https://orcid.org/0000-0002-8534-8860
Figures
Abstract
Background
The United States is experiencing a drug addiction and overdose crisis, made worse by the COVID-19 pandemic. Relative to other types of health services, addiction treatment and overdose prevention services are particularly vulnerable to disaster-related disruptions for multiple reasons including fragmentation from the general medical system and stigma, which may lead decisionmakers and providers to de-prioritize these services during disasters. In response to the COVID-19 pandemic, U.S. states implemented multiple policies designed to mitigate disruptions to addiction treatment and overdose prevention services, for example policies expanding access to addiction treatment delivered via telehealth and policies designed to support continuity of naloxone distribution programs. There is limited evidence on the effects of these policies on addiction treatment and overdose. This evidence is needed to inform state policy design in future disasters, as well as to inform decisions regarding whether to sustain these policies post-pandemic.
Methods
The overall study uses a concurrent-embedded design. Aims 1–2 use difference-in-differences analyses of large-scale observational databases to examine how state policies designed to mitigate the effects of the COVID-19 pandemic on health services delivery influenced addiction treatment delivery and overdose during the pandemic. Aim 3 uses a qualitative embedded multiple case study approach, in which we characterize local implementation of the state policies of interest; most public health disaster policies are enacted at the state level but implemented at the local level by healthcare systems and local public health authorities.
Discussion
Triangulation of results across methods will yield robust understanding of whether and how state disaster-response policies influenced drug addiction treatment and overdose during the COVID-19 pandemic. Results will inform policy enactment and implementation in future public health disasters. Results will also inform decisions about whether to sustain COVID-19 pandemic-related changes to policies governing delivery addiction and overdose prevention services long-term.
Citation: Eisenberg MD, McCourt A, Stuart EA, Rutkow L, Tormohlen KN, Fingerhood MI, et al. (2021) Studying how state health services delivery policies can mitigate the effects of disasters on drug addiction treatment and overdose: Protocol for a mixed-methods study. PLoS ONE 16(12): e0261115. https://doi.org/10.1371/journal.pone.0261115
Editor: Elisa Panada, UNITED KINGDOM
Received: November 24, 2021; Accepted: November 26, 2021; Published: December 16, 2021
Copyright: © 2021 Eisenberg et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: All authors received funding for this study from the National Institute on Drug Abuse (NIDA), grant number R01DA053232. https://www.drugabuse.gov/ Authors EEM and MDE serve as multi-principal investigators of this project and all other authors are co-investigators. The funders had and will not have a role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Disasters and the drug addiction and overdose epidemic are among the foremost public health issues facing the United States. Public health disasters, defined by the U.S. Centers for Disease Control and Prevention (CDC) as events that lead to a major disruption in health services [1], are predicted to increase in the coming decades due to climate change and other sources of geopolitical instability [2–4]. At the same time, the U.S. faces an ongoing drug addiction and overdose crisis. In 2019, one in five Americans aged 12 years or older used drugs non-medically; 8.3 million people met criteria for a substance use disorder involving drugs other than alcohol, and 70,630 died of a drug overdose [5, 6]. Provisional drug overdose death data for 2020 and 2021 show a sharp increase in overdose deaths spanning the COVID-19 pandemic [5], which has increased risk factors for drug use and overdose including psychological distress, poverty, unemployment, housing instability, and disruption of social networks [7–10].
Public health disasters pose major threats to addiction treatment delivery, and even brief treatment disruptions can prompt a return to drug use and attendant risk of overdose [11–15]. Relative to other types of healthcare services, addiction treatment is particularly vulnerable to disaster-related disruptions, for several reasons. People with drug addiction are disproportionately affected by socioeconomic barriers to treatment, such as poverty, unemployment, and housing instability [16–20], which may be exacerbated during disasters. In addition, drug addiction treatment is tightly regulated, which limits flexibility during disasters [7, 21–24]. For example, under federal law, methadone to treat opioid use disorder can only be dispensed in specialty clinics, which most patients must attend every day to receive their dose [25]. Addiction treatment is often disconnected from the general medical system, where critical on-the-ground disaster preparedness and response occurs; this fragmentation is a barrier to health system efforts to enhance addiction treatment access during disasters [26–29]. Finally, drug addiction treatment is stigmatized and under-resourced [30–36]. In disaster scenarios, which stretch health system resources and force choices about which services to prioritize, addiction treatment may be a low priority. All these factors also apply to overdose prevention services, such as distribution of naloxone or fentanyl testing strips by state and local public health departments.
Policies form the backbone of disaster response by delineating what health systems can and cannot do in the midst and aftermath of disasters. Most disaster policymaking occurs at the state level and is implemented at the local level by healthcare systems and local public health authorities [37–41]. The federal government’s role is primarily to allocate additional resources and provide authorization to make certain policy changes [40]. For example, in response to COVID-19, the federal government relaxed rules around in-person methadone dosing for opioid use disorder, allowing states to apply for a waiver to let clients take home a 14–28 days’ supply; some states opted to use that federal waiver, while others did not [42].
