Patient outcomes from implementing an enhanced services pharmacy network.

OBJECTIVE: To evaluate the effect of implementing a network of community pharmacies on medication adherence, health service utilization, and health care spending. DESIGN: Quasi-experimental difference-in-difference analysis with a nonequivalent control group. SETTING AND PARTICIPANTS: Eligible Medicaid-enrolled patients in North Carolina were attributed to intervention pharmacies between March 2015 and December 2016. A control group was propensity score-matched. Interventions consisted of enhanced services and a more intensive, comprehensive initial pharmacy assessment (CIPA). OUTCOME MEASURES: Outcomes included hospitalizations; emergency department (ED) visits; health care spending for total medical, inpatient, outpatient, and ED services; and adherence to renin-angiotensin system antagonists (RASA), statins, noninsulin diabetes medications (NIDM), and multiple medications for chronic conditions (MMCC). RESULTS: There were 31,509 patients who met eligibility criteria and were attributed to a participating pharmacy. Of these, 3897 received a CIPA. Before matching, patients attributed to participating pharmacies had greater Medicaid enrollment through aged, blind, or disabled status (49.2% vs. 31.5%, P < 0.001); greater case management (10.3% vs. 7%, P < 0.001); and worse rates of chronic disease (P < 0.001). Successful matching removed these differences. Adherence to RASA medications and MMCC increased by 9.5% and 10.3% (P < 0.05), respectively. Adherence did not change for statins and NIDM. The analysis also revealed a slower decline in average total medical spending of 5.7% (P < 0.01) relative to the control group over the same period, owed to a 9.6% (P < 0.001) slower decline in outpatient spending. ED utilization also decreased more slowly relative to controls by 4.8% (P < 0.05) following the intervention. CONCLUSION: The pharmacy intervention resulted in a statistically significant improvement in medication adherence to RASA and multiple chronic medications, but did not change or may have worsened utilization and spending outcomes. More research is needed to explore patient selection and variation in implementation and heterogeneity of treatment effects when evaluating pharmacy interventions.

Pharmacist and physician satisfaction and rates of switching to preferred medications associated with an instant prior authorization program for proton pump inhibitors in the North Carolina Medicaid program.

