Applying Population Health Management to Opiate Prescription Medication Misuse Joe Parks, MD National Council Senior Medical Advisor April 13, 2016
My Background Medicaid Director Previously DMH Medical Director 20 years Practicing Psychiatrist CMHCs 10 years FQHC 18 years Distinguished Professor, Missouri Institute of Mental Health, University of Missouri St. Louis Adjunct Professor of Psychiatry University of Missouri Columbia
Outline What Is Population health What is population health management Why do we need it for Persciption Drug Abuse? Missouriexample
Population Health Definitions The health of the population as measured by health status indicators and as influenced by social, economic and physical environments, personal health practices, individual capacity and coping skills, human biology, early childhood development, and health services (Dunn and Hayes, 1999). A conceptual framework for thinking about why some populations are healthier than others as well as the policy development, research agenda, and resource allocation that flow from it (Young, 2005).
Population Management Principles Population-based care Data-driven care Evidence-based care Patient-centered care Addressing social determinates of health Team care Integration of behavioral and primary care
Population-Based Care Don't rely solely on patients to know when they need care, what care to ask for and from whom use data analytics for outreach on high need/utilizer patients Don't focus on fixing all care gaps one patient at a time - Choose selected high prevalence and highly actionable individual care gaps for intervention across the whole population The population-based health care provider is the public health agency for their clinic population
Data-Driven Care Patient Registries Data Sharing Risk Stratification Performance Benchmarking Predictive Analytics
How do you deliver PHM? Assess Stratify Implement Solutions Measure & Report
Population Management Selects those from whole population: Most immediate risk Most improvement opportunities Aids in planning: Care for whole population New interventions and programs Early identification and prevention Choosing and targeting health education
Data Uses Individual drill-down care coordination Aggregate reporting performance benchmarking Disease registry care management Identify care gaps Generate to-do lists for action Understanding planning and operations Telling your story presentation like this
Principles Use the Data you have before collecting more Show as much data as you can to as many partners as you can as often as you can: Sunshine improves data quality They may use it to make better decisions It s better to debate data than speculative anecdotes When showing data, ask partners what they think it means Treat all criticisms that results are inaccurate or misleading as testable hypotheses
More Principles Tell your data people that you want the quick easy data runs first. Getting 80% of your request in one week is better than 100% in six weeks Treat all data runs as initial rough results Important questions should use more than one analytic approach Several medium data analytic vendors/sources is better than one big one Transparent bench marking improves attention and increases involvement
Most Important Principles Perfect is the enemy of good Use an incremental strategy If you try to figure out a comprehensive plan first, you will never get started Apologizing for a failed prompt attempt is better than apologizing for a missed opportunity
Increases Overdose Deaths Methadone is reported by the Centers for Disease Control and Prevention to be involved in 30 percent of prescription overdose deaths CDC also reports that the death rate from methadone overdoses was 6 times higher in 2009 than in 1999. While buprenorphine abuse and overdose deaths are much rarer, they are rapidly increasing in number.
