 To assess the cost-effectiveness of laparoscopic adjustable gastric banding (LAGB) and laparoscopic Roux-en-Y gastric bypass (LRYGB) as treatment for morbid obesity.

A Markov model was developed to simulate weight loss, health consequences, and costs for surgical treatment of morbid obesity. The model was used to estimate incremental cost-effectiveness ratios (ICERs) in terms of cost per quality-adjusted life-year (QALY) gained.

Estimates of procedure effectiveness were derived from published results of a head-to-head randomized controlled trial. Other model parameters, including complication rates, costs of treatment, adverse events and obesity, mortality rates, and utilities, were estimated from published literature and publicly available databases. Costs (2006 US dollars) and QALYs were discounted by 3% per annum.

Under conservative assumptions, both LAGB and LRYGB improved health outcomes, at a higher cost, compared with no treatment. ICERs for both LAGB and LRYGB versus no treatment were below $25,000 per QALY gained. ICERs were lower for individuals with higher initial body mass index and higher for older individuals. ICERs for men were generally higher than those of women. Sensitivity analyses showed these results to be robust to reasonable variation in model parameters and overall parameter uncertainty. Base-case ICERs for LRYGB versus LAGB were below $25,000 per QALY gained, but were highly sensitive to model assumptions.

Both LAGB and LRYGB provide significant weight loss and are cost-effective compared with no treatment at conventionally accepted thresholds for medical interventions.

	The majority of bariatric procedures in the United States are laparoscopic Roux-en-Y gastric bypass (LRYGB) and laparoscopic adjustable gastric banding (LAGB).

		Both LAGB and LRYGB lead to better healthcare outcomes and higher costs compared with no treatment. In the main analysis, LAGB results in less weight loss than LRYGB, but is associated with lower costs.

		Both LAGB and LRYGB are cost-effective alternatives to no treatment, with incremental cost-effectiveness ratios below $25,000 per quality-adjusted life-year gained; results are robust.

		Estimates of the cost-effectiveness of LAGB versus LRYGB are sensitive to assumptions about maintenance of weight loss in the medium to long term; long-term comparative studies of the 2 procedures are needed.

	Obesity, or excessive body fat, is a major risk factor for a host of detrimental medical conditions including type 2 diabetes, hypertension, dyslipidemia, osteoarthritis, and obstructive sleep apnea.1 Approximately 5.7% of American adults (>12 million people) are morbidly obese,2 with a body mass index (BMI) of >40 kg/m

	There are 3 established approaches to treating morbid obesity: lifestyle modification, pharmacologic therapy, and surgery. The first 2 approaches have been shown to have a limited impact on long-term weight control.

 The third option, bariatric surgery, has been shown to be an effective long-term treatment for morbid obesity, with mean excess weight loss exceeding 60%.

 However, the risks of complications with bariatric surgery are not trivial.

	The majority of bariatric surgeries currently performed in the United States are laparoscopic Roux-en-Y gastric bypass (LRYGB) procedures. An alternative to gastric bypass is laparoscopic adjustable gastric banding (LAGB), in which a silicone elastomer ring is placed around the upper part of the stomach, creating a new small stomach pouch so the food storage area in the stomach is reduced. Excess weight loss with LAGB, although significant, may be lower than that seen with LRYGB.

 However, LAGB, unlike LRYGB, does not involve stapling or partition of the stomach or intestine; thus, postoperative complication rates are lower, hospitalizations are shorter, and there is less risk of micronutrient deficiency compared with LRYGB.

 Moreover, the procedure is fully reversible. The objective of this study is to assess the cost-effectiveness of LAGB versus that of LRYGB, or no treatment, for patients with morbid obesity.

	We constructed and estimated a Markov model of treatment and outcomes for obese adults. The model focused on decreases in BMI and treatment-related adverse events as predictors of lifetime cost, health-related quality of life, and survival. Patients received initial treatment with LAGB or LRYGB, or received no treatment. After the initial treatment period, patients entered the Markov model in 1 of the 5 health states corresponding to their initial BMI status: not obese (BMI <30 kg/m

). The sixth health state in the model was death. Health states were associated with specific treatment costs, levels of “other” health expenditures, health-related quality of life, and mortality risks. Patients transition between Markov states by changing BMI status (losing or gaining weight) or dying. For the first 10 years of the model, surgical patients remained at risk of transient treatment-related adverse events that may be mild, moderate, or severe.

	We estimated model probabilities, costs, and utilities from a variety of published and publicly available sources, supplemented by input from a clinical expert in minimally invasive (laparoscopic) bariatric surgery. 

 reports model parameters and data sources.

The population for this analysis was limited to adults age 18 to 74 years who satisfied the clinical eligibility criteria for bariatric surgery: BMI >40 kg/m

Probabilities of treatment-related weight loss were estimated from a prospective study by Angrisani et al, in which morbidly obese patients were randomized to LAGB and LRYGB and followed for 5 years postsurgery.

 In base-case analyses, it was assumed that BMI remained constant after the 5-year follow-up period. Patients receiving no treatment were assumed to maintain a constant BMI for the duration of the model. The study by Angrisani et al was judged to be the strongest available based on its randomized head-to-head design and minimal attrition over the 5-year study period. The study limitations include a small initial sample (N = 51) and a European rather than US setting. For this reason, we also estimated the model using an alternative source for 5-year efficacy data: a systematic review of 36 English-language studies of weight loss outcomes following LAGB or LRYGB by O’Brien et al.

 Limitations of the O’Brien et al study include a preponderance of single-armed rather than comparative studies, and attrition rates that may have been underreported. Compared with the Angrisani et al study, O’Brien et al reported greater effectiveness for LAGB and shorter durability of weight loss for LRYGB (with weight regain beginning in year 3), as shown in 

	The calculation of transition probabilities between BMI categories also is described in 

Probabilities of Treatment-Related Adverse Events.

 The estimated incidences of adverse events were derived from a comparative study of the complications of LAGB and LRYGB by Parikh et al.

