Telepharmacy and Its Impact on Formulary Adherence
Telepharmacy is comparable to usual care in transitioning members new to an HMO onto formulary medications.
Formularies are used to promote the utilization of preferred medications based on cost, safety, and/or efficacy data. Formularies typically use 2- to 4-tiered systems to incentivize patients to use lower-cost medications by varying patient copayment by tier. Patient and prescriber awareness of formulary requirements can maximize formulary acceptance and adherence. Nevertheless, little information exists on factors associated with the use of formulary agents (formulary adherence).
The use of an electronic medical record (EMR), including e-prescribing, can provide guidance to prescribers to help promote formulary adherence.1 One retrospective analysis reported an association between the implementation of an eprescribing system and an increase in formulary adherence to H2-blockers from approximately 14% to 88%.2 Other factors associated modestly with formulary adherence include step therapy, >$15 copayment differentials between formulary and nonformulary agents, and patient outreach through mailings, the telephone, or the Internet.3,4 Additionally, pharmacy claim data reviews by pharmacists have been used to inform and direct prescribers when nonformulary medications can be changed to formulary alternatives.5 Conversely, an increasing count of current prescriptions and having a 3-tier copayment structure were associated with a small reduced likelihood of switching to a formulary medication.6 One consideration regarding these studies is that they were performed among existing health plan members, and information on clinical factors associated with formulary adherence was lacking.
New members transitioning into a health plan (ie, from another plan, fi rst time with insurance, or returning after a period of nonmembership) pose additional challenges to the encouragement of formulary adherence. Limited pharmacy and medical background information are available to physician prescribers when such patients present. Survey research suggests that physicians are more likely to prescribe a formulary medication for their new patients compared with their established patients.7 However, prescribers report that patient visits for new members are more difficult than those for established patients, and the time used for evaluating new patients and their medications prolongs wait times and adds to staff overload.8 In addition, national survey data indicate that physicians perceive new patients to require approximately 10% to 30% longer visits than established patients.9,10 Furthermore, 83% of HMO physicians reported in a national survey that they needed more time than allotted with new patients.11 Internal Kaiser Permanente evidence indicates that primary care providers (PCPs) require an additional 15 minutes of office visit time when seeing new patients who receive chronic medications. When clinical pharmacists offer medication recommendations for patients, prescribers will accept them.5,12 This evidence suggests that pharmacists and,
by extension, telepharmacy may have a role to play as lower-cost alternatives to and time-savers for PCPs in formulary promotion and adherence.
At Kaiser Permanente Colorado (KPCO), a groupmodel HMO, an existing telepharmacy infrastructure in the Clinical Pharmacy Call Center (CPCC) and an EMR provided an opportunity to facilitate the transition of members receiving chronic medications onto the HMO’s formulary.13 A pharmacist-managed service was developed utilizing the clinical abilities of the CPCC. The CPCC pharmacists work collaboratively with clinical call center physicians to change nonformulary medications, as appropriate, to formulary alternatives. The purpose of this investigation was to assess the feasibility of telepharmacy for transitioning new members of an HMO to formulary medications. We compared a telepharmacy service’s rate of formulary adherence with the rate for usual care, where PCPs manage the transition of new members receiving chronic medications.
METHODS
Study Design
This was a retrospective health services research investigation. It used a comparator group to assess formulary adherence among patients receiving chronic medications whose medications were transitioned into an HMO’s system either via telepharmacy or traditional usual care mechanisms. While a matched study design for a retrospective analysis would have been preferable, clinical and demographic factors decidedly linked to formulary adherence to use in a valid matching schema were not identified in the literature.14 This study used data from queries of integrated electronic medical, pharmacy, and administrative record databases and was reviewed and approved by the KPCO Institutional Review Board prior to data collection.
Study Setting
This study was conducted at KPCO, a not-for-profit HMO with approximately 480,000 members. Within KPCO, CPCC was established in 1996. The Clinical Pharmacy Call Center is a centralized service composed of 24 clinical pharmacists and 6 pharmacy technicians (at the time of this study) who work collaboratively with physicians, nurses, patients, and other healthcare providers.13 During the study period, KPCO used 2 formularies: the commercial and Medicare Part D formularies.
