Development and validation of a risk score to identify patients at high risk for opioid-related adverse drug events.

BACKGROUND: Opioid-related adverse drug events (ORADEs) are common causes of hospitalization and increased health care costs. OBJECTIVES: To (a) estimate rates of specific adverse drug events (ADEs) among gastrointestinal (GI) surgery patients receiving postoperative opioids; (b) examine the utility of a risk-scoring model in categorizing patients at high risk of experiencing ORADEs; and (c) quantify potential clinical/economic benefits of targeting high-risk GI surgical patients for opioid-sparing regimens in terms of hospitalization cost, length of stay (LOS), and 30-day readmission rates. METHODS: Using a retrospective design based on an administrative database, patients with an inpatient surgical procedure between January 1, 2010, and December 31, 2010, were included. GI surgical patients aged greater than 18 years followed from admission through 30 days postdischarge were characterized as high or low risk using clinical/demographic characteristics and were evaluated for several outcomes. Using multivariate logistic regression, the ORADE incidence, total hospitalization cost, LOS, and 30-day readmissions were compared for high-risk and low-risk patients. RESULTS: In 87.8% (n = 3,235) of the surgical population, there was a strong concordance between risk assignment and ORADE incidence. Among the remaining 12.2% (n = 449) of patients, 5.5% (n = 202) were low risk with an ORADE, and 6.7% (n = 247) were high risk without an ORADE. Overall, 20.6% (n = 344) of high-risk patients experienced ≥1 ORADE (mean cost: $31,988; LOS: 12.1 days) compared with only 5.3% (n = 107) of low-risk patients (mean cost: $25,216; LOS: 8.0 days). High-risk patients had higher hospitalization costs and longer LOS than low-risk patients, respectively (mean cost: $19,234 vs. $13,036; mean LOS: 6.8 days vs. 3.3 days). These differences correspond to 47.0% higher costs for high-risk patients and an LOS approximately twice as long compared with low-risk patients. CONCLUSIONS: Patient clinical/demographic characteristics influence the risk of developing ORADEs. Risk assessment tools can effectively identify high-risk patients, thereby enabling interventions that can reduce ORADEs, decrease hospital costs, and improve postsurgical experiences for patients.

Comparison of injectable anticoagulants for thromboprophylaxis after cancer-related surgery.

PURPOSE: The clinical and economic outcomes associated with using injectable anticoagulants for thromboprophylaxis after cancer-related surgery are evaluated. METHODS: This retrospective cohort analysis was conducted from an institutional perspective using hospital administrative data and examined patients age 18 years or older who received unfractionated heparin (UFH), enoxaparin, dalteparin, or fondaparinux after undergoing cancer-related surgery. Outcomes assessed included venous thromboembolism (VTE) and major bleeding (MB) rates; VTE-related, MB-related, and all-cause readmission rates; mean length of stay (LOS); and mean total cost of care during hospitalization. RESULTS: In the 4068 patients evaluated (1017 per group), VTE rates were similar for fondaparinux compared with the other anticoagulants. The risk of MB was 80% higher for enoxaparin (p = 0.035) and 2.5 times higher for UFH (p = 0.0004) but not significantly higher for dalteparin compared with fondaparinux. The mean LOS was 8% longer for patients taking enoxaparin (p = 0.03) and dalteparin (p = 0.0494) and 21% longer for those treated with UFH (p < 0.0001) compared with fondaparinux. The unadjusted mean ± S.D. total cost of care per patient was lower in the fondaparinux group compared with the enoxaparin and UFH groups but higher compared with dalteparin. CONCLUSION: A retrospective evaluation of hospital administrative data for patients who had received thromboprophylaxis after cancer-related surgery revealed a similar risk of VTE with fondaparinux compared with other injectable anticoagulants. Fondaparinux was associated with a lower risk of MB compared with enoxaparin and UFH but did not differ significantly from dalteparin in this regard. A shorter LOS was observed for patients who received fondaparinux compared with dalteparin, enoxaparin, and UFH. The total cost of care for patients who received fondaparinux was lower compared with enoxaparin or UFH but higher compared with dalteparin.

Lifetime costs to Medicare of providing care to patients with chronic lymphocytic leukemia.

