Aileen Rothbard, ScD

Philadelphia, Pennsylvania, is an urban epicenter of the opioid epidemic, and inappropriate opioid prescribing remains a top concern. To help address this issue, the Philadelphia Medicaid Opioid Prescribing Initiative, a type of community quality collaborative

mailed thousands of local Medicaid providers an individualized prescribing report card in 2017 and 2018. The report card featured details of providers’ opioid prescribing, including peer comparison measures and inappropriate prescribing measures like concomitant opioid and benzodiazepine prescribing. This case study describes the unique process of developing and distributing the opioid prescribing report cards, with a particular focus on the role of Medicaid managed care organizations. Using Medicaid pharmacy claims, the extensive variation in prescribing measures within and across specialties is also illustrated. The report card’s implementation points to the potential value of collaborations between public health departments and Medicaid managed care organizations and can provide insight for other locally grown policies.

This article describes the Philadelphia Medicaid Opioid Prescribing Initiative that was launched by a multidisciplinary team and mailed local Medicaid providers individualized prescribing report cards.

To our knowledge, Philadelphia is the first municipality to develop and deliver its own individualized opioid prescribing report cards to local Medicaid providers. We describe the unique process of creating and distributing the prescribing report cards, which can be replicated by other municipalities.

The implementation of these prescribing report cards showcases the unique role that municipalities have played in combating the opioid epidemic and points to the potential value of collaborations between public health departments and managed care organizations.

The prescribing report cards reveal substantial variation in all measures, within and across provider specialties.

Philadelphia, Pennsylvania, is an urban epicenter of the opioid epidemic.

 Heroin and synthetic opioids such as fentanyl resulted in more than 1000 overdose deaths in Philadelphia in 2019, representing one of the highest death rates among metropolitan areas in the United States.

 Beyond mortality, opioid use disorder (OUD) has far-reaching consequences across the community, with links to criminal justice involvement

 and higher rates of neonatal abstinence syndrome and foster care placements.

For many patients with OUD, their first experience with opioids was not illicit. Instead, it occurred in a physician’s office with a legal prescription for opioids.

 Billions of opioid prescriptions have been written in the past few decades, and surveys report that upward of 80% of new heroin users began with prescription opioids.

 Although many of these prescriptions adhere to clinical guidelines, others needlessly place patients (and their family and friends

To curb inappropriate opioid prescriptions, myriad state-level policies have been introduced (eg, pharmacies requiring patients to show identification to pick up opioid prescription fills).

 Yet states can be slow to implement policies that have been shown to be effective. For example, it took more than a decade for some states to institute prescription drug monitoring programs (PDMPs).

Locally grown policies may offer some advantages, including institutional knowledge of underlying population problems and existing relationships across stakeholder groups. One recent example is Philadelphia’s Medicaid Opioid Prescribing Initiative, which was intended to improve the appropriateness of opioid prescriptions in Philadelphia County. Many of the participants were already working together on other public health initiatives in the county. Its first task was instituting a performance feedback initiative in which thousands of local Medicaid providers received customized report cards detailing their opioid prescribing relative to their peers.

Here we present details of this community quality collaborative, which included public health officials and Medicaid managed care organizations.

 This case study points to the potential value of county-level collaboration in combating the opioid epidemic, particularly the unique role of managed care organizations, and provides insight for other policy interventions instituted at a local level. To our knowledge, Philadelphia is the first municipality to develop and deliver its own opioid prescribing report card.

Public and private reporting on the performance of health care providers is common practice.

 Whereas public reporting is often used to inform patients about the quality of health care services, private reporting tends to target providers.

 A common type of private reporting is performance feedback, which is sometimes presented in report cards.

 Performance feedback has been extensively studied, and a review of randomized trials points to a positive effect overall.

 However, the success of performance feedback depends on providers’ baseline performance and the method, source, and duration of feedback.

The major barrier to implementing performance feedback is the availability of verified information regarding provider performance. For opioid prescribing, PDMPs are one potential source of administrative data. Since the onset of the opioid epidemic, the number of states instituting PDMPs has grown rapidly, with programs now operational in almost every state.

 Yet only a few states have launched report cards using this information, and there is a paucity of academic literature on the implementation and impact of report cards using PDMPs.

Another source of performance feedback is administrative data collected by insurers and managed care organizations (MCOs).

 A growing number of state Medicaid programs are creating performance feedback initiatives, and many are involving Medicaid MCOs, which are also encouraged to introduce quality improvement programs.

 However, in states such as Pennsylvania, Medicaid MCOs are operated at the county level, presenting an administrative barrier to implementing statewide programs.

In 2017, Community Behavioral Health (CBH), the behavioral health MCO in Philadelphia County for Medicaid enrollees, approached Pennsylvania’s Office of Medical Assistance Programs to discuss the need to monitor opioid prescribing in Philadelphia. CBH was experiencing an increased demand for substance use disorder treatment. Informal investigations revealed inappropriate prescribing practices, such as the concurrent use of opioids and methadone maintenance.

Like several other states, Pennsylvania “carves out” behavioral health from physical health, and state Medicaid officials were interested in supporting CBH in developing a tool that could subsequentially be taken over by the state’s PDMP. CBH was in a unique position to lead the report card initiative because it was the sole behavioral health MCO in Philadelphia, with up-to-date Medicaid eligibility records, pharmacy claims, and physical health claims, in addition to behavioral health claims for the county’s Medicaid population. By comparison, there were 4 physical health MCOs that could access only the claims of their Medicaid enrollees.

Report cards were envisioned as a complementary initiative to the PDMP. Although providers could search for opioid prescriptions of their patients using the PDMP, many were not actively using the resource.