Intervention
Each year, approximately 50,000 to 70,000 members are transitioned into KPCO. New members to KPCO receive a welcome packet instructing them to contact CPCC if they are currently receiving a chronic medication. When scheduling a visit, new members are asked if they are on any medications and are directed to CPCC if the reply is “yes” and they agree to a transfer. When calling the KPCO advice line, the member is prompted to choose from a selection of menu items and may be transferred to CPCC if they (1) need prescription(s) filled or (2) have questions about medications. Additionally, pharmacy staff and PCPs may make recommendations for members to contact CPCC.
For new member transition interventions (NMTIs), CPCC clinical pharmacists call the member to gather patient information on medical history, current health problems, allergies, previous adverse drug reactions, and current medications. The pharmacists then review the information to assess for (1) potential drug-related problems, (2) opportunities to improve drug therapy, and (3) opportunities for therapeutic interchange to formulary alternatives. If patients were prescribed medications that are not on the KPCO formulary, CPCC pharmacists make recommendations to prescribers to change such medications to appropriate formulary alternatives.13 It is estimated that NMTIs take an average of 20 minutes of visit time with a patient.
If a new member did not contact CPCC, his/her medications would be reviewed during usual care. Usual care entails a member being seen by a PCP who would then be responsible for changing any nonformulary agents to formulary agents, when appropriate, with guidance from the KPCO EMR and/or pharmacy staff. Primary care physicians at KPCO do not provide pharmaceutical samples to patients.
All new members have to go through NMTI or usual care to receive prescription medications through KPCO. Previous internal auditing and surveying of PCPs indicated that new members transitioned via usual care had better than 90% formulary adherence and that the average member receiving chronic medications required approximately 15 minutes of additional physician time to evaluate and transition. For the purposes of this investigation, members transitioned through CPCC were assigned to the
intervention group (NMTI group), and those who were transitioned by a PCP were assigned to the control group (usual care group).
Study Population
Patients were included in this study if they (1) were members newly enrolled in KPCO between July 1, 2007, and June 30, 2008, with no KPCO membership during the 6 months prior to enrollment; (2) were 18 years of age or older at membership start date; (3) had purchased at least 1 chronic medication (eg, antiretroviral, hormone, antihypertensive, antihyperlipidemic, antihyperglycemic, antiasthmatic, antiulcer, anticonvulsant, mental health) in the 6 months after membership start; and (4) had continuous benefit eligibility during the 6-month period after membership start. Patients were excluded if they had nontraditional HMO benefit plans (ie, point of service, Medicaid, high deductible, plans with no prescription benefit) or if they were KPCO members outside of the Denver-Boulder market.
Outcomes
The primary study outcome was the formulary adherence rate during the 6-month period after membership start. The formulary adherence rate was defined as the percentage of formulary chronic medication purchases compared with all chronic medication purchases during the 6-month follow-up period. Secondary outcomes included formulary adherence rates for the antihypertensive, antihyperlipidemic, and antihyperglycemic medications. These therapeutic classes were chosen to be examined
since they are commonly prescribed chronic medications. In addition, we describe therapeutic classes of the nonformulary agents prescribed during the study period.
Data Collection
Information on patients who received NMTI was obtained from CPCC’s electronic call records. Information on membership, demographics, benefit plan, medical office visits, and ambulatory medication purchases during the 6 months after membership start was obtained through queries of the KPCO electronic integrated records databases. Chronic medications were determined
by a clinical review of generic product identifier codes. Controlled substances and medications with both acute and chronic indications (eg, antibiotics, corticosteroids) were excluded.
Data Analysis
Based on an estimate of 91% formulary adherence rate among patients in care settings with an EMR,15 400 patients would be required for each group to attain a power of 80% with an alpha of .05 to observe ±4% difference in adherence rates between groups. Patient age was determined as of the first medication purchase date after the membership start date. Prior KPCO membership was assessed for its influence on medication selection. Medicare beneficiaries at the time of membership
start were identified since they had a more expanded formulary than commercial members and could be provided with a temporary 30-day supply of a nonformulary medication when transitioning into a new health plan. Owing to the myriad pharmacy benefit designs available during the study period, benefits were categorized as ≤$10 generic copayment, >$10 generic copayment, and deductible plus copayment. The count of healthcare personnel encounters (both telephonic and office visits, but
excluding CPCC telephone call encounters) prior to the first purchase date and during the 6 months after membership start was calculated for each patient. In addition, we identified patients who initiated membership in January 2008, as this was the month with the highest volume of membership transition and possibly was a time period during which the increased transition processing allowed for greater approval of nonformulary agents.
