Real-world study of patients with locally advanced HNSCC in the community oncology setting.

Introduction: There is a need to understand the current treatment landscape for LA HNSCC in the real-world setting. Methods: This retrospective study assessed real-world outcomes and treatment patterns of 1,158 adult patients diagnosed with locally advanced (stage III-IVB) HNSCC initiating chemoradiotherapy (CRT) within the period January 2015 to December 2017 in a large network of US community oncology practices. Structured data were abstracted from electronic health records. Demographic, clinical and treatment characteristics were analyzed descriptively overall and stratified by index treatment (cisplatin + radiotherapy [RT], cisplatin + other chemotherapy + RT, or cetuximab + RT). Time to next treatment (TTNT) and overall survival (OS) were measured using the Kaplan-Meier method, and median duration of treatment was assessed. OS was compared across treatment cohorts using multinomial logistic regression with inverse probability treatment weighting. To identify covariates associated with OS, a multivariable adjusted Cox proportional hazard model was used. Results: This study examined 22,782 records, of which 2124 had stage III to stage IVB and no other cancers, and 1158 met all eligibility criteria. Among the treatment cohorts analyzed (cisplatin + RT, cisplatin + other chemotherapy + RT, or cetuximab + RT), cisplatin + RT was the most common concurrent chemotherapy (65.8%). Among 1158 patients, 838 (72.4%) did not initiate subsequent treatment and 139 (12.0%) died. The median TTNT and median OS were only reached by the cetuximab + RT cohort. Among patients with oropharynx primary tumor location, patients with human papilloma virus (HPV) positive status had the longest time on treatment and highest survival at 60 months. Covariates associated with improved survival were never/former tobacco use, HPV positive status, and overweight or obese body mass index. Covariates associated with poorer survival were age of 60+ years, primary tumor location of hypopharynx or oral cavity and Eastern Cooperative Oncology Group performance status score of 2+. Conclusion: These data describe real-world treatment patterns in locally advanced head and neck squamous cell cancer and sets the baseline to assess outcomes for future studies on the community oncology population.

Real-world relationship of early end points to survival end points in patients with resectable non-small-cell lung cancer.

Aim: Pathologic response has been shown to be a promising surrogate for survival in non-small-cell lung cancer. We examined the real-world relationship between these end  points in patients with resectable stage IB-IIIA non-small-cell lung cancer receiving neoadjuvant chemotherapy/chemoradiotherapy (CT/CRT). Methods: Electronic health records/medical charts were analyzed. Overall and event-free survival (OS/EFS) were assessed by Kaplan-Meier stratified by pathologic response. Associations between the end  points were assessed by Cox analyses. Results: A total of 425 patients were selected for the study; 147 and 278 received CT and CRT, respectively. Pathologic complete response (pCR) was associated with longer OS (adjusted HR = 0.50; 95% CI: 0.29-0.85) and EFS (adjusted HR = 0.44; 95% CI: 0.28-0.68) versus no pCR, and EFS was associated with OS (HR = 0.51, 95% CI: 0.38, 0.69). Conclusion: In patients receiving neoadjuvant CT/CRT, pCR and EFS were associated with improved survival in this real-world dataset.

Retrospective Observational Study of Outcomes in HER2-Positive Metastatic Breast Cancer (mBC) Patients Treated with Ado-Trastuzumab Emtansine (T-DM1) and Subsequent Treatments After T-DM1 in the United States.

BACKGROUND: Limited evidence exists on real-world outcomes with ado-trastuzumab emtansine (T-DM1) treatment and the effectiveness of subsequent therapies. OBJECTIVE: This study evaluated treatment patterns and outcomes of patients treated with T-DM1 and post-T-DM1 therapy in the United States. PATIENTS AND METHODS: Adult patients with HER2-positive (HER2+) metastatic breast cancer (mBC) initiating treatment with T-DM1 between 1/1/2013 and 9/30/2018 were included and followed through 12/31/2018. Data were obtained from the iKnowMed electronic health record. Demographic, clinical, and pre- and post-T-DM1 treatment characteristics were described. The Kaplan-Meier method was used to estimate time to treatment discontinuation (TTD) and overall survival (OS). RESULTS: Of 318 patients treated with T-DM1, 184 (57.9%) had prior treatment with pertuzumab. The median age was 58 years. Most patients had visceral disease (93.4%), and 62.3% had two or more prior treatments for mBC before T-DM1 (range 0-9). The most common subsequent regimens were trastuzumab + vinorelbine (22.5%), HER2-targeted monotherapy (22.5%), and trastuzumab + other chemotherapy (19.6%). Median TTD with T-DM1 was 5.9 months (95% confidence interval [CI] 4.6-6.9); median OS from the start of T-DM1 therapy was 19.2 months (95% CI 16.8-24.5). CONCLUSIONS: Patients treated with T-DM1 in this study appeared to have more advanced disease than patients in clinical trials and were treated in later lines of therapy. Variability was observed across subsequent therapy selections. Treatment patterns and outcomes appeared comparable for patients who received prior pertuzumab. The short treatment durations and survival with T-DM1 therapy in the real-world setting underscore the need for effective post-trastuzumab therapies.

