Search

Purpose To use medical record adjudication and predictive modeling methods to develop and validate an algorithm to identify anaphylaxis among adults with type 2 diabetes (T2D) in administrative claims. Methods A conventional screening algorithm that prioritized sensitivity to identify potential anaphylaxis cases was developed and consisted of diagnosis codes for anaphylaxis or relevant signs and symptoms. This algorithm was applied to adults with T2D in the HealthCore Integrated Research Database (HIRD) from 2016 to 2018. Clinical experts adjudicated anaphylaxis case status from redacted medical records. We used confirmed case status as an outcome for predictive models developed using lasso regression with 10-fold cross-validation to identify predictors and estimate the probability of confirmed anaphylaxis. Results Clinical adjudicators reviewed medical records with sufficient information from 272 adults identified by the anaphylaxis screening algorithm, which had an estimated Positive Predictive Value (PPV) of 65% (95% confidence interval [CI]: 60%-71%). The predictive model algorithm had a c-statistic of 0.95. The model's probability threshold of 0.60 excluded 89% (84/94) of false positives identified by the screening algorithm, with a PPV of 94% (95% CI: 91%-98%). The model excluded very few true positives (15 of 178), and identified 92% (95% CI: 87%-96%) of the cases selected by the screening algorithm. Conclusions Predictive modeling techniques yielded an accurate algorithm with high PPV and sensitivity for identifying anaphylaxis in administrative claims. This algorithm could be considered in future safety studies using similar claims data to reduce potential outcome misclassification.

Daina B Esposito,1 Vibha CA Desai,1 Judith J Stephenson,1 M Soledad Cepeda,2 Jennifer G Lyons,1 Crystal N Holick,1 Gregory P Wedin,3 Stephan Lanes1 On behalf of members of the REMS Program Companies Metrics Subteam1HealthCore, Inc., Wilmington, DE, USA; 2Janssen Research and Development, Titusville, NJ, USA; 3Upsher-Smith Laboratories, LLC, Maple Grove, MN, USACorrespondence: Judith J StephensonHealthCore Inc, 123 Justison Street, Suite 200, Wilmington, DE, 19801, USATel +1 302-547-5770Email jstephenson@healthcore.comBackground/Rationale: The US Food and Drug Administration (FDA) approved a Risk Evaluation and Mitigation Strategy (REMS) for extended release/long-acting (ER/LA) opioids in 2012. The purpose of this study was to assess patient knowledge of the safe use of these products following implementation of the REMS and to determine possible effects of the REMS, including impact on medication access.Objective: To assess patient knowledge of safe use of ER/LA opioids and use of REMS patient education tools such as the Medication Guide (MG) and Patient Counseling Document (PCD).Methods: This was a cross-sectional survey of commercially insured (Commercial) and Medicare Advantage-insured (Medicare) adults with ≥ 1 pharmacy claim for an ER/LA opioid (10/01/2015 – 02/28/2017) in the HealthCore Integrated Research Database and Medicaid-insured (Medicaid) adult members of a research panel, about their knowledge of safe use of ER/LA opioids and receipt/comprehension of the MG and PCD.Results: Survey respondents consisted of 382 Commercial, 43 Medicare and 40 Medicaid adults. While ≥ 95% of respondents received and read the MG, fewer were aware of the PCD (Commercial: 47%, Medicare: 65%, Medicaid: 53%). Almost 75% of the knowledge questions were answered correctly by ≥ 80% of all respondents; fewer respondents recognized that use of opioids as directed can lead to death (Commercial: 73%, Medicare: 56%, Medicaid: 63%), the MG should be read at each dispensing (Commercial: 78%, Medicare: 53%, Medicaid: 75%), opioids should not be stored in the medicine cabinet (Commercial: 77%, Medicare: 79%, Medicaid: 58%), missed doses should not be taken as soon as possible (Commercial: 56%, Medicare: 51%, Medicaid: 50%), and pills should not be crushed (Commercial: 85%, Medicare: 67%, Medicaid: 52%).Conclusion: Although most respondents reported reading and understanding the MG and exhibited knowledge of safe use of ER/LA opioids, providers’ use of the PCD and increased understanding of safe use core messages need reinforcement.Keywords: Risk Evaluation and Mitigation Strategy, opioids, patient knowledge

PURPOSE It is well documented that outcome misclassification can bias a point estimate. We aimed to understand current practice in addressing this bias in pharmacoepidemiology database studies and to develop an open source application (app) from existing methodology to demonstrate the impact and mechanism of this bias on results. METHODS Studies of an exposure and a clinical outcome were selected from all Pharmacoepidemiology and Drug Safety publications during 2017 and any reference to outcome misclassification described. An app to correct risk ratio (RR) and cumulative incidence for outcome misclassification was developed from a published methodology and used to demonstrate the impact of correction on point estimates. RESULTS Eight (19%) of 43 papers selected reported estimates of outcome ascertainment accuracy with positive predictive value (PPV) the most commonly reported measure (7 of 8 studies). Three studies (7%) corrected for the bias, 1 by exposure strata, and 5 (12%) restricted analyses to confirmed cases. The app (app http://apps.p-95.com/ISPE/) uses values of PPV and sensitivity (or a range of possible values) in each exposure strata and returns corrected point estimates and confidence intervals. The app demonstrates that small differences between comparison groups in PPV or sensitivity can introduce bias even when accuracy estimates are high. CONCLUSIONS Outcome misclassification is not usually corrected in pharmacoepidemiology database studies although correction methods using routinely measured indices are available. Error indices are needed for each comparison group to correct RR estimates for these errors. The app should encourage understanding of this bias and increase adjustment.