, defined as organ repairs such as resection of the bowel; 

, defined as operations not requiring organ repair (eg, laparoscopic repositioning of a gastric band); and 

, defined as drug therapy only (eg, antibiotic therapy for an infection). Parikh et al reported separate estimates for the probabilities of major reoperations and all adverse events for the first 30 days (early) and after 30 days (late). We estimated the early and late probabilities of moderate reoperations and medical treatments based on the relative proportions of total early and total late complications from that study. The average period of follow-up was approximately 12 months; we assumed that the yearly probability of surgical adverse events remained constant for the first 4 years postsurgery, was halved for years 5 to 10, and was zero after 10 years. Consistent with the literature, patients receiving no treatment were assumed to have no serious adverse events.

Patients undergoing LAGB or LRYGB have a 1-time probability of postoperative death from surgery in the first 30 days.

 BMI-specific mortality rates were estimated based on age- and sex-specific all-cause mortality rates derived from US life tables,

 to which we applied BMI-specific relative risk ratios from an analysis of the National Health and Nutrition Examination Survey by Flegal et al.

 Because Flegal et al reported a single aggregated mortality risk ratio for BMI >35 kg/m

, we estimated separate mortality risk ratios for the morbidly obese I, morbidly obese II, and super obese groups using linear extrapolation based on the trends at lower levels of BMI and the relative prevalence of morbid obesity categories reported by Sturm: morbidly obese I (67%), morbidly obese II (29%), and super obese (4%).

The physician costs for the LAGB and LRYGB procedures were estimated using 

codes: 43770, 90772, 90801, 97802, and 99214 for LAGB; and 43644, 90801, 97802, and 99214 for LRYGB.

 Associated inpatient costs were estimated using data from the Healthcare Cost and Utilization Project (HCUP) Database with 

International Classification of Diseases, Ninth Revision 

codes: 44.95 for LAGB, and 44.38 for LRYGB.

 The hospital charges were converted to resource costs using hospital cost-to-charge ratios published in the HCUP Database. The default schedule of follow-up visits was based on treatment guidelines of the American Society for Metabolic and Bariatric Surgery, supplemented by input from our clinical expert as: 9 visits in year 1, 2 visits in year 2, and 1 annual visit thereafter for LAGB; and 6 visits in year 1, 2 visits in year 2, and 1 annual visit thereafter for LRYGB.

	In consultation with our clinical expert, the costs of LAGB- and LRYGB-related adverse events for each level of severity were estimated using the Physician Fee Guide (physician costs),

 Within each severity category, a weighted average cost was derived, based on the relative prevalence of such complications as reported by Jan et al,

 supplemented by additional input from the clinical expert. The cost of a postoperative fatality was assumed to be equal to that of an early major reoperation. A list of the complications that were used to generate the weighted average costs is available from the authors on request.

	"Other" medical expenditures specific to each BMI level were calculated based on a study of the Medical Expenditure Panel Survey by Arterburn et al.

 Because this study reported aggregated expenditures for the BMI >40-kg/m

 group as a whole, we estimated the other medical costs for the morbidly obese II and super obese groups using linear extrapolation as described above.

 Where necessary, costs were updated to 2006 US dollars using the Medical Care Services component of the Consumer Price Index.

Estimates of BMI-specific utilities were derived from a regression analysis of EQ-5D data from the 2000 Medical Expenditure Panel Survey by Jia and Lubetkin,

 rescaled to replicate the US mean population EQ-5D score of 0.868 as reported by Sullivan and Ghushchyan.

 Jia and Lubetkin reported mean utilities for individuals with BMI >35 kg/m

 as a single group. We used linear extrapolation to estimate separate utilities for the morbidly obese I, morbidly obese II, and super obese groups, as described above.

	No estimates of bariatric-surgery—specific utility decrements were identified in the literature. Estimated utility decrements for the LAGB and LRYGB procedures therefore were derived from utilities reported for laparoscopic surgery for hernia repair in the United Kingdom,

 The utility decrement was applied at the time of surgery for 4 weeks for LAGB and 6 weeks for LRYGB, to reflect the slightly longer recovery time associated with LRYGB compared with LAGB. The utility decrement for a major reoperation was estimated as the decrement associated with a hysterectomy

 and was applied for 6 weeks. The decrement for a moderate reoperation was estimated as the decrement associated with laparoscopic hernia repair

 and was applied for 4 weeks. The decrement for medical treatment was estimated as the mean of the decrements associated with venous thrombosis, gastroesophageal reflux disease, ulcer, and gastritis as reported by Sullivan and Ghushchyan

	Cumulative costs associated with morbid obesity and its treatment, survival in life-years, and quality-adjusted life-years (QALY s) were estimated for each obesity treatment strategy. Cumulative costs include costs of initial treatment, follow-up, and treatment-related adverse events minus cost offsets generated by decreases in other health expenditures as patients lose weight. The treatments were ranked in order of increasing costs and the incremental cost-effectiveness ratios (ICERs) were calculated by comparing the more costly with the less costly, dividing the incremental cost by the incremental benefit. The base-case analyses were stratified by sex and initial BMI category, assuming an initial age of 40 years. Results also were aggregated to the population level using a patient distribution consistent with the patient sample in the LAGB arm of the Angrisani trial (80% female; initial BMI: 23% morbidly obese I, 66% morbidly obese II, and 11% super obese).9 Cost-effectiveness analyses were performed from the perspective of a third-party payer over a lifetime horizon. All costs, life-years, and QALY s were discounted at an annual rate of 3%. Base-case analyses were performed using treatment efficacy parameters derived from the Angrisani et al study.

