Impact of the Take ACTION Train-the-Trainer Model of Opioid Overdose Education with Naloxone Distribution– Who Benefits?

Abstract

Background:

Overdose education with naloxone distribution (OEND) is a key national strategy to reduce morbidity and mortality related to opioid overdoses. Train-the-trainer model has been one method to increase the pool of trainers to facilitate greater dissemination of OEND. This exploratory study seeks to 1) evaluate participant’s change in knowledge and confidence, 2) examine if pre- and post-training test outcomes differed by occupation and level of experience, and 3) determine if train-the-trainer participants trained others six months later.

Methods:

Fifteen train-the-trainer sessions were delivered to staff from community organizations who served high-risk clients in four counties whose overdose death rates ranged from 11.2–32.8 per 100,000. Participants were administered pre- and post-training tests from September 2017 to December 2018. A follow-up survey was conducted six months post-training to evaluate outcomes. Final paired pre-and post-training surveys of 109 participants were used for analysis. Paired sample t-tests were used to evaluate changes in the knowledge and confidence in teaching others. One-way ANOVA compared the change scores across groups with different demographic or experiential characteristics. Kruskal-Wallis Test was used for Likert scales.

Results:

The one and one-half hour train-the-trainer curriculum increased participants’ knowledge and confidence to teach others (p < .001). This was particularly true for participants who had no prior experience compared to those who had some experience with naloxone (p = .0003). Changes in confidence to teach others significantly improved among demographic subgroups of participants. At six months after completing the train-the-trainer curriculum, 14 participants had trained 243 new trainees.

Conclusions:

Implementing a train-the-trainer model for OEND increases knowledge and participants’ confidence to teach others. This demonstrates the important potential of the train-the-trainer model to respond to the growing opioid overdose epidemic.

INTRODUCTION

Deaths from opioid overdose have reached epidemic levels in the United States. Since 1999, nearly 400,000 deaths have been attributed to unintentional prescription and illicit opioid overdoses.¹ An estimated 130 Americans die daily from an opioid overdose.² The most recent statistics show that the total economic burden associated with opioid overdose is over $78.5 billion in the United States, with almost two-thirds of the costs related to health care, substance use treatment, and lost productivity from non-fatal overdoses.³

Naloxone hydrochloride is a pure opioid antagonist used to treat opioid overdose. Naloxone reverses respiratory depression caused by opioid ingestion and has been used by medical personnel for more than 40 years. Naloxone does not have any pharmacokinetic activity in the absence of narcotics and is considered safe if administered in low doses and titrated until the person resumes breathing.^4–6 The sooner naloxone is administered, the less likelihood there is of cerebral damage or organ failure. Because opioid overdoses can occur from seconds to hours after ingestion^7–10 and more likely to happen in the presence of others,^11,12 time-sensitive opportunities exist for intervention.

To address the growing opioid overdose problem, the Centers for Disease Control (CDC) recommend the following evidence-based strategies: 1) overdose education with naloxone distribution (OEND) to first responders and community laypeople, and 2) a co-prescription of naloxone when opioids are prescribed.¹³ Moreover, the Surgeon General has advised health care practitioners to learn how to use and make naloxone easily accessible.¹⁴ There is strong evidence showing that OEND programs are feasible,^15,16 reduce overdose death rates,¹⁷ and exhibit a dose-response relationship. Indeed, in areas with higher naloxone distribution there is lower opioid overdose mortality.¹⁸ However, despite the feasibility of OEND, there is a need to increase naloxone training in areas hit hardest by opioid overdoses.

