 To evaluate the effectiveness of a telephone-based “virtual” osteoporosis clinic in increasing the use of osteoporosis medication assessed at 1 year after receipt of a prescription.

Women 60 years and older with previously undiagnosed osteoporosis were randomized to evaluation and treatment by a dedicated telephone-based osteoporosis clinic with monthly telephone follow-up until medication was successfully started (intervention) or to usual care provided by their primary care physician (control). A successful outcome was defined as having filled a prescription for a 3-month supply of medication within 130 days, marking 1 year and 30 days since enrollment.

A total of 235 women underwent randomization, and 211 received the allocation. Of 109 women in the telephone-based osteoporosis clinic group, 75 (68.8%) were using osteoporosis medication at 1 year compared with 46 of 102 women (45.1%) in the usual care group (

 <.001). A poststudy questionnaire showed no significant differences between the groups in regard to knowledge about osteoporosis or attitude toward their osteoporosis care provider. The significant increase in osteoporosis medication use with the telephone intervention occurred at the same time that an independent health maintenance organization—wide program promoting osteoporosis treatment seemed to improve overall rates of use.

The use of osteoporosis medication among women with newly diagnosed osteoporosis may be significantly improved by a simple intervention based on monthly telephone followup. Overall use of osteoporosis medication in this trial may have been increased by a systemwide initiative to improve osteoporosis care conducted concurrently with the trial. (ClinicalTrials.gov Identifier: NCT00145067.)

Adherence to osteoporosis medication use is suboptimal, with most observational studies reporting persistence rates of 30% to 50% after 1 year. In this randomized controlled trial, we found that simple monthly telephone calls until the use of osteoporosis medication was confirmed significantly increased the use of medication 1 year after trial entry compared with usual care by a primary care physician.

The use at 1 year among women randomized to receive monthly telephone calls was 68.8% compared with 45.1% among women randomized to usual care.

The simple protocol in this study has been readily adopted into a disease management “virtual” clinic serving approximately 6000 patients with osteoporosis.

Osteoporosis is one of the most common diseases of the geriatric population, affecting more than 10 million persons 50 years and older in the United States. It is estimated that up to one-half of American women and one-quarter of American men will experience a fracture caused by osteoporosis in their lifetimes.

Despite the high prevalence of the disease and the increased media attention it has received in recent years, less than one-quarter of patients who would benefit from drug therapy for osteoporosis are offered treatment.

Further adding to low treatment rates, adherence to medication use is poor among patients who are prescribed osteoporosis therapy. Observational studies

 report persistence with treatment at 12 months (defined as continuing to fill prescriptions) to be in the range of 30% to 50% as determined by pharmacy records.

Adherence to medication use is generally defined as having 2 components, namely, compliance (taking medication as prescribed) and persistence (period over which prescriptions continue to be filled).

 Both components have been associated with improved outcomes in chronic diseases in general

 and specifically in regard to osteoporosis medication use, for which studies

 have shown that increased adherence is directly associated with decreased risk of fractures. Despite this, few interventions designed to increase adherence to osteoporosis medication use have been studied. The use of weekly rather than daily dosing of osteoporosis medication has been associated with better adherence to bisphosphonate use in some studies

 although overall adherence remained suboptimal regardless of dosing frequency. In other studies,

 increased direct contact with osteoporosis care providers through additional clinic visits has improved adherence.

Telephone-based interventions have also shown promise in increasing adherence to osteoporosis medication use in combination with increased clinic visits

 or decreased medication dosing frequency

 in randomized trials. Monthly telephone counseling alone achieved adherence rates that were significantly higher than historical rates in a recent nonrandomized trial.

The use of a telephone consultation protocol to encourage the use of medication among women diagnosed as having uncomplicated osteoporosis has been previously described.

The protocol involved telephone follow-up of patients until initiation of prescribed osteoporosis medication was confirmed. Given the promising results of the protocol, we conducted a study to compare the use of medication 1 year after prescription among women randomized to this “virtual” osteoporosis clinic with the use among women randomized to usual care for osteoporosis delivered in a staff-model not-for-profit health maintenance organization (HMO). We hypothesized that women receiving telephone follow-up would have a higher rate of medication use at 1 year than women receiving usual care.

The study was conducted at Kaiser Permanente San Diego Department of Preventive Medicine from January 2004 to March 2007 and was approved by the Kaiser Permanente Southern California Institutional Review Board. Women were eligible for the study who were 60 years and older, met the National Osteoporosis Foundation guidelines for uncomplicated osteoporosis (by having a T score of −2.5 or lower or by having sustained an osteoporotic fracture), and were not previously identified as having osteoporosis. Most patients were referred to the study through the Health Appraisal Center of the Department of Preventive Medicine or from the Orthopedics Department. Baseline laboratory values for creatinine, thyroid-stimulating hormone, calcium, parathyroid hormone, and 25-hydroxyvitamin D levels were obtained. Patients were excluded from participating if they had secondary osteoporosis other than vitamin D deficiency, were unable to provide informed consent, or spoke a language that precluded a conversation by telephone with the study staff. Recruitment occurred over the first 2 years of the 3-year study. Patients meeting entry criteria who gave informed consent were randomized by sequential enrollment according to a computer-generated list to usual care or to care provided by the osteoporosis clinic. The random assignment was performed by the physician assistant (ALB) who staffed the osteoporosis clinic.

Patients randomized to usual care received a referral to their usual primary care physician. Standard paper or electronic referrals were issued by the Health Appraisal Center, and the primary care physician was not informed of the patient’s participation in the study. The patient was advised that she would be contacted by her primary care physician for follow-up. All subsequent evaluation and treatment were performed by the primary care physician, and no further contact with the patient was initiated by the osteoporosis clinic until the end of the study. Baseline laboratory values were available to the primary care physician through electronic laboratory records.

Patients randomized to the osteoporosis clinic received care from the physician assistant under the supervision of a preventive medicine physician (EMB). The clinic was staffed by a provider (ALB) working full-time for the first 2½ years of the 3-year study and part-time thereafter. Patients with 25-hydroxyvitamin D levels below 30 ng/mL (to convert 25-hydroxyvitamin D level to nanomoles per liter, multiply by 2.496) were given a prescription for vitamin D (50,000 IU orally weekly). All other patients were advised to take vitamin D (800-1200 IU) and calcium (1000-1500 mg) daily. Patients received educational handouts in a one-time mailing and were contacted by telephone for an open-ended discussion with the osteoporosis clinic provider about osteoporosis treatment. Patients were then given a prescription for osteoporosis treatment with a second-generation bisphosphonate to be taken weekly. All prescriptions were for a 3-month supply of medication.

Four to 6 weeks after osteoporosis medication was prescribed, the patient received a standardized follow-up telephone call from the clinic provider. Each call lasted approximately 5 minutes. The patient was asked the following 3 basic questions: (1) if they had filled the prescription for the osteoporosis medication; (2) if they had started taking the medication; and (3) if they answered no to the second question, why they were not taking the medication. Patients were also invited to ask questions, which typically involved concerns about the cost of medication and the potential adverse effects detailed in the medication insert. Strategies for remembering to take the medication such as designating 1 day of the week as “bone day” were discussed. Patients who indicated they would have trouble paying for the medication were assisted in obtaining the drug at no cost from the study sponsor (Merck & Co, Inc, Whitehouse Station, NJ). If a patient had problems with the medication or had not begun therapy, a second telephone call was planned for the following month. Two to 3 telephone calls per patient per month were required on average to make contact with the patient. Monthly telephone calls were continued only until the patient reported that she had filled a prescription for the medication, was taking the medication, and was not having any problems with the medication. No further contact was initiated by the osteoporosis clinic  provider.

