Research Briefs

Bertakis, K.D. and Azari, R. (2007). "Determinants of physician discussion regarding tobacco and alcohol abuse." (AHRQ grant HS06167). Journal of Health Communication 12, pp. 513-525.

The U.S. Preventive Services Task Force recommends that physicians screen adults for tobacco and alcohol abuse. Researchers investigated the impact of patient age, gender, education, income, patient health status, depression, alcohol abuse, and current smoking on the likelihood of physician discussion. They analyzed videotapes of initial primary care visits together with the results of previsit screening tests for over 500 patients. The results of the study showed that physicians discussed substance use with over 90 percent of patients who were either problem drinkers or smokers. Older patients were less likely to have their physician discuss health promotion behavioral changes. Physicians tended to address substance abuse with healthier, younger, male patients who currently abuse tobacco and alcohol.

Brown, S.E.S., Chin, M.H., and Huang, E.S. (2007). "Estimating costs of quality improvement for outpatient healthcare organizations: A practical methodology." (AHRQ grants HS10479 and HS13635). Quality & Safety in Health Care 16, pp. 248-251.

Many small outpatient health care organizations have not implemented quality improvement (QI) programs for chronic disease management due to uncertainties about the costs involved. Established standard methods for determining either costs or consequences are lacking for small organizations. Using a diabetes QI program already in place in some community health centers as a test case, researchers developed a practical set of methods for estimating both direct costs/revenues and cost/revenue consequences. They gathered data on direct costs/revenues such as staff time, purchase of items needed, and grants received through a self-administered survey. They collected cost/revenue consequences such as changes in patient use of services and physician activities using electronic billing data. These methods may help in comparing the relative cost-effectiveness of different QI programs and identifying those that should be widely disseminated.

Chou, A.F., Scholle, S.H., Weisman, C.S., and others (2007, May/June). "Gender disparities in the quality of cardiovascular disease care in private managed care plans." (AHRQ contract 290-04-0018). Women's Health Issues 17, pp. 120-130.

Process-of-care variables may account for some of the gender disparities in cardiovascular disease (CVD) outcomes in commercial managed care programs, concludes this study. It analyzed seven CVD quality of care indicators to examine the CVD care of members of a national sample of commercial managed care plans. Quality indicators ranged from use of beta blockers after a cardiac event to cholesterol screening and blood pressure control (140/90 mm Hg or less).

The researchers found low rates of adequate lipid control in both men and women, with a lower rate of control in women. Women with diabetes were 19 percent less likely than men to have low-density lipoprotein (LDL) cholesterol controlled at less than 100 mg/dL, and women with a history of CVD were 28 percent less likely to have their LDL cholesterol controlled. This suggests the possibility of less intensive cholesterol treatment in women. However, more women than men had controlled blood pressure, although the difference was only 2 percent (70.8 vs. 68.9 percent). Smaller gender differences were observed in measures related to screening and medication prescription.

Chou, A.F., Wong, L., Weisman, C.S., and others (2007, May/June). "Gender disparities in cardiovascular disease care among commercial and Medicare managed care plans." (AHRQ contract 290-04-0018). Women's Health Issues 17, pp. 139-149.

Researchers evaluated plan-level performance of seven quality of care measures (ranging from blood pressure and cholesterol control to comprehensive diabetes care) for CVD using a national sample of commercial and Medicare managed care plans. They also conducted key informant interviews with a subset of commercial plans. Over half of participating commercial plans showed a disparity of 5 percent or more in favor of men for cholesterol control measures among people with diabetes, a recent cardiovascular procedure, or heart attack.

Disparity was greatest (9.3 percent in favor of men) among those with recent acute cardiac events. Yet no commercial plans showed such disparities in favor of women. These gender differences favoring men were even larger for Medicare plans. For those plans, the mean differences in performance rates for men on the cholesterol control measure were 6.4 percent among people with diabetes and 8.5 percent among those with a recent cardiac event.

Clancy, C. (2007). "Mistake-proofing in health care: Lessons for ongoing patient safety improvements." American Journal of Medical Quality 22(6), pp. 463-465.