BACKGROUND: Proton pump inhibitors (PPIs) are among the highest expenditure drugs covered by health care plans. During fiscal year 2001-2002, Medicaid programs nationwide spent nearly $2 billion on PPIs. Although the costs of individual PPIs vary widely, there is little variation in therapeutic effectiveness. On June 1, 2007, the North Carolina Medicaid program implemented an "instant approval" option simultaneously with a prior authorization (PA) program for PPIs with the goal of managing costs and maintaining high-quality care. Preferred PPIs included generic omeprazole and Prilosec OTC. This instant approval process (IAP) was expected to impose less administrative burden than is typically associated with PA programs by permitting physician and nonphysician prescribers to either write the PA criteria directly on a prescription form or use "MD Easy," a preprinted form that could be faxed by the prescriber to the dispensing pharmacy. A previous study found that from the prescriber's perspective the IAP reduced practice-related administrative burden and was associated with a reduced gap in PPI therapy when compared with traditional PA. OBJECTIVE: To evaluate the acceptability and effectiveness of this IAP for PPIs as assessed by the outcome measures of (a) pharmacist satisfaction with the IAP; (b) physician and pharmacist satisfaction with the MD Easy form; and (c) utilization rates for preferred PPIs, comparing medical practices that used the MD Easy form with practices that did not. METHODS: A cross-sectional design was used to assess pharmacist and physician satisfaction. A stratified random sample of 240 pharmacies was selected from 1,561 North Carolina pharmacies with claims in the Medicaid claims data file during state fiscal year 2006. Additionally, a stratified random sample of 240 medical practices was selected from 1,045 primary care practices serving Medicaid beneficiaries during 2006. Surveys were administered to pharmacists using either in-person interviews or self-administered questionnaires and to physicians using a mailed questionnaire with follow-up to nonrespondents. An interrupted time series analysis was used to evaluate the effect of the MD Easy form on switching to preferred PPIs using paid Medicaid claims of surveyed practices from calendar year 2007. Practices that reported both using the IAP and receiving the MD Easy form were defined as MD Easy users. Monthly market share data were analyzed using log negative binomial regression models to account for autocorrelation in the time series data. RESULTS: The pharmacy survey was completed by 202 (84.2%) pharmacies selected for participation. Of 198 permanently employed pharmacists, 140 (70.7%) reported experience with the IAP for PPIs. More than two-thirds (68.6%) of the pharmacist respondents with IAP experience indicated that the IAP is better (34.3%) or much better (34.3%) than traditional PA with RESEARCH respect to overall administrative burden of phone calls, faxes, patient interactions, and doctor contacts. Surveys were completed by 171 (71.3%) of selected physician practices, of which 56 (32.7%) reported experience with the MD Easy forms. Of practices that recalled receiving the MD Easy forms, 52 of 56 (92.9%) reported that the forms "very much" or "somewhat" helped prevent gaps in PPI therapy; 54 of 55 (98.2%) reported that they helped identify patients affected by Medicaid PPI PA; and 100% reported that they helped physicians to follow PA requirements. Immediately after implementation of the IAP and MD Easy form, the observed market share of preferred PPIs increased by 4.1 times (95% CI = 3.57-4.62). From May to June 2007, the preferred PPI market share increased by 64.0 percentage points, from 19.3% to 83.3% (P < 0.001), for practices that reported using the IAP and receiving the MD Easy form (n = 56) and by 55.4 percentage points, from 21.8% to 77.2% (P < 0.001), for practices that either (a) reported not receiving the MD Easy form (n = 25) or (b) reported not using the IAP (n = 84) or (c) did not respond to the survey item asking about the MD Easy form (n = 4). The overall increase in preferred PPI market share after implementation of the IAP was 1.29 times higher for practices that used the MD Easy form than for those that did not based on negative binomial regression modeling; this difference approached statistical significance (95% CI = 1.00-1.68; P = 0.053). CONCLUSION: This study suggests that an IAP for PPIs using either handwritten prescriptions or a preprinted form is an effective alternative to traditional PA. The IAP was associated with an increase in market share for preferred PPIs and was perceived by pharmacists as less administratively burdensome than traditional PA. Additional studies are needed to determine sustainability and the applicability to other prescription drugs.

Evaluation of product switching after a state Medicaid program began covering loratadine OTC 1 year after market availability.