Treatment Patterns and Healthcare Resource Utilization among Patients with Advanced or Metastatic Soft Tissue Sarcoma in US Community Practices.

INTRODUCTION: This study was designed to describe demographic and clinical characteristics of patients diagnosed with advanced or metastatic soft tissue sarcoma (STS) and to examine treatment and healthcare resource utilization patterns of this patient population in a United States (US) community-based oncology practice setting over time. METHODS AND MATERIALS: A retrospective observational study was conducted within the US Oncology Network (USON). Patients were eligible if they were diagnosed with advanced or metastatic STS and were treated at a USON site between 01 July 2015 and 31 August 2018. Demographic, clinical, and treatment characteristics were described for the overall study population. Comparisons between patients by time period (prior to and after October 2016) were evaluated using the T test for continuous variables and chi-squared test for categorical variables. Data were available for analysis through 31 August 2018. RESULTS: Demographic and clinical characteristics of the eligible study cohort (N = 376) were similar between patients who initiated treatment before and after October 2016 (all p > 0.05). Forty-three unique regimens were observed in the first-line setting, with the predominant regimen (gemcitabine + docetaxel) received by 33.2% (n = 125) patients. Prior to October 2016, 45.4% of patients received first-line gemcitabine + docetaxel, while 29.0% received this regimen after October 2016. CONCLUSIONS: While demographic and clinical characteristics were similar, treatment patterns changed in 2016. Future research should evaluate the impact of changing drug approvals and clinical trial results on future treatment patterns.

An evaluation of the THIN database in the OMOP Common Data Model for active drug safety surveillance.

BACKGROUND: There has been increased interest in using multiple observational databases to understand the safety profile of medical products during the postmarketing period. However, it is challenging to perform analyses across these heterogeneous data sources. The Observational Medical Outcome Partnership (OMOP) provides a Common Data Model (CDM) for organizing and standardizing databases. OMOP's work with the CDM has primarily focused on US databases. As a participant in the OMOP Extended Consortium, we implemented the OMOP CDM on the UK Electronic Healthcare Record database-The Health Improvement Network (THIN). OBJECTIVE: The aim of the study was to evaluate the implementation of the THIN database in the OMOP CDM and explore its use for active drug safety surveillance. METHODS: Following the OMOP CDM specification, the raw THIN database was mapped into a CDM THIN database. Ten Drugs of Interest (DOI) and nine Health Outcomes of Interest (HOI), defined and focused by the OMOP, were created using the CDM THIN database. Quantitative comparison of raw THIN to CDM THIN was performed by execution and analysis of OMOP standardized reports and additional analyses. The practical value of CDM THIN for drug safety and pharmacoepidemiological research was assessed by implementing three analysis methods: Proportional Reporting Ratio (PRR), Univariate Self-Case Control Series (USCCS) and High-Dimensional Propensity Score (HDPS). A published study using raw THIN data was selected to examine the external validity of CDM THIN. RESULTS: Overall demographic characteristics were the same in both databases. Mapping medical and drug codes into the OMOP terminology dictionary was incomplete: 25 % medical codes and 55 % drug codes in raw THIN were not listed in the OMOP terminology dictionary, representing 6 % condition occurrence counts, 4 % procedure occurrence counts and 7 % drug exposure counts in raw THIN. Seven DOIs had <0.3 % and three DOIs had 1 % of unmapped drug exposure counts; each HOI had at least one definition with no or minimal (≤0.2 %) issues with unmapped condition occurrence counts, except for the upper gastrointestinal (UGI) ulcer hospitalization cohort. The application of PRR, USCCS and HDPS found, respectively, a sensitivity of 67, 78 and 50 %, and a specificity of 68, 59 and 76 %, suggesting that safety issues defined as known by the OMOP could be identified in CDM THIN, with imperfect performance. Similar PRR scores were produced using both CDM THIN and raw THIN, while the execution time was twice as fast on CDM THIN. There was close replication of demographic distribution, death rate and prescription pattern and trend in the published study population and the cohort of CDM THIN. CONCLUSIONS: This research demonstrated that information loss due to incomplete mapping of medical and drug codes as well as data structure in the current CDM THIN limits its use for all possible epidemiological evaluation studies. Current HOIs and DOIs predefined by the OMOP were constructed with minimal loss of information and can be used for active surveillance methodological research. The OMOP CDM THIN can be a valuable tool for multiple aspects of pharmacoepidemiological research when the unique features of UK Electronic Health Records are incorporated in the OMOP library.

