Melanoma associated with tumour necrosis factor-α inhibitors: a Research on Adverse Drug events And Reports (RADAR) project

BACKGROUND

Tumour necrosis factor-α inhibitors (TNFαIs) are used for treatment of inflammatory disorders. There is evidence linking these agents with occurrence of malignancies. For four out of five TNFαIs the Food and Drug Administration (FDA) label states, 'melanoma has been reported in patients treated with these agents'.

OBJECTIVES

To determine whether a statistically significant association exists between administration of TNFαIs and development of malignant melanoma.

METHODS

We searched the FDA Adverse Event Reporting System (FAERS) database for terms related to melanoma and TNFαIs for detection of safety signals. We also searched a large urban academic electronic medical record (EMR) database for which we calculated the relative risk (RR) of melanoma in subjects exposed to TNFαIs vs. nonexposed subjects.

RESULTS

There were 972 reports of melanoma associated with a TNFαI identified in the FAERS database, with 69 reports among individuals using more than one TNFαI. A safety signal was detected for infliximab, golimumab, etanercept and adalimumab, but not certolizumab pegol. For TNFαIs as a class of drugs, a safety signal was detectable in the FAERS database, and RR was significant in the EMR database. For the EMR cohort, 6045 patients were exposed to TNFαIs and 35 cases of melanoma were detected. Significance for RR was detected for adalimumab (RR 1·8, P = 0·02) and etanercept (RR 2·35, P = 0·0004 < 0·001).

CONCLUSIONS

We identified a significant association between exposure to TNFαIs and malignant melanoma in two different analyses. Our findings add to existing evidence linking these agents with the occurrence of malignant melanoma. Additional investigations are required to explore this association further along with the risk of melanoma with TNFαI therapy.

Advancing pharmacovigilance through academic-legal collaboration: the case of gadolinium-based contrast agents and nephrogenic systemic fibrosis-a Research on Adverse Drug Events and Reports (RADAR) report

OBJECTIVE

To compare and contrast three databases, that is, The International Centre for Nephrogenic Systemic Fibrosis Registry (ICNSFR), the Food and Drug Administration Adverse Event Reporting System (FAERS) and a legal data set, through pharmacovigilance and to evaluate international nephrogenic systemic fibrosis (NSF) safety efforts.

METHODS

The Research on Adverse Drug events And Reports methodology was used for assessment-the FAERS (through June 2009), ICNSFR and the legal data set (January 2002 to December 2010). Safety information was obtained from the European Medicines Agency, the Danish Medicine Agency and the Food and Drug Administration.

RESULTS

The FAERS encompassed the largest number (n = 1395) of NSF reports. The ICNSFR contained the most complete (n = 335, 100%) histopathological data. A total of 382 individual biopsy-proven, product-specific NSF cases were analysed from the legal data set. 76.2% (291/382) identified exposure to gadodiamide, of which 67.7% (197/291) were unconfounded. Additionally, 40.1% (153/382) of cases involved gadopentetate dimeglumine, of which 48.4% (74/153) were unconfounded, while gadoversetamide was identified in 7.3% (28/382) of which 28.6% (8/28) were unconfounded. Some cases involved gadobenate dimeglumine or gadoteridol, 5.8% (22/382), all of which were confounded. The mean number of exposures to gadolinium-based contrast agents (GBCAs) was gadodiamide (3), gadopentetate dimeglumine (5) and gadoversetamide (2). Of the 279 unconfounded cases, all involved a linear-structured GBCA. 205 (73.5%) were a non-ionic GBCA while 74 (26.5%) were an ionic GBCA.

CONCLUSION

Clinical and legal databases exhibit unique characteristics that prove complementary in safety evaluations. Use of the legal data set allowed the identification of the most commonly implicated GBCA.

ADVANCES IN KNOWLEDGE

This article is the first to demonstrate explicitly the utility of a legal data set to pharmacovigilance research.

Rituximab-associated hepatitis B virus (HBV) reactivation in lymphoproliferative diseases: meta-analysis and examination of FDA safety reports

BACKGROUND

Rituximab has been associated with hepatitis B virus reactivation (HBV-R). However, the characteristics and scope of this association remain largely undefined.

METHODS

We completed a comprehensive literature search of all published rituximab-associated HBV-R cases and from the Food and Drug Administration (FDA) Adverse Event Reporting System (AERS) MedWatch database. Literature and FDA cases were compared for completeness, and a meta-analysis was completed.

RESULTS

One hundred and eighty-three unique cases of rituximab-associated HBV-R were identified from the literature (n = 27 case reports, n = 156 case series). The time from last rituximab to reactivation was 3 months (range 0-12), although 29% occurred >6 months after last rituximab. Within FDA data (n = 118 cases), there was a strong signal for rituximab-associated HBV-R [proportional reporting ratio = 28.5, 95% confidence interval (CI) 23.9-34.1; Empiric Bayes Geometric Mean = 26.4, 95% CI 21.4-31.1]. However, the completeness of data in FDA reports was significantly inferior compared with literature cases (P < 0.0001). Among HBV core antibody (HBcAb(+)) series, the pooled effect of rituximab-based therapy showed a significantly increased risk of HBV-R compared with nonrituximab-treated patients (odds ratio 5.73, 95% CI 2.01-16.33; Z = 3.33, P = 0.0009) without heterogeneity (χ(2) = 2.12, P = 0.5473).

CONCLUSIONS

The FDA AERS provided strong HBV-R safety signals; however, literature-based cases provided a significantly more complete description. Furthermore, meta-analysis of HBcAb(+) series identified a more than fivefold increased rate of rituximab-associated HBV-R.