Several types of state health services delivery policies have the potential to enhance access to drug addiction treatment and prevent overdose during and after disasters, including state telehealth and Medicaid policies, state essential service designations, and policies explicitly targeting drug addiction treatment (Table 1). To date, however, these policies’ effects on drug addiction treatment and overdose during and following public health disasters have not been well studied. In addition, no prior studies have comprehensively considered strategies for effective implementation of state policies designed to mitigate disruptions to addiction treatment and overdose prevention services during public health disasters. The effectiveness of state policies depends upon local-level implementation because local healthcare systems and public health departments are the front-line of disaster response. Thus, it is critical to understand how local system-level policies and practices influence the implementation and effectiveness of the state-level policies of interest.
Materials and methods
Aim 1.
Study Aim 1 is to examine how state policies designed to mitigate the effects of the COVID-19 pandemic on health services delivery influenced receipt of addiction treatment. We will use a difference-in-differences analysis of insurance claims and addiction treatment program admissions data from the 50 U.S. states. We expect the policies of interest—all of which are designed to enhance health service access—to mitigate disruptions to addiction treatment during and following the COVID-19 public health disaster.
Aim 2.
Study Aim 2 is to examine how state health services delivery policies influenced the effects of the COVID-19 crisis on fatal and non-fatal overdose. We will use the same difference-in-differences approach in Aim 1 to analyze insurance claims and mortality data. We expect that the policies of interest will mitigate increases in overdose during and following the COVID-19 disaster.
Aim 3.
Study Aim 3 is to characterize local-level implementation of state health services delivery policies put in place in response to COVID-19 and to identify policy gaps and changes needed to enhance addiction treatment access and prevent overdose during and following public health disasters. In this Aim, we will conduct interviews with state and local policy implementation leaders and review relevant local government, healthcare system, and public health system policy documents. We will characterize local healthcare and public health system policies and practices designed to support access to addiction treatment and overdose prevention services during the COVID-19 disaster; consider how these local health system policies and practices may interact with state policies to influence on-the-ground implementation of policies enacted at the state level; and examine policy implementation leaders’ perceptions of policy and practice gaps and strategies to address them. We will also explore front-line policy implementers’ perceptions of how state health services delivery policies designed to mitigate disruptions to health services during the COVID-19 pandemic could generalize to other disasters and examine leaders’ perceptions of whether and how the state policies of interest, if sustained long-term, would enhance access to addiction treatment and prevent overdose post-COVID-19 pandemic.
Methods
Study design and sample
The study uses a concurrent-embedded design. Aims 1 and 2 quantitative analyses guide the research, and the secondary Aim 3 qualitative study plays a supportive role. Quantitative Aims 1–2 use a difference-in-differences design to analyze the effects of policies on outcomes using secondary data sources that include data on all 50 U.S. states. Qualitative study Aim 3 uses an embedded multiple case study approach, in which we characterize policy implementation in counties embedded within states. Aims 1–3 will be conducted concurrently.
Study period
The study period is 2015–2023, encompassing the five years prior to the COVID-19 pandemic and four years following onset on the pandemic in the U.S. Aims 1–2 quantitative analyses will include data for each of the nine years in the study period. The Aim 3 qualitative study will primarily focus on implementation of state health services delivery policies put in place in response to the COVID-19 pandemic. Aim 3 will also characterize local disaster preparedness policies that were in place pre-COVID and explore leaders’ perceptions of how those pre-pandemic policies influenced COVID-19 disaster response.
Data sources
State health services delivery policy data.
Data on the state health services delivery policies of interest in Table 1 will be assembled by the public health lawyers on the study team using legal research and legislative history techniques [43], including full-text searches of the Westlaw database and identification of state executive orders, session laws, regulatory materials, and sub-regulatory guidance. The data will include each policy’s provisions, effective date, and end date (if applicable). While many policies put in place in response to the COVID-19 disaster were designed to be temporary, as of November 2021, there are ongoing policy dialogues to consider making some of these policy changes permanent, as they have the potential to enhance service access long-term [44–47].
Administrative claims data.
Aims 1–2 will use IQVIA and OptumLabs® Data Warehouse (OLDW) administrative claims to measure addiction treatment (Aim 1) and overdose-related treatment utilization (Aim 2) outcomes. The IQVIA data captures 93% of all U.S. retail prescriptions, as well as outpatient services delivered by ≈75% of U.S. licensed physicians. Using a portal embedded in their billing software, pharmacies and outpatient clinics generate daily data that is automatically transmitted to IQVIA. A key strength of the IQVIA data is that it captures services from all payers, including services paid by any insurer (e.g., commercial insurance, Medicaid, Medicare) or by cash. The OLDW contains de-identified administrative claims data, including medical and pharmacy claims on commercial insurance enrollees and patients, representing a diverse mixture of ages, ethnicities and geographical regions across the United States [48]. Both data sources include claims for patients of all ages and have information on patient age, sex, state, 5-digit zip-code and medications, diagnoses, and procedures received, along with dates of receipt. The IQVIA and OLDW data are complementary in that the OLDW data includes inpatient and emergency department claims as well as a flag to identify commercial insurance beneficiaries in fully insured plans (as noted in Table 1, some policies of interest apply to this subset of beneficiaries), information not included in the IQVIA data.
Specialty addiction treatment program admissions data.
