Abstract
BACKGROUND:
Overprescribing of opioids has led to hundreds of thousands of overdose deaths and substantial health care costs. In response, the US Food and Drug Administration (FDA) implemented a revised Risk Evaluation and Mitigation Strategy (REMS) for opioids in 2018.
OBJECTIVE:
To evaluate trends in opioid prescribing by oncologists for Medicare Part D beneficiaries from 2014 to 2022.
METHODS:
This cross-sectional study used data from the 2014-2022 Medicare Part D Prescriber Public Use Files. Opioid claims and prescribing trends were assessed by opioid types, oncologist subspecialty, geographic region, and rurality status. An interrupted time series analysis was conducted to assess the changes in oncologists’ prescribing patterns before and after the 2018 REMS modifications.
RESULTS:
The analysis included 25,371 unique oncologists, with the majority being male (66%) and specializing in hematology-oncology (47%). Over the study period, oncologists issued more than 9.4 million opioid prescriptions, with long-acting opioids accounting for 18% of these claims. Hematology-oncology specialists were responsible for the largest share of the prescriptions (67%). Oncologists practicing in the South and rural areas exhibited higher prescribing rates and longer average supply durations than those in other regions. A national sustainable decline in opioid prescribing was observed among oncologists between 2014 and 2022, with a significant immediate decline following 2018 in which the REMS changes were implemented.
CONCLUSIONS:
The 2018 FDA REMS update coincided with significant declines in opioid prescribing by oncologists treating Medicare beneficiaries. Although other factors, such as the COVID-19 pandemic, may have also contributed to this decline, the sustained downward trend over time highlights the need for targeted policies and tailored provider education to ensure effective cancer pain management and to address persistent regional and rural-urban disparities in prescribing practices.
Plain language summary
Opioid prescribing by oncologists for Medicare patients declined significantly from 2014 to 2022, with notable geographic disparities. Findings reflect regulatory impacts and health care inequities, underscoring the importance of balancing opioid restrictions with ensuring adequate evidence-based pain management in cancer care.
Implications for managed care pharmacy
This study demonstrates declining trends in opioid prescribing by oncologists treating Medicare beneficiaries, suggesting that regulatory measures may be influencing prescribing in oncology. These trends have important implications for pain management policies and clinical practices to ensure access for patients with legitimate needs. Higher prescribing rates observed among oncologists in the South and rural areas emphasize the need for targeted policies and tailored provider education to address geographic disparities and promote equitable evidence-based prescribing practices.
Prescribing the right medication at the optimal dose and duration is crucial for effective disease management and prevention.1 However, overprescribing has recently increased, negatively impacting patient care and adding considerable costs to health care systems.2,3 Opioids have been overprescribed in the United States for decades and represent a major public health crisis,4 resulting in billions of dollars in health care costs and hundreds of thousands of hospitalizations and deaths. Between 1999 and 2020, more than 263,000 deaths were attributed to prescribed opioids,5 with nearly 90,000 overdose deaths recorded in 2021 alone.6
In response to the national crisis, the US Food and Drug Administration (FDA) implemented a Risk Evaluation and Mitigation Strategy (REMS) in 2012 for long-acting opioids.7 This program aimed to promote evidence-based prescribing practices and reduce unnecessary patient exposure through provider education.7 In 2018, the REMS was expanded to include immediate-release opioids, broadened training to additional health care providers, updated the educational content, and introduced new safety labeling.8
These regulatory initiatives have been associated with reductions in opioid misuse, overdose deaths, and overall prescribing trends.9–12 However, the resulting declines have impacted both cancer and noncancer patients alike,11,12 raising concerns about the potential under prescribing of opioids for cancer pain. Multidisciplinary expert groups have worked to align opioid restriction policies with clinical practice guidelines to ensure adequate and appropriate pain treatment in cancer populations.13 Although the evidence on opioid misuse risk in cancer patients is limited,14,15 unclear guidelines and regulatory challenges may contribute to oncologists’ hesitancy to prescribe opioids and limit access for patients with legitimate pain management needs.