Quality not quantity: Key success factors from the first decade of safety reports from the Research on Adverse Drug Events and Reports project (RADAR)

6559

Background: RADAR is the only independent academic pharmacovigilance organization funded exclusively by peer-reviewed grants. We describe the role of high quality case reports in the detection of drug safety signals. Methods: RADAR has identified 11 cancer-related adverse drug reactions (ADRs). Initial reports for small numbers of cases were obtained from our own institution, NU, (4 ADRs) or from referral centers (7 ADRs). Clinicians at these centers voluntarily provided brief case reports to RADAR, who submitted detailed case reports to the FDA/manufacturer. Clinicians were promised that patient/provider data would be kept confidential and that these data would be submitted as peer-reviewed manuscripts. Results: See Table. Conclusions: RADAR was successful at signal generation and amplification because it focused on quality, not quantity of case reports. Pharmacovigilance efforts that allow clinicians to complete brief forms, maintain confidentiality of patient and provider, and result in submission of collaborative manuscripts may improve early detection of drug safety signals initiatives in oncology.

[Table: see text]

No significant financial relationships to disclose.

Life-threatening dermatologic toxicities to cancer drug therapy: An assessment of the published peer-reviewed literature

e20592

Background: Stevens-Johnson syndrome (SJS) and Toxic Epidermal Necrolysis (TEN) lie on opposite ends of a clinical spectrum characterized by skin detachment, mucosal involvement and bullous skin lesions. Currently, reporting is in the range of 1–7 cases of SJS and 0.4–1.2 of TEN/million persons/year, with mortality as high as 35%. The association of cancer drug therapy with SJS/TEN has been uncommonly addressed. Methods: We searched Ovid (incl.1950 through Nov 2008), PubMed (incl.1948 to present) and Litt's Drug Eruption Reference Manual (incl.1980 through Oct 2007) using search terms and combination of terms relevant for SJS/TEN and cancer drug therapy. Only primary case reports were included. Histology reporting was not required for inclusion. Results: Of 448 results, there were 20 reports of SJS associated with 12 FDA-approved drugs and 22 reports of TEN associated with 16 FDA-approved drugs. In addition, all cases (cancer therapy and non-cancer therapy) were obtained from Medwatch (through Mar 2008). ( Table ) Conclusions: Whereas mild-moderate dermatologic toxicities including alopecia, nail abnormalities or rashes are frequent with cancer drug therapies, it is essential for the oncologist to be familiar with life-threatening dermatologic toxicities such as SJS and TEN, and the agents that have been associated with such reactions.

[Table: see text]

[Table: see text]

Complementary systems: Quality and timeliness of passive vs. active adverse event reporting and regulatory vs. peer-reviewed adverse drug reaction (ADR) dissemination

6062

Background: RADAR, a NIH funded R01 project, obtains reports of serious ADRs by conducting hypothesis-driven active surveillance efforts in hematology/oncology. The FDA and pharmaceutical sponsors conduct passive surveillance efforts based on review of reports voluntarily submitted by health care workers. Methods: Completeness and timeliness of ADR reports and dissemination efforts by RADAR versus the FDA/pharmaceutical sponsors were compared. Results: Individual reports were more complete in RADAR databases with fewer total reports (341 versus 1,341). Pharmaceutical sponsors disseminated ADR data 1 year (median) earlier as revised package inserts (PIs), but RADAR publications were more complete. Conclusion: Compared to pharmacovigilance efforts by the FDA/pharmaceutical suppliers, RADAR is more complete but less timely.

[Table: see text]

No significant financial relationships to disclose.

Do post-approval phase III trials for accelerated approved cancer drugs violate equipoise?

6026

Background: Since 1992, the Food and Drug Administration (FDA) has allowed accelerated approval of novel cancer drugs based on improvements in surrogate outcomes, provided that subsequent phase III trials, in compliance with subpart H, show evidence of clinical benefits. However, drugs that receive accelerated approval must already show promise, making it difficult to recruit for randomized studies in which patients might get other drugs which are likely to be inferior. We evaluated whether drugs granted accelerated approval were just as likely to be superior as inferior to standard therapy during phase III clinical trials, a necessary condition known as equipoise, which is used as the ethical basis for recruitment. Methods: Descriptions of marketing approval decisions and subpart H commitments for all drugs that received accelerated approval for oncology indications between 1992 and 2005 were obtained from the FDA website, transcripts of the Oncologic Drug Advisory Committee of the FDA, and PubMed searches. Results: Accelerated approval has been granted for 25 drugs and 29 indications. These approvals have been based on phase II clinical trials (23 indications) or phase III trials (6 indications). 14 approvals were for novel cancer therapeutic drugs. Post-approval phase III clinical trials, outlined in subpart H commitments, have been reported for 9 indications associated with common cancers of the colon, lung, or breast, and 1 indication associated with multiple myeloma, a less common cancer, for which 9 studies identified improved clinical outcomes with the accelerated approved drug. Of 15 drugs that received accelerated approval prior to 2003 for cancers that affect small numbers of patients, 13 are years behind planned recruitment milestones for post-approval phase III trials. Conclusion: While the equipoise theory would predict that 50% of the completed phase III trials would support the novel therapy, empirical data have identified that 90% of the studies required by subpart H commitments support the novel therapy. Therefore, it is likely to hinder recruitment to ongoing phase III trials evaluating other accelerated approved cancer drugs.

No significant financial relationships to disclose.

Variability in the assessment of adverse events in a multicenter clinical trial

BACKGROUND

Consistent documentation, characterization, and evaluation of adverse events (AEs) are needed during multicenter clinical trials to ensure accuracy of data reported to the US Food and Drug Administration and in the medical literature.

OBJECTIVE

The purpose of this study was to identify and characterize variations in the assessment of AEs by clinical trial personnel.

METHODS

During the annual meeting of personnel from a multicenter, controlled clinical trial of an investigational new drug treatment for opioid dependence, an oral presentation of procedures for AE data collection was given to 25 principal investigators and ancillary study personnel who assessed AEs for the study. A post-test using 3 hypothetical AE cases in which AEs were categorized by type of reaction, relatedness to study drug, severity, action taken, and outcome was completed by study participants. Cases and expected responses were reviewed for content and validity by clinical research pharmacists who were not involved with the study. The level of agreement with expected responses was assessed using McNemar symmetry chi-square tests.

