A coordinated approach for managing polypharmacy among children with medical complexity: rationale and design of the Pediatric Medication Therapy Management (pMTM) randomized controlled trial

The trial design is conceptualized based on the Shed-MEDS model of deprescribing, which posits that adult patients with potentially inappropriate polypharmacy will benefit from a patient-centered deprescribing intervention to reduce polypharmacy and improve health [60]. With permission, the model is modified to include the broader core activity of pMTM, during which parents and providers review medication changes, continuation, proper use, monitoring, and follow-up (Fig. 2) [30, 60]. For CMC with polypharmacy, our model specifies that the pMTM intervention (which accounts for and prioritizes safety, quality of life, and parental considerations) will lead to patient-centered optimization of medications [4, 30, 61].

Following baseline patient and caregiver assessments, patients randomized to the intervention will take part in a PharmD visit for application of the pMTM intervention, occurring in person or via telehealth within 3 days prior to the planned PCP clinical visit. The pMTM intervention will be applied during a 30- to 45-min visit in which the PharmD will work collaborative with the caregiver to complete 3 major activities (Table 4).

First, the PharmD will perform with the patient and caregiver a comprehensive medication review (CMR) of the patient’s current personal medication list (PML). CMR is a systematic process of collecting patient-specific data and assessing for potential medication-related problems. Subsequent clinical decisions are reliant on accuracy of available data; therefore, the first step of CMR is to conduct a thorough medication history using all available resources, updating the patient’s PML where necessary, and documenting that such activities have occurred. The PharmD will gather a list of active medications (including prescriptions, over-the-counter medications, dietary supplements, and complementary medicines), determine and confirm active disease states, and identify providers involved in the prescribing and management of current medication therapy. Next, the PharmD will review each disease therapy with the caregiver and patient, if appropriate, to determine current goals of therapy. Caregiver understanding of goals of therapy is important to their ability to provide confident care to CMC. Education related to mismatch of therapy goals and known medication effects may be addressed at this time. Additionally, the PharmD will determine if any barriers may be affecting adherence. Barriers to adherence in CMC often include taste aversion, difficult or confusing administration techniques, burdensome dosing schedule, or medication cost, among others. Such barriers may be addressed in subsequent steps of the pMTM intervention. Following review of therapy goals and adherence, the PharmD will evaluate any available laboratory data and discuss ongoing clinical symptoms to determine current medication effectiveness or lack thereof. The PML will be appraised for potential therapeutic duplication and therapies for which resolution of symptoms may render ongoing treatment unnecessary. Common examples of duplication of therapy within this population include use of multiple NSAID agents or acetaminophen-containing products or use of multiple medications within the same therapeutic class (e.g., clonidine and guanfacine). Such duplication may cause potentiation of adverse effects or contribute to excessive medication costs without conferring additional therapeutic benefit. As the final component of CMR, the PharmD will review current patient symptoms to identify those which may be attributable to medication use or toxicity, which may result in recommendations for alternative therapy or deprescribing. Gaps in therapy for guideline-directed care of various disease states (e.g., asthma) may also be identified at this stage.

The second essential element of the pMTM intervention involves optimization of the medication regimen to address those concerns or opportunities identified within the preceding CMR. With the caregiver, a list of all ongoing concerns will be prioritized according to goals of therapy. Considerations related to safety, patient quality of life, and caregiver or family quality of life will be measured during this process and communicated within subsequent recommendations for medication optimization. Next, the PharmD will formulate a plan for recommended medication-related changes, including potential dose or frequency adjustment, discontinuation of therapy, or initiation of alternate medications to manage untreated or ongoing disease symptoms. Recommendations will be classified according to type (Table 5), and rationale will be provided. All recommendations will be included in a structured pMTM provider note within the EHR which is intended to be informative and suggestive, as well as concise and respectful. Recommendations may be provided in the form of changes recommended for urgent action by the PCP (at the impending clinic visit) or communication with subspecialists for those diseases states primarily managed by alternate providers. Expectations and recommendations for both subjective monitoring and objective labs or tests (e.g., ECG) will be communicated to the caregiver and documented within the structured pMTM provider note. Any medications changes recommended during the visit will ultimately be made at the discretion of the PCP during clinical care.

