Original article
Intraindividual change in SF-36 in ambulatory clinic primary care patients predicted mortality and hospitalizations

https://doi.org/10.1016/j.jclinepi.2003.08.004Get rights and content

Abstract

Objective

We sought to determine whether change in SF-36 scores over time is associated with the risk of adverse outcomes.

Study Design and Setting

7,702 participants in the Ambulatory Care Quality Improvement Project who completed a baseline and 1-year SF-36. Using logistic regression methods we estimated the 1-year risk of hospitalization and death based on previous 1-year changes in the physical (PCS) and mental (MCS) component summary scores.

Results

After adjusting for baseline PCS scores, age, VA hospital site, distance to VA, and comorbidity, a >10-point decrease in PCS score was associated with an increased risk of death (OR 2.3, 95% CI 1.6–3.4) and hospitalization (OR 1.8, 1.4–2.2). An increased risk was also seen with a >10-point decrease in the MCS (OR for death, 1.6, 1.1–2.3; OR for hospitalization 1.5, 1.2–1.8).

Conclusion

Change in SF-36 PCS and MCS scores is associated with mortality and hospitalizations, and provides important prognostic information over baseline scores alone.

Introduction

Health-related quality of life (HRQoL) is an important outcome in the assessment of interventions in the study of chronic diseases [1]. In addition, patient-centered HRQoL measures are used as global assessments of functional status and overall perceptions of health [2], and have been proposed as measures of disease severity. For example, the SF-36 predicts mortality in patients following coronary artery bypass graft surgery [3], successful treatment of chronic low back pain [4], and postdischarge medical services [5]. The SF-36 also has been used to predict 1-year total health plan expenses in a health maintenance organization [6], [7]. We have found that the component summary scores of the SF-36 also can be used to predict mortality in a large cohort of patients enrolled in primary care clinics [8].

For many individuals, however, HRQoL changes over time. These observed changes may reflect a true change in the HRQoL of an individual, or may result from random measurement error due to lack of precision in the instrument [9], [10], [11]. One approach to identify meaningful change has been to use the standard error of the measurement (SEM) [9], a distributional method that has the theoretical advantage of being sample independent [12]. Using the SEM criterion to define significant change, 31–64% of disadvantaged older adults had a change in one of the eight SF-36 subscales over a 12-month period [13].

Because a significant number of patients experience a change in HRQoL, we hypothesized that intraindividual change in HRQoL over time would be a better predictor of mortality and hospitalizations than a single observation. Specifically, we sought to determine whether a decrease in SF-36 scores over 1 year provides information about the risk of mortality and hospitalizations above and beyond that conveyed by a measure at one point in time using data obtained from the Ambulatory Care Quality Improvement Project (ACQUIP).

Section snippets

Subjects and setting

ACQUIP was a multicenter, randomized trial of patients receiving primary care in General Internal Medicine Clinics at seven Department of Veterans Affairs (VA) medical centers (Birmingham, AB; Little Rock, AR; San Francisco, CA; West Los Angeles, CA; White River Junction, VT; Richmond, VA; Seattle, WA) [14]. As part of this study, patients were asked to provide regular assessments of their health and satisfaction with care. This information was linked to the Veterans' Health Information System

Results

Among subjects who were sent the first SF-36, those who returned the questionnaire, compared to those that did not, were older (mean age 64.6 vs. 62.2 years), more likely to be Caucasian (78.8 vs. 68.0%), married (61.0 vs. 54.8%), not working (80.0 vs. 77.7%), and to have an annual income >$10,000 (70.8 vs. 68.5%).

Furthermore, patients who returned both the first and second SF-36 questionnaires, compared to those who returned only the first, were also older (mean age 65.4 vs. 62.7 years), more

Discussion

In this large, clinic-based study, change in SF-36, summarized in the PCS and MCS scores, was associated with both mortality and hospitalizations. We found that for many outpatients, HRQoL changed over a 1-year period, and that the direction and magnitude of intraindividual changes in measured health status provided information about the likelihood of future adverse events above and beyond that provided by a measurement at a single point in time. The relationship between change in SF-36 and

Acknowledgements

The research reported here was supported by the Department of Veteran Affairs, Health Services Research, and Development Service Grants SDR 96-002 and IIR 99-376. The views expressed in this article are those of the authors and do not necessarily represent the views of the department of Veterans Affairs.

References (31)

  • R.J. Gatchel et al.

    The association of the SF-36 health status survey with 1-year socioeconomic outcomes in a chronically disabled spinal disorder population

    Spine

    (1999)
  • D.G. Fairchild et al.

    A prediction rule for the use of postdischarge medical services

    J Gen Intern Med

    (1998)
  • M.C. Hornbrook et al.

    Assessing relative health plan risk with the RAND-36 health survey

    Inq

    (1995)
  • M.C. Hornbrook et al.

    Chronic disease, functional health status, and demographics: a multi-dimensional approach to risk adjustment

    Health Serv Res

    (1996)
  • C.A. McHorney et al.

    Individual-patient monitoring in clinical practice: are available health status surveys adequate?

    Qual Life Res

    (1995)
  • Cited by (0)

    View full text