Renal impairment, hemoglobinuria, and hemoglobinemia among patients with idiopathic thrombocytopenic purpura.

Renal impairment (RI) and events potentially leading to RI were reported in idiopathic thrombocytopenic purpura (ITP) patients with specific medications. This study was conducted to estimate the incidence rate (IR) of RI, hemoglobinuria and hemoglobinemia (HE) and characterize baseline risk factors in ITP and ITP-free patients. Incident ITP and matched non-ITP patients were identified from an electronic medical record database from 1990 to 2002. ITP patients were classified by the treatment first received (initiators) or ever received (users). All cohorts were followed for study outcomes. IRs were calculated and standardized by age and gender. A total of 881 ITP and 4,496 ITP-free patients yielded 3,044 and 16,006 person-years, respectively. The ITP cohort had a slightly higher prevalence of autoimmune diseases and infections than the ITP-free cohort. The IR (/10,000 person-years) for RI, hemoglobinuria and HE was 14.2, 35.7, and 7.1 in the ITP cohort; 10.0, 48.8, and 0 in the ITP-free cohort; and 18.3, 37.1, and 6.1 in untreated ITP patients, respectively. The risk of RI, HE or hemoglobinuria was not found to differ substantially between ITP and non-ITP patients or across ITP treatments.

Association between extended-release niacin treatment and glycemic control in patients with type 2 diabetes mellitus: analysis of an administrative-claims database.

The aim of the study was to evaluate trends in antihyperglycemic agents (AHAs) use in patients with type 2 diabetes mellitus (T2DM) newly initiating extended-release niacin (ERN) compared with other lipid-modifying therapy (LMT). United States administrative-claims data identified adults with T2DM on AHAs who received a new prescription for ERN or another LMT between January 2001 and June 2003 (index date), and these adults were followed for 12 months. Inclusion criteria were (1) stable T2DM as defined by International Classification of Diseases, Ninth Revision, codes and also receiving at least 2 AHA prescriptions 12 to 24 months before initiating ERN or LMT treatment and (2) at least 2 prescriptions within 12 months before the onset of ERN or LMT. Trends in AHA prescriptions 12 months before (baseline) and after (follow-up) index date were defined as (1) no change (ie, stable T2DM), (2) increased (ie, worsening T2DM), or (3) reduced (ie, improved T2DM). Among 3799 patients with T2DM, 392 (10.3%) were treated with ERN and 3407 (89.7%) were treated with other LMT. In the ERN cohort, 82.1% of patients experienced no change in AHA prescriptions between baseline and follow-up compared with 79.4% of patients in the LMT cohort (P = .20); 13% of the ERN cohort and 16% of the LMT cohort (P = .17) experienced a dose increase or the addition of another AHA; and 5% of both cohorts were prescribed fewer AHAs or switched to a lower dose (P = .92). Treatment with ERN (vs other types of LMT) did not significantly increase AHA use, implying that T2DM status did not worsen in this cohort.

Loss of early gains in low-density lipoprotein cholesterol goal attainment among high-risk patients.