Aim 1 will use Treatment Episode Data Set (TEDS) admissions data to measure inpatient and ambulatory admissions to specialty addiction treatment facilities. The data capture two-thirds of specialty addiction treatment programs in the U.S., including programs in all 50 U.S. states, and contain records on those aged 12 or older. TEDS includes clinical characteristics (e.g., substances used, frequency of use) as well as information regarding the date and type of treatment admission (e.g., detoxification, intensive outpatient). Patient demographic information includes age, sex, race/ethnicity, state of residence, and core-based statistical area (CBSA) of admission. The TEDS data is complementary to the claims data sources described above: as addiction treatment is often delivered outside the general medical sector and paid for through federal and state grant programs, the specialty programs included in TEDS are underrepresented in the claims data.
Mortality data.
Aim 2 will use CDC multiple cause of death data to measure fatal overdoses. The data includes underlying cause of death, state, zip-code, and decedent demographics including age, sex, and race/ethnicity. Where the administrative claims data captures treated overdose episodes—many of which are non-fatal—the mortality data captures all fatal overdoses, including those that occur outside of the healthcare system.
Mobility data.
Measuring mobility is critical to disentangling the effects of the state health services delivery policies of interest from the effects of COVID-19-related disruptions to in-person health service delivery on outcomes. We will use cell phone tracking data to measure area-level mobility. This data comes from signals, or ‘pings,’ that identify the location of smartphones at a moment in time, and includes information about duration, origin, and destination of trips made by people with smartphones, including specific information on points of interest (e.g., healthcare settings) visited.
Area-level characteristics data.
Aims 1–2 will also use data on area-level characteristics from the U.S. Census Bureau data including rural/urban status, racial/ethnic, income, education, and home ownership distribution, data from the Area Health Resource File on health provider density, and data from the National Survey of Substance Abuse Treatment Services (N-SSATS) on specialty addiction treatment program density.
Qualitative data.
Aim 3 qualitative data will be collected through semi-structured interviews and document collection. In the states and counties included in the Aim 3 sample (see section below), we will recruit state-level, county-level, and healthcare and public health system-level leaders in addiction treatment, overdose prevention, and disaster preparedness and response. The interview guide will open with an overview of the study and an opportunity to ask questions, followed by innocuous and grand tour questions to establish rapport, and then by researcher-driven questions [49–55]. The researcher-driven portion of the interview guide will be structured in two sections: a state policy section and a local policy and practices section. In the state policy section, interview guide domains will include perceived importance and feasibility of the specific state health services delivery policies put in place in an interviewees’ state in response to the COVID-19 pandemic; policy implementation strategies, barriers, and facilitators; perceptions of policy gaps and needed changes; views on how policies used in COVID-19 generalize to other disasters; and thoughts on whether and how policies should be sustained long-term. The local policy section of the guide will be focused on identifying local healthcare and public health system policies and practices put in place to support health services delivery during COVID-19 and exploring leaders’ perceptions of how these local policies interact to support or impede implementation of the state-level policies of interest in Aims 1–2.
The semi-structured interview guide will be developed by the study team in close consultation with the study’s advisory board, which includes experts in addiction treatment, overdose prevention, healthcare administration, and health system disaster preparedness and response, as well as individuals with lived experience of drug addiction during the COVID-19 pandemic. Interviews will be conducted by a single master’s-level study team member trained in qualitative interviewing, via videoconference. Table 2 provides additional details regarding our qualitative research methods within the Consolidated Criteria for Reporting Qualitative Studies (COREQ) framework. We will ask interviewees to provide supplementary documentation regarding COVID-19 disaster response in their jurisdiction, e.g., state or local public health department disaster response memoranda or healthcare system emergency management plans. We will identify additional publicly available policy documents by searching local system websites.
Study sample
Aims 1–2 analytic samples will include data from all 50 U.S. states and Washington, D.C. Aims 1–2 claims data analyses will use continuous cohorts of patients of all ages diagnosed with a substance use disorder during the 2015–2023 study period, with estimated sample sizes of 24,000,000 individuals in the IQVIA data and 475,000 individuals in the OLDW data. The Aim 1 analysis of specialty drug treatment admissions will be conducted using all admissions among people aged 12 years or older included in the TEDS data from 2015–2023, with an estimated sample size of 8,000,000 admissions. The Aim 2 overdose analyses will be conducted using all overdose deaths in the OLDW claims (same estimated sample size as above) and CDC mortality data (estimated sample size 644,000 drug overdose deaths) from 2015–2023. As shown in Table 1, analyses of policies applying to specific subsets of individuals (e.g., Medicaid beneficiaries) will be limited to those groups. The Aim 3 study sample will include eight states: the two states with highest per-capita COVID-19 death rate in each of the four U.S. census regions at the start of the qualitative study. Within each of those states, we will conduct embedded case studies of two counties: the urban county with the largest population and the rural county with the largest population (16 counties total). We expect to interview approximately 115 total state and local-level policy implementation leaders. Interviews will be conducted until data saturation, defined as no new key themes emerging from the data, is reached.
Measures
Aims 1–2.
Final analytic measures will be constructed at the person-month (or admission-month in the TEDS data, which identifies admissions rather than individuals) level in Aim 1 and state-month level in Aim 2. Aims 1–2 independent variables are binary indicators of the state health services delivery policies of interest, which change from zero to one starting the first state-month a policy is enacted.
