Impacts of 2024 Western Pennsylvania Floods on Opioid Use Treatment Clinics

Publication Date: 2026

Abstract

Click to View the Research Brief

This study investigated how opioid use disorder (OUD) treatment clinics in Pittsburgh, PA, responded to severe flooding in April 2024. Through 13 qualitative interviews with clinic staff and service providers we explored how flooding affected patients’ access to care and medication. We found that flooding was significant but the disruption to OUD care was minimal. This resilience of OUD clinics was largely attributable to the rapid and widespread adoption of remote care modalities established during the COVID-19 pandemic. Respondents described a smooth pivot to virtual services and alternative medication delivery methods during flood recovery, which helped circumvent challenges posed by transit disruptions and facility damage. However, the benefits of remote care were unevenly distributed. While remote care increased access for some—such as mothers with small children—it also introduced new inequities. People lacking digital literacy, stable internet, or private space at home often struggled with virtual care. Moreover, the loss of in-person interaction raised concerns about the diminished relational components of care. Moreover, interviewees emphasized the equal importance of cultural identity in disaster response. Many cited a strong culture of resourcefulness, flexibility, and a deep commitment to mutual aid among patients and providers alike as a key source of hazard resilience in the OUD recovery community. Our study demonstrates that remote care modalities should be viewed not just as conveniences but as critical infrastructures that can buffer vulnerabilities and advance behavioral health resilience before, during, and after disasters. However, digital equity must be prioritized to ensure inclusive preparedness. These insights carry implications for emergency planning, public health policy, and the design of equitable care systems in hazard-prone settings.

Introduction

In April 2024, southwestern Pennsylvania experienced two heavy precipitation events (April 2-3 and April 11-12) that led to significant flooding throughout Pittsburgh and surrounding areas. The second flooding event resulted in the U.S. Small Business Administration issuing a disaster declaration for Allegheny County, where Pittsburgh is located, and the surrounding counties of Armstrong, Beaver, Butler, Washington, and Westmoreland (U.S. Small Business Administration, 2024¹). A federal disaster declaration was not issued for either event, meaning fewer external resources were available to assist impacted communities with recovery.

Flooding is one of the most frequently occurring and damaging natural hazards in southwestern Pennsylvania (Southwestern Pennsylvania Commission, n.d.²), and climate change is expected to increase the frequency of flood events (Pennsylvania Department of Environmental Protection [DEP], 2025³; Sharma et al., 2021⁴). Pittsburgh—located in the heart of the American Rust Belt—is a city increasingly affected by flooding and profoundly stricken by the opioid epidemic. In the substance abuse treatment community, there exists an understanding that people who use or are in recovery are more vulnerable to the negative effects of hazard events, such as floods. However, there is limited knowledge about who is most affected within this group, how negative impacts are geographically distributed, and what measures could be put in place to protect these populations (El Ibrahimi et al., 2023⁵), especially for flooding events that are not declared federal disasters. To fill this gap, this research study examined how the flooding events in April 2024 affected people in treatment for opioid use disorder (OUD). More specifically, this study sought to understand how small-scale flooding—that is, a disaster that is not severe enough to justify a federal disaster declaration—impacted people in the opioid recovery community.

Through interviews with staff at OUD treatment clinics across the greater Pittsburgh area, we sought to understand how this small-scale disaster disrupted OUD care and exacerbated vulnerability in the OUD community, particularly the uniquely vulnerable subgroup who relies on medication-assisted therapy regularly as part of their treatment protocol. By identifying the issues that this vulnerable population experienced during the aftermath of flooding, and how the event affected their access to OUD recovery services, this study has advanced knowledge of the impacts of natural hazards on people in OUD recovery and provided insights on how to better support this community during disasters.