Medications were classified as formulary/nonformulary as well as into specific therapeutic classes. Additionally, a Chronic Disease Score (CDS) was calculated for each patient based on ambulatory prescription medication data for the 6 months after the membership start date. The CDS is a validated measure ranging from 0 to 35, with a higher score representing increased burden of chronic disease and therefore an increased risk of mortality and healthcare utilization.16 In addition, the count of unique prescription medications (including all individual acute and chronic medications) at the first medication purchase date after the membership start date was determined for each patient.
The analysis was performed at the unique medication level (ie, one patient could contribute more than 1 purchase observation to the data). Nominal and ordinal data are reported as percentages, with differences between the groups assessed using the Pearson’s χ2 test of association. Interval data are reported as means, medians, and intraquartile ranges, with differences assessed by t test and rank-sum test as appropriate based on the normality of distribution of the data. The percentage (ie, rate)
of nonformulary purchases was calculated along with the 95% confidence interval of the rate estimate. Rates were compared between the groups using conditional logistic regression to allow for adjustment of the intracorrelation of patients who purchased more than 1 agent and potential confounding variables identified in the bivariate analyses. All analyses were performed with SAS
version 9.1 (SAS Institute, Cary, North Carolina).
RESULTS
A total of 28,950 new members met the inclusion criteria (
Figure
). Of these, 6600 and 7723 were excluded for being outside the Denver- Boulder KPCO market and not having a traditional HMO plan, respectively. Thus, the records of 14,627 patients
were included, with 5128 and 9499 patients assigned to the NMTI and usual care groups, respectively. The NMTI group patients had a higher median age (P <.001), count of unique medications purchased at first purchase date (P <.001), count of healthcare personnel encounters during the 6 months after membership start date (P <.001), and CDS (P <.001); and they were more likely be Medicare beneficiaries (P <.001) and have a $10 or less generic medication copayment (P <.001) (
Table 1
). Patients in the usual care group were more likely to have had a healthcare personnel encounter prior to their first purchase date (P <.001), to have had a prior KPCO membership (P <.001), and to have a pharmacy benefit with both a deductible and copayment (P <.001).
A total of 73,966 chronic medication dispensings were made to the groups during the 6-month follow-up period (37,641 and 36,325 to NMTI and usual care patients, respectively) (
Table 2
). Formulary adherence rates were high for both groups (99.3%), with no significant difference in the rates between the 2 groups (unadjusted P = .491; adjusted for age, sex, CDS, being a Medicare beneficiary, copayment level, count of medications at first purchase, healthcare encounters, and previous KPCO membership, P = .061). Formulary adherence rates were equivalent between the groups in the antihypertensive (adjusted P = .572), antihyperlipidemic (adjusted P = .843), and antihyperglycemic (adjusted P = .073) therapeutic classes.
Nonformulary agents purchased (n = 523) were primarily in the endocrine (4.2% of nonformulary purchases), antihyperlipidemic (14.7%), antiulcer (17.4%), antidepressant (11.7%), psychotherapeutic (22.9%), and anticonvulsant (6.2%) therapy classes. Formulary adherence rates for these therapy classes were equivalent between the groups, except for antidepressants, for which
NMTI patients had a lower rate of formulary adherence (99.1% vs 99.6%, P <.001).
DISCUSSION
In this retrospective, health services research study of more than 14,000 new members in an HMO, we found equivalent and very high rates of formulary adherence among both new members who received a telepharmacy NMTI and members who were transitioned via usual care. The comparable rates identified in the telepharmacy and usual care observation groups support using
telepharmacy physician extenders to change new members to appropriate formulary alternatives and/or maintain new members on formulary agents. We are aware of no other studies that have examined the transition of new members into an HMO and their subsequent formulary adherence with or without a telepharmacy intervention.
High formulary adherence rates (>98%) were identified in the antihypertensive, antihyperlipidemic, and antihyperglycemic therapeutic classes in both groups. This result suggests that even among chronic medications with many nonformulary agents available, telepharmacy was able to provide rates of formulary adherence comparable to those obtained with usual care. We identified a lower formulary adherence rate in the NMTI group for the antidepressant therapy class. This may have occurred
because CPCC pharmacists were not as likely to change mental health medications, preferring to have the patient be seen by his or her PCP or mental health specialist prior to switching. Nevertheless, we did not find rate differences between the groups in the psychotherapeutic therapy class.
