Page 1 of 1

Factors affecting enrollment in literacy studies for English- and Span

Research Activities, February 2009, No. 342

Alexander, G.L. (2008, Summer). "A descriptive analysis of a nursing home clinical information system with decision support." (AHRQ grant HS16862). Perspectives in Health Information Management 5 (12), pp. 1-9.

Nursing home care quality can be improved by the use of information technology (IT). One example is the use of clinical information systems with decision support that make the most appropriate treatment options available for providers, increase accessibility of evidence-based protocols specific for nursing home residents, enhance data collection, and allow more rigorous analysis of outcomes. In this study, the researchers sought to determine the frequency and types of triggers in active alerts in a clinical decision support system for residents with specified diagnoses over a 6-month period. Alerts and triggers were recorded for conditions such as dehydration, constipation, skin integrity, decline or improvement in condition, and weight loss or gain. A total of 172 residents with 8 primary diagnoses were included in the analysis. This analysis using nursing homes that are early adopters of clinical information systems with decision support provides significant details on how these tools are being used for resident care.

Anstrom, K.J., Kong, D.F., Shaw, L.K., and others (2008, August). "Long-term clinical outcomes following coronary stenting." (Contract no. 290-05-0032). Archives of Internal Medicine 168(15), pp. 1647-1655.

Clinical trial patients with one clogged coronary artery have been shown to be less likely to need an additional operation to reopen that artery (revascularization) when they receive drug-eluting stents (DES) versus bare metal stents (BMS) to prop open the artery. A new study shows that patients in community practices, even those with multiple clogged coronary arteries (and thus multiple stents), had nearly 10 percent fewer revascularization procedures with DES than BMS 2 years after their initial surgery, similar to clinical trial results. The researchers examined outcomes of patients undergoing their first revascularization with DES or BMS from January 1, 2000 through July 31, 2005.

After adjustment for other factors affecting revascularization, DES reduced target vessel revascularization (TVR) rates at 6, 12, and 24 months compared with BMS. For example, 24-month rates of TVR for DES were 6.6 percent compared with 16.3 percent for BMS. The TVR benefit for DES was even better among patients with multiple vessel coronary artery disease (CAD): 1-vessel CAD, 8.3 percent; 2-vessel CAD, 9.7 percent; 3-vessel CAD, 16.2 percent. However, the type of stent did not significantly affect rates of death or nonfatal heart attack. Although the current U.S. Food and Drug Administration-approved indication for DES is limited to discrete, new lesions in native vessels with diameters of 2.5 to 3.5 mm, these findings indicate that the TVR benefits associated with DES versus BMS occur in patients with 1-, 2-, and 3-vessel CAD and are greater in patients with multivessel disease than those with single-vessel disease.

Clancy, C. (2008, July). "H-CAHPS survey reflects patient experiences." Managed Healthcare Executive, p. 34.

Patients choosing a hospital have not, until recently, been able to find information about the experiences other patients have had. Since March 2008, this incomplete state of hospital quality data has been changing, according to Carolyn Clancy, M.D., Director of the Agency for Healthcare Research and Quality (AHRQ). Data on patient experiences in the hospital have been added to the Federal government's Hospital Compare Web site (http://www.hospitalcompare.hhs.gov). This data is drawn from the Hospital Consumer Assessment of Healthcare Providers and Systems (H-CAHPS), a 27-question survey developed by AHRQ, and reflects the issues of greatest concern to patients. Questions focused on such concerns as the quality of communication and the helpfulness of the hospital staff. Patient feedback from the H-CAHPS survey 2007 shows both that hospitals are doing some things very well and that they clearly have more work to do. By the end of 2008, information from most of the nation's hospitals should be available. This information will help patients make more informed decisions and guide providers in their ongoing quality improvement efforts.

Corbie-Smith, G., Blumenthal, C., Henderson, G., and others (2008, August). "Studying genetic research participants: Lessons from the 'Learning about research in North Carolina' study." (AHRQ grant T32 HS00032). Cancer Epidemiology Biomarkers & Prevention 17(8), pp. 2019-2024.