RESULTS

Assessments of type of AE, relatedness to study drug, and severity were less frequently aligned with expected responses than were action taken and outcome (P < 0.013). Less consistency with expected responses was found in I case than in the other 2, suggesting that certain types of AEs may be more difficult to assess.

CONCLUSIONS

There was considerable variability in categorization of AEs in an exercise following training for AE data collection. Type of report, relatedness, and severity were found to have more variability in reporting than did action taken or outcome. The results suggest that unless data are gathered to verify reliability of reporting, subcategorization of AE data should be undertaken cautiously. Further research is needed regarding methods for improving consistency in reporting of AEs.

Impact of a formulary change in proton pump inhibitors on health care costs and patients' symptoms

Patients may fail to successfully undergo a switch in therapy associated with a formulary change. The aim of this study was to measure health care costs and outcomes among patients who failed a formulary change in proton pump inhibitors in a VA medical center. Patients who failed a switch from omeprazole to lansoprazole (N = 51) were matched with patients who were successfully switched (N = 51). Health care utilization data was gathered from VA electronic databases and medical records for six months before and after the switch and, for failure patients, during the lansoprazole trial period. Statistical comparisons between failure and success patients were performed on changes in health care costs between these time periods. Health outcome data for the lansoprazole trial period and subsequent omeprazole reinstatement period were obtained through a telephone questionnaire of failure patients. Changes in total health care utilization costs did not differ significantly between failure and success groups for any of the time periods. Failure patients had significantly poorer health outcomes during their lansoprazole trial periods with significantly greater severity of heartburn and severity and frequency of acid regurgitation (P < 0.001). In conclusion, the formulary change had a negative impact upon health outcomes among failure patients but did not significantly affect their health care utilization costs. Identification of failure patients early in their lansoprazole trial periods could improved their health outcomes and satisfaction with medical care.

Computerized system for outcomes-based antiemetic therapy in children

OBJECTIVE

To introduce a computerized data collection system used for an outcomes-based approach to antiemetic therapy in children, and to present data collected with this system in support of a new antiemetic dosing regimen.

METHODS

A validated nausea/vomiting survey was used to collect data on nausea severity (NSEV), vomiting severity (VSEV), daily activity interference (DAI), and the number of vomiting episodes. NSEV, VSEV, and DAI were rated as 0 = none to 3 = severe. All children and/or their parents were surveyed following the completion of each highly emetogenic chemotherapy regimen. This survey was computerized and transferred to a handheld data entry unit. Time and motion studies were performed to compare the time required to collect nausea/vomiting data and transfer the data to a computerized database with the hand-held system versus traditional paper (manual) surveys. The hand-held technology was used to collect survey data for children receiving a new antiemetic regimen (daily ondansetron and dexamethasone [OD]), which was then compared with data obtained with a previously employed regimen (thrice-daily ondansetron and daily methylprednisolone [OM]). Statistical analysis and a cost-effectiveness analysis (CEA) were performed to compare the two antiemetic regimens.

RESULTS

The mean time required for total data entry with the manual system was 5.2 minutes per survey versus 2.4 minutes with the hand-held technology (p = 0.0026). A total of 376 nausea/vomiting surveys in 78 children receiving the OM antiemetic regimen were compared with 153 surveys in 38 children treated with the OD regimen. The mean survey scores were as follows: NSEV (1.2 vs. 0.8), VSEV (1.0 vs. 0.7), DAI (1.0 vs. 0.7), and number of vomiting episodes (4.3 vs. 2.1) for OM and OD, respectively; all were significantly lower with the OD regimen (p < 0.05). The percentage of patients with complete control of nausea and vomiting (19.2% vs. 39.2%) and good control (55.6% vs. 65.4%) were significantly greater with the OD regimen (p < 0.05). The CEA revealed that the OD resulted in a reduction of approximately $31 per patient for good protection and a $258 reduction for complete protection from nausea and vomiting.

CONCLUSIONS

A computerized outcomes-based system aided by handheld technology allowed for more prompt and efficient collection of nausea/vomiting data. The OD antiemetic regimen was shown to be a more cost-effective alternative for children receiving severely emetogenic chemotherapy.

Understanding quality-adjusted life years and their application to pharmacoeconomic research

OBJECTIVE

To provide a basic overview of quality-adjusted life years (QALYs) and their application in cost-effectiveness analysis (CEA), compare and contrast QALYs with other health-related quality-of-life (HRQOL) assessments, describe current controversies regarding QALYs, and provide comparisons between QALY instruments.

METHODS

The literature regarding HRQOL and QALYs was reviewed and key issues are summarized.

RESULTS

QALYs provide relative preferences of patients for different health states. They range from 0, representing death, to 1.0, representing optimal health. QALYs are distinguished from other HRQOL assessments in that they provide a summary measurement that incorporates quantity of life in addition to HRQOL. When QALYs are used as the outcome measure in CEA, a cost per QALY is calculated. The Panel on Cost Effectiveness in Health and Medicine has recommended that QALYs be used as outcome measures in CEA; when QALYs are used in CEA, comparisons between treatments for different illnesses as well as within an illness are possible. The three most commonly used preference measurement techniques in determining QALYs are visual analog scales, time trade-off, and standard gamble. Controversies regarding QALYs include which preference measurement technique is most appropriate, whether QALY assessments should be obtained from patients or the community, and how to address states, such as coma, that individuals sometimes assess as worse than death. QALY instruments can be compared regarding preference measurement technique, HRQOL domains assessed, ease of administration, validity, reliability, and sensitivity.

CONCLUSIONS

When used appropriately, QALYs provide valuable outcome measures for pharmacoeconomic research.