High formulary adherence was achieved by NMTI even though patients who chose to go through the call center tended to be older with an increased burden of chronic disease and accompanying prescription load. Our formulary adherence rate was similar to that reported in a primary care practice setting in Grampian, Scotland, where the rate for antihypertensive medications was also in the 90% range.15 As KPCO PCPs have reported that the average new member receiving a chronic medication requires 15 minutes of additional office visit time and that NMTI requires 20 minutes of visit time, salary savings from NMTI, using 2010 mean salary estimates for PCPs (~$86.53/ hour)17 and clinical pharmacists (~$54.55/hour),18 were estimated at ~$3.42 per new member transition. Thus, the 5128 NMTIs performed during the study equated to ~$17,537 in salary savings. In addition, we estimated that the NMTIs performed during the study period avoided 1282 hours of PCP visit time (approximately 5128 office visits at 15 minutes per visit). Reducing demand on physician office visit time is likely to become increasingly important as the shortage of PCPs increases and puts PCPs’ time at a premium.19
Two studies have reported that patient outreach increases formulary adherence.3,4 In KPCO, all new members (and by extension, their dependents) received a letter informing them to contact CPCC if they were receiving medications so that the medications could be brought into the KPCO system. However, only those patients who initiated contact with CPCC received further outreach. Thus, it appears that the additional outreach by CPCC resulted in formulary adherence rates comparable to those obtained with usual care and that this outreach might be able to reduce PCP office visit time.
Telepharmacy allows for the provision of pharmaceutical care to distant patients through the use of telecommunications and information technologies.20,21 Although there are limited data examining the impact of telepharmacy, patient satisfaction with services and improved clinical outcomes have been found to be associated with its use.22-24 This study adds to that limited knowledge base by providing additional context for the use of telepharmacy services to augment healthcare service and potentially improve healthcare efficiency.
Although telepharmacy may provide a new avenue for patient care, there are challenges because telepharmacy providers do not see patients face-to-face. As a result, these providers must rely on verbal information provided by the patient (eg, patients might read from medication labels rather than show the provider the bottle directly), which might result in the patient feeling isolated from the provider. Despite this potential drawback, 94.6% of survey respondents enrolled in a telepharmacy lipid management
service reported satisfaction with the care they received from the service. Furthermore, 86.7% and 90.0% of patients reported that they were content receiving care over the telephone or by mail, respectively.25
This study is not without limitations. We did not match intervention and observation patients on propensity to receive a formulary medication because information on reliable factors to perform such matching was unavailable. Nevertheless, we did perform adjusted analysis on the formulary adherence rate in an attempt to ameliorate any potential confounding from nonequivalent clinical and demographic characteristics between the groups. Because the study design was retrospective, we were unable to evaluate the medications members were receiving prior to becoming KPCO members. Therefore it is unknown whether
patients were already receiving formulary medications. However, it has been reported that 70% of new members in a managed care health plan were receiving a branded (nonformulary) medication at the time of transition into the plan.26 We were unable to assess whether therapeutic interchange to formulary alternatives had untoward consequences (eg, adverse reactions, overdosing/underdosing). Selection/channeling bias may have occurred in that older, sicker patients with a greater perceived need for a nonformulary agent would have initiated contact with CPCC. Despite this potential bias, we found that the management such patients received from NMTI was comparable to that for patients receiving usual care. In addition, primary care clinical pharmacy specialists are located at each medical office in the KPCO Denver-Boulder market. We were unable to assess whether and/or how often primary care clinical pharmacy specialists were consulted by PCPs to make formulary recommendations for new members. Furthermore, we assessed formulary adherence at the first purchase only in the 6 months after patients became members and did not determine whether members subsequently switched to a nonformulary agent. A subsequent switch could result in the patient incurring unaccounted-for PCP office visit time. Additional research opportunities include assessment of the economic and clinical outcomes and patient-reported preferences regarding this transition.
CONCLUSIONS
To our knowledge, this is the first study that evaluated the rate of formulary adherence among members of an HMO with a pharmacist-managed telepharmacy program. Our study found a very high rate of formulary adherence for both the telepharmacy and usual care groups. Our findings support the use of telepharmacy to transition members receiving chronic medications into an HMO, as telepharmacy may decrease PCP time requirements for new patients and result in salary savings for health plans. Future research is needed to elucidate further the role of telepharmacy with regard to physician time and salary savings and prospective clinical outcomes.
