Tania Winzenberg1, Kelly A Shaw. 1. Menzies Research Institute, University of Tasmania, Australia. tania.winzenberg@utas.edu.au
Abstract
BACKGROUND: It is unclear what is the best method of accurately identifying physically inactive patients in general practice. This study aimed to compare the performance of different methods of assessing patient physical activity levels in general practice. OBJECTIVE:Thirteen general practitioners were randomly allocated to perform either their usual assessment, or this with a Lifescripts tool, on consecutive patients. The authors measured patients' physical activity by accelerometer over 1 week, including steps per day, then calculated agreement, kappa specificity, sensitivity, positive and negative predictive value (PV) and ROC characteristics for each assessment method (GPs' usual assessment, Lifescripts tool and steps per day) against the reference standard of accelerometer classification. RESULTS: Data from 29 patients was included. Agreement between subjective assessments was highest for GPs' usual assessment (agreement 73%; kappa 0.47; p=0.03), which also gave the highest area under the ROC curve (0.75, 95% CI: 0.52-0.98). However, this still had low specificity (67%) and positive PV (63%). Using a cut-off of 7500 steps/day maximised the area under the ROC curve at 0.91 (95% CI: 0.82-1.00), 19.2% greater than GPs' usual assessment. CONCLUSION: Measuring steps per day may be a feasible and more effective way to screen for physically inactive patients than self report. A large scale study to confirm these results is necessary.
RCT Entities:
BACKGROUND: It is unclear what is the best method of accurately identifying physically inactive patients in general practice. This study aimed to compare the performance of different methods of assessing patient physical activity levels in general practice. OBJECTIVE: Thirteen general practitioners were randomly allocated to perform either their usual assessment, or this with a Lifescripts tool, on consecutive patients. The authors measured patients' physical activity by accelerometer over 1 week, including steps per day, then calculated agreement, kappa specificity, sensitivity, positive and negative predictive value (PV) and ROC characteristics for each assessment method (GPs' usual assessment, Lifescripts tool and steps per day) against the reference standard of accelerometer classification. RESULTS: Data from 29 patients was included. Agreement between subjective assessments was highest for GPs' usual assessment (agreement 73%; kappa 0.47; p=0.03), which also gave the highest area under the ROC curve (0.75, 95% CI: 0.52-0.98). However, this still had low specificity (67%) and positive PV (63%). Using a cut-off of 7500 steps/day maximised the area under the ROC curve at 0.91 (95% CI: 0.82-1.00), 19.2% greater than GPs' usual assessment. CONCLUSION: Measuring steps per day may be a feasible and more effective way to screen for physically inactive patients than self report. A large scale study to confirm these results is necessary.