Abstract
Dear Editor,
We want to thank Benedetti et al. for their time and effort to respond to our paper “Advocacy for use of the modified Iowa Level of Assistance Scale (mILAS) for clinical use in patients after hip replacement: an observational study”. With this reply, we will provide additional information to clarify the questions and issues raised. However, before we provide a detailed response, we believe it is important to point out that the mILAS data used in this work were primarily collected as part of day-to-day practice as formerly advised by van Genderen et al. [1]. This means that the mILAS data were actually used as part of clinical decision-making, and thus not purely for the sake of psychometric properties testing. From the perspective of ecological validity [2], it makes sense to test psychometric properties of outcome measures as part of their real-world setting, however, this typically does result in a number of (mainly statistical) challenges. Some of those challenges were raised by Benedetti et al.
In order to properly use the ILAS in day-to-day care (e.g., to assess the progress of functional recovery, to inform therapeutic management and to aid in discharge planning), we (in collaboration with the clinical staff, among which the physiotherapy team, and patients) felt the need to reduce and alter some of its original items, namely:
- (1)
Remove the item “walking speed”. Walking speed was not considered a relevant item to determine if a person could safely walk without assistance. Moreover, we believed the walking speed item would hamper the implementation of the mILAS in daily practice, as this would require additional, therapeutically-irrelevant actions by the participating physiotherapists;
- (2)
Alter the item requirement “stair climbing”. We changed the item stair climbing from 3 steps to a complete staircase with the use of a guardrail which better reflects a home-situation and therefore makes it a more suitable therapeutic goal;
- (3)
Add the item “sit to supine”. From a clinical perspective, therapists noted that a proportion of people needing assistance getting back in bed while in the hospital and they believed this was a relevant milestone to function independently at home;
- (4)
Remove the scoring option “walking aids”. Although physiotherapists did score this item, we did not include it in the overall mILAS score, as we were solely interested whether a patient was able to independently perform those activities that he/she would need at their discharge destination.
Abovementioned changes have consequences for the analysis and interpretation of results compared to, for example, studies who particularly focus on the psychometric properties of the instrument, but not on the instrument in relation to its real-world context and use.
In the first point raised by Benedetti et al., they ask us why we treated the score 6 as missing. During regular care the scoring option 6 [not tested] is useful, after all, items like stair climbing are not typically performed during the first sessions of therapy. However, and here we full-heartedly agree with Benedetti et al., from a statistical analysis point of view, the scoring option 6 is far from ideal. First of all, because it substantially reduces the variability in item responses necessary for psychometric analyses, such as internal consistency and Rasch analysis. Additionally, as Benedetti et al. rightly note, it may obscure different reasons for missingness on items. Both of which may explain some of the differences in the outcomes between our study and that of our colleagues Benedetti et al. [3].
In their second point, Benedetti et al. point out an eye-catching difference between our study and the current evidence-base on the ILAS [3, 4, 5], namely the difference in item difficulty of “stair climbing”. We believe that our choice of wording was inconvenient and led to misunderstanding, however, we agree with the comment that stair climbing is indeed the most difficult item of all mILAS items (see Table 4) looking at the Rasch statistics alone [6]. However, the exact item difficulty estimate was very unreliable, given the poor fit-statistics and large error (difficulty −2.91, SE 1.31, infit 2.22 (Z-std 2.1) and outfit 4.27 (Z-std 1.8)). The statement that stair climbing was the easiest item, was based on the observation that the vast majority of the people could fulfill the item stair climbing without any assistance (see Table 2) [6]. As stated in our discussion, we considered to omit stair climbing from the mILAS because it seems to measure a different construct but concluded not to do so because it is a clinically relevant item to assess in order to evaluate if a person can be discharged based on their recovery of activities.
The third point raises the question if the operational definition of stairclimbing was comparable to the original ILAS. As stated earlier, we altered the original operational definition to reflect a patient’s living context (mostly home)-situation better and therefore make it a more suitable therapeutic goal.
The final point regarded the scale that measures the assistive devices. As indicated earlier, we did score the assistive device the patient needed during the mILAS. However, we found it irrelevant in assessing if the patient was ready for discharge and did not adopt it to the sum score of the mILAS. We were solely interested if the patient could perform all mILAS items required at his discharge destination regardless of the assistive device needed.
In conclusion, we all want a clinically useful tool that measures the mutual goal(s) of the patient and healthcare professional the best during hospitalization. In daily practice, this required us to adjust the ILAS measurement tool in order to properly fit the individual patient and his living context. We strongly believe that developing a clinical useful scale to properly inform day-to-day clinical practice with good psychometric properties should be advocated over creating a perfect standardized scale from the perspective of statistical uniformity with limited clinical usefulness. We therefore stand by our conclusion that the mILAS is a clinically sound measurement tool to assess the ability of patients to perform five functional tasks safely during hospitalization and at the same time monitor progress in order to anticipate and organize their discharge.
Citation
Response to letter to the Editor re ‘Advocacy for the use of the modified Iowa Levels of Assistance Scale for clinical use in patients after hip replacement: an observational study’