Background: We conducted a study to assess whether testosterone therapy (TT) alters prostate cancer risk using a large U.S. commercial insurance research database. Methods: From the HealthCore Integrated Research Database (HIRD), we selected men ages 30 years or greater who were new users of TT during 2007 to 2015. We selected two comparison groups: (i) unexposed (matched 10:1) and (ii) new users of phosphodiesterase type 5 inhibitor (PDE5i). Incident prostate cancer was defined as diagnosis of prostate cancer within 4 weeks following prostate biopsy. Propensity scores and inverse probability of treatment weights were used in Poisson regression models to estimate adjusted incidence rates, incidence rate ratios (IRR), and 95% confidence intervals (CI). Subgroup analyses included stratification by prostate cancer screening, hypogonadism, and follow-up time. Results: The adjusted prostate cancer IRR was 0.77 (95% CI, 0.68–0.86) when comparing TT with the unexposed group and 0.85 (95% CI, 0.79–0.91) in comparison with the PDE5i group. Inverse associations between TT and prostate cancer were observed in a majority of subgroup analyses, although in both comparisons estimates generally attenuated with increasing time following initial exposure. Among TT users, duration of exposure was not associated with prostate cancer. Conclusions: Men who received TT did not have a higher rate of prostate cancer compared with the unexposed or PDE5i comparison groups. The inverse association between TT and prostate cancer could be the result of residual confounding, contraindication bias, or undefined biological effect. Impact: This study suggests that limited TT exposure does not increase risk of prostate cancer in the short term.

Daina B Esposito, 1 M Soledad Cepeda, 2 Crystal N Holick, 1 Caitlin Knox, 1 Vibha CA Desai, 1 Nianya Liu, 1 Shiva-Krishna Vojjala, 1 Jennifer G Lyons, 1 Gregory P Wedin, 3 Stephan Lanes 1 On Behalf of the REMS Program Companies Metrics Subteam 1HealthCore, Inc, Andover, MA, USA; 2Janssen Research and Development, Titusville, NJ, USA; 3Upsher-Smith Laboratories, LLC, Maple Grove, MN, USACorrespondence: Stephan LanesSafety and Epidemiology, HealthCore, Inc., 300 Brickstone Square, 8th Floor, Suite 801A, Andover, MA 01810, USATel +1 302-384-0673Fax +1 978-247-6650Email slanes@healthcore.comIntroduction: The United States (US) Food and Drug Administration (FDA) required a Risk Evaluation and Mitigation Strategy (REMS) for extended-release and long-acting (ER/LA) opioid analgesics on 09 July 2012.Methods: This study compared the incidence of opioid overdose before (July 2010-June 2012) and after (July 2013-September 2016) the initiation of the Risk Evaluation and Mitigation Strategy (REMS) for extended-release and long-acting (ER/LA) opioid analgesics. We identified patients with ≥ 1 ER/LA opioid dispensing in either time period in national data from the HealthCore Integrated Research DatabaseSM (HIRD) and in United States (US) Medicaid claims data from four states. We described each population, calculated the incidence rate (IR) of opioid overdose, and assessed crude and propensity score adjusted incidence rate ratios (IRR) comparing the overdose rate after vs before implementation of the REMS.Results: A total of 121,229 commercially insured and 11,488 Medicaid patients were included in the analysis. Rates of overdose were substantially higher in Medicaid patients than in the commercially insured patients (IR 192.0, 95% confidence interval [CI] 162.60– 225.18 versus 102.60, 95% CI 93.0– 112.93 in the active period). The IRRs for opioid overdose were 1.01 (95% CI 0.87– 1.17) in the commercially insured population and 0.70 (95% CI 0.52– 0.93) in Medicaid.Conclusion: This leveling off of overdose rates among commercially insured patients and decline among Medicaid patients is encouraging, but it is difficult to disentangle the specific impact of the REMS from many other ongoing initiatives with similar goals.Keywords: opioid, overdose, REMS, Medicaid, claims

Background There is limited real-world safety information on palbociclib for treatment of advanced stage HR+/HER2- breast cancer. Methods We conducted a cohort study of breast cancer patients initiating palbociclib and fulvestrant from February 2015 to September 2017 using the HealthCore Integrated Research Database (HIRD), a longitudinal claims database of commercial health plan members in the United States. The historical comparator cohort comprised patients initiating fulvestrant monotherapy from January 2011 to January 2015. Propensity score matching and Cox regression were used to estimate hazard ratios for various safety events. For acute liver injury (ALI), additional analyses and medical record validation were conducted. Results There were 2445 patients who initiated palbociclib including 566 new users of palbociclib-fulvestrant, and 2316 historical new users of fulvestrant monotherapy. Compared to these historical new users of fulvestrant monotherapy, new users of palbociclib-fulvestrant had a greater than 2-fold elevated risk for neutropenia, leukopenia, thrombocytopenia, stomatitis and mucositis, and ALI. Incidence of anemia and QT prolongation were more weakly associated, and incidences of serious infections and pulmonary embolism were similar between groups after propensity score matching. After adjustment for additional ALI risk factors, the elevated risk of ALI in new users of palbociclib-fulvestrant persisted (e.g. primary ALI algorithm hazard ratio (HR) = 3.0, 95% confidence interval (CI) = 1.1–8.4). Conclusions This real-world study found increased risks of several adverse events identified in clinical trials, including neutropenia, leukopenia, and thrombocytopenia, but no increased risk of serious infections or pulmonary embolism when comparing new users of palbociclib-fulvestrant to fulvestrant monotherapy. We observed an increased risk of ALI, extending clinical trial findings of significant imbalances in grade 3/4 elevations of alanine aminotransferase (ALT).

