|| Checking for direct PDF access through Ovid
There are questions about the treatment of HIV/AIDS that can only be answered by asking the patient. Some of the questions are clinically important, such as: among treatments that are equivalent in terms of virologic and immunologic outcome, which are best tolerated? How can we identify unsuspected symptoms? Other questions are important to policy makers, such as: what is the extent of need in communities? What is the natural history of disability? Answering these questions requires collecting patient-reported data on quality of life.In research, when we refer to `quality of life', we usually mean health-related quality of life. At a minimum, this includes several dimensions: physical function and symptoms, performance of social roles, emotional status, cognitive functioning, and how the patient feels about his or her health . All of these concepts can be measured adequately using sets of questions which are deployed in standardized questionnaires.I have a basic science colleague who chuckles every time he hears quality of life questionnaires referred to as `instruments'. But these questionnaires are in fact measurement tools, with coefficients of variation that compare favorably to those for parameters such as CD4 cell count and cholesterol [2,3]. As with all tools, quality of life instruments are important to the extent that they provide precise and unbiased answers to our research questions.In HIV, the first applications of quality of life assessment coincided with the advent of antiretroviral therapy . Since that time, quality of life data have added useful information to treatment trials, sometimes in support of clinical findings [5–9], sometimes at odds , sometimes revealing the only difference among treatment regimens . Some of the results make more sense today than they did at the time of publication, such as the finding that antiretroviral monotherapy decreased quality of life in patients with early disease [12,13].Although quality of life assessments are less time-consuming and less invasive than many clinical assessments, obtaining answers from patients poses its own challenges. Data collection requires adequate preparation of study personnel to assure co-operation. Since patient expectations can affect their responses, blinded designs are preferable. Protocols are needed to reduce missing data, which are more likely to occur when patients' quality of life is worse.Questionnaire administration may need to be tailored for patients with lower literacy and educational levels . Analytic strategies must be deployed to handle missing data. More data are needed to aid interpretation of what scores mean.In this issue of the Journal, an industrious research team from Barcelona reports on the head-to-head comparison of two questionnaires: the Medical Outcomes Study-HIV (MOS-HIV) Health Survey [15,16], a brief, widely used measure of health-related quality of life, and the Multidimensional Quality of Life (MQOL)-HIV, a newer instrument [17,18]. Like other studies comparing the test characteristics of different assays, the paper is a little technical and not terribly newsworthy. However, it provides valuable evidence on the `responsiveness' of the questionnaires.Responsiveness can be defined as the ability of a measurement instrument to detect change. In this study, the team used both patient reports of improvement and changes in clinical parameters to define change, making the results easier to interpret. For researchers selecting a measure to use in a study, or performing sample size calculations, the specific results of the study are useful since greater responsiveness means more power. It should be noted that this was not precisely a horse-race, as the two questionnaires do not aim to measure exactly the same things. Indeed, the results provide ammunition to both camps to support the clinical utility of their methods.