Literature Review

People in substance abuse recovery are particularly vulnerable to disasters due to increased risk of treatment disruption and difficulty managing stress and mental health (Olayinka & Alemu, 2023⁶). Several studies report that in times of disaster, patients in recovery experience reduced access to treatment (Eisenberg et al., 2021⁷; Matusow et al., 2018⁸; Tofighi et al., 2014⁹). Additionally, increased stress and trauma from a disaster can result in a decline in mental health, potentially resulting in people with prior substance abuse issues relapsing or those without that prior history turning to substance use to cope with stress (Cepeda et al., 2010¹⁰; North & Pfefferbaum, 2013¹¹; Vetter et al., 2008¹²; Waters & Copeland, 2021¹³).

Among those in recovery, patients who must regularly travel to treatment facilities to receive medical attention or medication are uniquely vulnerable during hazard events. Some patients in treatment for OUD travel on a daily or near daily basis to treatment clinics to access medications, including drugs such as methadone, buprenorphine, and naltrexone, which work to block the euphoric effects of opioids as well as relieve the symptoms of opioid withdrawal (Frank et al., 2021¹⁴). Some of these medications are classified by the U.S. Food and Drug Administration as controlled substances, meaning there are additional requirements both prescribers and pharmacies must meet to make this medication available to patients (Pennsylvania Department of Health, n.d.-b¹⁵). In order to prevent abuse of medications for OUD treatment, patients are typically required to be present in person to receive a limited supply (e.g., 1 to 3 days in the case of methadone) of their medication. Additionally, some forms of medications for OUD must be taken as subcutaneous injections administered by a doctor (U.S. Food and Drug Administration, 2024¹⁶). During and after disasters, researchers find that patient access to medication-assisted treatment is often disrupted, resulting in patients engaging in potentially risky behaviors (e.g., adjusting their dosage to “stretch” the medication or procuring additional doses from non-approved sources) to maintain access to their medication (Carlisle Maxwell et al., 2009¹⁷; Caruana, 2024¹⁸; Griffin et al., 2018¹⁹; Matusow et al., 2018; Tofighi et al., 2014). Recommendations for improving post-disaster access include developing take-home dosing plans for eligible patients and improving relationships with neighboring clinics where patients could be transferred if their home clinic is damaged or otherwise not operational post-event (Caruana, 2024; Griffin et al., 2018; Matusow et al., 2018).

Few existing studies consider how disaster preparedness recommendations for OUD treatment clinics and patients could be implemented for small-scale disasters with short lead time. Much of the research OUD in disaster contexts in the United States focuses on hurricane events, specifically Katrina (2005), Rita (2005), Sandy (2012), and Harvey (2017) (Carlisle Maxwell et al., 2009; Cepeda et al., 2010; Griffin et al., 2018; Karaye et al., 2022²⁰; Matusow et al., 2018; McCann-Pineo et al., 2021²¹). The National Hurricane Center issued accurate landfall forecasts 2-3 days in advance of each of these hurricanes (Willoughby et al., 2007²²), giving treatment clinics in their paths time to prepare through such actions as dispensing take-home OUD medication doses (allowing patients to stay home or evacuate to a different location without risk of missing doses) and/or taking mitigation actions to safeguard the physical clinic facility from storm damage. Preparedness and response needs for these groups may be very different for hazards with limited forecast lead time, such as flash flooding or wildfires.

Much of the existing literature examining the intersection of OUD recovery and disasters is limited to events that have received a federal disaster declaration, or an international equivalent. There are few studies that examine smaller scale hazard events that did not cause enough damage to merit this designation. Smaller events can be as destructive as federally declared disasters to individuals directly impacted, but there are typically fewer response and recovery resources made available to them.

Research Questions

We selected the April 2024 floods in Pittsburgh for this study because they could help fill this research gap; they were small enough in scale that they did not cause enough damage to merit a federal disaster declaration and the public had a much shorter forecast lead time. Climate change models also predict that these types of events will become more prevalent, which can potentially place OUD recovery communities at greater risk in the future. Our research questions were as follows:

• How did flooding impact the OUD recovery community’s access to treatment?
• What structural vulnerabilities – referring to the risks built into systems and places – affected the OUD recovery community?

