Outcomes research is an emerging discipline in health care today. A major feature of outcomes research is its focus on broad definitions of health, including all aspects of an illness that might be directly affected by the disease itself or as a result of treatment. These outcomes are especially important to men with prostate cancer, in whom the disease and its therapies have a profound impact on physical function, sexual and marital well-being, and mental health.[1-7] Although recent efforts have measured outcomes of prostate cancer treatments among predominantly white populations, there is little information about racial/ethnic minorities, persons with low literacy skills, or the poor, groups that constitute a disproportionately large proportion of the prostate cancer population. In addition, few outcome studies include data on a large percentage of individuals prior to the initiation of therapy.
Prostate cancer is the most common cancer in men in the United States, accounting for almost 200,000 new cases and 40,000 deaths in 1998. The incidence of prostate cancer increases with age; 50-year-old men have almost a 10% risk of developing clinically significant prostate cancer during their lifetime. Although previous studies of prostate cancer focused on survival as the primary outcome measure, more recent efforts have centered on health-related quality of life (QOL).[9-11]
For example, the Cancer of the Prostate Strategic Urological Research Endeavor (CaPSURE) database effort has amassed information on quality of life (QOL) and other outcomes in 1,419 patients with prostate cancer seen in private medical offices and university clinics throughout the United States. One-fifth of the CaPSURE subjects are newly diagnosed patients with prostate cancer, 92% are white, and almost all are covered by Medicare or private health insurance plans. The CaPSURE database is designed to help practicing urologists make clinical decisions on the basis of such data as the results of treatment in typical care settings for privately insured or Medicare patients. Much of the information is provided from patients self-reports, using standardized QOL instruments.
Data similar to those provided by the CaPSURE project are especially important in the Veterans Affairs (VA) medical system, the largest integrated medical health system in the United States. Despite the fact that over 20,000 veterans underwent a radical prostatectomy for the treatment of localized prostate cancer from 1993 to 1997, little is known about their subsequent QOL. Similarly, medical castrating agents, such as luteinizing hormonereleasing hormone (LHRH) agonists, are among the most commonly prescribed therapies in the entire VA system, but their effects on health status have not been evaluated.
In order to obtain more information on prostate cancer in the VA system, we initiated an observational database effort, the VA Cancer of the Prostate Outcomes Study (VA CaPOS). Unlike CaPSURE, this VA study includes information on prostate cancer patients who are predominantly of lower socioeconomic status, the majority of whom are poor and over half of whom are African-American. Many participants have a low literacy level and, thus, are unable to complete self-administered standardized QOL instruments.
In this article, we provide an overview of observational database efforts of this large study of lower socioeconomic status patients, present baseline information on individuals in the VA CaPOS, and summarize results of the study to date
The VA CaPOS, initiated in 1995, is based on well-tested elements of chronic disease databases and includes information on clinical, epidemiologic, health-related QOL, and economic outcomes associated with prostate cancer and its treatment. The VA CaPOS includes both cross-sectional and longitudinal databases. The cross-sectional databases are used to assess feasibility, reliability, and validity of instruments and methods, while the longitudinal database is used to evaluate the time course of health outcomes, including QOL, preferences for health states, satisfaction with care, health literacy, and reasons for transferring to the VA health care system.
The organization of the VA CaPOS data is shown in Table 1. The individual assessments and results of the VA CaPOS are described in more detail in our prior publications.[13-15] All aspects of the data collection effort are overseen by a coordinating committee consisting of two medical oncologists, one urologist, two psychologists (one clinical and one educational), one urology nurse, one statistician, one database manager, and one project manager. Monthly investigators meetings and weekly operations meetings providea forum to discuss: (1) study design, (2) report preparation, (3) recruitment goals, (4) participant retention in the longitudinal data collection effort, and (5) quality control for data collection and management.
Because the VA CaPOS is a multicenter effort, medical oncologists and urologists and data collectors from sites at five collaborating VA hospitals attend quarterly conference call meetings and annual meetings at the VA Chicago Health Care System site for training and review of methods and procedures. The project coordinator at the VA Chicago Health Care System makes weekly telephone calls to all site data collectors to ensure the attainment of recruitment and retention goals and the standardization of questionnaire administration.
Patients are invited to enroll in the study by their VA physician at the time of a clinic visit. Patients with stable or progressive metastatic prostate cancer are invited to participate in the QOL assessment, irrespective of disease duration, severity, or type of treatment. A second effort to obtain longitudinal QOL data on patients with localized or metastatic prostate cancer targets all newly diagnosed prostate cancer patients, with enrollment occurring after the diagnosis of prostate cancer but prior to the initiation of any treatment. Six VA medical centers (Long Beach, California; Shreveport, Louisiana; Durham, North Carolina; Hines, Westside, and Lakeside, Chicago, Illinois) are participating in the project.
Clinical data for each patient includes a history of the prostate cancer diagnosis; biopsy and pathology results; treatments (including surgery, hormonal therapy, irradiation, and watchful waiting), and treatment dates; Karnofsky performance status; medications; and laboratory values (eg, prostate specific antigen [PSA] and hemoglobin). Other variables of interest include QOL, preferences for health states, health literacy, satisfaction with care, perceived involvement in care, and reasons for transferring to the VA for health care. Blood samples are also obtained to evaluate racial variations in the androgen receptor gene, through analysis of CAG repeat length. Newly diagnosed patients are asked to participate in an interviewer-assisted CD-ROMbased educational effort describing prostate cancer, alternative therapeutic options for the individual patient, and potential outcomes.