 Even when they did use the PDMP, providers could not compare their own prescribing with that of their peers, nor could they see all controlled substances. Prior to Act 191 of 2014, Pennsylvania’s PDMP reported Schedule II controlled substances only. Also, prescriptions for benzodiazepines, a Schedule IV controlled substance, were reportable beginning in 2017.

After receiving support from state officials, CBH convened a project team of stakeholders across Philadelphia, including Philadelphia’s Department of Public Health and Department of Behavioral Health and Intellectual disAbility Services, the 4 physical health Medicaid MCOs, and researchers from the University of Pennsylvania’s Center for Mental Health (“Penn”). CBH created an internal management team consisting of physicians, pharmacists, epidemiologists, health economists, and data managers. It also appointed an administrative assistant to coordinate the project’s activities.

The project team members agreed that report cards could help close gaps in information regarding providers’ prescribing, but they felt that the report card should be educational rather than punitive. To design the report card, researchers from Penn reviewed the literature on existing report card initiatives and performed preliminary analyses of the pharmacy claims. The literature review identified a technical assistance report prepared by Brandeis University researchers that included templates of opioid report cards, which were developed by several states in conjunction with their PDMPs.

 Arizona’s model served as the basis for Philadelphia’s report because it had design features and measures that were consistent with the interests of the project team. It consisted of a single page displaying various prescribing measures, some with peer comparison.

In addition to aggregate opioid prescribing, other measures were considered. The team was interested in benzodiazepine prescribing and concomitant opioid-benzodiazepine prescribing. Since 2012, more than 2000 overdose deaths in Philadelphia have involved benzodiazepines, which are among the most abused psychotropic drugs and should not be used concurrently with opioids due to an increased risk of respiratory depression.

 The team also added high-dose opioid prescribing as a measure. Clinical guidelines suggest a daily dose lower than 50 to 90 morphine milligram equivalents (MME), but study results have shown that nearly 15% of patients have higher-than-recommended doses.

 A final measure flagged opioid prescriptions to patients with a history of substance use treatment.

Once measures were agreed upon, the methodology for constructing peer comparison was determined. The team used the National Provider Identifiers, located in pharmacy claims, and linked them to self-identified taxonomy codes of specialties in the National Plan and Provider Enumeration System (NPPES). A group of physicians, pharmacists, and epidemiologists at CBH grouped 266 specialties into 27 broader categories using the Health Care Prescriber Taxonomy Code and Specialty crosswalk from the National Uniform Claim Committee.

Providers received a report card only if they prescribed opioids to 10 or more Medicaid patients over a 12-month period. Any prescription written to Philadelphia Medicaid enrollees could be included in the report cards, even if providers did not practice in Philadelphia County.

Over the course of a few months, the project team agreed on a report card design (

). For the first 6 measures, report cards included a ranking system for peer comparison purposes. Data were aggregated within specialties and a mean and SD were estimated. Rankings were categorized as extreme (3 SDs above the mean), severe (2 SDs above the mean), high (1 SD above the mean), and normal (the mean or below). Providers who prescribed opioids to fewer than 10 patients were still included in peer comparison calculations.

The first iteration of report cards was mailed in December 2017 to 2632 providers. The mailings included recommendations for best prescribing practices from the CDC and information on an upcoming continuing medical education program focused on evidence-based tools to fight the opioid epidemic.

 Links to the PDMP and a list of Philadelphia’s substance use treatment providers were also included.

Report cards contained contact information for the Department of Public Health, where comments and questions could be submitted. All feedback was reviewed by the pharmacy director and medical director at CBH. Fewer than 10 providers responded, with some saying that the report cards included valuable information. Providers also referred to a key limitation: Report cards included only Medicaid patients. Although Medicaid patients have been disproportionally affected by the opioid epidemic,

 the providers argued that prescribing report cards would be more meaningful if they included a provider’s complete patient panel. This would have been possible had the data source been Pennsylvania’s PDMP or an all-payer claims database, the latter of which is growing in popularity but has not been used widely for performance feedback.

After the first mailing, the project team redesigned some aspects of the report card. Inconsistencies across waves make an evaluation more complex, so only 3 changes were ultimately instituted: report cards included 6 months of prescribing rather than 12; the threshold of 90 MME/day was lowered to 50 MME/day; and the proportion of opioid prescriptions lasting longer than 7 days was added. An excessive duration of opioid prescriptions was another concern because there is little evidence of the effectiveness of opioids for noncancer pain in the long term.

 In July 2018, a second wave of prescribing report cards was sent to 1495 providers.

Report cards revealed substantial variation in opioid prescribing, highlighting why opioid-related policies cannot be one size fits all. While a provider’s peer comparison categorization could be normal for one measure and extreme for another, specialties explained much of the variation in prescribing (

). The mean number of Medicaid patients filling an opioid prescription over 12 months (among all Medicaid providers who prescribed at least 1 opioid or benzodiazepine) was only 0.6 for psychiatry but higher in pain medicine (20.5 patients), emergency medicine (11.8 patients), and surgery (43.1 patients for surgical dentists, 18.0 for surgical podiatrists, and 12.9 for general surgeons). Benzodiazepine prescribing, which was also included in report cards, was common in psychiatry (28.8 patients on average) and rarer in pain medicine (2.8 patients), emergency medicine (2.0 patients), and surgery (0.2 patients for surgical dentists, 0.2 for surgical podiatrists, and 0.6 for general surgeons).