Background Background incidence rates (IRs) of potential safety outcomes among vaccine eligible individuals can inform assessment of vaccine safety. Vaccine safety surveillance often uses claims databases, but the impact of outcome definitions on background IR estimates is largely unexplored. Using two definitions for each outcome, we estimated background IRs of 32 cardiac, metabolic, allergic, autoimmune, neurologic, hematologic and nephrologic outcomes among individuals eligible to receive pneumococcal vaccination. Methods We defined a cohort of individuals aged 6–100 years in US commercial health plans who had ≥12 months of health plan enrollment between January 2007 and August 2014 and no previous record of conjugate or simple polysaccharide pneumococcal vaccination. We developed a sensitive and a specific definition for each outcome, with the specific definition requiring evidence of additional care consistent with the outcome. IRs per 100,000 person-years for each outcome were presented overall and stratified by age, gender, and invasive pneumococcal disease (IPD) risk category. Results We followed 19.9 million individuals for a median of 2.5 years. Wide variation was seen in IRs across different definitions of the 32 outcomes, with 19 (59%) outcomes having a specific definition IR less than half of the sensitive definition IR. IRs were particularly variable by definition for outcomes categorized as either hematologic/nephrologic or neurologic (mean ratio of specific IR to sensitive IR = 0.26 and 0.30, respectively). Across definitions, the IRs of the 32 outcomes were often highest in females, adults ≥65, and those at higher IPD risk. Conclusions Background IRs of safety outcomes relevant to populations indicated for pneumococcal vaccine varied by outcome definitions and population subgroups in this large US commercially-insured population. Given large differences in estimated IRs using sensitive versus specific case definitions, neurologic, and hematologic/nephrologic safety outcomes as compared to allergic and autoimmune outcomes may warrant more refined definitions and medical record validation.

Background The accuracy with which hemophilia A can be identified in claims databases is unknown. Objective Develop and validate an algorithm using predictive modeling supported by machine learning to identify patients with hemophilia A in an administrative claims database. Methods We first created a screening algorithm using medical and pharmacy claims to identify potential hemophilia A patients in the US HealthCore Integrated Research Database between January 1, 2006 and April 30, 2015. Medical records for a random sample of patients were reviewed to confirm case status. In this validation sample, we used lasso logistic regression with cross-validation to select covariates in claims data and develop a predictive model to estimate the probability of being a confirmed hemophilia A case. Results The screening algorithm identified 2,252 patients and we reviewed medical records for 400 of these patients. The screening algorithm had a positive predictive value (PPV) of 65%. The predictive model identified 18 predictors of being a hemophilia A case or noncase. The strongest predictors of case status included male sex, factor VIII therapy, office visits for hemophilia A, and hospitalizations for hemophilia A. The strongest predictors of noncase status included hospitalizations for reasons other than hemophilia A and factor VIIa therapy. A probability threshold of ≥0.6 resulted in a PPV of 94.7% (95% CI: 92.0–97.5) and sensitivity of 94.4% (95% CI: 91.5–97.2). Conclusions We developed and validated an algorithm to identify hemophilia A cases in an administrative claims database with high sensitivity and high PPV.

Temozolomide (TMZ) is used to treat adult patients with glioblastoma multiforme (GBM). Cases of hepatotoxicity have been reported among patients using TMZ. The objective of the study was to assess the relation, if any, between exposure to TMZ and serious acute liver injury (SALI). We used the HealthCore Integrated Research Database to perform a case-control study nested within a retrospective cohort of adult patients aged 18-100 years with at least two diagnoses of brain cancer anytime between 2006 and 2014. Patients without continuous eligibility or with a SALI diagnosis within 6 months prior to the date of incident brain cancer diagnosis were excluded. Medical records were sought for potential SALI cases and reviewed by two hepatologists. Five controls were selected for each case using incidence density sampling, matched on age and calendar year of index date. The analysis included 61 confirmed SALI cases and 305 selected controls. Exposure to TMZ was classified according to dispensing date and days supply of medication dispensed. We estimated odds ratios using conditional logistic regression models. The odds ratio for any exposure to TMZ was 0.91 (95% CI 0.44-1.91), for recent exposure to TMZ was 0.62 (95% CI 0.21-1.85). There was no increased risk of SALI with increasing duration of exposure to TMZ. When patients with unconfirmed SALI were included in the analysis, results were similar (OR 1.04; 95% CI 0.70-1.54). In conclusion, this study did not find an association between TMZ and SALI risk among patients with brain cancer.

Background: Immunological cross-reactivity between common cold coronaviruses (CCC) and SARS-CoV-2 might account for the reduced incidence of COVID-19 in children. Evidence to support speculation includes in vitro evidence for humoral and cellular cross-reactivity with SARS-CoV-2 in specimens obtained before the pandemic started. Method: We used retrospective health insurance enrollment records, claims, and laboratory results to assemble a cohort of 869,236 insured individuals who had a PCR test for SARS-CoV-2. We estimated the effects of having clinical encounters for various diagnostic categories in the year preceding the study period on the risk of a positive test result. Findings: After adjusting for age, gender and care seeking behavior, we identified that individuals with diagnoses for common cold symptoms, including acute sinusitis, bronchitis, or pharyngitis in the preceding year had a lower risk of testing positive for SARS-CoV-2 (OR=0.76, 95%CI=0.75, 0.77). No reduction in the odds of a positive test for SARS-CoV-2 was seen in individuals under 18 years. The reduction in odds in adults remained stable for four years but was strongest in those with recent common cold symptoms. Interpretation: While this study cannot attribute this association to cross-immunity resulting from a prior CCC infection, it is one potential explanation. Regardless of the cause, the reduction in the odds of being infected by SARS-CoV-2 among those with a recent diagnosis of common cold symptoms may have a role in shifting future COVD-19 infection patterns from endemic to episodic. Funding: All authors are employees or consultants for Anthem Inc. or HealthCore Inc. There was no funding for this study. Declaration of Interests: Drs. Beachler and Lanes are employees of HealthCore Inc., a subsidiary of Anthem Inc. Dr. Ovehage is a consultant for HealthCore. Dr. Aran is a consutlant for Anthem Innovation Israel. None of the authors have any competing interests. Ethics Approval Statement: This study was designed as an analysis based on medical claims data, and there was no active enrollment or active follow-up of study subjects, and no data were collected directly from individuals. The study was not required to obtain additional IRB approval, as the HIPAA Privacy Rule permits protected health information (PHI) in a limited data set to be used or disclosed for research, without individual authorization, if certain criteria are met.

