Adams, A.S., Soumerai, S.B., and Ross-Degnan, D. (1999). "Evidence of self-report bias in assessing adherence to guidelines." (AHCPR grant HS07357). International Journal for Quality in Health Care 11(3), pp. 187-192.
This study found that physicians tend to overestimate their adherence to practice guidelines, and thus, their self-reports of guideline use are not a reliable measure of quality of care. The researchers conducted a meta-analysis of 10 studies conducted between 1980 and 1996 of actual and self-reported physician adherence to practice guidelines. They found that clinicians overestimated their adherence to clinical practice guidelines by a median of 27 percent in 87 percent of comparisons of self-reported versus objectively obtained rates. Guidelines ranged from cancer and cholesterol screening to prevention of sexually transmitted diseases and drug treatments for conditions ranging from anxiety and heart attack to diarrhea. Physicians consistently overestimated their adherence to practice guidelines for four clinical services: mammography, breast exam, rectal exam, and testing for occult blood. This may be due to interviewer bias, or it may mean that physicians are simply reacting to what they think is expected of them. The researchers conclude that the increasing reliance on physician reports of adherence to practice guidelines as a measure of quality of care appears to greatly overestimate actual performance.
Goldie, S.J., Kuntz, K.M., Weinstein, M.C., and others (1999, May). "The clinical effectiveness and cost-effectiveness of screening for anal squamous intraepithelial lesions in homosexual and bisexual HIV-positive men." (NRSA training grant T32 HS00060). Journal of the American Medical Association 281(19), pp. 1822-1829.
Homosexual and bisexual men infected with HIV are at increased risk for human papilloma virus-related anal neoplasia and anal squamous cell carcinoma (SCC). These researchers developed a model to calculate lifetime costs, life expectancy, and quality-adjusted life expectancy for no screening versus several screening strategies for anal squamous intraepithelial lesions (ASIL, precursor to anal SCC) and anal SCC using anal Pap testing at different intervals for a hypothetical group of homosexual and bisexual HIV-positive men in the United States. They found that screening with anal Pap tests every 2 years, beginning in early HIV disease (CD4 cell count greater than 500) resulted in a 2.7-month gain in quality-adjusted life expectancy for an incremental cost-effectiveness ratio of $13,000 per quality-adjusted life year saved. Yearly screening with anal Pap tests provided additional benefit at an incremental cost of $16,600 per quality-adjusted life year saved. If screening was not initiated until later in the course of HIV disease (CD4 count less than 500), then yearly Pap test screening was preferred due to the higher prevalence of anal disease (cost-effectiveness ratio of less than $25,000 per quality-adjusted life years saved compared with no screening). Screening every 6 months provided little additional benefit over yearly screening.
Lenert, L.A., Treadwell, J.R., and Schwartz, C.E. (1999). "Associations between health status and utilities: Implications for policy." (AHCPR grants HS08349 and HS08582). Medical Care 37(5), pp. 479-489.
This study examines the impact of depression and health status on how individuals evaluate their current and hypothetical health states. It shows that patients in poor health tend to overvalue their current health relative to the most similar hypothetical state. In contrast, patients in good health tend to undervalue their current health state. The researchers did a cross-sectional study of 139 patients from three large primary care practices with various medical illnesses complicated by symptoms of depression. They measured the patients' health status, responses to a standard gamble, and scale preference measurements for patients' current health and for three hypothetical health states. Utilities for the best and worst states were similar across different levels of health status. However, standard gamble utilities for intermediate health states were higher for patients in poorer health than patients in better health. In patients with depressive illnesses, there were significant interactions between health and values that could result in systematic undervaluation of the health effects of treatments that primarily benefit more severely ill patients.
Monheit, A.C., Schone, B., and Taylor, A.K. (1999, Spring). "Health insurance choices in two-worker households: Determinants of double coverage." Inquiry 36, pp. 12-29.
Sixty percent of two-earner non-elderly couples chose double health insurance coverage in 1987, according to an analysis of data from the 1987 National Medical Expenditure Survey (NMES), a
national household survey conducted by researchers at the Agency for Health Care Policy and Research. In these cases, a working spouse is the policyholder of an employment-related health plan, and, also, is covered by a spouse's health plan. Double coverage was sensitive to the costs of insurance and the employment circumstances of working spouses, especially the employer's contribution to coverage. Households with double coverage had richer insurance benefits than other two-worker households. Also, attitudes toward risk and health status affected demand for double coverage. The likelihood of double coverage increased when either spouse was in fair or poor health and declined for those couples characterized as risk takers. Reprints (AHCPR Publication No. 99-R072) are available from the
AHCPR Clearinghouse.
Wyrwich, K.W., Nienaber, N.A., Tierney, W.M., and Wolinsky, F.D. (1999). "Linking clinical relevance and statistical significance in evaluating intra-individual changes in health-related quality of life." (AHCPR grant HS07763). Medical Care 37(5), pp. 469-478.
Numerous health-related quality-of-life instruments have emerged during the past two decades. These research instruments permit the cross-sectional comparison of groups and the longitudinal monitoring of groups or individuals. This article focuses on the SEM (standard error of measurement), a theoretically fixed test characteristic within a population. It shows that the SEM may be a viable link between the issues of clinically relevant and statistically meaningful intra-individual change on health-related quality-of-life (HRQL) measures. The researchers conducted a secondary analysis of data from a randomized controlled trial involving 605 outpatients with a history of cardiac problems attending the general medicine clinics of a major academic medical center. Baseline and followup interviews included a modified version of the Chronic Heart Failure Questionnaire (CHQ) and the SF-36. One-SEM changes in this population corresponded well to the patient-driven MCID (minimum clinically important difference) standards on all CHQ dimensions. The distributions of outpatients who improved, remained stable, or declined were generally consistent between CHQ dimensions and SF-36 subscales. The researchers caution, however, that the use of the SEM to evaluate individual patient change should be explored among other HRQL instruments with established standards for clinically relevant differences.
Zhou, X-H, Catelluccio, P., Hui, S.L., and Rodenberg, C.A. (1999). "Comparing two prevalence rates in a two-phase design study." (AHCPR grant HS08559). Statistics in Medicine 18, pp. 1171-1182.
An epidemiological study often uses a two-phase design to estimate the prevalence rate of a mental disease. In a two-phase design study, the first phase assesses a large sample with an inexpensive screening test. The second phase selects a subsample for a more expensive diagnostic evaluation. Disease status may not be ascertained for all subjects who are selected for disease verification, because some subjects are unable to be clinically assessed, while others may refuse. Since not all screened subjects are selected for diagnostic assessments, there is a potential for verification bias. In this paper, the authors propose the maximum likelihood and bootstrap methods to correct for verification bias for estimating and comparing the prevalence rates under the missing-at-random (MAR) assumption for the verification mechanism. They also propose a method to test this MAR assumption. Finally, they apply these methods to a large-scale prevalence study of dementia disorders.
Return to Contents
AHCPR Publication No. 99-0048
Internet Citation:
Research Activities newsletter. September 1999, No. 229. Agency for Health Care Policy and Research, Rockville, MD. http://www.ahrq.gov/research/sep99/
540 Gaither Road Rockville, MD 20850