Refining Low Physical Activity Measurement Improves Frailty Assessment in Advanced Lung Disease and Survivors of Critical Illness

Abstract

Rationale: The frail phenotype has gained popularity as a clinically relevant measure in adults with advanced lung disease and in critical illness survivors. Because respiratory disease and chronic illness can greatly limit physical activity, the measurement of participation in traditional leisure time activities as a frailty component may lead to substantial misclassification of frailty in pulmonary and critical care patients.

Objectives: To test and validate substituting the Duke Activity Status Index (DASI), a simple 12-item questionnaire, for the Minnesota Leisure Time Physical Activity (MLTA) questionnaire, a detailed questionnaire covering 18 leisure time activities, as the measure of low activity in the Fried frailty phenotype (FFP) instrument.

Methods: In separate multicenter prospective cohort studies of adults with advanced lung disease who were candidates for lung transplant and older survivors of acute respiratory failure, we assessed the FFP using either the MLTA or the DASI. For both the DASI and MLTA, we evaluated content validity by testing floor effects and construct validity through comparisons with conceptually related factors. We tested the predictive validity of substituting the DASI for the MLTA in the FFP assessment using Cox models to estimate associations between the FFP and delisting/death before transplant in those with advanced lung disease and 6-month mortality in older intensive care unit (ICU) survivors.

Results: Among 618 adults with advanced lung disease and 130 older ICU survivors, the MLTA had a substantially greater floor effect than the DASI (42% vs. 1%, and 49% vs. 12%, respectively). The DASI correlated more strongly with strength and function measures than did the MLTA in both cohorts. In models adjusting for age, sex, comorbidities, and illness severity, substitution of the DASI for the MLTA led to stronger associations of the FFP with delisting/death in lung transplant candidates (FFP-MLTA hazard ratio [HR], 1.42; 95% confidence interval [CI], 0.55-3.65; FFP-DASI HR, 2.99; 95% CI, 1.03-8.65) and with mortality in older ICU survivors (FFP-MLTA HR, 2.68; 95% CI, 0.62-11.6; FFP-DASI HR, 5.71; 95% CI, 1.34-24.3).

Conclusions: The DASI improves the construct and predictive validity of frailty assessment in adults with advanced lung disease or recent critical illness. This simple questionnaire should replace the more complex MLTA in assessing the frailty phenotype in these populations.

Keywords: critical illness; disability; frailty; lung transplantation; motor activity.

Figure 1.

Assessment of content validity. Histograms of Duke Activity Status Index (DASI) scores and Minnesota Leisure Time Physical Activity (MLTA) questionnaire kilocalories expended per week for lung transplant candidates with both DASI and MLTA measured (n = 360) (A and B) and for older intensive care unit (ICU) survivors with both DASI and MLTA measured (n = 124) (C and D), respectively.

Figure 2.

Unadjusted fractional polynomial prediction plots of the associations between Duke Activity Status Index scores and concurrent disability scores in male and female lung transplant candidates (A and B) and male and female older intensive care unit (ICU) survivors (C and D), respectively. The red line represents the sex-specific lowest quintile of activity observed in the Lung Transplant Body Composition cohort. ADL = activities of daily living; CI = confidence interval; VLA = valued life activities.

Figure 3.

Kaplan-Meier survival curves for frail (score, ≥3) and nonfrail (score, 0–2) lung transplant candidates (A and B) and older intensive care unit (ICU) survivors (C and D) using the Duke Activity Status Index scores and Minnesota Leisure Time Physical Activity questionnaire kilocalories per week expended for the low-activity criterion, respectively. P values are for the log-rank test.