Research Design

At the time this report was prepared in August 2025, we had interviewed 13 opioid treatment clinic workers for this qualitative study. The interviews provided information on how clinics responded to a flood event and other hazards. We used a semi-structured interview format which allowed us to probe new insights volunteered by participants during interviews, as well as to capture unexpected information not included in the original interview guide. In this report, we highlight preliminary findings from completed interviews with clinic workers employed at medication-dispensing opioid treatment programs. Our findings explore clinic professionals’ perceptions of the impacts of the April flooding event with a focus on their perceived change in community treatment needs.

Study Site and Access

We selected the greater Pittsburgh area, specifically Allegheny (urban) and Washington (rural) counties, because the region is severely impacted by the opioid epidemic Decades of economic decline in the U.S. Rust Belt, which involved the loss of jobs in manufacturing and steel production since the 1970s, produced economic despair in the region. The blue-collar workforce in Pittsburgh has been further diminished by the more recent, uneven transition to a predominately white-collar and service-oriented workforce (Vitale, 2015²³). This extended period of economic fragility and precariousness created the conditions for the opioid epidemic to gain a strong foothold in the region. In their hazard mitigation plans, the emergency managers for Allegheny and Washington Counties—the two counties included in this study—identified opioid addiction as a major hazard affecting their county (Allegheny County, 2020, p. 64 ²⁴; Washington County, 2021, p. 40²⁵). Opioids have caused severe mortality in Pennsylvania, with an average of one Pennsylvanian dying every 2 hours from a drug overdose—82% of which involved opioids (Pennsylvania Department of Health, n.d.-a²⁶). Western Pennsylvania, especially the counties of Allegheny, Westmoreland, and Washington, are deeply affected, with the majority of overdose deaths across western Pennsylvania occurring in these three counties (Miller et al., 2016²⁷).

Regarding flood risk, a report from the Pennsylvania Department of Environmental Protection forecasts that southwestern Pennsylvania will experience an increase in heavy precipitation events, like the ones that occurred in April 2024, due to climate change over the course of the 21st century (DEP, 2025). According to the National Weather Service (NWS), the total rainfall in April 2024 (7.93 in.) almost broke the monthly record of 8.11 in. set in 1901 (National Weather Service, n.d.²⁸). Most of that rain fell during the April 2-3 and April 11-12 floods. According to data from the Cooperative Observer Program, there is already a verified trend of increasing precipitation in the area (DEP, 2025). Southwestern Pennsylvania will also experience an increase in heavy precipitation events in the coming decades (DEP, 2025), the end result being more floods of all magnitudes in the region’s future.

Semi-Structured Interviews

Sampling Strategy and Participant Recruitment

This study built on existing research relationships with the opioid treatment community to help recruit interview participants. Additionally, we identified potential interviewees through two online databases: (a) the list of facilities designated by Pennsylvania as Centers of Excellence for providing both treatment and wrap-around services for OUD (Commonwealth of Pennsylvania, n.d.²⁹) and (b) an online treatment provider database of self-selected providers (Shatterproof, n.d.³⁰). These two sources provided contact information for 63 facilities in Allegheny County and 15 facilities in Washington County. Since our goal was to cast a wide net during recruitment, we also conducted snowball sampling to connect with additional participants. Recommendations made through snowball sampling added 27 additional facilities in Allegheny and three in Washington county to our list of potential interviewees. We completed semi-structured interviews with 13 staff members. All interviews were conducted between February and May 2025.

In total, we contacted 108 facilities across both counties. We made a considerable effort to recruit members by calling each facility an average of two times (approximately 216 total phone calls) and following up with an average of two emails (approximately 216 total emails). If a facility indicated that they did not wish to participate in our study, we removed them from the potential contact list.