Inappropriate prescribing measures showed similar variation. Among the subset of providers who met the 10-patient threshold to receive report cards, the share of patients filling concurrent opioid-benzodiazepine prescriptions ranged from less than 10% to nearly 50% across specialties. High-dose opioid prescriptions occurred less frequently. The rate of opioid prescribing to patients with substance use disorder treatment was the lowest (< 20% across all specialties). The stark heterogeneity suggests that interventions aimed at inappropriate opioid prescribing may need to be tailored for specialties.

The introduction of report cards in Philadelphia County illustrates the potential trade-offs involved in state- vs county-level initiatives. On the one hand, interventions may be slower to implement by states given the increased number of stakeholders and competing priorities. The smaller size of local governments means fewer stakeholders, many with existing relationships, thus making quality improvement interventions easier to implement. For example, CBH was already a member of the Philadelphia Opioid Task Force in 2017, which gave it ready access to public health officials.

On the other hand, state-level initiatives benefit from larger budgets and data infrastructure that can support the development and sustainability of programs. The use of the PDMP to create provider report cards would allow for prescriptions from multiple insurers to be used. Perhaps the best scenario is a partnership between state and local organizations, an approach that was used by Pennsylvania Medicaid during the development of a dashboard for antipsychotic prescribing to youth in foster care.

 In this program, children with inappropriate prescriptions are flagged by the state, then followed up by the county’s Medicaid MCO as part of its quality improvement efforts.

Some limitations are worth noting. Report cards included only prescriptions filled and reimbursed by Philadelphia Medicaid, and some specialties had few providers, which can be problematic for peer comparison. In addition, a small number of providers claimed that their specialty was inaccurate (although regulations require providers to update the NPPES within 30 days of any changes, it appeared that some providers had not done so

). Also, some measures involved multiple providers, such as concurrent opioid-benzodiazepine prescribing, which can undermine individualized interventions. For example, nurse practitioners or physician assistants may have been the designated prescribers for physicians (in Pennsylvania, nurse practitioners must adhere to a written collaborative agreement with physicians to prescribe

). Finally, the report cards issued to prescribers had no financial incentives but were meant to increase awareness and provide a comparison that would guide further practice.

Critical questions remain, including around the report cards’ effectiveness. An evaluation using neighboring counties as a comparison group is in progress. There are also questions of cost-effectiveness, although report cards are relatively low cost compared with other types of performance feedback, which can justify even small changes in prescribing outcomes.

 Other potential consequences include substitution to nonopioid pain medications or improper tapering from opioids and benzodiazepines that precipitates withdrawal, which can be dangerous to patients.

There is also the COVID-19 epidemic to consider, which caused major disruptions as providers drastically limited face-to-face visits.

 COVID-19 may have exacerbated the opioid epidemic due to the dramatic drop in health care utilization, including the provision of medication-assisted treatment and the loss of employer-sponsored insurance.

 There are concerns that a perfect storm is looming as impending state budget shortfalls put additional pressure on public-sector programs to provide services such as substance use disorder treatment.

 As with the opioid epidemic, COVID-19 requires tailored responses to prevention, testing, and vaccination, which may benefit from existing community quality collaboratives. Medicaid MCOs are in a unique position to help municipalities respond, given their access to rich, up-to-date patient-level information.

In the meantime, inappropriate opioid prescribing remains a top concern and municipalities are taking novel approaches to address it. The Philadelphia Medicaid Opioid Prescribing Initiative, a community quality collaborative involving the Philadelphia Department of Public Health and Medicaid MCOs, demonstrates the potential value of these approaches.

 Center for Mental Health, University of Pennsylvania (SS, MC, DSM, OO, AR), Philadelphia, PA; Community Behavioral Health (OF, KK, CT), Philadelphia, PA; Department of Behavioral Health and Intellectual disAbility Services (SL), Philadelphia, PA; Community Care Behavioral Health Organizations (GN), Pittsburgh, PA; Philadelphia Department of Public Health (LNP, RW), Philadelphia, PA; School of Social Policy and Practice, University of Pennsylvania (AR), Philadelphia, PA.

 Ms Shen, Dr Candon, and Dr Rothbard received salary support from a services contract with Community Behavioral Health and Penn Center for Mental Health. Dr Neimark owns shares of Pfizer. Ms Shen, Dr Candon, and Dr Rothbard are employed by the Center for Mental Health, which receives funds to perform QA and evaluation. The remaining authors report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.

 Concept and design (SS, MC, OF, KK, DSM, GN, LNP, RW, AR); acquisition of data (KK, SL, CT, RW); analysis and interpretation of data (SS, MC, OF, LNP, RW, AR); drafting of the manuscript (SS, MC, OF, DSM, RW, AR); critical revision of the manuscript for important intellectual content (SS, MC, OF, SL, GN, OO, LNP, CT, RW, AR); statistical analysis (SS); provision of patients or study materials (OF); obtaining funding (CT); administrative, technical, or logistic support (SS, MC, OF, KK, SL, DSM, GN, OO, CT, RW); and supervision (MC, OF, GN, OO, CT, AR).

 Siyuan Shen, MCIT, Center for Mental Health, University of Pennsylvania, 3535 Market St, 3rd Fl, Rm 3087, Philadelphia, PA 19104. Email: Lisa.Shen@pennmedicine.upenn.edu.

1. Whelan A. How Philly plans to combat the nation’s worst big-city opioid crisis in 2020. 

January 21, 2020. Accessed January 22, 2021. https://www.inquirer.com/health/opioid-addiction/whats-next-2020-opioid-crisis-overdoses-philadelphia-kensington-20200121.html

2. Underlying cause of death 1999-2017. CDC WONDER Online Database. 2018. Accessed January 22, 2021. https://wonder.cdc.gov/Deaths-by-Underlying-Cause.html

3. Winkelman TNA, Chang VW, Binswanger IA. Health, polysubstance use, and criminal justice
involvement among adults with varying levels of opioid use. 