Survey ranking of emetogenic control in children receiving chemotherapy

PURPOSE

To determine the effect of standard antiemetic treatment in children receiving various combination chemotherapy regimens.

METHODS

A validated nausea/vomiting survey was administered to 78 patients receiving 13 different emetogenic chemotherapy regimens. Patients received antiemetic prophylaxis with ondansetron (0.3 mg/kg/day) alone for moderately emetogenic chemotherapy regimens or ondansetron (0.45 mg/kg/day) and methylprednisolone (4 mg/kg/day) for severely emetogenic chemotherapy regimens. A total of 324 different courses of chemotherapy were surveyed. Nausea and vomiting severity, duration, number of emetic episodes, appetite, daily activity interference, and rates of both complete and good antiemetic protection were determined for each chemotherapy protocol. Differences between genders and ages were analyzed.

RESULTS

Chemotherapy combinations containing platinum compounds were found to be highly emetogenic and nauseating despite antiemetic therapy with ondansetron plus a corticosteroid. In addition, complete antiemetic protection for the combination of vincristine, cyclophosphamide, and dactinomycin was poor. For most of the severely emetogenic chemotherapy protocols, patients experienced good protection from nausea and vomiting less than 60% of the time, despite the use of ondansetron plus methylprednisolone. Significant differences were found in rates of residual nausea and vomiting and failure of antiemetic protection among the severely emetogenic chemotherapy protocols despite identical antiemetic therapy. Good protection rates were higher for moderately emetogenic chemotherapy treated with ondansetron alone.

CONCLUSIONS

The currently recommended prophylactic therapy for pediatric patients receiving severely emetogenic chemotherapy fails to provide protection for many patients receiving commonly administered chemotherapy regimens and for most pediatric patients receiving platinum-containing chemotherapy combinations. New and refined antiemetic strategies are needed to improve efficacy in the pediatric population.

The impact of pharmaceutical services in community and ambulatory care settings: evidence and recommendations for future research

OBJECTIVE

To review and evaluate research on pharmaceutical services in community and ambulatory care pharmacy settings, specifically study designs and patient outcome measures, and to provide recommendations to improve future research on pharmaceutical services in community and ambulatory care pharmacy settings.

DATA SOURCE

English-language articles were identified by searching MEDLINE (1966-December 1998) and International Pharmaceutical Abstracts (1970-December 1998), using a combination of search terms: pharmacist services, pharmacist interventions, community pharmacy, ambulatory care, primary care, and patient outcomes. Relevant studies were selected based on article abstracts.

DATA EXTRACTION

From each relevant study, we extracted the study objectives, sample size, study period, study design, major tasks performed by pharmacists, and economic, clinical, and humanistic outcomes (ECHO). Results were tabulated separately for research on community pharmacy and ambulatory care pharmacy settings.

RESULTS

We identified 95 relevant studies. Of these, 21 studies were conducted in community pharmacy settings and 74 in ambulatory care settings. Ten community pharmacy studies used prospective, single group, pretest/posttest, or posttest only designs; seven used prospective two or more group comparison designs; and four used randomized, controlled designs. Nine studies on community pharmacies measured clinical outcomes, two measured humanistic outcomes, and five measured economic outcomes. Four studies measured both clinical and humanistic outcomes and one measured humanistic and economic outcomes. No study measured all three ECHO variables. Twenty-three studies in ambulatory care settings used prospective or retrospective, single group, pretest/posttest or posttest only designs; 21 used prospective or retrospective two-or-more group comparison designs; and 30 used randomized, controlled designs. Thirty-six measured clinical outcomes, five measured humanistic outcomes, and 15 measured economic outcomes. Fifteen studies measured clinical and economic outcomes and three measured clinical and humanistic outcomes.

CONCLUSIONS

Only 21 of 95 selected studies were conducted in community pharmacy settings and measured the impact of pharmaceutical services on patient outcomes. Few studies employed adequate research designs to control threats to internal and external validity. In order to obtain a comprehensive and accurate picture of the impact of pharmaceutical services on patient outcomes, an attempt must be made to measure all three ECHO variables while employing adequate research design.

Thrombotic thrombocytopenic purpura associated with ticlopidine in the setting of coronary artery stents and stroke prevention

BACKGROUND

One of the most unusual causes of thrombotic thrombocytopenic purpura (TTP), a life-threatening disease, is ticlopidine hydrochloride, an antiplatelet agent used to prevent strokes in high-risk populations or following coronary artery stent placement. Recently, Hoffman-LaRoche Pharmaceuticals, following reports of 20 deaths from ticlopidine-associated TTP, updated the information about the hematologic adverse effects of the drug.

OBJECTIVES

To review our recent findings on ticlopidine-associated hematologic toxic effects, which served as the impetus for the revised warnings, and to discuss the implications of these findings.

METHODS

Data were obtained from the Food and Drug Administration's MedWatch program, published phase 3 clinical trials and case reports, hematologists, and plasmapheresis centers.

RESULTS

No cases of TTP have been reported in phase 3 ticlopidine trials. In contrast, postmarketing surveillance has identified serious adverse drug reactions to ticlopidine, resulting in 259 deaths, with TTP accounting for 40 of these deaths. Detailed information was available on 98 cases of ticlopidine-associated TTP. Compared with 42 patients in the coronary artery stent setting, 56 patients with ticlopidine-associated TTP in the stroke prevention setting were more likely to be women (62.5% vs 28.6%; P = .01). Before the onset of TTP in patients receiving stroke prevention therapy and patients with stent placement, ticlopidine had been used for less than 2 weeks in 5.4% and 2.4%, between 2 and 3 weeks in 17.9% and 21.4%, between 3 and 4 weeks in 30.4% and 38.1%, and between 4 and 12 weeks in 46.4% and 38.1%, respectively. Death occurred in almost 60% of all patients not receiving plasmapheresis compared with 21.9% of patients receiving plasmapheresis for stroke prevention and 14.3% of patients receiving plasmapheresis in the stent setting.