To scale up OEND and increase naloxone distribution, a train-the-trainer model for OEND has been implemented.^19,20 The goal of the train-the-trainer model is to build a pool of trainers who can teach others to implement public health interventions, thereby facilitating wider dissemination. A train-the-trainer model has been used successfully in a variety of health contexts to deliver education on opioid prescribing,²¹ HIV prevention,²² and tobacco cessation treatment.²³

Typically, a train-the-trainer model involves a master trainer to coach new trainers, and then these new trainers will teach the information to others in their local contexts. The advantage of a train-the-trainer model is that it has the potential to quickly build capacity.^24,25 Madah-Amiri and colleagues were able to effectively train over 500 staff using three central trainers to distribute nearly 2000 naloxone kits over an 18-month period. From the pre- and post-training tests and surveys of 54 train-the-trainer participants, significant improvements in knowledge and attitudes were reported.¹⁹ Although this study highlights the possible reach of a train-the-trainer model, Madah-Amiri et al, did not examine differences in training and knowledge outcomes by type of occupation and levels of professional experience. Furthermore, the authors did not report if the train-the-trainer participants were able to train others after completing their training program.

The purpose of this exploratory study was to further expand on OEND by using a train-the-trainer model to increase the number of individuals able to train others to use naloxone, thereby increasing naloxone distribution in the community. We aim to: 1) evaluate the change in pre- and post-training test knowledge and confidence after receiving a one and one-half hour train-the-trainer naloxone curriculum, 2) determine if the change in pre- and post-training test outcomes (e.g., knowledge and confidence) differed by experiential characteristics and occupation of the participants, and 3) determine whether educating a cohort of learners through the train-the-trainer curriculum generates an increase in the number of individuals who are trained to deliver naloxone six months later.

METHODS

Setting and Participants

From September 2017 to December 2018, train-the-trainer sessions were held in various locations within four counties in Southeast Michigan. Opioid overdose death rates per 100,000 residents in each of the counties ranged from 11.2–32.8.²⁶ Organizations were targeted based on their service to high-risk clients. Similar to Madah-Amiri et al, low threshold facilities such as day/overnight shelters, medical facilities, and jails were selected.¹⁹ Furthermore, non-profit organizations such as substance use treatment facilities, charitable organizations, school nurses, and government organizations including law enforcement agencies, community mental health, and public health departments were selected due to the at-risk clients each of these organizations serve. The majority of law enforcement agencies were trained prior to September 2017, leaving only a few law enforcement agencies in need of training. Each organization sent between one and ten participants to receive the train-the-trainer naloxone curriculum. The didactic portion of the training lasted for one hour. One central trainer, a doctorally-prepared nurse practitioner and lead author, facilitated 15 train-the-trainer sessions. All trainings were held in a central location in the community.

Train-the-Trainer Naloxone Curriculum

The train-the-trainer naloxone curriculum titled, “Take A.C.T.I.O.N.: Opioid Overdose Prevention Education”,²⁷ was adapted from Substance Abuse and Mental Health Services Administration (SAMHSA) Opioid Overdose TOOLKIT²⁸ and Harm Reduction Coalition.²⁹ Modifications to the train-the-trainer curriculum were made prior to this study after receiving feedback from a law enforcement agency and a substance use treatment organization. The curriculum consisted of: 1) background and root causes associated with the opioid epidemic; 2) overdose prevention legislation; 3) risk factors; 4) pathophysiology of overdoses; 5) pharmacology of naloxone and opioids; 6) how to respond to opioid overdoses using A.C.T.I.O.N. acronym (Arouse the person, Check signs of an overdose, Telephone 911, Intranasal/Intramuscular naloxone, Oxygen, and Naloxone again; See Appendix for patient education brochure); 7) post-overdose care; 8) documentation procedures for future trainings, reporting overdose reversals, and obtaining additional naloxone kits; and 9) how to have layperson conversations about naloxone and overdose prevention. All participants had basic life support certification, but this was not a pre-requisite for participation.