Aim 1 dependent variables include measures of receipt of any inpatient, ED, or outpatient drug addiction treatment service (claims data sources); any receipt of a prescription for a medication used to treat opioid use disorder (claims data sources); and any specialty drug addiction treatment admission (TEDS). In claims data analyses, among people who receive any of these services, we will also measure the number of services received. Aim 2 dependent variables include measures of fatal and non-fatal drug overdose rate per 100,000 population.
To measure area-level mobility in Aims 1–2, we will use the cell phone data to create measures of the ratio of the volume of travel in a person’s zip-code (CBSA in TEDS) in each month relative to the same month in 2019, pre-COVID-19 pandemic. We will construct separate measures of overall mobility and traffic to healthcare settings. In Aims 1–2 we will also construct zip-code/CBSA-level measures of urban/rural, racial/ethnic, income, education, and home ownership distribution, as well as state-level measures of physician and specialty addiction treatment program density. For Aim 2, where final analytic measures will be constructed at the state-month level, we will link zip-code/CBSA data in the granular overdose-level data and then ‘roll-up’ the data to the state-month level, e.g., the percent of people in a given state-month who lived in a zip-code with 25%, 50%, or 75% mobility relative to the same most recent pre-COVID state-month.
Aim 3.
Aim 3 qualitative interviews will yield key themes aligning with the interview guide domains described above. In addition, we will characterize local (e.g., county government, healthcare system, public health system) policies put in place to mitigate disruptions to health services generally or addiction services specifically in the eight states in the Aim 3 sample.
Analysis
Aims 1–2.
In Aims 1–2, we will use a difference-in-differences approach to compare trends in outcomes before and after the implementation of state health services delivery policies in states with versus without these policies. To illustrate the general model specification, the Aim 1 model for administrative claims data analyses—in which the unit of observation is person-months nested in zip-codes, which are nested in states—will take the following form:
where yist is one of the outcomes of interest, Policyst is an indicator variable equal to one if state s had the health service delivery policy of interest in effect during time t, Mobilityzt is the overall level of mobility in individual’s i’s zip code during time t, Xi,t-1 is a vector of individual-level pre-period covariates, Zzt is a vector of zip code-level covariates, Pst is a vector of state-level provider availability measures, Montht are a series of fixed effects for each calendar month, and States is a vector of state fixed effects.
To address the multilevel nature of the data, we will calculate standard errors via the delta method and employ two-way clustering at the zip-code-state level [56, 57]. Aim 1 TEDS data analyses and Aim 2 analyses follow a conceptually similar form. In Aim 1 analyses of the TEDS specialty addiction treatment program admission data, the unit of analysis is admission-month and rather than zip-code we use CBSA, the most granular geographic identifier in TEDS. In Aim 2, where the unit of analysis is state-month, the mobility measure and all area-level covariates will be aggregated to the state level, and we will cluster standard errors by state. In both Aims 1–2, area-level characteristic measures will be employed as either covariates, as depicted in the model specification above, or effect modifiers. For example, we will conduct effect modification analyses to examine whether the effects of the state health services delivery policies differ in urban versus rural areas.
We will also consider multiple conceptualizations of the overall and healthcare setting-specific mobility measures. Over the course of the COVID-19 pandemic, mobility has varied across states and localities, and within jurisdictions over time, due to spiking/waning of COVID-19 cases; in response to jurisdictions imposing, lifting, and reinstating physical distancing rules; and as virus fatigue has set in and led some people to increase their mobility irrespective of case rates or policy mandates. In a difference-in-differences framework, covariates are defined as variables that vary by treatment group and could cause variation in outcome trends over time. Given variation in mobility across states and the fact that mobility could influence our treatment and overdose outcomes of interest, mobility could operate as a covariate. We will also examine mobility as an effect modifier, e.g., by assessing whether policy effects differ in areas with low versus high mobility. In addition, we will explore the possibility that mobility to healthcare settings could be on the causal pathway between the state policies of interest and outcomes by examining the relationship between the policies and mobility during the pandemic as a dependent variable. For example, in a scenario where a healthcare system offers both in-person and telehealth visits, a state policy offering enhanced coverage of telehealth services could lead some people to choose telehealth, leading to a lower level of traffic to healthcare settings than would be observed if the policy were not in place.
Given that some states implemented multiple policies of interest at the same time, we will analyze the effects of individual policies as well as combined effects of grouped policies, e.g., at least one telehealth policy in Table 1, all telehealth policies, etc. Our ability to make causal inferences about specific policies/groups of policies is enhanced by the fact that different policies are expected to effect different populations and outcomes (Table 1). Recent work shows that in scenarios where policy adoption is staggered across geographic units, traditional difference-in-differences specifications with two-way state and time fixed-effects, like the model above, can produce biased results. While adoption of the state health services delivery policies of interest consistently occurred in March or early April 2020, there may be staggered policy phase-out across states. If this is the case, as identified through our legal mapping, we will analyze treatment heterogeneity following Goodman-Bacon 2021 [58] and, as necessary, employ alternative approaches to address two-way fixed-effects-related biases [59–62].
Aim 3.