CONCLUSIONS

Use of ticlopidine requires frequent physician visits and laboratory tests for at least 3 months in the stroke prevention setting, while, with short-term use in the coronary artery stent setting, adverse events are less likely to occur. These factors, as well as competition from clopidogrel bisulfate, a new antiplatelet agent, potentially limit the feasibility of ticlopidine as a stroke prevention agent, while having less impact on its use following coronary artery stent placement.

Using feedback letters to influence the use of antiulcer agents in a Medicaid program

OBJECTIVE

To determine the impact of printed patient-specific feedback regarding potential misprescribing of antiulcer agents (AUAs). Measures of impact included improvements in patients' dispensing profiles, assessed according to predetermined criteria, and decreases in cost and quantity of AUAs dispensed.

DESIGN

Controlled study. After evaluation for compliance with predetermined criteria, prescribers identified as having one or two patient profiles with potential errors were assigned alternatively to control or experimental groups. An intervention was mailed to the experimental group.

SETTING

Outpatient setting in the New Mexico Medicaid population.

PARTICIPANTS

Patients and prescribers identified as having potential misprescribing of AUAs.

INTERVENTION

The intervention consisted of a cover letter describing the purpose of the drug utilization review program, an educational fact sheet regarding prescribing AUAs, patient profiles with potential misprescribing, and physician response forms.

MEASUREMENTS AND MAIN RESULTS

There were greater improvements in dispensing to patients in the intervention group (chi2, p <.001). Significant odds ratios for the intervention group were 2.29 for AUAs discontinued, 1.98 for all improvements combined, 13.13 for improvement in listing of proper diagnosis for AUAs, and 2.84 for appropriate indication when prescribing the higher acute daily dosage. Using data from 3 months before and after the intervention, we found greater decreases in mean monthly costs (p =.044) and mean monthly quantity of AUAs dispensed (p =.049) in the intervention group.

CONCLUSIONS

This intervention significantly decreased AUA dispensing to patients whose prescribers were mailed the patient-specific feedback intervention.

Effect of a pharmacist-education initiative on ketorolac use and costs in a Medicaid program

The effects of a pharmacist-education initiative on the use and costs of ketorolac in a state Medicaid program are reviewed. An intervention letter describing changes in the manufacturer's prescribing guidelines for ketorolac and providing suggestions for interacting with physicians regarding the use of ketorolac was sent to 150 of the 301 pharmacies that participate in New Mexico's Medicaid program. The remaining 151 pharmacies served as a control group. Ketorolac claims records for three months before and after the intervention were reviewed. The mean quantity of ketorolac tablets, total days' supply, and number of prescriptions filled per pharmacy per month were calculated for both periods. The number of prescriptions not filled as a result of the intervention as well as the number that could have been avoided, the number of cases of peptic ulceration (ketorolac's major adverse effect) that would be avoided, and the associated cost savings if all the state's Medicaid pharmacies had been included in the intervention were estimated. A total of 167 pharmacies (90 intervention and 77 control) dispensed ketorolac for Medicaid patients during the study period. Ketorolac dispensing rates declined during the postintervention period in both the intervention group and the control group, but the reduction was greater in the intervention group. It was predicted that if all pharmacies were included in an intervention, 135.6 fewer prescriptions for ketorolac would be filled each year; as a result, 1.14 cases of peptic ulceration would be avoided and net Medicaid costs would be reduced by $1638. Sending educational letters to pharmacists was associated with a modest reduction in ketorolac use in a state Medicaid program; a net reduction in Medicaid costs if the intervention were extended to all pharmacies that participate in the state's Medicaid program was projected.

Assessment of the emetogenic potential of intrathecal chemotherapy and response to prophylactic treatment with ondansetron

The purpose of this study was to document the emetogenic potential of intrathecal chemotherapy (IC) in children and to evaluate the efficacy of ondansetron in reducing nausea and vomiting with this chemotherapy treatment. Patients less than 18 years of age with acute lymphoblastic leukemia were eligible to participate in a survey project measuring the emetogenic potential of various chemotherapy treatments. Patients surveyed for 1 or more IC treatments were included in this report. The IC consisted of methotrexate, hydrocortisone and cytarabine, dosed according to patient age. A nausea/vomiting survey instrument was completed by each patient and/or parent following IC treatment. The instrument rated nausea, vomiting and daily activity interference (DAI) on a 4-point scale of 0 = none, 1 = mild, 2 = moderate and 3 = severe, and collected data on the number of vomiting and/or retching episodes in addition to the child's appetite following the chemotherapy treatment. When ondansetron was employed, it was administered in an i.v. infusion at a dose of 0.15 mg/kg before and after chemotherapy or as an oral dose of 4 mg or 8 mg before chemotherapy. Courses of IC without antiemetics were analyzed to determine the emetogenic potential of IC. For patients receiving IC both with and without ondansetron, courses were compared with each patient used as their own control to determine the influence of ondansetron upon survey responses. Statistical analysis consisted of nonparametric Friedman 2-way ANOVA for ordinal variables and a paired t-test for continuous variables. The binomial test was employed to analyze for differences between ondansetron and no antiemetic in the number of patients with complete control of both nausea and vomiting or vomiting alone. A total of 63 children with a mean age of 7.6 +/- 4.2 years were each studied on one or more occasions. Thirty-seven children were surveyed for 87 IC treatments without antiemetics (group I), and 17 children from this group were surveyed for 48 IC courses with i.v. ondansetron (group IA). An additional 18 children were subsequently surveyed for 39 IC courses with i.v. ondansetron (group II). Fifteen patients (7 of whom were members of group I) were surveyed following 33 IC courses with oral ondansetron (group III). The survey scores for group I patients were: nausea severity 1.3 +/- 1.1, vomiting severity 1.2 +/- 1.1, DAI 1.2 +/- 1.0 and mean number of emetic episodes 4.7 +/- 8.4. The mean appetite score was 1.5 +/- 1.1. For patients in group IA, nausea severity (0.8 +/- 0.9), vomiting severity (0.5 +/- 0.8), DAI (0.7 +/- 0.8), and the number of emetic episodes (1.4 +/- 2.8) were all significantly lower than with prior IC treatments without ondansetron. For complete protection, children receiving i.v. ondansetron had greater complete protection rates from both nausea and vomiting or vomiting alone than did patients receiving no antiemetic. Survey responses were also lower for patients receiving oral ondansetron, but insufficient control data did not allow for statistical analysis. IC results in mild to moderate nausea and vomiting in children. The emetogenic potential of IC is significantly reduced by i.v. ondansetron.

