Skip Navigation U.S. Department of Health and Human Services
Agency for Healthcare Research Quality
CERTs banner

Annual Report

Year 7

CERTs Progress

Herbal Supplement Use and Interactions in Diabetes

Treating and controlling diabetes usually includes diet, exercise, blood glucose testing, and in many cases, using oral or injected medications such as insulin to control hyperglycemia. However, herbal remedies are used by many Mexican Americans in the Southwestern United States, where there is a long history of using plants to treat a variety of ailments. Because of this practice, it is important for physicians to understand the possible adverse effects and potential drug interactions that may occur in persons with diabetes who use herbal remedies.

CERTs investigators examined the use of herbal remedies by Hispanic women with type II diabetes in two community health centers in the southwestern United States.1 A total of 91 percent of the participants interviewed reported using one or more herbal remedies, and 6.7 percent used herbal remedies intended specifically for diabetes.

Researchers found that most of the 77 herbal remedies identified were included as part of commercial preparations and were used by patients to supplement standard medical therapies for diabetes. Although herbal supplements may have adverse effects or interactions with other medications, those found in use in this study were unlikely to have a significant effect on outcomes in diabetes.

Categorizing Antibiotic Use To Understand Antibiotic Resistance

Antibiotic resistance is an important concern. Many studies have sought to identify risk factors for resistant infections. While many of these studies explore the association between antibiotic use and resistance, little attention has been paid to the methods by which prior antibiotic use is defined (e.g., by agent, class, and/or the spectrum of activity against different organisms). To best identify targets for intervention, it is critical to establish whether a resistant pathogen is associated with use of a specific class of antibiotics or whether the specific agents used are associated with certain spectra of activity.

The objective of this project2 was to explore the variability and possible impact of methods of categorizing antibiotic use on the association between prior antibiotic use and resistant infection. Using infections due to extended-spectrum beta-lactamase-producing Escherichia coli and Klebsiella species (ESBL-EK) as a model, researchers conducted a systematic review of the existing literature and reanalysis of the database from a previous study of risk factors for ESBL-EK infection.

The literature revealed tremendous variability across studies in the categorization of prior antibiotic use, with no study justifying its method for categorization. The reanalysis of a past data set also showed great variability across bivariate and multivariate analyses, depending on which antimicrobial use categorization (i.e., class of agent vs. spectrum of activity) was employed.

Identifying modifiable risk factors is paramount in efforts to curb the further emergence of antibiotic resistance. This study is the first to investigate the impact of different methods of antibiotic use categorization on the epidemiological study of resistance.

Future studies assessing risk factors for antibiotic resistance should describe which antibiotics were assessed and how they were categorized. Such information is essential to inform antibiotic use interventions and to facilitate comparison of results across studies.

Patient Adherence to Preventive Cardiovascular Medications

Beta-blockers—Although beta-blockers have been shown to improve survival after myocardial infarction (MI) and reduce the frequency of recurrent MI, little attention has been paid to the need to sustain beta-blocker therapy long term.

In collaboration with the Council for Affordable Quality Healthcare (CAQH), a not-for-profit organization of health plans that promotes collaborative initiatives for administrative simplification and quality, CERTs investigators examined the long-term use of beta-blockers over the first year after MI in patients with commercial health insurance and prescription drug benefits.3 The investigators also looked at adherence in relation to time from discharge, type of health plan, age, sex, and region. In this study, patients were considered adherent if they had prescription claims for beta-blockers for 75 percent or more of the days during the first year after MI.

Over the first year after discharge, only 45 percent of the 17,000 patients adhered to beta-blocker therapy. Lower adherence occurred in patients with Medicare-Plus-Choice plans, patients in the Southeast, and young females in commercial health plans. The study suggests that factors in addition to cost affect long-term adherence.

Beyond targeting physicians who prescribe beta-blockers, educational efforts need to target patients and their families to encourage continued use of beta-blocker therapy during long-term outpatient care. Since this study showed that compliance fell off the most between 30 and 90 days after discharge, interventions to improve adherence need to be initiated early after discharge.

Quality measures that focus on long-term persistence of beta-blocker therapy can help move practice in the right direction. The first example of such a tool is the new 2005 HEDIS® (Healthcare Effectiveness Data and Information Set) measure from the National Committee for Quality Assurance (NCQA), derived by the same methods used in the CAQH study. This HEDIS measure looks at persistence of beta-blocker therapy over the first 6 months after MI. Certainly, the interests of all parties (physicians, patients, health plans, pharmaceutical companies, pharmacists, and nurses) should be well aligned to promote patients' long-term adherence to this life-saving medication.

