The prevalence of overdose deaths in the United States has continued to increase in recent years, eclipsing 100,000 such deaths (year over year) for the first time in June 2021 [1]. The majority of such deaths involve opioids; more specifically, in the current wave of the overdose epidemic, they involve heroin, fentanyl, and its analogues—the latter 2 of which are often introduced as adulterants in other substances (eg, stimulants) [1,2]. Largely related to such deaths, the United States renewed a declaration of an opioid public health emergency in April 2022 [3]. In the state of Indiana, where the proposed study will take place, provisional data indicated 2554 fatal drug overdoses in 2021, of which more than 78% involved opioids [4].

Naloxone, an opioid antagonist, has become a key component of the US response to the crisis [5], as it can reverse the effects of an overdose and prevent death [6,7]. However, the effectiveness of naloxone depends on the time it is administered relative to the time of the overdose. Deaths from heroin overdose can occur relatively quickly [8], and even when they do not, they typically happen within 1 to 3 hours of overdose [9]. Deaths from fentanyl overdose may occur even more rapidly [10], potentially due to differences in pharmacology between fentanyl and other opioids [11,12]. Unfortunately, individuals who witness an overdose may be hesitant to contact first responders [13,14], and even when they do, there is still variable lag time between the call for assistance and the arrival of assistance due to a variety of factors (eg, distance from the dispatch site to the overdose location). In some cases involving fentanyl, death may occur before first responders arrive [10]. Thus, there is an ongoing need to better ensure that naloxone is readily available to prevent overdose deaths.

In 2018, the US Surgeon General recommended that all people in the United States be prepared by having naloxone on hand to help in case of an overdose [15]. That recommendation followed, and has been followed by, evidence that overdose education and layperson and/or bystander administration of naloxone are safe and effective approaches to reducing the likelihood of opioid overdose fatality [16-19]. Similar to many states, Indiana has adopted laws to facilitate community overdose reduction using naloxone [20], including “good Samaritan” laws [21]. Although the means by which people obtain naloxone and other information related to harm reduction remain diverse, our focus in this proposed study is on community pharmacies.

Community pharmacies are well positioned to provide harm reduction services to all community members, along with more targeted services to individuals at risk of overdose. Almost 9 in 10 Americans live within 5 miles of a community pharmacy [22], meaning pharmacists are among the most accessible health care professionals [23]. Among some populations (eg, Medicare beneficiaries), it has been shown that pharmacists interact with patients more often than their primary care physicians [24].

In addition to their accessibility, pharmacies increasingly are perceived as effective venues for both clinical and nonclinical harm reduction owing to pharmacists’ expertise in fields such as medication management and pharmaceutical education [25,26]. A study of Arizona pharmacists found that even when uncomfortable with most forms of harm reduction, pharmacists were generally comfortable with dispensing naloxone [27]. In Indiana, qualitative data from managing pharmacists found conceptual support for harm reduction but, simultaneously, concerns about systemic barriers to implementation, including a lack of time for interventions as well as concerns about role clarity and patients’ expectations of what should happen in a pharmacy [28]. Similar patterns of support and concern were raised in a more recent study of key informants (including pharmacists) in Connecticut, Kentucky, and Wisconsin, reflecting willingness to discuss naloxone with patients but barriers related to role clarity as well as the financial cost of naloxone [29].

Thus, despite comfort and willingness to support harm reduction and treatment, as well as faciliatory state laws (eg, standing orders permitting naloxone dispensing without a patient-specific prescription), a recent systematic review found that naloxone was stocked on average in 62.8% of community pharmacies, with both the likelihood of stocking and willingness to dispense without a prescription significantly lower in independent pharmacies than in chain pharmacies [30]. That finding was echoed by a contemporary analysis of naloxone nasal spray stocking in 11 US states, which reported a similar rate (69.5%) and a significantly lower likelihood of stocking among independent pharmacies [31].

The proposed study will be a pilot randomized controlled assessment of the effectiveness of PharmNet, a harm reduction intervention for community pharmacies focused on naloxone distribution, awareness building, and referral. As described in the write-up of our single-site pilot study, “the intention was to study procedures that have as minimal an impact as possible on pharmacy costs and operational functioning while maximally facilitating harm reduction from opioid overdose—in other words, to find an optimal intersection point of those concerns” [32]. In brief, the intervention includes [32]:

Our previous single-site pilot study found that during the intervention, compared to the mean of the 3 months of practice prior to the intervention, the monthly rate of naloxone sold had a relative increase of 34.4% (+3.33 doses/month), and the overall naloxone dispensing rate through any mechanism (eg, including newly added no-cost dispensing) increased 96.48% (+9.33 doses/month) [32].