Differences among raters evaluating the success of EMLA cream in alleviating procedure-related pain in children with cancer

We evaluated the analgesic efficacy of EMLA cream after repeated bone marrow aspirations or lumbar punctures (LPs) in children with cancer, and compared the ratings among patients, their parents, physicians, and nurses. Data from LPs were analyzed at the last procedure without EMLA (T1) and the first and last procedures with EMLA (T2 and T3). Friedman's nonparametric analysis of variance was used for statistical analysis. A total of 272 procedures in 29 children were analyzed. For 179 procedures without EMLA, physicians rated pain lower than other raters, and for the 93 with EMLA physicians rated pain less than the children. Children rated pain at T2 lower than at T1 or T3. Physicians rated pain at T2 less than at T3. Both children and physicians rated pain at T3 as not different from that at T1. No differences were noted at these time points for other raters in LP distress ratings, or in bone marrow aspiration pain or distress ratings. Thus EMLA was associated with decreased pain ratings for LPs, but this effect was not sustainable with repeated procedures. The cream alone should not be relied on to control pain of bone marrow aspiration or repeated LPs in children. Physicians underestimated pain, which may have implications for undertreatment in this patient population.

Economic impact with home delivery of chemotherapy to pediatric oncology patients

OBJECTIVE

To examine the economic impact of a home chemotherapy program (HCP) for pediatric oncology patients.

RATIONALE

Factors that led to initiation of an HCP included availability of specially trained nurses and programmable ambulatory infusion devices at local home care agencies, routine central venous catheter placement, inpatient bed space shortages, and the availability of ondansetron.

SETTING

Chemotherapy delivery in the home setting from June 1991 through June 1994.

DESIGN

Charge data and nausea and vomiting severity data were collected for patients treated through the HCP.

METHODS

Economic impact was calculated by incorporating and summing all charge categories associated with hospital admission for chemotherapy (HAC) versus delivery by the HCP. All data were adjusted for 1993 dollars, and reflect changes for the average patient size (1 m2). Charge data for each chemotherapy protocol delivered in the home were analyzed by calculating the differences between HAC and HCP charges using the following formula: charge difference (HAC - HCP) per protocol times the number of courses. Total economic impact was calculated by summing the differences in charges for each protocol.

RESULTS

A total of 262 chemotherapy courses were given to 44 patients (mean age 9.5 +/- 5.1 y) through the HCP, which represented 1012 patient care days and 24 different chemotherapy protocols. Monetary savings from the HCP ranged from $5180 per course of ifosfamide plus etoposide to $367 per course for high-dose methotrexate. Total monetary savings from the HCP during the 3-year period was $640,793. Successful control of nausea and vomiting with a combination of ondansetron plus methylprednisolone was achieved in approximately 80% of the patients receiving highly emetogenic chemotherapy protocols.

CONCLUSIONS

HCP for pediatric oncology patients results in substantial monetary savings to payors. Effective control of nausea and vomiting can be accomplished at home in the majority of patients with an ondansetron-based antiemetic regimen.

Assessment of chemotherapy-induced emesis and evaluation of a reduced-dose intravenous ondansetron regimen in pediatric outpatients with leukemia

OBJECTIVE

To measure the severity of nausea and vomiting in pediatric patients receiving intravenous or intrathecal chemotherapy for acute lymphoblastic leukemia and to evaluate the effectiveness of 2 intravenous doses of ondansetron for this condition.

DESIGN

Patients were surveyed during repeated treatments of maintenance chemotherapy, given with or without ondansetron, using a repeated measures pretest/posttest design.

SETTING

Outpatient pediatric oncology clinic.

PATIENT POPULATION

Sixteen pediatric patients (aged 2-15 years, mean 6.2) with acute lymphoblastic leukemia.

METHODS

Surveys to assess nausea and vomiting and the extent of interference with daily activities were administered following emetogenic chemotherapy with or without ondansetron.

RESULTS

A total of 255 surveys following emetogenic chemotherapy with daunorubicin, cyclophosphamide, carmustine, and etoposide and cytarabine combined, as well as intrathecal therapy with methotrexate, hydrocortisone, and cytarabine, were analyzed. Analysis was performed on surveys of 149 courses without antiemetic therapy and 106 courses after 2 doses of ondansetron 0.15 mg/kg iv. The most emetogenic chemotherapy treatment was the etoposide/cytarabine combination (p < 0.05). Ondansetron completely protected patients (defined as no nausea or no vomiting) during most (> 50%) of the chemotherapy treatments, except for those in which cyclophosphamide was used. Ondansetron provided greater control of nausea and vomiting, a higher percentage of complete protection, and decreased the daily activity interference rating for carmustine and etoposide/cytarabine compared with courses of chemotherapy without antiemetics (p < 0.05). Two intravenous doses of ondansetron also provided durable antiemetic efficacy over time for the most emetogenic chemotherapy treatment (etoposide/cytarabine).

