Research Briefs

Atkins, D. (2002). "Principles of preventive care." Primary Care Clinical Office Practice 29, pp. 475-486.

This review explores how clinicians can find recommendations on clinical preventive services that are up-to-date and evidence based, understand reasons for conflicting recommendations, and identify other issues that may affect the appropriateness of specific recommendations for their particular practice. The author, David Atkins, M.D., M.P.H., manages the activities of the U.S. Preventive Services Task Force on behalf of the Agency for Healthcare Research and Quality. He suggests that clinicians identify a source of prevention recommendations that is evidence-based and has a perspective that agrees with their practice. They should periodically review current recommendations to keep up with changing science. When a new screening test is introduced, they should ask whether it has been proven to improve important health outcomes or simply been shown to detect more disease. Clinicians should acknowledge the role of the patient in decisions in areas of uncertainty, approach prevention as a package of important services, and examine how well they are delivering high priority services.

Reprints (AHRQ Publication No. 03-R002) are available from the AHRQ Publications Clearinghouse.

Beckham, J.C., Calhoun, P.S., Glenn, D.M., and Barefoot, J.C. (2002). "Posttraumatic stress disorder, hostility, and health in women: A review of current research." (National Research Service Award training grant T32 HS00079). Annals of Behavioral Medicine 24(3), pp. 219-228.

There is increasing evidence that both posttraumatic stress disorder (PTSD) and hostility adversely affect health, but most of this research has been done on men. The impact of PTSD and hostility on the health of women is less clear. This study examines the current literature on PTSD, hostility, and health in women and discusses possible mechanisms underlying the relationship between PTSD and hostility on health outcomes in the context of a theoretical model. Hostile individuals tend to have greater levels of interpersonal conflict and lower levels of social support, more frequent and severe daily stressors, and greater levels of major negative life events. Studies show that among women, hostility is a risk factor for hypertension, coronary heart disease, and heart attack. However, the mechanisms responsible for this relationship are unclear. Similarly, the few studies of PTSD and health in women suggest that PTSD is associated with increased health problems, including arthritis, bronchitis, migraines, and gynecological complaints. However, more rigorous, focused research is lacking.

Bowers, M.R., and Kiefe, C.I. (2002, July). "Measuring health care quality: Comparing and contrasting the medical and marketing approaches." (AHRQ grants HS09446 and HS08843). American Journal of Medical Quality 17(4), pp. 136-144.

Clinicians typically measure care quality by medical outcomes such as improved patient mobility after surgery. Health care administrators often use managerial input measures such as the average number of nursing hours required for an outpatient surgery. Governments, insurers, and other payers may measure quality as the reduction of some disease in a given population. Patients tend to measure quality by personal outcomes or interactions with providers. These researchers summarized definitions and basic conceptual approaches to care quality in both health care administration and marketing research and then compared them on several attributes: basic goals, sources of measurement, role of patient perceptions, role of health care personnel, and need for risk adjustment. They developed a conceptual model combining the two approaches. They conclude that both clinicians and administrators could benefit from broader outcome measures. Patient satisfaction deserves more attention from medical researchers, whereas marketing approaches should go beyond patient satisfaction as the only outcome of interest.

Friedman, D.S., Tielsch, J.M., Vitale, S., and others (2002). "VF-14 item specific responses in patients undergoing first eye cataract surgery: Can the length of the VF-14 be reduced?" (AHRQ grant HS06280). British Journal of Ophthalmology 86, pp. 885-891.

The 14-item Visual Function questionnaire (VF-14) is designed to measure the impact of impaired vision on a patient's ability to perform daily tasks such as sewing, reading, or watching television. These authors sought to determine whether the VF-14 could be reduced in length without compromising its ability to act as an index of cataract-related visual impairment. To do this, the researchers analyzed the item-specific responses to the VF-14 before (771 patients) and 4 months after (552 patients) cataract surgery in one eye among patients enrolled in the Cataract Patient Outcomes Research Team (PORT) study. They found that 10 items correlated moderately with change in vision trouble, and 11 correlated moderately with change in vision satisfaction at 4 months after cataract surgery. Three items (recognizing people, cooking, and reading large print) were less responsive to cataract extraction and were more strongly associated with other coexisting ocular problems. The limited time saved by the VF-11 does not justify altering the already validated VF-14, conclude the researchers.