This study estimated the average lifetime costs of patients with chronic lymphocytic leukemia (CLL) relative to similar patients without cancer. An analysis of the Medicare 5% database (1999-2007) was conducted. Each patient with CLL was matched up to three patients without cancer based on year of birth, gender, race and state. Average total lifetime costs were estimated using the Kaplan-Meier sample average estimator and stratified by treatments. For patients who died, average monthly costs during continuing and terminal phases were compared between cohorts. A total of 7463 patients with CLL and 22 331 matched controls were identified (mean age: 76 years; proportion of women: 49%). The mean observation period was 39.4 months for patients with CLL and 45.9 months for matched controls. Patients with CLL incurred average costs of $87,151 compared to $47 642 for matched controls (p < 0.001). Among common CLL treatments, average costs per patient were $5140 for rituximab and $953 for radiation therapy. Compared to matched controls, patients with CLL had significantly higher monthly costs during the continuing and terminal phases. This study showed that average lifetime costs to Medicare were $87,151 for patients with CLL compared to $47,642 for matched controls without cancer, for a significant difference of $39,509.

Impact on daily functioning and indirect/direct costs associated with chemotherapy-induced nausea and vomiting (CINV) in a U.S. population.

PURPOSE: Chemotherapy-induced nausea and vomiting (CINV) is a debilitating side effect of chemotherapy, but it may be prevented or mitigated with medications. Uncontrolled CINV can lead to reduced quality of life and can result in increased costs (due to health care utilization and missed work). We prospectively assessed the prevalence and burden of CINV in a US population. METHODS: Final analysis was performed on 178 patients, beginning chemotherapy during 2007-2008 at oncology specialty settings. Patients kept a diary recording use of antiemetic medications just before the start of chemotherapy and use of antiemetic medications, health care resources, and episodes of nausea and vomiting during the 5 days following. In addition, they completed a Functional Living Index-Emesis (FLIE) questionnaire and a Work Productivity and Assessment Inventory-Nausea and Vomiting assessment, to determine the impact of CINV on daily functioning and on work productivity, respectively. Physicians independently recorded prescribed medications and health care utilization. RESULTS: Of the patients, 61.2% reported experiencing CINV (34.3% with acute CINV and 58.4% with delayed CINV). Based on the FLIE assessment, 37.2% of all patients reported reduced daily functioning, and of those with poorly managed CINV, about 90% reported a significant impact on daily functioning. Total costs due to CINV were on average $778.58 per patient from the day of administration through the 5 days following the first cycle of chemotherapy; patients with more severe CINV typically had higher costs. CONCLUSIONS: CINV remains a significant problem among US patients, suggesting a need for more effective prophylaxis use in clinical practice.

Consequences of major bleeding in hospitalized patients with non-ST segment elevation acute coronary syndromes receiving injectable anticoagulants.

OBJECTIVE: To evaluate the burden of major bleed in patients with non-ST segment elevation acute coronary syndromes (NSTE ACS) receiving injectable anticoagulation from the hospital perspective. METHODS: Retrospective analysis of inpatient medical and pharmacy data from the Premier Perspective Comparative Database between 1/1/2003 and 3/31/2006. Hospitalized patients aged >or=18 years with a diagnosis of UA or NSTEMI who received an injectable anticoagulant agent during the same hospital stay were stratified into two cohorts: those who experienced a major bleed during hospitalization and those who did not, defined by the presence of >or=1 pre-specified ICD-9 codes. Length of hospital stay (LOS), inpatient mortality, 30-day readmissions, and hospitalization costs over 30 days were assessed between the cohorts using statistical models to control for covariates which may have impacted the outcomes. RESULTS: Patients with a major bleed had significantly longer length of stay (13.8 days vs 5.6 days), higher readmission rates (31.3% vs 14.7%), and increased all-cause mortality (15.0% vs 4.5%) compared with patients who did not bleed. After controlling for covariates, major bleeding was significantly associated with increased length of stay, readmission rate, and mortality. Adjusted costs were $13,856 higher on average for patients with a major bleed (95% CI: $13,828-$18,884; p < 0.0001). Subanalyses conducted on patients aged >or=65 years and those undergoing invasive procedures demonstrated higher occurrence of bleed than the general population and a similar impact on outcomes assessed. CONCLUSION: In conclusion, the study showed that patients with UA or NSTEMI who experience a major bleed have significantly longer hospital stays, higher readmission rates, increased costs, and increased mortality than those without a major bleed. These data emphasize the importance of considering the safety profile in context of the efficacy of the recommended agents. The findings from this study are limited by the retrospective study design and certain endpoints, such as readmissions, may have been underreported.