During the time of the study, an independent HMO—wide program to improve osteoporosis treatment was launched. Endocrinologists identified as thought leaders met with primary care physicians, obstetrician/gynecologists, and orthopedists to emphasize the importance of osteoporosis treatment. Screening and treatment guidelines were promulgated, with brochures and Web sites targeted toward the same physician groups.

The use of medication was measured at 1 year and 30 days from entry into the study by querying the Kaiser pharmacy database. A successful outcome was defined as having filled a prescription for a 3-month supply of osteoporosis medication within 130 days preceding the query. Assessment of prescriptions filled was facilitated by the fact that most Kaiser enrollees fill all prescriptions through a Kaiser pharmacy. Patients who did not seem to be filling prescriptions at Kaiser were contacted by telephone to ascertain if they were filling medication at an outside pharmacy.

After 1 year and 30 days from enrollment, all patients in the study were sent a validated questionnaire to assess their understanding of osteoporosis and their satisfaction with their osteoporosis care (McHorney et al

; and T. W. Weiss, DrPH, personal written communication). The questionnaire was designed to evaluate osteoporosis treatment in the osteoporosis clinic group or the usual care group and not fracture care in the Orthopedics Department.

A sample size was not calculated a priori. The decision to stop the trial was based on results of a preplanned 2-year interim analysis demonstrating highly significant differences in the use of medication at 1 year between the study groups.

 test were used to test for differences in categorical variables and continuous variables, respectively. All statistical tests were 2-sided and had an a level of .05. Baseline variables and the use of medication at 1 year were compared between the treatment groups in the intent-to-treat analysis, as well as among the subset of subjects who filled at least 1 osteoporosis medication prescription during the trial. Comparisons of knowledge and attitudes between the treatment groups were limited to questionnaire respondents. All analyses were performed using SAS 9.1 (SAS Institute, Inc, Cary, NC).

A total of 442 women were referred to the study and identified as eligible over the 2-year enrollment period. Of these, 207 patients (46.8%) declined to participate, most frequently because of a preference to be treated by their primary care physician (61 patients), followed by having other medical conditions (18 patients), already starting treatment for osteoporosis (13 patients), having insufficient time (12 patients), and moving out of the Kaiser San Diego system (11 patients).

Of 235 patients undergoing randomization, 110 were assigned to usual care and 125 to the osteoporosis clinic. The 2 randomized groups were similar in age, provider-recorded race/ethnicity, percentage with incident fracture, and percentage with a documented T score of −2.5 or lower at baseline (

). Of the randomized subjects, 24 did not complete the study, primarily because of moving out of the Kaiser system or death (

). The outcome was measured in 211 patients (102 patients in the usual care group and 109 patients in the osteoporosis clinic group). Of 109 patients randomized to the osteoporosis clinic who did not move out of the Kaiser system, 10 requested to be dropped from the study at the first telephone call but were included in the intent-to-treat analysis.

Once-weekly alendronate sodium accounted for 80.2% of prescriptions filled by the usual care group and 92.9% of prescriptions filled by the osteoporosis clinic group. Estrogen and selective estrogen receptor modulators accounted for an additional 10% of prescriptions in both groups, and risedronate sodium was taken by another 10% of the usual care group. Eight subjects in the osteoporosis clinic group received assistance in obtaining the medication at no cost from the study sponsor.

Of 109 women in the telephone-based osteoporosis clinic group, 75 (68.8%) were using osteoporosis medication at 1 year compared with 46 of 102 women (45.1%) in the usual care group (

 <.001). Patients in the usual care group who were not using osteoporosis medication at the end of 1 year included 29 (28.4%) who never filled a prescription for any osteoporosis medication during the study period. A similar proportion of patients in the osteoporosis clinic group reported not filling their prescription at the 1-month intervention telephone call, with most reporting compliance at 1 month and not requiring additional follow-up calls. All 99 patients in the osteoporosis clinic group who completed the trial filled at least 1 prescription sometime during the year. There was no statistically significant difference in the use of osteoporosis medication at 1 year between patients who completed the study and filled at least 1 prescription during the trial in the osteoporosis clinic group (72 of 99 [72.7%]) compared with the usual care group (46 of 73 [63.0%]) (

). Only 2 of 8 patients in the osteoporosis clinic group who received sponsor-supplied free medication were still using it at 1 year.

A completed poststudy questionnaire was returned by 136 subjects (73 [67.0%] in the osteoporosis clinic group and 64 [62.7%] in the usual care group) (

). For both treatment groups, the use of osteoporosis medication at 1 year in women who responded to the questionnaire was higher than that in the overall sample (80.8% in the osteoporosis clinic group and 54.0% in the usual care group). No statistically significant differences were seen between the 2 groups in regard to knowledge about osteoporosis or self-reported health behaviors (

). More patients in the osteoporosis clinic group reported that their healthcare provider was knowledgeable all or most of the time (91.9% vs 81.5% in the usual care group, 

). No other differences in attitude toward the provider were found.

Our study was performed among women with newly diagnosed uncomplicated osteoporosis. Women who were followed up by the telephone-based osteoporosis clinic had a rate of use of osteoporosis medication at 1 year after diagnosis that was 52.5% higher than that of women undergoing usual care.

Our osteoporosis clinic used a structured follow-up protocol designed to ensure that patients who were prescribed medication filled the prescription and initially took the medication without complications. All women who completed the trial in the osteoporosis clinic group filled at least 1 prescription during the 1-year period. By contrast, approximately 28% of subjects in the usual care group did not fill any prescription for osteoporosis medication during the period. We do not know if failure to fill a prescription was due to the woman not keeping the appointment with her primary physician or due to osteoporosis medications not being offered by the primary care physician. Therefore, usual care patients categorized as not using osteoporosis medication at 1 year may include patients who never received a prescription and did not have the opportunity to be “adherent.” However, a similar initial rate of not filling the prescription occurred in the intervention group (approximately one-fourth reported not having filled the prescription at the first monthly telephone call) and has been reported by others for osteoporosis

 and other asymptomatic diseases such as hyperlipidemia and hypertension.

 Therefore, the primary effect of the intervention is a combination of ensuring that medication is prescribed and that at least 1 prescription is filled.

In addition to the structured follow-up, several other factors likely contributed to the success of the osteoporosis clinic compared with usual care. Patients were given a telephone number that enabled them to directly contact the osteoporosis clinic provider. By contrast, telephone calls made to primary care physicians generally went through a large centralized call center. As a specialty clinic, the osteoporosis clinic may also have been perceived by patients as more authoritative and robust.

Despite these provisions, the poststudy questionnaire (returned by approximately 65% of women in both groups) showed no statistical difference between the osteoporosis clinic and usual care groups in patient satisfaction with their ability to reach a physician or provider by telephone. Patient satisfaction did not significantly differ between the groups, except for a greater percentage of women in the osteoporosis clinic group reporting that their provider was knowledgeable all or most of the time.

The effect of the intervention also did not seem to be related to improved knowledge about osteoporosis among patients in the osteoporosis clinic arm. Despite the increased educational efforts by the osteoporosis clinic, knowledge about osteoporosis, as assessed by our poststudy questionnaire, was similar between the 2 groups of patients. This apparent lack of effect of patient education is in agreement with findings of others who have reported little or no increased adherence to or persistence with medication use among those receiving more robust patient education,

Our findings are consistent with other studies of interventions that were found to improve adherence to or persistence with osteoporosis medication use, which have generally included some form of direct patient contact after receipt of a prescription. Successful interventions have included additional in-person clinic visits,

 which some have combined with less frequent dosing

The effectiveness of the intervention in this study has led to adoption of the protocol by the Kaiser San Diego Department of Population-Based Medicine to manage approximately 6000 patients with osteoporosis in its “Healthy Bones” program. The protocol may also be adapted for improving the use of medications in other chronic diseases for which adherence is low

 and for which other interventions have had mixed success. In a comprehensive review of these efforts,

 it was noted that most interventions that were effective for increasing adherence to medication use for chronic diseases were complex and often labor intensive and did not lead to large improvements in adherence and treatment outcomes. In addition, most studies of successful complex interventions have not assessed the separate effects of the components, so it is unknown which components are required. However, interventions that included telephone contact were most often successful in increasing adherence to medication use for hyperlipidemia,

 conditions like osteoporosis that are largely asymptomatic. Two of these protocols included automated calls from computerlinked systems with

 telephone follow-up from a healthcare provider.