Latent conditions can create the opportunity for medical errors within health care settings through a lack of standardized equipment and procedures, poor visibility, or distraction. Hospitals can overcome many of these conditions by the use of innovative facility design. In new hospitals, the design of all patient rooms—inpatient, emergency, postrecovery, and ambulatory/diagnostic—can be standardized. Equipment and technology, such as infusion pumps and beds can also be standardized. Multimillion dollar investments are not required to reduce the likelihood of mistakes, notes Carolyn Clancy, M.D., director of the Agency for Healthcare Research and Quality (AHRQ). She points to a new AHRQ research synthesis that shows that mistake-proofing can be achieved through relatively small innovations such as wristband medical records and revamped medication bottles. Dr. Clancy asserts that mistake-proofing is a supplement to, not a replacement for, existing patient safety efforts.

Reprints (AHRQ Publication No. 08-R016) are available from the AHRQ Publications Clearinghouse.

Clancy, C. (2007). "The performance of performance measurement." HSR: Health Services Research 42(5), pp. 1797-1801.

Since the late 1980s, interest in and support for core performance measurement has been growing among State governments, accrediting organizations, public and private payers, and the Agency for Healthcare Research and Quality (AHRQ). Thus far, there have been few efforts to compare the validity of various approaches to performance measurement, according to AHRQ Director Carolyn Clancy, M.D.

A new study appearing in this issue of Health Services Research reports on an unprecedented comparison of three approaches to performance measurement: condition-specific measures, global explicit measures, and implicit measures based on professional judgments of overall quality of care. The researchers found a high level of agreement among them for summary measures of quality as well as substantial agreement across the three approaches for diabetes and preventive care. Also important for performance measurement is the ease of identifying available data sources, which increases as more and more providers adopt electronic health records. The growing use of health information technology allows the sharing of information across settings of care, decreases the cost of data collection, and improves the completeness of clinical data, notes Dr. Clancy.

Reprints (AHRQ Publication No. 08-R017) are available from the AHRQ Publications Clearinghouse.

Clancy, C. (2007, September). "Sleepless in the hospital: Evidence mounts that tired caregivers may compromise quality." Journal of Patient Safety 3(3), pp. 125-126.

Asking young doctors to work long hours without a break is deeply ingrained in our culture. We believe that it teaches dedication, stamina, and responsibility. Yet in doing so, we are placing patients at risk, which is unacceptable, asserts Carolyn M. Clancy, M.D., Director of the Agency for Healthcare Research and Quality (AHRQ), in a recent commentary. She cites several examples of AHRQ-supported and other research that demonstrates this. For example, one study found that first-year doctors-in-training who worked five extra long shifts (of 24 hours or more at a time without rest) per month had a 300 percent increase in their chances of making a fatigue-related preventable error that contributed to the death of a patient. One study found that sleep-deprived interns working 24-hour shifts made many more serious medical errors while working in intensive care units and crashed their cars more often than interns whose work was limited to 16 consecutive hours. Another AHRQ-supported study found that even the 12-hour shifts worked by most U.S. hospital nurses greatly increased their risk of making a medical error.

At the request of the U.S. House Committee on Energy and Commerce, AHRQ will be working with the Institute of Medicine of the National Academies to review the evidence to make recommendations for health care provider training and work hours.

Reprints (AHRQ Publication No. 08-R007) are available from the AHRQ Publications Clearinghouse.

Clancy, C. (2007, July). "TeamSTEPS: Optimizing teamwork in the perioperative setting." AORN Journal 86(1) pp. 18-22.

Working with the Department of Defense, the Agency for Healthcare Research and Quality (AHRQ) developed a toolkit to improve patient safety through improved communication and teamwork. Team Strategies and Tools to Enhance Performance and Patient Safety (TeamSTEPPS™) is built upon AHRQ's research in patient safety and health care quality and the U.S. military's expertise in team building.