Data for the VA CaPOS are collected from several sources: the health care provider, patient, patients spouse, and paper and electronic medical records. At the time of enrollment in the longitudinal database, the patient and physician provide information on preferences for alternate QOL states. A research assistant completes information on medical history, demographics, comorbidities, stage of disease, and prior treatments, and performs a 30-minute interview to assess health literacy and preferences for health states. In several of the cross-sectional studies, we have also conducted face-to-face interviews in the clinic to assess QOL, reasons for seeking care in the VA, and satisfaction with care. Patients in the longitudinal database respond to these questions in a telephone interview following the initial clinic visit.
Patients with spouses or significant others are asked for their permission to contact this individual. After the spouse or significant other consents to participate, an interviewer collects information on perceived involvement in care, caregiver QOL, and caregiver perception of patient QOL. These interviews are conducted within a week of the patient interview. Patients with newly diagnosed prostate cancer enrolled in VA CaPOS are followed at 3 and 12 months during the first year and at 6-month intervals thereafter.
Quality of life is assessed using four instruments. In the cross-sectional databases, we are using the European Organization for Research and Treatment of Cancer Quality of Life Questionnaire (EORTC-QLQ-30), the Functional Assessment of Cancer Therapy-General scale (FACT-G), and the Quality of Life Index (QLI).[17-19] In the longitudinal database, we are including the EORTC-QLQ-30, QLI, and the Medical Outcomes Study Short Form
36 items (SF-36). The EORTC-QLQ-30 and FACT-G allow for determination of subscale scores to describe important QOL domains of daily functioning and well-being. The QLI is a brief measure that allows for the estimation of a global composite QOL score.
The SF-36 is used primarily to assess QOL of the spouse or significant other. This measure allows for determination of subscale scores to describe various QOL domains and permits comparison with healthy adults of comparable ages. To assess spousal perception of patient QOL, we are using the EORTC-QLQ-30 and FACT-G with instructions to the spouse to provide her judgment of patient QOL. Physician ratings include the Karnofsky Performance Status Scale, which yields a single score measuring physical and functional impairment on a scale from 100 (no impairment) to 0 (death).
Preferences for Health States
Based on previous QOL research, we identified six attributes associated with prostate cancer: pain, sexual function, bladder and bowel function, fatigue and energy, mood, and appetite.[3-5,9,22,23] We established three levels for each attribute that were then used to develop three clinically realistic health states for prostate cancer (health state A, stable; health state B, early progression; and health state C, advanced metastatic disease). For each health state, patients determine the number of years in perfect health that would be preferred to 10 years in that particular health state.
In the second part of the interview, patients are asked to rank the six attributes according to importance and assess their current health. Patients rank the attributes by considering a hypothetical health state in which each of the six attributes is at the worst level and choosing a medication to restore one attribute to perfect health.
Current health assessments are obtained by having the patient select one of the three levels for each attribute that best describes his health over the past month. In the current health state preference, patients are asked to determine the number of years of perfect health that they would prefer to their current state of health.
Physicians and nurses who are treating the patient on the day of the interview receive a clinician questionnaire that is similar to the patient instrument. These health professionals are asked to determine the number of years in perfect health that would be equivalent to the patients assessments of the three health states and their current health status, priority ranks of the six attributes, and current health status for the six attributes.
Satisfaction With Care
In the VA CaPOS, the 12-item Attitudes Toward Care Questionnaire has been adapted from the 43-item Patient Satisfaction Questionnaire used in the Medical Outcomes Study. The Patient Satisfaction Questionnaire is based on a multidimensional conceptualization of satisfaction with care. According to this conceptualization, satisfaction involves a number of domains that are relevant to health care interactions; these include access, convenience, technical quality, interpersonal aspects, and general satisfaction. The 12-item Attitudes Toward Care surveys dimensions of satisfaction, including access, convenience, interpersonal aspects, and technical quality, but the measure is brief and easy for the patient to complete with the assistance of a face-to-face or telephone interviewer.
The Attitudes Toward Care Questionnaire can be used to evaluate a specific treatment visit or a treatment course. In our studies, we use the Attitudes Toward Care Questionnaire for patient assessment of the course of treatment for prostate cancer. Although the MOS Patient Satisfaction Questionnaire has been widely used for this purpose, our reports are the first to employ the Attitudes Toward Care Questionnaire.
The Rapid Estimate of Adult Literacy in Medicine (REALM) is being used to assess the literacy ability of patients with prostate cancer, serving as an indicator of functional literacy skills. The REALM has been validated previously, and it can be administered and scored in 1 to 3 minutes by personnel with minimal training. It is comprised of 66 words that patients are expected to recognize, pronounce, and know the meaning of in order to participate effectively in their own health care. The REALM scores (0 to 66) can be converted into four reading grade levels: 0 to 3rd grade (0 to 18), 4th to 6th grade (19 to 44), 7th to 8th grade (45 to 60), and 9th grade and above (61 to 66). REALM scores are highly correlated with the Test of Functional Health Literacy in Adults.