Given the high cost of recruiting diverse populations for genetic research on cancer, the researchers decided to query existing cohorts of research participants who had taken part in the North Carolina Colorectal Cancer Study. The existing literature did not provide sufficient guidance on how best to capture the views of potential or former participants in genetic research. The Learning about Research in North Carolina (LeARN) study is an interview study aimed at eliciting research participants' perceptions of disease causality, perceptions of the pluses and minuses of genetic research, and the purposes of genetic research. Asking questions about genetic research was challenging and required extensive pretesting of open-ended questions. Between 6.2 percent and 9.2 percent of respondents either said that they did not understand the survey's four basic questions or answered in a manner that raised questions about whether they understood. However, the data collected were rich and varied and represented a broad spectrum of views.

Curtis, J.R., Cheng, H., Delzell, E., and others (2008, September). "Adaptation of Bayesian data mining algorithms to longitudinal claims data. Coxib safety as an example." (AHRQ grant HS10389). Medical Care 46(9), pp. 969-975.

Assessing drug safety after a product appears on the market is of great interest to patients, clinicians, pharmaceutical companies, and regulatory agencies. In evaluating adverse event reports and clinical trial results on medications, a Bayesian data mining method known as the Multi-item Gamma Poisson Shrinkage (MGPS) algorithm has been used. However, this type of data does not place much importance on the element of time. The researchers conducted a pilot study to evaluate Bayesian data mining for the analysis of longitudinal administrative claims data. Outcomes of cyclooxygenase-2 selective (coxib) nonsteroidal anti-inflammatory drugs (NSAIDS) were compared with outcomes of nonselective nonsteroidal anti-inflammatory drugs (NS-NSAIDS). Using the Medicare Current Beneficiary Survey to identify a group of 9,431 Medicare patients using NSAIDS, the researchers found that an empirical Bayes MGPS identified an association between current celecoxib use and acute myocardial infarction. Results from using traditional epidemiological methods were similar. Bayesian data mining methods seem useful and should be extended to different types of drug exposures, conclude the researchers.

Du, H., Valenzuela, V., Diaz, P., and others (2008). "Factors affecting enrollment in literacy studies for English- and Spanish-speaking cancer patients." (AHRQ grant HS10333). Statistics in Medicine 27, pp. 4119-4131.

The researchers sought to determine what factors affected the enrollment in two studies of literacy and health-related quality of life for 651 English-speaking and 487 Spanish-speaking ambulatory cancer patients. The purpose of the studies was to develop and validate a bilingual multimedia touchscreen program that allows patients with diverse literacy and computer skills to self-administer health-related quality of life (HRQL) questionnaires. The participants were not told in advance about the literacy component of the studies to minimize refusals due to shame about literacy skills. The researchers found that Spanish-speaking patients enrolled at a much higher rate (91 percent) than English-speaking patients (65 percent). For English-speaking patients, the recruitment barriers were older age and lower educational levels; for both English- and Spanish-speaking groups, lack of time and recruiting from private hospital sites were barriers. Recruiting sites with more indigent patients and longer clinic waiting times had higher enrollment, suggesting that financial compensation (they were paid $20 to participate) and time availability may be important recruitment factors.

Fried, L.E., Cabral, H., Amaro, H., Aschengrau, A. (2008, November/December). "Lifetime and during pregnancy experience of violence and the risk of low birth weight and preterm birth." (AHRQ grant HS08008). Journal of Midwifery and Women's Health 53(6), pp. 522-528.

Determining the relationship between a mother's experience with violence and its effect on her pregnancy may provide the medical community with strategies to prevent poor pregnancy outcomes. A study in Boston found that women who were exposed to violence either before or during their pregnancies were not at increased risk of delivering early or having babies born with low birth weights compared with women who never experienced violence. However, married women who suffered violence were more at risk for delivering babies with low birth weights than women who never experienced violence. Massachusetts researchers used data from 1,555 women who enrolled in Boston's Healthy Baby Program, which provides services to pregnant women living in areas with high rates of infant deaths.

Huang, C.X., Plantinga, L.C., Fink, N.E., and others (2008, July). "Phosphate levels and blood pressure in incident hemodialysis patients: A longitudinal study." Advances in Chronic Kidney Disease 15(3), pp. 321-331.

