-overdose intervention, we found that respondents identified the potentially life-saving nature of a post-overdose inter- vention. However, fears related to using the 911 system to summon emergency medical assis- tance were pervasive. Incorporating a patient-centered perspective in intervention design may help improve outcomes and reduce opioid overdose mortality. Acknowledgments We offer thanks to the research participants and community partners who made this research possible. We would like to acknowledge the contributions of Patrick Reuther, MPH; Leah Baker, MPH; Krysti Smith, MPA; and our local EMS agencies and uniformed first responders. Author Contributions Conceptualization: Karla D. Wagner, Robert W. Harding, Richard Kelley, Brian Labus, Silvia R. Verdugo, Elizabeth Copulsky, Jeanette M. Bowles, Maria Luisa Mittal, Peter J. Davidson. Data curation: Robert W. Harding. Formal analysis: Karla D. Wagner. Funding acquisition: Karla D. Wagner, Brian Labus, Silvia R. Verdugo, Peter J. Davidson. Investigation: Karla D. Wagner, Robert W. Harding, Richard Kelley, Peter J. Davidson. Methodology: Karla D. Wagner, Richard Kelley, Brian Labus, Silvia R. Verdugo, Peter J. Davidson. Project administration: Karla D. Wagner, Robert W. Harding. Resources: Karla D. Wagner. Validation: Robert W. Harding, Brian Labus, Silvia R. Verdugo, Elizabeth Copulsky, Jeanette M. Bowles, Maria Luisa Mittal, Peter J. Davidson. Writing – original draft: Karla D. Wagner. PLOS ONE | https://doi.org/10.1371/journal.pone.0223823 October 17, 2019 12 / 14 PWUDs’ perspectives on post-overdose interventions Writing – review & editing: Karla D. Wagner, Robert W. Harding, Brian Labus, Silvia R. Ver- dugo, Elizabeth Copulsky, Jeanette M. Bowles, Maria Luisa Mittal, Peter J. Davidson. References 1. Jalal H, Buchanich JM, Roberts MS, Balmert LC, Zhang K, Burke DS. Changing dynamics of the drug overdose epidemic in the United States from 1979 through 2016. Science. 2018; 361(6408). 2. Hedegaard M, Miniño A, Warner M. NCHS Data Brief: No. 329: November 2018: Drug overdose deaths in the United States, 1999–2017. 2018. 3. Kariisa M. Drug Overdose Deaths Involving Cocaine and Psychostimulants with Abuse Potential— United States, 2003–2017. MMWR Morbidity and mortality weekly report. 2019;68. 4. Tyndall MJ, Craib KJ, Currie S, Li K, O’Shaughnessy MV, Schechter MT. Impact of HIV Infection on Mortality in a Cohort of Injection Drug Users. Journal of Acquired Immune Deficiency Syndromes. 2001; 28(4):351–7. https://doi.org/10.1097/00126334-200112010-00008 PMID: 11707672 5. Maxwell S, Bigg D, Stanczykiewicz K, Carlberg-Racich S. Prescribing naloxone to actively injecting her- oin users: a program to reduce heroin overdose deaths. J Addict Dis. 2006; 25(3):89–96. https://doi.org/ 10.1300/J069v25n03_11 PMID: 16956873. 6. Wheeler E, Jones TS, Gilbert MK, Davidson PJ, Centers for Disease C, Prevention. Opioid Overdose Prevention Programs Providing Naloxone to Laypersons—United States, 2014. MMWR Morb Mortal Wkly Rep. 2015; 64(23):631–5. PMID: 26086633. 7. Boyer EW. Management of Opioid Analgesic Overdose. New England Journal of Medicine. 2012; 367 (2):146–55. https://doi.org/10.1056/NEJMra1202561 PMID: 22784117. 8. Institute of Medicine. Emergency Medical Services: At the Crossroads. Washington, DC: The National Academies Press; 2007. 9. Skolarus LE, Murphy JB, Zimmerman MA, Bailey S, Fowlkes S, Brown DL, et al. Individual and commu- nity determinants of calling 911 for stroke among African Americans in an urban community. Circ Cardi- ovasc Qual Outcomes. 2013; 6(3):278–83. https://doi.org/10.1161/CIRCOUTCOMES.111.000017 PMID: 23674311; PubMed Central PMCID: PMC3779662. 10. Sasson C, Haukoos JS, Ben-Youssef L, Ramirez L, Bull S, Eigel B, et al. Barriers to calling 911 and learning and performing cardiopulmonary resuscitation for residents of primarily Latino, high-risk neigh- borhoods in Denver, Colorado. Ann Emerg Med. 2015; 65(5):545–52 e2. https://doi.org/10.1016/j. annemergmed.2014.10.028 PMID: 25481112; PubMed Central PMCID: PMC4866505. 