. 2018;1(3):e180558. doi:10.1001/jamanetworkopen.2018.0558

4. Patrick SW, Davis MM, Lehmann CU, Cooper WO. Increasing incidence and geographic distribution of neonatal abstinence syndrome: United States 2009 to 2012. 

2015;35(8):650-655. doi:10.1038/jp.2015.36

5. Radel L, Baldwin M, Crouse G, Ghertnew R, Waters A. Substance use, the opioid epidemic, and the child welfare system: key findings from a mixed methods study. Office of the Assistant Secretary for Planning and Evaluation, HHS. March 7, 2018. Accessed January 22, 2021. https://aspe.hhs.gov/sites/default/files/migrated_legacy_files//179966/SubstanceUseChildWelfareOverview.pdf

6. Edlund MJ, Martin BC, Russo JE, Devries A, Brennan Braden J, Sullivan MD. The role of opioid prescription in incident opioid abuse and dependence among individuals with chronic noncancer pain: the role of opioid prescription. 

. 2014;30(7):557-564. doi:10.1097/AJP.0000000000000021

7. Muhuri PK, Gfroerer JC, Davies MC. Associations of nonmedical pain reliever use and initiation of heroin use in the United States. Substance Abuse and Mental Health Services Association. August 2013. Accessed January 22, 2021. https://www.samhsa.gov/data/sites/default/files/DR006/DR006/nonmedical-pain-reliever-use-2013.htm

8. U.S. opioid dispensing rate maps. CDC. Accessed October 12, 2018. https://www.cdc.gov/drugoverdose/maps/rxrate-maps.html

9. Lankenau SE, Teti M, Silva K, Bloom JJ, Harocopos A, Treese M. Initiation into prescription opioid misuse amongst young injection drug users. 

. 2012;23(1):37-44. doi:10.1016/j.drugpo.2011.05.014

10. Liu Y, Logan JE, Paulozzi LJ, Zhang K, Jones CM. Potential misuse and inappropriate prescription practices involving opioid analgesics. 

11. Menu of state prescription drug identification laws. CDC. Accessed December 12, 2018. https://www.cdc.gov/phlp/docs/menu-pdil.pdf

12. Technical assistance guide: history of prescription drug monitoring programs. Prescription Drug Monitoring Program Training and Technical Assistance Center. March 2018. Accessed December 12, 2018. https://www.pdmpassist.org/pdf/PDMP_admin/TAG_History_PDMPs_final_20180314.pdf

13. Shaller D, Kanouse D. Private “performance feedback” reporting for physicians: guidance for community quality collaboratives. Agency for Healthcare Research and Quality publication No. 13-0004. November 2012. Accessed December 12, 2018. https://www.ahrq.gov/sites/default/files/publications/files/privfeedbackgdrpt.pdf

14. Faber M, Bosch M, Wollersheim H, Leatherman S, Grol R. Public reporting in health care: how do consumers use quality-of-care information? a systematic review. 

47(1):1-8. doi:10.1097/MLR.0b013e3181808bb5

15. McNamara P, Shaller D, De La Mare D, Ivers N. Confidential physician feedback reports: designing for optimal impact on performance. Agency for Healthcare Research and Quality publication No. 16-0017-EF. March 2016. Accessed December 12, 2018. https://www.ahrq.gov/sites/default/files/publications/files/confidreportguide_0.pdf

16. Campanella P, Vukovic V, Parente P, Sulejmani A, Ricciardi W, Specchia ML

 The impact of public reporting on clinical outcomes: a systematic review and meta-analysis. 

2016;16:296. doi:10.1186/s12913-016-1543-y

17. Ivers N, Jamtvedt G, Flottorp S, et al. Audit and feedback: effects on professional practice and healthcare outcomes. 

. 2012;(6):CD000259. doi:10.1002/14651858.CD000259.pub3

18. Soumerai SB, McLaughlin TJ, Avorn J. Improving drug prescribing in primary care: a critical analysis of the experimental literature. 

19. Technical assistance guide: prescriber report cards. Prescription Drug Monitoring Program Training and Technical Assistance Center. March 15, 2016. Accessed July 10, 2017. http://2m2q501o69ol2wiszw2thkas-wpengine.netdna-ssl.com/wp-content/uploads/2016/11/Report_Card_TAG_20160315_final.pdf

20. 2016 QRUR and 2018 value modifier. CMS. Updated July 26, 2019. Accessed January 1, 2020. https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment/PhysicianFeedbackProgram/2016-QRUR

21. Gadbois EA, Gordon SH, Shield RR, Vivier PM, Trivedi AN. Quality management strategies in Medicaid managed care: perspectives from Medicaid, plans, and providers. 

. 2021;78(1):36-47. doi:10.1177/1077558719841157

22. Managed care in Pennsylvania. Medicaid.gov. Accessed January 20, 2020. https://www.medicaid.gov/medicaid-chip-program-information/by-topics/delivery-systems/managed-care/downloads/pennsylvania-mcp.pdf

23. Leichtling GJ, Irvine JM, Hildebran C, Cohen DJ, Hallvik SE, Deyo RA. Clinicians’ use of prescription drug monitoring programs in clinical practice and decision-making. 

