Numerous studies have now found that amyloid PET scans can alter a person’s diagnosis and disease management—but do they reduce overall medical costs? According to data from the final phase of the IDEAS study, presented at the virtual Alzheimer’s Association International Conference this past week, the answer is no. Gil Rabinovici at the University of California, San Francisco, reported that patients who received a scan had about 5 percent fewer hospitalizations in the next year than did unscanned, matched Medicare participants. This small benefit missed the prespecified outcome of 10 percent, and leaves unclear whether insurers will consider the benefits of scanning to be worth the cost. Meanwhile, the researchers continue to dig through the dataset for additional insights, and IDEAS participants have joined additional studies.
- Amyloid PET scans led to a small drop in hospitalizations, but missed the prespecified endpoint.
- Researchers are still analyzing the data and running add-on studies in this huge cohort.
- IDEAS 2 will recruit a racially diverse cohort and bank DNA and plasma.
“Our goal was to provide a resource to the field, beyond just addressing the issue of amyloid PET coverage,” Rabinovici told Alzforum. “We hope this will be a platform to pioneer the implementation of biomarkers and other research advances into the real world.”
The first phase of the IDEAS study found that physicians changed treatment plans for about two-thirds of the 11,409 people who underwent amyloid scans (Aug 2017 conference news; Nov 2018 conference news; Apr 2019 news). So far, so good. In the second phase of the study, however, the researchers compared hospitalizations and emergency-room visits of 12,684 IDEAS participants with those of the same number of matched controls from the Medicare database. In the year after an amyloid scan, IDEAS participants were 4.5 percent less likely to be hospitalized than controls were, but equally likely to visit an emergency room. The difference in hospitalizations was driven by the 4,379 participants with dementia, not the 8,305 with mild cognitive impairment.
Curiously, among the IDEAS participants, having a positive amyloid scan was associated with better health outcomes afterward than was having a negative scan. For both the MCI and dementia groups, people with positive scans were 22 percent less likely to be hospitalized in the next year than were those with negative scans. The amyloid-positive dementia group was also 15 percent less likely to visit an emergency room.
One possible explanation for this counterintuitive finding is that a positive scan and Alzheimer’s diagnosis led to an improved management plan that benefited patients’ health. Participants with a negative scan, on the other hand, continued without a diagnosis for their neurological issues, and thus their conditions were less well-managed and they were likelier to need hospitalization. That idea is consistent with the medical literature for other diseases, which links the presence of a clear diagnosis and multidisciplinary treatment plan to a decline in hospitalizations, Rabinovici noted.
However, an alternate explanation is that amyloid-negative participants simply have conditions that are more likely to require hospitalization. Rabinovici said their model tried to control for that by matching participants based on previous hospitalizations. The researchers are now analyzing the nature of the hospitalizations to determine if there are consistent health differences between amyloid-negative and -positive participants.
It is unclear if these data will convince the Centers for Medicare and Medicaid Services to extend coverage for amyloid PET scans (Jan 2013 news; Jul 2013 conference news). Rabinovici noted that CMS is expected to base their decision on the data from IDEAS and other studies to date. He believes the advent of accurate blood-based biomarkers might actually make PET scans more economical on a population level, by cutting down the number of people who would need one (Jul 2020 conference news). People could first be screened with a cheaper blood test, and then only those who tested positive would go on to receive PET scans or lumbar punctures for confirmation.
“I hope the totality of advances in biomarkers will be considered in coverage [decisions], and be part of an algorithm we’ll use in the future about how to deploy these biomarkers in clinical practice,” Rabinovici said. If aducanumab or another disease-modifying treatment is approved, the cost/benefit calculation would change (Jul 2020 news).
The researchers are digging further into the IDEAS data. “This is a treasure trove of data on over 18,000 people studied in real-world memory clinics,” Rabinovici said. He believes this dataset holds answers to many questions. For example, are claims-based diagnoses used in epidemiologic studies less accurate than case-report forms from dementia specialists? What conditions do amyloid-negative memory clinic patients have? By examining racial and ethnic groups in the IDEAS cohort, the researchers have found preliminary evidence of differences between whites and people of color. White participants were more likely to enter the study at the MCI stage, while blacks and Latinos were more likely to join at the dementia stage, suggesting a delayed detection of disease. At every disease stage, whites were more likely to be amyloid-positive than were blacks and Latinos, hinting that other conditions such as vascular cognitive impairment might be more prevalent in minority populations.
The next IDEAS study will address racial disparities by recruiting a diverse cohort. Whereas IDEAS participants were predominantly white, IDEAS 2 will include at least 2,000 blacks and 2,000 Latinos among the 7,000 planned participants. Outreach will be led by Consuelo Wilkins at Vanderbilt University in Nashville, Tennessee, and Peggye Dilworth-Anderson at the University of North Carolina, Chapel Hill. The study will also collect DNA and plasma samples and bank them in a biorepository for future research on predictive genetics and blood-based biomarkers. “There’s a need to understand MCI and dementia better in these communities that seem to be at higher risk of cognitive decline,” Rabinovici noted. The study is set to start this fall, and researchers are adjusting the protocols to adapt to increased use of telemedicine and heightened safety precautions during the coronavirus pandemic.
Meanwhile, the original IDEAS cohort has proved a fertile ground for recruitment to other studies. About 2,000 IDEAS participants contributed DNA to a study led by Andrew Saykin and Tatiana Foroud at Indiana University in Indianapolis. Another 1,000 joined the Brain Health Registry run by Michael Weiner at UCSF. Others are participating in a study of how amyloid PET scan results affect caregivers, while an add-on study will compare the accuracy of the plasma Aβ test developed by C2N Diagnostics with PET scans (Jul 2017 conference news; Aug 2019 conference news). “This unique cohort will allow us to address other research questions in the field and contribute to our knowledge about AD and related disorders,” Rabinovici said.—Madolyn Bowman Rogers
- In Clinical Use, Amyloid Scans Change Two-Thirds of Treatment Plans
- Amyloid PET Aids Diagnosis. But Could CSF Do Just as Well?
- Results from IDEAS Study Published
- Not So Fast: Amyloid PET Needs More Data Before Insurance Pays
- Coverage Denial For Amyloid Scans Riles Alzheimer’s Community
- Plasma p-Tau217 Set to Transform Alzheimer’s Diagnostics
- Biogen Asks FDA To Approve Aducanumab
- Finally, a Blood Test for Alzheimer’s?
- Are Aβ Blood Tests Ready for Prime Time?