In total, 13 staff members agreed to be interviewed. The response rate was relatively low—at about 12%—likely because substance use facilities have limited resources. It was challenging to reach staff members who had the time or training to respond to our invitation. We were also informed by these facilities that our study fell during a particularly difficult time for substance use services because of federal budget cuts. The people that we interviewed included directors and staff of opioid treatment centers, counselors, and nurse practitioners.

Interview Guide and Procedures

Interview guide questions asked treatment facility staff to describe how the April 2024 flood impacted clinic operations, including both the facility itself and their patients’ ability to access treatment and services. For inpatient facilities, staff had intimate knowledge of the experiences of their patients because they were housed in the facility during the flood. For outpatient facilities, staff learned about patient experiences when patients came for counseling sessions or to receive medication. We also asked about general preparedness for weather hazards. After conducting the first few interviews, we added additional questions related to other hazards experiences, including the COVID-19 pandemic, as early participants had much to share on this unexpected topic. We included core questions in the interview guide, with the semi-structured guide design allowing participants the flexibility to provide additional relevant information that we in the research team were not expecting. All interviews were conducted via Zoom, which was preferred by all participants over in-person interviews, and recorded with participant permission, then transcribed and anonymized. Interviews typically lasted 40-60 minutes.

Data Analysis Procedures

Interviews from treatment facility staff were transcribed with Otter.ai and then edited for accuracy. We manually inductively coded finalized transcripts using NVivo software. First, an initial read through of the transcripts by the first and second authors identified a preliminary list of themes. Then, the second author conducted a second read through of the transcripts to condense the initial themes into a more concise set of codes. These were recorded in a code book, with accompanying definitions to ensure consistency during the process of coding. We tested the codes with two interviews and then edited the code book according to any discrepancies or oversights. The process of coding transcripts is ongoing, as some interviews were conducted just recently, but the majority have been completed. The codes inform the subsections in the findings of this report, and an expanded version of these codes will inform a manuscript targeted for publication in an academic journal.

Ethical Considerations

All study procedures and interview guides were approved by the Institutional Review Board (IRB) of the Pennsylvania State University (#25822) and the study received exempt status on October 25, 2024. Before conducting an interview, we asked staff members to give verbal consent to participate in accordance with our IRB protocol. We chose not to interview patients in OUD treatment, as no one on the research team has a clinical background in trauma-informed care to help patients if we inadvertently hit a trauma trigger during the interview.

Findings

Effects of April Floods

Our respondents described in detail the effects of April floods on the operations of OUD recovery clinics, which can be characterized as disruptive but not catastrophic. While some respondents initially said operations were not affected at all, upon further probing most said disruptions were not noticeable because they were immediately quelled by switching to remote care modalities. However, one clinic was flooded, and its outpatient unit had to be temporarily relocated.

The relatively mild impacts of flooding, while surprising considering the substantial scope of the event, are empirically significant. Our respondents’ accounts pointed us to an unexpected finding: flooding impacts were largely mitigated by the availability of remote services which became common during COVID-19. This helped equalize patients’ access to care. However, there were still variations in access to remote services, which our research highlights.

COVID-19 Era Remote Care Modalities Increase Resiliency but Introduces New Disparities

The post-COVID landscape is characterized by a greater resilience to certain types of hazards. We did not expect to hear how smoothly the community responded to flooding, however, we learned that remote care modalities established during COVID-19 contributed, in large part, to their adaptability and resilience. In other words, clinics were prepared to deliver services, and in some cases drugs, remotely, which made all the difference when the floods hampered transit and other services. As one participant said, “We're going to switch you to Zoom. No problem. We'll take care of it.”

Remote care brought about new variations in access to treatment, especially as populations benefited differently. For instance, one respondent explained that mothers with small children benefitted substantially from remote care options, describing how it removed complications around transportation and childcare: “I mean, mothers, definitely like mothers of small children, telemedicine is exponentially easier for them. They typically are the full-time caretakers, the people staying at home with their children.” According to this respondent, remote care allowed her clients to attend their medical appointments without having to worry as much about childcare.