The majority of patients undergoing hemodialysis have high blood levels of phosphate because it is difficult to remove phosphate by dialysis. Elevated serum phosphate levels in hemodialysis patients have been associated with calcium deposition in blood vessels. This could potentially lead to vascular wall stiffness and resistant high blood pressure. Thus far, no clinical trial has examined the relationship between serum phosphate and blood pressure. Using a national group of hemodialysis patients, the researchers tested the hypothesis that these patients, if exposed to high levels of serum phosphate, would have higher blood pressure early in dialysis and over time. This prospective study found that elevated serum phosphorus levels at the beginning of dialysis were strongly and independently associated with higher systolic and diastolic blood pressure, as well as higher pulse pressure. This was also true 3 months after the baseline measurements. Further analyses showed that it was serum phosphate and not calcium that was associated with blood pressure at baseline and over time. These results reinforce the importance of serum phosphate control in reducing all-cause mortality in hemodialysis patients.

Johnson, K.B., Serwint, J.R., Fagan, L.A., and others (2008). "Computer-based documentation: Effects on parent-provider communication during pediatric health maintenance encounters." (AHRQ grant HS10363). Pediatrics 122(3), pp. 590-598, 2008.

The introduction of computer-based documentation (CBD) of patient encounters has raised concerns that use of computers in exam rooms may distract from or interfere with patient-physician interaction. However, a new study found that CBD improved parent-physician communication when it was used for pediatric health maintenance visits. Researchers used videotapes and audiotapes to compare verbal and nonverbal communication dynamics between clinicians and parents/children in health maintenance visits before and after implementation of the ClicTate CBD system. CBD visits were slightly longer than control visits (32 vs. 27 minutes). After controlling for visit length, amounts of conversation were similar during both types of visits. However, CBD visits were associated with a greater proportion of open-ended questions (28 vs. 21 percent), more use of partnership strategies, more social and positive talk, and a more patient-centered interaction style, but fewer orienting and transition phrases. These results support the integration of CBD into primary care pediatric visits. ClicTate, like most CBD tools, provides prompts designed to improve visit completeness.

Kemper, A.R., Boyle, C.A., Aceves, J., and others (2008, April). "Long-term follow-up after diagnosis resulting from newborn screening: Statement of the US Secretary of Health and Human Services' Advisory Committee on Heritable Disorders and Genetic Diseases in Newborns and Children." Genetics in Medicine 10(4), pp. 259-261.

In the interest of public health, newborn screening is provided to all newborns in the United States. The U.S. Secretary of Health and Human Services' Advisory Committee on Heritable Disorders and Genetic Diseases in Newborns and Children provides guidance to reduce the morbidity and mortality associated with heritable disorders. It focuses in particular on those conditions detectable through newborn screening. To improve the variable and inconsistent long-term followup of problems diagnosed through newborn screening, the Advisory Committee identifies key features of followup. The Committee also describes the four components central to achieving long-term followup.

Reprints (Publication No. 09-R006) are available from the AHRQ Publications Clearinghouse.

Lu, C., Frank, R.G., and McGuire, T.G. (2008). "Demand response of mental health services to cost sharing under managed care." (AHRQ grant HS10803). The Journal of Mental Health Policy and Economics 11, pp. 113-125.

This study compares how cost-sharing affects demand for mental health services under managed health care and traditional indemnity plans. Using the Agency for Healthcare Research and Quality's Medical Expenditure Panel Survey (MEPS) data from 1996, the researchers focused on privately insured employees (and their dependents) who have no choice of health plan. They found that although managed care itself decreased rates of care use, the effect of the coinsurance rate on demand for mental health services in managed care plans was almost nonexistent. By contrast, the coinsurance rate in traditional indemnity plans had a significant negative effect on rates of utilization. Deductibles had no effect on spending for any type of plan. The researchers conclude that in the presence of rationing mechanisms, i.e., managed care, the reduction of cost-sharing does not trigger an increase in mental health care use. A further implication is that any parity law focused on benefit expansion will have a smaller impact on equal access to mental health care than it would have had in a pre-managed care era.

McDonald, K.M., Davies, S.M., Haberland, C.A., and Geppert, J.J. (2008). "Preliminary assessment of pediatric health care quality and patient safety in the United States using readily available administrative data." (AHRQ contract no. 290-04-00200). Pediatrics 122, pp. e416-e425.