After each interview, we will create summary memos identifying preliminary themes. These memos, along with the interview guide, will contribute to the development of a codebook. Using a randomly selected sub-sample of transcripts, two team members will pilot the codebook. It will then be further refined and analyzed with input from the advisory board. The final codebook will be applied to all interview transcripts. Text segments will be organized in QSR NVivo v11 and analyzed according to themes and sub-themes. We will compare themes across states and counties, for example to identify varying patterns in themes in urban versus rural counties. We will identify local healthcare and public health system policies of interest through combined analyses of interviews (when we will ask interviewees to identify policies) and policy documents. We will conduct member-checking with a random sample of interviewees who will review key themes.
Ethical considerations
Aims 1–2 of this study involve analysis of secondary data sets that meet limited data set criteria and do not include any individual identifying information such as name or medical record number. Aim 3 involves qualitative interviews with state and local health system leaders and focuses on topics related to their professional roles. The study was deemed exempt by the Johns Hopkins Bloomberg School of Public Health Institutional Review Board on May 11, 2021 (IRB #16400). Consent was waived for the Aims 1–2 limited data sources. Oral consent will be obtained from participants in qualitative interviews.
Status and timeline of the study
As of November 2021, participant recruitment and data analyses have not yet begun. These activities will begin in winter 2022. Aims 1–2 secondary data analysis will be conducted iteratively from 2022–2025 as the latter years of data included in the study become available. Aim 3 qualitative data collection will be conducted over an approximately 15-month period beginning in winter 2022. The study will be completed by 2026.
Discussion
This study will provide important insights into the implementation and effects of state health services delivery policies on addiction treatment and overdose during the COVID-19 pandemic. Aim 3 qualitative interview and document review results will inform the final design and interpretation of Aims 1–2 differences-in-differences analyses. For example, if we find in Aim 3 that interviewees’ view a given type of state policy as having effects on outcomes in urban but not rural areas, we could conduct an analysis limiting the sample in the difference-in-differences analysis to residents of urban counties. If we find consistently minimal local-level implementation of a given type of state policy in the 16 counties included in the Aim 3 case study, that may help to explain null findings for that policy in Aims 1–2. In addition, the Aim 3 qualitative study is designed to identify disaster-response health services delivery policy implementation best-practices, gaps, and strategies to overcome those gaps, as well as insight into whether and how the state policies evaluated in Aims 1–3 could generalize to other types of public health disasters. This study provides an example of a mixed-methods design supplementing econometric policy evaluation results with qualitative data characterizing details about policy implementation that are highly relevant to decision-makers and front-line policy implementers.
A key contribution of this study is consideration of local implementation of state public health policies. As in Aims 1–2 in the study described in this protocol, quantitative policy evaluations typically assess a policy of interest at a single level, e.g., state laws. However, on-the-ground policy implementation is often influenced by policies and practices at more granular levels. The disaster-response health services delivery policies in this study are an excellent example of this phenomenon, in that state policies must be implemented within counties and cities by local public health and healthcare systems. Policies and practices within those local jurisdictions and systems—for example, county-level designations of what services are deemed “essential” and can therefore be delivered in-person during disasters, a healthcare system policy delineating staffing for a virtual addiction consult service during the COVID-19 pandemic, or a local public health department policy laying out procedures for filling online requests for naloxone during the pandemic—likely influence implementation of state-level policies. Our study is designed to characterize policy implementation leaders’ perceptions of how these local factors influence implementation of the state health services delivery policies of interest. This analysis could be hypothesis-generating and inform the design of future quantitative studies assessing how state-local disaster-response policy interactions influence addiction treatment and overdose—or other health services and outcomes—during public health disasters.
A limitation of the study described in this protocol is our inability to measure policies’ effects on overdose prevention service delivery. No comprehensive data on delivery of these services exists. In a sensitivity analysis, we will explore policies’ effects on receipt of naloxone prescriptions in the IQVIA and OLDW data; however, most naloxone is distributed by health departments and other entities under standing orders with no prescriptions. To address the multi-level nature of our Aims 1 insurance claims data, we employ a two-way clustering of standard errors at the zip-code-state level. While our main approach uses fixed intercepts, which can be preferable for teasing out policy effects, we will also conduct sensitivity analyses using multi-level (hierarchical) modeling with random state and zip-code level intercepts. Another key consideration in this study is the complex policy environment in which states and localities implemented multiple policies designed to mitigate the effects of the COVID-19 pandemic on health services delivery at or around the same time. The study is designed to both acknowledge and unpack this complex environment by aligning specific policies with specific target populations and outcomes (e.g., state Medicaid policies should only affect Medicaid beneficiaries; state telehealth policies should only affect addiction treatment services delivered via telehealth technology) and by triangulating quantitative policy evaluation results from Aims 1–2 with in-depth qualitative data explaining front-line implementers’ perceptions of how policies work on the ground.
Little is known about how various state policies designed to mitigate the adverse effects of the COVID-19 pandemic on health services delivery have influenced drug addiction treatment and overdose. People who experience addiction and/or overdose risk are at particularly high risk of morbidity and mortality due to disaster-related service disruptions. Our study’s results will fill this gap and inform the design and implementation of policies in future public health disasters.
References
- 1.
