Lessons learned from implementation of a pharmacist-delivered medication therapy management intervention in primary care can inform future studies and be adopted into real-world clinical settings.
Lessons learned from the implementation of a pharmacist-delivered medication therapy management (MTM) intervention in primary care (PC) can inform future MTM studies and be adopted into real-world clinical settings. We sought to describe the variations and challenges of patient recruitment, enrollment, MTM pharmacist visits, and telephone follow-up in a 3-arm randomized trial of MTM interventions conducted at 3 health centers.
Using a post-study structured interview, we interviewed study personnel, clinical pharmacists, and investigators about 5 study domains: recruitment, enrollment visits, MTM pharmacist visits, telephone follow-up, and data collection.
All centers screened clinic schedules and conducted queries of administrative databases to identify eligible participants. Patients were recruited either during existing primary care visits or by mailing letters with telephone follow-up. Patients with many medical problems, with transportation difficulties, or who were unaccompanied by a family member were less likely to enroll. MTM visits scheduled separately from other clinic appointments had higher cancellation or no-show rates. Provider response to pharmacist recommendations was low overall but better when the provider was acquainted with the pharmacist who was making contact.
Off-site implementation of MTM services results in lower participation by patients and providers. Future MTM studies should consider integrating MTM services within the clinic during existing appointments by a pharmacist familiar to the primary care provider.
(Am J Manag Care. 2012;18(7):e238-e244)This paper describes the lessons learned from implementing a pharmacist-delivered medication therapy management (MTM) intervention.
The Medication Evaluation and Drug Use Problem Identification to Improve Safety in High Risk Medicare Beneficiaries (MEDIS-MB) study was a randomized, multisite trial of different medication therapy management (MTM) strategies conducted at the University of Illinois at Chicago (UIC), the Baylor Health Care System in Dallas, Texas, and the Duke Primary Care Research Consortium (PCRC) in Durham, North Carolina. Patients 65 years or older with 3 or more chronic illnesses, 6 or more medications, and 1 or more risk factors for development of a drug-related problem (DRP) (eg, recent hospitalization or multiple providers) were randomly assigned to 1 of 3 treatment arms: usual care; basic MTM by patient interview only; or enhanced MTM with access to a clinical synopsis of medical history, laboratory data, and medications from the patient’s medical record.1 The overall results showed that MTM reduced DRPs and increased patient satisfaction. Access to the clinical synopsis in the enhanced MTM arm resulted in fewer medication list discrepancies.1 Touchette and colleagues discussed the potential benefits of expanding MTM services to include platforms for clinical record data sharing for community pharmacist access, thus potentially improving patient outcomes by reducing adverse drug events (ADEs).2
Communicating the study implementation issues that we experienced can inform clinicians, administrators, researchers, and payers who may be interested in 1) clinical adoption of the intervention, or 2) conducting future MTM studies. In this paper, we intend to describe the variations and challenges of patient recruitment, enrollment, MTM pharmacist visits, and telephone follow-up within this comparative effectiveness trial.
The detailed MEDIS-MB study design has been previously described.1 Institutional review boards at participating health systems approved the study; written informed consent was obtained from all patients. During the enrollment phase, weekly screening and enrollment reports, including reasons for ineligibility and patient refusal, were sent from Duke and Baylor to the UIC coordinating center and reported to the study sponsor. At the end of study, the investigators developed a questionnaire of 5 study domains—recruitment, enrollment visits, MTM pharmacist visits, telephone follow-up, and data collection issues—to understand the challenges of study implementation and gather feedback on how to improve future studies.