This study examines opioid prescribing trends among oncologists treating Medicare Part D beneficiaries from 2014 to 2022. Oncologist prescribing patterns are hypothesized to remain relatively stable following the 2018 REMS modifications. By contextualizing these trends within broader regulatory changes and pandemic-related disruptions, this analysis offers insights into the intersection of oncology care and national opioid policies.
Methods
DATA SOURCE
This cross-sectional study used data from the Medicare Part D Prescriber Public Use Files for years 2014-2022.16 These files contain annual provider-level aggregated information on prescription drug claims. We included providers specializing in oncology who prescribed at least 1 medication covered by the Medicare Part D program across the United States. Oncology subspecialties included hematology-oncology, medical oncology, surgical oncology, gynecological oncology, and radiation oncology, as classified in the dataset based on Medicare specialty codes. Institutional review board approval was not required as the data are publicly available.
VARIABLES DEFINITION
The primary outcomes were (1) opioid prescribing rate, expressed as the number of opioid claims filled per 100 Medicare beneficiaries, and (2) average days supply, defined as the mean number of days supply of opioids per prescriber. Both outcomes were calculated annually at the prescriber level and then averaged across all oncologists over the entire study period as well as within the premodification (2014-2017) and postmodification (2019-2022) periods. In cases in which patient or claim counts were suppressed because of small cell sizes (counts <11), we imputed values as zero to simplify calculations and avoid overestimation.
Additional variables included prescriber characteristics (eg, gender, subspecialty), opioid prescription type (any opioid vs long-acting opioids), and practice location (southern states vs other regions; rural vs urban areas). Long-acting opioids per Medicare Part D definitions included buprenorphine, fentanyl, hydrocodone, hydromorphone, levorphanol, methadone, morphine, oxycodone, oxymorphone, tapentadol, and tramadol. States classification followed the US Census Bureau regional definitions,17 and rurality was determined using Rural-Urban Commuting Area Codes, with codes 9 or 10 indicating rural areas.
Practice size was defined as the number of Medicare beneficiaries with at least 1 prescription (for any medication) filled by the oncologist. Patient characteristics included the mean age and risk score (assigned by the Centers for Medicare and Medicaid Services [CMS] based on patient demographics and health status), and both were averaged across each prescriber’s patient population.
STATISTICAL ANALYSIS
Descriptive statistics, including means and percentages, were calculated to summarize prescriber characteristics, opioid claim types, and geographic distributions. Opioid prescribing trends were evaluated over the full study period and stratified by oncologist subspeciality, region, and rurality status to identify disparities.
To assess and quantify the impact of the 2018 FDA REMS update, we used an interrupted time series model with segmented linear regression. This model estimated level changes (intercepts) and trend changes (slopes) to detect immediate shifts and long-term differences in prescribing rates, respectively. The premodification baseline period was defined as 2014-2017 and the postmodification period as 2019-2022. The year 2018 was excluded from the model to account for a potential transition period during policy implementation. All data analyses were performed using SAS software version 9.4 TS1M6 (SAS Institute Inc.). Statistical significance was set at P < 0.05.
Results
PRESCRIBER CHARACTERISTICS
Of Medicare prescribers, our analysis included 25,371 unique oncologists. The majority were male (66%) and specialized in hematology-oncology (47%). Approximately 34% of oncologists were based in the South, whereas less than 1% practiced in rural areas. On average, each oncologist had a practice size of 113 Medicare beneficiaries, with a mean patient age of 72 years and an average CMS risk score of 2 (approximately twice the average score for the general Medicare population).
CLAIMS CHARACTERISTICS
Over the study period, oncologists issued more than 9.4 million opioid prescriptions totaling 175 million days of supply, with a mean duration of 1,432 days per prescriber. Long-acting opioids were prescribed by 46% of oncologists and constituted 18% of all opioid claims. Hematology-oncology specialists accounted for the largest share of claims (67%), followed by medical oncologists (21%). Nearly half of all opioid prescriptions were dispensed in southern states where oncologists had higher mean total claims of 80 and supply days of 1,803 compared with 47 claims and 1,210 days in other regions. Similarly, oncologists in rural areas had nearly twice the average number of opioid claims and longer average supply durations than those in urban areas (100 claims and 1,855 days vs 58 claims and 1,431 days, respectively).