Secondary prevention therapies—A great deal of research is directed toward understanding and improving in-hospital and discharge prescription practices of doctors, but less goes toward determining long-term adherence to these medications by patients. To address this issue, CERTs researchers analyzed data from the Duke Databank for Cardiovascular Disease from 1995 to 2002 in order to determine adherence to several secondary preventive medications, including beta-blockers, aspirin, angiotensin-converting enzyme (ACE) inhibitors, and lipid-lowering therapies.4

Starting in 1995, a routine followup questionnaire was sent to patients who had undergone a cardiovascular procedure at 6 and 12 months after their procedure and yearly thereafter. Patients were prompted to report all medications they were using. Based on these surveys, patterns of long-term medication use were described. Consistent use was defined as use reported on two or more consecutive surveys up to an endpoint of death, withdrawal, or the end of the study.

From one perspective, the results of this analysis were encouraging, showing that patient-reported use of medications after discharge has increased steadily since the beginning of the study. In 2002, use of aspirin was 83 percent; beta-blockers, 61 percent; and lipid-lowering agents, 63 percent. However, these overall rates of medication use are still suboptimal. In addition, many patients do not use these medications consistently. Aspirin is the oldest and least expensive drug used in secondary prevention and the one with the fewest absolute contraindications and least controversy over its indication for secondary prevention, but only 71 percent of these patients used it consistently. Only 46 percent of patients consistently used prescribed beta-blockers, 44 percent used prescribed lipid-lowering agents, and only 21 percent consistently used a prescribed combination of aspirin, beta-blockers, and lipid-lowering agents. Further, consistent use was lowest among patients who had the greatest potential for benefit from these agents, including the elderly, patients with diabetes, and patients with heart failure. Although quality improvement initiatives have shown modest success in encouraging adherence to medications, much more research is needed to determine the most effective means of bolstering long-term medication adherence.

Physician Oversight of NSAIDs Use

Nonsteroidal anti-inflammatory drugs (NSAIDs) are among the most common medications prescribed to older Americans, and they often are prescribed inappropriately. Adverse events associated with long-term NSAID use are common and often result in substantial illness and death. Physician adherence to guidelines recommending NSAID safety monitoring practices is low.

Researchers conducted a randomized trial to determine whether providing audit/feedback and educational materials to prescribing physicians would improve adherence to three recommendations: laboratory monitoring for blood counts, laboratory monitoring for kidney function, and use of cytoprotective agents (medications to prevent stomach problems).5

There were no significant differences in performance between intervention and control physicians before or after the intervention for any of the three recommendations. However, results suggested a trend toward higher performance among physicians who reported receipt of the intervention materials and participated in the continuing medical education activities. Provider and patient factors and health services use were more strongly associated with adherence to NSAID safety-monitoring guidelines than were the educational materials tested.

The results highlight the challenges of interventions designed to improve adherence to NSAID safety-monitoring guidelines. No single intervention or set of interventions has proven effective in altering physician practice patterns in all settings. Although multimodal interventions have been effective for changing practice in other chronic diseases, no significant difference in NSAID safety-monitoring practices was found in this study.

The clinical and methodological insights gained from this work may lead future studies to greater success in improving physician adherence to evidence-based practice guidelines.

ACE Inhibitor Use and Risks in Early Pregnancy

Use of angiotensin-converting enzyme (ACE) inhibitors during the second and third trimesters of pregnancy has been associated with an increased risk of fetal abnormalities. In contrast, until now, based on several small, uncontrolled studies and unpublished reports, first-trimester use of ACE inhibitors had not been linked to adverse fetal outcomes. However, because angiotensin II receptors are widely expressed in fetal tissue and could play an important role in fetal development, it was postulated that first-trimester exposure to ACE inhibitors could increase the risk of congenital malformations.

Using a large Medicaid database, investigators assessed the association between exposure to ACE inhibitors during the first trimester of pregnancy and the risk of congenital malformations.6

Infants with exposure to ACE inhibitors solely in the first trimester had an increased risk of major congenital malformations compared with infants who had no exposure to antihypertensive medications. Use of other antihypertensive medications did not confer an increased risk. Infants exposed to ACE inhibitors were at increased risk of malformations of the cardiovascular system and the central nervous system.

As the approved uses for ACE inhibitors have expanded, their use among women of childbearing age has increased. This increase in use is likely to result in an increase in first-trimester fetal exposures. These data suggest that such exposures cannot be considered safe and should be avoided.