Procedure

For each train-the-trainer session, an anonymous pre-test survey was administered, with a post-test survey completed after the train-the-trainer curriculum was delivered. One hundred and forty-three participants completed the pre-test survey and 128 completed the post-test survey immediately after the training. The final paired pre- and post-test surveys consisted of 109 participants resulting in a 76% participation rate. The trainings were delivered through a PowerPoint presentation that included videos of a live overdose scenario and naloxone administration. Each participant completed a hands-on opioid overdose response demonstration. The A.C.T.I.O.N. acronym was used as a memory aid and participants demonstrated with a manikin as their final check-off. All participants received naloxone kits to take to their agencies as well as a folder that contained information such as reporting procedures, materials for training others, and for ordering naloxone. A follow-up email survey was sent to the participants (N = 54) who provided email addresses six months post-training to evaluate if train-the-trainer participants trained others (i.e., staff, laypeople, or others). Twenty-two participants responded to the follow-up survey resulting in a response rate of 40%. This study was deemed exempt by the University of Michigan Institutional Review Board.

Measures

Pre- and Post-Test Assessment.

Opioid overdose knowledge (OOKS) was assessed using an adapted OOKS scale.³⁰ The OOKS is a 45-item survey that assesses for four domains: 1) Risk - overdose risk factors (range 0–8); 2) Signs - signs of an overdose (range 0–10); 3) Actions - actions to take in an overdose (range 1–11); and 4) Naloxone use - appropriate use of naloxone (range 0–9). The OOKS has been used widely in multiple contexts to assess for opioid overdose knowledge and is internally reliable (Cronbach’s alpha = 0.83).³¹ For the purposes of our study, the OOKS was modified for readability to an American audience and type of naloxone formulation used for the training. We omitted questions pertaining to intramuscular naloxone. The modified OOKS contained 38 items with the total knowledge score that ranged from 0–38.

Trainer confidence to teach others was assessed using an adaptation of Madah-Amiri and colleagues’¹⁹ 5-point Likert scale (1 = strongly disagree to 5 = strongly agree). Confidence questions focused on: 1) teaching others about overdose risk factors, 2) recognizing signs of an overdose, 3) responding to an overdose, 4) using naloxone, and 5) knowing the next steps to care post-naloxone use. The individual confidence scores were summed and the mean was calculated to represent the total confidence score (range 1–5).

Post-Train-the-Trainer Outcome Assessment.

Outcomes were assessed with a 10-item survey at six months after the trainings. The outcome assessment asked the train-the-trainer participant about the numbers of people trained, the numbers of naloxone distributed, the barriers to training others, the training resources and lack thereof, and how to facilitate naloxone distribution in their local area. For participants that did not train others, questions regarding barriers to training and ways to facilitate training were asked.

Demographic variables included age, sex, race/ethnicity, range of years working with drug users, and occupation. The occupation variable was collapsed into four types of occupation: 1) human services, 2) health care, 3) law enforcement, and 4) school system.

Prior experience variables included dichotomous responses (yes or no) to: 1) previous naloxone training, 2) ever witnessed an overdose, and 3) ever used naloxone? We calculated a total prior experience variable by summing these three questions and then collapsed to two categories of “no experience” and “some experience.” A score of 0 was categorized as “no experience” and scores of 1–3 were categorized as “some experience.”

Data Analysis

We used SAS software version 9.4³² for all analyses. Analytic techniques were relevant to the study purpose and included descriptive statistics to describe the participants’ characteristics. Pearson’s correlations were used to examine changes in pre- and post-training test scores in knowledge gained. We also examined pre- and post-training test knowledge scores across the four knowledge domains (i.e., overdose risk factors, signs of overdose, actions to be taken, and naloxone use) and across confidence to teach others (i.e., about opioid overdose risk factors, how to recognize the signs of an opioid overdose, how to respond to an opioid overdose, how to use naloxone, and what to do after giving naloxone). Paired sample t-tests were used to evaluate changes in the knowledge and the total confidence in teaching scores. Effect sizes were calculated and interpreted using the Cohen’s criteria (0.2 = small, 0.5 = medium, 0.8 = large, 1.2 = very large, 2.0 = huge).³³ And finally, one-way ANOVA was used to compare the change scores across groups of participants with different demographic or experiential characteristics. Kruskal-Wallis Test was used for Likert scales.