OBJECTIVE: The conversion of loratadine from prescription (Rx)-only to over-the-counter (OTC) status on November 27, 2002, brought about the question of how OTC products may influence utilization of both OTC and Rx-only low-sedating antihistamines (LSAs) simultaneously. North Carolina (NC) Medicaid initially did not cover loratadine OTC but subsequently changed the policy 1 year after OTC conversion, on November 23, 2003. The objective of this study was to determine patterns of LSA utilization in relation to changes in OTC availability and Medicaid coverage policy and to assess the rate of product switching associated with these policies. METHODS: Administrative pharmacy claims from the NC Medicaid population of approximately 1.1 million eligible recipients were used to study the 3 years of LSA use between July 1, 2001, and June 30, 2004. Two general methods were employed to evaluate the extent of product switching. First, monthly rates of incident use, new starts (i.e., no LSA use in the prior 12-month period) and product switching in time series were determined. These series were constructed to include a baseline period of no OTC availability, a period of OTC availability without coverage, and a period of OTC availability with coverage. Second, product switching was assessed through the use of rate-ratio calculations. Three equal 12-month periods were compared using rate ratios: (1) a baseline referent period (July 1, 2001, to June 30, 2002) during which loratadine OTC was not yet available, (2) a noncoverage period (July 1, 2002, to June 30, 2003) during which loratadine OTC was introduced to the market but not covered by NC Medicaid, and (3) a coverage period (July 1, 2003, to June 30, 2004). The primary comparison periods for the 3 years were the 5-month periods from February to June of each year. RESULTS: The use of individual drugs within the LSA class responded to coverage changes as expected, with alternative LSAs replacing loratadine use in the loratadine noncoverage period. Switching behavior for individual drugs within the LSA class was strongly associated with coverage changes. Recipients using loratadine were 2.16 times more likely to switch to an alternative Rx-only antihistamine in the noncoverage period (95% confidence interval [CI], 2.10-2.22) as compared with the baseline period. Yet they were only 1.11 times as likely not to use an Rx LSA during the last 5 months of the noncoverage period (95% CI, 1.09-1.13), as compared with the baseline period, suggesting minimal OTC uptake. The largest 12-month percentage increase in market share was observed for cetirizine (13.4%) although desloratadine accounted for the largest switch rate from loratadine at 3.10 (95% CI, 2.91-3.30), as compared with the baseline period, with a total market share increase of 7.8%. This suggests that new users of LSAs were most likely to initiate therapy with cetirizine, while existing loratadine users were most likely to switch to desloratadine. Compared with baseline switch rates, LSA users were only 0.34 (95% CI; 0.32-0.37) times as likely to switch to loratadine OTC from another (Rx-only) LSA during the subsequent OTC coverage period. LSA expenditure per member per month (PMPM) was essentially constant over time, at dollar 3.03 in the 5-month pre-OTC period, dollar 2.96 in the 5-month loratadine noncoverage period, and dollar 2.93 in the 5-month coverage period for loratadine OTC. Total LSA utilization increased slightly, from 1.37 days PMPM in the 5-month pre-OTC period to 1.41 in the 5-month loratadine noncoverage period and 1.45 in the 5-month coverage period for loratadine OTC. Loratadine OTC accounted for only 4.1% of the total LSA days of therapy and 4.2% of the LSA patients in the 5-month OTC coverage period from February to June 2004. CONCLUSION: Medicaid recipients switched to another covered (Rx) LSA when loratadine became available as an OTC and was not covered. After the subsequent policy change 1 year later to cover loratadine OTC, there was little switching to loratadine OTC. Though the average cost per LSA claim dropped dollar 4.15 (6.6%), from dollar 62.79 in the baseline period to dollar 58.64 in the OTC coverage period, time-series and rate-ratio results suggest that an additional dollar 6.01 (10.2%) could have been saved per LSA claim had OTC coverage been in effect at the time of the conversion of loratadine to OTC status. Although coverage of loratadine OTC offers a substantial cost-savings opportunity for the Medicaid program compared with Rx-only LSAs, not covering the OTC product immediately at the time of OTC availability contributed to (a) increased switching to Rx-only LSA products and (b) little use of loratadine OTC in the subsequent OTC coverage period.

Pharmacist response to alerts generated from Medicaid pharmacy claims in a long-term care setting: results from the North Carolina polypharmacy initiative.