Under the recruitment domain we asked questions about the recruitment method (mail, phone, or in person), time spent on recruitment, previous MTM program enrollment, differences in patient participation, whether incentives were adequate, consent form problems, and how the various study team members approached the clinics and providers. Under the enrollment visit domain we asked questions about the length of visit and space available. Under the MTM pharmacist visit domain we asked about scheduling the visit and the number of visits needed/required, work flow issues, contacting or faxing the provider with recommendations, access to pill bottles, and completing the clinical synopsis, medication lists, and drug-related problem surveys. Under the telephone follow-up domain we asked about the length of the phone call and issues with filling out the symptom survey scale, resource utilization survey, and pharmacist satisfaction form. Finally, under the data collection domain, we asked questions about the methods of data collection (ie, filling out, making copies of, and mailing paper forms to the coordinating center), monthly study calls, and overall clinic participation. An initial conference call with the study personnel, clinical pharmacists, and investigators was held to gather initial responses to the questionnaire items; a second call confirmed their responses and gathered additional feedback. Personnel who were unable to attend the calls submitted written responses to the questionnaire.
The MEDIS-MB study was conducted at 3 institutions to reach the enrollment goal of 600 participants (approximately 200 participants per site). UIC worked with 1 family medicine (FM) and 1 internal medicine (IM) clinic, Baylor enrolled patients from 2 senior health centers, and Duke recruited participants from 6 primary care clinics (1 FM, 5 IM) within its practice-based research network (PBRN). Enrolling at 3 institutions resulted in ethnic diversity among participants (51% black, 48% white, 1% Asian/American Indian).2 outlines the number of patients that were screened, contacted, and enrolled. Of the 3084 patients who were screened, we enrolled 637 (21%) participants. Patients were most often deemed ineligible for participation because they did not meet the inclusion criteria of taking at least 6 chronic medications (42%); patients most commonly refused to participate due to lack of interest (69.5%). Enrollnment lasted 9 months at Duke, 12 months at Baylor, and 13 months at UIC. Both Duke and Baylor enrolled more than 200 participants to help reach the total enrollment goal. Additional subjects (more than 600) were enrolled to replace patients who were lost to follow-up.
Feedback about the 5 study domains from investigators, coordinators, and pharmacists is shown in . Under the recruitment domain, study staff reported that administrative and pharmacy databases were useful for identifying eligible patients. All 3 sites initially used recruitment letters and phone calls to mimic real-world pharmacy implementation; however, Baylor and UIC switched to in-person clinic recruitment because response rates were low with the mail approach. Duke was able to continue the letter/telephone recruitment strategy given the larger patient population from 6 participating clinics. The differences in recruitment approach translated into the differences in the proportion of participants enrolled to the patients screened for this study (35.3% at UIC, 30.7% at Baylor, and 13.3% at Duke). Patients who were approached at the clinic visit and accompanied by a caregiver were more likely to participate. These caregivers viewed the MTM intervention as a benefit. The presence of a provider champion and a motivated clinic staff was felt to be useful for enhancing recruitment.
For the enrollment visit domain, patients with more chronic illnesses and multiple clinic appointments were more likely to decline participation due to perceived study visit burden. Patients who were concerned about the consent form language, cost of gas, or transportation difficulties were also less likely to participate. The free MTM intervention was perceived as a benefit; however, the study payment ($10 per completed visit, or $30 total) was perceived as too low by the patients. In addition, some patients in the control group inquired whether the MTM intervention could be offered to them after all study follow-up was completed. The amount of space available to conduct the enrollment and MTM visits varied at the 3 sites. UIC had ample space in the academic clinics, but limited space in the outpatient pharmacy. Baylor
used the clinic room space prior to the provider entering the room to see the patient, so timing was essential to limit interference with clinic work flow. Duke held all study visits at a central location that had available rooms depending on the day of the week.