PRESCRIBING TRENDS
Overall, oncologists prescribed 47 opioid claims on average per 100 Medicare beneficiaries, which included 8 long-acting opioid claims per 100 beneficiaries. Hematology-oncology specialists showed the highest mean opioid prescribing rate at 55 claims per 100 patients, whereas radiation oncologists showed the lowest rate at 29 claims per 100 patients. Throughout the study period, oncologists in the South and rural areas consistently had higher prescribing rates, averaging 58 vs 41 claims per 100 patients compared with other regions, and 61 vs 47 claims per 100 patients compared with urban areas, respectively.
In the postmodification period (2019-2022), the mean opioid prescribing rates declined nationally from 57 to 38 claims per 100 beneficiaries, with a corresponding decrease of approximately 330 days in the average supply per oncologist, compared with the premodification period. These downward trends were consistent across different opioid types, oncologist subspecialities, geographic regions, and rural-urban settings (Table 1).
TABLE 1.
Rate of Opioid Prescribing by Oncologists per 100 Medicare Beneficiaries
| Premodification (2014-2017) | Postmodification (2019-2022) | Overalla (2014-2022) | |
|---|---|---|---|
| By opioid type | |||
| Any opioid | 57 | 38 | 47 |
| Long-acting opioids | 10 | 5 | 8 |
| By oncologist subspeciality | |||
| Hematology-oncology | 67 | 44 | 55 |
| Medical oncology | 62 | 40 | 50 |
| Surgical oncology | 56 | 44 | 50 |
| Gynecological oncology | 47 | 33 | 39 |
| Radiation oncology | 36 | 23 | 29 |
| By geographic region | |||
| South | 68 | 48 | 58 |
| Other regions | 50 | 32 | 41 |
| By rurality status | |||
| Rural | 70 | 54 | 61 |
| Urban | 57 | 38 | 47 |
Overall values represent averages across the entire study period.
The interrupted time series analysis revealed a sustained and statistically significant overall decline in opioid prescribing by oncologists throughout the study period. Compared with the premodification period, there was a significant immediate reduction of approximately 3,302 opioid claims per year across the cohort, which corresponded to an estimated drop of 5 claims per 100 patients in the postmodification period (change in intercept estimate = –5.4033; 95% CI = –6.0130 to –4.7935; P < 0.0001). Although prescribing continued to decline after the 2018 REMS update, the rate of decrease was significantly less steep than that during the premodification period as reflected by a positive change in slope (interaction estimate = +1.6349; 95% CI = 1.3232 to 1.9466; P < 0.0001). The premodification slope was –4.0947 (95% CI = –4.3250 to –3.8645), whereas the postmodification slope was –2.4598 (95% CI = –2.6454 to –2.2742). Figure 1 illustrates these temporal trends in opioid prescribing.
FIGURE 1.
Trends in Opioid Prescribing by Oncologists From 2014 to 2022
Discussion
This study demonstrates a consistent and sustained national decline in opioid prescribing by oncologists between 2014 and 2022, with a notable immediate discontinuity after 2018. This shift aligns temporally with the implementation of the FDA REMS modifications and the onset of the COVID-19 pandemic and may also reflect broader regulatory measures and growing clinical hesitancy in opioid prescribing. These trends were evident across both total and long-acting opioids as well as in the average durations of supply. Persistent higher prescribing rates and longer supply durations were observed in the South and in rural areas. Although these patterns might suggest greater access to opioids in those settings, they may instead reflect systemic gaps (eg, inconsistencies in guidelines application or limited access to nonopioid alternative resources) and underscore ongoing geographic inequities in cancer pain management that merit targeted policy and practice interventions.