OBJECTIVE: In response to burgeoning drug costs, North Carolina (NC) Medicaid encouraged pharmacists and prescribers to develop collaborative programs to reduce drug expenditures. One of these programs, the North Carolina Polypharmacy Initiative, was a focused drug therapy management intervention aimed at reducing polypharmacy in nursing homes. This intervention targeted patients with more than 18 prescription fills in 90 days, beginning in November 2002. These patients were believed to have a high likelihood of experiencing potential drug therapy problems (PDTPs). Consultant pharmacists were asked to utilize profiles displaying alerts generated from pharmacy claims to guide interventions in addition to usual-care drug regimen reviews. The pharmacists documented their reviews, recommendations, and resulting changes in drug therapy. Our objectives were to determine (1) the persistence of PDTP alerts following interventions by consultant pharmacists and (2) the impact of these interventions on patient drug costs from a payer perspective. METHODS: A before-after study with comparison group design was used. Medicaid prescription claims data were compared for the 90-day periods prior to the intervention (June-August 2002) and following the intervention (March-June 2003). The 90-day post-intervention period allowed for 2 to 3 follow-up prescriptions and reduced the drop-out rate. The 5 categories of potential problem alerts included potentially inappropriate medications (Beers criteria), substitution opportunity for a lower-cost drug, 16 drugs or drug classes with specific quality improvement opportunities (Clinical Initiatives list), therapeutic duplication, and length of drug therapy evaluation. RESULTS: A total of 253 nursing homes, involving 110 consultant pharmacists and 6,344 patients, were in the intervention arm, with 5,160 patients (81.3%) remaining at the end of the follow-up period. At baseline, study-group patients used an average of 9.7 prescriptions per month, costing the NC Medicaid program 517 US dollars per patient per month (PPPM). There were 6,360 recommendations offered for 3,400 patients, or an average of 1.87 recommendations per patient. Physicians concurred with 59.8% (3,801 of 6,360) of all recommendations to change drug therapy, about half involving a switch to a lower-cost drug. Two of 5 alert categories had significant (P <0.01) reductions in alert persistence: -10.8% for the study group versus -0.7% for the comparison group for the Clinical Initiatives list and -29.7% for the study group versus -14.1% in the comparison group for the drug substitution opportunity. Median drug costs per patient in the study group decreased by 12.14 US dollars (-0.92%), from 1,329.46 US dollars to 1,317.32 US dollars, and increased in the comparison group by 44.98 US dollars (3.35%), from 1,341.25 US dollars to 1,386.23 US dollars, creating a relative cost reduction of 57.12 US dollars per patient in the 3-month follow-up period, or 19.04 US dollars PPPM. CONCLUSION: A supplemental program of medication reviews for nursing home patients targeted by high drug utilization resulted in a reduction in the persistence of PDTP alerts and was cost beneficial based solely on drug cost savings. This intervention may be a model for future medication therapy management services provided by prescription drug plans under Medicare Part D for patients in long-term-care settings and possibly ambulatory patients.

Transitional care cut hospital readmissions for North Carolina Medicaid patients with complex chronic conditions.

Recurrent hospitalizations represent a substantial and often preventable human and financial burden in the United States. In 2008 North Carolina initiated a statewide population-based transitional care initiative to prevent recurrent hospitalizations among high-risk Medicaid recipients with complex chronic medical conditions. In a study of patients hospitalized during 2010-11, we found that those who received transitional care were 20 percent less likely to experience a readmission during the subsequent year, compared to clinically similar patients who received usual care. Benefits of the intervention were greatest among patients with the highest readmission risk. One readmission was averted for every six patients who received transitional care services and one for every three of the highest-risk patients. This study suggests that locally embedded, targeted care coordination interventions can effectively reduce hospitalizations for high-risk populations.

Analysis of the North Carolina long-term care polypharmacy initiative: a multiple-cohort approach using propensity-score matching for both evaluation and targeting.