An Epidemic of Opioid Misuse w w w. T h e N a t i o n a l C o u n c i l. o r g Drug overdose death rates in the US have more than tripled since 1990 In 2008 more than 36,000 people died from drug overdoses 42 people die everyday from prescription painkiller overdose in US Most of those deaths were caused by prescription drugs 16
w w w. T h e N a t i o n a l C o u n c i l. o r g Most Cases of Opioid Misuse Start with a Doctor s Prescription 17
Opioid Misuse Varies Regionally w w w. T h e N a t i o n a l C o u n c i l. o r g 18
Risk Factors for Opioid Misuse: Demographic w w w. T h e N a t i o n a l C o u n c i l. o r g Young age (18-25) Misuse rising among elderly Male Rural area Living in south, Maine, New Hampshire Low income, unemployment 19
Risk Factors for Opioid Misuse: Clinical w w w. T h e N a t i o n a l C o u n c i l. o r g Previous or current psychiatric illness Previous alcohol or other substance abuse Benzodiazepine use 20
Risks for Misuse of Opioids: Doctor Shopping w w w. T h e N a t i o n a l C o u n c i l. o r g Multiple pharmacies Multiple prescribers Lost pills or prescriptions Frequent requests for dose increase Frequent requests for early refill 21
Risk for Misuse of Opioids: Prescriber Issues w w w. T h e N a t i o n a l C o u n c i l. o r g Influenced by patient request Inadequate training Ignore PMPs REF: McKinlay et al, Med Care, 2014 22
Evidence-based Use of Opioids w w w. T h e N a t i o n a l C o u n c i l. o r g Control of post-traumatic pain Control of post-surgical pain Treatment of pain in patients with malignant cancer Palliative, end-of-life care Treatment of opioid addiction (methadone & buprenorphine) 23
Contraindications of Opioid Use w w w. T h e N a t i o n a l C o u n c i l. o r g Migraine headache Cough suppression 24
Uncertain Use of Opioid Analgesics w w w. T h e N a t i o n a l C o u n c i l. o r g Chronic non-malignant pain Defined as: pain lasting more than three months or past the time of normal tissue healing One-third of American adults report chronic pain Neuropathic pain REF: AHRQ, 2013 25
Solutions to Opioid Misuse w w w. T h e N a t i o n a l C o u n c i l. o r g Mandatory PMP use Follow practice guidelines Screen for substance abuse and psychiatric illness Do not co-prescribe benzodiazepines Evaluate prescribing data on a regular basis, including in ED Use Population Data Based Analytics Prescription Monitoring Programs (PMP); CMT Model 26
Role of Data Analytics in Risk Analysis and Treatment Stratification w w w. T h e N a t i o n a l C o u n c i l. o r g Population Stratification Resource Consumption 5% 25% 70% Poly-chronic BH/SA comorbidities Elderly and disabled At risk for major intervention/dependency Healthy/minor issues 45%-50% 30%-35% 20% ED visits Avoidable events Polypharmacy Higher volume of preventable acute episodes Complications and readmissions Unmanaged and unengaged Opportunities for integrative medicine Opportunities to enhance value through greater BH access and engagement Opportunities to reduce/prevent dependency; Improve quality of life Source: Blended MarketScan Commercial, Medicare 5% LDS, and representative payer Medicare data 27
Emerging Evidence for Use of Data to Improve Opioid Dependency/Misuse w w w. T h e N a t i o n a l C o u n c i l. o r g Lowered rates of doctor shopping Improved clinical decision making Increased discussion with patients about pain/opioid use Increased referrals for SA and brief intervention services *Sproule, B. Prescription Monitoring Programs in Canada: Best Practice and Program Review, CCSA, June 2014. 28
Audit and Feedback w w w. T h e N a t i o n a l C o u n c i l. o r g Cochrane Review: More efficacious than academic detailing Best value when: o Includes peer comparisons o Communicated by a peer-verbally and in writing o Targeted goals and action plans o Patient specific information tied to outcomes Reflective of CMT s Core Quality Indicator Approach 29