Previous studies have reported a comparable steady decline in opioid prescribing over time across both oncologic and nononcologic providers prior to the 2018 FDA opioid strategy changes.12,18,19 This broad downward trend reflects, in part, the impact of national regulatory initiatives on mitigating opioid overprescribing. However, concerns remain that the reductions in opioid prescribing within cancer care may result from prescriber hesitancy and regulatory challenges rather than clinically appropriate adjustments or evidence-based practices. Some investigations have noted that the reduction in opioid prescribing has coincided with a shift toward nonopioid alternatives (eg, gabapentin) to minimize opioid dependence.11,18,19 Nevertheless, the effectiveness of these alternatives in managing severe or chronic cancer pain is uncertain.
Multiple policy changes have had unintended consequences on cancer patients. For example, while Prescription Drug Monitoring Programs have been effective in curbing inappropriate opioid use, these programs have also imposed administrative burdens on oncologists, such as dosage caps and prior authorization requirements that often delay timely access to opioids and undermine patient outcomes.20 States without exemptions for cancer patients in their Prescription Drug Monitoring Programs have experienced sharper declines in opioid prescribing.21 Moreover, the Centers for Disease Control and Prevention (CDC) opioid prescribing guideline for chronic noncancer pain has been inappropriately applied to cancer populations, despite the recommendations not being intended for them.22,23 The development of oncology-specific guidelines and policies could help balance regulatory objectives with the need for effective cancer pain management.
Reductions in opioid prescribing cannot be attributed solely to the FDA REMS modifications or other evolving regulatory measures. It is important to acknowledge that the health care system disruptions caused by the COVID-19 pandemic may have independently influenced opioid prescribing, particularly between 2020 and 2022, and contributed to the observed deceleration in the declining trend during the postmodification period. During the pandemic, routine health care and pain management were affected by reduced in-person visits, increased reliance on telemedicine, and broader shifts in prescribing practices.24,25 Researchers have reported pandemic-related delays in cancer treatment and follow-up care, which likely impacted clinical decision-making for pain management among oncologists.26 Compared with nononcologic providers during the same period, studies showed that prescribing patterns across the health care system fluctuated considerably, reflecting broad uncertainty and changes not only in opioid prescribing but also in prescribing of other medications.27,28
Hematology-oncology specialists maintained the highest prescribing rates during the study period, indicating the complex and severe pain management needs of patients with hematologic malignancies. Differences in prescribing behaviors across oncologist subspecialties emphasize the importance of tailored approaches that consider specific clinical contexts and patient needs.
Regional trends revealed consistently higher opioid prescribing rates and longer supply durations among oncologists practicing in the South and rural areas compared with those in other locations. These disparities may reflect systemic inequities in access to comprehensive pain care likely due to variations in health care infrastructure, sociocultural and economic factors that shape prescribing norms, the availability of nonopioid alternatives, as well as a higher prevalence of chronic pain conditions among rural populations.4,29,30 Addressing these challenges through targeted policies, improved resource allocation, and tailored provider education is essential to promote equitable and effective cancer pain management across geographic settings.
LIMITATIONS
This study uniquely examines opioid prescribing trends among cancer patients during a period marked by the FDA REMS modifications and the disruptions of the COVID-19 pandemic. However, the study has limitations, including reliance on aggregate claims data, which precludes an assessment of the clinical appropriateness of prescribing practices. It also does not account for opioid prescriptions from nononcologic providers and may underestimate the full extent of opioid access among cancer patients. Finally, findings are specific to Medicare Part D prescribers and may not be generalized to other populations. Future research should explore qualitative barriers to opioid prescribing from the oncologist perspective and incorporate patient-level outcomes (eg, pain control level, quality of life) to comprehensively assess the clinical implications of declining opioid prescribing trends within oncology care.
Conclusions
These findings highlight the potential impact of opioid restriction policies on prescribing practices in oncology. Efforts to mitigate the opioid crisis must consider the unique and legitimate pain management needs of cancer patients and ensure that regulatory measures do not inadvertently compromise oncologic care. Targeted policies and tailored provider education are critical to address regional and rural-urban disparities in prescribing practices and promote equitable access to pain resources across settings.