In the MTM pharmacist domain, MTM visits occurring separate from an existing clinical visit had higher no-show or cancellation rates. Reminder phone calls helped reduce missed visits. A total of 186 study participants (88.6%) in the basic MTM group attended the first MTM visit, and 155 (73.8%) participants completed their second MTM visit. A similar proportion of participants in the enhanced MTM group completed their first (n = 196, 89.9%) and second (n = 165, 75.7%) visits. The first MTM visit lasted 45 to 60 minutes, and the second MTM visit lasted 15 to 30 minutes. Access to pill bottles was essential for delivering the MTM intervention because patients had trouble recalling medication names and dosages. The second visit was often unnecessary because most drug-related problems were identified during the first visit. Therefore, the second visit was often used to reinforce or educate patients on the previous medication recommendations. The DRP and ADE forms were straightforward but perceived as impractical for real-world (non-study) settings. DRP form was based on the Modified Pharmaceutical Care Network Europe (PCNE) Drug Assessment Form V 5.01 (see ) and served as both a checklist and as a documentation tool for this study.3 The ADE form was used to assess symptoms potentially related to medications and contained questions from parts 2 and 3 of a validated research tool developed by Jarernsiripornkul and colleagues.4 Part 2 () of this questionnaire assesses potential side effects of medications through a system-by-system approach. Part 3 () of the questionnaire assesses the status of the side effect if the drug was stopped. Non-study MTM providers are unlikely to be able to use these surveys for assessing DRPs and ADEs in routine clinical practice, given the length and detail of questions contained in these documents. After the MTM visit, pharmacists sent medication recommendations via facsimile to patients’ primary care providers (PCPs). Providers often returned the study facsimiles without indicating whether they accepted the pharmacist’s recommendation, or they failed to return the form. Response to e-mail, phone, or face-to-face communication better ensured receipt of the recommendation and implementation of a plan of action.
Within the telephone follow-up domain, study coordinators reported that these calls ranged from 10 to 60 minutes (average, 20 minutes); the length depended on the number of symptoms discussed. Research coordinators asked patients questions from the symptom, utilization, and patient satisfaction surveys at 90 and 180 days. The ADE symptom survey (19 survey items with multiple potential responses to each item, followed by a 10-item survey for each symptom identified by the participant; see Appendix B) was difficult to administer by phone because the interviewer had to maintain a patient’s focus on whether the symptom was related to medication. Also, patients did not refer to the paper copy of the ADE symptom surveys while they were being asked the questions by the interviewer on the telephone. A study folder with a copy of the ADE symptom survey, patient satisfaction survey, patient visit log, consent form, medication list, and contact information of study personnel was given to each patient at the enrollment visit. For the visit (clinic/emergency department/inpatient) utilization form, patients had difficulty recalling the dates of these visits, but having access to the scheduling system allowed the coordinator to find the information. Patients were given a visit log to write down their visit dates, and patients
were either fully compliant or noncompliant with the visit log. Patients with higher socioeconomic status were more likely to complete the visit log.
Finally, under the data collection domain, the coordinators noted that paper case report forms (CRFs) required them to copy and send forms to the coordinating center, which took a lot of time and effort. Suggestions included 1) using carbonless (no carbon required [NCR]) paper to keep 1 copy of the CRF at the site and send the other to the coordinating center for data entry, or 2) creating a web-based data entry system for electronic data capture (EDC). The EDC system could be complemented by telephone follow-up surveys housed in a computer-assisted telephone interview system, which would allow immediate data entry at the time of the call.
We describe our experience implementing a prospective, randomized study to inform clinicians, researchers, and funders about the challenges and successes of communitybased MTM trials. As noted above, successes included identifying potentially eligible patients via medication and billing databases, participation of 10 community clinics resulting in enrollment of a diverse patient population, and monthly study calls that helped standardize operational issues. Challenges included the need to contact numerous patients (5 screened for every 1 enrolled), identifying space to conduct and schedule the MTM visit, contacting PCPs with medication recommendations, telephone follow-up using lengthy symptom questionnaires with variability in patients’ recall of clinic/emergency department visits or hospitalizations, and a paper-based data collection system. A description of the MTM intervention and tool kit with copies of the study forms is available on the Agency for Healthcare Research and Quality website.5
Our enrollment challenges are similar to those experienced by pharmacy benefit managers (PBMs) in contacting a population at high risk for a medication-related adverse event due to multiple chronic conditions and medications. MTM interventions by PBMs who contact eligible patients by phone or mail also experience low participation by highly motivated patients and even lower participation by patients who would most benefit from pharmacy coaching.6,7 A medication review survey packet mailed to 4000 US Department of Defense beneficiaries resulted in 1469 responses (38.1%) to the consent letter, 606 consents (15.7%) to participate, and 373 (9.3%) completed surveys.8 In this study, mailed letter and telephone contact resulted in less participation (13%) than recruitment from within the clinic during existing appointments (30%-35%). Interestingly, about 70% of eligible patients stated that they were not interested in study participation. Non-participation is likely due to the presence of a consent form, required study visits, and telephone followup. Participation rates in a real-world community MTM intervention will hopefully be higher if conducted outside the context of a clinical research project.