Applying Population Health to Prescription Drug Abuse Use claims to identify patients who appear to be at high risk for prescription drug abuse Identify prescribers with high portions of their patients at risk for prescription drug abuse Identify high risk patients to the prescribers involve and provide benchmark feedback and recommendations for change Report selected high risk prescribers who did not respond to feedback for regulatory investigation
Applying Population Health Prescription Drug Abuse Pharmacy Data Alone Multiple opiates or benzodiazepines for over 60 days High-dose benzodiazepines Multiple opiates prescribers Pharmacy and Claims Data Multiple pharmacies filling opiates or benzodiazepines SUD Related diagnoses Co-prescribed benzodiazepines and opiates
Evidence-based Decision Support System w w w. T h e N a t i o n a l C o u n c i l. o r g What gets measured, gets done. Big Data Engine Clinical lnsights 32
Emerging Best Practices w w w. T h e N a t i o n a l C o u n c i l. o r g Includes ALL drugs Ensure patient, prescriber, pharmacist linkage in data Full patient profile available Unsolicited reports to stakeholders Standard data collection method Patient privacy safeguards Evaluate intended and unintended consequences Allow for encrypted data for research/outcomes analyses 33
Core Foundation w w w. T h e N a t i o n a l C o u n c i l. o r g Team of Clinical Experts Cohort Studies Published Research & Outcomes Expert Consensus Guidelines 200 Evidence-based BH Algorithms 34
Quality Indicator Overview w w w. T h e N a t i o n a l C o u n c i l. o r g Behavioral Pharmacy 164 Age Banded Child, Adult, Elderly Opioid Pharmacy 56 Age Banded Child, Adult, Elderly 35
CMT s Quality Indicators Common Foundation w w w. T h e N a t i o n a l C o u n c i l. o r g Evidence or consensus based. Involve significant cost and/or health and safety. A small proportion of providers responsible for a large proportion of suspected errors. Compelling empirical support for the indicator. Are actionable. 36
Identifying High Risk Patients w w w. T h e N a t i o n a l C o u n c i l. o r g Using multiple prescriber Using multiple pharmacies Monthly total methadone equivalent of all opiate prescriptions Diagnoses of substance use disorders Diagnoses of factitious disorder, malingering, and somatization 37
Identifying High Risk Prescribers Percent of opioid patients flagged for substance use diagnosis Monthly average number of opiate prescriptions per opioid patient Average daily methadone equivalent of opioid prescriptions to SUD flagged patients
Peer Comparator Unsolicited Report w w w. T h e N a t i o n a l C o u n c i l. o r g 39
Sample Opioid Clinical Consideration w w w. T h e N a t i o n a l C o u n c i l. o r g 40
Overview Original Mailing QIs Use of Buprenorphine with another Opioid (prescribed by another physician). Use of Buprenorphine with a Benzodiazepine (prescribed by another physician). Patient s use of 5 or more prescribers for Opioid prescriptions. Use of Opioids for 60 or more days with a diagnosis suggesting Opioid, alcohol or other substance abuse in the last year. Use of Opioids for 60 or more days with two or more diagnoses of malingering, somatization or factitious disorder. # 41
New Mailing QIs (6/13 Initiation) Following Expansion Use of cough and cold medications containing Opioids - Adult / Child / Elderly Patient's use of 4 or more pharmacies for Opioid Rxs Adult Patient's use of 5 or more prescribers for Opioid Rxs - Child / Elderly Use of Opioids for 60 or More Days in Absence of a Diagnosis Supporting Chronic Use - Adult / Child / Elderly # 42
Overview Mailing Statistics - Overview Mailing Date Patients Providers Phase 2/25/2013 2,627 1,786 4/22/2013 3,219 1,273 Original 5 QIs 6/27/2013 33,780 1,675 8/30/2013 28,422 1,594 10/30/2013 27,238 1,566 12/20/2013 26,093 1,585 2/14/2014 25,900 1,646 5 Original QIs + Expansion QIs 4/21/2014 27,572 1,603 6/20/2014 27,897 1,756 # 43
Overview After mailing summary reports: Following each mailing, summary reports are emailed to Missouri administrative staff. Those reports include: OPI QI Summary (list of QIs with patient and prescriber counts and percentages) High Risk Substance Abuse Patient Report Prescriber Identified High Risk Patients Report Prescribers More Likely to Treat Patients with SA Report OPI Intervention Report (selected QIs and counts of patients and prescribers) OPI CMHC and CMHC-Prescriber Benchmark Reports # 44