Fuhrer, R., Shipley, J.J., Chastang, J.F., and others (2002, August). "Socioeconomic position, health, and possible explanations: A tale of two cohorts." (AHRQ grant HS06516). American Journal of Public Health 92(8), pp. 1290-1294.

Some common susceptibility may underlie the social gradient in health and disease. This may explain why inequalities occur in cultures with different patterns of morbidity and mortality, concludes this study. The researchers examined whether the social gradient for measures of morbidity was comparable in English and French public employees and investigated risk factors that may explain this gradient. They studied 5,825 London civil servants and 6,818 French office-based employees using two health outcomes: long spells of sickness absence during a 4-year followup and self-reported health. Health behaviors showed different relations with socioeconomic position in the two groups. Psychosocial work characteristics showed strong gradients in both groups. In the presence of a similar social gradient in ill health in two culturally different groups, the different distributions of smoking, alcohol intake, and fruit and vegetable intake make it unlikely that these are major explanatory variables for the social gradient. On the contrary, the consistency of the gradient in early childhood environment factors and adult psychosocial work characteristics makes it plausible that these factors have universal importance in explaining social gradients in poor health.

Ginsburg, K.R., Forke, C.M., Cnaan, A., and Slab, G.B. (2002, August). "Important health provider characteristics: The perspective of urban ninth graders." (AHRQ grant HS07876). Developmental and Behavioral Pediatrics 23, pp. 237-243.

The researchers surveyed 2,602 urban, multiethnic 9th graders from 39 Philadelphia high schools and asked students to rate each of 36 provider characteristics on a 5-point scale ranging from 1 (not at all important to me) to 5 (extremely important to me). Four underlying factors accounted for 52 percent of the variance in item rating: interpersonal relationship, physical safety, emotional safety, and counseling ability. The interpersonal relationship factor included such descriptions as someone who is honest, trustworthy, and helpful; someone who understands teenagers and spends time with them; and someone who is friendly and nice. It was the strongest factor in the model, followed by physical safety and emotional safety. The fourth most important factor was the provider's counseling ability. The researchers conclude that communication with adolescent patients should be a standard component of health care training.

Hill, S.C., Thornton, C., Trenholm, C., and Woolridge. (2002, summer). "Risk selection among SSI enrollees in TennCare." Inquiry 39, pp. 152-167.

The issue of risk selection is especially important for States that enroll blind and disabled beneficiaries of Suppplemental Security Income (SSI) in Medicaid managed care. SSI beneficiaries have persistent needs for care, have a wide variety of chronic conditions, and often need atypical and complex services. Risk selection occurs when the health care needs of beneficiaries enrolled in a specific plan differ systematically from the needs of the overall beneficiary population, and payments do not reflect those needs. These authors assess the extent of risk selection among managed care plans for SSI beneficiaries over the first 3 years of Tennessee's Medicaid managed care program, TennCare. Using claims data containing fee-for-service expenditures prior to enrollment in managed care, they found substantial evidence of persistent risk selection among plans. Results are robust to most alternative measures of risk selection for most plans.

Reprints (AHRQ Publication No. 02-R092) are available from the AHRQ Publications Clearinghouse.

Kilbourne, A.M., Asch, S., Andersen, R.M., and others (2002). "Development and application of a method to assess timeliness of medical care for HIV symptoms." (AHRQ grant HS08578). Health Services & Outcomes Research Methodology 2, pp. 101-115.