Barriers to the initiation of, and persistence with, insulin therapy.

OBJECTIVES: Many patients with Type 2 diabetes mellitus (DM) delay and/or discontinue the use of insulin. This study determined patient-perceived barriers to the initiation of, and persistence with, insulin therapy. RESEARCH DESIGN AND METHODS: Patients > or = 18 years with Type 2 DM and > or = 1 elevated HbA(1c) test result (> or = 9%) were identified from computerized laboratory test results within a single healthcare system, a study limitation. Insulin use patterns were characterized by automated pharmacy claims data, and patients were classified into those who discontinued insulin use (discontinuers) or those who did not initiate insulin use (non-initiators). Telephone interviews were conducted to determine the barriers to initiation of, and persistence with, insulin therapy. RESULTS: Response rates were 80.0% (73/91) for discontinuers and 82.0% (129/157) for non-initiators. Pharmacy claims data indicated that discontinuers stopped filling prescriptions for insulin; 46.6% of patients self-reported discontinuing insulin. The average time between first and last prescription for insulin among discontinuers was 4.9 years (SD = 4.4). The most common reasons for discontinuation were related to insulin injection (74.0%) and a doctor's advice not to use insulin (47.1%). In the non-initiator group, 86.1% were never advised by a healthcare provider to take insulin. CONCLUSIONS: These findings suggest that issues related to insulin injection are the primary reason patients with Type 2 DM discontinue insulin therapy. Understanding these patterns is important to develop interventions to overcome barriers to treatment and improve the medical outcomes of patients with Type 2 DM.

Hypoglycemic events and glycosylated hemoglobin values in patients with type 2 diabetes mellitus newly initiated on insulin glargine or premixed insulin combination products.

PURPOSE: Rates of hypoglycemic events and their associated costs were compared among patients with type 2 diabetes mellitus newly initiated on insulin glargine or a premixed insulin fixed-combination product. METHODS: Patients newly initiated on insulin glargine or premixed insulin fixed-combination products (including pen delivery systems) between June 1, 2001, and February 29, 2004, were identified using an administrative claims database. Hypoglycemic events were identified from International Classification of Diseases, 9th Revision, Clinical Modification codes. Multivariate analyses were performed. RESULTS: A total of 2315 patients met the inclusion criteria. Of those, 1212 received insulin glargine and 1103 received a premixed fixed-combination insulin product. The mean +/- S.D. treatment duration was 13.7 +/- 8.1 months. Patients treated with premixed insulin had a higher hypoglycemic event rate than glargine patients (13.8 versus 7.0/100 patients/year; p = 0.027), which yielded a number needed to treat of 15 patients. The mean cost per hypoglycemic event was $1049 (95% confidence interval, $426-1672). The mean annual cost of all insulin use was $46 more for the insulin glargine cohort than for those who received premixed insulin ($534 versus $488, respectively) (p < 0.05). Mean postindex insulin use was higher in patients receiving premixed insulin than in those treated with insulin glargine (48.1 versus 43.8 units per day) (p < 0.05). CONCLUSION: Patients with type 2 diabetes mellitus who were newly initiated on insulin glargine had a lower rate of hypoglycemic events compared with patients newly initiated on a premixed fixed-combination insulin product. Treatment of 15 patients with insulin glargine instead of premixed insulin for one year would avoid one hypoglycemic event per year.

Incremental cost savings 6 months following initiation of insulin glargine in a Medicaid fee-for-service sample.