In this commentary, AHRQ Director Carolyn M. Clancy, M.D., describes the toolkit's features. Each phase of the training mirrors the effort that goes into executing a military campaign: assessment; planning, training, and implementation; and sustainment. Lessons in the kit apply to any health care setting and can be used sequentially or as standalone modules. They stress leadership, situational awareness, anticipating and meeting other team members' needs, and communicating. The uniqueness of TeamSTEPPS™ lies in the fact that it gives users tools they can use to improve patient safety. It offers more than 130 mini case scenarios that can be customized for different specialty areas, such as the surgical suite.

Ordering information for the toolkit is available at the AHRQ Web site, http://www.ahrq.gov/qual/teamstepps. Reprints (AHRQ Publication No. 08-0002) are available from the AHRQ Publications Clearinghouse.

Clancy, C. and Slutsky, J.R. (2007, October). "Commentary: A progress report on AHRQ's Effective Health Care Program." HSR: Health Services Research 42(5), pp. xi-xix.

The goal of the Effective Health Care Program, supported by the Agency for Healthcare Research and Quality (AHRQ), is to provide up-to-date, evidence-based information about treatment options so that patients, clinicians, purchasers, and policymakers can make informed health care decisions. To date, AHRQ has released eight comparative effectiveness reviews, which can be found on the program's Web site http://effectivehealthcare.ahrq.gov.

These reviews can aid better treatment decisions. For example, Gastroesophageal Reflux Disease found that medications called proton pump inhibitors can be as effective as surgery in relieving disease symptoms and improving quality of life. The review Oral Diabetes Drugs found that most oral medications prescribed for type 2 diabetes are similarly effective for reducing blood glucose. However, the drug metformin is less likely to cause weight gain and may be more likely than other treatments to decrease so-called bad cholesterol. The Effective Health Care Program is already having an impact. For example, Consumer Reports Best Buy Drugs, a public education project of Consumers Union, uses findings from the program to help clinicians and patients determine which drugs and other medical treatments work best for certain health conditions. Also, the National Business Group on Health uses findings from the program in its evidence-based benefit design program.

Reprints (AHRQ Publication No. 08-R006) are available from the AHRQ Publications Clearinghouse.

Crane, P.K., Cetin, K., Cook, K.F., and others (2007). "Differential item functioning impact in a modified version of the Roland-Morris Disability Questionnaire." (AHRQ grant HS09499). Quality of Life Research 16, pp. 981-990.

Measuring the effects of back pain consistently across different demographic groups requires an assessment of scale bias. This assessment is needed to determine if scores for a given group may be artificially high or low, thus exaggerating the actual differences. Researchers sought to determine if the modified 23-item Roland-Morris Disability Questionnaire was subject to this problem. They evaluated it for differential item functioning (DIF) to see if particular test items yielded different responses from different demographic groups. The groups were based on the variables of gender, age, education, marital status, employment status, surgical status, and self-rated general health. Of the 23 items in the questionnaire, 18 showed DIF related to at least one variable, and all of the variables except gender were associated with DIF in at least one item. However, mean scores across demographic groups varied minimally. Thus, for most purposes, the DIF in the modified Roland-Morris Disability Questionnaire may be ignored without threatening the validity of the results.

Dobalian, A., Tsao, J.C., Putzer, G.J., and others (2007). "Improving rural community preparedness for the chronic health consequences of bioterrorism and other public health emergencies." (AHRQ grant H13110). Journal of Public Health Management Practice 13(5), pp. 476-480.

In the wake of a major bioterrorism event, there are likely to be extensive chronic health needs requiring long-term attention from health care providers. Little is known about how rural health care providers will respond to these needs following a public health emergency. In order to prepare for these events, a needs assessment of the public health system in rural communities is required. The assessment of existing resources and response mechanisms will rely on both qualitative (e.g., key informant interviews, focus groups with stakeholders) and quantitative (e.g., surveys that examine the knowledge-based testing of health care providers) methods. A preparedness needs assessment will incorporate key thematic elements based on the 10 Essential Public Health Services model, according to the authors. The needs assessment should also include an analysis of the surge capacity of facilities available within the rural community. The approach to needs assessment suggested by the authors is a potentially expensive one and likely to require at least 6 months.