Centers for Disease Control and Prevention. Disaster Epidemiology and Response. https://wwwcdcgov/nceh/hsb/disaster/traininghtm, Accessed May 27, 2020. 2020.
- 2. Frumkin H, Hess J, Luber G, Malilay J, McGeehin M. Climate change: the public health response. American journal of public health. 2008;98(3):435–445. pmid:18235058
- 3. Haines A, Kovats RS, Campbell-Lendrum D, Corvalán C. Climate change and human health: impacts, vulnerability and public health. Public health. 2006;120(7):585–596. pmid:16542689
- 4. Lindert J, Bilsen J, McKee M. Terrorist attacks: a public health issue. European journal of public health. 2018. pmid:30107459
- 5.
Centers for Disease Control and Prevention. Provisional Drug Overdose Death Counts. https://wwwcdcgov/nchs/nvss/vsrr/drug-overdose-datahtm, Accessed October 26, 2021. 2021.
- 6.
Substance Abuse and Mental Health Services Administration. Key substance use and mental health indicators in the United States: Results from the 2019 National Survey on Drug Use and Health (HHS Publication No. PEP20-07-01-001, NSDUH Series H-55). Rockville, MD: Center for Behavioral Health Statistics and Quality https://wwwsamhsagov/data/sites/default/files/reports/rpt29393/2019NSDUHFFRPDFWHTML/2019NSDUHFFR1PDFW090120pdf Accessed October 25, 2020. 2020.
- 7.
Alexander GC, Stoller KB, Haffajee RL, Saloner B. An epidemic in the midst of a pandemic: opioid use disorder and COVID-19. In: American College of Physicians; 2020.
- 8.
American Medical Association. Taking action on opioid use disorder, pain & harm reduction during COVID-19. https://wwwama-assnorg/delivering-care/opioids/taking-action-opioid-use-disorder-pain-harm-reduction-during-covid-19, Accessed May 27, 2020. 2020.
- 9.
Becker WC, Fiellin DA. When Epidemics Collide: Coronavirus Disease 2019 (COVID-19) and the Opioid Crisis. In: American College of Physicians; 2020.
- 10. Czeisler MÉ, Lane RI, Petrosky E, et al. Mental health, substance use, and suicidal ideation during the COVID-19 pandemic—United States, June 24–30, 2020. Morbidity and Mortality Weekly Report. 2020;69(32):1049. pmid:32790653
- 11.
National Academies of Sciences EaM. Medications for Opioid Use Disorder Save Lives. Washington, DC: The National Academies Press https://doiorg/1017226/25310. 2019.
- 12. Bradizza CM, Stasiewicz PR, Paas ND. Relapse to alcohol and drug use among individuals diagnosed with co-occurring mental health and substance use disorders: a review. Clinical psychology review. 2006;26(2):162–178. pmid:16406196
- 13. Scott CK, Dennis ML, Gustafson DH. Using ecological momentary assessments to predict relapse after adult substance use treatment. Addictive behaviors. 2018;82:72–78. pmid:29499393
- 14. Clarke PB, Lewis TF, Myers JE, Henson RA, Hill B. Wellness, Emotion Regulation, and Relapse During Substance Use Disorder Treatment. Journal of Counseling & Development. 2020;98(1):17–28.
- 15. Panebianco D, Gallupe O, Carrington PJ, Colozzi I. Personal support networks, social capital, and risk of relapse among individuals treated for substance use issues. International Journal of Drug Policy. 2016;27:146–153. pmid:26520237
- 16. Dasgupta N, Beletsky L, Ciccarone D. Opioid crisis: no easy fix to its social and economic determinants. American journal of public health. 2018;108(2):182–186. pmid:29267060
- 17. Saloner B, McGinty EE, Beletsky L, et al. A public health strategy for the opioid crisis. Public Health Reports. 2018;133(1_suppl):24S–34S. pmid:30426871
- 18. Henkel D. Unemployment and substance use: a review of the literature (1990–2010). Current drug abuse reviews. 2011;4(1):4–27. pmid:21466502
- 19. Molina KM, Alegría M, Chen C-N. Neighborhood context and substance use disorders: A comparative analysis of racial and ethnic groups in the United States. Drug and alcohol dependence. 2012;125:S35–S43. pmid:22699095
- 20. Carriere G, Garner R, Sanmartin C. Social and economic characteristics of those experiencing hospitalizations due to opioid poisonings. Health reports. 2018;29(10):23–28. pmid:30329146
- 21. McClure B, Mendoza S, Duncan L, Rotrosen J, Hansen H. Effects of regulation on methadone and buprenorphine provision in the wake of Hurricane Sandy. Journal of urban health. 2014;91(5):999–1008. pmid:25163931
- 22.
Levander X, Wakeman SE. COVID-19 will worsen the opioid crisis if we don’t prepare now. Stat. 2020;https://www.statnews.com/2020/03/17/covid-19-will-worsen-the-opioid-overdose-crisis-if-we-dont-prepare-now/, Accessed May 27, 2020.
- 23. Matusow H, Benoit E, Elliott L, Dunlap E, Rosenblum A. Challenges to opioid treatment programs after Hurricane Sandy: Patient and provider perspectives on preparation, impact, and recovery. Substance use & misuse. 2018;53(2):206–219.