Alternate Scenarios and Sensitivity Analyses

	We assessed the effect of uncertainty in comparative weight loss estimates by reestimating the model with an alternative source for efficacy estimates.10 In a second alternative analysis, patients were assumed to regain half of the cumulative BMI lost after 5 years over the following 5-year period. Multiple 1-way sensitivity analyses then were conducted to test the robustness of model results to changes in key parameters, particularly efficacy, rates of adverse events, and costs. Additional 2-way sensitivity analyses are discussed in 

	A second-order probabilistic Monte Carlo sensitivity analysis was conducted to assess the sensitivity of study findings to overall parameter uncertainty, particularly with regard to treatment efficacy, rates of complications, costs, and utilities. We constructed a probability distribution for the estimates of treatment efficacy by bootstrapping the simulated LAGB and LRYGB cohorts—repeatedly sampling with replacement from each cohort of patients. Uncertainty in other key model parameters was characterized by probability distributions with means equal to the default estimates. Probabilities of adverse events were assumed to follow a beta (β) distribution, with distributional parameters reflecting the underlying trial samples. Estimates of treatment costs were assumed to follow a gamma (g) distribution with standard errors assessed, subjectively, as 10% of the mean value. The parameters and distributions of the variables included in the probabilistic analysis are summarized in 

	Base-case results from the analysis using efficacy estimates derived from the article by Angrisani et al9 are reported in 

. Results of analyses are presented weighted over the patient population, and stratified by sex and initial BMI. Neither LAGB nor LRYGB is cost saving compared with no treatment; however, both LAGB and LRYGB are more beneficial than no treatment, providing an additional 2.04 (LAGB) and 2.9 (LRYGB) discounted lifetime QALY s for each treated patient. The ICERs for LAGB versus no treatment and for LRYGB versus no treatment are similar and relatively low (<$10,000 per QALY gained). Although LAGB costs an average of $5324 less than LRYGB in the aggregate population, it also is less effective, providing 0.42 fewer life-years saved and 0.86 fewer QALY s; the ICERs for LRYGB versus LAGB across patient groups lie below $15,000 per QALY gained. As the stratified results show, bariatric surgery is most cost-effective for females and for patients who have higher initial BMI. In addition, bariatric surgery is more cost-effective for younger patients (not shown).

Results for Alternate Scenarios and Sensitivity Analyses

	For comparison, results of the analysis using the O'Brien et al data

 also are presented in Table 2. Under these alternative efficacy assumptions, both procedures remain cost-effective compared with no treatment, with ICERs less than $20,000 per QALY gained; however ICERs for LRYGB versus LAGB are significantly greater than those estimated using efficacy estimates from Angrisani et al.

reports the results of a series of 1-way sensitivity analyses for the default analysis based on efficacy estimates from Angrisani et al.

 We report the results for analyses for which the range in the relevant ICERs exceeds $1000. In all instances, the magnitude of the variation in the aggregate ICER is less than $10,000 per QALY gained, and the upper limits of our estimates of incremental cost-effectiveness for LAGB and LRYGB versus no treatment lie below $25,000 per QALY . The model was not sensitive to small changes in efficacy over the initial 10-year period. All ICERs for LRYGB versus LAGB fell below $15,000 per QALY .

	However, in an analysis in which patients receiving LRYGB were assumed to partially regain weight over years 5 to 10 (as suggested by the O'Brien et al

 analysis), LAGB dominated LRYGB when using estimates of efficacy based on either Angrisani et al

 or O’Brien et al (not shown). When weight regain was modeled for patients receiving either LAGB or LRYGB, the ICER for LRYGB versus LAGB was $19,700 using estimates derived from Angrisani et al; however, LRYGB was dominated by LAGB using estimates derived from O'Brien et al.

Articles

WEB EXCLUSIVE

Laparoscopic adjustable gastric banding and gastric bypass are cost-effective treatments for morbid obesity compared with no treatment.

Cost-Effectiveness of Laparoscopic Gastric Banding and Bypass for Morbid Obesity

 To develop guidelines to monitor high-risk medications and to assess the prevalence of laboratory testing for these medications among a multispecialty group practice.

We developed guidelines for the laboratory monitoring of high-risk medications as part of a patient safety intervention trial. An advisory committee of national experts and local leaders used a 2-round Internet-based Delphi process to select guideline medications based on the importance of monitoring for efficacy, safety, and drug—drug interactions. Test frequency recommendations were developed by academic pharmacists based on a literature review and local interdisciplinary consensus. To estimate the potential effect of the planned intervention, we determined the prevalence of high-risk drug dispensings and laboratory testing for guideline medications between January 1, 2008, and July 31, 2008.

Consensus on medications to include in the guidelines was achieved in 2 rounds. Final guidelines included 35 drugs or drug classes and 61 laboratory tests. The prevalence of monitoring ranged from less than 50.0% to greater than 90.0%, with infrequently prescribed drugs having a lower prevalence of recommended testing (P <.001 for new dispensings and P <.01 for chronic dispensings, nonparametric test for trend). When more than 1 test was recommended for a selected medication, monitoring within a medication sometimes differed by greater than 50.0%.

Even among drugs for which there is general consensus that laboratory monitoring is important, the prevalence of monitoring is highly variable. Furthermore, infrequently prescribed medications are at higher risk for poor monitoring.

This article adds to the existing literature and informs clinical decisions in the following ways:

By describing a process to develop guidelines for the laboratory monitoring of high-risk medications within a multispecialty group practice. 

By estimating the potential effect of an intervention to improve laboratory testing of high-risk medications by reporting the use frequency of high-risk medications requiring monitoring in 2008 and the prevalence of test completion for high-risk drugs in 2008. 

By demonstrating that infrequently used medications have lower rates of recommended laboratory test monitoring.

Drug-induced injury is common. Independent risk factors for adverse drug events in the ambulatory setting include advanced age, multiple comorbid conditions, and the use of high-risk medications requiring close monitoring. For example, failure to appropriately monitor older patients receiving drug therapy accounts for 36% of preventable adverse drug events in the ambulatory setting.

Efforts to improve monitoring within organizations are hampered by a lack of comprehensive and specific guidelines for the laboratory monitoring of high-risk medications.

 Monitoring guidelines for selected drugs exist in recommendations of organizations such as the American Heart Association guidelines to manage heart failure,

 the American Geriatrics Society Assessing Care of Vulnerable Elders medication quality indicators,

 and the National Quality Forum—endorsed measures.

A nationwide baseline monitoring assessment study

 developed more comprehensive guidelines for medications used through 2000, among which a subset of 14 were updated and adopted for an intervention trial between 2002 and 2003 within a single health maintenance organization.

We sought to develop an updated and comprehensive list of drugs requiring laboratory monitoring for an electronic medical record—embedded clinical decision support intervention at a multispecialty group practice for medications in clinical use during 2008. The intent of the updated guidelines was to include drugs introduced to the market since December 2000, the date of the literature review for the original drug and laboratory monitoring recommendations published by Raebel et al

 in 2005, as well as to update laboratory test and frequency recommendations based on 2007 changes in monitoring recommendations.