Background Prostate cancer is a commonly studied outcome in administrative claims studies, but there is a dearth of validated case identifying algorithms. The long-term development of the disease increases the difficulty in separating prevalent from incident prostate cancer. The purpose of this validation study was to assess the accuracy of a claims algorithm to identify incident prostate cancer among men in commercial and Medicare Advantage US health plans. Methods We identified prostate cancer in claims as a prostate cancer diagnosis within 28 days after a prostate biopsy and compared case ascertainment in the claims with the gold standard results from the Georgia Comprehensive Cancer Registry (GCCR). Results We identified 74,008 men from a large health plan claims database for possible linkage with GCCR. Among the 382 prostate cancer cases identified in claims, 312 were also identified in the GCCR (positive predictive value [PPV] = 82%). Of the registry cases, 91% (95% confidence interval = 88, 94) were correctly identified in claims. Claims and registry diagnosis dates of prostate cancer matched exactly in 254/312 (81%) cases. Nearly half of the false-positive cases also had claims for prostate cancer treatment. Thirteen (43%) false-negative cases were classified as noncases by virtue of having a biopsy and diagnosis >28 days apart as required by the algorithm. Compared to matches, false-negative cases were older men with less aggressive prostate cancer. Conclusions Our algorithm demonstrated a PPV of 82% with 92% sensitivity in ascertaining incident PC. Administrative health plan claims can be a valuable and accurate source to identify incident prostate cancer cases.

Purpose Long-acting beta agonists (LABAs) when used without concomitant inhaled corticosteroids (ICS) increase the risk of asthma-related deaths, but the effect on asthma-related death of LABA used in combination with ICS therapy is unknown. To address this question, we explored the feasibility of conducting an observational study using multiple US health care data sources. Methods Retrospective cohort study to evaluate the likelihood of getting an upper 95% confidence limit ≤1.4 for the asthma mortality rate ratio and ≤0.40 per 10 000 person-years for the mortality rate difference, assuming no effect of new use of combined LABA + ICS (versus non-LABA maintenance therapy) on asthma mortality. Ten research institutions executed centrally distributed analytic code based on a standard protocol using an extracted (2000-2010) persistent asthma cohort (asthma diagnosis and ≥4 asthma medications in 12 months). Pooled results were analyzed by the coordinating center. Asthma deaths were ascertained by linkage with the National Death Index. Results In a cohort of 994 627 persistent asthma patients (2.4 million person-years; 278 asthma deaths), probabilities of the upper 95% confidence limit for effect estimates being less than targeted values, assuming a null relation, were about 0.05. Modifications in cohort and exposure definitions increased exposed person-time and outcome events, but study size remained insufficient to attain study goals. Conclusions Even with 10 data sources and a 10-year study period, the rarity of asthma deaths among patients using certain medications made it infeasible to study the association between combined LABA + ICS and asthma mortality with our targeted level of study precision. Copyright © 2016 John Wiley & Sons, Ltd.

Background. Endometrial adenocarcinoma is the most prevalent type of endometrial cancer. Diagnostic codes to identify endometrial adenocarcinoma in administrative databases, however, have not been validated. Objective. To develop and validate an algorithm for identifying the occurrence of endometrial adenocarcinoma in a health insurance claims database. Methods. To identify potential cases among women in the HealthCore Integrated Research Database (HIRD), published literature and medical consultation were used to develop an algorithm. The algorithm criteria were at least one inpatient diagnosis or at least two outpatient diagnoses of uterine cancer (International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) 182.xx) between 1 January 2010 and 31 August 2014. Among women fulfilling these criteria, we obtained medical records and two clinical experts reviewed and adjudicated case status to determine a diagnosis. We then estimated the positive predictive value (PPV) of the algorithm. Results. The PPV estimate was 90.8% (95% CI 86.9–93.6), based on 330 potential cases of endometrial adenocarcinoma. Women who fulfilled the algorithm but who, after review of medical records, were found not to have endometrial adenocarcinoma, had diagnoses such as uterine sarcoma, rhabdomyosarcoma of the uterus, endometrial stromal sarcoma, ovarian cancer, fallopian tube cancer, endometrial hyperplasia, leiomyosarcoma, or colon cancer. Conclusions. An algorithm comprising one inpatient or two outpatient ICD-9-CM diagnosis codes for endometrial adenocarcinoma had a high PPV. The results indicate that claims databases can be used to reliably identify cases of endometrial adenocarcinoma in studies seeking a high PPV.

Background Although healthcare databases are a valuable source for real-world oncology data, cancer stage is often lacking. We developed predictive models using claims data to identify metastatic/advanced-stage patients with ovarian cancer, urothelial carcinoma, gastric adenocarcinoma, Merkel cell carcinoma (MCC), and non-small cell lung cancer (NSCLC). Methods Patients with ≥1 diagnosis of a cancer of interest were identified in the HealthCore Integrated Research Database (HIRD), a United States (US) healthcare database (2010–2016). Data were linked to three US state cancer registries and the HealthCore Integrated Research Environment Oncology database to identify cancer stage. Predictive models were constructed to estimate the probability of metastatic/advanced stage. Predictors available in the HIRD were identified and coefficients estimated by Least Absolute Shrinkage and Selection Operator (LASSO) regression with cross-validation to control overfitting. Classification error rates and receiver operating characteristic curves were used to select probability thresholds for classifying patients as cases of metastatic/advanced cancer. Results We used 2723 ovarian cancer, 6522 urothelial carcinoma, 1441 gastric adenocarcinoma, 109 MCC, and 12,373 NSCLC cases of early and metastatic/advanced cancer to develop predictive models. All models had high discrimination (C > 0.85). At thresholds selected for each model, PPVs were all >0.75: ovarian cancer = 0.95 (95% confidence interval [95% CI]: 0.94–0.96), urothelial carcinoma = 0.78 (95% CI: 0.70–0.86), gastric adenocarcinoma = 0.86 (95% CI: 0.83–0.88), MCC = 0.77 (95% CI 0.68–0.89), and NSCLC = 0.91 (95% CI 0.90 – 0.92). Conclusion Predictive modeling was used to identify five types of metastatic/advanced cancer in a healthcare claims database with greater accuracy than previous methods.

