A Silver Lining for Negative Dementia Trials?

Two separate clinical trials of dementia care programs report negative results in terms of patient neuropsychiatric symptoms and caregiver strain or distress, but the news may not be all bad.

Experts are pointing to the COVID-19 pandemic and the choice of primary outcomes as potential reasons for the null results.

However, while the overall trial results were disappointing, there were some encouraging signs. For example, the x study, which examined investigating dementia palliative care, showed emergency department (ED) use and hospitalizations were cut in half for those receiving this intervention compared with usual care.

The Dementia Care (D-CARE) study showed caregivers assigned to a targeted healthcare-based or community-based intervention felt significantly more empowered than those receiving usual care.

“The take-home message is not to expect these interventions to have dramatic effects on patient outcomes or caregiver strain and distress and depressive symptoms, but it may empower caregivers to better care for their loved one,” investigator of the D-CARE study, David B. Reuben, MD, David Geffen School of Medicine at the University of California, Los Angeles, told Medscape Medical News.

Findings of both studies were published online on January 29 in JAMA.

Good Science

As the population continues to age and dementia becomes increasingly prevalent, researchers and policymakers are trying to develop interventions to address both the medical and social support needs of affected families. In 2024, the Center for Medicare and Medicaid Innovation released an alternative payment model called Guiding an Improved Dementia Experience. It offers Medicare payments to interdisciplinary care teams, provides training for care partners, ensures access to 24/7 dementia support, and respite care.

The D-CARE study included 2176 individuals living with dementia (mean age, 80.6 years; 58.4% women) and their caregivers (mean age, 65.2 years; 75.8% women) at four sites. Almost 21% of all participants were Hispanic or Black.

Researchers randomly assigned individuals (in a 7:7:1 ratio) to one of three treatment approaches — health system–based care, community-based care, or usual care.

In the health system–based group, dementia care was delivered by advanced practice providers, mostly nurse practitioners, working within the health system to provide coordinated care. They collaborated with partnering clinicians and connected with community-based organizations to arrange additional services as needed. These practitioners were authorized to write orders and maintained open communication with physicians, said Reuben.

Initially, care was provided in person until COVID hit after, which it was delivered remotely.

Community-based care was provided over the phone by a social worker, nurse, or licensed therapist affiliated with a community-based organization partnered with a D-CARE clinical site. A significant portion of the care involved “coaching,” noted Reuben.

Participants in the usual care group did not receive any additional treatments as part of the study.

The primary outcome for those living with dementia was caregiver-reported severity on the Neuropsychiatric Inventory Questionnaire (NPI-Q), with scores ranging from 0 to 36 and higher scores indicating more severe symptoms. Here, group differences were: 0.30 (97.5% CI, −0.18 to 0.78; P = .33) for health system vs community-based care; −0.33 (97.5% CI, −1.32 to 0.67; P = .46) for healthcare system vs usual care; and −0.62 (97.5% CI, −1.61 to 0.37; P = .33) for community-based vs usual care.

The primary caregiver outcome was the Modified Caregiver Strain Index with 13 items (score ranges from 0 to 26, with higher scores indicating greater strain). Results comparing the three groups were also insignificant.

The two care models investigated in this study were shown to be effective in previous research. However, several factors could explain the trial’s negative results. One significant factor was the impact of COVID-19, during which many community-based organizations had to shut down, making it impossible to deliver key behavioral interventions, explained Reuben.

In addition, participants in the usual care group may have had access to more resources than is typical in clinical settings, such as the Alzheimer’s Association helpline. Trial sites may also have implemented quality improvement efforts, such as enhanced care coordination, which could have contributed to better outcomes for those in the usual care group.

The choice of primary outcomes — patient neuropsychiatric symptoms and caregiver strain — may also have affected between-group differences, said Reuben. While these symptoms typically increase over time, in D-CARE, they remained relatively stable over 18 months in all groups.

The negative results were “somewhat surprising,” Reuben admitted. “We came into it with expectations — or what we call hypotheses — about how the three different arms of the study would perform and the comparisons between them.”

“[While] it would have been great if one of the interventions, or both of them, hit a home run,” this is sometimes what good science looks like, said Reuben. “Good science is coming up with a well-designed study and letting the chips fall where they fall. Good science is respecting the results, interpreting them, and going on to the next step.”

However, both interventions led to greater caregiver self-efficacy — feeling more empowered or in control — over 18 months than usual care. Differences vs usual care were 0.70 (95% CI, 0.26-1.14; P = .01) for health system–based care and 0.85 (95% CI, 0.42-1.29; P = .001) for community-based care. The differences between the two treatment groups did not differ significantly.

Reuben said his team plans to look at the use of healthcare services, nursing home placement, and costs of care among study participants. “Even if the interventions didn’t have a benefit on clinical outcomes, if it saves money or reduces nursing home placement or allows more days spent at home, then it’s definitely worthwhile,” he said.

Lower ED Visits, Hospitalization

The second paper, the Indiana Palliative Excellence in Alzheimer Care Efforts (IN-PEACE) study, compared usual care with a dementia care program integrating palliative care among 201 community-dwelling individuals with dementia and their caregivers at two Indiana sites.

Participants had a mean age of 83.6 years, 67.7% were women, and 96% had moderately severe to severe dementia. Caregivers had a mean age of 60.5 years, and 81.1% were women. Over 40% of both patients and caregivers were African American, a notably higher percentage than in past studies.