 Herein, we describe the development of guidelines to monitor high-risk medications in the ambulatory setting using a 2-step consensus-based approach, including a national expert advisory panel and local leaders to select candidate medications for monitoring and to determine the frequency of laboratory monitoring. To estimate the potential effect of our guidelines on actual practice, we determined the use frequency of the guideline drugs and the prevalence for each of the recommended laboratory tests. The specific objectives of this study were (1) to develop recommendations to guide the monitoring of high-risk medications in the ambulatory setting, (2) to assess the use prevalence of candidate medications for monitoring, and (3) to determine completion of recommended testing for medications dispensed among the patient population.

This study was conducted in a large multispecialty group practice that provides most medical care to members of a closely associated health plan based in the New England area of the United States. The group practice employs 250 outpatient clinicians at 30 ambulatory clinic sites. The practice uses the EpicCare Ambulatory (Epic, Verona, WI) electronic medical record system and provides medical care to approximately 180,000 individuals. For this analysis, we included patients if they received care from the multispecialty group practice, were 18 years or older, and obtained insurance coverage from the health plan between January 1, 2007, and July 31, 2008. Patients had to be continuously enrolled during the observation period and not residing in a long-term care facility. Data about medication exposure were derived from the prescription drug claims of the health plan. Data about laboratory test completion were derived from the multispecialty group practice electronic medical record.

Determination of High-Risk Drugs and Laboratory Monitoring Guidelines

Laboratory monitoring guidelines for high-risk drugs were developed using a sequential process adapted from an approach created and tested for a study

conducted within the Health Maintenance Organization Research Network Center for Education and Research on Therapeutics. This approach used an advisory committee of national experts and local health plan leaders, including practicing clinicians and pharmacists and experts in patient safety and geriatric pharmacotherapy. The charge for this group was as follows: (1) to review a comprehensive initial list of medications requiring monitoring; (2) to assess the importance of including monitoring recommendations to evaluate efficacy, safety, and clinically relevant drug—drug interactions; and (3) to determine the need to include infrequently prescribed medications in the guidelines. The initial list of high-risk drugs included those commonly implicated in adverse events among patients in the ambulatory setting

 and those associated with adverse events leading to emergency department visits,

 as well as drugs with low monitoring rates,

 drugs included in national quality guidelines,

We asked panel members to participate in an Internet-based questionnaire administered in a 2-round modified Delphi process

 between August and October 2008. Panelists were asked whether electronic monitoring alerts should be sent to primary care physicians, specialists, or both and whether monitoring alerts should be generated for infrequently dispensed or effectively obsolete medications. Panelists were also asked to rate the importance of monitoring each candidate medication for efficacy, toxic effects, and drug—drug interactions. Each question was answered based on a 5-point Likert-type scale to evaluate agreement or disagreement with statements concerning the importance of monitoring each medication or medication class for the domain assessed. The scale ranged from 1 (indicating “strongly agree”) to 5 (indicating “strongly disagree”).

After the first round of the survey, we eliminated questions for which there was agreement and readministered questions for which there was lack of consensus. Consistent with other modified Delphi methods, 

 for a question was defined by agreement on categorization by at least a majority (>50.0%) of respondents.

 We then administered a second questionnaire to participants. In this round, panelists were reminded of their original responses to individual questions and were given the group’s aggregated response to the questions in the first round. At this stage, each participant was given the opportunity to revise his or her response to increase consensus with the succeeding round. The results of the Delphi process informed the selection of the final high-risk drug list.

Determination of Laboratory Test Monitoring Frequency

After selection of the final high-risk drug list, 2 research pharmacists (JLD and AOK) reviewed the literature to determine the appropriate frequency of laboratory monitoring for each drug. The review included monitoring recommendations provided in manufacturer labeling information, nationally available published guidelines, and clinical guidelines from national organizations and initiatives (eg, American Heart Association guidelines

 and National Committee for Quality Assurance—developed measures and guidelines

). Because manufacturers’ warnings often describe a nonspecific frequency for monitoring such as “periodically,” we reviewed other authoritative sources that pharmacists and clinicians commonly use to guide decisions about the frequency of monitoring. These sources included the following: (1) the published literature; (2) the Micromedex Healthcare Series

 (Thomson Reuters [Healthcare] Inc, Greenwood Village, CO) database, which sources the primary case report and experimental literature; (3) UpToDate Inc

 (Waltham, MA), an online peer-reviewed reference database; and (4) the Pharmacist’s Letter

 (Therapeutic Research Center, Stockton, CA). The final list of laboratory monitoring tests and the associated monitoring frequencies were reviewed by a local panel of academic pharmacists, the clinical pharmacist for the multispecialty group practice, and the medical director of the multispecialty group practice to determine local acceptability and congruence with local quality standards.

We used drug dispensing claims between January 1, 2007, and July 31, 2008, to identify the first dispensing of a high-risk medication for a patient after January 1, 2008. 

 was defined as an initial dispensing on or after the index date of January 1, 2008, and no drug dispensing in the 6 months preceding the index date. 

 was defined as a dispensing on or after January 1, 2008, with evidence of drug dispensing in the 6 months before that date. Because clinicians might rely on laboratory test results for up to 6 months before initiation of a drug or might order a test within 2 weeks after initiation of a drug, we defined 

test completion for a new dispensing in 2008

 as having occurred if there was at least 1 associated monitoring test completed 180 days before dispensing to 42 days after dispensing (test ordered 

14 days after dispensing plus 28 days for test completion). Completion of a test of serum drug levels (eg, serum carbamazepine) for new prescriptions was measured from the day of dispensing to 30 days after initial dispensing. 