. 2017;18(6):1063-1069. doi:10.1093/pm/pnw251

24. Achieving Better Care by Monitoring All Prescriptions Program (ABC-MAP) Act – Enactment. Pennsylvania General Assembly. October 27, 2014. Accessed December 12, 2018. https://www.legis.state.pa.us/cfdocs/legis/li/uconsCheck.cfm?yr=2014&sessInd=0&act=191

25. Schmitz A. Benzodiazepine use, misuse, and abuse: a review. 

. 2016;6(3):120-126. doi:10.9740/mhc.2016.05.120

26. About CDC’s Opioid Prescribing Guideline. CDC. February 17, 2021. Accessed March 30, 2021. https://www.cdc.gov/drugoverdose/prescribing/guideline.html

27. Orgera K, Tolbert J. The opioid epidemic and Medicaid’s role in facilitating access to treatment. Kaiser Family Foundation. May 24, 2019. Accessed January 22, 2021. https://www.kff.org/medicaid/issue-brief/the-opioid-epidemic-and-medicaids-role-in-facilitating-access-to-treatment/

28. Bardach NS, Lin GA, Wade E, et al. All-payer claims databases measurement of care: systematic review and environmental scan of current practices and evidence. Agency for Healthcare Research and Quality. June 2017. Accessed January 22, 2021. https://www.ahrq.gov/sites/default/files/publications/files/envscanlitrev.pdf

29. Chou R, Turner JA, Devine EB, et al. The effectiveness of long-term opioid therapy for chronic pain: a systematic review for a National Institutes of Health Pathways to Prevention Workshop. 

. 2015;162(4):276-286. doi:10.7326/M14-2559

30. Malone M, Zlotnik S, Miller D, Kreider A, Rubin D, Noonan K. Psychotropic medication use by Pennsylvania children in foster care and enrolled in Medicaid: an analysis of children ages 3-18 years. Children’s Hospital of Philadelphia PolicyLab. 2015. Accessed December 12, 2018. https://policylab.chop.edu/sites/default/files/Psychotropic_Medication_Use_by_PA_Children_in_Foster_Care_and_Medicaid_Spring_2015.pdf

31. CMS, HHS. HIPAA administrative simplification: standard unique health identifier for health care providers; final rule. 

32. Pennsylvania state policy fact sheet. American Association of Nurse Practitioners. 2019. Accessed January 1, 2020. https://www.aanp.org/advocacy/state/state-practice-environment/state-policy-fact-sheets/pennsylvania-state-policy-fact-sheet

33. Flottorp SA, Jamtvedt G, Gibis B, McKee M. Using audit and feedback to health professionals to improve the quality and safety of health care. World Health Organization. 2010. Accessed December 12, 2018. https://www.euro.who.int/__data/assets/pdf_file/0003/124419/e94296.pdf

34. Haley DF, Saitz R. The opioid epidemic during the COVID-19 pandemic. 

. 2020;324(16):1615-1617. doi:10.1001/jama.2020.18543

35. Vaccaro AR, Getz CL, Cohen BE, Cole BJ, Donnally CJ III. Practice management during the COVID-19 pandemic. 

. 2020;28(11):464-470. doi:10.5435/JAAOS-D-20-00379

36. Acs G, Karpman M. Employment, income, and unemployment insurance during the COVID-19 pandemic. Urban Institute. June 2020. Accessed January 22, 2021. https://www.urban.org/sites/default/files/publication/102485/employment-income-and-unemployment-insurance-during-the-covid-19-pandemic.pdf

37. Temporary enhanced federal Medicaid funding can soften the economic blow of the COVID-19 pandemic on states, but is unlikely to fully offset state revenue declines or forestall budget shortfalls. News release. Kaiser Family Foundation; July 22, 2020. Accessed January 22, 2021. https://www.kff.org/coronavirus-covid-19/press-release/temporary-enhanced-federal-medicaid-funding-can-soften-the-economic-blow-of-the-covid-19-pandemic-on-states-but-is-unlikely-to-fully-offset-state-revenue-declines-or-forestall-budget-shortfalls/

Articles

The Implementation of Opioid Prescribing Report Cards in Medicaid Managed Care: A Community Quality Collaborative

 The Pennsylvania Chronic Care Initiative (CCI) was a statewide patient-centered medical home (PCMH) initiative implemented from 2008 to 2011. This study examined whether the CCI affected utilization and costs for HIV-positive Medicaid patients with both medical and behavioral health comorbidities.

302 HIV-positive Medicaid patients treated in 137 CCI practices and 2577 HIV-positive Medicaid patients treated elsewhere.

 All patients had chronic medical conditions (diabetes, chronic obstructive pulmonary disease, asthma, or congestive heart failure) and a psychiatric and/or substance use disorder. Analyses used Medicaid claims data to examine changes in total per patient costs per month from 1 year prior to 1 year following an index episode. Propensity score weighting was used to adjust for potential sample differences. Secondary outcomes included costs and utilization of emergency department, inpatient, and outpatient/pharmacy services.

 We identified an average total cost savings of $214.10 per patient per month (

 = .002) for the CCI group relative to the non-CCI group. This was a function of decreased inpatient medical (—$415.69; 

 = .007) and outpatient substance abuse treatment (—$4.86; 

 = .001) costs, but increased non-HIV pharmacy costs ($158.43; 

= .001). Utilization for the CCI group, relative to the non-CCI group, was correspondingly decreased for inpatient medical services (odds ratio [OR], 0.619; 

 = .002) and inpatient services overall (OR, 0.404; 

 = .001), but that group had greater numbers of outpatient medical service claims when they occurred (11.7%; 

 = .003) and increased non-HIV pharmacy claims (9.7%; 

 There was increased outpatient service utilization, yet relative cost savings, for HIV-positive Medicaid patients with medical and behavioral health comorbidities who were treated in PCMHs.