These findings provide specific opportunities to improve the design of future MTM studies and disseminate and implement MTM within the community setting. Under study design considerations, recruiting at-risk patients during an existing clinical appointment and delivering the MTM intervention on the same day can reduce participant burden (removes the need for a visit on a separate day or location) and allow the pharmacist to contact the prescriber onsite (not via facsimile) with any newly identified problems related to medication. Flexibility for the MTM pharmacist to decide whether a second MTM visit (or additional visits for complex cases) is required and whether it is done in person or by phone can optimize the effectiveness and efficiency of visits. Lengthy patient-reported outcome surveys (drug-related adverse event reporting, satisfaction, etc) should be shortened to facilitate administration by phone or at a follow-up clinic visit. Healthcare utilization (clinic or emergency department visits, hospitalizations) is better obtained from review of an electronic health record than from patient recall. Finally, the use of a web-based data entry system is preferred to transmitting paper forms to a central location.
For the dissemination and implementation of this MTM intervention within a usual care setting, the ideal situation would be the colocation of a clinical pharmacist within the primary care clinic to deliver the intervention at the point of care. Incorporating this pharmacist as a team member within the patient-centered medical home may be feasible in integrated health systems. Smaller independent practices without the resources required for a dedicated pharmacist may consider collaboration with the MTM clinicians from PBMs, as long as this includes communication from the PCP to the patient about the importance of participation to reduce ADEs. In a usual care setting, the barriers of patient consent and lengthy study-related forms would be removed; therefore, we expect that patient buy-in would be greater than within a trial. Creation of a quality improvement strategy to reduce drug-related problems by implementing MTM would allow prospective measurement of the intervention over time.
In summary, we found that implementation of MTM services not directly linked to a primary care visit (either geographically or temporally) resulted in lower participation by patients and providers. Therefore, participation in future MTM studies or usual care implementation can be improved by integrating these MTM services within the clinic during existing appointments by a pharmacist familiar to the primary care provider.Acknowledgments
The authors acknowledge the following study sites and collaborators:
From the University of Illinois at Chicago (coordinating site), Chicago, IL:
Medication Therapy Management Clinicians: Shiyun Kim, PharmD; Jessica Michaud, PharmD; Annette Pellegrino, PharmD; Daphne Smith Marsh, PharmD; Jessica Tilton, PharmD; and Lori Wilken, PharmD.
Clinical Specialists: Vicki Groo, PharmD; and Mary Ann Kliethermes, PharmD.
Research Assistants: Ketsya Amboise, PharmD; Adriana Bautista; Zenobia Dotiwala; Yash Jalundhwala; Xiaochen Luo; Alexandra Perez; Sapna Rao; and Funda Tiryaki, PharmD.
From Duke University, Durham, NC:
Duke University DEcIDE Center Principal Investigator, past: David B. Matchar, MD.
Clinical Specialist: Philip Rodgers, PharmD.
MTM Clinician: Catherine Murphy, PharmD.
Clinical Research Coordinator: Kathlene Chmielewski.
From RTI International, Research Triangle Park, NC:
RTI International DEcIDE Center Principal Investigator, past: Kathleen Lohr, PhD.
Administrative Coordinator: Linda Lux, MA.
From Baylor Health Care System, Dallas, TX:
Clinical Specialist and MTM Clinician: Jennifer Craft, PharmD.
Research Coordinator: Nadine Rayan, MHA. Research Assistants: Louann Cole, BA; Teri Cowling, MS; Kate Newman, RN; Bruce Koehler, MPH.
Data Collection Tool Development: Robert Page, BA.