BACKGROUND: Guidelines recommend a low-density lipoprotein cholesterol (LDL-C) measurement of <70 mg/dL as a reasonable goal in high-risk patients with coronary heart disease (CHD) or atherosclerotic vascular disease (AVD). METHODS: This retrospective, cross-sectional study examined LDL-C goal attainment monthly trends from January 1, 2004, to August 31, 2008, in a large, managed-care claims database in the United States. High-risk CHD or AVD patients who had at least one LDL-C test during that time period were included (N = 284,915). Average LDL-C values and percent of patients not achieving LDL-C goal (LDL-C ≥70 mg/dL) were obtained by averaging patient level LDL-C values for each month. A linear trend analysis with first-order autocorrelated errors was conducted. RESULTS: The proportion of patients treated with lipid-lowering therapy gradually increased from 58.5% in 2004 to 70.5% in 2008. Mean LDL-C values in patients treated with lipid-lowering therapy decreased from 100.4 to 96.4 mg/dL, whereas LDL-C remained relatively constant in untreated patients (114.3 mg/dL). In treated patients, the percentage with LDL-C ≥70 mg/dL decreased from 87.5% in January 2004 to 73.8% in December 2006 (P < .0001), then gradually declined between January 2007 (79.6%) and August 2008 (76.2%; P < .0001). Among untreated patients, 92.9% had LDL-C levels ≥70 mg/dL in January 2004 and 93.0% in August 2008. CONCLUSION: In conclusion, the percentage of high-risk patients with CHD or AVD treated with lipid-lowering therapy who achieve LDL-C <70 mg/dL levels has increased since 2004, although a large proportion of these patients still do not meet this goal. Additionally, 1 of 4 high-risk patients otherwise eligible for lipid-lowering therapy remains untreated. These data suggest the need for renewed efforts to support guideline-based LDL-C lowering in high-risk patients.

Association between dyslipidemia and vascular events in patients treated with statins: report from the UK General Practice Research Database.

OBJECTIVE: A retrospective cohort study was conducted to evaluate the association between low high-density lipoprotein cholesterol (HDL-C) and/or elevated triglycerides (TG) and cardiovascular (CV) and/or cerebrovascular (CB) events among patients with elevated low-density lipoprotein cholesterol (LDL-C) despite statin treatment. METHODS: Patient demographics, clinical characteristics, laboratory data, and CV/CB events, were collected from the UK General Practice Research Database. Abnormal lipid levels were defined using US and European clinical guidelines. The association between the frequency of CV/CB events among patients with HDL-C/TG abnormalities versus patients with isolated low LDL-C was estimated using multivariate Cox proportional hazards regression. RESULTS: Of 19,843 statin-treated patients, 6823 had elevated LDL-C despite therapy for a mean follow-up of 1.99+/-1.06 years. Among these patients, 3115 (45.7%) also had HDL-C/TG abnormalities. A total of 715 patients (10.5%) experienced CV/CB events. In statin-treated patients not at LDL-C goal, the relative risk of a vascular event was 24% higher in patients with HDL-C/TG abnormalities (HR=1.24, 95% CI: 1.06-1.46, p=0.006) than in patients without HDL-C/TG abnormalities. Additional variables that were associated with a significantly increased risk of CV/CB events included age (p<0.0001), gender (p=0.027), and medication possession ratio (p<0.0001), while diabetes mellitus (p<0.0001), hypertension (p<0.0001), 10-year Framingham risk score>30% (p=0.005), statin dose (p<0.0001), and LDL-C level at baseline (p<0.0001) were associated with a significantly decreased risk of CV/CB events. CONCLUSION: Among statin-treated patients with elevated LDL-C from UK clinical practices, reduced HDL-C and/or elevated TGs were associated with a significantly increased relative risk of CV/CB events.

Utilization and dosage pattern of echinocandins for treatment of fungal infections in US hospital practice.