Although not formally evaluated by our study, it is our opinion that the personalized encouragement and answers to questions provided by a conversation with the clinician would be more effective than automated telephone calls. This concept is supported by Cook et al,

 who used monthly telephone calls from nurses trained in motivational interviewing to encourage adherence among women who were prescribed risedronate for treatment or prevention of osteoporosis and achieved rates of use similar to ours of 69% at 6 months compared with 40% among historical control subjects.

A limitation of our study is that we were unable to calculate the medication possession rate, a standard measure of persistence with medication use. Therefore, we are unable to directly compare our results with those of other studies using these standard definitions. An additional limitation is that patients in the intervention group were explicitly assisted in obtaining study medication at no cost from the study sponsor. The provision of this assistance was not measured in the usual care group and may have been less easily obtained. However, given that only 2 of 8 patients in the osteoporosis clinic group who received assistance were still using the medication at 1 year, it is unlikely that this assistance was a major factor in the success of the intervention. Finally, interpretation of the results is complicated by the fact that a HMO—wide effort to increase osteoporosis treatment by primary care physicians occurred concurrently with our intervention. Although this program may have increased baseline rates of osteoporosis medication use (estimated to be approximately 25% before the trial), our intervention seemed to have an additional effect.

Therefore, we have demonstrated that a simple protocol based on monthly telephone calls until a prescription is filled was effective in increasing rates of the use of osteoporosis medication at 1 year by 52.5%, from 45.1% under usual care to 68.8% with the intervention. If combined with medications requiring less frequent dosing, the protocol may be even more effective at ensuring that patients at high risk for fracture will receive adequate osteoporosis treatment.

This is manuscript number MEM-19483 from The Scripps Research Institute.

Articles

WEB EXCLUSIVE

In this randomized controlled trial, women receiving monthly telephone calls had significantly higher use of osteoporosis medication at 1 year versus women receiving usual care.

A Telephone-Based Intervention for Increasing the Use of Osteoporosis Medication: A Randomized Controlled Trial

To minimize out-of-pocket prescription drug plan (PDP) expenditures by Medicare beneficiaries.

Trained student pharmacists from 6 California pharmacy schools provided expert guidance on Medicare Part D PDPs to beneficiaries through interventions at statewide outreach events. Demographic and insurance information for 2008 was collected via survey. Cost information for the beneficiary’s current PDP for 2008 and for the least expensive PDP for 2008 was obtained using the Medicare Plan Finder tool (http://www.medicare.gov).

Data were collected from 250 beneficiaries at 22 outreach events. For the cost-minimization analysis, data were excluded from 72 beneficiaries who were not enrolled in a stand-alone PDP before the intervention and from another 23 beneficiaries for whom information regarding their current PDP or prescription drug profile was incomplete. Of the remaining 155 study participants, 39.4% were male, the mean (SD) age was 74.6 (8.7) years, and they were taking a mean (SD) of 5.3 (3.5) prescription drugs each month. In addition, 68 beneficiaries (43.9%) had limited or no English proficiency, and 85 beneficiaries (54.8%) were enrolled in both Medicare and Medicaid. In total, 89.7% of beneficiaries could have realized cost savings by switching to a different PDP. The median annual potential cost savings was $98 per beneficiary but this varied as a function of subsidy level.

Targeted community outreach services to Medicare Part D beneficiaries can help optimize patient selection of a PDP, thereby resulting in lower out-of-pocket expenditures.

Targeted community outreach services to Medicare beneficiaries can help optimize selection of a prescription drug plan, thereby resulting in lower out-of-pocket expenditures.

Trained student pharmacists who are proficient in using the Web-based Medicare Plan Finder tool can systematically evaluate Medicare Part D plan choices available to beneficiaries.

Beneficiaries should reevaluate their Part D plans annually, as 9 of 10 could have saved money by doing so.

Targeted community outreach could not only lower beneficiaries’ out-of-pocket prescription drug plan costs, but also help identify those who may qualify for additional governmental assistance (receipt of the low-income subsidy).

Medicare Part D (Part D), the prescription drug program for Medicare beneficiaries, went into effect on January 1, 2006.

 In its third year of operation, 25.4 million beneficiaries were enrolled in a Part D plan, with 68% of enrollees in a stand-alone prescription drug plan (PDP) and the remainder in a plan that combines prescription drug coverage with hospital and medical coverage (ie, Medicare Advantage Prescription Drug plan).

 The program has enjoyed some important successes. Since the introduction of Part D, the proportion of Medicare-eligible beneficiaries without drug coverage has decreased by 60%.

 Research suggests that Part D has improved some beneficiaries’ access to and use of prescription medicines, while decreasing out-of-pocket expenditures.

Nonetheless, some important challenges remain. Despite having lowered out-of-pocket costs for some beneficiaries, Part D plans may still burden many with high out-of-pocket costs, creating a significant barrier to accessing prescription drugs. Compared with those having employersponsored or Department of Veterans Affairs PDPs, more seniors with Part D have high out-of-pocket prescription drug expenditures.

 More than 25% of Medicare-eligible individuals reported spending at least $100 per month on their prescription medications; these same investigators found that 8% of such individuals reported spending at least $300 per month on their prescription medications.

 of national patient-level retail pharmacy claims data found that 14% of Part D enrollees (excluding those receiving federal subsidies)—3.4 million beneficiaries&mdash;reached the coverage gap in 2007, with monthly out-of-pocket spending almost doubling from $104 before the gap to $196 during the gap. Among those who reached the gap, 15% stopped taking their medications, and 5% switched to an alternative drug in the same drug class. These findings are especially important because many Medicare beneficiaries are senior citizens with fixed incomes or limited resources, and high out-of-pocket expenses may result in increased rates of noncompliance (eg, forgoing filling a prescription or delaying to fill a prescription).

 Finally, other evidence suggests that access to the benefit remains problematic for some. Of 12.5 million most vulnerable Medicare beneficiaries who may qualify for extra help (“low-income subsidy” [LIS]) to pay for their prescription drug costs, 2.6 million (20.8%) have yet to apply for the subsidy. Many of these individuals may be unaware of the existence of or may fail to fully understand the Part D benefit and the extra help available to them.

These challenges notwithstanding, perhaps the greatest barrier to accessing the Part D benefit is the enrollment process. Part D allows the beneficiary to choose and enroll in a privately sponsored Medicare-approved PDP. However, the number of PDP plan offerings can be overwhelming for a beneficiary and differs among 34 Medicare PDP regions set by the Centers for Medicare & Medicaid Services (CMS). The number of PDPs per region in 2008 ranged from a low of 47 in Alaska to a high of 63 in the Pennsylvania and West Virginia region.