BACKGROUND: The high cost and undesirable consequences of polypharmacy are well-recognized problems among elderly long-term care (LTC) residents. Despite the implementation of the 1987 Omnibus Budget Reconciliation Act, which requires pharmacist review of drug regimens in this setting, medical and drug costs for LTC residents have continued to increase. OBJECTIVE: This study evaluates the North Carolina Long-Term Care Polypharmacy Initiative, a large-scale medication therapy management program (MTMP) that combined drug utilization review activities with drug regimen review techniques. METHODS: This was a prospective records-based study that used a difference-in-difference model with both historical and nonintervention group controls. To ensure equivalence among subjects, propensity scoring was used to match study subjects from participating LTC facilities with comparison subjects from nonparticipating facilities. Residents with interventions were grouped for analysis by intervention type-retrospective only, prospective only, or dual type (residents with both prospective and retrospective interventions)-and by intervention stage-review, recommendation, and drug change-plus an all-inclusive "all types" grouping that aggregated groups by intervention type, for a total of 10 total cohorts. RESULTS: In the overall population of 5255 study subjects identified, a US $21.63 per member per month drug-cost savings was observed. Although only 1 of 10 cohorts had a change in the number of drug fills, substantial reductions in 2 of 5 types of drug alerts were observed in all 10 cohorts. A reduction in the relative risk for hospitalization (0.84 [95% CI, 0.71-1.00]) was observed in the cohort of residents receiving a retrospective review. CONCLUSIONS: This Initiative suggests that an MTMP can be quickly launched in a large number of LTC facility residents to produce monetary drug-cost savings and improved health outcomes. Additionally, the evaluation of this program illustrates the utility of using propensity scoring techniques to target future intervention groups in a cost-effective manner.

A physician-friendly alternative to prior authorization for prescription drugs.

OBJECTIVE: To determine if the instant approval (IA) process differs from the traditional prior authorization (PA) process in preferred drug channeling, resultant gaps in therapy, and provider dissatisfaction. STUDY DESIGN: An interrupted time series analysis using pharmacy claims and a retrospective cohort study. METHODS: The study assessed changes in preferred drug use and subsequent cost reductions. A retrospective cohort study determined if the IA process produced fewer gaps in therapy than the PA process. Provider acceptance of the IA process was assessed using a brief survey of 240 randomly selected primary care practices. RESULTS: Market share for preferred proton pump inhibitors quadrupled from a range of 17.6% to 19.3% at baseline to 76% in the first month after implementation of the new IA policy. Most practices (81.1%) reported reduced administrative burden with the IA process. The median gaps between medication fills for patients using IA were approximately one-half those of patients using PA (P <.001) and were one-fourth in a subset of highly adherent, regularly filling patients (P <.001). CONCLUSIONS: Instant approval may be more patient friendly and prescriber friendly than PA as assessed by a proxy measure for access (gap in therapy) and physician-reported acceptance. Despite its ease of use, IA does not seem to reduce switching to preferred drugs.

Trends in emergency department antibiotic prescribing for acute respiratory tract infections.

BACKGROUND: Injudicious use of antibiotics is associated with the reported rise in antibiotic-resistant bacteria. With an estimated 26 million antibiotics being prescribed annually in the emergency department (ED), the ED represents an important setting for targeting interventions. OBJECTIVE: To provide national estimates of potentially inappropriate antibiotic prescribing during ED visits for acute respiratory tract infections (ARTIs) and examine associations between patient, provider, visit characteristics, and antibiotic prescribing patterns. METHODS: A cross-sectional study was conducted of ED visits for ARTIs, identified from pooled 1995-2000 National Hospital Ambulatory Medical Care Survey data. National estimates, descriptive statistics, and multivariate analyses were used to assess antibiotic prescribing patterns. RESULTS: An estimated 51.3 million ED visits for ARTIs occurred during the study period, 62% of which had an antibiotic prescribed. For a narrowly defined subset of ARTIs, where antibiotic therapy is nearly always inappropriate (eg, nasopharyngitis, ARTI of multiple or unspecified sites, acute bronchitis), the percentage decreased over the 6-year period from 57% to 44% (p < 0.01). For children ED visits, however, the downward trend occurred almost exclusively in urban EDs. Compared with visits in which a resident or intern physician was involved, the odds of antibiotic prescribing for child ED ARTI visits were 2.2 times higher for staff physicians (95% CI 1.3 to 3.6) and 1.8 times higher for nonphysicians with prescribing privileges (95% CI 1.3 to 2.4). CONCLUSIONS: ED antibiotic prescribing for ARTIs has decreased from 1995 to 2000, but still is occurring in well over half of ED visits for ARTI. Further research assessing knowledge and attitudes of patients and providers about antibiotic prescribing is needed.