OBJECTIVE: With the availability of multiple echinocandins in the US, recommended dosages and dosing schedules vary by agent but actual utilization practices are unknown. The purpose of this study was to describe the utilization and dosage pattern of intravenous echinocandins for treatment of fungal infections in US hospitals. METHODS: The Premier Perspective Database was used to describe echinocandin use in 332 US hospitals. Adult patients hospitalized from January, 2006 through June, 2007 with at least one billing record for anidulafungin (Eraxis ** ), caspofungin (Cancidas dagger ), or micafungin (Mycamine double dagger ) were included. Hospitalizations with > 1 echinocandin or >or= 1 dosage with an FDA approved indication for fungal prophylaxis were excluded. Mixed multivariable models were developed to identify factors associated with mean daily dose. ** Eraxis, a registered trade name owned by Pfizer, Inc., New York, NY, USA dagger Cancidas, a registered trade name owned by Merck & Co., Inc., Whitehouse Station, NJ, USA double dagger Mycamine, a registered trade name owned by Astellas Pharma US, Inc., Deerfield, IL, USA. RESULTS: The number of unique patient hospitalizations was 708 for anidulafungin, 15 739 for caspofungin, and 1199 for micafungin. A single echinocandin was utilized at 88.6% of hospitals. Micafungin patients had the highest prevalence of cancer, bone marrow transplant, solid organ transplant, HIV/AIDS, fungal infection, and neutropenia. Mean day 1 dose of echinocandin therapy was 171.2 +/- 85.4 mg, 79.7 +/- 25.6 mg, and 154.3 +/- 67.3 mg; and mean day 2 onwards dose was 98.7 +/- 39.4 mg, 53.1 +/- 12.5 mg, 122.6 +/- 39.4 mg for anidulafungin, caspofungin and micafungin, respectively. Commonly used loading doses were 200 mg (55.6%) for anidulafungin, 70 mg (57.2%) for caspofungin, and 200 mg (21.2%) for micafungin. The first-day dose of echinocandin therapy (vs. subsequent days) was most strongly associated with mean daily dose. CONCLUSIONS: In hospital practice, the mean dosages were consistent with the recommended loading and maintenance dosages for caspofungin and anidulafungin. Patients frequently received a loading dose of > 150 mg on day 1 of micafungin which was inconsistent with recommended dosing schedules. Micafungin maintenance dosages > 100 mg were also commonly used. Lack of information on reason for initiating echinocandin therapy was an important study limitation.

Lipid testing among patients beginning statin therapy in general practice in the United Kingdom.

OBJECTIVES: To determine the frequency of lipid testing and to identify the factors predictive of lipid-testing frequency over a 1-year period in patients beginning statin treatment. METHODS: Retrospective cohort study performed using the UK General Practice Research Database. The patients were selected if they were > or = 35 years of age, received first-ever statin between January 2000 and December 2004, had at least one total cholesterol (TC), high-density lipoprotein cholesterol (HDL-C), or triglyceride (TG) test conducted in the 1-year period before statin initiation, and had at least 1 year of follow-up data. The main outcome measures were TC, HDL-C, and TG testing frequencies in the year after initiating statins. Poisson regression was used to assess baseline factors associated with testing frequency for each lipid. RESULTS: In the year after initiating statins, the patients received a mean (+/-SD) of 1.3 (+/-1.0) TC tests, 0.9 (+/-1.0) HDL-C tests, and 0.9 (+/-1.0) TG tests; however, 22.7%, 44.3%, and 39.1% of the patients did not receive any TC, HDL-C, and TG tests, respectively. In multivariate analyses, a high coronary heart disease (CHD) risk (odds ratio [OR] 1.04; 95% confidence interval [CI] 1.01-1.07) and elevated baseline TC (> or = 6.2 vs. < 5.0 mmol/L; OR 1.12; 95% CI 1.06-1.18) were significantly associated with greater TC testing frequency. CONCLUSIONS: High risk of CHD and elevated baseline TC were associated with greater rates of TC testing in the year after statin initiation. Lack of TC testing in approximately one in five patients, and infrequent HDL-C and TG testing may be barriers to comprehensive lipid management.

Discontinuation of lipid modifying drugs among commercially insured United States patients in recent clinical practice.

Although several lipid-modifying drug (LMD) treatments and strategies are available to successfully manage patients at risk for cardiovascular events, the benefits of drug treatment can be realized only if these therapies are continued on a long-term basis. Previous observational studies examining rates of discontinuation with LMDs are not generalizable to current clinical practice in the United States. In this study, the discontinuation of newly initiated LMD classes in recent clinical practice was compared in a geographically diverse, commercially insured United States population. Administrative claims from the Ingenix Lab/Rx Database were used to identify patients aged >or=20 years who were newly prescribed statins, extended-release niacin, fibrates, bile acid sequestrants, or ezetimibe from January 1, 2001, to June 30, 2003. An LMD class was considered discontinued if a patient did not receive a prescription from the same LMD class within 180 days of the most recent prescription expiration date. The median time to discontinuation was 8.2 months in the bile acid sequestrant group, followed by 12 months in the extended-release niacin group, 17.4 months in the fibrate group, and 27.5 months in the statin group. By the end of 1 year, the adjusted cumulative incidence of discontinuation was 68.3% in bile acid sequestrant users, 55.4% in extended-release niacin users, 39.9% in fibrate users, 33.0% in ezetimibe users, and 28.9% in statin users (p <0.001 for all LMD classes vs statins). In conclusion, despite the changes in lipid treatment paradigms and the importance of long-term lipid therapy, this study found high discontinuation rates of LMD classes in recent United States clinical practice.

