The study was conducted in a mixed medical-surgical 6-bed ICU of a university hospital of 710 beds located in the Northeast of Italy. This ICU serves all thoracic, vascular, and high-risk abdominal surgery patients and about half of the medical ward patients of the hospital.
All consecutive adult ICU patients discharged between March 2013 and October 2014 were considered for enrolment. We included patients aged more than 18 years, staying in ICU for at least 72 h, without pre-existing cognitive dysfunction or language barrier. One year after hospital discharge, we assessed the vital status of the patients through the Vital Statistics Offices. We excluded patients residing farther than 30 km from the hospital, and we contacted the ICU survivors by phone to ask them to participate in the study. Those who agreed received an appointment at the hospital for a date and time convenient for them.
The demographics and clinical data of the study patients were retrieved from the ICU database. The software for this (PROSAFE) is made available by the Italian GiViTI group, and has been used in the study ICU since 2008 for daily collection and storage of patient data. For each participant, the following information was retrieved: gender, age, Body Mass Index (BMI), pre-existing comorbidities (like hypertension, COPD, cardiac failure of NYHA class III or IV, and diabetes), reason for ICU admission, type of admission (elective surgery, emergency surgery, or medical admission), date of hospital and ICU admission, presence of severe sepsis or septic shock (at ICU admission or during ICU stay), main treatments received (administration of vasoactive amines, mechanical ventilation, renal replacement techniques), date of ICU and hospital discharge and vital status at hospital discharge. The following computed data were also retrieved: Simplified Acute Physiology Score (SAPS) II [18] and Sequential Organ Failure Assessment (SOFA) Score [19] which referred to the first 24 h in ICU, and Length Of Stay (LOS) referring to the number of days in ICU and in hospital after ICU discharge.
Instruments
At the time of the appointment at the hospital, a researcher not involved in the care of the patient during his/her stay in ICU welcomed the patient, obtained signed consent, and administered the following questionnaires in a calm and confidential environment. The time scheduled for each appointment was 30 min and it was generally observed.
The Functional Assessment of Chronic Illness Therapy for Fatigue (FACIT-F) scale is a sub-scale of a general questionnaire developed to assess anaemia-related symptoms in cancer patients [10]. FACIT-F consists of 13 items referring to the previous seven days. Each item allows five response options from ‘Not at all’ (scored 4) to ‘Very much’ (scored 0) with two items needing a reverse score. The scores are summed, multiplied by 13, then divided by the number of items actually answered, thereby allowing calculation for missing items. The final (raw) score ranges from 0 to 52, with higher scores representing less fatigue, and lower scores more fatigue. The raw scores can be transformed into interval measures (FACIT-F-trans) ranging from 0 to 100, according to an interval metric proposed by Cella et al. [11].
The Short-Form 36 (SF-36 version 1) [20] was administered in the previously validated Italian version [21]. The questionnaire refers to the four previous weeks, and consists of eight multi-item dimensions, i.e., physical functioning (PF), role limitation due to physical problems (RP), bodily pain (BP), general health (GH), vitality (VT), social functioning (SF), role limitation due to emotional problems (RE), and mental health (MH). Scores with a range from 0 to 100 are obtained for each dimension. Higher scores represent better functioning. Differences in SF-36 scores of more than 5 points were reported as clinically meaningful [22].
The Medical Research Council Dyspnoea (MRCD) Scale is suitable to assess the extent to which breathlessness affects patient mobility during daily activities [23]. It consists of five statements about perceived breathlessness, from grade 1 (breathless with strenuous exercise) to grade 5 (too breathless to leave the house). We administered the MRCD Scale to check for a correlation between this scale and both FACIT-F and VT of SF-36. We predicted that FACIT-F scores would be significantly worse (lower) in patients with mobility affected by breathlessness than in those without it. This analysis was performed to demonstrate the construct validity of both FACIT-F and VT.
The Ethics Committee of the Province of Ferrara approved the study protocol (n. 140696), and required the written consent of each participant. We received permission for the appointed institution to use the FACIT-F and SF-36 questionnaires.
Statistics
Categorical data are presented as count (%). The Shapiro-Wilk test was used to assess whether continuous variables were normally distributed, and data are shown as mean (±SD), or median with 1st and 3rd quartiles [Q1-Q3], accordingly. A Chi squared test, or a Fisher’s exact test when appropriate, was used for categorical comparisons, and a Mann-Whitney U test for comparisons of non-normally distributed variables. We investigated the internal consistency of the FACIT-F questionnaire using Cronbach’s α coefficient. Spearman’s correlation coefficient was used to analyse correlations between FACIT-F and VT of SF-36, and between both FACIT-F and VT and the MRCD Scale. The Bland Altman plot was used to assess the agreement between FACIT-F-Trans and VT of SF-36 (both ranging 0-100). Linear regression between the differences and the means, which should be close to zero, was used to evaluate the presence of a proportional bias.
For each SF-36 dimension, we computed the normal values for a population matched to our study patients for gender and age [24], and we compared mean values of our study patients with the Italian adjusted normative population using an unpaired t test.
A Type I error in two-tailed tests was considered significant (α 0.05). Statistical analysis was performed using the software packages SPSS v. 11.5 (IBM, New York, USA) and STATA 12.1 (StataCorp, Texas, USA).
Due to the lack of studies specifically assessing fatigue in long-term ICU survivors, we relied on the raw values of FACIT-F (43.6 ± 9.4) for the general USA population [11]. To detect a 10 % difference (4.36) between the mean score of our patients and that of the general USA population (α = 0.05, power = 0.80), the necessary sample size was calculated as 37 patients. Considering that long-term mortality and dropouts could markedly decrease the number of the patients who can be assessed at one year, we decided that the time hypothetically requested for reaching the sample size was about 16 months (March 2013 - October 2014).