11. Strang J, Best D, Man L, Noble A, Gossop M. Peer-initiated Overdose Resuscitation: Fellow Drug Users Could Be Mobilised to Implement Resuscitation. International Journal of Drug Policy. 2000; 11 (6):437–45. PMID: 11099924 12. Tobin KE, Davey MA, Latkin CA. Calling Emergency Medical Services During Drug Overdose: An Examination of Individual, Social and Setting Correlates. Addiction. 2005; 100(3):397–404. https://doi. org/10.1111/j.1360-0443.2005.00975.x PMID: 15733253 13. Ambrose G, Amlani A, Buxton JA. Predictors of seeking emergency medical help during overdose events in a provincial naloxone distribution programme: a retrospective analysis. BMJ open. 2016; 6(6): e011224. https://doi.org/10.1136/bmjopen-2016-011224 PMID: 27329442 14. Pollini RA, McCall L, Mehta SH, Celentano DD, Vlahov D, Strathdee SA. Response to Overdose Among Injection Drug Users. American Journal of Preventive Medicine. 2006; 31(3):261–4. https://doi. org/10.1016/j.amepre.2006.04.002 PMID: 16905039 15. Wagner KD, Valente TW, Casanova M, Partovi SM, Mendenhall BM, Hundley JH, et al. Evaluation of an overdose prevention and response training programme for injection drug users in the Skid Row area of Los Angeles, CA. Int J Drug Policy. 2010; 21(3):186–93. https://doi.org/10.1016/j.drugpo.2009.01. 003 PMID: 19268564; PubMed Central PMCID: PMC4291458. 16. Darke S, Ross J, Zador D, Sunjic S. Heroin-related deaths in new south wales, Australia, 1992–1996. Drug and alcohol dependence. 2000; 60(2):141–50. https://doi.org/10.1016/s0376-8716(99)00147-7 PMID: 10940541 17. Karamouzian M, Kuo M, Crabtree A, Buxton JA. Correlates of seeking emergency medical help in the event of an overdose in British Columbia, Canada: Findings from the Take Home Naloxone program. Int J Drug Policy. 2019. Epub 2019/01/25. https://doi.org/10.1016/j.drugpo.2019.01.006 PMID: 18. Davis C, Webb D, Burris S. Changing law from barrier to facilitator of opioid overdose prevention. J Law Med Ethics. 2013; 41 Suppl 1:33–6. Epub 2013/04/23. https://doi.org/10.1111/jlme.12035 PMID: PLOS ONE | https://doi.org/10.1371/journal.pone.0223823 October 17, 2019 13 / 14 PWUDs’ perspectives on post-overdose interventions 19. Lankenau SE, Wagner KD, Silva K, Kecojevic A, Iverson E, McNeely M, et al. Injection drug users trained by overdose prevention programs: responses to witnessed overdoses. J Community Health. 2013; 38(1):133–41. Epub 2012/08/01. https://doi.org/10.1007/s10900-012-9591-7 PMID: 22847602; PubMed Central PMCID: PMC3516627. 20. Latimore AD, Bergstein RS. "Caught with a body" yet protected by law? Calling 911 for opioid overdose in the context of the Good Samaritan Law. Int J Drug Policy. 2017; 50:82–9. Epub 2017/10/16. https:// doi.org/10.1016/j.drugpo.2017.09.010 PMID: 29040841. 21. Bohnert AS, Nandi A, Tracy M, Cerda M, Tardiff KJ, Vlahov D, et al. Policing and risk of overdose mor- tality in urban neighborhoods. Drug Alcohol Depend. 2011; 113(1):62–8. https://doi.org/10.1016/j. drugalcdep.2010.07.008 PMID: 20727684; PubMed Central PMCID: PMC3008306. 22. Koester S, Mueller SR, Raville L, Langegger S, Binswanger IA. Why are some people who have received overdose education and naloxone reticent to call Emergency Medical Services in the event of overdose? Int J Drug Policy. 2017; 48:115–24. https://doi.org/10.1016/j.drugpo.2017.06.008 PMID: 28734745. 23. Formica SW, Apsler R, Wilkins L, Ruiz S, Reilly B, Walley AY. Post opioid overdose outreach by public health and public safety agencies: Exploration of emerging programs in Massachusetts. Int J Drug Pol- icy. 2018; 54:43–50. Epub 2018/02/08. https://doi.org/10.1016/j.drugpo.2018.01.001 PMID: 29414484. 24. Wagner KD, Bovet LJ, Haynes B, Joshua A, Davidson PJ. Training law enforcement to respond to opi- oid overdose with naloxone: Impact on knowledge, attitudes, and interactions with community mem- bers. Drug Alcohol Depend. 