Inflammatory bowel disease (IBD) is a chronic condition commonly requiring lifelong care. Both IBD and IBD-related treatments can cause significant morbidity, and it is often difficult to differentiate their relative etiologic contribution to adverse events (AEs). The objectives of this study were to assess the rates of select AEs among patients with IBD as a function of disease severity and of the use of anti-tumor necrosis factor alpha (anti-TNFα) medications.We conducted a retrospective cohort study of IBD patients in the HealthCore Integrated Research Database (HIRD(TM)) between January 2004 and January 2011 to determine rates of AEs in patients with mild and moderate to severe IBD. Key study endpoints were select prespecified malignant neoplasms, infections, and other AEs of interest.A total of 33,386 IBD patients (52.7% ulcerative colitis; 47.3% Crohn's disease) met the inclusion criteria, and 60% had been followed for ≥1 year. Patients with moderate to severe IBD had increased rates of infections, lymphatic and digestive tract cancers, gastrointestinal (GI) perforations, and myocardial infarctions versus patients with mild IBD. Patients with IBD who used anti-TNFα therapies during the study had increased incidence of many types of infections, certain GI cancers (including rectal and anal cancer), intestinal perforations, and kidney stones compared with patients who had never used anti-TNFα therapies.Results from this large US cohort provide descriptive information on AE rates in a population of IBD patients undergoing routine care, estimating background incidence rates of AEs that are not readily available in the published literature. Our study findings may be limited owing to a lack of generalizability and potential for misclassification due to reliance on medical diagnosis and treatment and procedure codes to identify disease, comorbidities, and treatments. Further research and validation of our findings in other populations and databases are needed.

Purpose The purpose of this review is to assist researchers in developing, using, and interpreting case-identifying algorithms in electronic healthcare databases. Methods We review clinical characteristics of health outcomes, data settings and informatics, and epidemiologic and statistical methods aspects as they pertain to the development and use of case-identifying algorithms. Results We offer a framework for thinking critically about the use of electronic health insurance data and electronic health records to identify the occurrence of health outcomes. Accuracy of case ascertainment in database research depends on many factors, including clinical and behavioral aspects of the health outcome, and details of database construction as it pertains to completeness and reliability of database content. Existing methods for diagnostic and screening tests, misclassification, validation studies, and predictive modelling can be usefully applied to improve case ascertainment in database research. Conclusions Good case-identifying algorithms are based on a sound understanding of care-seeking behavior and patterns of clinical diagnosis and treatment in the study population and details about the construction and characteristics of the database. Researchers should use quantitative bias analyses to take into account the performance characteristics of case-identifying algorithms and their impact on study results. Copyright © 2015 John Wiley & Sons, Ltd.

Purpose Claims databases offer large populations for research, but lack clinical details. We aimed to develop predictive models to identify estrogen receptor positive (ER+) and human epidermal growth factor negative (HER2-) early breast cancer (ESBC) and advanced stage breast cancer (ASBC) in a claims database. Methods Female breast cancer cases in Anthem's Cancer Care Quality Program served as the gold standard validation sample. Predictive models were developed from clinical knowledge and empirically from claims data using logistic and lasso regression. Model performance was assessed by classification rates and c-statistics. Models were applied to the HealthCore Integrated Research Database (claims) to identify cohorts of women with ER+/HER2- ESBC and ASBC. Results The validation sample included 3184 women with ER+/HER2- ESBC and 1436 with ER+/HER2- ASBC. Predictive models for ER+/HER2- ESBC and ASBC included 25 and 20 factors, respectively. Models had robust discrimination in identifying cases (c-stat = 0.92 for ESBC and 0.95 for ASBC). Compared with a traditional a priori algorithm developed with clinical insight alone, the ER+/HER2- ASBC-predictive model had better positive predictive value (PPV) (0.91, 95% CI, 0.90-0.93, vs 0.69, 95% CI, 0.66-0.73) and sensitivity (0.54 vs 0.35). Models were applied to the claims database to identify cohorts of 33 001 and 3198 women with ER+/HER2- ESBC and ASBC. Conclusion We conducted a validation study and developed predictive models to identify in a claims database cohorts of women with ER+/HER2- ESBC and ASBC. The models identified large cohorts in the claims data that can be used to characterize indications in the evaluation of targeted therapies.

To describe patient and provider characteristics for patients with type 2 diabetes (T2DM) initiating basal insulin and describe basal insulin's impact on sulfonylurea (SU) discontinuation.A retrospective cohort study was conducted using the HealthCore Integrated Research Database. Patients had ≥12 months of continuous coverage prior to initiating insulin, and were utilizing at least one anti-hyperglycemic drug at the time of insulin initiation. Predictors for SU discontinuation were evaluated utilizing Cox proportional hazards models.Among the 74,334 individuals aged ≥18 years with T2DM who initiated basal insulin from 2006-2015, 30% were taking metformin (MET) and SU when initiating insulin. Among the 22,418 MET/SU patients, 31% discontinued SU within 3 months of insulin initiation and, by 12 months, 55% had discontinued SU. Sulfonylurea discontinuation was similar among many patient and provider characteristics, while being modestly positively associated (p .05; HRs1.5) with female gender, more co-morbidities, cardiac revascularization, chronic liver disease, hospitalizations with a T2DM diagnosis, and hypoglycemia prior to insulin initiation. SU discontinuation was modestly inversely associated with receiving an insulin prescription from an endocrinologist (HR = 0.90, 95% CI = 0.85-0.95).Roughly half of commercially-insured T2DM patients discontinued SU within 1 year after insulin initiation, and SU discontinuation was not strongly associated with a range of patient and provider characteristics.