Qualitative responses from the open-ended question in the pre-test survey, “What is (are) the reason(s) for taking this course?”, were coded into categories using open coding analysis.

RESULTS

Over a 15-month period, 143 participants completed the pre-test survey and 128 participants completed the post-test survey immediately following the training session. The total paired pre- and post-test surveys included 109 participants (Table 1). Of the 109 participants, the majority were female (n = 66, 60.6%), White (n = 66, 60.6%), and between the ages of 30–49 (n = 60, 60%). We had a limited number of other ethnic groups (Arab [n = 1], Asian [n = 2], Hispanic [n = 1], and Other [n = 2]). As a result, the ethnic groups were combined to indicate “Other.” Thirty-three percent (n = 36) of the participants had greater than 10 years of working with drug users and 37% (n = 40) reported witnessing an opioid overdose. However, only 16 (15%) indicated ever using naloxone. Over one-third of the participants (n = 39, 36%) had previously received naloxone training.

Table 1:

Demographic Characteristics

	N	%
Gender
Male	42	38.5
Female	66	60.6
I prefer not to say	1	0.9
Race/Ethnicity
African American	5	4.6
White	66	60.6
Other	6	5.5
Missing	32	29.3
Age
20 – 29	19	17.4
30 – 39	30	27.5
40 – 49	30	27.5
50 – 59	25	22.9
60 – 69	4	3.7
70 – 79	1	0.9
Number of years working with people who use drugs
0–1 year	21	19.3
1–2 years	7	6.4
3–5 years	12	11.0
6–10 years	20	18.4
>10 years	36	33.0
Missing	13	11.9
Have you witnessed a drug overdose
Yes	40	36.7
No	68	62.4
Missing	1	0.9
Have you used naloxone?
Yes	16	14.7
No	92	84.4
Missing	1	0.9
Have you received naloxone training before?
Yes	39	35.8
No	67	61.5
Missing	3	2.7
Type of Occupation
Human Services	39	35.8
Healthcare Field	31	28.4
Law Enforcement	23	21.1
School System	9	8.3
Other	4	3.7
Missing	3	2.7

Qualitative comments reported by the participants (N = 96) for the reason(s) of taking this course included six broad categories: 1) train others, 2) increase knowledge and be confident in responding to overdoses, 3) help clients in the workplace, 4) save lives and serve the community, 5) satisfy job requirements, and 6) distribute naloxone. Select comments included, “I would like to be better informed and be able to train the men at the recovery home I manage”, “I am a home visitor going into homes of drug users”, and “I would like to be confident in what to do in case of an opioid overdose.”

One participant’s comment encapsulated the majority of the categories: “I work with a county funded program…We talk to schools, treatment providers, and community stakeholders about the importance of providing treatment and options to those who are facing addiction. As an active voice for recovery in my community, it’s important to me to be able to train others in my community on how to save a life if someone is overdosing on opiates. I am also a probation officer, and I work with addicted individuals daily. It’s important to me to be able to provide individuals and their families with all the education possible. Naloxone is an important tool in the struggle of opiate/opioid addiction.”

In order of preference for future formats, participants preferred the in-person training, followed by web-based training, web-based training with check-off, and video PowerPoint presentation. Other suggested topics to include in the training were CPR, how to talk to patients after surgical procedures, protocols for aftercare, and how to respond to other drug overdoses. The majority of participants believed it was a thorough training and were grateful to have received the training stating, “I feel very confident and informed now”, “Wonderful presentation, thank you!”

Opioid Overdose Knowledge and Confidence in Teaching Score

The overall knowledge scores as well as scores for each individual knowledge domain improved significantly after the training (Table 2). The effect size was small to very large in all domains with a very large effect seen in naloxone use (d = 1.144) and total knowledge score (d = 1.47). The effect size was smaller for risks (d = 0.271). Similarly, participants’ confidence to teach others about overdose risk factors, identify signs of overdose, respond (action) to an overdose, use naloxone, and deliver post-naloxone care increased significantly (p < .001).