As the final component of the pMTM intervention, after the clinical visit, the PharmD will create a written, patient-centered, and caregiver-friendly MAP using a template populated from the EHR. This MAP will describe a prioritized list of specific action items resulting from the interactive pMTM consultation, which empowers the caregiver to be personally involved in the administration of the proposed optimization(s). The document was developed from the CMS standardized format to allow for tracking of patient progress, clarification of intended patient response, and documentation of the perceived clinical effects of all changes [62]. The MAP is designed assist the caregiver with resolving current drug therapy concerns and to help achieve the goals of medication treatment but is not intended to provide the level of detailed communication provided to the PCP or other healthcare providers. Items reinforcing compliance, maintaining caregiver actions, and acknowledging success in the child’s medication therapy may be included. The caregiver will be encouraged to bring the MAP with them to future healthcare visits and to request update of the document as necessary. A plan for follow-up all changes with the PharmD of other appropriate providers will be outlined within the MAP and communicated to the caregiver. Additionally, the reconciled PML (created within the CMR stage of the pMTM visit) will be provided to the caregiver to assist in the understanding of current medication treatment and the tracking of potential medication changes, such as addition of over-the-counter medications or redaction of discontinued products. Information about appropriate disposal of unneeded medications will be provided by the PharmD if applicable.

Patients randomized to usual care will undergo medication history review performed by study personnel prior to the PCP clinical visits as previously described. The goal of this medication review process is to ensure accuracy of the baseline medication-related data without recommendations related to medication management. We selected a usual care comparator because there are no current established standards for centralized medication management strategies within the population of CMC. All medications decisions for the control group are at the discretion of the PCP.

Research personnel will use a standardized approach to extract information from the EHR related to basic patient demographics (age, gender, complex chronic conditions, level of polypharmacy). Complex chronic condition data is generated using the CCC V2 published classification system based on ICD-10 diagnosis codes [5]. Prior work has demonstrated that CMC with some complex chronic conditions, such as technology dependence (e.g., tracheostomy dependence, gastrostomy tube), may be exposed to higher levels of polypharmacy, and subsequent analysis will seek to determine if the medical complexity of CMC is associated with varying trial outcomes [1].

Caregiver health literacy will be assessed through parental completion of the Short Assessment of Health Literacy (SAHL-E) [63]. This test consists of 18 items, for which participants are instructed to read a medical term aloud before associating each term to another word with similar meaning to demonstrate comprehension. In this study, scores > 14 will indicate adequate health literacy, while scores ≤ 14 will indicate inadequate health literacy. Studies of medication management in adults have demonstrated a clear link between health literacy and medication self-management skills [64, 65]. Parents with different levels of health literacy may have different levels of engagement with the pMTM intervention, especially because medication optimization is comprised of multiple activities and not solely medication discontinuation. Ultimately, interventions to manage polypharmacy must support parents of all levels of health literacy [66-73]. While the pMTM intervention uses patient-centric communication modalities, understanding differences in the outcome by level of parental health literacy will guide post-trial refinements.

Finally, attitudes toward medication management will be assessed through parental completion of the Patients’ Attitudes Toward Deprescribing (PATD) tool [74]. This scoring system consists of 15 items used to classify the participant’s feelings towards polypharmacy, their own medication history, and comfort with discontinuation of medications.

The primary outcome is the MRP count at 90 days after the clinical visit during which the PCP finalizes any clinical recommendations for either the pMTM intervention or usual care groups. Because medication changes require time to effect change in clinical outcomes, we will collect outcome measurements at 90 days after the clinical visit, consistent with adult literature using the MRP outcome [12, 13, 16-28]. Robust evidence exists to support the utility of using MRPs as an outcome to evaluate MTM. As related to this outcome, we will follow established guidelines for analyzing and reporting composite measures [75-77]. To facilitate blinded assessment of MRPs, we will generate an EHR-based clinical summary at 90 days, including the current weight, active medication list, symptom report, lab values, serum levels, and any diagnostic codes related to ADEs (Table 6). Trained study personnel will contact parents to verify the medication history, symptom data, and any adverse events or acute healthcare utilization. Blinded outcome assessments will be made by ≥ 2 pediatric pharmacists not involved in the pMTM intervention using our published standardized approach [61].

