Currently Viewing:
The American Journal of Managed Care October 2012
Change to FIT Increased CRC Screening Rates: Evaluation of a US Screening Outreach Program
Elizabeth G. Liles, MD, MSCR; Nancy Perrin, PhD; Ana Gabriela Rosales, MS; Adrianne C. Feldstein, MD, MS; David H. Smith, RPh, MHA, PhD; David M. Mosen, PhD, MPH; and Jennifer L. Schneider, MPH
Toward Tailored Disease Management for Type 2 Diabetes
Arianne M. J. Elissen, MSc; Inge G. P. Duimel-Peeters, PhD; Cor Spreeuwenberg, PhD; Marieke Spreeuwenberg, PhD; and Hubertus J. M. Vrijhoef, PhD
Implementation of EHR-Based Strategies to Improve Outpatient CAD Care
Stephen D. Persell, MD, MPH; Janardan Khandekar, MD; Thomas Gavagan, MD; Nancy C. Dolan, MD; Sue Levi, RN, MBA; Darren Kaiser, MS; Elisha M. Friesema, BA, CCRP; Ji Young Lee, MS; and David W. Baker, MD, MPH
Pediatric Integrated Delivery System's Experience With Pandemic Influenza A (H1N1)
Evan S. Fieldston, MD, MBA, MSHP; Richard J. Scarfone, MD; Lisa M. Biggs, MD; Joseph J. Zorc, MD, MSCE; and Susan E. Coffin, MD, MPH
Medicare Part D Claims Rejections for Nursing Home Residents, 2006 to 2010
David G. Stevenson, PhD; Laura M. Keohane, MS; Susan L. Mitchell, MD, MPH; Barbara J. Zarowitz, PharmD, FCCP, BCPS, CGP, FASCP; and Haiden A. Huskamp, PhD
Currently Reading
Identification of and Intervention to Address Therapeutic Gaps in Care
Daniel R. Touchette, PharmD, MA; Sapna Rao, BPharm, MS; Purna K. Dhru, PharmD; Weihan Zhao, PhD; Young-Ku Choi, PhD; Inderpal Bhandari, PhD; and Glen D. Stettin, MD
Financial Incentives and Physician Commitment to Guideline-Recommended Hypertension Management
Sylvia J. Hysong, PhD; Kate Simpson, MPH; Kenneth Pietz, PhD; Richard SoRelle, BS; Kristen Broussard Smitham, MBA, MA; and Laura A. Petersen, MD, MPH
Identifying Frail Older People Using Predictive Modeling
Shelley A. Sternberg, MD; Netta Bentur, PhD; Chad Abrams, MA; Tal Spalter, MA; Tomas Karpati, MD; John Lemberger, MA; and Anthony D. Heymann, MB BS
Application of New Method for Evaluating Performance of Fracture Risk Tool

Identification of and Intervention to Address Therapeutic Gaps in Care

Daniel R. Touchette, PharmD, MA; Sapna Rao, BPharm, MS; Purna K. Dhru, PharmD; Weihan Zhao, PhD; Young-Ku Choi, PhD; Inderpal Bhandari, PhD; and Glen D. Stettin, MD
A program identifying and resolving care gaps involving community pharmacists resulted in increased adherence and omission gaps closure and fewer adherence gap reopenings.
Sensitivity analyses were run to assess the effects of certain operational decisions on the models. In the initial models, we chose to include multiple gap categories for each patient if these existed. For example, if a patient was non-adherent with multiple medications, he or she could potentially contribute to statin, beta-blocker, and ACE-inhibitor adherence gaps. Since this violates the assumption of independence of observations, we ran a sensitivity analysis including only 1 of the gaps selected at random. We also ran a sensitivity analysis examining only those gaps generated after the study start date and excluding all those that existed when the study began.

An analysis of refill compliance, as an insight into adherence during the study, was conducted using a continuous multiple interval measures of medication gaps (CMG) method described by Steiner and Prochazka.23 Specifically, we identified the number of days after which medication supplies were exhausted between the start of an adherence gap and the end of the study divided by the number of days in the same period. The numerator was not adjusted for oversupplies. Rather, the minimum of days supplied or time between fills was used to calculate the number of days for which medication supplies were available. The intervention CMG was compared with the control group CMG using a Mann-Whitney U test.