Table 2:

Opioid Overdose Knowledge and Confidence in Teaching Scores

	Pre-test		Post-test		Change		Paired Sample t-test		Effect size
	Mean	SD	Mean	SD	Mean	SD	T	p value
Knowledge 1
Actions (out of 11)	10.17	0.92	10.63	0.59	0.47	1.03	−4.77	<0.001	0.608
Signs (out of 10)	7.57	1.51	9.06	1.14	1.49	1.48	−10.51	<0.001	1.328
Naloxone Use (out of 9)	6.68	1.95	8.72	0.51	2.05	1.95	−10.93	<0.001	1.144
Risks (Out of 8)	7.46	1.17	7.74	0.81	0.28	0.96	−3.08	0.003	0.271
Total (Out of 38)	31.87	3.51	36.16	1.72	4.28	3.33	−13.42	<0.001	1.47
Confidence in Teaching Others 2
Risk (out of 5)	3.24	1.18	4.59	0.55	1.36	1.12	−12.60	<0.001	1.398
Sign (out of 5)	3.23	1.14	4.65	0.50	1.42	1.12	−13.30	<0.001	1.537
Action (out of 5)	3.22	1.13	4.65	0.50	1.43	1.11	−13.48	<0.001	1.626
Naloxone use (out of 5)	2.95	1.22	4.70	0.48	1.74	1.24	−14.64	<0.001	1.829
What to do after giving naloxone (out of 5)	2.98	1.20	4.66	0.51	1.68	1.19	−14.79	<0.001	1.741
	Mean	SD	Mean	SD	Mean	SD	T	p value
Total (1–5)	3.12	1.10	4.64	0.48	1.52	1.06	−14.90	<0.001	1.706

¹Our Train-the-Trainer Opioid Overdose Knowledge scale was revised from Williams et al., 2013 which contained 38 items with a total score range of 1–38 items.

²The confidence questions based on 1–5 Likert scale were revised from Madah-Amiri et al., 2016. The total confidence score was a sum of the individual scores and calculated as a mean with a range of 1–5.

Changes in Overdose Knowledge and Confidence in Teaching by Experience and Occupation

There was a statistically significant change in pre- and post-training test knowledge scores among participants by differing levels of experience and by type of occupation (Table 3). When examining participants’ levels of experience (no experience vs. some experience) with their gains in knowledge, those who had no experience had significantly greater gains in knowledge than participants with some experience. However, the change in knowledge scores did not depend upon the type of occupation (Table 3). When examining participants’ levels of experience with their changes in confidence in teaching, participants who had no experience had significantly greater improvement in their confidence to teach than those who had some experience. Unlike knowledge gains, significant differences in confidence to teach were observed among different types of occupation. Participants who worked in the school system had significantly higher changes in their confidence to teach others than the other occupation types (Table 4).

Table 3:

Change in Overdose Knowledge Based on Experience Level and Type of Occupation

	N	Pre-Test Total Knowledge		Post-Test Total Knowledge		One-way ANOVA1		Total Knowledge Change		One-way ANOVA Change2
		M	SD	M	SD	F	p	M	SD	F	p
Experience Level 3						−2.24	0.03			3.75	<.001
No Experience	49	30.22	3.27	35.76	1.85			5.53	3.35
Some Experience	59	33.25	3.14	36.49	1.56			3.24	3.00
Occupation Type						2.58	0.04			1.22	0.31
School System	9	30.54	3.73	34.56	1.94			4.89	2.85
Law Enforcement	23	32.13	3.44	36.00	2.02			3.87	2.78
Health Field	31	31.10	3.75	36.22	1.45			5.13	3.85
Human Services	39	32.74	3.11	36.51	1.57			3.77	3.16

¹One-way ANOVA was conducted to compare pre-and post-test knowledge scores on level of experience and type of occupation.