Test completion for chronic medication use

 was defined as having occurred if there was at least 1 recommended test for the drug test pair that occurred up to 365 days before the index dispensing in 2008 through 42 days after the dispensing if the test was indicated annually (or 180 days before to 42 days after index dispensing if the test was indicated every 6 months). For each drug—laboratory test combination, the proportion of completed recommended tests was determined for all index dispensings in the observation period. We used a nonparametric test for trend across ordered groups by Cuzick17 to examine whether more frequently dispensed drugs had a higher prevalence of recommended test completion. All analyses were conducted using commercially available software (SAS 9.2; SAS Institute Inc, Cary, NC). The study was approved by the institutional review boards of the University of Massachusetts Medical School, Worcester, and the participating group practice.

A panel of 3 pharmacists and 6 physicians participated in the consensus survey. All physicians were certified by the American Board of Internal Medicine, 2 had an added qualification in geriatric medicine, and 3 had subspecialty board certifications. All respondents completed the 2 rounds of the survey. After the second round of the survey, consensus was achieved for all 40 medications with respect to parameters (ie, efficacy or toxic effects) for laboratory monitoring and for alerts to physicians (ie, primary care physician vs specialist). The consensus panel agreed that alerts for high-risk medications were indicated to monitor efficacy and toxic effects alone or in the presence of significant drug—drug interactions. The panel also agreed that alerts should be sent to specialists and primary care physicians and should be developed for medications that were frequently and infrequently prescribed. Because strict national prescribing policies already exist to guide the use and monitoring of isotretinoin, we did not include this medication in the newly developed guidelines. The resulting list of drugs included 39 drugs or drug classes.

Pilot Testing of Monitoring Recommendations

Once the medication list was established, we examined overall dispensings of the selected medications. The chronic index dispensings ranged from 0 (procainamide and ganciclovir)to 13,351 (statins), and new index dispensings ranged from 0 (procainamide and ganciclovir) to 2463 (statins) during the observation period of January 1, 2008, to July 31, 2008. Based on these results, 4 drugs with fewer than 5 new or chronic users combined (ticlopidine, tolcapone, ganciclovir, and procainamide) were eliminated from the final drug list, resulting in 35 drugs or drug classes in the final list of monitoring guidelines. There were 61 separate drug—laboratory test combinations for monitoring, with 15 requiring more than 1 laboratory test (eg, amiodarone–thyroid-stimulating hormone [TSH] and amiodarone–aspartate aminotransferase [AST]) and 20 requiring a single test; for the purpose of this study, we considered a basic metabolic panel that includes sodium, potassium, chloride, bicarbonate, blood urea nitrogen, and creatinine levels as a single test.

Overall, the rates of test completion ranged from 0.0% for aspartate aminotransferase or alanine aminotransferase [ALT] among new users of nefazodone, phenobarbital, and quinidine at baseline to 96.8% for AST or ALT test among chronic users of niacin (

).Drug—laboratory test pair completion differed between new and chronic medication dispensings. For new medication dispensings, less than 20.0% of the recommended 61 drug–laboratory test pairs were monitored in the majority (>75.0%) of patients. In contrast, close to 40.0% of the recommended drug–laboratory test pairs for chronic dispensings were monitored in greater than 75.0% of patients.

For drugs with multiple laboratory tests, the percentage test completion often varied within the drug (Table 2). For example, for new amiodarone use, 52.7% had an AST or ALT  test, while 74.6% had a TSH test; for azathioprine use,28.0% had a baseline AST or ALT test, while 72.0% had a complete blood count (CBC). We also found that the frequency of medication use was associated with the frequency of recommended test completion (

). Infrequently dispensed medications had a lower prevalence of test completion. Nonparametric test for trend showed a significant trend toward more frequently dispensed medications having a higher prevalence of appropriate testing (

Herein, we describe the development of drug—laboratory test monitoring guidelines based on manufacturers’ recommendations and published guidelines. Recently published guidelines to monitor drugs in the long-term care setting included a similar list of medications and laboratory tests.

 To determine the potential effect of our guidelines on actual practice, we determined the use frequency of the guideline drugs and the prevalence of testing for each of the recommended laboratory tests. We found that (1) certain high-risk drugs had very low rates of use among our target population, (2) rates of testing for high-risk drugs varied widely, and (3) infrequently used drugs had lower testing rates. Our study adds to the literature by identifying ways to inexpensively improve patient safety and quality of care (describing a method for formulating monitoring recommendations) and by providing final monitoring recommendations (detailed guidelines are available from the authors) that are useful to practitioners considering similar interventions at their local institutions.

It is important to point out that the level of evidence for each recommended drug test pair varied and that this likely explains some of the variation in testing among our sample. For example, there is increasing controversy about the role of evaluating renal function in patients taking metformin.

Among our sample, we found that 45.0% to 60.0% of metformin users had completed a serum creatinine test. Similarly, there is a difference in the role of digoxin-level monitoring depending on indication, with digoxin serum drug levels of less importance when prescribed for heart rate control and of greater importance when used for heart failure.

 Among our sample, we found digoxin levels completed for 30.0% to 60.0% of users. We did not discern the indication for digoxin use in our study, but previous investigations have shown similarly low rates of digoxin-level testing.

 Variance in the level of evidence for testing may explain some of the differences we observed in the prevalence of completion among different tests within the same drug.

We are concerned about the low prevalence of testing for drug test pair recommendations that are much less controversial such as evaluation of thyroid function among amiodarone users. Amiodarone-induced thyroid dysfunction, including thyrotoxicosis and hypothyroidism, can occur in up to 20% of users.22 For this drug test pair, we found that 75.0% of new users and 50.0% of chronic users had completed a TSH test. The reason for the lower rate of testing among new drug users is unclear. Although a possible explanation may be the initiation of drugs in the hospital, as may be the case for amiodarone among using patients with arrhythmias, the prevalence of hospitalizations did not differ among those with versus without appropriate monitoring of amiodarone in a prior study.

 Further investigation is necessary to determine whether hospitalizations explain low testing rates among our study population.

An important finding of our study was that patients using infrequently prescribed drugs were less likely to complete recommended laboratory tests. This is consistent with literature showing an association between patient volume and quality of care

 but is not supported by at least 1 study

 examining patient volume in association with process measures such as laboratory testing in diabetes. The rationale supported by the human factors model of medical errors is that physicians will likely have better familiarity of the literature for drugs that they prescribe more frequently than for drugs that they do not prescribe.

