The primary objective is to assess the effectiveness of training a clinician to be a 'value champion' within clinical settings to decrease the use of three classes of potentially inappropriate prescription medications (PIMs) among people living with dementia (PLWD). Secondary objectives include determining if the intervention is associated with a reduction in emergency department (ED) visits or hospitalizations due to a fall, and examining five implementation outcomes: appropriateness, feasibility, fidelity, penetration, and equity.

This study is a pragmatic cluster-randomized trial to test the effectiveness of a primary care clinician value champion for de-implementing PIMs among patients 65 years of age and older with a diagnosis of dementia. Medicare Part D pharmacy claims data will be analyzed at the end of the 12-month intervention for the primary outcome, the medication possession rates (MPR) for three groups of potentially inappropriate medications: antipsychotic medications, benzodiazepines, and hypoglycemic medications (sulfonylureas and insulin). In a similar fashion, a hospital admission, or an emergency department visit for a fall will be assessed at the end of the intervention using Medicare claims data. Finally, the five implementation outcomes will be evaluated at the end of the intervention from notes entered by the value champions in project workbooks.

Primary care clinics within each of the two participating ACOs will be randomized to either the intervention or control arms of the study. Prior to random assignment, the investigators will stratify practices based on high versus low historic prescribing rates. A primary care clinician from each clinic selected for the trial in the intervention arm (n=30 across the two ACOs) will be recruited as a clinician value champion for each intervention clinic. The clinician value champion will participate in twice monthly value champion web-based training sessions for six months and then launch a 12-month initiative within the clinician value champions' clinics to reduce PIM prescribing among PLWD. Study outcomes will be assessed 12 months after the clinician value champions launch the initiative.

The hypothesis is that for each medication class, the intervention will produce clinically relevant decreases in mean possession rates of 10% of a standard deviation in patients seen in intervention clinics compared to those who are seen in control group clinics.

Background on Condition, Disease, or Other Primary Study Focus:

For people living with dementia (PLWD) the overuse of Potentially Inappropriate Medications (PIMs), those for which the potential for harm outweighs benefit, remains a persistent problem despite evidence-based guidelines supporting de-adoption. A group of geriatric experts convened by the Choosing Wisely initiative identified three classes of PIMs for PLWD: antipsychotics, benzodiazepines, and hypoglycemics (sulfonylureas and insulin) with adequate glycemic control. In a systematic review the prevalence of PIMs when cognitive impairment was reported ranged from 20.6% to 80.5%. Approximately 14.3% of Medicare Part D enrollees with dementia residing in the general community are prescribed an antipsychotic. The prevalence of potentially inappropriate benzodiazepine prescriptions has been reported to be as high as 20% among elderly persons with dementia living in the community. The proportion of elderly patients with an A1c < 7% who received a prescription for sulfonylurea, insulin or combined insulin and sulfonylurea therapies was 35.2%, 24.2% and 16.3% respectively and was as prevalent in those with dementia as in those without. Park and colleagues compared rates of prescribing low-value medications in the elderly from 2006-2015 in both traditional Medicare and Medicare Advantage. Not only was there no difference in rates between the two groups, there was also no evidence of any decline in rates of prescribing over time, including use of benzodiazepines in PLWD.

Study Rationale:

The rationale for decreasing the use of PIMs is that use in this population of patients results in a greater likelihood of harm than benefit. Documented harms in the medical literature includes falls, worsening cognitive impairment, hospital admission, functional impairment, and death.

Name and Description of the Intervention:

One clinician value champion from each clinic randomized to the intervention arm will complete a value champion training program led by the P.I. and then implement care redesign activities in the clinical practice setting to reduce the use of low value prescribing in older adults with dementia. The 6-month training phase will consist of twice monthly web-based training sessions. A recently completed Robert Wood Johnson Foundation (RWJF)-funded Value Champion Fellowship program resulted in the development of a training curriculum comprised of 10 learning modules for the training phase of the intervention and a project workbook to guide clinician value champions during the 12-month project phase. Following the 6 months of training, clinician value champions will participate in a monthly 1-hour shared learning sessions via video conference to share successes, challenges, and brainstorm solutions for 12 months (months 10-22 of the study). The investigators will invite former value champion fellows and faculty from the RWJF fellowship to participate in these meetings to support this new cohort of value champions.

Inclusion Criteria - clinician practices:

• Clinical practices with 3 or more primary care providers (defined as a primary care physician (specialty code of 08 or 11), nurse practitioner (specialty code = 50) or physician's assistant (specialty code = 97), and
• Clinical practices with clinical encounters with 10 or more Medicare beneficiaries with Alzheimer's or Alzheimer's related dementia in the base years (2019-2020).

Inclusion Criteria - Medicare beneficiaries:

• Seen by a clinician at a participating practice as evidenced by one or more evaluation and management claim,
• Continuous coverage in Medicare Parts A, B and D and no months of Part C (Medicare Advantage),
• Two or more claims with an International Statistical Classification of Diseases (ICD-10) diagnosis for Alzheimer's or Alzheimer's related dementia 30 days apart or 1 inpatient stay with a principal diagnosis of Alzheimer's.

Exclusion Criteria- Medicare beneficiaries:

• Medicare beneficiaries with a diagnosis of metastatic cancer or
• Medicare beneficiaries enrolled in hospice any time in the 6 months before the start of the intervention

There were too few participants meeting inclusion criteria for analysis at one of the two study sites (site 2).

1. Dementia prevalence was only 3% for site 2, vs 20% at site 1.
2. By the time all exclusions and eligibility criteria were applied, the number of cases got even smaller;
3. Site 2 data did not support reliable attribution to study national provider identification number (clinician identifier or NPI). We could not perform a meaningful analysis with site 2.

10 months

We surveyed clinical Value Champions administering the intervention monthly on whether any adverse events occurred with their patients receiving the intervention.

All-Cause Mortality, Serious Adverse Events, and Other (Not Including Serious) Adverse Events were not monitored/assessed in clinicians