129 Background: Our main objective was to identify, in healthcare claims data, patients with advanced or metastatic: urothelial carcinoma (amUC), gastric cancer (amGC) and non-small cell lung cancer (amNSCLC) and to report on their characteristics, treatments, and survival rates using contemporaneous real-world data. Methods: This cohort study was conducted in the HealthCore Integrated Research Database (HIRD), from January 2010 to January 2018, which contains healthcare claims data from commercial health plans across the US (60 million lives). We applied algorithms, previously validated on registry data, to the HIRD to define 3 cohorts of advanced stage cancer. Cohort characteristics and treatment patterns were described. Patient vital status was captured through probabilistic linkage with the National Death Index (NDI) and survival was assessed using the Kaplan-Meier method. Results: Algorithms to predict advanced stage cancer resulted in the following cohorts: 1,501 amUC, 6,253 amGC and 38,451 amNSCLC cases. Most patients in each cohort were de novo advanced or metastatic, but subsets were diagnosed at early stage and progressed to advanced stage (ranging from 15.1% for amNSCLC to 23.1% for amUC). Patient characteristics, treatments and survival outcomes are described in Table 1. Not all received systemic treatment; Immune Checkpoint inhibitors (ICI) were used in 5.3%, 2.2% and 10.8% of treated amUC, amGC and amNSCLC patients, respectively. Conclusions: In these cohorts of advanced or metastatic cancer patients, median survival time was limited despite most receiving treatment: radiation, systemic therapy or surgery. Treatment with ICI was low despite recent data in amUC and amNSCLC.[Table: see text]

Background: Testosterone supplementation (TS) has dramatically increased in the United States (US). We conducted a study of TS and prostate cancer risk using a large US commercial insurance research database. Methods: From the HealthCore Integrated Research Database (HIRDSM), we selected men aged 30 years or greater who were new users of TS during 2007-2015. We selected two male comparison groups: 1) unexposed (matched 10:1); 2) new users of a phosphodiesterase type 5 inhibitor (PDE5i). Incident prostate cancer was defined as diagnosis of prostate cancer within four-weeks following prostate biopsy. Propensity scores and inverse probability of treatment weights were used in Poisson regression models to estimate adjusted incidence rates, incidence rate ratios (IRRs) and 95% confidence intervals (CI) using doubly robust estimation. Subgroup analyses included stratification by prostate cancer screening, hypogonadism, and follow-up time. Results: The adjusted prostate cancer IRR was 0.77 (95%CI: 0.68, 0.86) when comparing TS with the unexposed group and 0.85 (95%CI: 0.79, 0.91) in comparison with the PDE5i group. Inverse associations between TS and prostate cancer were observed in a majority of subgroup analyses, although in both comparisons estimates generally attenuated with increasing time following initial exposure. Amongst TS users, duration of exposure was not associated with prostate cancer. Conclusion: In this study, men who received TS did not have a higher rate of prostate cancer compared with the unexposed or PDE5i comparison groups. The inverse association between TS and prostate cancer could be the result of residual confounding, contraindication bias, or undefined biologic effect. Citation Format: Michael B. Cook, Daniel C. Beachler, Lauren E. Parlett, Philip T. Cochetti, William D. Finkle, Stephan Lanes, Robert N. Hoover. Testosterone supplementation in relation to prostate cancer in a US commercial insurance claims database [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 5050.

e13082 Background: Our main objective was to identify, in healthcare claims data, patients with advanced or metastatic: urothelial carcinoma (amUC), gastric cancer (amGC) and non-small cell lung cancer (amNSCLC) and to report on their characteristics, treatments, and survival rates using contemporaneous real-world data. Methods: This cohort study was conducted in the HealthCore Integrated Research Database (HIRD), from January 2010 to January 2018, which contains healthcare claims data from commercial health plans across the US (60 million lives). We applied algorithms, previously validated on registry data, to the HIRD to define 3 cohorts of advanced stage cancer. Cohort characteristics and treatment patterns were described. Patient vital status was captured through probabilistic linkage with the National Death Index (NDI) and survival was assessed using the Kaplan-Meier method. Results: Algorithms to predict advanced stage cancer resulted in the following cohorts: 1,501 amUC, 6,253 amGC and 38,451 amNSCLC cases. Most patients in each cohort were de novo advanced or metastatic, but subsets were diagnosed at early stage and progressed to advanced stage (ranging from 15.1% for amNSCLC to 23.1% for amUC). Patient characteristics, treatments and survival outcomes are described in Table. Not all received systemic treatment; Immune Checkpoint inhibitors (ICI) were used in 5.3%, 2.2% and 10.8% of treated amUC, amGC and amNSCLC patients, respectively. Conclusions: In these cohorts of advanced or metastatic cancer patients, median survival time was limited despite most receiving treatment: radiation, systemic therapy or surgery. Treatment with ICI was low despite recent data in amUC and amNSCLC. Characteristics, treatments, and survival from estimated advanced stage date. [Table: see text]

Purpose The impact of i.v. drug delivery via point-of-care (POC)-activated and closed systems versus traditional manual admixture systems on the risk of hospital-acquired bloodstream infection (BSI) is examined. Methods Using data from a proprietary hospital database, a retrospective observational cohort study of patients receiving one or more i.v. drug administrations via POC-activated or closed systems during a three-year period (2007–09) was conducted. Cases of hospital-acquired BSI were identified using diagnosis codes and billing charges for blood cultures and antibiotic use. The risk of BSI in patients with exposure to POC-activated systems, closed systems, or both relative to that of patients exposed to open systems was estimated by odds ratios (ORs) calculated by multivariate logistic regression analysis. Results The evaluated data indicated that of the 4,073,864 patients included in the study cohort, 0.5% ( n = 20,251) experienced hospital-acquired BSI. After adjusting for selected confounding variables, the use of POC-activated systems was associated with a 16% reduction in BSI risk relative to the use of open systems (OR, 0.84; 95% confidence interval [CI], 0.76–0.93), and the use of closed systems correlated with a 12% risk reduction (OR, 0.88; 95% CI, 0.82–0.96). Patients who received i.v. drugs via both POC-activated and closed systems appeared to derive the greatest relative risk reduction benefit (OR, 0.12; 95% CI, 0.06–0.23). Conclusion Use of POC-activated and closed systems for i.v. drug delivery was associated with a significantly reduced risk of hospital-acquired BSI compared with exclusive use of open systems in a large population of hospitalized patients.