Dohan, D. and Levintova, M. (November 2007). "Barriers beyond words: Cancer, culture, and translation in a community of Russian speakers." (AHRQ contract 290-2006-00231). Journal of General Internal Medicine 22(Suppl 2), pp. 300-305.

A study of Russian-speaking cancer patients in San Francisco revealed several language and cultural barriers to their care that may best be overcome with professional Russian interpreters. Researchers collected data from patients, providers, and families using focus groups at two medical centers with cancer clinics, individual interviews, and observations. The local cancer clinics did not provide adequate Russian language resources compared with Spanish and Chinese, even though more than 30,000 Russian-speaking immigrants reside in San Francisco. Also, Russian family members, especially patients' children, often did not have adequate Russian language skills to interpret medical and technical information and often edited messages to protect family members from the stress or despair.

The demanding nature of Russian patients also created barriers to care. Many would demand to see a certain doctor or to be seen without an appointment or complain about the lack of hands-on care that they were used to in their home countries.

Filardo, G., Nicewander, D., Hamilton, C., and others (2007, November). "A hospital-randomized controlled trial of an educational quality improvement intervention in rural and small community hospitals in Texas following implementation of information technology." (AHRQ grant HS15431). American Journal of Medical Quality 22(6), pp. 418-427.

Rural and small community hospitals typically have fewer resources and poorer quality of care than larger hospitals. They also tend to have little experience with quality improvement (QI). This article describes the planned methodology of the first study to apply a formal QI education program in rural U.S. hospitals. The research team will randomly assign 47 rural and small community Texas hospitals that had received a Web-based quality benchmarking and case review tool to either just that information technology (control group) or to those IT tools plus an educational QI intervention. The team will compare composite quality care scores for congestive heart failure and community-acquired pneumonia 2 years following implementation of the QI intervention.

Holmes-Rovner, M., Nelson, W.L., Pignone, M. and others (2007). "Are patient decision aids the best way to improve clinical decision making? Report of the IPDAS Symposium." (AHRQ grant HS 16486). Medical Decision Making 27, pp. 599-608.

This article reports on the International Patient Decision Aid Standards Symposium held in 2006 by the Society for Medical Decision Making. The subject of debate at the symposium was whether patient decision aids are the best way to improve clinical decisionmaking. Among the related topics discussed were the history and philosophy of decision aids (DAs), the problems addressed by DAs, the growth and scope of DAs, and standards setting for DAs. The symposium highlighted several major areas of controversy in DA research: the correct theoretical framework and the gold standard for a good decision, the efficacy of DAs, and the feasibility of DAs in clinical practice. There was consensus that the purpose of DAs is to inform patients about choices and their outcomes. However, the questions of efficacy and feasibility of DAs remained controversial throughout the discussion. At the end of the symposium, participants were split down the middle about whether DAs are the best way to improve clinical decisionmaking.

Reed, P.L., Rosenman, K., Gardiner, J., and others (2007). "Evaluating the Michigan SENSOR surveillance program for work-related asthma."(AHRQ grant HS14206). American Journal of Industrial Medicine 50, pp. 646-656.

Workplace inspections by the Occupational Safety and Health Administration (OSHA) are normally conducted on the basis of employee complaints or the targeting of industry groups. An alternate basis for inspections is the Sentinel Event Notification System for Occupational Risks (SENSOR), which relies on reports of diseases such as workplace-related asthma (WRA). This study compares Michigan worksites inspected in the normal program with those inspected under the SENSOR program as a result of reported WRA cases. From a sample of 12,813 inspections carried out between 1989 and 2002, the researchers found that the occurrence of violations and issuance of penalties was similar under both systems of inspections. However, SENSOR-inspected worksites received fewer citations than non-SENSOR worksites. Also, under both programs of inspection, worksites receiving citations for any violation, as well as worksites cited for serious violations, were more likely to have fewer employees and to lack unions. The researchers suggest that SENSOR inspections are as valuable with respect to public health as inspections undertaken for other reasons.