In addition, because each plan can differ with regard to premium, deductible, cost-sharing provisions, formulary coverage, and other design features, beneficiaries may find it difficult to select a PDP that best meets their needs. To enroll in the plan with the best coverage and the lowest  out-of-pocket costs, Part D enrollees need to understand factors beyond premium and deductible costs, including copayment and coinsurance structure, coverage gap (doughnut hole) protection, and catastrophic coverage costs. However, many Medicare beneficiaries, particularly low-income and other vulnerable populations, lack the knowledge and experience to select a PDP that best suits their needs. The CMS will facilitate autoassignment and enrollment of LIS and dualeligible (dual-eligible individuals are those with full Medicaid and Medicare benefits) recipients into a Part D plan. This autoassignment and enrollment is a random process, and a limited number of plans qualify for having LIS or dual-eligible beneficiaries assigned to them. These plans are called benchmark plans. The autoassignment and enrollment process fails to take into account the prescription history or medication needs of the beneficiary. However, those who receive the LIS or are dual eligible have increased flexibility when selecting a Part D plan and in fact may switch plans at any time throughout the course of the year.

For those who must choose a Part D plan or those who want to reevaluate their Part D plan annually, the sole way to compare plans is through the Plan Finder tool (http://www.medicare.gov) located on the Medicare Web site, a challenge for 54% of beneficiaries who lack Internet access.

 Beneficiaries who rely on the CMS hotline (1-800-MEDICARE [1-800-633-4227]) may find obtaining information difficult if they have low literacy skills or are not fluent in English. A recent study

 of 112 seniors, all English-speaking with Internet experience, found that 72% were unable to select a drug plan or encountered difficulties navigating the site to find drug plan information. These findings are likely to be even more pronounced among those with limited or no English proficiency, as well as those having little familiarity with the Internet.

Given their accessibility, knowledge of prescription drugs, and understanding of Part D, pharmacists are in an ideal position to serve as Part D patient advocates. In addition, the American Association of Colleges of Pharmacy10 states that pharmacy educators should teach their students to serve society as caring ethical professionals and as enlightened citizens. Finally, the most recent Accreditation Council for Pharmacy Education

 accreditation standards reflect the increasing value of training student pharmacists to provide patient-centered care during their professional graduate education. Given these factors, Partners in D, a statewide research program among 7 schools of pharmacy in California, was developed by the faculty at the University of California, San Francisco School of Pharmacy to train student pharmacists to assist underserved Medicare beneficiaries in various ways, including distinguishing which PDP may result in the lowest out-of-pocket cost, explaining plan formulary changes, describing a plan’s cost-sharing structure, and identifying duplicative or unnecessary drug therapy.

The research reported herein examines our efforts to enable seniors who were enrolled in a Part D plan to select a PDP that minimized their out-of-pocket expenditures. This study used a cost-minimization analysis (CMA) to determine whether and to what extent student pharmacists’ interventions reduce out-of-pocket PDP costs for Medicare beneficiaries.

Cost-minimization analysis is a pharmacoeconomic method that has been increasingly adopted by studies reported in the medical literature. Examination of the number of articles using CMA revealed an increase of almost 100% in the 5-year period from 2001 to 2005 (212 articles in PubMed and 657 articles in SCOPUS) compared with 1996 to 2000 (122 articles in PubMed and 366 articles in SCOPUS).

 A CMA is undertaken when outcomes between comparator groups are equivalent. In clinical studies in which a CMA is performed, investigators have assumed

 that outcomes between differing treatments or interventions were equivalent. As such, the ultimate goal of a CMA is to determine which treatment or intervention minimizes cost.

This study was a cross-sectional CMA of beneficiary annual out-of-pocket Part D expenditures. We sought to reduce beneficiary spending by identifying the PDP resulting in the lowest estimated annual cost given the beneficiary’s current prescription drug regimen. This was compared with the estimated out-of-pocket costs that the beneficiary would have incurred if he or she did not change his or her current PDP. Potential cost savings for each beneficiary were calculated as follows:

= the cost of staying in the current PDP in 2008

- the cost of the least expensive PDP in 2008.

The study intervention was performed without making any changes to a beneficiary’s prescription drug regimen. As such, outcomes were not expected to change, and a CMA was conducted.

One of the collaborating schools of pharmacy was unable to participate in this part of the study. Therefore, in total, 22 outreach events were conducted by 6 schools of pharmacy in California (Loma Linda University, University of California—San Diego, University of California–San Francisco, University of the Pacific [Stockton], University of Southern California [Los Angeles], and Western University of Health Sciences [Pomona]) between November 3, 2007, and December 31, 2007 (the end of open enrollment for Part D PDPs in 2007). Most beneficiaries may only change their Part D PDP during the open enrollment window (November 15 to December 31). Approval to conduct this research was obtained by the institutional review boards at all 6 schools of pharmacy.

Data were collected from 250 beneficiaries at the time they attended 1 of 22 statewide outreach events. Beneficiaries were excluded if it was discovered that they were enrolled in a Medicare Advantage Prescription Drug plan or had other creditable coverage provided from a current or former employer, union, or other source as good as or better than Medicare Part D (n = 72). In addition, those beneficiaries who were not recognized by the Medicare Web site when attempting to conduct a personalized plan search or whose prescription drug profile was incomplete at the time of the intervention were excluded from analysis, as their current PDP and cost information could not be determined (n = 23). Although outreach events were targeted toward vulnerable populations (low income and low educational level, limited or no English proficiency, etc), beneficiary demographic or clinical characteristics were not used as inclusion or exclusion criteria. Ultimately, data from 155 outreach participants were included in the CMA.

Beneficiary consultations were provided by trained student pharmacists throughout California during community outreach events under the supervision of faculty pharmacist preceptors. Students typically worked in pairs during an outreach event, with one completing the study-specific data collection form, while the other entered beneficiary demographic information and specific prescription drug data (eg, drug name, strength, and quantity dispensed per interval) into the Plan Finder tool. During the intervention, a personalized plan search was performed to ensure that the beneficiary’s current PDP and specific subsidies recognized by Medicare were compared with other plan options available to the beneficiary. When conducting a personalized plan search, the beneficiary’s current PDP is listed first. Thereafter, all other PDP offerings are listed in ascending order based on estimated annual cost (ie, least to most expensive). The estimated annual cost includes beneficiary copayments or coinsurance, deductible, plan premium, and, if applicable, the amount paid in the coverage gap and for catastrophic coverage.

Beneficiary prescription drug information (eg, medication name, strength, dosage, and frequency of use) was entered into the Plan Finder tool in a systematic and rigorous manner, as minor variations in such information are likely to result in the reporting of different estimated annual plan costs. If the patient had a preference for a particular pharmacy, this was entered into the Plan Finder tool as well to ensure that he or she could continue to receive care from that pharmacy. Once all of the required information was entered into the Plan Finder tool, available Part D PDP options that were available through their pharmacy of choice (including mail-order options) were presented to the beneficiary and evaluated in consultation with the student pharmacists. Plan-specific quality ratings were inconsistently reported during the 2007 open enrollment period, so these were not considered for this study.

Demographic and clinical information was collected from each beneficiary through the use of a standardized survey. The survey also provided space to record the name of and cost information for the beneficiary’s current PDP, as well as for the lowest-cost plan for 2008. The survey also contained questions designed to assess each beneficiary’s eligibility for the LIS and the outcomes of each consultation (eg, whether the PDP was changed or not). The survey was available in several different languages (English, Chinese, Korean, Vietnamese, Tagalog, Russian, and Spanish) and when possible was administered by students who were fluent in the beneficiary’s native language so as to best accommodate the needs of outreach participants.

