search
ghc.org |
|
|
| site map
search
ghc.org |
|
Feature ArticleThrough CERT, Group Health contributes to powerful studies in patient safetyBy Katie Saunders Medications are often powerful weapons against disease—think depression, hypertension, and AIDS, to name just a few. Precisely because of this inherent strength, a delicate balance often exists between a drug’s ability to do good or to cause harm. The scales can tip precariously if any of the following are absent:
Given the immense variety and complexity of medications available today, it’s not surprising that lapses occur in all of these areas. Often, there’s simply not enough information to answer critical questions about a drug’s safety. For example, the effects of many types of medications on pregnant women, or on children, are unknown. And when information about drug safety is available, it’s not always judiciously applied, whether due to poor dissemination of information, system problems, or medical mistakes. To increase our understanding of pharmacoepidemiology—the study of the use and effect of drugs among large numbers of people—and to improve patient safety, the Agency for Healthcare Research and Quality (AHRQ) set up the Centers for Education and Research on Therapeutics (CERT) in 1999. The purpose of the CERTs is to advance better use of drugs, medical devices, and biological products. Seven CERTs exist today, including the "HMO CERT" which is comprised of 10 HMOs, including Group Health. Robert L. Davis, MD, MPH, a senior investigator at Group Health Center for Health Studies, is the Group Health principal investigator. Davis describes the impetus for the CERT as follows: "From cradle to grave, we’ve had the best health care system in the world, but its application was falling far short of what it should be." Davis believes that the HMO CERT has several unique advantages over a single health plan in conducting research to improve the safe and effective use of medications:
These advantages were put to good use in one of the first studies undertaken by the HMO CERT—a descriptive assessment of antibiotic prescribing patterns among children in the late 1990s. Davis and colleagues tracked antibiotic prescriptions among 225,000 children in nine health plans between 1996 and 2000. This study was conducted amidst a background of growing concern that over-prescribing of antibiotics in children was resulting in increased resistance to antibiotics—not to mention elevated health care costs. The researchers found that antibiotic prescribing during this time period had actually decreased, concurrent with a reduction in diagnoses of otitis media, a bacterial ear infection that is a major target of antibiotics. The scientists speculated that due to increased awareness on the part of parents and physicians, the threshold for diagnosing otitis media was raised, resulting in fewer antibiotic prescriptions. Since up to half of otitis media cases resolve in one day, Davis supports the "wait-and-see" approach, saying, "By simply waiting one day, we can perhaps reduce antibiotic use for otitis media by 50 percent." Davis and colleague Heather McPhillips at the University of Washington are heading up a series of CERT studies examining prescribing among children. Their primary focus is "off-label" prescribing—that is, drugs are prescribed in a population, or for a disease, for which they have not been specifically approved, or "labeled," by the FDA. For example, antidepressants are becoming much more commonly used for children. This might be a good thing, Davis points out, but the necessary studies that would allow proper labeling, and thus, standardized use of these medications are for the most part non-existent. "Off-label prescribing is a common problem in pediatrics," says Davis, adding that it has large implications for patient safety. In another study, Davis and McPhillips are looking at prescription-writing errors, such as over- and under-dosing, in the outpatient setting. The initial findings suggest that mistakes—some of them substantial—are more common than the researchers expected, and are in line with studies of prescription-writing errors involving hospitalized children. Other "vulnerable" populations for whom drug safety is largely an unknown are pregnant women and their offspring. Among a cohort of 130,000 pregnant women in nine HMOs, CERT researchers found that over half were taking at least one drug other than vitamins during pregnancy. While some drugs are specifically contraindicated during pregnancy, the safety of others—to both the mother and child—is largely a mystery. Davis and colleagues are examining the association of about 30 different drugs, including antidepressants, antihypertensives, and antibiotics, with short-term (up to one month after birth) infant outcomes, such as jaundice and hypoglycemia, and longer-term serious congenital malformations. In a yet-to-be funded upcoming CERT study, Davis hopes to collaborate with CHS Research Associate Denise Boudreau, PhD, to assess the long-term health outcomes of patients with psoriasis, a serious skin disease. These patients may be at a higher risk for developing certain cancers, and it is unknown whether the cause is psoriasis itself, or the medications used to treat the disease. While all studies described above contain elements related to patient safety, the CERT Prescribing Safety Program deals with a specific aspect of safety—prescribing errors. HMO CERT researchers are looking at the prevalence of four types of errors: 1) prescribing at variance with Food and Drug Administration (FDA) "black box" warnings (see below); 2) inappropriate concurrent use of drugs that interact with each other; 3) inadequate adjustment of dosages for renal (kidney) insufficiency; and 4) inadequate monitoring for toxicity (e.g., liver function tests when using statins). Davis explains that FDA "black box" warnings are mailed to physicians cautioning them against prescribing a certain drug unless they definitely feel it will benefit their particular patient. In essence, these are drugs that have almost been withdrawn from the market due to safety concerns, and that should be prescribed sparingly, if at all. Studies have shown repeatedly that the mailed alerts have no effect on physician prescribing behavior. The CERT scientists will estimate the rates at which prescriptions are at variance with these warnings. Proper dosing for patients with renal insufficiency poses another challenge for providers. When the