²One-way ANOVA was conducted to compare change in overdose knowledge scores on level of experience and type of occupation.

³Total prior experience variable was calculated from 3 dichotomous variables of having prior naloxone training (Yes/No), ever witnessed an overdose (Yes/No), and ever used naloxone (Yes/No). A score of 0 was categorized as “no experience” and scores of 1–3 were categorized as “some experience.”

Table 4:

Change in Confidence to Teach Based on Experience Level and Type of Occupation

	N	Pre-Test Score		Post-Test Score		One-way ANOVA1		Change Score		One-way ANOVA Change2
		M	SD	M	SD	F	p	M	SD	F	p
Experience Level 3						−1.67	0.10			3.78	<.01
No Experience	49	2.65	1.00	4.56	0.48			1.90	0.98
Some Experience	59	3.54	1.01	4.71	0.47			1.17	1.01
Occupation Type						1.24	0.30			3.09	<.05
School System	9	2.46	1.22	4.47	0.49			2.53	0.93
Law Enforcement	23	3.40	1.09	4.58	0.57			1.18	1.23
Health Field	31	2.83	0.99	4.57	0.48			1.74	0.90
Human Services	39	3.45	1.02	4.76	0.40			1.32	0.99

¹One-way ANOVA was conducted to compare pre-and post-test confidence to teach scores on level of experience and type of occupation.

²One-way ANOVA was conducted to compare change in confidence to teach scores on level of experience and type of occupation.

³Total prior experience variable was calculated from 3 dichotomous variables of having prior naloxone training (Yes/No), ever witnessed an overdose (Yes/No), and ever used naloxone (Yes/No). A score of 0 was categorized as “no experience” and scores of 1–3 were categorized as “some experience.”

Six Month Post Follow-Up Survey

In the six month follow-up survey, 22 participants out of 54 surveyed responded regarding their naloxone kit distribution and training behaviors. From these 22 participants, a total of 243 new participants were trained and 137 naloxone kits distributed. The 22 participants represented 18 out of 36 different agencies, resulting in a 50% response rate of agencies who participated in the train-the-trainer model. Out of the 22 participants who responded, 14 (63.6%) train-the-trainer participants trained others. The end trainees were staff from the agencies or community laypeople. Seventy-seven naloxone kits were distributed to community laypeople, 43 kits to law enforcement officers, and 19 kits to agencies such as public health departments and social service agencies to have on site. For participants that did not train others, barriers to training included no training needs, time constraints, and multiple trainers in their agency. Other barriers included fears about public speaking, logistical issues in coordinating a training, and the use of another agency to deliver the training. Facilitators to training included access to additional naloxone kits for distribution, sample naloxone kits for teaching, condensed training that is 30 minutes or less, and having an outside staff conduct training. Suggestions for increased naloxone distribution included advertising on Facebook, increased access to naloxone kits, increased distribution to health centers, coalitions, and community events, and improved identification of individuals who would need the naloxone kits.

DISCUSSION

Opioid-involved overdose deaths continue to rise in the U.S. driven primarily by synthetic opioids such as fentanyl.^1,34 As a result, the need to evaluate and increase naloxone training has never been greater. In fact, a recent review³⁵ recommended more studies on the best way to feasibly and effectively educate community laypersons on the use of naloxone and with the recommendation to standardize the core elements of overdose prevention curriculum.

Our study addresses a gap in knowledge by describing the effectiveness of a train-the-trainer curriculum. We demonstrated that the train-the-trainer naloxone curriculum was effective in increasing the knowledge and participants’ confidence to teach others about overdose prevention. Similar studies by Madah-Amiri et al and others^19,36–38 showed significant improvements in overdose knowledge. Although Madah-Amiri et al assessed comfort in teaching others, they only assessed the variable confident to train others at post-test. Our study extends Madah-Amiri et al’s study by demonstrating that participants’ confidence to teach others significantly improved at post-test, especially among those who had no prior experience with naloxone and overdose prevention. Moreover, our study showed that the participants’ teaching confidence increased after completing the train-the-trainer naloxone curriculum.

