Background: Multimorbidity, expressed as the number of chronic conditions (CC), tends to increase over time at various speed among individuals. The impact of speed of multimorbidity accumulation on health outcomes in older adults remains unclear.
Objectives: We identified trajectories of multimorbidity in older people and investigated their value in predicting health services use, polypharmacy and death over the following year.
Methods: We performed a population-based study using a 5% random sample of the 1.5 million community-dwelling adults ages >65 years in Quebec, Canada, on April 1, 2019 (index-date). The number of CC was assessed yearly from April 2009 to April 2019 using a list of 31 conditions from the Charlson/Elixhauser combined indices. We used latent class growth modeling to identify multimorbidity trajectory patterns. We chose the final model based on fit statistic, diagnosis criteria, parsimony and interpretability. We compared the risk of 1-year hospitalizations, emergency visits, polypharmacy (≥10medications/year) and death between trajectories using Poisson model with robust variance estimator while adjusting for age, sex, and material or social deprivation.
Results: We identified 8 trajectories: 3 were “stable” with few CC (0, 1, 2.5); 1 was “decreasing” with declining number of CC (0 at index-date); 2 were “progressive” with gradual increment in multimorbidity (6 and 10 CC at index-date); and 2 were “high/recent increase” with abrupt increase in multimorbidity in the last few years (2 and 6 CC at index-date). The risk of each outcome increased globally with the number of CC at index-date (e.g., compared to “stable” 0 CC, risk of death was 11 times higher [RR=10.63; 95%CI: 8.62-13.11] in the “progressive” 10 CC). The trajectory pattern seemed to be most associated with the polypharmacy outcome. In comparison with the “stable” 0 CC, the risk of polypharmacy was higher among the “stable” or “progressive” trajectories than the “high/recent increase” although the number of CC at index-date was similar (e.g. “progressive” 6 CC [RR=9.90; 9.30-10.54] vs “high/recent increase” 6CC [RR=7.91; 7.41-8.45]). On the other hand, the risk of death was similar for these same trajectories (“progressive” 6 CC: RR=6.31; 5.16-7.72; “high/recent increase” 6 CC: RR=6.30; 5.11-7.77). Similar RRs for these same trajectories where also observed for hospitalisations and emergency visits.
Conclusions: The speed of multimorbidity accumulation was strongly associated with the polypharmacy risk but little with deaths and health services use, for which the number of CC at index-date seemed to be a better predictor. Deprescribing efforts should be prioritized in people with long-term multimorbidity, to reduce the economic burden of medication use.
.jpg "Marc Simard (he/him/his) photo")