The implication of this finding can be considered in terms of potential relative and absolute effects of an intervention to improve monitoring. There is potential for significant effect from interventions aimed at improving infrequently used medications because baseline rates of testing are low. This is important for drugs such as phenobarbital, whose laboratory monitoring is a Healthcare Effectiveness Data and Information Set measurement for assessing health plan quality of care.

 In contrast, the number of patients affected by infrequently prescribed drugs is small from a population health perspective. For example, the number of patients who failed to receive monitoring of statins was 2136, more than 20 times greater than the number of patients affected by failure to monitor the 7 infrequently prescribed drugs combined (ie, imidazole antifungals, terbinafine, angiotensin II receptor blockers, isoniazid, rifampin, theophylline,and nefazodone). Health plans considering the implementation of interventions to improve medication monitoring need to weigh the absolute and relative potential effects of their program. Although the marginal cost of targeting many medications in a health information technology—based intervention to improve monitoring is minimal and allows for the inclusion of many medications, programs using high-cost resources such as pharmacist time need to limit target medications by weighing the relative overall effect carefully to optimally select target medications.

Other important drugs for inclusion in monitoring guidelines are those with a high likelihood of a serious adverse outcome. Our guidelines incorporate several drugs commonly implicated in adverse drug events leading to emergency department visits,

 including hypoglycemics, warfarin, anticonvulsants, digoxin, theophylline, and lithium. Of these, we found high rates for warfarin, which is often monitored through special anticoagulation programs, but low-to-moderate rates for digoxin, anticonvulsants, lithium, and theophylline.

Limitations of our study should be noted. First, our study was conducted in a single multispecialty group practice. Second, laboratory tests may have been ordered for another reason (ie, not for high-risk medication monitoring), so that we may have overestimated the prevalence of recommended testing. Third, we were unable to confirm patient adherence to drugs and were unable to identify patients who did not complete tests because they were no longer using the medication. Fourth, in many cases, despite the panel’s conclusions, limited evidence was found to support testing frequency or importance, and that may explain the low prevalence of test ordering for some drugs. For many drugs, further study is necessary to determine whether laboratory test monitoring improves health outcomes. Furthermore, lower testing for new use may also be explained by drug initiation in the hospital setting for drugs such as amiodarone and antiepileptics. In these cases, it is possible that baseline testing was obtained after initiation in the hospital and was not repeated after discharge to the ambulatory setting. However, similarly low rates of TSH and aminotransferase testing have been reported for chronic users of amiodarone

 and for determining serum levels of antiepileptics.

 Fifth, we did not assess all guideline recommendations such as monitoring after dose changes, drug—drug interactions, changing renal function, and clinical status.

Our study updates the laboratory monitoring recommendations for high-risk drugs to include drugs used in ambulatory practice until 2008. Our findings and guidelines are generalizable to other institutions seeking to improve monitoring of high-risk medications to enhance the safety for their patient population after review and adaptation by local medical and pharmacy leadership to ensure consistency with local standards. Health plans considering similar interventions can use our findings to weigh the potential benefit of targeting drugs of high-frequency and low-frequency use in clinical practice, as our study demonstrates potential for improvement in the laboratory monitoring of drug types. Further research is necessary to understand factors contributing to the low monitoring of high-risk drugs, including physician factors resulting in lack of test ordering and patient factors leading to incomplete performance of tests.

We acknowledge the contributions of the national expert advisory panel, including Susan Andrade, ScD; Steven Simon, MD; Jerry H. Gurwitz, MD; and Marsha A. Raebel, PharmD. Local experts included Josie Cambia-Kiely, PharmD; Mojgan Hajji, PharmD; Michael Kelleher, MD; Leslie Harrold, MD; Sarah McGee, MD; and Robert Yood, MD.

From the Division of Geriatric Medicine (JT, TSF, SJG, SHF, JHG), University of Massachusetts, Worcester, MA; Meyers Primary Care Institute (JT, TSF, YZ, JCF, JHG), Worcester, MA; Fallon Clinic (LDG), Worcester, MA; Massachusetts College of Pharmacy (JLD, AOK), Worcester, MA; and Kaiser Permanente Colorado (MAR), Denver, CO.

CLINICAL

Pilot testing of guidelines for the laboratory monitoring of high-risk medications shows that monitoring is highly variable and that there is room for improvement.


Development and Pilot Testing of Guidelines to Monitor High-Risk Medications in the Ambulatory Setting

 To test the hypotheses that African American patients and older patients with stage IV colorectal cancer were less likely to receive newer chemotherapy agents.

Among 5068 Surveillance, Epidemiology, and End Results—Medicare patients diagnosed as having stage IV colorectal cancer between 2000 and 2002, a total of 2466 received chemotherapy and were included in the analysis. Irinotecan hydrochloride was the first of the “newer” chemotherapy agents and was marketed in 2000 as a first-line add-on agent. Descriptive statistics were generated, and a multivariable logistic regression was run to estimate the odds of receiving irinotecan among African American patients and older patients and within 2 months of chemotherapy initiation.

African American patients had lower odds of initiating treatment with a newer chemotherapy than white patients (adjusted odds ratio, 0.641; 95% confidence interval, 0.453-0.907). An age disparity was also found, with all older age groups being significantly less likely to initiate treatment with a newer chemotherapy than the youngest age group: the adjusted odds of receiving newer chemotherapy agents (relative to patients aged 66-70 years) were lower and significant among patients aged 71 to 75, 76 to 80, and older than 80 years (odds ratios, 0.708, 0.527, and 0.213, respectively).

Disparities in chemotherapy selection exist among patients receiving chemotherapy for stage IV colorectal cancer. On initiating chemotherapy, African American patients and older patients were less likely to receive a newer agent.

In a population-based study of older Medicare patients with stage IV colorectal cancer, racial/ethnic and age disparities were detected in the receipt of newer chemotherapy agents (using irinotecan hydrochloride as an example of a newer therapy) for first-line management of the disease. The following specific findings were observed:

African American patients had lower odds of initiating treatment with a newer chemotherapy than white patients. 