Descriptive statistics were used to summarize demographic characteristics of the participants. Cost data were not normally distributed. As such, Kruskal-Wallis and Mann-Whitney rank sum tests were used to determine if differences in cost data existed as a function of beneficiary demographic characteristics. All significance calculations were based on a 95% confidence interval at 

 = .05. Data analyses were performed using commercially available statistical software (SPSS, version 15.0; SPSS Inc, Chicago, IL).

 gives the demographic and clinical characteristics of 155 study participants. While most were aged between 65 and 85 years, 7.1% were younger than 65 years and able to qualify for Medicare by some reason other than age (eg, end-stage renal disease or permanent disability), and 14.8% of study participants were classified as the oldest old (>85 years). Race/ethnicity was similarly diverse. The most common race/ethnicity of study participants was Asian (44.5%). When asked their preferred language, participants identified 12 different languages. More than 50% of study participants were unmarried, and most of those (55.7%) were widowed. The education variable revealed a bimodal distribution in which 40.0% of participants had less than a high school education and 42.6% had completed at least some college. The mean (SD) number of prescriptions each month per participant was 5.3 (3.5), with 2 individuals reporting taking 0 prescription medications (minimum) and 1 individual reporting taking 26 prescription medications (maximum). Most outreach participants (54.8%) were dual-eligible beneficiaries.

of beneficiary demographic and clinical characteristics showed statistically significant differences in median annual potential cost savings by race/ethnicity (

<.001), and total number of prescription medications (

 = .02). For those who reported taking fewer than 3 medications (16.9%), the median potential annual savings totaled $79 versus a median potential annual savings of $318 (>4-fold increase) for those who reported taking more than 7 medications (16.9%). The contrast is even greater when the maximum potential savings are considered, with almost a 17-fold differential ($580 vs $9849) between these 2 groups.

, dual-eligible and LIS beneficiaries were likely to incur significantly lower out-of-pocket PDP costs than beneficiaries who were not receiving subsidized financial assistance. In addition, beneficiaries who were not dual eligible or enrolled in the LIS had greater individual potential PDP cost savings (when looking at total US dollar savings) by switching to the least expensive plan identified for 2008. However, those who were dual eligible or enrolled in LIS had the highest potential savings when the proportional decrease in cost was compared with baseline. In fact, dual-eligible and LIS-enrolled individuals had a statistically significant decrease (45.6% and 48.5%, respectively) in costs from baseline compared with those who were not dual eligible or LIS enrolled (28.6%).

In total, 89.7% of beneficiaries could have realized cost savings by switching to a different PDP. Differences were observed in annual potential cost savings based on whether study participants switched their Part D PDP during the outreach intervention (

). Those beneficiaries who chose to switch plans on-site during an outreach intervention were likely to realize significantly higher annual cost savings compared with those who did not (

 <.001). Because some outreach events occurred before the onset of open enrollment (November 15), not all 155 participants were eligible to switch Part D plans during an outreach event. However, 44.7% of 123 participants who were eligible to switch plans during an outreach intervention did so. Trends between those who did or did not switch plans were independent of any beneficiary demographic or clinical characteristic (data not shown).

The median annual potential cost savings for all study participants were $98 (Table 2). However, the mean annual potential cost savings for these same 155 participants were much higher ($431 [data not shown]), reflecting the positively skewed nature of these data. We posit that the $98 median annual potential cost savings are largely due to our efforts to target the dual-eligible (Medicare and Medicaid) population through our outreach efforts. In fact, beneficiaries who were dual eligible were expected to save less (median savings, $47) by switching their PDP to the least expensive plan in 2008 than beneficiaries who were ineligible to receive some type of government assistance (median savings, $319) (Table 3). This is understandable because dual-eligible beneficiaries already receive financial assistance to offset the cost of their prescription medications. However, the potential benefit of switching to the least expensive PDP is a finding that is not limited to those beneficiaries without a subsidized Part D benefit. In fact, the expected proportional annual mean PDP cost savings to dual-eligible beneficiaries (45.6%) or to those already enrolled in the LIS (48.5%) are significantly greater than the cost savings to those who were not LIS or Medicaid eligible (28.6%). All individuals receiving extra help have a subsidized benefit at 2 of 4 LIS levels. Three LIS categories have fixed copayments for brand and generic medications. Therefore, one might expect beneficiaries in these 3 categories to have the same benefit if enrolled into any one of the available benchmark plans. However, our findings reveal that this was not the case. Significant cost savings could be achieved by recommending a particular benchmark plan, as

each had a unique formulary, and the beneficiary may have been autoassigned to a plan that did not cover all of their medications, resulting in higher out-of-pocket expenses. There are also benchmark plans that offer $0 copayments for generic medications, another potential cost savings. Therefore, while the outreach participants who were not receiving any type of governmental assistance with the cost of their prescription medications were expected to save more in absolute US dollars, the low-income beneficiaries (dual eligible or LIS enrolled) were expected to save more in relative (percentage) US dollars with regard to their annual PDP costs as a result of the student pharmacist interventions. In summary, the aggregate annual potential cost savings if each study participant (n = 155) switched from his or her current PDP to the least expensive plan identified for 2008 totaled $66,805 (cost minimization). In addition, the 55 individuals who switched plans at an outreach event were expected to save a total of $28,658 on their out-of-pocket costs in 2008.

While most outreach participants were receiving some type of government-subsidized financial assistance with the cost of their prescription medications, 36.8% (n = 57) of 155 participants were not receiving any such financial aid at the time of the intervention (Table 3). During community outreach efforts, 24.6% (14 out of 57) of outreach participants who were not already receiving any type of governmental assistance with their Part D plan costs (eg, dual-eligible or LIS enrolled) were identified as potentially eligible for the LIS. Based on income and asset information provided by study participants, pharmacy students were able to provide information about the LIS to all 14 eligible applicants and were able to help 6 of the 14 complete their LIS applications online. Enrollment in LIS is a cost-minimization strategy, as receiving the subsidy can dramatically decrease patients’ out-of-pocket costs and eliminates the coverage gap for those enrolled.

Most noteworthy is our finding that only 10.3% of participants were enrolled in the lowest-cost PDP available to them. We were able to collect feedback from 41.9% (n = 65) of beneficiaries with regard to their suboptimal plan selection. Nineteen individuals (29.2%) were autoenrolled into a plan by Medicare, 14 individuals (21.5%) decided to pick a Part D plan after consulting with family or friends, and 13 individuals (20.0%) chose their Part D plan after feedback from sources such as insurance representatives or agents, while 12 individuals (18.4%) provided other reasons for selecting their plan and 7 individuals (10.8%) reported being unaware of how they picked their plan or received help from healthcare professionals (eg, physicians, pharmacists, or social workers) when selecting a Part D plan. We posit other reasons for the extremely low percentage of outreach participants who were enrolled in the lowest-cost plan. The responsibility may lie with the participants, who may be unwilling to search for a better plan. Many beneficiaries express satisfaction with their current plan and may see no need to delve into the complexity of the Part D enrollment process to search for another plan on an annual basis.19 This is most likely the case among those for whom potential savings were modest. In addition, the Medicare program may share in the responsibility. As previously mentioned, the CMS will automatically assign or facilitate enrollment for those who receive LIS or have both Medicare and Medicaid coverage. In such cases, an individual’s medications are not matched to any plan formulary; therefore, patients may be assigned to plans that do not cover their prescription drugs. In addition, plans change on an annual basis, and an individual who was enrolled in a plan that covered all their medications one year may find that the plan does not cover all their medications or places them on a different tier the following year. Finally, the CMS will reassign a patient to a new plan if his or her plan does not fit the established benchmark threshold for the coming year. Therefore, individuals who were assigned to a plan by the CMS may not have been assigned to the lowest-cost plan. To the extent that the low percentage of lowest-cost enrollment is due to a lack of knowledge or a lack of access to proper information, the Medicare program must find ways to improve communication (eg, explanatory and educational materials and the use of appropriate media) to these vulnerable populations.

Our findings further revealed the cost of inertia for Part D beneficiaries. Specifically, participants who switched plans during an outreach intervention were likely to save significantly more (median savings, $150) than participants who did not make the switch on-site (median savings, $43) (Table 4). It can be gathered that those who chose to remain in their current PDP, despite counseling that another PDP may result in lower out-of-pocket costs, thought that the potential money saved by switching PDPs did not justify such a change.