- 24. Rutkow L, Vernick JS, Mojtabai R, Rodman SO, Kaufmann CN. Legal challenges for substance abuse treatment during disasters. Psychiatric services. 2012;63(1):1–9. pmid:22227749
- 25.
Yarmolinsky A, Rettig RA. Federal regulation of methadone treatment. National Academies Press; 1995.
- 26. Saloner B, Stoller KB, Alexander GC. Moving addiction care to the mainstream—improving the quality of buprenorphine treatment. New England Journal of Medicine. 2018;379(1):4–6. pmid:29972745
- 27. Madras BK, Ahmad NJ, Wen J, Sharfstein J. Improving Access to Evidence-Based Medical Treatment for Opioid Use Disorder: Strategies to Address Key Barriers Within the Treatment System. NAM Perspectives. 2020.
- 28. O’Connor PG, Nyquist JG, McLellan AT. Integrating addiction medicine into graduate medical education in primary care: the time has come. Annals of internal medicine. 2011;154(1):56–59. pmid:21200039
- 29. Townley C, Dorr H. Integrating substance use disorder treatment and primary care. The National Academy for State Health Policy. 2017;2017:1–16.
- 30. Kennedy-Hendricks A, Busch SH, McGinty EE, et al. Primary care physicians’ perspectives on the prescription opioid epidemic. Drug Alcohol Depend. 2016;165:61–70. pmid:27261154
- 31.
Stone EA, Barry CL, Kennedy-Hendricks A, Bachuber M, McGinty EE. Primary Care Provider Attitudes toward Medication Treatment for Opioid Use Disorder. Presentation at 2019 NIDA Addiction Health Services Research Conference, October 17, 2109. 2019.
- 32. Van Boekel LC, Brouwers EP, Van Weeghel J, Garretsen HF. Stigma among health professionals towards patients with substance use disorders and its consequences for healthcare delivery: systematic review. Drug and alcohol dependence. 2013;131(1–2):23–35. pmid:23490450
- 33. McGinty EE, Barry CL. Stigma Reduction to Combat the Addiction Crisis—Developing an Evidence Base. The New England journal of medicine. 2020;382(14):1291–1292. pmid:32242352
- 34.
Richter L, Vuolo L, Salmassi MS. Stigma and Addiction Treatment. In: The Stigma of Addiction. Springer; 2019:93–130.
- 35. Allen B, Nolan ML, Paone D. Underutilization of medications to treat opioid use disorder: What role does stigma play? Substance abuse. 2019;40(4):459–465. pmid:31550201
- 36. Wakeman SE, Rich JD. Barriers to medications for addiction treatment: How stigma kills. Substance use & misuse. 2018;53(2):330–333. pmid:28961017
- 37. Rutkow L. An analysis of state public health emergency declarations. American journal of public health. 2014;104(9):1601–1605. pmid:25033156
- 38. Walsh L, Garrity S, Rutkow L, et al. Applying a behavioral model framework for disaster recovery research in local public health agencies: a conceptual approach. Disaster medicine and public health preparedness. 2015;9(4):403–408. pmid:25912850
- 39. Hodge JG Jr, Rutkow L, Corcoran AJ. A hidden epidemic: assessing the legal environment underlying mental and behavioral health conditions in emergencies. Louis UJ Health L & Pol’y. 2010;4:33.
- 40. Hodge JG, Gostin LO, Vernick JS. The Pandemic and All-Hazards Preparedness Act: improving public health emergency response. JAMA: the journal of the American Medical Association. 2007;297(15):1708–1711.
- 41. Gostin LO, Hodge JG Jr. Public health emergencies and legal reform: implications for public health policy and practice. Public Health Reports. 2003;118(5):477. pmid:12941861
- 42.
Substance Abuse and Mental Health Services Administration. FAQs: Provision of methadone and buprenorphine for the treatment of Opioid Use Disorder in the COVID-19 emergency. https://wwwsamhsagov/sites/default/files/faqs-for-oud-prescribing-and-dispensingpdf, Accessed May 27, 2020. 2020.
- 43.
Wagenaar AC, Burris S. Public Health Law Research: Theory and Methods. Jossey-Bass; 2013.
- 44.
Podulka J, Blum J. Which Medicare Changes should Continue Beyond the COVID-19 Pandemic? Four Questions for Policymakers. Commonwealth Fund Issue Brief https://wwwcommonwealthfundorg/publications/issue-briefs/2021/may/which-medicare-changes-should-continue-beyond-covid-19-pandemic?utm_source=facebook&utm_medium=social&utm_campaign=Improving%20Health%20Care%20Quality. 2021.
- 45. Thomas EE, Haydon HM, Mehrotra A, et al. Building on the momentum: Sustaining telehealth beyond COVID-19. Journal of telemedicine and telecare. 2020:1357633X20960638. pmid:32985380
- 46. Wosik J, Fudim M, Cameron B, et al. Telehealth transformation: COVID-19 and the rise of virtual care. Journal of the American Medical Informatics Association. 2020;27(6):957–962. pmid:32311034
- 47. Butler SM. After COVID-19—Thinking Differently About Running the Health Care System. JAMA Health Forum. 2020;1(4):e200478–e200478.
- 48.
