Validation of the SOPLAY direct observation tool with an objective accelerometry-based physical activity monitor
The System for Observing Play and Leisure Activities (SOPLAY) is a direct observation instrument designed to facilitate observation of groups and environmental contexts. To date, no field-based studies have been done to test validity of SOPLAY using objective criteria. Purpose: The purpose of this study was to test the validity of the SOPLAY with corresponding data from an accelerometry-based activity monitor (Biotrainer pro) in a sample of 9-12 years old children. Methods: Data was collected from 19 different physical activity sessions with a total of 160 children (mean age 10.5y0.8). The SOPLAY direct observation tool and activity monitors were used to evaluate the activity levels of the groups on two different occasions. The agreement between instruments and the influence of SOPLAY sampling rate on its validity were tested. The primary outcome measure for all comparisons was the percent of youth categorized as being "active" but different measures were used to reflect this variable. One measure (MVPA1) used the sum of "walking" + "very active" to reflect activity while another measure (MVPA2) was based on the percentage of youth coded as "very active". Results: Difference between observed and recorded activity levels varied depending on what coding was used. There were large and significant differences when the standard scoring system was used for interpreting the SOPLAY (MVPA1: 50.55y26.41%, p-value<0.001). There was an overall better agreement (non-significant, p-value>0.01) when the alternative measure was used (MVPA2: 1.33y22.06%). The combination of "walking" and "very active" was found to have good correspondence when compared with a parallel measure from the accelerometer based on the sum of light and moderate to vigorous activity (MVPA3: -2.02y29.00%). Correlations for the different MVPA classifications followed the same pattern (MVPA1: 0.404, MVPA2: 0.562, MVPA3: 0.575). Evaluation of the impact of scan rate on validity was evaluated by comparing agreement with different number of scans. Although observation scans every 2, 4, 6, 8, 12 and 20 had systematic increases in error, there was a substantial absolute error (21.76%) associated when there was only 1 scan per session (20 minutes scans). Conclusions: Observations provide valid indicators of MVPA if coding is based on the percentage of youth classified as "very active". The results demonstrate that more frequent scans can improve the validity of the estimations.