From a policy perspective, much work remains to be done to increase the percentage of individuals enrolled in the least expensive Part D plan, thereby minimizing beneficiary out-ofpocket expenditures. Annual plan changes and reassignment by the CMS of LIS or Medicare and Medicaid recipients render the need for annual plan evaluation even more important. As our study findings suggest, student pharmacists have the ability to provide community-based interventions that reduce out-of-pocket costs among Medicare populations by guiding beneficiaries to the lowest-cost PDP plan. If these initial results are replicated in a larger second-generation study, pharmacy schools across the country should consider developing similar community-based Part D interventions into their curricula.

By expanding this training program to schools of pharmacy across the nation, academe can create a cohort of trained pharmacists who will enter practice with the knowledge and skills to assist these vulnerable populations. Given the level of savings achieved, a potential collaboration in this endeavor between pharmacy schools and the CMS seems to be a mutually beneficial partnership that should be considered.

A potential limitation of this cross-sectional study design is that PDP costs were estimated based on a patient’s drug regimen at the time of the outreach intervention. A patient’s prescription drug needs may change throughout the course of the year (eg, treatment of an acute condition); therefore, recorded costs were an estimation of the actual 2008 costs incurred. This study did not identify and record the results of other student pharmacist—driven interventions such as recommendation of an over-the-counter product or a lower-cost therapeutic equivalent. Because of the nature of this study, participant follow-up was not possible. As such, we are unable to determine how many individuals who were identified by student pharmacists as potentially eligible for the LIS did in fact go on to receive the subsidy. Furthermore, we are unaware of whether participants decided to switch plans outside of the outreach event based on the student pharmacist recommendations. Finally, though these initial results are provocative, a larger study should be conducted to determine if these findings are reproducible and generalizable.

METHODS

Study results indicate that outreach interventions performed by trained student pharmacists assist Medicare beneficiaries to effectively navigate their Part D benefit, thereby reducing out-of-pocket costs.

Cost Minimization of Medicare Part D Prescription Drug Plan Expenditures

: To evaluate whether patients with coronary artery disease (CAD) discharged from the Clinical Pharmacy Cardiac Risk Service (CPCRS) would maintain their lipid goals with use of an electronic laboratory reminder system.

: A 2-year, randomized study at Kaiser Permanente Colorado. 

: Patients with prior CAD (acute myocardial infarction, coronary artery bypass graft surgery, percutaneous coronary intervention) who had been enrolled in the CPCRS for at least 1 year and who had 2 consecutive low-density lipoprotein cholesterol (LDL-C), non—high-density lipoprotein cholesterol, and blood pressure readings at goal within 6 months before enrollment were randomized to remain in the CPCRS (CPCRS care) or to receive usual care from primary care physicians plus laboratory reminder letters (usual care). The primary outcome was maintenance of LDL-C goal at study end. The 

 test and χ2 test of association were used to assess differences in mean and categorical values, respectively. 

: A total of 421 patients (214 CPCRS care, 207 usual care) were randomized. Their mean age was 72 years; 74% were male. After 1.7 years of follow-up, the proportions of patients maintaining their LDL-C goal of <100 mg/dL were 91% and 93.1% in the CPCRS care and usual care groups, respectively (

 = .46). The proportions maintaining their LDL-C goal of <70 mg/dL were 68.6% and 56.8% in the CPCRS care and usual care groups, respectively (

: This study demonstrated that LDL-C measures can remain controlled in most patients discharged from a cardiac disease management program.

Patients with coronary artery disease can maintain their lipid levels after discharge from a disease management program.

Disease management programs for patients with heart disease improve processes of care and risk factor profiles for the patients enrolled; however, they are resource intensive.

Patients can maintain lipid levels after discharge from such programs through the use of electronic laboratory reminder letters.

Such a system of follow-up will allow disease management programs to provide care to other high-risk populations without the need for additional resources.

Treatment of elevated low-density lipoprotein cholesterol (LDL-C) is part of standard care for patients with coronary artery disease (CAD). National treatment guidelines recommend an LDL-C goal of <100 mg/dL for all patients with CAD and a goal of <70 mg/dL for those at particularly high risk.

 In addition, the National Committee for Quality Assurance Healthcare Effectiveness Data and Information Set (HEDIS), a tool used by the majority of health plans to measure performance on various dimensions of care and service, includes the proportion of patients with cardiovascular conditions who have LDL-C values of <100 mg/dL as part of the performance measures.

Statins are the mainstay treatment option for lowering LDL-C and have been shown to improve clinical outcomes and survival among patients with CAD.

 As a result of this and other treatment strategies, there has been a decrease in both the numbers of deaths and the rate of death from CAD.

 However, real-world data consistently demonstrate that a significant portion of patients with CAD are not screened, initiated on lipid-lowering therapy, or treated to recommended LDL-C goals.

 Although the reasons for undertreatment have not been fully elucidated, healthcare provider, patient, and healthcare system factors likely all contribute.

In an effort to improve treatment patterns, multidisciplinary cardiac disease management programs have been established.

 A meta-analysis of randomized controlled trials demonstrated that such programs improve processes of care, risk factor profiles including lipid and blood pressure levels, appropriate use of secondary prevention medications, and quality of life, and result in fewer hospitalizations for the patients enrolled.

 The majority of studies evaluating these programs have focused their evaluations on a prespecified enrollment time in a program, but have not investigated the sustainability of a program’s interventions over time once patients are discharged. Thus, few studies have explored the maintenance of treatment goals after discharge from such programs.

Given limited resources within healthcare systems, care delivery must be focused and efficient in order to provide care to the greatest number of high-risk patients and to be able to expand care to other patients who could derive benefit, such as those with diabetes, peripheral arterial disease, or cerebrovascular disease. However, methods to maintain the gains achieved within cardiac disease management programs need to be evaluated because such services are resource intensive. The purpose of this study was to evaluate whether patients with CAD who were at their LDL-C goal would maintain lipid control after discharge from a cardiac disease management program.

This was a 2-year, randomized controlled study conducted at Kaiser Permanente Colorado (KPCO). Kaiser Permanente Colorado is a group-model, closed-panel, nonprofit health maintenance organization that provides integrated healthcare services to more than 480,000 members at 18 medical offices in the Denver-Boulder metropolitan area. Kaiser Permanente Colorado utilizes electronic medical records that contain all office visit, vital, laboratory, and pharmacy data. Approval to conduct the study with a waiver of informed consent was obtained from the KPCO Institutional Review Board.

In 1998, KPCO established the Clinical Pharmacy Cardiac Risk Service (CPCRS), a centralized, clinical pharmacy specialist—managed, physician-directed disease management program that utilizes a systems-based approach to focus on the long-term management of more than 12,000 patients with CAD.

 All KPCO patients with a history of acute myocardial infarction, percutaneous coronary interventions, or coronary artery bypass graft surgery are enrolled into the program through direct referral by primary care physicians, referral from a registered nurse—managed cardiac rehabilitation program, or administrative queries of 

International Classification of Diseases, Ninth Revision

 codes. The CPCRS staff work closely with primary care providers, cardiologists, and nurses to ensure that lipids are controlled, evidence-based secondary prevention treatment strategies are initiated and adhered to, and follow-up laboratory tests are completed. The CPCRS focuses on patient follow-up to ensure that appropriate lipidlowering, antihypertensive, and smoking cessation medications are initiated and titrated, as necessary. Patient care is delivered primarily over the telephone.

All interventions by the CPCRS are made using collaborative drug therapy management protocols for lipid, hypertension, and laboratory management. With regard to fasting lipid profiles (FLPs), the CPCRS reorders laboratory tests approximately 8 to 12 weeks after any changes to lipid-lowering medications or therapeutic lifestyle changes and does so until the patients have achieved their LDL-C goal. Once patients are at their goal, FLPs are ordered every 6 months until at least 2 consecutive LDL-C values are at goal, at which time patients are monitored yearly. With regard to blood pressure, the specialists review blood pressure at the time FLP results are available and adjust medications accordingly. The patient’s primary care physician is notified of any medication changes via the electronic medical record.