CONCLUSIONS

Etoposide/cytarabine proved to be the most emetogenic of the chemotherapy treatments studied. A reduced-dose regimen of intravenous ondansetron was shown to be an effective antiemetic for the outpatient treatments with etoposide/cytarabine and carmustine, but not with cyclophosphamide.

Barriers to providing cognitive services

Seventy-three pharmacists were surveyed to evaluate their perceptions of barriers to performing cognitive services in community pharmacy; 64 responded. Perceived barriers to counseling patients were excessive workload, lack of privacy, patient attitudes, and store layout. Perceived barriers to interacting with physicians were difficulties contacting them, negative physician attitudes toward pharmacists' recommendations, excessive workload, and inadequate patient information. Rates of interaction with prescribers correlated with greater job satisfaction; rates of patient counseling did not. Methods to address these barriers include structuring reimbursement and workload around cognitive services rather than drug product dispensing, redesigning pharmacies to provide counseling areas, changing attitudes and abilities through pharmacist education, and using technology to promote information transfer between pharmacies and health care facilities.

Patient counseling in community pharmacy and its relationship with prescription payment methods and practice settings

OBJECTIVE

To determine whether patient counseling activities in community pharmacies are related to payment method and practice setting.

DESIGN

Randomly selected pharmacists voluntarily performed self-data collection (39 percent participation rate) of counseling activities for 40 hours. Observational data were also collected by trained observers.

SETTING

Community pharmacies.

RESULTS

Forty-seven pharmacists in national chain pharmacies and 26 in independent pharmacies collected data on 3766 counseling events. Patient counseling rates, defined as percentage of patients counseled per prescription per pharmacist, were significantly higher (p < 0.05) for self-pay (median 12.9 percent) and Medicaid (median 13.0 percent) than for capitation patients (median 7.9 percent). Rates were somewhat higher (p = 0.06) for chain (median 17.7 percent) than for independent pharmacists (10.7 percent). Although there was no difference in terms of pharmacist-initiated counseling, chain pharmacists had higher rates of patient-initiated counseling per prescription (median 4.3 percent) than independent pharmacists (median 2.7 percent) (p < 0.01). Counseling information was related to payment method and practice setting. A significantly higher percentage (p < 0.01) of counseling events for capitation patients pertained to chronic therapies (33.6 percent) compared with counseling events involving patients using other payment methods (21.8-24.8 percent). A strong correlation between observational data and self-reported data was found (Pearson r = 0.696, p < 0.001).

CONCLUSIONS

The finding that pharmacists counsel capitation patients less frequently may be attributable to financial incentives or to pharmacists' perception that these patients do not need to be counseled. Independent pharmacists' higher number of capitation patients and increased workloads may have affected their counseling activities. These findings have implications for pharmacy service reimbursement programs, practice conditions, and continuing education programs.

Relationships among prescription payment methods and interactions between community pharmacists and prescribers

OBJECTIVE

This research was performed to examine community pharmacists' interactions with prescribers and to determine if these interactions are related to payment method.

DESIGN

Randomly selected pharmacists (47 in chain pharmacies and 26 in independent pharmacies) collected data concerning prescriber interactions for a 40-hour period. These interactions were analyzed in terms of payment methods, prescriber acceptance, and types of information discussed.

RESULTS

Information concerning 730 interactions by 72 pharmacists was obtained. Payment methods were related to the frequency of interactions per pharmacist (p less than 0.01). There were higher percentages of interactions for self-pay (median 2.2 percent) and Medicaid (median 1.8 percent) prescriptions than for third-party fee-for-service (mode 0 percent) or capitation prescriptions (mode 0 percent). Type of information discussed was related to payment method. The vast majority of information provided by pharmacists (91 percent) was accepted by prescribers. The rate of acceptance was 97 percent for prescriber-initiated interactions versus 88 percent for pharmacist- or patient-initiated interactions (p less than 0.01).

CONCLUSIONS

Relationships between payment method and interactions were identified. These findings may be attributable to prescribing policies and reimbursement policies. Prescribing policies that restrict prescribers to a formulary may help make them become more adept at using those products; thus, they will make fewer prescribing errors. Reimbursement policies that require patients to consistently use a specific pharmacy (i.e., capitation) may help pharmacists become more familiar with the patient's prescription history. Therefore, interactions with prescribers are needed less frequently for these patients' prescriptions.

Availability of narcotics and pharmacists' attitudes toward narcotic prescriptions for cancer patients

OBJECTIVE

To determine narcotic availability and pharmacist apprehension toward dispensing narcotics for cancer patients.

DESIGN

Mailing of a one-page survey.

SETTING

All pharmacies in the state of New Mexico.

PARTICIPANTS

Of the 309 pharmacies, 76.1 percent (235) completed the survey.

MEASUREMENTS AND MAIN RESULTS

Apprehension was significantly elevated for methadone, hydromorphone, and for morphine doses greater than 500 mg/d (p less than 0.001). The most common reasons listed for apprehension were forgery (46.4 percent), theft (40.4 percent), high dosages (23.8 percent), narcotic investigations (18.7 percent), and patient addiction (9.4 percent). Hydromorphone was available (in stock) in 80.5 percent of the pharmacies, followed by morphine 30-mg sustained-release tablets (64.5 percent) and methadone (53.4 percent). The vast majority of pharmacists were willing to order strong narcotics for specific patients. Pharmacists working in chain stores expressed a higher level of apprehension toward dispensing methadone and more frequently cited forgery as a reason for apprehension. Distance from a metropolitan area and practice setting best predicted apprehension to forgery (p = 0.01).

CONCLUSIONS

Pharmacists do not appear to be a major obstruction to adequate analgesia for cancer patients in New Mexico but may require further education regarding lack of narcotic addiction and dosing in patients with cancer.

