Chronic obstructive pulmonary disease (COPD) is characterized by airflow obstruction that is progressive over many years. Advanced COPD is associated with oxygen desaturation leading to a series of complications including decreased survival. Standard methods to evaluate the oxygen status of these patients underestimate their oxygen needs when measured by ambulatory techniques while patients are in the home environment. Our prior work has extended ambulatory monitoring by simultaneously measuring physical activity with accelerometers to link oxygen desaturation to specific categories of daily activity. This pilot project studied whether the Continuous-Scale Physical Function Performance (CS-PFP) test, an assessment utilizing a range of domestic tasks, could more accurately predict oxygen desaturation in the home environment than the standard assessment. Standard assessment consisted of a standardized hallway walk test (HWT), six-minute walk test (6MWT) and saturation measurements at rest. The CS-PFP consisted of 10 common household functional tasks including: stair climbing, transferring laundry, carrying groceries, carrying a pot of weight, sweeping, donning and doffing a jacket, reaching to a high shelf, picking scarves up from the floor, getting down on and up from the floor, and 6MWT.
- Estimate the sensitivity, specificity, positive predictive value, negative predictive value, total prediction error and Akaike information criterion (AIC) of each of the standard tests and the 10 tasks of the CS-PFP to predict oxygen saturation during four categories of daily living activities defined by accelerometers: Walking, Slow/Intermittent Walking (SIW), Active-not-Walking (ANW) and Rest.
- Determine the optimal combination of tests to achieve the best prediction of oxygen desaturation in each daily activity category.
Veterans at VA Medical Center Brooklyn with severe COPD were recruited to undergo standard assessment for oxygen desaturation, CS-PFP testing, and a 2-day ambulatory study wherein continuous measurements of oxygen saturation and daily activity were monitored. Fifty-four subjects completed testing. Oxygen desaturation during testing was defined as <90% saturation. Oxygen desaturation during the ambulatory study was defined as the percent time that saturation was <90% during the accelerometer-defined activity categories. Logistic regression analysis using stepwise selection technique with the total prediction error and AIC was used to determine the best combination of tasks to predict oxygen desaturation during daily living activities.
Desaturation during a standard HWT best predicted desaturation during walking in daily life. For SIW, the two tests that best predicted desaturation were the HWT and the laundry task; combining these tasks improved the AIC so that the model was >100 times more likely to be correct, compared to the standard testing. For ANW, the laundry and reach tasks best predicted desaturation; the combined model was >2500 times more likely to be correct than the standard tests alone. The best predictors of desaturation during Rest were the pot carry and the scarves tasks. The probability of these tasks to correctly classify desaturators during Rest was >15 times more likely than the standard tests, but did not achieve a statistically significantly difference from the standard test of oxygen saturation <90% at rest.
We have assessed the value of the CS-PFP to predict oxygen desaturation during categories of daily living defined by activity monitors. We have identified the best combination of tasks and standard tests that accurately predicted desaturation during daily living. We propose that this new test can accurately define the individual profile of diurnal oxygen desaturation, categorize COPD disease severity, determine outcomes of mild, moderate and severe hypoxemia and provide a new type of assessment on which to base an oxygen prescription.
External Links for this Project
None at this time.