Patients aged 71 to 75, 76 to 80, and older than 80 years had lower odds of initiating treatment with a newer chemotherapy agent than patients aged 66 to 70 years.

 have evaluated racial/ethnic and age disparities in whether patients receive chemotherapy for nonmetastatic colo-rectal cancer (CRC) treatment and have found that patients of nonwhite race/ethnicity and older patients are less likely to receive chemotherapy. However, these studies did not evaluate specific types of chemotherapy agents used.

 Management of metastatic (stage IV) CRC historically has involved 5-fluorouracil—leucovorin as the standard chemotherapy,

 but the introduction of various novel add-on chemotherapies with improved efficacy has led to an increase in the chemotherapeutic options available to manage stage IV CRC.

 Irinotecan hydrochloride, approved in 1998 as a second-line agent for metastatic CRC treatment and approved in 2000 as a first-line agent for use with 5-fluorouracil—leucovorin,

 was the first of these novel agents to show improved efficacy relative to 5-fluorouracil—leucovorin alone.

 Irinotecan inhibits the enzyme necessary for the reversible breakage and relegation of DNA strands during DNA synthesis, leading to significant improvement in tumor progression slowing and median survival time (approximately 3 additional months relative to the 5-fluorouracil—leucovorin regimen).

 Its availability, as well as the subsequent availability of other novel agents such as oxaliplatin, bevacizumab, and cetuximab, has raised expectations for more efficacious regimens than 5-fluorouracil—leucovorin alone.

 Nevertheless, these novel agents also place patients at higher risk of experiencing adverse effects as a result of their increased toxic effects.

 Knowledge is scarce about racial/ethnic and age disparities in the receipt of newer chemotherapies. Addressing this gap by examining the first of the newer agents marketed for treatment of advanced CRC disease can provide an understanding of healthcare disparities among specific services for CRC management. In addition to knowing whether patients are receiving chemotherapy in real-world settings, it is important to know whether newer chemotherapies are being selectively utilized among chemotherapy users, particularly elderly patients in whom incidence of the disease is high.

 This study sought to determine whether racial/ethnic and age disparities existed in the selection of specific types of chemotherapy among elderly patients with stage IV CRC. Given that irinotecan was the first novel agent to be available at the start of the study period, newer therapy was designated as irinotecan-containing chemotherapy. It was hypothesized that patients of African American race/ethnicity and older patients with stage IV CRC would have a lower likelihood of receipt of newer chemotherapy than white patients and younger patients.

 The linked Surveillance, Epidemiology, and End Results (SEER)-Medicare database was used for this study.

 SEER data provided clinical and demographic information about patients with cancer from SEER registries across the following 16 geographic areas: Connecticut, Detroit (Michigan), Hawaii, Iowa, New Mexico, Seattle (Washington), Utah, Atlanta (Georgia), rural Georgia, Kentucky, Louisiana, New Jersey, greater California, and Los Angeles, San Jose, and San Francisco (California). Health utilization information for these patients was obtained from their Medicare Part A and Part B claims, which include hospital services, physician and other noninstitutional provider services, and outpatient institutional providers.

 A total of 112,345 patients with at least 1 diagnosis of nonappendiceal colon or rectal cancer occurring between 1998 and 2002 were initially identified from SEER. Patients diagnosed at death were excluded (n = 1104 patients), as were patients diagnosed as having nonadenocarcinomas or a stage of disease that did not require any chemotherapy treatment (n = 39,755). Patients younger than 66 years also were excluded to allow a complete year of data during the year before CRC diagnosis to identify prediagnosis comorbidities (n = 10,600).

 Patients who were ineligible for Medicare fee for service were excluded, as were those with no utilization information because of death or loss of Medicare eligibility occurring during the same month of diagnosis (n = 22,140). To ensure that chemotherapy use was specifically for stage IV CRC, patients with a history of cancer dating 5 years or earlier before the current CRC diagnosis or patients with a concurrent cancer occurring in the same year as the CRC diagnosis were excluded (n = 3883). The final sample was further refined to those eligible to receive first-line irinotecan-containing chemotherapy (ie, patients with stage IV disease diagnosed after 1999 [n = 5068]).

 The main variable of interest was the receipt of newer chemotherapy (irinotecan vs all other chemotherapy) within 2 months of chemotherapy initiation. Chemotherapy administration was identified through specific chemotherapy-related medical codes in physician, outpatient facility, and durable medical equipment Medicare claims. Various resources and published studies

 that have measured chemotherapy through Medicare claims were used to identify the relevant codes that represented chemotherapy use. These codes are listed in 

 Patients with any chemotherapy-related Medicare claim were identified as chemotherapy users, and the earliest chemotherapy-related claim on or after the first day of the diagnosis month was used to identify the chemotherapy start date. A 2-month window beginning on the chemotherapy start date was searched for any claims indicating irinotecan use (Healthcare Common Procedure Coding System code J9206). If no irinotecan claims were identified, the chemotherapy user was identified as having received “other” chemotherapy.

SEER-Medicare data were used to identify several clinical variables. Year of diagnosis was obtained from the SEER file. Other SEER clinical variables included a history of cancer that occurred at least 5 years before the cancer diagnosis under study. Furthermore, although all patients had stage IV disease, variability in tumor characteristics such as tumor site and lymph node involvement was observed, and these variables were included in the descriptive analysis.

 The Charlson Comorbidity Index (CCI), a measure of disease burden, was used as a proxy for health status (prediagnosis). The index has been validated as a reliable tool for the prediction of morbidity in patients with cancer.

 Calculation of the index involved using physician and hospital claims in the year preceding cancer diagnosis to search medical codes that reflected the following: myocardial infarction, congestive heart failure, peripheral vascular disease, chronic obstructive pulmonary disease, dementia, paralysis, diabetes, chronic renal failure, liver disease, ulcers, rheumatoid arthritis, and AIDS.

 Because few patients had CCI scores exceeding 2, the final CCI obtained was used to categorize patients into the following 3 groups: 0 CCI, 1 CCI, or more than 1 CCI.

 Other cancer-related procedures occurring within the first 4 months of diagnosis were measured. Codes reflecting surgery to the colorectal region, surgery to metastatic sites (eg, liver, lung), and radiation therapy were identified from published resources (Table 1).

