Physical function in critical care (PACIFIC): A multi-centre observational study

Introduction: Impairment in physical function is a significant problem for survivors of critical illness [1,2]. There is growing urgency to develop a core set of outcome measures which can be adopted in clinical and research practice to evaluate efficacy in response to interventions such as rehabilitation. There is currently not a single outcome measure which can be used across the continuum from ICU admission to hospital discharge for individuals with critical illness [3]. Objectives: (1) To determine the clinical utility of two physical function measures: De Morton Mobility Index (DEMMI) and Physical Function in Intensive Care test-scored (PFIT-s) when used in isolation across the hospital admission; and (2) To transform the (15-item) DEMMI and (4-item) PFIT-s into a single measure to evaluate function in ICU survivors using rasch analytical principles. Methods: Multi-centre prospective observational study conducted across four sites internationally. Consecutive eligible participants were recruited who met inclusion criteria; ; Adults > 18 years of age whom were mechanically ventilated > 48 hours and were ambulant at least 10 metres independently prior to their ICU admission. Physical function was evaluated at ICU awakening, and both ICU and hospital discharge using the PFIT-s and DEMMI, administered in a randomised sequence using concealed allocation on each measurement occasion to minimise bias in testing order. Results: 128 participants have been recruited into the study to date across the four sites. 61 % were male (n = 78) with median age of 65 [53–73]; and moderate severity of illness (median [IQR] APACHE II: 22 [17–27]). Median [IQR] ICU and hospital LOS were 9 [5–14] and 21 [13– 37] days respectively. The incidence of ICU-acquired weakness was 50 % (n = 67). Aim 1: On awakening mean ± SD PFIT-s was 4.9 ± 2.5 (out of 10) and DEMMI was 19 ± 21 (out of 100). In isolation the PFIT-s had a floor effect of 9 % (n = 11) at ICU awakening, and 1 % (n = 1) at both ICU and hospital discharge; and a large ceiling effect at hospital discharge of 42 % (n = 40). The DEMMI in isolation had a large floor effect in the ICU of 23 % at awakening, and a small ceiling effect at hospital discharge of 14 % (n = 14). Both the PFIT-s and DEMMI were demonstrated to be highly responsive to change in functional recovery over the acute hospitalisation period (p < 0.005). Aim 2: Preliminary exploration of a subgroup with complete data at hospital discharge (n = 73) was evaluated.The data fit the Rasch model Chi squared =10.4, df = 24, p = 0.99 with no item misfit or differential item functioning based on age, gender, BMI, severity of illness (APACHE II) or comorbidity. A new single measure (12- items) has been proposed combining the DEMMI and PFIT-s. Conclusions: The PFIT-s and DEMMI have limitations when used in isolation. A new transformed scale based on rasch analytical principles is promising combining features of both tools for evaluation of functional recovery of critically ill.