Economic Analysis

The cost of providing the intervention, excluding administrative costs, was calculated from the insurer’s perspective. Specifically, the counseling reimbursement was calculated as total cost and as a cost per eligible patient per month (ie, those patients who were on medications and were eligible for the program). The costs associated with generating and communicating gaps, pharmacist educational sessions, administrative support, and program evaluation were not included in this analysis.


A total of 96 pharmacies were subject to randomization. Three pharmacies in the intervention group were unable to have a pharmacist complete the training program and 1 control pharmacy closed before the study began. As a result, 92 pharmacies participated in the study. Of these, 45 pharmacies were assigned to conduct the intervention and 47 were assigned as the control group. A total of 2571 patients (1445 intervention and 1126 control patients) with therapy gaps were assigned to these pharmacies. Pharmacy, patient, and provider characteristics are shown in Table 2. During the 90- day intervention period after the study start, 2614 existing and new adherence gaps were identified that met inclusion criteria and were included in the study (1433 intervention and 1181 control) while 1211 omission gaps were included

in the study (677 intervention and 534 control gaps). Out of 1433 adherence and 677 omission gaps sent to intervention pharmacies, pharmacists intervened on 639 (44.6%) and 506 (74.7%) gap alerts respectively. Gaps that were unaddressed by pharmacists included 441 (30.8%) adherence and 57 (8.4%) omission gaps recalled because the criteria required for generating the gap were no longer met. For example, to generate an alert, a patient must have filled the medication at least 1 time in the past 6 months. An additional 335 (23.4%) adherence and 16 (2.4%) omission gaps were recalled because a claim for the medication was received prior to pharmacist intervention, thereby automatically closing the gap. The patient declined intervention in 10 (0.7%) adherence and 20 (3.0%) omission gaps. The remaining 8 (0.6%) adherence and 78 (11.5%) omission gaps remained unaddressed for unknown reasons.

Adherence Gap Closure

In total, gaps were closed from the alert notification more often in the intervention group than control at 30 days (55.5% closure rate in intervention vs 50.6% in control) 45 days (61.1% vs 58.4%, respectively), 60 days (66.1% vs 65.2%, respectively), and 90 days (73.0% vs 72.9%, respectively). When evaluated using the Cox proportional hazards model, including control variables for the adjusted analysis, a significant difference was observed at 90 days (hazard ratio [HR] = 1.242; P = .022; 95% confidence interval [CI] 1.115-1.385) and at all of the time points (30-, 45-, and 60-day adjusted P values all <.001). Figure 1 shows the KM survival curves for gap closure. All final models included 2 covariates: standardized number of pharmacy claims and pharmacy patient count (these variables were standardized for better convergence in SAS 9.2).

Adherence Gap Reopenings

Including only those gaps that had sufficient time to reopen (45 days) prior to the end of the study, adherence gaps reopened in 59.6% of intervention and 67.0% of control patients. The difference was statistically significant when compared using a Cox proportional hazards model (HR = 0.863; P = .012; 95% CI 0.769-0.968). The final model included 2 covariates: standardized number of pharmacy claims and standardized pharmacy patient count.

Estimated Short-term Change in Adherence

The proportion of days in which medication supply was exhausted was 52.3% in the intervention group using a CMG method. The corresponding CMG was 55.6% in the control group. This difference in CMG was statistically significant (P = .001). When only those patients filling a medication during the study period were included (ie, those who remained persistent with therapy), the CMG was 39.5% in the intervention group and 43.5% in the control group (P <.001).

Omission Gap Closure

There were 89 (13.1%) intervention and 29 (5.4%) control gaps that had closed within 90 days of the gap alerts being sent to the pharmacists (unadjusted χ2 P <.001). This statistical significance was maintained when evaluated using the Cox proportional hazards model (HR = 1.770; P = .005; 95% CI 1.182-2.653). The KM survival curve for open omission gaps is shown in Figure 2. The final omission gap model included the covariates patient age, patient gender, and practitioner’s prescription claim count.