With health care costs for hospitalized children at almost $50 billion per year, health care quality evaluation for children has become an increasingly important priority. The authors aimed to develop indicators using inpatient administrative data to assess quality of inpatient pediatric care and access to quality outpatient care. The authors adapted the Agency for Healthcare Research and Quality's (AHRQ's) quality indicators for a pediatric population. They performed a systematic literature review and then convened four expert panels to review and discuss the evidence and then rate each indicator through a two-stage modified Delphi process. Using data from AHRQ's Healthcare Cost and Utilization Project Kids' Inpatient Database, they generated national estimates for provider- and area-level indicators. The panelists recommended 18 indicators for inclusion in the pediatric quality indicator set, including 13 hospital-level indicators, 11 of which were based on complications. Trends in rates varied by age: short-term complications of diabetes increased with age, while gastroenteritis admissions decreased with age.

Meier, F.A., Zarbo, R.J., Varney, R.C., and others (2008). "Development and validation of a taxonomy of defects." (AHRQ grant HS13321). American Journal of Clinical Pathology 130, pp. 238-246.

It is important to develop a consistent understanding of the reasons why pathology reports are sometimes amended, i.e., changed, after the reports have been released. Surgical pathology and cytopathology reports record the information extracted from tissue and cell specimens. Amendments of pathology reports document defects in the diagnostic process that lose information or add misinformation. The aim of the researchers was to develop a taxonomy of defects that would yield a high level of agreement by analysts of amended reports working at different provider organizations. The researchers identified four types of classification defects (misinterpretations, misidenti-fications, specimen defects, and report defects) that led to amendments being made by either clinicians or pathologists. Each of the 4 types of defects was, in turn, divided into 3 to 5 subtypes, and 430 cases of amended reports were then reviewed by 4 reviewers from a single institution and 30 cases by 7 reviewers from 7 organizations. Among the seven institutions, agreement was highest with misinterpretation fractions (23 percent to 29 percent) and lowest with report defects (29 percent to 48 percent). The researchers concluded that their taxonomy yielded excellent reproducibility and good agreement across institutions

Puggioni, G., Gelfand, A.E., and Elmore, J.G. (2008). "Joint modeling of sensitivity and specificity." (AHRQ grant HS10591). Statistics in Medicine 27, pp. 1745-1761.

Sensitivity and specificity are two customary performance measures associated with medical diagnostic tests. These measures are typically modeled independently as a function of risk factors. The potential clinical relevance of the joint modeling of sensitivity and specificity is to encourage those who study performance of screening tests (e.g., mammography) to think jointly about screening results and disease outcome. The authors argue that a model for the four cell probabilities that determine the joint distribution of screening test result and outcome result is needed. Using two different models, the authors assess the nature of the dependence between these two performance measures. They start by reviewing the dataset taken from three registries that are part of the Breast Cancer Surveillance Consortium. They then discuss coherent modeling for the joint distribution of screening outcome and disease outcome. This is followed by a discussion of computational issues associated with fitting the models. Finally, the authors analyze the dataset under two different models with regard to the dependence between sensitivity and specificity.

Ranji, S.R., Steinman, M.A., Shojania, K.G., and Gonzales, R. (2008, August). "Interventions to reduce unnecessary antibiotic prescribing: A systematic review and quantitative analysis." (Contract no. 290-02-0017). Medical Care 46(8), pp. 847-862.

Unnecessary antibiotic prescribing for outpatient acute respiratory infections (ARIs) such as bronchitis, ear infections, sinusitis, and pharyngitis can be reduced by active clinician education and targeting management of all ARIs, rather than single conditions in single age groups (such as ear infections in young children), concludes a new study. Researchers analyzed 30 studies on quality improvement (QI) approaches to reducing unnecessary antibiotic prescribing for ARIs. These approaches reduced the number of patients receiving antibiotics for ARIs an average of nearly 10 percent over the 6 months median followup. No single QI strategy or combination of strategies was clearly superior. However, active clinician education strategies trended toward greater effectiveness than passive strategies. Also, studies that targeted specific conditions or patient populations did not show as large an impact on the large community of antibiotic use as broad-based interventions such as those targeting adult patients with ARIs. These broad-based approaches saved an estimated 17 to 117 antibiotic prescriptions per 1,000 person-years.

Reid, R.J., Bonomi, A.E., Rivara, F.P., and others (2008, June). "Intimate partner violence among men: Prevalence, chronicity and health effects." (AHRQ grant HS10909). American Journal of Preventive Medicine 34(6), pp. 478-485.

