FREQUENCY AND METHODOLOGIC RIGOR OF QUALITY-OF-LIFE ASSESSMENTS IN THE CRITICAL CARE LITERATURE

Objectives: To present a framework for describing/measuring quality of life (QOL) and health-related quality of life (HRQL), and to assess t he frequency and methodologic rigor of QOL studies in the adult critic al care literature. Data Sources: Computerized bibliographic search of published research, manual search of key intensive care unit (ICU) jo urnals, and citation review of relevant articles. Study Selection: We manually searched Critical Care Medicine American Journal of Respirato ry and Critical Care Medicine, and Intensive Care Medicine for the per iod January 1992 to July 1995 to assess the frequency of published QOL studies. Combined with a computerized bibliographic search, we found 64 studies that met our criteria that reported on patient-related outc omes (other than mortality) after hospital discharge. Data Extraction: We abstracted data on the nature of each study, the instruments used to measure QOL, and the methodologic rigor of the QOL assessments. We evaluated each study using criteria we developed to assess the validit y of HRQL measurements. Measurements and Main Results: in our manual s earch, we found 1,073 articles relevant to the practice of adult criti cal care. Of these, 19(1.7%)/1073 included QOL measurements. Combined with our comprehensive search strategy, we found 64 papers that evalua ted QOL in ICU patients. These papers dealt primarily with issues of: prognosis(32 [50%]), resource allocation (23 [36%]), and clinical pred iction (9 [14%]). We found no randomized trials that included post-ICU QOL as an outcome. With respect to the validity of the HRQL assessmen ts, 63/64 (98%) studies evaluated aspects at patients' lives that we c onsidered important. Thirty-three (52%) studies were limited in scope to specific aspects of HRQL, and 31 (48%) studies covered broad areas of HRQL. In these 64 studies, there were 108 different instruments use d. The reliability and validity of the instruments used were reported in 7 (6%) and 15 (14%) cases, respectively. For 81 (75%) instruments, the investigators either explained the results or used informative, qu alitative descriptors whose interpretation was self-evident. Overall, three (5%) of the included studies met all four methodologic criteria to assess the validity of HRQL assessments. Conclusions: QOL assessmen ts occur infrequently in the ICU literature and are of limited methodo logic quality. More studies using valid and reliable instruments are n ecessary to document the long-term QOL of critically ill patients, esp ecially those at risk of a ''poor'' outcome.