The Academic Medical Center Linear Disability Score for evaluation of physical reserve on admission to the ICU: can we query the relatives?

To assess the effects of critical illness and treatment in the ICU on changes in disability status, measurements evaluating health-related quality of life (HRQOL) should be performed at ICU admission. As most patients are not able to complete questionnaires at the time of admission, proxies must frequently be used as a surrogate. Assessment of the functional status on admission provides valuable information that could support the physician in decision-making regarding ICU admission and withholding life-sustaining treatment in a critically ill patient. However, can proxies provide useful and reliable information on functional status in critically ill patients? Some studies in critically ill patients reported moderate agreement between individual patients and their proxies, although lower levels of agreement may be reported for psychological functioning [1, 2]. In a previous study performed by our group, we found that proxies adequately assess the patients' quality of life on admission to the ICU when using the Short Form 36 (SF-36) [3]. The SF-36 is based on calculating sum-scores, however, and there are some disadvantages with the use of sum-scores such as the unclear clinical significance of the sum-scores and the difficulty to compare the results of repeated health measures [4].

The assessment of the functional health status on ICU admission is important to judge the biological reserve of a patient in relation to final outcome [5, 6]. Current measures such as the physical component score (PCS) of the widely used SF-36, however, have several limitations; for instance, the SF-36 scores are used to calculate a summarised score with an inherent loss of information. Using a questionnaire based on item response theory may overcome these problems. The Academic Medical Center Linear Disability Score (ALDS) is a recently developed generic item bank that measures the disability status of patients with a broad range of diseases, as expressed by the ability to perform activities in daily living [7]. Using the ALDS, it is possible to arrange both the item difficulty and the patient's ability on a single hierarchical linear scale. This method makes it possible to present different sets of items to different groups of patients. Since all of the items are calibrated, the measurements remain comparable. Another advantage of the ALDS is its simplicity to use for both patients and their relatives directly after ICU admission. The ALDS has been validated in a large, mixed patient population [8], and in patients with rheumatoid arthritis [9] and with Parkinson's disease [10]. The ALDS has not, however, been validated for critically ill patients or for use by proxies to estimate ICU preadmission functional status.

The aims of our study were to evaluate whether proxy scoring with the ALDS and the Short Form 12 (SF-12) can be used to assess the patients' functional status on admission to the coronary care unit (CCU) or ICU and, secondly, to evaluate how the physical disability score of the ALDS relates to the PCS using the SF-12.

Between January 2010 and October 2010, 370 admissions were screened on the ICU. In total, 300 patients were excluded for a variety of reasons (Figure 1). Between January 2010 and June 2010, 215 admissions were screened on the CCU. A total of 145 patients were excluded (Figure 1). We included a total of 140 patients in the study (88 men, 52 women) for acute ICU admissions (n = 56), scheduled ICU admissions (n = 14) and acute CCU admissions (n = 70). Those patients and proxies completed the ALDS and the SF-12. Demographic and clinical characteristics of all study patients are shown in Table 1.

The present study shows that relatives in close contact with the patient can adequately reflect the patient's functional status on admission to the ICU by use of the ALDS item bank and the SF-12. When comparing the physical disability score of the ALDS with the PCS of the SF-12, the highest intra-class correlation between patients and their proxies was found using the ALDS item bank. The ALDS could therefore be a useful alternative for the PCS of the SF-12.

To our knowledge the ALDS item bank has not yet been used in critically ill patients. In earlier studies by our group we used the SF-36 to measure HRQOL [3, 21‐23]. A major advantage of the ALDS item bank for measuring the functional status compared with a sum-score-based questionnaire like the SF-36 is the possibility to arrange both the item difficulty and the patient's ability on a single hierarchical linear scale [4]. Items can be selected for individual patients using small sets of items tailored to the disability level of the patients [14]. Another advantage of the ALDS item bank is its simplicity to use for both patients and their relatives directly after ICU or CCU admission.

The results of the present study are in line with an earlier study by our group, where we used proxies to assess the patient's HRQOL, and a significant correlation was found between the patient's HRQOL and the assessment made by the proxies [3]. In this study the proxies adequately reflect the patient's quality of life (especially the physical functioning) on admission to the ICU when the SF-36 questionnaire was used [3]. Rogers and colleagues [2] and Crispin and colleagues [24] showed that proxies reliably assessed the patient's HRQOL at time of discharge from the ICU. However, in specific dimensions - especially in the area of mental well-being - agreement between patients and proxies was moderate. Other studies have reported similar results [1, 2, 25, 26].