From the Agency for Healthcare Research and Quality, Bethesda, MD:
Task Order Officer: Carmen Kelly, PharmD.
Author Affiliations: From the Department of Medicine (RJD), Duke University Medical Center, Durham, NC; Institute for Health Care Research and Improvement (ALM), Baylor Health Care System, Dallas, TX; Center for Pharmacoeconomic Research (DRT, GTS), College of Pharmacy, University of Illinois at Chicago, Chicago, IL; Agency for Healthcare Research and Quality (SRS), Rockville, MD.
Author Disclosures: The authors (RJD, ALM, DRT, SRS, GTS) report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.
Authorship Information: Concept and design (ALM, DRT, SRS, GTS); acquisition of data (RJD, DRT, GTS); analysis and interpretation of data (RJD, ALM, DRT, SRS, GTS); drafting of the manuscript (RJD, DRT, GTS); critical revision of the manuscript for important intellectual content (RJD, ALM, DRT, GTS); statistical analysis (RJD, DRT); provision of study materials or patients (ALM, DRT, SRS); obtaining funding (DRT, GTS); administrative, technical, or logistic support (DRT, GTS); and supervision (ALM, DRT, SRS, GTS).
Funding Source: This study is funded under Contract Numbers: HHSA290-05-0032 (Duke University DEcIDE Center), HHSA290-05-0036 (RTI International DEcIDE Center), and HHSA290-05-0038 (University of Illinois at Chicago, Chicago-area DEcIDE Center) from the Agency for Healthcare Research and Quality, US Department of Health and Human Services, as part of the Developing Evidence to Inform Decisions about Effectiveness (DEcIDE) program. The authors of this report are responsible for its content. Statements in the report should not be construed as endorsement by the Agency for Healthcare Research and Quality or the US Department of Health and Human Services. Clinicaltrials.gov registration number: NCT00773942.
Address correspondence to: Rowena J. Dolor, MD, MHS, Duke Primary Care Research Consortium, Box 3850, Duke University Medical Center, Durham, NC 27710. E-mail: firstname.lastname@example.org. Masica AL, Touchette DR, Dolor RJ, et al. Evaluation of a medication therapy management program on patient safety in Medicare beneficiaries at high risk of adverse drug events: study methods. In: Henriksen K, Battles JB, Keyes MA, Grady ML, eds. Advances in Patient Safety: New Directions and Alternative Approaches (Vol 4: Technology and Medication Safety). Rockville, MD: Agency for Healthcare Research and Quality; 2008 (PMID: 21249958).
2. Touchette DR, Masica AL, Dolor RJ, et al. The Medication Evaluation and Drug Use Problem Identification to Improve Safety in High Risk Medicare Beneficiaries (MEDIS-MB) Study: Main Results. Bethesda, MD: Agency for Healthcare Research and Quality. In press.
3. van Mil JW, Westerlund LO, Hersberger KE, Schaefer MA. Drugrelated problem classification systems. Ann Pharmacother. 2004;38(5):859-867.
4. Jarernsiripornkul N, Krska J, Capps PA, Richards RM, Lee A. Patient reporting of potential adverse drug reactions: a methodological study. Br J Clin Pharmacol. 2002;53(3):318-325.
5. Touchette DR, Stubbings J, Schumock G. Improving Medication Safety in High Risk Medicare Beneficiaries Toolkit. Bethesda, MD: Agency for Healthcare Research and Quality. In press.
6. Dlugosz CK. Building the bridge to partnership: fostering the community and managed care pharmacy relationship. Pharm Today. 2010;16(10):80-91.
7. Brooks JM, Unni EJ, Klepser DG, Urmie JM, Farris KB, Doucette WR. Factors affecting demand among older adults for medication therapy management services. Res Social Adm Pharm. 2008;4(4):309-319.
8. Linton A, Bacon TA, Trice S, et al. Results from a mailed promotion of medication reviews among Department of Defense beneficiaries receiving 10 or more chronic medications. J Manag Care Pharm. 2010;16(8):578-592.