Rubenstein, L.V., Rayburn, N.R., Keeler, E.B. and others (2007). "Predicting outcomes of primary care patients with major depression: Development of a Depression Prognosis Index." (AHRQ grant HS08349). Psychiatric Services 58(8), pp. 1049-1056.

Mental health conditions, unlike general medical conditions, have generally lacked prognosis indices. The authors of this study developed and tested a Depression Prognosis Index (DPI) to predict 6-month outcomes. They enrolled 1,471 patients with major depression being treated in 108 primary care practices. They identified the following factors as predictors of poor depression outcome: physical and mental comorbidities, a history of depression treatment, low social functioning and support, being older and male, being unemployed, and being a member of a racial or ethnic minority group. The principal outcome measure was depression symptom severity after 6 months. At the outset, patients were ranked in quartiles based on their self-reported characteristics.

At the 6-month followup, 64 percent of those with the poorest prognosis had a likely diagnosis of major depression while only 14 percent of those in the healthiest group had a similar diagnosis. Thus, the ability of the DPI to predict depression outcomes compares favorably with that of prognostic indicators for general medical problems.

Russell, R.B., Green, N.S., Steiner, C.A., and others (July 2007). "Cost of hospitalization for preterm and low birth weight infants in the United States." Pediatrics 120, pp. e1-e9.

According to hospital discharge data from the 2001 National Inpatient Sample from the Healthcare Cost and Utilization Project, 8 percent of infants were born premature or with low birth weights. Their hospital stays accounted for nearly half (47 percent) of infant hospitalization costs in the United States. Hospital stays for premature or low-birthweight (less than 5.5 pounds) infants averaged 12.9 days and cost $15,100 compared with 1.9 days and $600 for uncomplicated births. The higher costs reflected the infants' need for intensive care for acute and chronic conditions, with respiratory illnesses being the most common and costliest. Very premature infants, born at less than 28 weeks of gestation or with birth weights of less than 2.2 pounds, incurred the highest costs, averaging $65,600 and stays of 42.2 days.

The true costs for care of these vulnerable infants are actually higher because physician fees, rehabilitation, outpatient expenses, and the mother's hospital costs were not included in the study's total costs. Both public (42 percent) and private insurers (50 percent) bore the costs of infant hospitalizations.

Reprints (AHRQ Publication No. 07-R066) are available from the AHRQ Publications Clearinghouse.

Shen, J.J., Washington, E.L., Chung, K., and Bell, R. (Spring 2007). "Factors underlying racial disparities in hospital care of congestive heart failure." (AHRQ grant HS13056). Ethnicity & Disease 17, pp. 206-213.

Researchers analyzed data from the 1995-1997 National Inpatient Sample on 373,158 patients discharged with heart failure from U.S. hospitals. Blacks with heart failure were nearly two times more likely and Hispanics were 30 percent more likely than their white counterparts to be admitted to the hospital through the emergency department. Blacks and Hispanics were less likely than their white counterparts to have other coexisting medical conditions. They were also more likely than whites to be admitted to teaching hospitals, which generally have better facilities and capability than nonteaching hospitals. Yet blacks and Hispanics were 34 and 30 percent, respectively, less likely than whites to receive invasive cardiovascular services such as cardiac catheterization, angioplasty, or bypass surgery. These procedures tend to improve outcomes for heart failure patients. Blacks and Hispanics also stayed in the hospital longer and had higher total charges (which is more typical of teaching hospitals).

Singer, S., Meterko, M., Baker, L, and others (2007). "Workforce perceptions of hospital safety culture: Development and validation of the patient safety climate in healthcare organizations survey." (AHRQ grant HS13920). HSR: Health Services Research 42(5), pp. 1999-2021.

This paper describes the development and psychometric evaluation of the Patient Safety Climate in Healthcare Organizations (PSCHO) survey by a Stanford-based patient safety research program. The theory underlying the survey was based on research regarding high-reliability organizations (HROs), whose successful operations require a culture of reliability centering on safety. The 38-item PSCHO survey was distributed in 105 hospitals to over 42,000 individuals. The response rate was 51 percent (21,494 completed surveys).

