Responsiveness of goal attainment scaling in a randomized controlled trial of comprehensive geriatric assessment

https://doi.org/10.1016/S0895-4356(03)00132-XGet rights and content

Abstract

Background and Objective: Frail elderly patients have complex problems that require a multidimensional assessment and a range of treatment goals. Goal Attainment Scaling (GAS) measures multiple, individualized goals, but its responsiveness in comparative clinical trials has not been established.

Methods: We assessed the responsiveness of GAS in a randomized, controlled trial of an interdisciplinary Mobile Geriatric Assessment Team (MGAT) in 265 rural frail older adults. Sensitivity to change was compared with standard measures; clinical meaningfulness was assessed in relation to a patient and a blinded physician global measure.

Results: At 3 months follow-up, GAS was the most responsive measure (standardized response mean 1.22, Norman's responsiveness statistic 0.58) compared with the Barthel Index (1.13, 0.46), Physical Self-Maintenance Scale (0.10, 0.16, 0.02), Instrumental Activities of Daily Living (0.23, 0.00), and modified Spitzer Quality of Life Index (−0.04, 0.00).

Conclusions: Only GAS detected clinically important change associated with the MGAT intervention in these frail elderly patients. Clinometric measures can offer a responsive means of evaluating complex interventions.

Introduction

The systematic evaluation of specialized geriatric programs faces challenges that arise from the nature of the patients being served. These patients are typically frail, have multiple problems, and vary in what they need and expect of treatment [1]. Programs usually respond to this heterogeneity by offering a range of services such as medicine, nursing, social work, and physiotherapy, delivered by multidisciplinary teams. In contrast to these programs' many interventions, their evaluation tends to focus on a limited number of outcomes [2], [3]. The choice of outcomes is usually constrained by tradition, availability, and feasibility. Consequently, the usual choice yields only about two primary outcomes, which typically have been developed using a psychometric, rather than a clinometric, approach [4].

Despite the theoretical advantages of clinical relevance and patient-centeredness, clinometric measures have been used much less than psychometric ones [5]. In earlier studies, we tested the feasibility, validity, reliability, and responsiveness of one clinometric measure, Goal Attainment Scaling (GAS), in frail elderly patients [4], [6], [7]. These tests supported the use of GAS and, in particular, its markedly greater responsiveness compared with measures using a psychometric paradigm. We further tested GAS by cross-validating it in patients with dementia [8], [9], [10] and in patients with traumatic brain injury [11]. Such patients, who have many goals with varying levels of achievement being judged successful, share important attributes with frail elderly persons.

Because of its responsiveness, GAS may be useful in evaluative studies of frail older adults. On these grounds, and because in our experience GAS is a useful patient management tool, we chose GAS as the primary outcome measure in an evaluation of a specialized geriatric intervention, known as the Mobile Geriatric Assessment Team (MGAT), targeted at frail, community-dwelling, older adults.

A randomized, controlled trial, although chiefly an opportunity to test an intervention, provides the opportunity to test the outcome measures that are used. The measures are varied, and the patients and interventions are held constant. Our hypothesis, with GAS as the primary outcome, was that GAS would be the most responsive measure of those usually used to test specialized geriatric interventions. We have described the results of the effectiveness of the MGAT intervention compared with usual care elsewhere [12]. Here we present our planned analysis of the responsiveness of the primary outcome measure (GAS) compared with the secondary measures, which were chosen, in part, because of their use in other trials of specialized geriatric interventions.

Section snippets

Design, patients, and setting

Most trials of specialized geriatric interventions have been based in urban teaching centers [13], [14]. Because we were interested in extending our understanding of how these interventions work, we focused on rural, community-dwelling, frail elderly patients as a rigorous test of comprehensive geriatric assessment. The design was a randomized, controlled trial, comparing usual care by family physicians with care in which their efforts were supplemented by a specialized geriatric intervention.

Results

Table 2 presents the results of the baseline measures and the change in the mean scores at 3 months by treatment status. At baseline, the groups were comparable on all measures, and in each group a similar number of goals were set (5.2±1.9 in the usual care group and 5.7±2.1 in those receiving the intervention). At 3-month follow-up, GAS was the only measure to demonstrate a statistically significant difference (Table 1). The clinical importance of the GAS score is supported by the proportion

Discussion

The effectiveness of a specialized geriatric intervention compared with usual care alone had been demonstrated using a clinometric measure, GAS, as the primary outcome [12]. GAS detected important differences in patient's function, safety, activity, and medication use, whereas several standard measures failed to capture these clinically important changes. In the current report we formally tested the responsiveness of GAS and found that it was more responsive than any of four standard tests.

To

Acknowledgements

This study was supported with grants from the National Health Research Development Program (NHRDP). Dr. Rockwood is supported by the Canadian Institutes of Health Research through an Investigator award.

References (31)

  • J.G Wright

    Evaluating the outcome of treatment. Shouldn't we be asking patients if they are better?

    J Clin Epidemiol

    (2000)
  • K Rockwood et al.

    Conceptualization and measurement of frailty in elderly people

    Drugs Aging

    (2000)
  • S.C Hedrick et al.

    Working group recommendations: measuring outcomes of care in geriatric evaluation and management units

    J Am Geriatr Soc

    (1991)
  • A.R Feinstein

    Clinimetrics

    (1987)
  • P Stolee et al.

    The use of goal attainment scaling in a geriatric care setting

    J Am Geriatr Soc

    (1992)
  • Cited by (0)

    View full text