Table 2.

Description of the outcome measures used in the Guelph Family Health Pilot Study

Outcome	Description of measure
Child percentage of fat mass	Child body composition was measured before and after the intervention using whole-body, single-frequency, 50-kHz bioelectrical impedance analysis (BIA, Quantum IV—Body Composition Analyzer™, RJL Systems, Clinton Township, MI) per manufacturer protocol. Preparation for BIA required the children to have fasted and avoided vigorous physical activity for 30 min prior to the test and removed any clothing with metal pieces (e.g., pants with zipper). Children lay supine on a hospital bed, with head on a pillow, and with arms and legs abducted 30° from midline. Children were instructed to lie still until the test was complete (approx. 1 min). The equation of Kushner et al. was used to calculate TBW in litres as follows, where H is height in centimetres (cm), R is resistance in ohms (Ω), and W is weight in kilograms (kg) (Fomon et al. 1982): TBW = 0.593H2/R + 0.065W + 0.04 TBW was then divided by an age- and gender-specific hydration constant to determine the percentage of fat mass (Kushner et al. 1992; Frisancho 2004).
Child dietary intake: Fruit, vegetable, and sugar-sweetened beverage	The primary parent (the first parent to sign up for the study) completed 3-day food records (Burrows et al. 2010; Gibson 2005) for their children before and after the intervention. Parents were instructed on how to complete food records by the study coordinator. Parents were instructed to provide as much detail as possible in describing foods and beverages consumed by their children. Parents recorded information on timing, eating occasion, description, and amounts of foods and beverages consumed. Food record data were inputted into ESHA, The Food Processor version 11.0.110. (ESHA Research, Salem, OR, USA, 2016) using standard operating procedures for data input, checking, and export to ensure accuracy of reported data. Food records were analyzed using ESHA Food Processor (version 11.0.110) for 3-day average intakes of food group servings (vegetables and fruit). Research assistants reviewed food records to calculate 3-day average intakes of sugar-sweetened beverages (SSBs), including juice.
Physical activity and sedentary behaviour—accelerometry	For assessment of physical activity (PA) and sedentary behaviour, ActiGraph GT3X accelerometers were used (Actigraph 2015). Parents were instructed to keep the accelerometer on the non-dominant wrist of their children for 24 h a day for 3–7 days, only removing it if the child was in water for a long period of time (e.g., swimming lessons). The ActiLife software was used as a data analysis platform. Within the ActiLife software, the accelerometer data was downloaded in 1-sec epochs. The Choi et al. (2007) algorithm was used to identify and remove any periods of non-wear time. Only subjects with a minimum of 360 min (6 h) of valid wear time for a minimum of 3 days were included in the analysis. Wear time was then divided into the different PA intensities (SED, LPA, and MVPA) using age-dependent cut points. For toddlers (less than 3 years old), the Trost et al. (2012) cut points were applied, and for preschoolers (3 years old or greater), the Butte et al. (2014) cut points were used. Data were presented in the form of percentage of wear time spent in each intensity (SED, LPA, MVPA) to account for variable wear times. Physical activity and sedentary behaviour variables include - Percent of wear time spent in sedentary (SED) - Percent of wear time spent in light PA (LPA) - Percent of wear time spent in moderate to vigorous PA (MVPA)
Sleep quantity	Parents were instructed to keep the accelerometers on the non-dominant wrist of their children for 3–7 days, only removing it if the child was in water for a long period of time (e.g., swimming lessons). To determine sleep periods, the Sadeh and Tudor-Locke algorithms were applied to the raw sleep data in 1-min epochs. Participants were included if they had at least two consecutive nights of sleep data. To measure sleep duration, the software provides a “total sleep time” variable for each night and these values were averaged for each participant to provide average nighttime sleep duration. This value represents the total time the participant spends sleeping (in minutes), excluding the time they spend awake in bed (Buysse et al. 1989).
Family meal frequency	The primary parent reported frequency of family meals with a single item on the baseline and 6-month post-intervention questionnaire that assessed frequency that either parent ate a meal with their children over the past week. Response options were one to two times, three to four times, five to six times, and seven or more times. Due to high level of responses towards the upper end of the response options, we dichotomized this variable as ≥ 7 times per week and ≤ 6 times per week.