• No results found

General MLIP Outcomes—Unpublished MLIP Evaluation Data 28

In document Roberts_unc_0153D_15527.pdf (Page 48-52)

CHAPTER II: BACKGROUND, LITERATURE REVIEW, AND CONCEPTUAL MODEL 11

2.2   LITERATURE REVIEW—MLIP Outcomes and Evaluations 24

2.2.4   General MLIP Outcomes—Unpublished MLIP Evaluation Data 28

Multiple sources of MLIP evaluation data were identified through comprehensive search of the grey literature. Most of these sources took the form of commissioned reports or presentation materials that

communicate results from internal state MLIP evaluations. First, a recent evaluation of the North Carolina Medicaid Recipient Management Lock-In Program is highlighted. Then, additional unpublished MLIP evidence is reported, grouped by economic, utilization, and clinical MLIP outcomes.

2.2.4.1 General MLIP Outcomes—Internal North Carolina MLIP Report

The State of North Carolina commissioned a report in 2011 to evaluate the impact of their MLIP, implemented in October 2010, on healthcare utilization and outpatient pharmacy costs.105 The evaluation assessed changes in outcome measures between the six month period before the MLIP start date—April through September 2010—and the six month period following—October 2010 through March 2011. Because enrollment of qualifying beneficiaries into the MLIP was steadily phased in at a rate of roughly 200 additional enrollees monthly, this evaluation captured a total of 2,909 person-months of exposure to the MLIP policy. They also utilized a comparator group of beneficiaries meeting the state’s criteria for MLIP enrollment but were not yet enrolled into the program during the study period.

The report estimated the MLIP was associated with $279,500 in total gross outpatient pharmacy savings in the first six months of MLIP implementation.105 Fourteen percent of MLIP enrollees incurred no pharmacy costs to North Carolina Medicaid following their lock-in, which accounted for almost 40% of the estimated savings. (The authors do not speculate about the lack of subject observation in Medicaid claims data following lock-in, but this finding could indicate circumvention behaviors that will be investigated in Aims 1 and 2of this dissertation). A total of 55% of the estimated $279,500 in total pharmacy savings were attributed to reduced claims for MLIP qualifying medications, including opioid analgesics and benzodiazepines. The MLIP resulted in a mean reduction of 2.8 total prescription claims per enrollee and a mean reduction in controlled substance claims per enrollee of 2.0 fills. Utilization of inpatient, outpatient, emergency, and radiology services were unchanged among MLIP enrollees during the post-implementation period. Subsequent internal evaluations found that 2.3 million fewer opioid analgesic tablets were dispensed in the first three months of MLIP implementation, with a total of $5.2 million in savings to the state in its first year.116

2.2.4.2 General MLIP Outcomes—Unpublished Economic Findings

Much of the grey literature—and for that matter, peer-reviewed literature—focus on MLIPs’ economic impact on state expenditures. The State of Washington conducted two large-scale evaluations of its MLIP; one in 2005 that assessed the first three years of the program’s operations when enrollment increased from 200 to 3,000 beneficiaries117 and again in 2012 with a cost-effectiveness analysis when enrollment topped out at 3,800 individuals.118 The state studied a nine month pre-MLIP and six month post-MLIP observation period and reported monthly savings of $1.5 million due to its MLIP117, with a return on investment of $6.61 for each dollar spent administering the MLIP. 118

In addition, presentation materials from Best and colleagues also describe findings from an unpublished randomized controlled trial that was conducted in 2008 and included economic outcomes.118 The study assigned 503 qualifying Washington Medicaid beneficiaries to MLIP enrollment and 188 beneficiaries meeting enrollment criteria to an observation-only control group.118 The researchers employed a difference-in-difference intention to treat analysis looking at changes in outcomes between pre- and post-enrollment observation periods of 12 months. Enrollment in the Washington MLIP was associated with significant decreases in physician, emergency care, planned hospitalization, and total Medicaid costs. No difference was found between MLIP and control groups in total pharmacy costs.

Many other states reported economic savings in internal MLIP evaluations. In 2012, Keast presented results of a pre-post MLIP enrollment analysis among 52 Oklahoma Medicaid beneficiaries who were locked in to a single pharmacy.119 This investigation found significant decreases in pharmacy- and emergency department-related costs. A 1995 audit of Utah’s MLIP found a decrease in total

healthcare costs of 50% for their 22 enrollees. Kentucky MLIP administrators forecasted savings to the state of over $6,300 per lock-in enrollee.120 Iowa Medicaid’s lock-in program saved the state $2 million annually.121 Florida’s MLIP saved the state nearly $13 million over the programs first three years.122 Similar economic findings are available for Connecticut123,124, Illinois125, Massachusetts126, South Carolina127, and West Virginia128.

2.2.4.3 General MLIP Outcomes—Unpublished Utilization Findings

In addition to economic outcomes from MLIPs, much of the unpublished MLIP evidence centers on healthcare utilization measures. And similar to the unpublished economic findings, MLIPs generally resulted in significant decreases in utilization of Medicaid services.

In the 2005 report from the State of Washington, the average number of narcotic prescription claims per enrollee lowered from 3.1 to 1.6, the total dosage of opioid analgesics received by enrollees decreased over 40%, and the average number of prescribers per person reduced from 4.8 to 2.8 following lock-in.117 In 2012, the State of Washington found that, in the six months following an expansion of the MLIP, lock-in enrollees exhibited a 37%, 33%, and 24% decrease in outpatient care, emergency care, and prescription utilization, respectively.118 Best and colleagues also reported significant decreases in

physician visits, emergency care visits, planned hospitalization costs, total narcotic prescription claims, and total narcotic tablets dispensed in the MLIP arm of their unpublished 2008 randomized controlled trial. No significant changes were found for total hospitalizations.

Utilization outcomes from the 2012 Oklahoma MLIP evaluation included significant decreases in total prescription claims, narcotic prescription claims, number of pharmacies used, number of prescribers used, and emergency department visits.119 Interrupted time series analysis showed that the MLIP

enrollment not only caused a reduction in the level of average monthly narcotic prescription claims per enrollee, but it also caused the trend line to flip from an upward trend of 0.05 fills per month prior to lock- in to a downward trend of -0.04 fills per month.

Lastly, an early internal Utah MLIP evaluation conducted in 1995 reported significant decreases in emergency room visits, multiple pharmacy use, and narcotic prescription claims in the magnitude of 65-75%.129

2.2.4.4 General MLIP Outcomes—Unpublished Clinical Findings

Only one unpublished source of MLIP outcomes evidence—the 2012 presentation materials from Best and colleagues in Washington State—investigated any type of clinical outcomes from MLIP

outcomes among roughly 500 MLIP enrollees and 200 assigned to observation only, found no significant changes in all-cause mortality between the MLIP intervention and control groups.118

In document Roberts_unc_0153D_15527.pdf (Page 48-52)