Standards of care are needed that identify inappropriate delays in receiving care for serious symptoms of infection with the human immunodeficiency virus (HIV). Toward that end, these researchers evaluated the reliability and construct validity of provider-based standards for timely care for HIV symptoms and applied these standards to data from the AHRQ-sponsored HIV Cost and Services Utilization Study (HCSUS), a national probability sample of HIV-infected people in care. HIV physicians estimated the appropriate time to care for headache, cough, diarrhea, or weight loss for patients with HIV disease. They defined timely care as receiving care by the mean provider-specified acceptable number of days to care by specific CD4 count, compared with receiving care after the mean days to care. Up to 70 percent of HIV-infected individuals from the HCSUS sample did not receive timely care for their most bothersome symptom. Even though physicians rated patients with lower CD4 counts (indicating weaker immune systems) as requiring care within a shorter period of time, they were no more likely or less likely to receive timely care than others.

LaPointe, N.M., Kramer, J.M., Weinfurt, K.P., and Califf, R.M. (2002). "Practitioner acceptance of the dofetilide risk-management program." (AHRQ grant HS10548). Pharmacotherapy 22(8), pp. 1041-1046.

Dofetilide was approved in the United States in October 1999 as an antiarrhythmic agent used to treat atrial fibrillation or heart flutter. However, it can cause a dose- and concentration-dependent increase in the QT interval that can lead to torsades de pointes, which is potentially fatal. The FDA requires practitioners to complete an education program before they prescribe dofetilide, in part because it found that post-marketing labeling changes and warning letters to health care practitioners were ineffective in changing prescribing patterns and reducing the risk of torsades de pointes. These researchers assessed the opinions and knowledge retention of 91 practitioners at a large academic medical center after participation in the risk management program. Practitioners agreed the program was necessary but were undecided about whether the prescribing guidelines were easily understood or implemented. Identification of seven drugs that should not be taken with dofetilide differed significantly across groups (mean accuracy score was 41 percent for nurses, 80 percent for pharmacists, and 86 percent for physicians).

Radwin, L. (2000). "Refining the quality health outcomes model: Differentiating between client trait and state characteristics." (AHRQ grant K08 HS11625). Nursing Outlook 50, pp. 168-169.

The Quality Health Outcomes Model (QHOM) is a conceptual framework that was developed by the American Academy of Nursing Expert Panel on Quality Health Care to guide health services research. The model focuses on reciprocal relations among interventions, health care system characteristics, client characteristics, and outcomes. In the QHOM, interventions do not act directly on outcomes but rather are affected by and affect both client characteristics and health care system characteristics in producing client outcomes. However, the authors of this commentary propose that, although client state characteristics remain reciprocally related to the other QHOM components, client trait characteristics can only be unidirectionally related to those components. For example, the client's age will typically influence the selection of patient teaching strategies (interventions) or the unit to which the patient is assigned when admitted to a hospital (health care system characteristic). But, of course, neither teaching strategies nor unit assignment will affect the patient's age.

Riddle, D.L., Freburger, J.K., and the North American Orthopaedic Rehabilitation Research Network. (2002, August). "Evaluation of the presence of sacroiliac joint region dysfunction using a combination of tests: A multicenter intertester reliability study."(National Research Service Award training grant T32 HS00032). Physical Therapy 82(8), pp. 772-781.

These investigators examined the intertester reliability of assessments of sacroiliac joint (SIJ) region dysfunction (pain near the joint, presumably due to misalignment or abnormal joint movement) based on a composite of four tests of pelvic symmetry or SIJ movement, which are advocated for identifying people with this problem. They used the four tests on 65 patients with low back pain and unilateral buttock pain who were examined by one of 34 randomly assigned physical therapists. For the composite test results, percentages of agreement ranged from 60 to 69 percent. The researchers conclude that the reliability of measurements obtained with the four tests appears to be too low for clinical use. They suggest that the four tests not be used to examine patients suspected of having SIJ region dysfunction, although the role of therapist training in use of the procedures is unclear.

Salomon, J.A., Weinstein, M.C., Hammitt, J.K., and Goldie, S.J. (2002). "Empirically calibrated model of hepatitis C virus infection in the United States." (National Research Service Award training grant T32 HS00020). American Journal of Epidemiology 156(8), pp. 761-773.

