Search

Objectives: Although widely used and reported in research for the evaluation of groups, measures of health status and health-related quality of life have had little application in clinical practice for the assessment of individual patients. One of the principal barriers is the demonstration that these measures add clinically significant information to measures of function or symptoms alone. Here, we review the methods for evaluation of construct validity in longitudinal studies and make recommendations for nomenclature, reporting of study results, and future research agenda., Methods: Analytical review., Results: The terms "sensitivity" and "responsiveness" have been used interchangeably, and there are few studies that evaluate the extent to which health status or health-related quality-of life measures capture clinically important changes ("responsiveness"). Current methods of evaluating responsiveness are not standardized or evaluated. Approaches for the assessment of a clinically significant or meaningful change are described; rather than normative information, however, standardized transition questions are proposed. They would be reported routinely and as separate axes of description to capture individual perceptions., Conclusions: Research in methods to assess the subject's evaluation of the importance and magnitude of a measured change are critical if health status and health-related quality-of-life measures are to have an impact on patient care.

Responsiveness, the ability to detect meaningful clinical change, is a critical attribute of instruments used to evaluate outcomes of treatments. The authors hypothesized that self-administered symptom severity and functional status questionnaires are more responsive to clinical improvement after carpal tunnel release than traditional physical examination measures of strength and sensibility. Data were obtained from a randomized clinical trial of endoscopic versus open carpal tunnel release conducted in four university medical centers. Patients were evaluated before surgery and 3 months after surgery. Seventy-four patients indicating that they were more than 80% satisfied with the results of surgery were assumed to have clinically meaningful improvement and were the focus of the analysis. Evaluations included questionnaires assessing symptom severity, functional status, and activities of daily living as well as measurement of grip, pinch, and abductor pollicus brevis strength, and 2-point discrimination and Semmes-Weinstein pressure sensibility. Responsiveness was calculated with the standardized response mean (mean change/standard deviation of change) as well as the effect size (mean change/standard deviation of baseline values). The symptom severity scale was four times as responsive, and the functional status and activities of daily living scales were twice as responsive, as the measures of strength and sensibility. Self-administered symptom severity and functional status scales are much more responsive to clinical improvement than measures of neuromuscular impairment and should severe as primary outcomes in clinical studies of therapy for carpal tunnel syndrome.

Short measures of health status are used increasingly in health services research, yet their sensitivities to clinical change have not been compared with longer, established instruments. In this study, 5 health status measures were administered preoperatively and 3 months postoperatively to 54 patients undergoing total hip arthroplasty. These instruments included the Sickness Impact Profile (SIP)--an established, long measure--and 4 short forms: the SF-36, Functional Status Questionnaire, shortened Arthritis Impact Measurement Scales, and Modified Health Assessment Questionnaire. Scores for physical, psychological, and global dimensions were constructed by aggregating subscales. Sensitivity to change, or responsiveness, was expressed with the standardized response mean (SRM), calculated as the mean change in score divided by the standard deviation of the change in score. The sampling distribution of the SRM was estimated with a jackknife procedure. Preoperative scores were moderately to highly correlated across instruments. The physical and global dimension SRMs of the brief health status measures ranged from 0.85 to 1.27 and were as large as or larger than the corresponding SIP SRMs. The SIP had the highest SRM on the psychological dimension. None of the instruments was significantly more sensitive than the others at the critical value (P = 0.005) adjusted for multiple comparisons. The brief health status measures were equally or more responsive than the SIP after total hip arthroplasty in the physical and global dimensions. Much larger samples are required to demonstrate statistically significant differences in SRMs among instruments.

This study represents a long-term effort to find optimal techniques for evaluating outcome in patients who have undergone total joint arthroplasty. Sensitivity of five health status questionnaires was studied in a longitudinal evaluation of orthopedic surgery. The questionnaires (Arthritis Impact Measurement Scales [AIMS], Functional Status Index [FSI], Health Assessment Questionnaire [HAQ], Index of Well Being [IWB], and Sickness Impact Profile [SIP]) were administered to 38 patients with end-stage arthritis at three points in time: two weeks before hip or knee arthroplasty, and at three-month and 12- to 15-month follow-up. Response values (i.e., changes within patients) were calculated on four scales: global health, pain, mobility, and social function. By the three-month follow-up, most instruments detected large mean responses in global health, pain scores, and mobility. Smaller changes on these scales were found between three and 12 to 15 months. Social function showed small to modest gains at successive follow-ups. Standardized response means were calculated to assess sensitivity to detect change. Confidence intervals for these indices were constructed using a jackknife procedure, and significance tests were performed by pairing selected indices. Finally, the study projected sample sizes required to assess a new therapy, using each response. These statistical tools facilitated comparisons among instruments and may prove useful in other settings.

A natural experiment provided an opportunity to determine whether physical therapy (PT) service reduction would affect the outcomes of total joint arthroplasty. A sample of 200 patients with rheumatoid arthritis or osteoarthritis who underwent total hip or total knee replacement during successive reduction of services was randomly selected. The demographic and clinical characteristics of patients in each period were similar. Despite large differences between groups in hours of PT per patient, there were no major differences in length of stay between the groups, or compared to similar hospitals, functional status at discharge, or numbers of surgical complications. Implications for health services organization are discussed.