Relatives may be more appropriate in their assessment of the patient's physical health rather than their mental health estimation [1‐3, 27, 28]. These results are also in line with our current study where the correlation between proxies and patients was lower in the MCS of the SF-12 compared with the PCS in all patients and the highest in the functional disability score (ALDS). Additionally, proxies may underreport good quality of life and overestimate poor quality of life [29].

Scales and colleagues reported poor agreement between proxies and patients in a cohort following ARDS survivors [30]; however, patients were asked to estimate their pre-admission HRQOL 3 months following ICU discharge while proxies rated the HRQOL around the time of ICU admission. This delay prior to patients completing the pre-admission HRQOL will probably influence their pre-admission responses to some extent [30]. Furthermore, different from our study population, these proxy estimations were studied in a population with high disease severity [30]. Interestingly, Capuzzo and colleagues found in their study that patients with planned ICU admission had a good memory of their health status before ICU admission after 3 months, suggesting that patients could be asked later to recall their pre-admission health status. Most ICU patients are acute admissions, however, and consequently the results of this study do not demonstrate that the findings reported are of value for patients with acute admissions [31].

Additionally, the gender of the proxy and the amount of contact between the proxy and the patient could influence their capacity to answer questions about the patient's HRQOL before ICU admission [32]. Holman and colleagues showed that some of the ALDS items may have different measurement characteristics for males and females and for younger and older respondents [32]. In our study, however, we found no significant differences between male and female or age regarding the ALDS. Furthermore, when comparing the relationship of the proxies regarding the patients we found no significant differences when the ALDS item bank was used. Existing research demonstrates also that the nature of the patient-proxy relationship does not affect patient-proxy agreement for HRQOL before ICU admission [1, 2, 28, 30]. However, the group of patients in our study was possibly too small to perform this analysis satisfactorily. Proxies can either report from the patient's perspective or they can report from their own perspective. Mc Phail and colleagues found in their study that it was important to give clear instructions regarding the perspective from which the proxies should provide their response [33]. In keeping with this recommendation, proxies in the present study were provided with standard instructions in which they were asked to view the patient's perspective. This prospective was selected because most of our proxies knew the patient very well and they were reporting on behalf of patients with sound cognition.

Finally, we speculated that the acute illness and emergency admission could bias the proxies in the adequacy to assess the patient's functional status. In an earlier study of our group using the SF-36 we found no differences between elective and emergency ICU admissions [3]. This is in line with our current data using the ALDS, which showed no differences between elective and emergency ICU admission.

Several limitations of our study should be mentioned. Firstly, some investigators have raised concerns about proxy estimations of HRQOL in populations with high disease severity [30]. The same study suggested that predictions of poor ICU outcome may be exaggerated if proxies underestimate HRQOL [30]. In contrast to the situation in our previous validation study as well as in the current study, however, where patients and their proxies were interviewed within 72 hours of ICU admission, those investigators interviewed patients 3 months after ICU discharge and interviewed their proxies at study entry. This difference makes it entirely possible that survivors of critical illness may overestimate pre-admission HRQOL. Second, the presence of delirium or cognitive dysfunction at the time of assessment could have influenced the response, although we made an effort to identify and exclude those patients by performing delirium screenings at least twice a day. Third, we included both ICU and CCU patients in our study because the number of patients on the ICU we could ask to fill out the questionnaire was limited due to sedation or mechanical ventilation. We only included acutely admitted CCU patients and think that the severity of disease was comparable with the ICU admissions. Fourth, unfortunately only patients who were conscious and well enough to complete the ALDS item bank could be included in the study. This limitation makes it difficult to generalise the results of our study to all ICU patients. Fifth, the high level of agreement observed in this study could be a potential contributor to the high levels of agreement between patient and proxy. Finally, these data are from a single centre in the Netherlands. Results may differ in other settings and parts of the world.

Relatives in close contact with the critically ill patient can adequately reflect the patient's functional status on admission to the ICU and CCU by use of the ALDS item bank and the SF-12. When comparing the physical disability score of the ALDS with the PCS of the SF-12, the highest intra-class correlation between patients and their proxies was found using the ALDS item bank. The ALDS could therefore be a useful alternative for the PCS of the SF-12.