Hepatitis C virus (HCV) is the most common blood-borne infection in the United States. It is not clear what the risks are of HCV leading to advanced liver disease, and treatments are effective in only 30 to 60 percent of patients. These authors developed a comprehensive epidemiologic model of HCV infection in the United States in order to gain insights into key uncertainties around the natural history of HCV and to improve the basis for projecting the future course of the epidemic. They reviewed the published literature to define plausible ranges around model parameters and used goodness-of-fit criteria to identify the range of parameter values that were consistent with available epidemiologic data on infection prevalence and mortality from liver cancer. Results indicated that rates of HCV progression to advanced liver disease may be lower than previously assumed. The authors also found that a wide range of plausible assumptions about heterogeneity in these rates, beyond that explained by age and sex, was consistent with observed epidemiologic trends.

Schneeweiss, S., Maclure, M., Soumerai, S.B., and others (2002). "Quasi-experimental longitudinal designs to evaluate drug benefit policy changes with low policy compliance." (AHRQ grants HS09855 and HS10881). Journal of Clinical Epidemiology 55, pp. 833-841.

A causal relation between drug benefit policy change and an increase in adverse outcomes can be tested by comparing the experience of a group of patients affected by the policy versus the (counterfactual) experience of the same patients if the policy had not been implemented. Quasiexperimental longitudinal designs with repeated measures can provide valid observational estimates of the counterfactual outcomes by comparing extrapolated prepolicy outcome trends with the observed postpolicy outcomes. If compliance with a policy is low, results may be biased toward the null (no policy effect), but a subgroup analysis of those who comply may be biased by nonignorable treatment selection. Using the example of reference drug pricing in British Columbia, these authors discuss assumptions for causal interpretations of such analyses and provide supplementary analyses to assess and improve the validity of findings.

Subak, L.L., Caughey, A.B., and Washington, A.E. (2002). "Cost-effectiveness analyses in Obstetrics & Gynecology: Evaluation of methodologic quality and trends." (AHRQ grant HS10856). Journal of Reproductive Medicine 47, pp. 631-639.

These authors used ten methodologic principles that should be incorporated in cost-effectiveness analyses (CEAs) to evaluate the methodologic quality and trends of CEAs included in studies published in the journal Obstetrics & Gynecology from 1966 through 1999. Thirty-four CEAs in the journal's studies met the inclusion criteria. Seven (20 percent) adhered to 10 of the principles, 5 (15 percent) to 9, and two (6 percent) to 8 of the 10 methodologic principles. The mean number of principles to which studies adhered was 5.7. Studies had high compliance (85 percent or more) with principles of research questions, probabilities, and effectiveness measures. They significantly improved over time in adherence to principles of time frame, perspective, costs, incremental analysis, sensitivity analysis, discounting, and total score. The investigators conclude that the CEAs evaluated adhered to only half the methodologic principles for performing CEAs but showed significant improvements in quality over time.

Zaslavsky, A.M., and Buntin, M.J. (2002, Summer). "Using survey measures to assess risk selection among Medicare managed care plans." (AHRQ grant HS10803). Inquiry 39, pp. 138-151.

These authors quantify risk selection among competing Medicare managed care plans, using beneficiary survey data from the Consumer Assessments of Health Plans Study (CAHPS®). Selection, measured by variation in plan-level prevalence of health conditions and predicted costs, was substantial. A plan with moderate adverse selection would have predicted costs 11.6 percent above an average plan. Only a small part of this variation was explained by geographic differences in the prevalence of health conditions among or within metropolitan statistical areas, indicating that the selection was driven by plan attributes. Plans serving members with greater health needs have the potential to establish programs to serve these sick members well, yet this places plans at financial risk. Hence, improved risk adjustment for chronic conditions may be warranted. Moreover, survey measures have the potential to measure the prevalence of such conditions reliably and consistently across plans.

Return to Contents

Current as of December 2002
AHRQ Publication No. 03-0011