People who declined to participate in a prospective study of aging developed dementia more quickly than those who joined the study, a researcher said here, suggesting study participants may not be a representative sample.
Among participants in the study, dementia was diagnosed at a rate of 19.1 per 1,000 person-years, compared with 23.6 per 1,000 person-years among those who refused participation but who had previously allowed their medical records to be reviewed for research purposes, said David Knopman, MD, of the Mayo Clinic in Rochester, Minn.
The difference may be important for other prospective studies of cognitive decline, insofar as it pegs study participants as not fully representative of the population at large.
In particular, said Knopman, who presented the study at the American Academy of Neurology’s annual meeting, the findings indicate that individuals with undiagnosed cognitive impairments are more likely to decline participation in prospective research.
The comparison – which Knopman said was unique, as far as he knew – was possible because most people in Olmsted County, Minn., the county surrounding Rochester, have agreed to make their medical records available on an anonymous basis to researchers.
Beginning in 2004, elderly area residents were asked to participate in the Mayo Clinic Study of Aging, a longitudinal analysis examining the natural history of cognitive decline, which involved a lengthy interview, cognitive assessment, and neurological exam at baseline. To be eligible, people had to be free of overt dementia at enrollment.
A total of 4,298 people age 70 to 89 were approached for the study; 2,050 agreed to participate and 1,679 did not.
Of those willing to be involved, 81 were found to have dementia at baseline or did not provide baseline data and were excluded. Among those declining participation, 272 also refused access to their medical records and another 78 had records indicating prevalent dementia.
That left 1,969 participants (including 237 who dropped out but with accessible records) and 1,329 nonparticipants with data available for the current comparison.
Their ages at the time they were approached for study participation were similar, as were the ages at diagnosis for those developing incident dementia.
But the two groups differed significantly in educational attainment (mean 13 years of schooling for participants versus 12 for nonparticipants) and in their overall illness burden (mean Charlson comorbidity score 3 for participants versus 4 for nonparticipants; P<0.005 for both comparisons). Knopman said those factors both would predict faster cognitive decline in an elderly population. And, in fact, a log-rank comparison of the two groups produced a hazard ratio for incident dementia of 1.27 for nonparticipants relative to participants (95% CI 1.03 to 1.56, P=0.02). Knopman noted one significant limitation to the analysis - it was "sensitive to assumptions about time of onset." He said he and other researchers on the project believed that most of nonparticipants developing dementia during the follow-up period probably would have been diagnosed with either dementia or mild cognitive impairment immediately "had we seen them in person at baseline."
Knopman reported payments for service on a data safety and monitoring board for Eli Lilly and has received research funding from Elan, Forest, and Baxter.
Primary source: American Academy of Neurology
Source reference: Knopman D, et al “Risk of dementia in persons who refused to participate in a longitudinal study of cognitive aging” AAN 2012; Abstract P01.083.