Overview Three Month Pre/Post Analysis Original QIs: Estimated savings of $217,034 in opioid pharmacy cost avoidance an average of $20.69 per intervened patient per month for 3,496 individuals eligible for 3 months follow-up. Significant decrease of emergency department visits by 37.84%* and hospital admits decreased by 37.82%* Average patient usage of 37.1%* fewer opioid prescribers and 31.2%* fewer opioid pharmacies Average monthly dose of opioids (in morphine equivalents) dispensed fell 17.9%* Three Month Pre/Post analysis includes MO HealthNet clients from the eight OPI mailing interventions from 2/25/2013 through 4/21/14. * comparing 3 months pre-intervention to 3 months post-intervention p<.001 # 45
Multiple Baseline Regression Analyses Methodology Study included eligible adult/child/elderly first intervened in 2013/2014. Patient/Months in the analysis included where spend is greater than zero or subject is Medicaid-eligible for the entire month. Study excludes subjects who: Were part of a BPM intervention in this report Were included in any BPM or OPI intervention in 2012 Are dual eligibles End date for claims analysis was 8/22/2014. Cost avoidance related to Hospitalizations, ER and Opioid Rx estimate is $40 million. Note: A multiple baseline analysis has multiple cohorts and multiple study periods, hence no single reporting period applies to all cohorts. # 46
Multiple Baseline Regression Analyses Cohort Original QIs # 47 Multiple Baseline Analysis includes multiple cohorts with variable amounts of follow-up based on eligibility and claims activity. Dataset includes up to 12 months of pre/post for each cohort. Eligible Patient- Mailing Date Cohort Total Months After Mailing Date 2/25/2013 788 8,155 4/22/2013 810 8,574 6/27/2013 659 6,894 8/30/2013 403 4,093 10/30/2013 311 2,645 12/20/2013 320 2,216 2/14/2014 306 1,665 4/21/2014 246 921 6/20/2014 328 615 Total 4,171 35,778
Multiple Baseline Analysis FY2014 Original QIs $3,000 Combined Hospital, ER and Opioid Spend PMPM Original 5 QIs Adults PharmacyTotalPaid_Cost_Hosp_Cost_ER (method 2) PRE y = 81.59x + 559.33 $2,500 PharmacyTotalPaid_Cost_Hosp_Cost_ER (method 2) POST PharmacyTotalPaid_Cost_Hosp_Cost_ER (method 2) PROJECTED $2,000 Linear (PharmacyTotalPaid_Cost_Hosp_Cost_ER (method 2) PRE) $1,500 $1,000 $500 $0-11 -10-9 -8-7 -6-5 -4-3 -2-1 0 1 2 3 4 5 6 7 8 9 10 11 12 # 48
Multiple Baseline Analysis FY2014 Original QIs Opioid Spend PMPM Original 5 QIs Adults $250 PharmacyTotalPaid_OPI PRE y = 8.0306x + 31.423 $200 PharmacyTotalPaid_OPI POST PharmacyTotalPaid_OPI PROJECTED $150 Linear (PharmacyTotalPaid_OPI PRE) $100 $50 $0-11 -10-9 -8-7 -6-5 -4-3 -2-1 0 1 2 3 4 5 6 7 8 9 10 11 12 # 49
Original QIs: # 50
Overview New QIs: Expansion started with mailings in June 2013 Substantial increase in mailing volume Approximately 20x as many clients impacted as before per mailing Estimated $454K savings on Opioids for FY 2013-2014 Adult: $7.16 PMPM x 17,142 individuals x 3 Months = $368K Child: $16.54 PMPM x 1,530 individuals x 3 Months = $76K Elderly: $1.36 PMPM x 2,439 individuals x 3 Months = $10K Cost avoidance related to Hospitalizations, ER and Opioid Rx estimate is $15.3 million # 51
New QIs Significant decrease of emergency department visits by 10.2%* for adults and a decrease of 45.3%* for children. Significant decrease in hospitalizations by 43.7%* for children. Average usage of opioid prescribers dropped by 12.6%* for adults, 11.8%* for elderly and 77.6%* for children Average number of pharmacies used to obtain opioids decreased 13.1%* for adults, 77.1%* for children and 10.0%* for elderly. Average monthly dose of opioids (in morphine equivalents) dispensed fell 8.1%* for adults, 52.3%* for children and 11.3%* for elderly. * comparing 3 months pre-intervention to 3 months post-intervention p<.001 # 52