Although the intervention is conceptually simple, the decisions made in its development involved highly intentional design. Changes to the program were rooted in the Consolidated Framework for Implementation Research because it allows for considerations to be addressed within complex settings, such as pharmacies [33]. The PharmNet team conducted numerous information gathering, feasibility, and methodological studies in preparation for the single-site pilot to obtain increasingly specific information and context, including both quantitative and qualitative work. Much of this work was conducted in conjunction with academic and community-based pharmacists and involved data collection from hundreds of managing pharmacists in 2 states [27,28,34-42]. The single-site pilot itself was informed by independent community pharmacists at a more granular level to maximize likely acceptability and feasibility [32].

Our study will be designed to facilitate an improved understanding of whether it is reasonable to believe that the PharmNet intervention causes increased dispensing of naloxone by randomizing independent pharmacies that are part of the same small pharmacy chain to either (1) the PharmNet intervention or (2) operations as usual (control). It will also explore multiple secondary outcomes and questions.

This study will comply with all ethical regulations as outlined and approved by the Indiana University institutional review board (IRB; 12339 and 13956). The pharmacy employee data collection for this study was approved with a designation of “exempt” from the Indiana University IRB (13956).

Pharmacy employees who choose to participate in the staff-specific components of the study will complete an electronic consent form prior to providing any data. The completion of the pretest survey will be compensated by a US $5 digital gift card, and the completion of the posttest survey will be remunerated by a US $10 digital gift card. All other components of PharmNet have been deemed an organizational level research project and received a not human subjects research designation from the Indiana University IRB (12339). Participating pharmacy sites will each be paid US $1000 following the successful completion of this pilot study based on estimated cost data provided to author JA by the parent organization following the single-site pilot study.

Pharmacy employees’ responses will be linked to their provided email addresses in a survey platform (QualtricsXM) that is approved by Indiana University for “critical data,” the highest security designation. Survey responses will not be shared outside of the research team, and data will only be reported and analyzed in aggregate.

No other identifiable data will be obtained by the research team (eg, naloxone-dispensing data will be obtained only in aggregate for the prespecified time periods).

As noted above, we will conduct this study among 7 independent community pharmacies, with a 4:3 allocation to intervention and control arms. With 80% power (α=.05, 2-tailed), this sample will allow us to detect large effect sizes (Cohen d=2.68). In our single-site pilot study, we observed a large increase in the average number of monthly naloxone doses dispensed (from 9.67 doses/month prior to the intervention to 19 doses/month during the intervention). We further do not expect that the control pharmacies will substantially vary in the number of naloxone doses dispensed over time.

We made the decision to conduct this smaller, pilot randomized controlled trial, because although our single-site pilot study produced a sizeable effect on naloxone dispensing, we could not be certain that it was not attributable in part to unmeasured confounding (eg, external trends in naloxone dispensing). In addition, recruiting and providing materials (especially naloxone) to a large sample of pharmacies represents a substantial cost, and so, prior to investing in a more conservative, fully powered, and pragmatic study, we wanted to obtain additional data. Thus, we plan to report all P values exactly rather than by threshold and undertake a nuanced interpretation of the findings [45], including a visual inspection and presentation of the trends in intervention and control pharmacies, with the purpose of determining whether a large, randomized study appears to be an appropriate and valuable investment.

Since the pharmacy manager has agreed in principle to participation among all 7 individual pharmacies, we do not intend to apply any inclusion or exclusion criteria but note that individual-level data collection will still be subject to exclusion via declining to participate at the point of informed consent.

This study is designed to be a pragmatic, pilot randomized controlled trial of the PharmNet intervention. We plan to move forward with preparations for a much larger trial if either 1 of the following 2 cases are true: (1) Hypothesis 1 is upheld and there is evidence of a strong, significant effect of the intervention compared to control; or (2) Hypothesis 1 is not upheld (eg, the primary effect is nonsignificant), but the visual and numeric inspection of differences suggests a reasonable inference that favorable differences might be observed given a fully powered study.

The proposed study is a pragmatic pilot trial, which imposes some limitations on the types of data and nature of information that can be collected, as our goal is to demonstrate real-world effectiveness. In addition, we expect that limitations may be identified during the course of study completion, and they will be identified and described in the results-oriented manuscript to follow the experimental intervention.