All patients are entered into a Web-based electronic database (HealthTrac), which is integrated with the KPCO’s administrative, laboratory, pharmacy, diagnosis/procedure, vital sign, and demographic data. The HealthTrac database is used to identify patients due for laboratory tests in order to send reminder letters and minimize missed laboratory appointments. The CPCRS has achieved high participation rates (>90% of the entire CAD population at KPCO). The service also has achieved top-tier HEDIS scores, with 98% of the population screened and 82.8% with LDL-C of <100 mg/dL, and has demonstrated improvements in both surrogate and clinical outcomes of the patients enrolled.

 Prior to this investigation, patients were monitored by the CPCRS indefinitely or until they terminated their health plan membership with KPCO.

Patients identified for inclusion in the study were those who required only yearly follow-up per the CPCRS treatment protocol. Patients had to have at least 2 consecutive LDL-C and non—high-density lipoprotein cholesterol (non–HDLC) values at goal; the most recent value had to have been recorded within 6 months prior to study enrollment. There was no prespecified required time frame between the 2 FLPs for eligibility. The LDL-C and non–HDL-C goals were defined as <100 mg/dL and <130 mg/dL, respectively, for all patients except those with diabetes, multivessel coronary disease, at least 1 recurrent coronary event, or current smokers. For these patients, the LDL-C and non–HDL-C goals were <70 mg/dL and <100 mg/dL, respectively.

Patients also were required to have controlled blood pressure within 6 months prior to study enrollment. The blood pressure goal was defined as <140/90 mm Hg for all patients except for those patients with diabetes or chronic kidney disease, for which the goal was <130/80 mm Hg.

 Patients were excluded if they had a glycosylated hemoglobin measurement of ≥9% within the previous 6 months, had dementia, died within 30 days of randomization, or had a life expectancy of less than 3 years.

All patients eligible for study enrollment were identified administratively using the HealthTrac database and randomized on September 1, 2005. Eligible patients were randomly assigned in a 1:1 fashion using a computer-generated list of random numbers to either continue to receive CPCRS care or to be discharged from the CPCRS to receive usual care from the primary care team plus laboratory reminder letters (usual care).

Patients randomized to CPCRS care continued to receive the standard care delivered by the CPCRS for all CAD patients in the program. This care consisted of clinical pharmacy specialists telephoning patients to review the results of their annual FLP, blood pressure measurements, and medications and adherence; counseling on diet and exercise regimens; and making medication adjustments to maintain treatment goals. The specialists ordered follow-up laboratory tests through the electronic medical record. Additionally, all patients were mailed letters informing them of their results, were scheduled for a follow-up FLP, and received laboratory reminder letters. Each patient contact was documented in the electronic medical record for other healthcare providers to review and was estimated to take the clinical pharmacy specialist approximately 15 minutes.

Patients randomized to usual care were discharged from the CPCRS at the time of study enrollment (September 1, 2005), and an FLP was ordered for 1 year in the future with results to be returned electronically to each patient’s primary care physician. It was the responsibility of the physician to address the results as they became available and order follow-up laboratory tests, as needed, during the study period. Patients in the usual care group received laboratory reminder letters that were generated electronically using the HealthTrac database. However, unlike the CPCRS care group, patients in the usual care group received no contact or counseling by the CPCRS specialists or letters informing them of their FLP or blood pressure results. Other than the aforementioned reminder letters, physicians received no additional clinical assistance from CPCRS specialists with respect to the usual care patients. In rare instances, physicians were able to contact the CPCRS with specific questions regarding patients with significant medication side effects. Physicians did, however, have access to regional guidelines for the treatment of CAD risk factors.

Patients were followed for 2 years (until August 31, 2007) or until the date of termination of KPCO membership, a recurrent cardiovascular event, or death, whichever came first (study end). The primary outcome measure was the percentage of patients maintaining the LDL-C goal at study end. A number of secondary outcome measures were evaluated: the occurrence of fatal/nonfatal coronary events, all-cause mortality, and any-cause hospitalization over the 2-year study period between groups. Coronary events were defined as acute myocardial infarction, percutaneous coronary intervention, and/or coronary artery bypass graft. In addition, blood pressure control, maintenance of non—HDL-C goal, and persistence with lipid-lowering therapy were assessed at study end. The lipid and blood pressure values assessed for the outcomes were those reported most recently but no later than 1 year prior to the study end date. Persistence was confirmed if a patient had a day’s supply of lipid-lowering therapy, based on the fill date and supply, which overlapped the study end date.

Based on dichotomizing the LDL-C goal and assuming that 99% of the CPCRS care group and 90% of the usual care group would remain at LDL-C goal at follow-up, a total of 150 study patients per group (300 study subjects) were required to provide a power of 80% and a 2-sided alpha of .05. To account for study subjects who terminated KPCO membership over the 2-year study period (estimated at approximately 20%), the sample size was increased to 400 patients.

All statistical analyses were performed on an intention-totreat basis utilizing SAS statistical software (SAS, Cary, NC). Baseline characteristics were reported as means, medians, and standard deviations for interval-level and ratio-level variables (eg, age, length of time between events, mean lipid levels) and proportions for nominal-level and ordinal-level data (eg, sex, cardiovascular event history). Interval-level outcome variables were assessed for normality of their distributions, and appropriate tests (eg, 

 test, rank-sum test) were used to assess differences in mean values between groups. To assess differences in proportions between groups on categorical variables, the Pearson 

 test of association was utilized. Patients were excluded from an analysis if they had missing data in the follow-up period. The alpha level for statistical significance was set at .05.

A total of 421 patients were randomized; 207 to usual care and 214 to CPCRS care. The reasons for study exclusion are detailed in the Figure. There were no differences between the groups at baseline (

). The mean age of the population enrolled was 72 years and 74% were male. The majority of patients had a diagnosis of hypertension and a history of acute myocardial infarction. The majority of patients (76%) had an LDL-C goal of <100 mg/dL and 89% had a blood pressure goal of <140/90 mm Hg. The mean time of enrollment in the CPCRS prior to randomization was 4 years. The majority of patients (79%) were followed for the entire 2-year study duration; however, 8.4% terminated membership in KPCO (average of 9 months after randomization) and 12.6% died or had a recurrent cardiovascular event (average of 12 months after randomization) before the study end. The mean duration of follow-up was 1.8 years in the CPCRS care group and 1.6 years in the usual care group (

. The overall percentage of patients with at least 1 FLP performed during the study follow-up was 92.8% (94.3% CPCRS care vs 91.3% in usual care; 

 = .21). The percentage of patients who maintained their LDL-C goal at study end was 85%. The percentages of patients who maintained their LDL-C goal of <100 mg/dL at study end were 91% and 93.1% in the CPCRS care and usual care groups, respectively (

 = .46). The percentages of patients who maintained their LDL-C goal of <70 mg/dL at study end were 68.6% and 56.8% in the CPCRS care and usual care groups, respectively (

 = .23). There were no differences between groups in maintenance of the non—HDL-C goal at study end (88.7% CPCRS care and 88.2% usual care; 

Overall, 75.7% of patients maintained their blood pressure goal by study end. A higher proportion of patients in the usual care group maintained a blood pressure goal of <140/90 mm Hg compared with those in the CPCRS care group (75% CPCRS care vs 84.2% usual care; 

 = .03). Among those with a blood pressure goal of <130/80 mm Hg, there were no differences between groups in the proportion that maintained their goal (60% CPCRS care vs 54.5% usual care; 

There were few coronary events during the study period. A total of 3.3% of patients in CPCRS care compared with 5.8% in usual care experienced a fatal/nonfatal coronary event during the study period (

 = .21). Similarly, there were no differences in all-cause mortality (8.4% CPCRS care vs 7.7% usual care; 

 = .80) or any-cause hospitalization (33.6% CPCRS care vs 27.1% usual care; 

 = .14) between groups. There were no differences in persistence with lipid-lowering therapy between groups at the end of the study (86.5% CPCRS care vs 85.5% usual care; 

This 2-year, randomized study demonstrated that both LDL-C and blood pressure can remain controlled among patients with CAD discharged from our cardiac disease management program to usual care with laboratory reminder letters. We found that the majority of patients maintained their LDLC and blood pressure goals at study end. In addition, there were no differences in the rates of recurrent coronary events and deaths at study end, although our study was not powered to evaluate these differences.