The Pennsylvania Chronic Care Initiative (CCI) was a patient-centered medical home (PCMH) initiative implemented from 2008 to 2011. This study examined whether the CCI affected utilization and costs for HIV-positive Medicaid patients with both medical and behavioral health comorbidities compared with similar patients treated in non-CCI practices.

Relative cost savings are evident for HIV-positive patients seen in PCMHs due to reduced inpatient medical and outpatient substance abuse treatment costs.

HIV-positive patients seen in a PCMH had higher outpatient healthcare service utilization and an increased number of non-HIV pharmacy claims.

Implementation of a PCMH for HIV-positive patients should be targeted to those with behavioral health and medical comorbidities.

Individuals with HIV have increased rates of depression, substance use, and other serious psychiatric disorders.

 A variety of medical comorbidities, including diabetes,

 are also common. Moreover, recent evidence suggests that psychiatric and substance use disorders may be especially important risk factors for cardiovascular and other medical comorbidities among HIV-positive individuals.

Within the general population, having more than 1 chronic disease is associated with increased mortality, poor functional status, decreased quality of life, unnecessary hospitalizations, and increased medical costs.

 The presence of psychiatric and substance use disorders in those with HIV has also been linked to increased healthcare utilization.

 Despite the prevalence and importance of both medical and psychiatric comorbidities among HIV-positive patients, comorbid disorders often go unrecognized and untreated.

The patient-centered medical home (PCMH) has the potential to identify and address comorbidity through core elements such as a whole-person orientation, behavioral health integration, emphasis on quality, enhanced access, use of health information technology, and coordination of outpatient care.

 Although evidence for cost savings and improved clinical outcomes has been mixed when PCMH models have been applied in the general population,

 implementation of a PCMH model for those with chronic diseases, such as diabetes, hypertension, and coronary heart disease, and high levels of comorbidity has been associated with relatively better clinical outcomes and reduced total healthcare costs.

The federal Ryan White HIV/AIDS Program brings elements of a medical home to those with HIV by providing funding for primary medical care and additional support services. However, despite being targeted toward those who do not have sufficient healthcare coverage or financial resources to cope with HIV/AIDS, Ryan White funding does not apply to those with Medicaid coverage.

Beginning in 1997, Pennsylvania introduced HealthChoices, a risk-based managed care program that was initially offered in certain counties and then expanded to cover more counties and services. By mid-2009, 72% of all Pennsylvania Medicaid beneficiaries were enrolled in some form of managed care. Enrollment became mandatory for most Medicaid beneficiaries by 2013 in the counties where the program operates.

During 2008 to 2011, the state conducted one of the largest statewide multipayer PCMH experiments in the United States. This experiment, the Pennsylvania Chronic Care Initiative (CCI),

 was in place from 2008 to 2011. The Pennsylvania Department of Health provided leadership and financial support for practice transformation, requiring CCI practice managers to attend learning sessions, report monthly quality metrics, and use assigned practice coaches. These strategies were specifically based on the Chronic Care Model.

 All CCI sites received behavioral health training, integrated depression screening using validated tools, and were coached to develop a process for coordinating behavioral and medical care. Practices were selected for participation in the CCI through a voluntary application process. The PCMH model was implemented for all patients receiving services at participating practices. In total, 152 primary care practices involving 640 providers participated in the CCI, with more than 1.18 million patients receiving care.

Evaluations of the CCI using all-payer claims data have been mixed.

 Recent analyses found reductions in emergency department (ED) use overall, as well as reduced cost and hospitalization among high-risk patients with chronic conditions.

 Prior work suggests that Medicaid patients with medical and psychiatric comorbidities may benefit disproportionately from the structural advantages of the PCMH model due to the complexity of managing their illnesses, generating significantly lower costs.

We previously reported that reductions in healthcare utilization and costs were evident for Medicaid patients treated with the PCMH model implemented in CCI practices.

 However, no analysis to date has determined whether such cost savings would apply to HIV-positive patients with medical and psychiatric comorbidities. Patients with HIV were a small component of the matched samples examined in the previous study of cost and utilization outcomes in the CCI,

 representing only 2.3% of that sample. A larger group of HIV-positive patients was available for analysis in the Pennsylvania Medicaid database, with 17.5% of such patients included in the previous matched sample analyses. Thus, it was not clear if findings for a non-HIV sample would generalize to an HIV-positive sample.

The current study tested the hypothesis that significant reductions in healthcare utilization and costs would be evident for HIV-positive patients with medical and psychiatric comorbidities who were treated in a CCI practice compared with similar patients treated in a non-CCI practice.

We examined pre-post healthcare utilization and costs for HIV-positive Medicaid patients with at least 1 of 4 chronic medical conditions, plus at least 1 psychiatric and/or substance use disorder, comparing changes in healthcare utilization and costs among patients treated in CCI and non-CCI practices. This study was approved by the University of Pennsylvania Institutional Review Board with a waiver of informed consent and complied with the ethical standards of the Office for Human Research Protections.