This study assessed the role of insulin glargine use on the short-term costs of diabetes care from a state Medicaid fee-for-service reimbursement perspective. A retrospective claims analysis was performed for 20% of Medicaid recipients in California (Medi-Cal) between November 2000 and September 2002. Each patient with continuous enrollment at least 6 months before and after the insulin glargine index date was matched with 2 reference patients by age, gender, diabetes diagnosis, and enrollment dates. Costs were calculated for emergency department, inpatient, outpatient, and pharmacy claims paid, and changes from baseline were determined. Incremental cost savings were estimated from baseline cost in the insulin glargine group multiplied by the difference in percentage of changes from baseline. Of the 1018 insulin glargine users, 267 satisfied inclusion criteria and were matched to 534 reference patients. Baseline treatment costs were 2.7-fold higher in the insulin glargine group. Absolute cost reductions in the insulin glargine and reference groups were 185 dollars and 72 dollars per patient, respectively. Incremental cost savings were 69 dollars per insulin glargine user, including more than 200 dollars in incremental savings for inpatient claims. Hypoglycemia decreased from 9.5% to 3.8% of inpatient claims in the insulin glargine group but remained 1.1% throughout in the reference group. Insulin glargine use was associated with reduced inpatient hypoglycemia-related claims paid and short-term reductions in the inpatient and total costs of diabetes-related care. Additional studies are warranted to assess the influence of insulin glargine on diabetes-related costs.

Differences in hypoglycemia event rates and associated cost-consequence in patients initiated on long-acting and intermediate-acting insulin products.

OBJECTIVE: To compare hypoglycemia event rates in patients initiated on long-acting insulin analog (glargine) or intermediate-acting insulin (NPH) and to analyze the associated cost-consequence from a managed care perspective. STUDY DESIGN: A retrospective analysis of pharmacy and medical claims and electronic laboratory result data using a southeastern United States managed care health plan. METHODS: Patients newly initiated on glargine or NPH between July 1, 2000 and August 31, 2002 were included. Hypoglycemia events were identified from medical claims by their ICD-9CM codes. Multivariable techniques were used to compare hypoglycemia event rates between cohorts. RESULTS: A total of 1434 patients were eligible (glargine = 310, NPH = 1124). The mean age was 53 years +/- 17 years and 51% of patients were male. The mean treatment duration was 8.6 months +/- 4.5 months. Multivariate analyses showed that patients in the NPH group had a higher hypoglycemia event rate than the glargine group (18.3 versus 7.3 per 100 patients per year; p = 0.009). The number needed to treat (glargine versus NPH) to avoid one hypoglycemia event per patient per year was nine patients at an A1C of 7%. The mean annual index medication cost was $47 more for glargine ($390) than for NPH ($343) per patient per year (p = 0.042). The mean cost per hypoglycemia event was $1087 (95% CI: $764-$1409). CONCLUSIONS: Patients treated with glargine had significantly lower hypoglycemia event rates compared to the NPH group. The risk difference indicated that one hypoglycemia event would be avoided for every nine patients treated with glargine instead of NPH. The cost increase associated with treating nine patients with glargine rather than NPH is less than the cost of treating one hypoglycemia event. In this population, the savings associated with reduced hypoglycemic events more than offset the increased acquisition cost associated with glargine.

Reduced severe hypoglycemia with insulin glargine in intensively treated adults with type 1 diabetes.

BACKGROUND: The goal of new therapies introduced for type 1 diabetes should be to decrease hypoglycemic episodes while improving glycemic control. METHODS: A database was used to computer match the baseline A1C values in 196 subjects with type 1 diabetes receiving multiple daily injections (MDI) consisting of four or more injections per day. There were 98 patients transferred from NPH to insulin glargine (Lantus, Aventis Pharmaceuticals, Bridgewater, NJ), and 98 patients remained on NPH throughout the study. The gender distribution and mean age (approximately 32 years), duration of diabetes (approximately 16 years), and duration of treatment (approximately 13 months) were not significantly different between the groups. The majority of patients were well controlled (>50% in both groups had an A1C <7%). RESULTS: The mean A1c values were not significantly different in the groups at baseline or at follow-up. Severe hypoglycemic episodes per patient per year were significantly lower in the glargine group compared with the NPH group (0.5 vs. 1.2, respectively; P = 0.04). The mean end-of-study total (P = 0.03) and long-acting (P = 0.0001) doses were significantly reduced from baseline in the group that switched to glargine, but not in the group that remained on NPH, with no change in the short-acting dose in either group. The weight gain was significantly higher in the NPH group at the end of the study (P = 0.004) with no significant change in the glargine group. CONCLUSIONS: Transfer to glargine treatment from NPH in MDI regimens significantly reduces severe hypoglycemic episodes despite a decline in long-acting basal insulin without significant weight gain.