Patient-caregiver pairs (dyads) were randomly assigned to palliative or usual care groups. The palliative care intervention involved a trained care manager who focused on dementia management, nonpharmacological symptom approaches, and strategies like advance care planning and hospice care. Usual care dyads received educational materials from the Alzheimer’s Association and other resources.

Like the CARE-D study, the primary outcome was severity of symptoms on the NPI-Q. The difference between the intervention and usual care at 24 months was −0.24 (95% CI, −2.33 to 1.84). The rate of change in NPI-Q severity scores from baseline did not differ between groups over time (P = .87 for the group × time interaction).

Results of secondary outcomes of caregiver distress assessed by the NPI-Q distress scale and caregiver depression assessed with the Patient Health Questionnaire-8 depression scale were also not significant.

Several factors could explain the “disappointing” results, lead investigator Greg A. Sachs, MD, of the Indiana University Center for Aging Research and Regenstrief Institute, Indianapolis, told Medscape Medical News.

One key issue was that both the intervention and usual care arms had very low symptom levels at baseline and throughout the study. “If you don’t have much in the way of symptoms, it’s hard to show a difference — it’s the so-called ‘floor effect,’” said Sachs.

He also noted a potential selection bias, as patients with significant symptoms or caregivers experiencing high distress may have been less willing to commit to a study requiring regular assessments over 2 years.

“We may have had sort of a selection bias; some people who were already struggling may not have wanted to enroll, even though we think we could have helped them,” said Sachs.

Additionally, many best practices in dementia care may already have been “adopted into practice,” potentially reducing the observed impact of the intervention. Sachs further pointed out that participants in this study were older, frailer, and had more advanced dementia compared with those in most previous dementia care management trials.

But here again, there was something of a silver lining. Patients receiving dementia palliative care had lower mean ED and hospitalization events (combined) compared with usual care (mean, 1.06 vs 2.37 with a between-group difference of −1.31; 95% CI −1.93 to −0.69; relative risk, 0.45; 95% CI, 0.31-0.65; P

“This is a very important outcome that we’re very excited about,” said Sachs, adding this outcome was prespecified for the study. “It means avoiding the disruption of going to the hospital and the kinds of complications and decline of function that often accompany being hospitalized.”

From a health system perspective, reducing emergency room visits and hospitalizations by half “has significant potential for cost savings,” Sachs noted.

A post hoc analysis revealed that the reductions in ED visits and hospitalizations were more pronounced among patients with more severe baseline neuropsychiatric symptoms, those with lower incomes, and African Americans.

Patients with higher baseline neuropsychiatric symptoms appeared to benefit more from the intervention, suggesting this group could be a focus for future studies, said Sachs.

Rather than enrolling based solely on dementia stage, targeting those with significant symptoms at baseline might yield better outcomes, he noted.

For newly diagnosed or mildly affected patients, the intensive intervention “is likely more than what they need,” Sachs added. “It’s something to implement as the disease progresses and symptoms, caregiver distress, and hospitalization risks increase.”

The study’s mortality rate was high, with about 40% of dementia patients dying during the study period. Additionally, as participants were from a single metropolitan area and managed by two care managers, the findings may not be generalizable to other settings.

New Insights

In an accompanying editorial, Julie P.W. Bynum, MD, from the Department of Internal Medicine and the Institute for Healthcare Policy and Innovation at the University of Michigan, Ann Arbor, observed that “it’s reasonable to think” COVID-19 influenced the trials’ outcomes, as both were conducted at least partially during the pandemic. She noted that neither study was designed to address additional stressors, such as reduced access to medical care and paid caregivers, introduced by the pandemic.

Also potentially contributing to null results was the choice of neuropsychiatric symptoms as the primary outcome, Bynum noted. “Conceptually, just as we do not expect a care management program to slow progression of cancer, a dementia care program may not influence incidence of these symptoms,” she wrote.

It may be time to reconsider whether symptoms, which are highly variable across participants, should be de-emphasized in favor of consequential long-term outcomes, like nursing home placement, she said.

Meanwhile, “each study provides encouraging insights that help us move forward,” she said. For example, she noted nonsignificant point estimates in the D-CARE study were in the direction of benefit for the primary outcomes in people with more severe dementia, when the caregiver was a spouse, or the caregiver had greater strain.

In the IN-PEACE study, certain subgroups — participants with worse baseline NPI-Q scores, those with low income, and African Americans — experienced even greater reductions in combined ED visits and hospitalizations.

“These findings suggest that targeting of trials toward later-stage disease or dyads with more strain and using a larger usual care group may be needed to show definitive results.”

Bynum suggested an alternative approach might be a series of “stepped interventions” deployed as appropriate for the stage of disease and based on risk of negative experience, “rather than a single, albeit multifaceted, intervention offered to everyone with a dementia diagnosis,” she wrote.

The D-CARE study was funded by the Patient-Centered Outcomes Research Institute and the National Institute on Aging (NIA). The IN-PEACE study was supported by the NIA of the National Institutes of Health. Sachs reported receiving honoraria for serving on the National Pharmacy and Therapeutics Committee of CVS Health, and grants from the NIA outside the submitted work. Reuben and Bynum reported no relevant conflicts of interest.