Evaluation of a non-Food and Drug Administration-approved use of cimetidine: treatment of pruritus resulting from epidural morphine analgesia

The purpose of this study was to evaluate the efficacy of a frequently used treatment at our hospital, intravenous cimetidine hydrochloride for pruritus resulting from epidural morphine sulfate analgesia in cesarean-section patients. Fifty-two patients were randomly assigned to receive either placebo or cimetidine 300 mg i.v. Thirty-nine patients experienced itching. Thirty-three patients completed the study (17 in the cimetidine group, 16 in the placebo group). For pruritus unrelieved by the study drug, bolus doses of naloxone hydrochloride 0.2 mg i.v. were administered at 30-minute intervals, if needed. Pruritus was assessed by two subjective rating scales and the number of naloxone doses needed for unrelieved pruritus. Using independent Student's t-tests, we identified no significant differences between the two groups. The mean numbers of naloxone boluses administered were 1.9 for placebo and 2.5 for cimetidine (p = 0.49). Our study demonstrates the importance of objectively evaluating the use of drugs for non-Food and Drug Administration-approved indications.

Improving antiulcer agent prescribing in a health maintenance organization

A study was made of the effect of one-to-one educational meetings between prescribers and a pharmacist on the prescribing of antiulcer agents for outpatients in a health maintenance organization (HMO). Ten-minute presentations were developed that communicated the same basic facts through the use of case studies (vivid interventions) or statistical data (nonvivid interventions). After a control group (n = 8) was selected, prescribers (n = 16) were randomly assigned to receive vivid interventions or nonvivid interventions. Data on the prescribing of cimetidine, ranitidine, and sucralfate were collected for one month before the interventions and for two months afterward. Three clinical pharmacy professors independently evaluated prescriptions for appropriateness of indication, dosage, and duration. No differences in appropriateness were found between the two intervention groups, but in the first postintervention month the mean rate of inappropriate prescribing per control practitioner was 80%, versus less than 32% for the intervention groups (p less than 0.01). Each prescription in the first postintervention month entailed a mean cost of $31 per control practitioner for inappropriate prescribing, compared with less than $12 for the intervention groups (p less than 0.01). Mean costs of inappropriate prescriptions per practitioner per patient visit were $0.88 and less than or equal to $0.41 for the control and intervention groups, respectively (p less than 0.05). During postvisit month 2, inappropriate prescribing for both intervention groups increased slightly and was no longer significantly less than that in the control group. One-to-one educational meetings improved the prescribing of antiulcer agents for outpatients in an HMO.

A model of methods for influencing prescribing: Part II. A review of educational methods, theories of human inference, and delineation of the model

The purpose of this literature review is to develop a model of methods used to influence prescribing. This article reviews well-controlled studies of educational efforts, describes and relates theories of human inference to these studies, and delineates a model of methods for influencing prescribing. The model can be used to help explain the success or failure of various programs and as a framework for developing and testing methods to influence prescribing.

A model of methods for influencing prescribing: Part I. A review of prescribing models, persuasion theories, and administrative and educational methods

The purpose of this literature review is to develop a model of methods to be used to influence prescribing. Four bodies of literature were identified as being important for developing the model: (1) Theoretical prescribing models furnish information concerning factors that affect prescribing and how prescribing decisions are made. (2) Theories of persuasion provide insight into important components of educational communications. (3) Research articles of programs to improve prescribing identify types of programs that have been found to be successful. (4) Theories of human inference describe how judgments are formulated and identify errors in judgment that can play a role in prescribing. This review is presented in two parts. This article reviews prescribing models, theories of persuasion, studies of administrative programs to control prescribing, and sub-optimally designed studies of educational efforts to influence drug prescribing.

Pharmacy participation in third-party contracts: decision making through economic and financial analysis

Third-party prescription coverage is becoming an important facet of pharmacy practice. Pharmacy managers must be able to evaluate contracts and to decide which contracts are more profitable. This paper presents a methodology for evaluating and comparing third-party contracts. An example demonstrates the proposed methodology, and a specific contract is identified as more advantageous.

Association of length of stay and total hospital charges with antimicrobial regimen changes

The association of changes in antimicrobial therapy with length of stay and total hospital charges in a 600-bed general medical hospital was studied. Changes in antimicrobial regimens were examined in three diagnosis-related groups (DRGs): DRG 79 (respiratory infections and inflammations, age greater than 69 years, and/or secondary diagnosis), DRG 89 (simple pneumonia and pleurisy, age greater than 69 years, and/or secondary diagnosis), and DRG 416 (septicemia, age greater than 17 years). These changes were defined as follows: (1) switching from one drug or drug combination to another, (2) adding one or more drugs to a regimen, or (3) discontinuing one drug of a multiple-drug regimen. Variables examined for association with these changes were length of hospital stay and pharmacy, antimicrobial, intravenous therapy, and hospital charges. In DRGs 79 and 89, changes in antimicrobial regimens were associated with significantly longer hospital stay and significantly higher pharmacy, antimicrobial, intravenous therapy, and total hospital charges. In DRG 416, changes in antimicrobial regimens were not associated with higher charges or longer stay. Significant differences among DRGs were found with respect to the types of changes, with more changes from broad- to narrow-spectrum coverage occurring in DRGs 89 and 416 than in DRG 79. At this institution, patients whose antimicrobial regimens were changed had significantly longer hospital stays and higher charges in two of three DRGs.

The effects of DRGs and cost-containment: assessment by hospital pharmacy directors

With the advent of prospective reimbursement, cost containment has increasingly come into focus. Arizona hospital pharmacy directors were surveyed to learn what effect DRG-based payments and other capitation programs have had on their pharmacy department and hospital. Their responses concerning effects on staffing and services are outlined. Descriptions of successful and unsuccessful cost-containing measures are provided, and pharmacists' perceptions of the effect of DRGs and the problems they see with cost containment are detailed. It is evident that P & T Committees will play an ever-increasing role as an important interface with hospital staff in resolving many cost-containment issues.