In addition to age at diagnosis and race/ethnicity, sex and marital status were obtained from the SEER-Medicare enrollment files. The following 4 age groups were created: 66 to 70, 71 to 75, 76 to 80, or older than 80 years. Race/ethnicity was categorized as white, African American, or other. A variable indicating whether the patient was married at the time of diagnosis was also available. Finally, SEER registry geographic information was available as a demographic variable.

 Individual-level socioeconomic variables were unavailable. However, SEER data contained a group of aggregate zip code—level and census tract–level household income variables that were used to create a proxy variable for socioeconomic status.37 A compound median household income proxy variable for socioeconomic status was generated using an algorithm previously applied to SEER-Medicare data.38,39 In order of priority, the algorithm used information from the census tract median household income by race/ethnicity, by age, and by income, as well as the zip code median income.38,39 Patients were then ranked into quintiles based on the final composite median income value.

 Patient residence in a metropolitan (vs rural) area was also measured. This tested whether proximity to an urban area might influence the receipt of newer agents if adoption occurs earlier in urban environments.

 All analyses were conducted using commercially available software (SAS/STAT software version 9.1; SAS Institute Inc, Cary, NC). Bivariate descriptive statistics and likelihood ratio 

tests of significance were generated to compare patients initiating newer chemotherapy versus patients initiating any other chemotherapy with respect to demographic, socioeconomic, and clinical variables. Variables significantly associated with initiation of newer chemotherapy at the bivariate level (

<.05) were included in a multivariable logistic regression model to test the association between race/ethnicity or age and the odds of initiating on newer chemotherapy, while controlling for clinical and demographic factors.

The final sample among 5068 patients with stage IV CRC receiving chemotherapy was 2466 patients. Approximately 32% of the chemotherapy users initiated their therapy using an irinotecan-containing regimen. Based on healthcare claims,the patient group using non—irinotecan-containing chemotherapy primarily received (in approximately 93% of patients) 5-fluorouracil as the cytotoxic agent.

 Results of the bivariate analysis comparing patients who initiated treatment with newer chemotherapy versus those who initiated treatment with other chemotherapy are given in 

. Race/ethnicity, sex, age, year of diagnosis, median household income, presence of comorbidities, marital status, surgery to the metastatic site, and radiation therapy were significantly associated with the receipt of newer chemotherapy agents (

 <.05). Specifically, among patients initiating treatment with newer chemotherapy, there were higher proportions of white, male, and younger patients, as well as patients diagnosed in later years (2001 and 2002), patients in higher quintiles of median household income, patients with fewer comorbidities, and married patients, compared with patients not receiving irinotecan. With respect to clinical characteristics significantly associated with the receipt of newer chemotherapy agents, fewer patients who received radiation therapy or surgery to the metastatic site (within 4 months of diagnosis) received newer chemotherapeutic agents.

 Results of the final model developed for the multivariable logistic regression analysis are given in 

. After simultaneously controlling for demographic and clinical variables significant at the bivariate level, African American race/ethnicity and older age remained significantly (

<.05) negatively associated with the receipt of newer chemotherapy agents in the logistic regression analysis. Specifically, African American patients were 35.9% less likely to initially receive newer chemotherapy than white patients (odds ratio, 0.641; 95% confidence interval, 0.453-0.907). Patients aged 71 to 75, 76 to 80, and older than 80 years were 29.2%, 47.3%, and 78.7%, respectively, less likely to initiate treatment with newer chemotherapy than patients aged 66 to 70 years (

<.05). Male patients remained significantly more likely to receive newer chemotherapy.

This study reveals important findings with regard to the receipt of general chemotherapy and specific chemotherapeutic agent use among a Medicare sample of older patients with stage IV CRC. Approximately 49% of the final patient sample received any chemotherapy. This estimate is not directly comparable to other SEER-Medicare data because of differences in staging, disease site (colon vs rectal), and period of study, but evidence nevertheless indicates that a substantial proportion of Medicare patients with CRC do not receive chemotherapy: studies

 of SEER-Medicare older patients with stage III CRC (among whom chemotherapy benefits are widely acknowledged) indicate that the receipt of chemotherapy ranges from only 54% to 70%.

 Our study results also indicate that, among chemotherapy-receiving patients with stage IV CRC, African American patients and older patients were significantly less likely to initiate their chemotherapy with a newer agent such as irinotecan compared with white patients and younger patients. The direction of results is generally consistent with published studies

 that have evaluated racial/ethnic and age disparities among patients receiving treatment for locally advanced CRC.

 Disparities in the receipt of chemotherapy among patients with stage IV CRC have not been as widely researched as disparities among patients with earlier stages of CRC. However, McKibbin et al

 evaluated the receipt of irinotecan, oxaliplatin, or bevacizumab in patients with advanced CRC from 10 cancer centers and detected a significant association between younger age (<65 years) and increased receipt of newer chemotherapeutic agents. Furthermore, compared with patients younger than 65 years, the receipt of newer agents decreased among older patients (ie, >70 and >75 years).

 In addition, the direction of racial/ethnic and age disparities observed in the present studywas consistent with that in other observational studies

 addressing surgical management of patients with stage IV CRC.

Among Medicare enrollees with metastatic colorectal cancer, the use of newer chemotherapy agents was lower for African American patients and for older patients.

July 2010

Cost-Effectiveness of Laparoscopic Gastric Banding and Bypass for Morbid Obesity

Development and Pilot Testing of Guidelines to Monitor High-Risk Medications in the Ambulatory Setting

Racial/Ethnic and Age Disparities in Chemotherapy Selection for Colorectal Cancer

Using the Lessons of Behavioral Economics to Design More Effective Pay-for-Performance Programs

Can Outpatient Pharmacy Data Identify Persons With Undiagnosed COPD?

Screening Cardiac Surgery Patients for MRSA: An Economic Computer Model

Diabetes Disease Management in Medicare Advantage Reduces Hospitalizations and Costs

Mometasone Furoate Versus Beclomethasone Dipropionate: Effectiveness in Patients With Mild Asthma

US Cost Burden of Ischemic Stroke: A Systematic Literature Review

Cost Analysis Review of Stroke Centers, Telestroke, and rt-PA