This study found that more than one in four men (29 percent) had been victims of intimate partner violence (IPV) during their lifetime, 10 percent in the past 5 years, and nearly 5 percent in the past year. Researchers interviewed 420 English-speaking adult men enrolled in a large health care system for 3 or more years. The researchers used surveys that assessed types of IPV, overall health, and mental health. Results revealed that men aged 18 to 55 were twice as likely to be recently abused than men aged 55 and older (14.2 vs. 5.3), even though overall rates of physical (ranging from hitting, slapping, and shoving to choking or worse) and nonphysical IPV (threats, anger, and/or controlling behavior) were similar. Abuse was typically nonviolent or mildly violent, but occurred on multiple occasions and typically with only one intimate partner. Nearly one-third (32 percent) of men reported mildly violent IPV, and 39 percent reported moderately or extremely violent IPV. Compared with men who never suffered IPV, older men who had experienced IPV suffered from nearly three times more depressive symptoms and had low mental health scores on a standard scale.

Rhee, M.K., Ziemer, D.C., Caudle, J., and others (2008, July-August). "Use of a uniform treatment algorithm abolishes racial disparities in glycemic control." (AHRQ grant HS07922). The Diabetes Educator 34(8), pp. 655-663.

Differences in glycemic levels (A1c) disappear between blacks and whites in care settings where patient and provider behaviors are comparable and that emphasize intensification of diabetes medication, concludes this study. The researchers sought to determine the effects of using a uniform treatment algorithm on the glycemic levels of black and white patients in a setting where both patient and provider behavior could be assessed. The treatment algorithm was driven by point-of-care glucose measurement, facilitating immediate treatment implementation at a diabetes clinic, and providing specific instructions as to which medications to add and/or how to adjust dosages. The patients with type 2 diabetes (3,324 blacks, 218 whites) all made initial and 1-year followup visits, with a subset having a 2-year followup visit. Initially, the glycemic level (A1c) was higher in blacks than in whites. After 1 year of followup, the difference in A1c between blacks and whites narrowed but remained significant (7.7 percent vs. 7.3 percent). Patient adherence to medications and number of visits were similar for both groups. Provider behavior was similar as well, both with respect to the frequency and amount of medication intensification. Among the patients who returned for a 2-year followup visit (1,691 black, 114 white), A1c levels were no longer different.

Sharp, B.A. and Clancy, C.M. (2008, April/June). "Limiting nurse overtime, and promoting other good working conditions, influences patient safety." Journal of Nursing Care Quality 23(2), pp. 97-100.

In this commentary, Beth A. Collins Sharp, Ph.D., of the Agency for Healthcare Research and Quality (AHRQ), and Carolyn M. Clancy, M.D., Director of AHRQ, discuss several AHRQ-supported studies on the current shortage of registered nurses (RNs). One study found that elderly patients who were admitted to intensive care units (ICUs) with more nurse hours per day had significantly lower rates of central-line-associated bloodstream infection, 30-day mortality, ventilator-acquired pneumonia, and decubitus ulcers (pressure sores) than those admitted to ICUs with fewer nurse hours per day. Increased ICU nursing overtime was associated with catheter-associated urinary tract infection and decubitus ulcers (pressure sores).

These findings echo those of a second study on the impact of nurse fatigue on medical errors. In a 2004 study of nearly 400 nurses, nurses who worked more than 12.5 consecutive hours had 3 times the risk of making an error, such as incorrect medication or dosage, than nurses who worked fewer hours. A study of 502 ICU nurses over a 28-day period found that 27 percent of the nurses reported making at least 1 error, and 38 percent reported they almost made a mistake during the study period. Such extensive work hours conflict with recommendations that nurses provide direct patient care for no more than 12 hours in any given 24-hour period and less than 60 hours in a 7-day period.

Reprints (AHRQ Publication No. 08-R065) are available from the AHRQ Publications Clearinghouse.

Verrips, E., Vogels, T., Saigal, S., and others (2008, September). "Health-related quality of life for extremely low birth weight adolescents in Canada, Germany, and the Netherlands." (AHRQ grant HS08385). Pediatrics 122 (3), pp. 556-561.