The survey was divided into three broad factors (senior managers' engagement in patient safety, organizational resources for patient safety, and overall emphasis on patient safety) consisting of nine dimensions: three organizational, two unit-based, three individual and one that relates to report-type questions about the actual incidence of unsafe care. This nine-dimensional model of hospital safety culture was well supported by the empirical results of the survey and may be used in further studies to understand the impact of safety climate on patient safety outcomes.

Smith, S.R., Wahed, A.S., Kelley, S.S., and others (2007 July). "Assessing the validity of self-reported medication adherence in hepatitis C treatment." Annals of Pharmacotherapy 41, pp. 1116-1123.

Strictly following a drug regimen of ribavirin and peginterferon is critical to treat the infection and halt liver damage in individuals infected with hepatitis C virus (HCV). While a high number of patients continue to take the combination medication, despite the flu-like symptoms, hair loss, extreme fatigue, depression, and other side-effects that are common, adherence gradually wanes during therapy.

Patients tended to report higher drug compliance than indicated by electronic monitoring of pill bottle openings, but the discrepancy between the two was slight suggesting that self-report can be used a simple screening tool for nonadherence to HCV treatment. Based on patient self-report, the proportion of patients who were adherent prior to a given visit ranged from 85 to 97 percent for ribavirin and 97 to 100 percent for peginterferon. Adherence based on an electronic monitor placed inside the cap of prescription containers ranged from 69 to 90 percent for ribavirin and 84 to 100 percent for peginterferon. Adherence decreased over time for both medications, probably due to drug side-effects. For combination therapy, adherence was better for weekly injections of peginterferon than for twice daily doses of ribavirin. The findings were based on self-report and electronic monitoring of drug adherence by 196 black and 205 white patients enrolled in an HVC treatment study.

Reprints (AHRQ Publication No. 08-R003) are available from the AHRQ Publications Clearinghouse.

Tan, A., Freeman, J.L., and Freeman, D.H. (2007). "Evaluating health care performance: Strength and limitations of multilevel analysis." (AHRQ grant HS11618). Biometrical Journal 49, pp. 707-718.

Several statistical methods are available to evaluate variations in health care performance: single-level analysis, cluster analysis, and multilevel analysis. Choosing which method to use is complicated by the fact that multi-level analysis, a relatively new and popular method, has not yet been compared to cluster analysis with either a simulated or real dataset. To perform the comparison, the researchers used Medicare claims data in evaluating the extent and source of variation in false-positive rates of screening mammography among radiologists. They found no systematic difference between cluster sampling and multilevel analysis in estimating fixed effects of patient and radiologist characteristics. However, both approaches yield larger standard errors of fixed-effect estimates than single-level analysis. In addition, multilevel analysis identified much less between-radiologist variation and fewer outlier radiologists than the other two methods. The researchers conclude that the choice between cluster sampling and multilevel analysis should be based on the researcher's objectives.

Wackerbarth, S.B., Tarasenko, Y.N., Curtis, L.A., and others (2007). "Using decision tree models to depict primary care physicians CRC screening decision heuristics." (AHRQ grant 84503). Journal of General Internal Medicine 22(10), pp. 1467-1469.

Although guidelines suggest that those without other risk factors be screened for colorectal cancer (CRC) starting at age 50, merely mentioning this is not enough to motivate all patients and there is variation in the nature of recommendations. The researchers interviewed 66 primary care physicians to find out how they formulated recommendations about who should be screened for CRC and what type of tests should be performed. In addition to age 50, inclusion criteria used by physicians were symptoms, cancer history, history of colon problems, and a family history of other cancer. Exclusion criteria were short life expectancy, multiple coexisting illnesses, contraindications to testing, and cognitive impairment. The researchers found that, in making decisions about the timing of screening, physicians used one of four heuristics involving age and, sometimes, family history. In making decisions about the type of screening, they used five heuristics.

Return to Contents
Proceed to Next Article