A consensus emerging in the literature indicates that mortality and morbidity alone may not be adequate for assessing outcome after surgery [31
]. Outcome involving HRQOL is patient focused and denotes interest in the patient's perspectives in evaluating health status. In the present study, HRQOL was evaluated 6 months after ICU discharge. The time for assessing HRQOL in follow-up patients was chosen to minimize drop-outs and took account of previous studies suggesting that health problems leading to ICU admission after 6 months are due to chronic underlying conditions, or to new and unrelated health problems commonly encountered in an elderly population [34
Among the patients who completed the questionnaire 6 months after discharge from the ICU, 59% reported feeling better than one year before and 20% reported feeling worse. These findings agree with other reports, using different tools, on patients after ICU discharge [36
In the SF-36 domains, younger patients reported greater physical and mental problems. Perception of HRQOL in younger patients may be due to a willingness to accept functional limitations or to differences in expectations among younger patients, as stated in the study by Eddleston et al. [37
]. This could also explain why, in our results, younger men had the worst scores in bodily pain, role-physical, vitality and mental health. Female patients had significantly worse results in bodily pain, general health, vitality and social function. Impairment of HRQOL with age and gender has previously been reported [36
] but others have shown no influence of age on HRQOL [39
The ICU variables showed that only ASA-PS physical status was able to predict the lower median results in all SF-36 domains except bodily pain; ASA-PS III/IV patients had significantly lower results in all domains. Although ASA-PS was never intended to be a peri-operative risk score, large studies have suggested that a higher ASA-PS score is one of the best predictors of post-operative morbidity [41
Consistent with a previous study by Graf et al. [18
], we found that neither age nor type and magnitude of surgery was associated with differences in HRQOL. For severity of disease measured by differences in SAPS II score, statistical significance was found only for the general health domain.
Previous studies have concluded that pre-existing disease has a significant impact on HRQOL [40
] and although pre-existing diseases were not included among the variables considered in the present study, the ASA-PS classification indirectly measured this parameter [43
]. Ridley et al. showed that HRQOL scores 6 months after ICU are similar to the pre-ICU scores for patients with pre-existing diseases, but lower in patients suffering acute pathologies [40
]. This could also explain why logistic regression analysis showed that only SAPS II – when adjusted for age, gender, ASA, type and magnitude of surgery, ICU LOS, marital and employment status – was associated with a decline in self reported general health status, 6 months after ICU discharge compared to 12 months previously (6 months before ICU discharge).
There is no generally accepted definition of the term 'long-term intensive care'. Because of the markedly skewed distribution of LOS-ICU, no obvious cut-off exists and time periods of ≥ 7 days up to > 30 days have been used to define prolonged ICU stay[45
]. For the present study, greater LOS was defined as an ICU stay of more than 7 days. Longer ICU stay had a significant influence only on the role-physical, bodily pain and role-emotional domains of SF-36. Previous studies have suggested that prolonged ICU LOS does not affect HRQOL after ICU [4
In a systematic review of quality of life in adult survivors of critical illness, Dowdy et al. [3
] refer to six studies in which survivors of elective versus emergency surgical procedures were evaluated [39
]. In three of these studies [39
], quality of life of life was worse, and in two [36
], emergency surgical patients had a significantly worse quality of life in a minority of domains. In our study there was no significant association between surgical status (emergency versus elective) and overall HRQOL except in the vitality domain. This could be explained by the fact that 85% of the patients underwent elective surgery.
A limitation of this study was the lack of data about employment status before ICU admission. Our data indicate that most patients were retired 6 months after discharge from the ICU, in contrast to the study of Cuthbertson et al. [54
], in which fewer than 60% of patients returned to their previous work after discharge from the ICU.
Global results concerning ADL tasks showed that 6 months after discharge from the ICU, 60% of the patients were dependent in at least one activity in ADLI and 34% were dependent in at least one activity in ADLP. In the original studies of Katz [22
] and Hulter Asberg [30
] the authors' studied dependency on older patients and results showed these patients to be more dependent than the patients that we have studied. By contrast, the patients described in the study of Niskanen et al. [8
], comprised patients admitted to a multidisciplinary ICU and were less dependent.
As in other studies, age, was not a determinant of dependence in ADLP, but patients may still need assistance; and age appears to be determinant of ADLI [55
The results of the study of ADL may have been influenced by co-morbidities and concurrent diseases; indeed, ASA-PS III/IV patients were more dependent in ADL, in both the ADLI and ADLP tasks. The logistic regression model (adjusted for age, gender, ASA-PS, SAPS II, type and magnitude of surgery, ICU LOS, marital and employment status) showed that SAPS II correlated significantly with disability in ADLP and ADLI tasks, alone for ADLP tasks and combined with age for ADLI tasks. Higher SAPS II scores and higher LOS appeared not to predict disabilities in ADL 6 months after ICU discharge, which could reflect the burden imposed by acute alterations.
Among the hospital survivors, no patients were lost to 6 months follow-up for the assessment of survival, but 23% were lost for the assessment of quality of life because they did not respond the questionnaire. Although the overall characteristics of the non-respondents at discharge from the ICU were similar to those of the participants, it must be emphasized that a poor quality of life or a high incidence of psychological disturbance at the time of the follow-up survey could have contributed to the non-response to the questionnaire and may be seen as a limitation of this study [44
Admissions to ICU are not homogeneous, and generalizing findings to all ICU admissions may be misleading since our sample was representative only of a surgical population. The population studied was composed mainly of patients undergoing scheduled surgery, probably already having a reduced quality of life, surgery being performed as an attempt to improve quality of life and survival.
Eighty-five percent of the patients underwent scheduled surgery, and this should be considered when analyzing our data.
Because HRQOL before admission to the ICU appears to be an important determinant of outcome and HRQOL after discharge [56
], another limitation of this study was the absence of any evaluation of HRQOL prior to admission and of a sample to act as control group.