To date, few studies have examined the association between overdose knowledge and confidence based on either the demographic characteristics of participants or their prior experiences with overdose and naloxone. A study by Heavey et al, reported that those who had witnessed an overdose had greater overall overdose knowledge and more favorable attitudes than those who had never witnessed an overdose.³⁹ Our study showed that participants who had no prior experience with overdose benefitted most in their knowledge and confidence from the training. However, as evidenced by similar post-training test scores from participants who had no prior experience versus some experience, findings suggest that our train-the-trainer curriculum is beneficial for participants at any level of experience. Future research should examine if people with prior experiences may benefit from shorter training and which specific experience would better predict participants’ knowledge and confidence to teach.

Our results indicate the changes in overdose knowledge scores did not depend on occupation. We believe this supports the generalizability of our train-the-trainer naloxone curriculum and that the effect of being confident to teach others (p = .03) as a result of training is accurately observed despite the limitation in the sample size. One of the goals of our train-the-trainer naloxone curriculum is to increase participants’ confidence to teach others. Our results showed that participants who work in the school system had the greatest change in confidence to teach others. In comparison to law enforcement that would have more experiences with overdoses, the fact that participants in the school system who may be least likely to have prior experiences are now confident to teach others is another indicator of generalizability of this training. This is important as several states have laws allowing naloxone use in schools in the event of student or staff overdose.^40,41

Despite our positive findings, some study limitations exist. The majority of the trainees were White, female, and worked in the Midwestern region of the United States, thereby limiting the generalizability of our findings. However, our training included rural, urban, and suburban counties with high overdose death rates ranging from 11.2–32.8 per 100,000 residents. Rapidly expanding OEND programs to areas with high overdose fatality rates is important. Compared with communities with no OEND implementation, communities with low and high rates of OEND implementation were associated with lower rates of opioid overdose deaths.¹⁸ Other limitations to the generalizability of the study are that the participants may already have had a prior interest in overdose prevention education, there were small numbers of participants in the various occupation types, and the majority of the participants were White. Future large-scale studies can address these limitations by examining potential racial/ethnicity, occupation type, and experiential differences in training outcomes on a larger, more diverse population sample. Large-scale studies can also establish consensus for standard OEND guidelines and assist in tailoring the content to specific groups.

We did not assess fidelity in the delivery of the curriculum, nor did we collect pre- and post-training test outcomes for the new trainees at six months, including whether they could reverse an overdose. Future studies should assess the fidelity and different modes of train-the-trainer naloxone curriculum for feasibility and effectiveness. For instance, web-based format of overdose education has the capacity for greater dissemination,⁴² may be preferred among participants,⁴³ and have similar educational outcomes as in-person training.⁴⁴ Another limitation is the response bias that could have occurred in the follow-up survey. Participants who trained others may have been more likely to respond than those that did not train others. Thus, we do not know responses of barriers and facilitators of training and if further training occurred from the non-respondents.

Although we collected six month post-training outcome data, our participation rate was low. However, from the 22 participants that did respond, it represented 50% of the agencies that participated in the train-the-trainer training. Lastly, we did not assess the level of knowledge retained by the train-the-trainer participants after six months. In the future, it will be important to address these limitations using prospective designs that assess implementation as well as the short and long-term outcomes.

This study demonstrates the effectiveness of a train-the-trainer naloxone curriculum in high-risk counties, and provides a model to increase training, distribution, and accessibility of naloxone, ultimately leading to more lives saved. Not only did all participants improve their knowledge and confidence to teach others about overdose and naloxone, more people were trained (N = 243) by participants who participated in the train-the-trainer curriculum. This indicates the exponential potential of a train-the-trainer model of OEND to respond to the growing opioid overdose epidemic.