References
- 1.Maxwell SR. Rational prescribing: The principles of drug selection. Clin Med (Lond). 2016;16(5):459-64. doi: 10.7861/clinmedicine.16-5-459 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Safer DJ. Overprescribed medications for US adults: Four major examples. J Clin Med Res. 2019;11(9):617-22. doi: 10.14740/jocmr3906 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Herzstein J, Ebell M. Improving quality by doing less: Overtreatment. Am Fam Physician. 2015;91(5):289-91. [PubMed] [Google Scholar]
- 4.Lyden J, Binswanger IA. The United States opioid epidemic. Semin Perinatol. 2019;43(3):123-31. doi: 10.1053/j.semperi.2019.01.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Centers for Disease Control and Prevention, National Center for Health Statistics . Drug overdose: Overview. Accessed April 23, 2024. https://archive.cdc.gov/www_cdc_gov/drugoverdose/data/OD-death-data.html
- 6.Centers for Disease Control and Prevention . National Center for Health Statistics. U.S. overdose deaths in 2021 increased half as much as in 2020 – but are still up 15%. Accessed October 1, 2022. https://www.cdc.gov/nchs/pressroom/nchs_press_releases/2022/202205.htm#:∼:text = Provisional%20data%20from%20CDC’s%20National,93%2C655%20deaths%20estimated%20in%202020
- 7.US Food & Drug Administration . Questions and answers: FDA approves a Risk Evaluation and Mitigation Strategy (REMS) for extended-release and long-acting (ER/LA) opioid analgesics. Accessed October 1, 2022. https://www.fda.gov/drugs/information-drug-class/questions-and-answers-fda-approves-risk-evaluation-and-mitigation-strategy-rems-extended-release-and#:∼:text=On%20July%209%2C%202012%2C%20the,ER%2FLA)%20opioid%20analgesics
- 8.US Food & Drug Administration . Opioid analgesic Risk Evaluation and Mitigation Strategy (REMS). Accessed October 1, 2022. https://www.fda.gov/drugs/information-drug-class/opioid-analgesic-risk-evaluation-and-mitigation-strategy-rems
- 9.Bucher Bartelson B, Le Lait MC, Green JL, et al. Changes in misuse and abuse of prescription opioids following implementation of extended-release and long-acting opioid analgesic Risk Evaluation and Mitigation Strategy. Pharmacoepidemiol Drug Saf. 2017;26(9):1061-70. doi: 10.1002/pds.4257 [DOI] [PubMed] [Google Scholar]
- 10.Black JC, Bau GE, Rosen T, et al. Changes in mortality involving extended-release and long-acting opioids after implementation of a Risk Evaluation and Mitigation Strategy. Pain Med. 2020;21(1):92-100. doi: 10.1093/pm/pnz031 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Bandara S, Bicket MC, McGinty EE. Trends in opioid and non-opioid treatment for chronic non-cancer pain and cancer pain among privately insured adults in the United States, 2012-2019. PLoS One. 2022;17(8):e0272142. doi: 10.1371/journal.pone.0272142 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Agarwal A, Roberts A, Dusetzina SB, Royce TJ. Changes in opioid prescribing patterns among generalists and oncologists for Medicare Part D beneficiaries from 2013 to 2017. JAMA Oncol. 2020;6(8):1271-4. doi: 10.1001/jamaoncol.2020.2211 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Schatz AA, Oliver TK, Swarm RA, et al. Bridging the gap among clinical practice guidelines for pain management in cancer and sickle cell disease. J Natl Compr Canc Netw. 2020;18(4):392-9. doi: 10.6004/jnccn.2019.7379 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Ako T, Ørnskov MP, Lykke C, Sjøgren P, Kurita GP. Prevalence of opioid misuse in patients with cancer: A systematic review and meta-analysis. Br J Cancer. 2024;131(6):1014-20. doi: 10.1038/s41416-024-02802-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Le TT, Fleming SP, Simoni-Wastila L. Patterns of opioid use in commercially insured patients with cancer. Am J Manag Care. 2022;28(5):207-11. doi: 10.37765/ajmc.2022.89141 [DOI] [PubMed] [Google Scholar]