Remote care also brought about new challenges, especially for people with unstable access to the internet or populations that have difficulty navigating the technical nuances that remote care introduces. For instance, providers were required to use secure platforms. While good clinical practice, these platforms could exclude some patients who have poor internet access or are unable to easily use technology. Furthermore, respondents emphasized that the recovery community tends to be composed of people who may not have consistent access to resources, such as phones. One respondent summarized this by saying: “I think when we're doing that telehealth, that applies more to the mental health population, our folks, typically, they may have a phone one day and not have a phone the next.” Other populations that struggled to maximize the benefits of remote care modalities such as telehealth included older generations unfamiliar with technology. Furthermore, respondents pointed out that the recovery community is specifically vulnerable to domestic abuse, making it hard for them to find safe and private spaces in their homes to meet with health care providers.

Another downside to remote care was the loss of face-to-face intimacy, which respondents said weighed heavily on people in recovery. Many people in this community rely on home visits where caregivers can physically see, hear, and smell how they were doing:

You don't see the things … how they're interacting with the other people in the home. You know, I even say, like, look how their pets are acting. You can catch more mannerisms or behaviors [during home visits] that you wouldn't necessarily see [online].

According to this respondent, there are benefits of visiting clients in person that simply cannot be recreated through remote care.

Recovery Culture and Identity

Staff Resilience

Another set of findings centers around the importance of recovery culture and identity in shaping the response of the OUD community to the April floods. Staff referenced a strong internal culture which centered on a passionate and altruistic commitment to healing.

Additionally,, staff in the recovery community are accustomed to managing daily crises and perceived the small-scale natural hazard as “manageable” by comparison. Indeed, staff said they considered adapting to the disaster as all in a day's work:

We’re pretty passionate, and we like to take care of our community, so we've always been like that. If the power goes out, we move to rooms where there's light, we don't close ...we just keep moving!

The respondent quoted above went on to explain that it’s not the kind of field where you can be flippant about serving your clientele in the first place, so “above and beyond” commitment is already in place:

We see our clients. If somebody's out sick, we see their clients. We don't... cancel appointments. This [other] person is going to see you, so that people don't miss their therapy... So, I think our ability to just adapt has always been how we run our programs.

There was also an emphasis on the culture of helping one another within and outside of the clinic, and how that inherently prepared people for hazard events. One clinic staff described how the culture of solidarity permeated into life at work and beyond.

[...] yeah, this looks like a pretty scary community to be in. But what we find is that because we are such a great resource for so many people in this community, we find that they are more protective of us.

This informant described how resilience comes both from OUD staff members and the patients themselves, who all work together as a community.

Patient Resilience

Respondents described an inherent resilience among patients in OUD recovery which has been overlooked by other studies focusing on their challenges and vulnerability. That resilience was visible in the flood context, where patients struggled, but not catastrophically. For instance, clinic workers described how patients rose up to help and didn’t see themselves as victims: “Crisis clients, they like, wanted to help. So, it kind of turned their instincts on, I think, to kind of pull together. It wasn't so much they were panicked...but they were just really helpful in accommodating.” For this respondent, the patients switched roles. Instead of only receiving help from the staff, the patients helped the staff members get the clinic back into working order. They assisted with cleanup, adjusted their schedules, and helped one another replace lost belongings. In addition, respondents emphasized that even though clients lacked some formal knowledge about how to protect themselves against unexpected hazards, they were adept at responding creatively to hard times.

Conclusions

Implications for Practice or Policy

The findings of this study have several significant implications for disaster response and OUD treatment. First, they highlight the critical role remote care modalities can play in bolstering system resilience during extreme weather events. The ability of clinics to rapidly transition to remote service delivery mitigated severe disruptions in care. This suggests that maintaining and expanding remote care infrastructure—even outside of crisis periods—is not only a matter of convenience but a vital preparedness strategy. Policymakers and public health practitioners should consider integrating or enhancing remote care capacities into emergency preparedness protocols for behavioral health services.

