, 2007); and an item describing the financial state of the farm at the end of 2012 (5 categories, “large deficit” through “large surplus”). The three socio-economic variables were combined into an internally consistent summary index (Cronbach’s alpha = 0.82) and placed into “low”, “medium” and “high” tertiles.

Exposures to farm work. Average reported hours of farm work per week were estimated by season and then averaged over the full year (“none”, “part-time” (< 30 hrs/week), “full-time” (≥ 30 hrs/week)). We asked respondents to estimate exposure to mechanized farm work tasks for 2012 in hours/year (“operation of tractors”, “maintenance of tractors”, “operation of combines”, “maintenance of combines”) and days/year (“operation of NVP-BGJ398 all-terrain

vehicles”, “operation of power tools with hands more than one hour over the day”). These items were developed for our study and were subject to multiple pilot tests for face validity (Pickett et al., 2008; Day et al., 2008). Reported hours/ year were converted to days per year at an assumed average rate of 8 hours/day. For analytical purposes, each of these variables was classified into four groups (none, plus tertiles of the remainder). Items describing exposure to non-mechanized work included: “lift, lower, or carry heavy objects (over 20 lbs) more than 1 hour over the day ”; “using a shovel or pitchfork more than 1 hour over the day”; “work with hands over shoulder height more than 1 hour over the day”; “routine chores with

large I-BET-762 in vitro animals (e.g., cattle or pigs)”, ”routine chores with small animals”, “herd maintenance activities (e.g., branding, vaccinating, transporting)”, and aminophylline “veterinary activities (e.g., medications administration, breeding, birthing)”. These items were developed for this cohort and were subject to pilot tests for face validity (Pickett et al., 2008). Each was classified into four groups (none, plus tertiles of the remainder). We then created two additive scores, one for mechanized and one for non-mechanized farm work, to illustrate the cumulative effects of exposure. Indicator variables (1-“yes” vs. 0-“no”) were created according to whether participants were in the highest category of each of the specific work tasks. The summed additive scores varied between 0 (lowest activity) and 5 or more (highest activity) for mechanized and non-mechanized work. The energy expenditure rates of different work tasks were expressed using metabolic equivalent (MET) scoring. MET scores refer to the ratio of the energy expenditure rate for an activity compared to resting energy expenditure. Thus, a MET of 3.0 infers that the energy expended while doing that activity is three times that of rest. MET scores were abstracted from the Compendium of Physical Activities (Ainsworth et al., 2000).