Continuous Glucose Monitoring Devices and Their Effectiveness in Children

Controlling blood glucose levels delays the long-term complications from type 1 diabetes mellitus but increases the risk for hypoglycemia. The continuous glucose-monitoring system (CGMS) gives blood glucose readings every 5 minutes and is very accurate and reliable in adults. However, there is not much information available about how effective this system is in children.

CERTs researchers conducted a study to determine whether the CGMS would improve blood glucose control in 27 children with type 1 diabetes mellitus.7 Children were assigned to one of two groups. Both groups wore the CGMS for 72-hour periods. Therapy was adjusted for the intervention group based on the CGMS and self-monitoring data. Therapy for the control group was based only on self-monitoring data. Hemoglobin A1c levels were determined and the change from 0 to 6 months for the two groups was compared. At the beginning of the study, hemoglobin A1c levels were similar for the two groups, but by the end of the study, they were significantly lower for the intervention group. The decrease in the intervention group was statistically significant, but the decrease in the control group was not.

The CGMS may improve blood glucose control in children with type 1 diabetes mellitus without increasing the risk for hypoglycemia. However, the long-term benefits of the CGMS on metabolic control remain to be determined.

Vaccine Coverage – Unexpected Benefits and Challenges

Reduced pneumococcal bacteremia in adults when children in the home are vaccinated—A new pneumococcal conjugate vaccine containing seven serotypes was released in February 2000 for use in infants and children. In the following years, several observations were published linking the vaccine to a reduced rate of invasive pneumococcal disease among adults. CERTs investigators sought to determine if pediatric vaccination lowers the risk of pneumococcal bacteremia in adults living in the same home.8

The study had two parts: a 2-year prospective population-based surveillance for pneumococcal bacteremia in adults and a case-control study of risk factors for the disease. The study was carried out in a five-county region of Pennsylvania. The surveillance tracked all cases of pneumococcal bacteremia, while the case-control study involved contacting and interviewing both adults with the disease and a healthy control population.

Although the overall rate of pneumococcal bacteremia for the adults in the study decreased by a statistically insignificant 9 percent from 2002 to 2004, there was a significant 22-percent decrease in the cases that involved serotypes included in the new vaccine. The results of this study showed that vaccination of the youngest child in the home corresponded to an 80-percent reduction in the odds of the disease among adults in the same home.

Examining surgeons' attitudes about hepatitis B vaccination—In the 1990s, prior to the Occupational Safety and Health Administration (OSHA) mandate to offer hepatitis B vaccination to all health care workers, studies showed that many surgeons were not adequately vaccinated. It is unknown if vaccination rates have improved since the OSHA mandate. Inadequately vaccinated surgeons may be a risk to their patients.

CERTs investigators sought to identify the proportion of U.S. transplant surgeons adequately vaccinated against the hepatitis B virus, identify characteristics associated with inadequate vaccination, and assess the proportion that had been evaluated for immunization after potential exposures to the hepatitis B virus.9

Of the 619 eligible surgeons, 347 (56.1 percent) returned completed questionnaires. A total of 23 percent had received fewer than the recommended three injections. Some factors associated with inadequate vaccination were increasing length of time a surgeon was in clinical practice, an increased fear of infection from patients, and a lack of recent testing for the hepatitis C virus. Of the 94 surgeons (27.3 percent) reporting at least one needle-stick exposure while operating on a patient with the hepatitis B virus, 14 (14.9 percent) were inadequately vaccinated; of these 14, only 5 (35.7 percent) sought appropriate testing and counseling for active immunization. Surgeons underestimated both the risk of exposure while operating and the risk of becoming infected if exposed.

Despite the mandate from OSHA that all health care workers be offered hepatitis B virus vaccination, some transplant surgeons are inadequately vaccinated and do not seek appropriate post-exposure evaluation. This increases risks to both the surgeons and their patients.

Determining the reasons for inadequate vaccination may help target interventions that promote vaccination and a change in behavior. Because inadequately vaccinated surgeons underestimate their risk of exposure, although they also express fear of acquiring hepatitis B, education regarding true transmission risks may increase the number willing to be vaccinated.

Interventions to Change Physician Behavior

Changing prescribing for heart failure—Despite the incorporation of beta-blocker use into clinical practice guidelines for the treatment of heart failure, use of this proven therapy remains low. This may result from previously held beliefs that beta-blockers are harmful in heart failure and from the typically slow translation of evidence into practice.