However, this transition also underscored inequities in access to remote care. Not all patients benefited equally. Those without stable internet, digital literacy, or safe private spaces were at a disadvantage. In this light, our findings call for targeted investments in digital equity: providing devices, broadband, and technical support to marginalized populations. Moreover, regulatory frameworks should remain flexible to allow the continued use of secure but accessible platforms while accounting for the practical needs of vulnerable clients. While the study was conducted in Pittsburgh, these findings have broad relevance. Policymakers can learn from how prior exposure to crises like COVID-19 helped structure resilient responses to new hazards, reinforcing the importance of institutional memory and adaptive capacity in health care delivery systems.

Limitations

This study’s qualitative design limits the generalizability of its findings. We relied on in-depth interviews with a select group of practitioners and did not interview patients directly. As a result, the analysis reflects professional interpretations of patient experiences rather than firsthand accounts. Furthermore, because the study focused on a single city and hazard event, it may not fully capture how OUD recovery communities in different geographical or institutional contexts respond to similar challenges. While we identified disparities in remote care access, we were not able to quantify the scope or impact of these disparities.

The changing federal funding climate also significantly limited our access to research participants. During the course of our research, several clinics that were interested in participating lost funding and closed. This also made it more challenging to get access to official planning documents. We initially intended to interview a second cohort of participants made up of emergency medical services (EMS) staff and firefighters who respond to overdose calls. However, we found that this cohort was reluctant to commit to interviews, possibly due to more immediate concerns around budgetary constraints. We also found that, during these stressful times, relationship building was particularly important. Potential informants, who are already overwhelmed by changes in their work environment, were even less likely to respond to cold calls. We had to rely heavily on one-off references and referrals, which was slow going.

Future Research Directions

In addition to addressing the proposed research objectives, this project revealed several areas that require further examination for both research and practice. First, in the post-COVID landscape, the use and access to remote care as a mechanism for treatment expanded. Our research demonstrates that the type of hazard matters, as shifting to remote care or shifting the sites for medication delivery mitigated issues of flooding in and around the opioid treatment centers. Additionally, our informants emphasized the place-based dynamics, such as social networks and community identity, contribute to resiliency in the health care sector. Taken together, our results highlight the importance of analyzing compounding health challenges (opioid treatment and flooding) through the integration of methods that reveal the role of local context in shaping responses.

Second, our study identifies the need to interrogate different types of natural hazards in producing differential vulnerability to opioid treatment. During the research period, the city of Pittsburgh experienced an unprecedented windstorm on April 29, 2025. This resulted in intense precipitation and damage to infrastructure and power lines that lasted for five days. Some of our more recent conversations with informants indicate that this event was more challenging for health care delivery. This points to an important research gap that considers the differential and compounding effects of hazardous events.

To address these research gaps, we intend to extend our research beyond the scope of this grant to continue gathering data, specifically from firefighters and EMS staff, during the summer of 2025. In June, an email invitation was sent by an administrative assistant to 250 staff members of the City of Pittsburgh Bureau of EMS. We aim to recruit 20 EMS staff members to participate in individual interviews. Additionally, we are in the process of organizing focus groups with EMS professionals in Washington County, PA. We are also working with a community partner in Pittsburgh, UrbanKind, to share the study’s findings with the OUD community and translate our research to practice in ways that benefit the Pittsburgh community.

In future projects, this research could be scaled to address other regions or disaster types. This could help clarify whether the resilience observed in Pittsburgh is replicable elsewhere. In addition, researchers might explore how recovery identities and community bonds influence hazard response for smaller magnitude hazard events, potentially developing new frameworks that integrate cultural and social capital into resilience planning. Partnerships with practitioners could be expanded to co-design interventions that address the digital divide in remote care delivery.

References