We initially obtained a list of 147 CCI practices from the Pennsylvania Department of Health. Identifying information for 5 of the 152 practices in the CCI was not in the list provided. Of the remaining 147, we excluded 10 that did not serve Medicaid patients, reducing the sample to 137 CCI practices. Using a Medicaid claims dataset obtained from the Pennsylvania Department of Human Services, we identified HIV-positive patients treated in the CCI practices and those treated in non-CCI practices during the time period the CCI was active. We identified HIV-positive patients with an 

International Classification of Diseases, Ninth Revision

 code of 042 (attached to at least 2 claims) who also had primary or secondary diagnostic codes for at least 1 of 4 chronic medical conditions (in any type of claim) and at least 1 claim (outpatient or inpatient nonlaboratory) for a psychiatric and/or substance use disorder. The 4 comorbid chronic medical conditions (diabetes, chronic obstructive pulmonary disease [COPD], asthma, and congestive heart failure [CHF]) were selected because they were identified as sources of disproportionate health and financial burdens by the Pennsylvania Health Care Cost Containment Council.

 Comorbid behavioral health conditions included psychiatric (major depressive disorder, schizophrenia/schizoaffective disorder, bipolar disorder, posttraumatic stress disorder, and anxiety disorders) and substance use (opioid, cocaine, and alcohol) disorders. The psychiatric disorders chosen were the most prevalent and are associated with high healthcare costs within a Medicaid population.

 The substance use disorders chosen were the 3 most common nationally (other than marijuana) for patients presenting at substance use treatment facilities during the target years.

HIV-positive patients were considered part of the CCI group if they had at least 1 claim from a CCI practice. Patients with HIV who never had a claim from a CCI practice were placed in the comparison group. For each CCI patient, the first claim filed after the date the patient’s practice joined the CCI was identified as the “index episode.” For patients in the comparison (non-CCI) group, the index episode was defined as the first claim filed after commencement of the CCI program. In the original sample, there were 404 CCI patients and 4039 non-CCI patients diagnosed as HIV positive. To be included in the study, a patient needed to have at least 6 months of Medicaid eligibility during the year prior and the year following the index episode. This restriction resulted in a sample of 302 CCI HIV-positive patients and 2577 non-CCI HIV-positive patients. The date of the index episode was used to mark the first exposure of each patient to the “intervention” of the CCI, allowing for a pre-post intervention comparison, as has been done in other studies.

Patients eligible for both Medicaid and Medicare were included, as many patients with chronic mental health and substance abuse conditions have such dual eligibility. For the target sample, we identified the presence of both chronic medical conditions and behavioral health diagnoses from claims at any point during the 2005-2010 time period, whether before or after the index episode date. This was done because the medical and psychiatric disorders targeted in this study typically are chronic conditions that manifest first as subthreshold symptoms and, if managed well, can be prevented from evolving into full diagnoses.

Costs were calculated using standardized prices for Medicaid claims. Outpatient costs were standardized using the Medicaid outpatient fee schedule, regardless of fee-for-service or capitation. Notably, capitated Medicaid managed care plans also submit claims for provided services. Pharmacy costs were based on fee-for-service Medicaid pharmacy costs specified for each Hierarchical Ingredient Code. To standardize inpatient costs, we computed 2008 Pennsylvania Medicaid average costs by diagnosis-related groups using fee-for-service data. We separately calculated pharmacy costs for antiretroviral drugs and other medications.

We compared pre-post changes in healthcare utilization and costs for HIV-positive patients treated at CCI practices with changes in utilization and costs for non-CCI HIV-positive patients in the same year. Changes in utilization and costs were measured from the 1-year pre—index episode period to the 1-year post–index episode period, with costs and counts calculated per month eligible for Medicaid.

The primary outcome measure was the difference between CCI and non-CCI patients in total healthcare cost changes per month eligible between the 1-year preindex period and the 1-year postindex period. Secondary outcomes included between-group differences in pre-post changes within specific ED, inpatient, pharmacy, and outpatient cost and utilization variables. Negative values for differences favored the CCI group over the non-CCI group and vice versa for positive values.

A propensity score was derived to address the potential lack of comparability of patients in CCI and non-CCI practices at the time of treatment initiation using variables listed in 

. Other variables included in the propensity score (not listed in Table 1) were disability status or Supplemental Security Income benefits, months of Medicaid eligibility, dual Medicaid and Medicare eligibility, year of index episode, and region within Pennsylvania. The propensity score was used as an inverse weighting factor in all analyses.

 Covariates included in all models were pre—index year total costs for each patient, pre–index year total utilization for each patient, and year of preindex episode.

Primary analysis of the cost variables consisted of a weighted least squares model of difference-in-differences (DID) scores, including the covariates described above and the propensity score as a weighting factor. Analyses of healthcare utilization counts proceeded using either generalized Poisson (GP) models, zero-inflated GP models, negative binomial regression models, or zero-inflated negative binomial models with SAS PROC-GENMOD, PROC-GLIMMIX, and PROC-NLMIXED.

 The choice between zero-inflated and noninflated models depended on the magnitude of the zero counts, assessed using Vuong’s test.

 The choice between GP and negative binomial was a function of whether there was underdispersion or overdispersion of the data, respectively. For the zero versus nonzero component of the zero-inflated models, differences between the CCI and non-CCI groups were described using odds ratios (ORs) with 95% CIs. For the GP, negative binomial regression, and count portion of the zero-inflated models, the exponential of the regression coefficient for the comparison between CCI and non-CCI cost differences was interpreted as the percent increase or decrease in the expected count for CCI compared with non-CCI.

As an assessment of whether the propensity procedure successfully balanced the groups, we implemented the goodness-of-fit diagnostic steps described by Austin.

 First, we derived and visually examined quintile side-by-side boxplots of the propensity score for the CCI and non-CCI groups. Finally, we derived the weighted conditional standardized difference for each of the baseline predictors in Table 1 and compared each with the unconditional standardized difference.