Sensitivity Analyses

The results of the models developed in the sensitivity analyses were consistent with those in the main analyses. Because of the smaller sample size, the results analyzing only new gaps and excluding gaps that existed at the study start were not statistically significant. The KM survival curves were very similar, as were the final model parameter estimates. The sensitivity analysis evaluating only a single gap per patient, rather than allowing multiple gaps per person, showed similar results.

Economic Analysis

Pharmacists provided 639 first sessions and 427 follow-up (second and third) sessions for adherence gaps. Pharmacists conducted 506 first sessions and 506 follow-up sessions for omission gaps. The total cost of counseling over the first 3 months of the intervention period was $48,345. Given that 6398 patients were assigned to intervention pharmacies and monitored for gap identification, the mean cost per patient per month for the first 3 months of the program was $2.52.


This study demonstrates that a program designed to improve adherence and reduce omission gaps utilizing community pharmacists was effective in closing those gaps sooner. In addition, a fewer number of adherence gaps reopened during the time frame of the analysis. In the case of omission gaps, the result was fewer omission gaps open at 90 days. These results are promising and suggest that community pharmacists can effectively address medication underutilization when appropriately supported.

Patients who demonstrated suboptimal adherence (MPR less than 80% over the past 18 months) and were late to fill a prescription were identified for intervention and approached at a time when adherence was potentially problematic. A greater proportion of adherence gaps were closed at 30 and 60 days after alert generation. This finding indicates that study patients were likely to fill their prescriptions sooner in the intervention group. Additionally, patients in the intervention group were more likely (absolute difference 7.4%) to refill their prescriptions again during the study period, suggesting that adherence was actually impacted. However, we observed no difference in the proportion of adherence gaps closed at 90 days, indicating that persistence was not improved by this intervention. Lack of persistence would commonly be addressed by alerting the prescribing physician, recommending an alternative therapy, and discussing other (including non-prescription) options with patients. Our analysis was not designed to identify drug switching or the impact of this intervention on drug switching. Therefore, we do not know if this intervention had any impact on improving therapy for non-persistent patients.

We identified 3 previously published studies designed to improve medication adherence via patient and/or physician outreach. In 1 multisite, controlled clinical trial at a medinum-sized grocery chain, over 3000 patients were randomized to 1 of 2 interventions or control.19 One intervention involved directly notifying the physician that a patient was overdue for refills. The other intervention involved telephone outreach by a pharmacist or technician to discuss an overdue refill, assess barriers to therapy, and suggest potential solutions to non-adherence. Over half of the physicians contacted chose to either opt out of the study (45.1%) or did not respond (13.7%) to the study questionnaire. Pharmacists or technicians were successful in contacting 81% of patients. There were no significant differences in the time to refill medications among any of the groups.19 While the study was well designed, the results suggest that either the method of identifying poorly adherent patients or the intervention as conducted was not effective. Specifically, identification of poor adherence based on late to fill criteria alone may not be sufficient. Our study differed in that adherence over the past 18 months and late to fill criteria were simultaneously used to identify patients with long-standing adherence issues. In addition, intervention pharmacists in our study received training in motivational interviewing and were solely responsible for conducting the intervention. The intervention was reinforced over the course of up to 3 sessions, rather than in a single telephone call.

Another study, conducted within the outpatient pharmacies of 1 hospital system, incorporated automated telephone reminders, patient information pamphlets with pictorial representations of how and when to take medications, and training sessions for pharmacists on improving communication with patients.14 While an improvement in the adherence in the intervention group compared with baseline, the difference did not achieve statistical significance. Also, limitations in the study’s design, such as a lack of random assignment exacerbatedby important differences between the control and intervention groups at study baseline, make the results of this study difficult to interpret. Our study employed randomization to ensure similar study groups.

Copyright AJMC 2006-2020 Clinical Care Targeted Communications Group, LLC. All Rights Reserved.
Welcome the the new and improved, the premier managed market network. Tell us about yourself so that we can serve you better.
Sign Up