We will also measure changes in total parent-reported outcomes of symptoms (PRO-Sx) scores and 90-day acute healthcare utilization for both the intervention and usual care groups [6, 7, 78]. To ensure that symptoms are stable or improved after medication changes, we will track PRO-Sx scores, which we have experience measuring among CMC in the ambulatory setting [6, 7, 78]. Based on our prior work, it is feasible for parents to easily track symptoms from home via EHR functionality [6, 7, 78]. This will occur at scheduled time points, including the date of the pMTM visit for patients allocated to the intervention group, the date of the PCP clinical visit for all patients, and at 7 days, 30 days, and 90 days following the clinical visit. As part of the 90-day clinical summary, we will track counts of unplanned acute care utilization, including ambulatory sick visits, emergency room visits, and inpatient hospitalizations.

During the 90-day follow up period, we will assess additional exploratory outcomes including Medication Regimen Complexity Index (MRCI) scores and medication counts for all medications, including scheduled, as needed, and over-the-counter medications [79]. MRCI, a tool developed to measure medication complexity in adult and geriatric populations with polypharmacy, has demonstrated potential for application in pediatric populations and has been associated with increased acute healthcare utilization [4, 61, 80-82]. In addition to the previously described patient-level PRO-Sx symptom burden scores, parental completion of the Patient-Health Burden Scale for Family Caregivers (BSFC) and Patient Health Questionnaire (PHQ-9) will be performed at scheduled time points to measure caregiver burden and mental health [83, 84].

In addition to assessment of ADEs as described within the MRP primary outcome, several other measures of medication safety and adherence will be collected. First, the DrugBank database will be interrogated against baseline and 90-day patient medication lists for potential drug-drug interaction count [85]. High-alert medication counts will be assessed using published guidance from the Institute for Safe Medication Practices [86]. Finally, medication adherence will be measured using the Adherence to Refills and Medications Scale (ARMS) at similar study time points [87].

As defined within the RE-AIM framework of the study design, the aims of this study will address several goals which are not formally captured by the primary and secondary outcomes defined above, which primarily measure pMTM effectiveness. The outcomes that will be used to assess the impact of the pMTM intervention towards other aims are described briefly below:

Implementation fidelity will be evaluated through audio recording of a sample of visits from the intervention arm (pMTM visit and corresponding clinical visit) and the usual care arm (clinical visit). We will screen and recruit participants for recording of visits using permuted block randomization for a total of 100 audio-recorded encounters (50 encounters per arm). For in-person visits, study personnel will start the recorder and leave the room. The parent, child, or provider can stop recordings at any point. For telehealth-based pMTM study visits, audio recording will occur within the software. The audio-recorded clinical encounters will be used to compare whether the provider addresses pMTM-related components (medication review, optimizations, and action plan) during the clinical visits (binary outcome), and to estimate the time needed to implement the pMTM intervention or discuss medication-related issues (continuous outcome), focusing on differences between pMTM and usual care.

To measure aims related to pMTM maintenance, we will conduct 15 qualitative interviews with consented providers at time points including the beginning, middle, and end of the trial, for a total of 45 interviews. To reduce bias, we will attempt to interview all providers at least once during the study period. We will also attempt to interview some providers (specifically the pharmacists) at > 1 time point to evaluate how their experience with pMTM changed over time. Qualitative interview guides will be pilot tested prior to use with study subjects. We will elicit providers’ perceptions of the feasibility, acceptability, and barriers/facilitators of the pMTM intervention. At the final time point, we will specifically focus on providers’ perceptions and intentions of sustaining the interventions following the completion of the trial. Providers who participate in an interview will receive a $50 gift card. Recruitment will discontinue if ongoing analysis (described below) reveals thematic saturation [88].

Finally, to measure maintenance outcomes related to program replication costs, we will use time-driven activity-based costing approach to measure the cost related to implementation and maintenance of pMTM relative to usual care costs. Using best practices, we will develop process maps for patient/parent flow for both pMTM and usual care delivery and specify care activities and who (pharmacist, provider, other clinic staff) performs each activity [89]. The largest component of cost will be the time clinic staff devote to delivering pMTM and usual care, which we will measure using the audio recordings of clinical visits, annual surveys (questions about average time spent for pre-clinical visit preparation and post-visit documentation), and provider interviews (to explore reasons for variation) described above. Measures of time will be converted to cost using internal salaries and fringe benefits for each category of clinic staff. We will also value time using Bureau of Labor Statistics data to estimate more representative replication costs [90]. We will obtain cost information for other clinic and informatics resources directly and indirectly supporting the adoption, implementation, and maintenance of pMTM [91, 92].