Drug switching patterns among patients with rheumatoid arthritis and osteoarthritis using COX-2 specific inhibitors and non-specific NSAIDs.

PURPOSE: To compare RA and OA patients' time-to-switch after newly initiating treatment with three most commonly used non-specific (NS)-NSAIDs and two COX-2 inhibitors, celecoxib and rofecoxib. METHODS: Managed care enrollees newly prescribed celecoxib, rofecoxib, ibuprofen, naproxen or diclofenac were identified. Time to switch to a different NS-NSAID or COX-2 specific inhibitor was determined using time-to-event analysis and Cox proportional hazards models were used to estimate the odds ratio (OR) after controlling for potential confounders. RESULTS: The time to 25% of the cohort switching was longer for rofecoxib and celecoxib (159 and 205 days respectively) compared to the three NS-NSAIDs (49-78 days). Patients were at the highest risk of switching within the first 100 days of therapy. After adjusting for potential confounding factors, the OR for switching to another NS-NSAID or COX-2 specific inhibitor ranged from 1.74 to 2.35 for the three NS-NSAIDs compared to celecoxib (all comparisons, p < 0.01). Similar findings were obtained when comparing rofecoxib to each of the three NS-NSAIDS (all comparisons, p < 0.01). When COX-2 inhibitors combined were compared to NS-NSAIDS combined, the OR for switching was 1.53 (95% confidence interval = 1.42-1.65; p < 0.01) after adjusting for potential confounders. CONCLUSIONS: Patients on the COX-2 specific inhibitors (celecoxib and rofecoxib) were significantly less likely to switch their therapy than patients on NS-NSAIDS (ibuprofen, naproxen and diclofenac). These results suggest that COX-2 specific inhibitors may be a more effective treatment option when compared with NS-NSAIDs in usual clinical practice.

Physician-reported management of edema and destabilized blood pressure in cyclooxygenase-2-specific inhibitor users with osteoarthritis and treated hypertension.

BACKGROUND: The addition of a nonsteroidal anti-inflammatory drug to the regimen of a patient with treated hypertension can cause a destabilization of blood pressure. OBJECTIVE: The aim of this study was to describe physician-reported management of clinically significant edema and/or destabilized blood pressure in patients with osteoarthritis (OA) and hypertension when initiating therapy with rofecoxib or celecoxib. METHODS: A cross-sectional survey was administered to physicians who attended one of several arthritis consultant programs sponsored by Pharmacia Corporation, with attendees selected by local sales representatives. Each program included a clinical presentation by a physician concerning the cardiorenal safety of celecoxib, followed by a consultative presentation and session led by a Pharmacia Clinical Education Manager. RESULTS: A total of 828 physicians in the following specialties completed the survey: family practice (33.0%), internal medicine (25.0%), orthopedics (15.2%), and rheumatology (11.4%). Responding physicians expected that the majority of patients who experienced edema would contact them (68.4%). They reported that they schedule follow-up visits for blood pressure monitoring 65.6% of the time after initiating a cyclooxygenase-2 (COX-2)-specific inhibitor, with family practitioners and internists most likely to indicate that they would do so and orthopedists least likely. Responding physicians indicated that the presence of edema and destabilized blood pressure generally led to discontinuation of the COX-2-specific inhibitor (58%-82% of the time). Internists and family practitioners were most likely to report that they treat edema by initiating or modifying diuretic therapy (33%-51% of the time). For destabilized blood pressure, an antihypertensive drug was reported to be initiated or modified 40% to 55% of the time by family practitioners and internists, whereas orthopedists indicated that they referred patients to the primary care provider. The COX-2-specific inhibitor prescribed resulted in management differences: physicians indicated that they were more likely to switch from rofecoxib to celecoxib in the event of edema or destabilized blood pressure, whereas they were more likely to adjust the celecoxib dose than the rofecoxib dose. Because the data were captured from convenience samples of physicians attending sponsored meetings, it is possible that respondents provided the answers they thought the sponsor would want. Because this was a cross-sectional survey, reported behavior was not compared with actual behavior. CONCLUSIONS: A significant percentage of physicians reported that they monitor patients with OA and hypertension for the occurrence of destabilized blood pressure and edema after initiation of a COX-2-specific inhibitor. Physicians indicated that they would nearly always intervene when either event is identified.