Long-acting beta agonists (LABAs) when used without concomitant inhaled corticosteroids (ICS) increase the risk of asthma-related deaths, but the effect on asthma-related death of LABA used in combination with ICS therapy is unknown. To address this question, we explored the feasibility of conducting an observational study using multiple US health care data sources.Retrospective cohort study to evaluate the likelihood of getting an upper 95% confidence limit ≤1.4 for the asthma mortality rate ratio and ≤0.40 per 10 000 person-years for the mortality rate difference, assuming no effect of new use of combined LABA + ICS (versus non-LABA maintenance therapy) on asthma mortality. Ten research institutions executed centrally distributed analytic code based on a standard protocol using an extracted (2000-2010) persistent asthma cohort (asthma diagnosis and ≥4 asthma medications in 12 months). Pooled results were analyzed by the coordinating center. Asthma deaths were ascertained by linkage with the National Death Index.In a cohort of 994 627 persistent asthma patients (2.4 million person-years; 278 asthma deaths), probabilities of the upper 95% confidence limit for effect estimates being less than targeted values, assuming a null relation, were about 0.05. Modifications in cohort and exposure definitions increased exposed person-time and outcome events, but study size remained insufficient to attain study goals.Even with 10 data sources and a 10-year study period, the rarity of asthma deaths among patients using certain medications made it infeasible to study the association between combined LABA + ICS and asthma mortality with our targeted level of study precision. Copyright © 2016 John WileySons, Ltd.

Idiopathic pulmonary fibrosis is a life-threatening condition, and few data concerning the impact on healthcare utilization and associated costs are available. The objective of this study was to describe the burden of illness (comorbidity, healthcare resource utilization, and associated costs) in patients with idiopathic pulmonary fibrosis.Two cohorts (patients with idiopathic pulmonary fibrosis and matched controls) were retrospectively identified from US claims databases between January 1, 2001 and September 30, 2008. Cases with idiopathic pulmonary fibrosis were defined by age of 55 years or older and either two or more claims with a code for idiopathic fibrosing alveolitis (ICD-9 516.3), or one claim with ICD 516.3 and a subsequent claim with a code for post-inflammatory pulmonary fibrosis (ICD-9 515). The prevalence and incidence of pre-selected comorbidities, healthcare resource utilization (hospital, outpatient, drugs), and direct medical costs were assessed in each cohort.A total of 9286 patients with idiopathic pulmonary fibrosis were identified. When compared with age- and gender-matched controls, these patients were at significantly increased risk for comorbidities including pulmonary hypertension and emphysema. The all-cause hospital admission rate (0.5 per person-year) and the all-cause outpatient visit rate (28.0 per person-year) were both ∼2-fold higher than in controls. Total direct costs for patients with idiopathic pulmonary fibrosis were $26,378 per person-year; the incremental costs over controls were $12,124 (2008 value).Patients with idiopathic pulmonary fibrosis experience increased comorbidity, healthcare resource utilization, and direct medical costs compared to controls.

Background Topical retinoids have been in clinical use for the treatment of chronic skin conditions, including acne, photodamage, and psoriasis, for 30 years. Objective A systematic literature review was conducted to assess the incidence of noncutaneous adverse events (AE) among patients treated with topical retinoids with a focus on topical tretinoin studies reported before the Veterans Affairs Topical Tretinoin Chemoprevention trial. Methods Electronic literature searches were conducted in Embase and MEDLINE for literature reporting development of nonteratogenic, noncutaneous AE among patients treated with topical retinoids published through September 2008. Results The search yielded 2778 citations, of which 20 studies met inclusion criteria. Tretinoin was used in 14 of the studies. Other retinoids assessed included isotretinoin, adapalene, alitretinoin, and tazarotene. Within patients receiving topical tretinoin, 27.9% reported the occurrence of at least one noncutaneous AE. The majority of noncutaneous AE were transient and judged not to be related to tretinoin treatment. Limitations The conclusions of this study apply largely to tretinoin compared with other topical retinoids. Many of the included trials were designed to evaluate the efficacy of topical treatment and reporting of safety events concentrated on incidence of localized AE, rather than systemic or noncutaneous events. Conclusion We found no clear evidence of a relationship between the use of topical tretinoin and the development of noncutaneous AE before a recent report of excess mortality in a clinical trial. The majority of noncutaneous AE reported by patients receiving topical retinoids consisted of nonsevere, nonspecific symptoms that were judged not to be related to treatment.

e13603Background: Healthcare databases are valuable for epidemiological research in oncology. Our main objective was to identify patients (pts) with advanced or metastatic: ovarian cancer (mOC), ur...

Estimates of idiopathic pulmonary fibrosis (IPF) incidence and prevalence from electronic databases without case validation may be inaccurate.Develop claims algorithms to identify IPF and assess their positive predictive value (PPV) to estimate incidence and prevalence in the United States.We developed three algorithms to identify IPF cases in the HealthCore Integrated Research Database. Sensitive and specific algorithms were developed based on literature review and consultation with clinical experts. PPVs were assessed using medical records. A third algorithm used logistic regression modeling to generate an IPF score and was validated using a separate set of medical records. We estimated incidence and prevalence of IPF using the sensitive algorithm corrected for the PPV.We identified 4,598 patients using the sensitive algorithm and 2,052 patients using the specific algorithm. After medical record review, the PPVs of these algorithms using the treating clinician's diagnosis were 44.4 and 61.7%, respectively. For the IPF score, the PPV was 76.2%. Using the clinical adjudicator's diagnosis, the PPVs were 54 and 57.6%, respectively, and for the IPF score, the PPV was 83.3%. The incidence and period prevalences of IPF, corrected for the PPV, were 14.6 per 100,000 person-years and 58.7 per 100,000 persons, respectively.Sensitive algorithms without correction for false positive errors overestimated incidence and prevalence of IPF. An IPF score offered the greatest PPV, but it requires further validation.