OptumLabs. OptumLabs and OptumLabs Data Warehouse (OLDW) Descriptions and Citation. Eden Prarie, MN: n.p., July 2020. PDF. Reproduced with permission from OptumLabs.
- 49. Mealer M, Jones Rn J. Methodological and ethical issues related to qualitative telephone interviews on sensitive topics. Nurse researcher. 2014;21(4):32–37. pmid:24673351
- 50.
Spradley JP. The Ethnographic Interview. New York: Holt, RInehart and Winston 1979.
- 51. Leech BL. Asking questions: Techniques for semistructured interviews. PS: Political Science & Politics. 2002;35(4):665–668.
- 52. Jennings GR. Interviewing: A focus on qualitative techniques. Tourism research methods: Integrating theory with practice. 2005:99–118.
- 53.
Patton M. Qualitative evaluation and research methods. In: Beverly Hills, CA: Sage; 1990.
- 54.
Quinn Patton M. Qualitative research and evaluation methods. In: Sage; 2002.
- 55.
Rubin HJ, Rubin I.S. Qualitative Interviewing: The Art of Hearing Data. Thousand Oaks, CA: Sage Publications; 1995.
- 56. Cameron AC, Gelbach JB, Miller DL. Robust Inference With Multiway Clustering. Journal of Business & Economic Statistics. 2011;29(2):238–249.
- 57. Yoo HI. VCE2WAY: Stata module to adjust a Stata command’s standard errors for two-way clustering. 2020.
- 58. Goodman-Bacon A. Difference-in-differences with variation in treatment timing. Journal of Econometrics. 2021.
- 59. Callaway B, Sant’Anna PH. Difference-in-differences with multiple time periods. Journal of Econometrics. 2020.
- 60. Sun L, Abraham S. Estimating dynamic treatment effects in event studies with heterogeneous treatment effects. Journal of Econometrics. 2020.
- 61.
Borusyak K, Jaravel X, Spiess J. Revisiting event study designs: Robust and efficient estimation. arXiv preprint arXiv:210812419. 2021.
- 62.
Gardner J. Two-stage differences in differences.
- 63. Kraetschmer NM, Deber RB, Dick P, Jennett P. Telehealth as gatekeeper: policy implications for geography and scope of services. Telemedicine journal and e-health: the official journal of the American Telemedicine Association. 2009;15(7):655–663. pmid:19694587
- 64. Harvey JB, Valenta S, Simpson K, Lyles M, McElligott J. Utilization of Outpatient Telehealth Services in Parity and Nonparity States 2010–2015. Telemedicine journal and e-health: the official journal of the American Telemedicine Association. 2019;25(2):132–136. pmid:29847224
- 65. Neufeld JD, Doarn CR, Aly R. State Policies Influence Medicare Telemedicine Utilization. Telemedicine journal and e-health: the official journal of the American Telemedicine Association. 2016;22(1):70–74. pmid:26218148
- 66. Grecu AM, Sharma G. The effect of telehealth insurance mandates on health-care utilization and outcomes. Applied Economics. 2019;51(56):5972–5985.
- 67. Lewis CB. Private payer parity in telemedicine reimbursement: how state-mandated coverage can be the catalyst for telemedicine expansion. U Mem L Rev. 2015;46:471.
- 68. Neufeld JD, Doarn CR, Aly R. State policies influence medicare telemedicine utilization. Telemedicine and e-Health. 2016;22(1):70–74. pmid:26218148
- 69. Davoli M, Bargagli AM, Perucci CA, et al. Risk of fatal overdose during and after specialist drug treatment: the VEdeTTE study, a national multi-site prospective cohort study. Addiction. 2007;102(12):1954–1959. pmid:18031430
- 70. Newhouse JP. Free for all?: lessons from the RAND health insurance experiment. Harvard University Press; 1993.
- 71. Sommers BD. Health care reform’s unfinished work—remaining barriers to coverage and access. New England Journal of Medicine. 2015. pmid:26509829
- 72. Goldman DP, Joyce GF, Zheng Y. Prescription drug cost sharing: associations with medication and medical utilization and spending and health. JAMA: the journal of the American Medical Association. 2007;298(1):61–69. pmid:17609491
- 73. Hodgkin D, Horgan CM, Garnick DW, Merrick EL. Cost sharing for substance abuse and mental health services in managed care plans. Medical Care Research and Review. 2003;60(1):101–116. pmid:12674022
- 74. Cairncross ZF, Herring J, van Ingen T, et al. Relation between opioid-related harms and socioeconomic inequalities in Ontario: a population-based descriptive study. CMAJ open. 2018;6(4):E478. pmid:30337473
- 75. Miller GE, Moriya AS. Any Use and Frequent Use of Opioids among Non-Elderly Adults in 2015–2016, by Socioeconomic Characteristics. 2018.
- 76. Walley AY, Xuan Z, Hackman HH, et al. Opioid overdose rates and implementation of overdose education and nasal naloxone distribution in Massachusetts: interrupted time series analysis. Bmj. 2013;346:f174. pmid:23372174
- 77. Krieger MS, Goedel WC, Buxton JA, et al. Use of rapid fentanyl test strips among young adults who use drugs. International Journal of Drug Policy. 2018;61:52–58. pmid:30344005