A study was conducted to develop and test a multifaceted intervention using Internet-based academic detailing, physician feedback, and patient education to improve outpatient use of beta-blockers in patients with heart failure.10

Practices receiving intervention did not have a significantly higher use of beta-blockers than a control group of medical practices given written educational materials. However, a trend toward greater use of beta-blockers in the intervention group, coupled with results of prespecified secondary analyses demonstrating significantly more patients starting a beta-blocker than stopping a beta-blocker in intervention practices, put the findings in a more encouraging light.

The survival benefit associated with beta-blockers in patients with heart failure is large. Therefore, interventions with even modest increases in beta-blocker use could result in substantial benefits.

Investigators should continue to explore new types of interventions, physician incentives, and technological advances to stimulate adherence to guidelines for recommended therapies for heart failure.

Testing for tuberculosis before starting infliximab—After infliximab was approved by the Food and Drug Administration (FDA) to treat Crohn's disease (1998) and rheumatoid arthritis (1999), tuberculosis was reported in the FDA's spontaneous reports as one potential infliximab-related adverse event.

Researchers conducted a study to examine the impact of Federal, industry, and academic efforts to communicate this risk to users of infliximab.11 The effectiveness was evaluated by the proportion of patients getting tuberculin skin tests, which assess whether patients are harboring latent tuberculin infection before or soon after taking infliximab. The investigators also wanted to determine if there was a difference in tuberculin skin testing rates among the various medical specialists who treated patients requiring infliximab.

Tuberculin skin testing rates doubled over a 30-month period as the risk communication efforts continued. Rheumatologists were the only specialists who showed a significantly higher than average rate of tuberculin skin testing among their patients.

The study showed that risk-communication efforts can increase physician compliance in tuberculosis testing of their patients, and, in this case, might decrease the risk of activating latent tuberculosis infection potentially associated with infliximab.

The findings of this study suggest opportunities for further research to determine what types of communications elicit the greatest compliance, whether multiple communications are more effective, and whether repetition of messages further increases compliance, thus optimizing public health.

Patient Self-Reporting of Medication Use

NSAIDs use—Each year an estimated 16,000 people die from NSAID-induced gastrointestinal complications. Patients frequently fail to report to their doctors that they are taking over-the-counter NSAIDs, and serious problems may occur when nonprescription NSAIDs are used alone or in combination with other medications.

CERTs investigators sought to determine if self-reported use of nonprescription and prescription NSAIDs matched the information in the patient's medical records, and to determine which of two survey formats aided patients' recognition and recall of medication use.12

Patients in a rheumatology clinic completed 361 surveys: 182 in list format and 179 in open-ended format. Medical records review confirmed 26.1 percent of patient-reported nonprescription NSAID use and 59.1 percent of patient-reported prescription NSAID use. For patients using both prescription and nonprescription NSAIDs, both medications were confirmed in only 2.3 percent of the medical records. Fewer patients reported nonprescription NSAID use with the open-ended format survey (24.0 percent) than with a checklist format (41.8 percent).

Agreement between self-reported NSAID use and medical record documentation is especially poor for nonprescription NSAIDs. This lack of communication and understanding about concomitant prescription and nonprescription medication use may lead to serious adverse reactions, including death.

Although this study did not test an intervention, it suggests that using a checklist for obtaining history of medication use might increase patient-physician communication regarding NSAID use. Addressing NSAID use at the point of purchase (pharmacies) using a list-type survey may offer a unique opportunity to decrease risky concomitant prescription and nonprescription NSAID use.

Osteoporosis medications—Pharmacy and linked claims databases are used to determine whether patients receive medications as a quality-of-care measure. However, these databases have not been compared with self-reported data for patients receiving medications used to treat glucocorticoid-induced osteoporosis.

CERTs researchers examined agreement between filled prescriptions in a pharmacy database and self-reported current use of several osteoporosis medications among chronic glucocorticoid users enrolled in a large U.S. managed care plan.13 Chronic glucocorticoid use was defined as 60 or more days of filled outpatient oral glucocorticoid prescriptions from July 2001 to December 2002.

Agreement between the pharmacy data and the self-reported data was high, and underreporting of current osteoporosis medication was uncommon among chronic glucocorticoid users enrolled in managed care. Both sources of information seem to be appropriate in assessing exposure to osteoporosis medications.

CERTs investigators found that use of pharmacy data alone is unlikely to lead to an underestimate of osteoporosis treatment, but there is a need for a longer assessment period in studies that rely only on pharmacy data.

Return to Contents
Proceed to Next Section



AHRQAdvancing Excellence in Health Care