Arthritis & Rheumatism, Volume 62,
November 2010 Abstract Supplement

Abstracts of the American College of
Rheumatology/Association of Rheumatology Health Professionals
Annual Scientific Meeting
Atlanta, Georgia November 6-11, 2010.

Projecting the Effects of a 5-Point Reduction in Body Mass Index on Knee or Hip Osteoarthritis and Health-Related Quality of Life in a Closed Population to 2031: A Microsimulation Study.

Sayre1,  Eric C., Kopec4,  Jacek A., Sharif4,  Behnam, Flanagan3,  William M., Fines3,  Philippe, Rahman5,  Mushfiqur, Abrahamowicz2,  Michal

Arthritis Research Centre of Canada, Vancouver, BC, Canada
McGill University, Canada
Statistics Canada, Canada
University of British Columbia, Canada
University of British Columbia, Canada
University of Ottawa, Canada
University of South Australia, Canada


Body mass index (BMI) is a risk factor for knee or hip osteoarthritis (OA). Obesity (BMI>=30) increased from 10 to 15% in Canada between 1990 and 2000. BMI reduction is expected to reduce OA and improve health-related quality of life (HRQoL). The purpose of this study is to project in a national cohort the effect of a risk-targeted 5-point reduction in BMI on OA and HRQoL between 2001 and 2031.


We used the Population Health Model (POHEM) platform to develop a stochastic continuous-time microsimulation model of physician-diagnosed OA. Incidence rates were calibrated to agree with administrative data for British Columbia, Canada. Projected upward trends in BMI under the base case (no intervention), the effect of obesity on OA incidence and the impact of OA on HRQoL were modeled using Canadian national surveys and hospital data. The effect of BMI on OA was fit categorically: underweight (<18.5), normal (<25.0), overweight (<30.0) and obese (>=30.0). A closed population weighted for 2001 Canada that reflected projected mortality was followed in POHEM to 2031, aged 20+. BMI was reduced by 0 or 5 points in 2001 in persons with BMI>=25. Outcomes included OA prevalence and incidence (per 1000 person years), average Health Utilities Index Mark 3 (HUI), and health-adjusted life years (HALE). Monte Carlo error was estimated via subsamples. HRQoL models included BMI only indirectly through rates of OA.


There were 22,483,400 in-scope subjects in the simulated population in 2001, 14,317,129 in 2031. Average age in 2001 is 45.89 (95% Monte Carlo CI=45.79, 45.99), in 2031 is 67.43 (67.37, 67.50). In 2001, average BMI is 25.55 (25.53, 25.58), OA incidence 7.42 (7.35, 7.49), OA prevalence 12.23% (12.00, 12.46), average HUI 0.874 (0.872, 0.875), proportion obese 14.99% (14.77, 15.20) and proportion overweight 33.66% (33.38, 33.94).

Under the base case, in 2031, average BMI is 27.79 (27.75, 27.82), OA incidence 21.03 (20.89, 21.16), OA prevalence 33.11% (32.98, 33.25), average HUI 0.752 (0.750, 0.753), proportion obese 26.82% (26.6, 27.04) and proportion overweight 37.15% (36.97, 37.33), with the rest normal or underweight. HALE for 20 year-olds is 27.06 (26.92, 27.21), for 40 year-olds 24.71 (24.61, 24.80) and for 60 year-olds 16.94 (16.81, 17.07).

With a 5-point targeted reduction in BMI, in 2031, average BMI is 25.56 (25.53, 25.58), OA incidence 19.29 (19.16, 19.42), OA prevalence 30.69% (30.55, 30.82), average HUI 0.756 (0.754, 0.758), proportion obese 11.57% (11.4, 11.74) and proportion overweight 38.19% (38.01, 38.36). HALE for 20 year-olds is 27.14 (26.99, 27.28), for 40 year-olds 24.81 (24.72, 24.91) and for 60 year-olds 17.00 (16.87, 17.12).


Reducing at-risk BMI by 5 points in 2001 would lead to a moderate reduction in OA (1.74 incidence and 2.42% prevalence) by 2031, but only very small improvements in HRQoL attributable to rates of OA. (Lowering BMI will have additional direct benefits on HRQoL.) Monte Carlo error was small, but future studies should incorporate additional error such as around parameter estimates. Future studies should vary BMI reduction and the target group, and simulate open populations including births and immigration.

To cite this abstract, please use the following information:
Sayre, Eric C., Kopec, Jacek A., Sharif, Behnam, Flanagan, William M., Fines, Philippe, Rahman, Mushfiqur, et al; Projecting the Effects of a 5-Point Reduction in Body Mass Index on Knee or Hip Osteoarthritis and Health-Related Quality of Life in a Closed Population to 2031: A Microsimulation Study. [abstract]. Arthritis Rheum 2010;62 Suppl 10 :1555
DOI: 10.1002/art.29321

Abstract Supplement

Meeting Menu