The researchers sought to compare the effects of extremely low birth weight (ELBW-less than 1000 grams) on the health-related quality of life (HRQL) of adolescents aged 12 to 16 born in Ontario, Bavaria, and the Netherlands. ELBW children have been shown to have various physical, cognitive, affective, and behavioral problems. HRQL self-reported outcomes are important in evaluating treatment options and assessing the economic consequences of specific treatments. The HRQL of over 1,700 adolescents was assessed by using the Health Utility Index 3, a comprehensive generic measure of vision, hearing, speech, walking, dexterity, emotion, cognition, and pain. The Netherlands group had the highest mean gestational age and the lowest morbidity for disabling and nondisabling cerebral palsy. The summary scores showed a greater variety of health problems among adolescent children in Ontario and Bavaria than in the Netherlands. HRQL differences found in this study, especially in the cognitive domain, may affect the functioning of ELBW children later in life differently in different countries.

Volpp, K.G., and Landrigan, C.P. (2008). "Building physician work hour regulations from first principles and best evidence." (AHRQ grant HS15906). Journal of the American Medical Association 300(10), pp. 1197-1199.

In 2003, the Accreditation Council for Graduate Medical Education (ACGME) mandated that physicians in training work no more than 80 hours per week and no more than 30 consecutive hours. However, studies have not shown a consistent benefit of those work hour limitations. For example, one study showed no effect on mortality rates among Medicare patients from the ACGME standards while others showed some mortality rate reductions in both Veterans Administration (VA) hospitals and non-VA hospitals. The authors recommend several priorities or guiding principles for the medical community in developing specific alternatives for physician work hour regulation. They suggest rigorously studying alternatives for work hour reduction, measuring outcomes related to resident education, improving "sign-out" procedures, eliminating or minimizing situations in which residents work 24 to 30 continuous hours, improving monitoring of standards, increasing flexibility for implementation and enforcement, recognizing the importance of supervision and work intensity, and aligning incentive for payment with desired objectives.

Waitzkin, H., Schillaci, M., and Willgring, C.E. (2008). "Multimethod evaluation of health policy change: An application to Medicaid managed care in a rural State." (AHRQ grant HS09703). HSR: Health Services Research 43 (4), pp. 1325-1347.

The researchers studied the impacts of Medicaid managed care (MMC) in New Mexico at the individual, organizational/community, and population levels. They used a combination of sequential surveys to understand the impacts of MMC on individuals, ethnographic methods to assess effects for safety net institutions and local communities, and secondary databases to determine how sentinel events changed as population-level indicators. Three Statewide surveys, performed at 9-, 18-, and 27-month intervals, found barriers to care, access, and utilization. For example, the 27-month survey found that 25 percent of Medicaid respondents reported a cost barrier to care. However, in general, Medicaid patients had relatively favorable experiences following implementation of MMC. In contrast, the ethnographic interviews found that personnel in safety net organizations experienced the transition to MMC as stressful and chaotic. The organizations experienced heightened financial stress and mental health services declined sharply. Analyses of databases showed that immunization coverage levels decreased significantly after MMC implementation.

Weech-Maldonado, R., Elliott, M.N., Oluwole, A., and others (2008). "Survey response style and differential use of CAHPS® rating scales by Hispanics." (AHRQ grant HS00294). Medical Care 46(9), pp. 963-968.

These researchers examined differences in how Hispanic ethnicity and insurance status (Medicaid vs. commercial managed care) affect the use of the 0-10 rating scales in the Consumer Assessment of Healthcare Providers and Systems (CAHPS®) survey. Previous CAHPS® studies have shown that Hispanics report care that is similar to or less positive than for non-Hispanic whites, yet have more positive ratings of care. Other research has shown that blacks and Hispanics in the United States are more likely than whites to use the extreme responses in the scale. Hispanics had a greater tendency toward extreme responses in the CAHPS® ratings than non-Hispanic whites. In particular, they were more likely than whites in commercial plans to give the highest rating of "10." and often scores of 4 or less, relative to an omitted category of "5" to "8." These differences may be due to what the researchers call "extreme response tendency," a tendency to "respond systematically to questionnaire items on some basis other than what the items were designed to measure.

Return to Contents
Proceed to Next Article

 

Current as of February 2009
Internet Citation: Factors affecting enrollment in literacy studies for English- and Span: Research Activities, February 2009, No. 342. February 2009. Agency for Healthcare Research and Quality, Rockville, MD. http://archive.ahrq.gov/news/newsletters/research-activities/feb09/0209RA36.html