- 16.Centers for Medicare & Medicaid Services . Medicare Part D prescribers - by provider. Accessed March 12, 2025. https://data.cms.gov/provider-summary-by-type-of-service/medicare-part-d-prescribers/medicare-part-d-prescribers-by-provider
- 17.US Census Bureau . Census regions and divisions of the United States. Accessed November 11, 2024. https://www2.census.gov/geo/pdfs/maps-data/maps/reference/us_regdiv.pdf
- 18.Korst MR, Santos Teles M, Choudhry HS, et al. Characterizing opioid prescribing trends of medical oncologists from 2013 to 2019: analysis from the Centers for Medicare & Medicaid Services Medicare Part D Prescribers Database. JCO Oncol Pract. 2024;20(2):268-77. doi: 10.1200/OP.23.00285 [DOI] [PubMed] [Google Scholar]
- 19.Jairam V, Yang DX, Pasha S, et al. Temporal trends in opioid prescribing patterns among oncologists in the Medicare population. J Natl Cancer Inst. 2021;113(3):274-81. doi: 10.1093/jnci/djaa110 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Graetz I, Yarbrough CR, Hu X, Howard DH. Association of mandatory-access prescription drug monitoring programs with opioid prescriptions among Medicare patients treated by a medical or hematologic oncologist. JAMA Oncol. 2020;6(7):1102-3. doi: 10.1001/jamaoncol.2020.0804 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Graetz I, Hu X, Ji X, Wetzel M, Yarbrough CR. The effect of cancer exemption in mandatory-access prescription drug monitoring programs among oncologists. JNCI Cancer Spectr. 2023;7(2):pkad006. doi: 10.1093/jncics/pkad006 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Kroenke K, Alford DP, Argoff C, et al. Challenges with implementing the Centers for Disease Control and Prevention opioid guideline: A Consensus Panel Report. Pain Med. 2019;20(4):724-35. doi: 10.1093/pm/pny307 [DOI] [PubMed] [Google Scholar]
- 23.Dowell D, Haegerich T, Chou R. No shortcuts to safer opioid prescribing. N Engl J Med. 2019;380(24):2285-7. doi: 10.1056/NEJMp1904190 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Mehrotra A, Chernew ME, Linetsky D, Hatch H, Cutler DM, Schneider EC. The impact of the COVID-19 pandemic on outpatient visits: Practices are adapting to the new normal. The Commonwealth Fund. Accessed June 25, 2025. https://www.commonwealthfund.org/publications/2020/jun/impact-covid-19-pandemic-outpatient-visits-practices-adapting-new-normal
- 25.Alexander GC, Tajanlangit M, Heyward J, Mansour O, Qato DM, Stafford RS. Use and content of primary care office-based vs telemedicine care visits during the COVID-19 pandemic in the US. JAMA Netw Open. 2020;3(10):e2021476. doi: 10.1001/jamanetworkopen.2020.21476 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Patt D, Gordan L, Diaz M, et al. Impact of COVID-19 on cancer care: How the pandemic is delaying cancer diagnosis and treatment for American seniors. JCO Clin Cancer Inform. 2020;4(4):1059-71. doi: 10.1200/CCI.20.00134 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.French C, Jackson J, Monahan Z, Murray K, Hartwell M. Emergency department opioid prescribing trends among provider types: An analysis of the NHAMCS, 2019-2021. Intern Emerg Med. Published online March 31, 2025. doi: 10.1007/s11739-025-03923-5 [DOI] [PubMed] [Google Scholar]
- 28.Vaduganathan M, van Meijgaard J, Mehra MR, Joseph J, O’Donnell CJ, Warraich HJ. Prescription fill patterns for commonly used drugs during the COVID-19 pandemic in the United States. JAMA. 2020;323(24):2524-6. doi: 10.1001/jama.2020.9184 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Skinner J. Causes and consequences of regional variations in health. In: Pauly MV, Mcguire TG, Barros PP, eds. Handbook of Health Economics. Vol 2. Elsevier; 2011:45-93. [Google Scholar]
- 30.Baker MB, Liu EC, Bully MA, et al. Overcoming barriers: A comprehensive review of chronic pain management and accessibility challenges in rural America. Healthcare (Basel). 2024;12(17):1765. doi: 10.3390/healthcare12171765 [DOI] [PMC free article] [PubMed] [Google Scholar]