Hj 102

Age, sex, hip and knee pain were recorded using a questionnaire at baseline. The presence of knee and hip radiographic osteoarthritis (ROA) at baseline were assessed using Altman atlas as previously described (16) . Joint space narrowing (JSN) and osteophytes in the knee and hip were assessed on a scale of 0-3 (where 0=no disease and 3=most severe disease). Participants were dichotomised as having knee/hip ROA (presence of either JSN or osteophytes) or not (no JSN or osteophytes). Physical activity was measured objectively over seven consecutive days using a pedometer (Omron HJ-003 & HJ-102; Omron Healthcare, Kyoto, Japan) as previously described (17) .

AA was determined as steps/day using pedometers (Omron HJ-003 & HJ-102, Omron Healthcare, Kyoto, Japan), at baseline. The participants were informed on pedometer use, keeping a log/diary of step-count and the time during which those were worn. They were required to wear the pedometers for seven consecutive days as they conducted day-to-day activities except during water activities and sleeping. This was repeated after six months in order to account for habitual changes in different seasons. Therefore, 2 sets of logs were available per participant. Readings were screened and excluded if there was any evidence of artificial pedometer readings. Then, pedometer wear-time was determined for each day utilizing the pedometer logs. A 'valid pedometer wear-day' was considered as a day on which the pedometer was worn for at least 8 hours. Then, the steps/day count was calculated as the mean of the two pedometer logs, with a minimum of five valid wear days [3] .

Information on diagnosed medical conditions was obtained from baseline questionnaire.
Participants were asked to select any of the 10 conditions in the questionnaire: diabetes, heart attack, hypertension, thrombosis, asthma, bronchitis, osteoporosis, hyper-and hypothyroidism, rheumatoid arthritis, or "other illnesses". Heart attack included history of coronary artery disease and myocardial infarction. Weight was measured to the nearest 0.1kg using a single pair of calibrated electronic scales (Seca Delta Model 707) and a stadiometer was used to measure height to the nearest 0.1cm. Both these measures were used to compute Body Mass Index (BMI)
[weight(kg)/height(m) 2 ]. Physical activity was measured as steps per day using pedometer (Omron HJ-003 and HJ-102; Omron Healthcare, Kyoto, Japan) [31] (link). Information about cigarette smoking status and presence of backpain was obtained from baseline questionnaire.

Ambulatory activity was measured via pedometer data and questionnaire at baseline. Pedometer data was collected over 7 consecutive days following baseline assessment in which participants were familiarized with the use of the pedometer (Omron HJ-003 and HJ-102, Omron Healthcare, Kyoto, Japan). Pedometers were worn on the waist-band or belt and were calibrated using a 100-pace walking test.
Participants were provided with a pedometer diary and recorded their steps daily. The start and finish times of pedometer use were recorded on each day, and participants also reported the duration and reason for any periods in which they did not wear the pedometer. Finally, participants reported any circumstances that may have affected a pedometer reading. Total number of steps were summed, and divided by the number of days the pedometer was worn to give the average number of steps per day.
At baseline and follow-up, subjects were asked to nominate how many kilometres they walked recreationally per week on average over the past year. Responses were categorical with 1) less than 1 km; 2) 1e5 km; 3) 5e10 km and 4) Greater than 10 km. Baseline values were subtracted from follow-up values. An increase of two or more categories was considered to be an increase in ambulatory activity.

AA was assessed at baseline as steps per day determined by pedometer (Omron HJ-003 & HJ-102, Omron Healthcare, Kyoto, Japan). All participants were instructed to wear a pedometer for seven consecutive days and were re-assessed 6 months later to account for seasonal variation 18 . Verbal and written instructions were given regarding pedometer wear and how to keep a pedometer diary. In general we had good compliance of pedometer wear. Of the 408 participants in the current study, the average days worn was 13.6 (standard deviation (SD) 1.6). There were strong correlation between pedometer assessment at baseline and 6 month repeated (Supplementary Figure 2). Mean steps per day was calculated as the average of the days worn at both time points. 14 participants only wore the pedometer for one time thus steps/day value at this time point were used. We categorised AA into three groups based on the Tudor-Locke et al. 19 reommondations which we modified slightly based on the average steps per day taken amongst older adults with self-reported functional limitation, reported to be 7681 steps per day 20 . Participants in our study were considered to be less active ( 7499 steps per day); moderately active (7500/9999 steps per day), or highly active (!10,000 steps per day).