kidney cannot efficiently remove waste and extra fluids from the body, drug overdosing is a very real danger. Thus, doses of certain medications should be adjusted downward for patients with renal insufficiency. This does not always happen, according to Davis: "In some of our preliminary data, we are seeing that up to one of two people who should have their drugs renally dosed are not getting them renally dosed. This is not one in 10,000—this is one in two." Davis views this situation as a perfect application for better use of information technology. The goal of the CERT is not only to document prescribing errors, but also to develop a technology infrastructure to help solve the problem. This type of system-wide fix is in keeping with the philosophy of a high-profile report on medical errors, To Err is Human: Building a Safer Health System, published in 1999 by the Institute of Medicine. While the report contained sobering statistics on the number of medical errors that occur in the United States, it was quick to clarify where the blame belonged: "People working in health care are among the most educated and dedicated workforce in any industry. The problem is not bad people; the problem is that the system needs to be made safer." So, Davis posits, in the case of adjusting doses for renal insufficiency, an automated system could flag a patient as being renally insufficient and a drug as requiring renal dosing, resulting in an alert when the physician enters a prescription. Group Health is in the process of implementing an automated medical record, EpicCare, wherein such alerts would be feasible. Such systems also feature computerized physician order entry, which allows the physician to complete a prescription online and send it electronically to the pharmacist, thereby circumventing the error-prone method of handwritten prescriptions delivered manually. Davis views the development of a technology infrastructure as one of the key elements to ensuring prescribing safety, but he says, "I think it’s going to be a long road." "The initial experiences in the CERT have shown that when you change a system to fix one failure, that it has all these unintended consequences," he explains. "You almost have to do the studies to find out what the problems are." Davis recalls a study where physicians used a laser pen to electronically enter prescriptions. While the drug and dosage tended to be correct, the prescription was sometimes written for the wrong person because the pen selected a name so fast that the physician did not notice when s/he chose an incorrect name. Despite the complexity, Davis is optimistic about information technology. "If we don’t start now, we’ll never get there," he says. "And it’s fun, being in on the ground floor." Davis looks forward to collaborating on some of these issues with the Center’s new informatics investigator, James Ralston, MD, and Group Health staff, including Ted Eytan, MD, associate medical director of Clinical Informatics, and James Hereford, executive director, Health Informatics. Davis is also leading an effort to improve the rapid detection of adverse events associated with newly marketed drugs. Even though all drugs are required to go through pre-marketing approval procedures, these efforts are often inadequate to certify a drug’s safety. Pre-marketing studies are usually conducted among small numbers (1,000 to 2,000) of homogenous patients who have little underlying disease or simultaneous drug use. These conditions cause pre-marketing safety studies to miss adverse events that are either very rare or that only manifest themselves under conditions of actual use. Thus, monitoring of most drugs continues after they’re approved by the FDA. With its millions of members, the HMO CERT is an ideal, "real-world" setting to conduct such post-marketing drug surveillance. For example, the HMO CERT is currently monitoring Alosetron, a drug for irritable bowel syndrome that was withdrawn and reintroduced into the market. In the future, Davis expects the FDA to rely more and more on the HMO CERT to do studies such as these. To maximize the HMO CERT’s ability to perform this function, Davis is attempting to shorten the amount of time between introduction of a new drug and identification of associated adverse events. He is building on a "signal detection" method he developed while working on research through the Centers for Disease Control and Prevention’s Vaccine Safety Datalink (VSD) project. Like the CERT, VSD is a multi-HMO endeavor studying issues related to vaccine safety and effectiveness. Using five years of historical VSD data, Davis and colleagues assessed whether they could detect an increased risk of intussusception (a problem with the intestine or bowel) following a rotavirus vaccine. (This was a known adverse event of the vaccine). To do this, the researchers created weekly cohorts of children who had, and had not, been exposed to the rotavirus vaccine, and the diagnoses they received in the 30 days following. They then calculated weekly outcome rates. Using these retrospective data, the researchers detected an increased risk of intussusception about 17 weeks after the vaccine had been introduced. "We looked for known signals that we knew we should be able to find and we found them," says Davis. Based on these results, the VSD is implementing a real-time signal detection method. Davis hopes to develop a similar system for medications in a CERT-sponsored demonstration project based on retrospective data. If successful, the demonstration could lead to a collaborative project with the FDA to set up prospective monitoring systems to assess adverse events as often as every three months. Davis is gratified to be part of the CERT. He enjoys the multi-site collaboration, which allows him to share expertise with a broad range of researchers. He also relishes the opportunity to conduct studies that would otherwise be impossible, citing the cohort of 130,000 pregnant women. He explains that medication safety has been an interest of his since medical school, when scientists first made the connection between birth control pills and massive pulmonary embolisms. "I still remember wondering, ‘How did this slip through the system?’ Now, 25 years later, I realize how things like that happen and why we need to develop sophisticated systems to prevent them as quickly as possible." |
|
||||
| site map
search
ghc.org |
| Copyright 2008 Group Health Cooperative. Revised: June 03, 2008. Contact Us |