Few studies have evaluated the outcomes of patients discharged from cardiac risk reduction programs. One retrospective analysis of 68 patients (43.5% with prior CAD) evaluated follow-up and maintenance of lipid goals after being discharged back to usual care from a pharmacist-coordinated, multidisciplinary lipid clinic.

 All patients were advised to schedule an appointment with their physician within 6 months of discharge for review of their FLP, and a clinic discharge summary of these recommendations was sent to each patient’s physician. This study found that 53% and 43% of patients discharged for at least 6 months and 1 year, respectively, did not have a follow-up FLP completed; however, this study did not report the outcomes for the CAD patients specifically. Furthermore, LDL-C was maintained in only 61% of patients who had originally achieved their goal prior to clinic discharge.

 A nonrandomized study of 481 patients with CAD evaluated maintenance of LDL-C levels as a secondary outcome measure 18 months after discharge from a collaborative care intervention compared with usual care.

 During the follow-up period after discharge, LDL-C levels were maintained in the intervention group and remained statistically lower than those in the usual care group. Finally, a study of 411 primary and secondary prevention patients who were followed for an average of more than 2 years after discharge from a cardiac risk reduction program with no systematic follow-up found no statistically significant changes in LDL-C levels over time.

None of the aforementioned studies reported on the maintenance of blood pressure goals, and only one reported on the proportion maintaining their LDL-C goal after discharge. Our study provides further evidence from a prospective, randomized study that lipid levels and blood pressure can be maintained for at least 2 years.

In our study, rather than discharging patients from our CPCRS with no coordinated, systematic follow-up, we utilized an electronic laboratory reminder system for the usual care patients to ensure that, at a minimum, annual FLPs were obtained. Given the small proportion of patients with no FLPs over the 2-year study (7.2%), this system was effective. This study provides support for the position that most patients who achieve their LDL-C goal within a risk reduction program will likely maintain that goal once discharged. Although more than 90% of the patients in our study with an LDL-C goal of <100 mg/dL maintained that goal, the proportion who maintained their LDL-C goal of <70 mg/dL was much lower in both groups, although not statistically different between groups. This result suggests that a 1-year follow-up interval may not be sufficient for those patients with more aggressive LDL-C goals; and shorter periods of follow-up may be warranted to maintain these goals. Future studies are needed to address this issue.

Our study is not without limitations. We did not evaluate the length of time it took to get to treatment goals within the CPCRS prior to study enrollment and whether this had an impact on the maintenance of LDL-C control. Thus, we cannot provide recommendations as to the length of time patients should remain within a risk reduction program before being discharged. The results of this study may not be generalizable to healthcare systems that do not have electronic medical records with laboratory ordering and patient tracking capabilities. This study relied heavily on the use of these electronic systems for patient tracking and follow-up. It is possible that patients may have maintained their goals with no systematic follow-up, as suggested by prior observational studies.

 Lastly, we did not compare the economic outcomes between groups. Additional resources (eg, more intense lipid management pharmacotherapy, more frequent clinic visits) may have been required to maintain the lipid control observed.

As evidence continues to emerge supporting the aggressive use of secondary prevention medications and control of risk factors for not only patients with CAD but also other high-risk patients, healthcare systems need to determine methods to deliver comprehensive and focused care to all patients that is sustainable over time. Cardiac disease management programs deliver comprehensive and focused care. Resources typically limit delivery of care to select patients at any given time. This study demonstrates that the majority of patients will maintain lipid and blood pressure goals after discharge from such programs with minimal, systematic follow-up. The ability to discharge well-controlled patients from a disease management program may permit expanding care to other high-risk patients without the need for additional resources.

The authors would like to thank the entire staff of Kaiser Permanente Colorado’s Clinical Pharmacy Cardiac Risk Service.

: From the Pharmacy Department (KLO, TD, JR, TLH) and the Population and Prevention Services (JAM), Kaiser Permanente Colorado, Aurora; and the School of Pharmacy (KLO, TD, JR, TLH) and the School of Medicine (JAM), University of Colorado, Denver.

CLINICAL

Lipid levels can remain controlled with electronic laboratory reminder letters for most patients with coronary artery disease discharged from a cardiovascular disease management program.

Outcomes of Patients Discharged From Pharmacy-Managed Cardiovascular Disease Management

: Guided Care (GC) is a model of proactive, evidence-based comprehensive healthcare provided by physician-nurse teams for people with several chronic health conditions. Our objective was to evaluate the preliminary effects of GC on health service utilization and costs.

: Cluster-randomized controlled trial of GC involving 14 primary care teams (49 physicians) and 904 of their chronically ill patients age 65 years or older.

: Using insurance claims, we compared the health services used by patients who received GC with the health services used by patients who received usual care during the first 8 months of the study.

: After adjustment for baseline characteristics, GC patients experienced, on average, 24% fewer hospital days (95% confidence interval [CI]: 49% fewer, 13% more), 37% fewer skilled nursing facility days (95% CI: 65% fewer, 5% more), 15% fewer emergency department visits (95% CI: 38% fewer, 18% more), and 29% fewer home healthcare episodes (95% CI: 53% fewer, 8% more), as well as 9% more specialist visits (95% CI: 8% fewer, 29% more). Based on current Medicare payment rates and GC costs, these differences in utilization represent an annual net savings of $75,000 (95% CI: −$244,000, $150,900) per nurse, or $1364 per patient. 

: Initial introduction of GC into primary care practices may be associated with less use of expensive health services and a net savings in healthcare costs among older patients with several chronic health conditions. Final results from the remaining 2 years of this ongoing study will be published in 2011.

Guided Care is a model of proactive, comprehensive healthcare by physician-nurse teams for people with several chronic health conditions. If its promising early effects persist and reach statistical significance during the remainder of an ongoing cluster-randomized controlled trial and are replicated during upcoming demonstration programs, Guided Care may:

Attract public and private health insurance coverage in the future.

Become adopted widely by US primary care practices.

Healthcare for older persons with chronic conditions is very expensive.

The early months of a cluster-randomized trial suggest that Guided Care is associated with a trend toward a net reduction in total healthcare expenditures.

August 2009

A Telephone-Based Intervention for Increasing the Use of Osteoporosis Medication: A Randomized Controlled Trial

Cost Minimization of Medicare Part D Prescription Drug Plan Expenditures

Outcomes of Patients Discharged From Pharmacy-Managed Cardiovascular Disease Management

Guided Care and the Cost of Complex Healthcare: A Preliminary Report

Medicare Part D After 2 Years

Cost Evaluation of a Coordinated Care Management Intervention for Dementia

Burden of Alzheimer's Disease and Association With Negative Health Outcomes

Distal Upper and Lower Limb Fractures Associated With Thiazolidinedione Use

Can a Nationwide Media Campaign Affect Antibiotic Use?

Cost Comparison of Peritoneal Dialysis Versus Hemodialysis in End-Stage Renal Disease