2016; 165:22–8. https://doi.org/10.1016/j.drugalcdep.2016.05.008 PMID: 25. Schiff DM, Drainoni ML, Weinstein ZM, Chan L, Bair-Merritt M, Rosenbloom D. A police-led addiction treatment referral program in Gloucester, MA: Implementation and participants’ experiences. J Subst Abuse Treat. 2017; 82:41–7. Epub 2017/09/09. https://doi.org/10.1016/j.jsat.2017.09.003 PMID: 26. McNeil R, Kerr T, Pauly B, Wood E, Small W. Advancing patient-centered care for structurally vulnera- ble drug-using populations: a qualitative study of the perspectives of people who use drugs regarding the potential integration of harm reduction interventions into hospitals. Addiction (Abingdon, England). 2016; 111(4):685–94. Epub 01/08. https://doi.org/10.1111/add.13214 PMID: 26498577. 27. Patton MQ. Qualitative Research & Evaluation Methods. Thousand Oaks, CA: Sage Publications, Inc.; 28. Lofland J, Lofland LH. Analyzing Social Settings: A Guide to Qualitative Observation and Analysis. 3 ed: Wadsworth; 1995. 29. Forge J. A note on the definition of “dual use”. Science and Engineering Ethics. 2010; 16(1):111–8. https://doi.org/10.1007/s11948-009-9159-9 PMID: 19685170 30. Brunson RK. “Police don’t like black people”: African-American young men’s accumulated police experi- ences. Criminology & public policy. 2007; 6(1):71–101. 31. Wagner KD, Davidson PJ, Iverson E, Washburn R, Burke E, Kral AH, et al. "I felt like a superhero": the experience of responding to drug overdose among individuals trained in overdose prevention. Int J Drug Policy. 2014; 25(1):157–65. Epub 2013/08/13. https://doi.org/10.1016/j.drugpo.2013.07.003 PMID: 23932166; PubMed Central PMCID: PMC3946806. 32. Seal KH, Thawley R, Gee L, Bamberger J, Kral AH, Ciccarone D, et al. Naloxone Distribution and Car- diopulmonary Resuscitation Training for Injection Drug Users to Prevent Heroin Overdose Death: A Pilot Intervention Study. Journal of Urban Health. 2005; 82(2):303–11. https://doi.org/10.1093/jurban/ jti053 PMID: 15872192 33. LaSalle L. An Overdose Death is not Murder: Why Drug-induced homicide laws are counterproductive and inhumane. The Drug Policy Alliance, 2017. 34. Institute of Medicine. Crossing the Quality Chasm: A New Health system for the 21st Century. Wash- ington, D.C.: National Academy Press, 2001. 35. Gerteis M. Through the patient’s eyes: understanding and promoting patient-centered care. 1993. 36. Green TC, Heimer R, Grau LE. Distinguishing Signs of Opioid Overdose and Indication for Naloxone: An Evaluation of Six Overdose Training and Naloxone Distribution Programs in the United States. Addiction. 2008; 103(6):979–89. Epub 2008 Apr 16. https://doi.org/10.1111/j.1360-0443.2008.02182.x PMID: 18422830 37. Behar E, Santos G-M, Wheeler E, Rowe C, Coffin PO. Brief overdose education is sufficient for nalox- one distribution to opioid users. Drug and Alcohol Dependence. 2015; 148(0):209–12. http://dx.doi.org/ 10.1016/j.drugalcdep.2014.12.009. 38. Sussman S, Burton D, Dent CW, Stacy AW, Flay BR. Use of Focus Groups in Developing a Adolescent Tobacco Use Cessation Program: Collective Norm Effects. Journal of Applied Social Psychology. 1991; 21(21):1772–82. PLOS ONE | https://doi.org/10.1371/journal.pone.0223823 October 17, 2019 14 / 14
PLoS ONE – Public Library of Science (PLoS) Journal
Published: Oct 17, 2019
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.
All for just $49/month
Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly
Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.
Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.
Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.
All the latest content is available, no embargo periods.
“Hi guys, I cannot tell you how much I love this resource. Incredible. I really believe you've hit the nail on the head with this site in regards to solving the research-purchase issue.”Daniel C.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud
“I must say, @deepdyve is a fabulous solution to the independent researcher's problem of #access to #information.”@deepthiw
“My last article couldn't be possible without the platform @deepdyve that makes journal papers cheaper.”@JoseServera