Table 1 provides details on demographics and health conditions for HIV-positive patients with medical and psychiatric comorbidities treated in CCI versus non-CCI practices. Although the CCI group had higher percentages of patients who were African American, higher baseline comorbidity scores, and higher prevalence of cocaine, opioid, and alcohol use disorders, there were no statistically significant between-group differences on any of these characteristics after weighting by propensity score. Average (SD) per patient total healthcare costs during the 1-year preindex period per month eligible were $2721.29 ($6323.52) for the non-CCI group and $2951.48 ($4854.14) for the CCI group (

= .43, without adjusting for propensity score). The results of the weighted conditional standardized difference analyses revealed that for most variables in Table 1, except presence of any psychiatric disorder, the weighted conditional standardized difference was smaller than the standardized difference. For presence of any psychiatric disorder, the standardized difference was 0.026 and the weighted conditional standardized difference was 0.091. Although the weighted conditional standardized difference was slightly larger than the standardized difference, it still fell below the recommended 0.10 threshold for concern about a potential residual imbalance between the treatment groups.

 Visual assessment of the balance of propensity scores using quintile plots showed adequate balance between CCI and non-CCI over the quintiles. We conclude that the CCI and non-CCI groups were adequately balanced following the propensity adjustment.

The adjusted mean total cost DID scores for the CCI group relative to the non-CCI group were —$214.10 (95% CI, –$345.65 to –$82.55) per patient per eligible month (

). This effect did not vary significantly by region of the state (region by intervention group interaction, 

 = .14). The most significant contributors to this cost savings were relatively greater decreases in inpatient medical costs (—$415.69 per patient per eligible month; 

 = .007) and outpatient substance abuse treatment costs (—$4.86; 

 = .001) for the CCI group compared with the non-CCI group. Exploratory analyses examined the CCI versus non-CCI difference in outpatient substance abuse treatment costs separately for the subgroups of patients with and without a substance use disorder. The overall effect was carried by the patients with a substance use disorder (n = 141 and 940 for CCI and non-CCI groups, respectively), with a relative reduction in costs of $12.25 (95% CI, —$17.57 to –$6.93; 

 <.001) for the CCI group compared with the non-CCI group. There was no significant difference in outpatient substance abuse treatment costs for patients without a substance use disorder diagnosis.

The CCI group increased outpatient costs significantly from the pre- to postindex periods relative to the non-CCI group ($181.54; 95% CI, $68.94-$294.14; 

 = .002), with this effect driven by greater increases in non-HIV pharmacy costs ($158.43; 95% CI, $68.42-$248.44; 

 =.001). To further understand this increase in non-HIV pharmacy costs, these costs were subdivided into costs for medications for substance use disorder, psychotropic medications, and other medical non-HIV medication costs. No significant differences between the CCI and non-CCI groups were evident for psychotropic or substance use medications; medical (non-HIV) pharmacy costs were significantly higher for the CCI group compared with the non-CCI group ($153.94; 95% CI, $65.94-$241.91; 

Relative reductions in healthcare utilization for CCI compared with non-CCI patients were evident for inpatient services (

). For inpatient services of any type, the CCI group had a reduction in any usage from the pre- to postindex period, whereas the non-CCI group had an increase in any claims (OR, 0.404; 95% CI, 0.280-0.575; 

 = .001). The most significant contributor to this effect was inpatient medical services, for which the CCI group experienced a reduction in any services, whereas the non-CCI group had an increase in any claims (OR, 0.619; 95% CI, 0.446-0.837; 

 = .002). With respect to outpatient medical claims, there was a significantly greater increase in utilization for CCI patients compared with non-CCI patients when outpatient medical claims occurred, with the average difference in number of claims per month eligible increased by 11.7% (95% CI, 3.9%-20.3%; 

 = .003) for CCI compared with non-CCI patients. Additionally, there was a significant increase in pharmacy claims for CCI compared with non-CCI patients, with the average number of claims per month eligible increasing by 8.0% (95% CI, 3.2%-13.1%; 

 = .0009) for the CCI group relative to the non-CCI group. This increase was driven by significantly greater pharmacy claims for non-HIV medications for CCI compared with non-CCI patients (9.7%; 95% CI, 4.6%-15.1%; 

Among HIV-positive Medicaid patients with comorbid medical conditions (asthma, COPD, CHF, and/or diabetes) and psychiatric and/or substance use disorders, our data indicate that Pennsylvania’s Chronic Care medical home initiative resulted in substantial cost savings compared with non-CCI treatment in the state. Relative decreases in cost were apparent for outpatient substance abuse treatment, inpatient medical, and total inpatient services. Decreases in utilization of any inpatient medical services were evident. These decreases in costs and utilization occurred with concomitant increases in the cost and utilization of outpatient medical services and outpatient non-HIV pharmacy claims for CCI-treated patients relative to non-CCI patients. The CCI intervention apparently was successful in shifting inpatient costs and utilization to outpatient care and use of non-HIV medications.

Studies examining the PCMH model in general (non-HIV) medical populations have reported mixed effects in terms of both costs/utilization and clinical outcomes,

 although many of these studies have had methodological weaknesses.

 Initial smaller-sample reports on the PCMH model implemented in the Pennsylvania CCI have also been mixed.

Among HIV-positive Medicaid patients with comorbid medical and psychiatric disorders, there was increased outpatient service utilization, yet relative cost savings, for patients who were treated in patient-centered medical homes.

Aileen Rothbard, ScD

Articles

The Implementation of Opioid Prescribing Report Cards in Medicaid Managed Care: A Community Quality Collaborative

Impact of a Medical Home Model on Costs and Utilization Among Comorbid HIV-Positive Medicaid Patients