The purpose of this review is to assist researchers in developing, using, and interpreting case-identifying algorithms in electronic healthcare databases.We review clinical characteristics of health outcomes, data settings and informatics, and epidemiologic and statistical methods aspects as they pertain to the development and use of case-identifying algorithms.We offer a framework for thinking critically about the use of electronic health insurance data and electronic health records to identify the occurrence of health outcomes. Accuracy of case ascertainment in database research depends on many factors, including clinical and behavioral aspects of the health outcome, and details of database construction as it pertains to completeness and reliability of database content. Existing methods for diagnostic and screening tests, misclassification, validation studies, and predictive modelling can be usefully applied to improve case ascertainment in database research.Good case-identifying algorithms are based on a sound understanding of care-seeking behavior and patterns of clinical diagnosis and treatment in the study population and details about the construction and characteristics of the database. Researchers should use quantitative bias analyses to take into account the performance characteristics of case-identifying algorithms and their impact on study results.

To identify predictors of chronic obstructive pulmonary disease (COPD) severity and assess the relation between COPD severity and risk of cardiovascular outcomes. A␣cohort of patients with diagnosed and treated COPD was compiled from the Saskatchewan Health longitudinal databases. We used multivariate modeling to identify predictors of hospitalization for COPD as an indicator of COPD severity, and we used the model to characterize patients according to quintiles of COPD severity. These severity levels were used as independent variables in multivariate models of cardiovascular outcomes. Determinants of COPD severity included emphysema, recent nebulizer use, home oxygen services, corticosteroid use, frequent bronchodilator use, pneumonia and prior COPD exacerbation. The 20% of patients with the highest COPD severity were 1.27 (CI: 1.07–1.50) times more likely to have arrhythmia, 1.25 (CI: 1.07–1.46) times more likely to have ischemic heart disease, 1.38 (CI: 1.11–1.71) times more likely to have angina, 2.28 (CI: 1.95–2.66) times more likely to have congestive heart failure, and 1.63 (CI: 1.22–2.16) times more likely to die of cardiovascular causes than the least severe 20% of patients. Patients with more severe COPD, as defined by our model, had higher cardiovascular morbidity and mortality than patients with less severe COPD.

Medications used to treat respiratory diseases include beta-adrenoceptors, antimuscarinics, inhaled and oral corticosteroids, and theophyllines. Most of these drugs have been associated indirectly with cardiac rhythm disorders, but epidemiologic evidence is limited.To evaluate the association between respiratory drugs and the occurrence of rhythm disorders among patients with asthma and those with chronic obstructive pulmonary disease, we conducted a case-control study nested in a population-based cohort of individuals 10-79 years of age and registered in the U.K. General Practice Research Database after 1 January 1994. The analysis included 710 confirmed cases and 5000 controls frequency-matched to cases by age (interval of 1 year) and sex.No increased risk of arrhythmias overall was found among users of inhaled steroids (relative risk = 1.0; 95% confidence interval = 0.8-1.3). Short-term use of theophylline was weakly associated with arrhythmia (1.8; 1.0-3.3). An increased risk was found among users of oral steroids, and the relative risk was greater at the beginning of therapy (2.6; 2.0-3.5). The risk of atrial fibrillation was increased, especially for short-term use of oral steroids (2.7; 1.9-3.8), and a weak association was seen for theophyllines, especially short-term use (1.8; 0.9-3.7). Supraventricular tachycardia was associated with long-term use of oral steroids (2.1; 0.8-5.7), long-term use of antimuscarinics (1.7; 0.7-4.1), and short-term use of theophylline (4.0; 0.9-18.1). Ventricular arrhythmias were associated with oral steroids (3.2; 0.8-13.3) and beta-adrenoceptors (7.1; 0.8-65.9).Oral steroids and theophylline were the therapeutic groups associated with risk of developing atrial fibrillation, especially with new courses of therapy. Results from this study also are consistent with certain suspected dysrhythmic effects of theophyllines, with supraventricular tachycardia associated with antimuscarinics, and with ventricular arrhythmias associated with beta-adrenoceptors.

To assess the risk of serious gastrointestinal and thromboembolic complications with approved doses of meloxicam.We pooled data from clinical trials of meloxicam at doses of 7.5 or 15 mg/d. A blinded gastrointestinal adjudication committee used prespecified criteria to identify gastric or duodenal perforation, gastric outlet obstruction, or hemodynamically important upper gastrointestinal bleeding. For analysis of thromboembolic complications, investigator-reported events were analyzed without adjudication.We analyzed data from 24,196 patients from 28 trials, most of whom had been followed for up to 60 days. Of these patients, 13,118 received meloxicam (10,158 received a daily dose of 7.5 mg and 2960 received 15 mg), 5283 were treated with diclofenac 100 mg, 181 received diclofenac 150 mg, 5371 were treated with piroxicam 20 mg, and 243 received naproxen 500 mg twice daily. Patients who received 7.5 mg of meloxicam daily had a 0.03% risk of serious upper gastrointestinal events, which was significantly lower than the risk in those who received diclofenac, naproxen, or piroxicam (P0.02). With the 15 mg daily dose of meloxicam, this risk was significantly different only when compared with piroxicam (P = 0.03). The risk of thromboembolic events in patients treated with meloxicam at either dose was lower than with diclofenac, but similar to that observed with piroxicam and naproxen.This pooled analysis of 24,196 patients demonstrates that meloxicam has a favorable gastrointestinal and thromboembolic safety profile. However, only a small number of patients were followed for more than 60 days, and meaningful comparisons were not possible in this subgroup.

Background The accuracy of vaccine administration information recorded in administrative claims databases is uncertain. Methods We conducted a retrospective cohort study using the HealthCore Integrated Research Database SM among infants who received at least 1 RotaTeq ® (RV5) dose during the first year of life between February 1, 2006 and November 30, 2012 and were enrolled in the health plan at birth. We reviewed medical records for a sample of infants to validate vaccine administration information. Results We identified 169,560 infants who received at least 1 RV5 dose. Medical records were obtained for 85 infants, of which 74 (PPV1 87.1%; 95% CI 78.0–93.4%) had a corresponding first RV5 vaccination in the medical record with the same or similar administration date. Conclusions Administrative claims contained inaccuracies in dose number or administration date for 13% of RV5 first doses identified.

Meloxicam is a nonsteroidal anti-inflammatory drug (NSAID) introduced in the UK and othercountries in 1996. Clinical trials have suggested that meloxicam could have better gastrointestinal(GI) tolerability than other NSAIDs, possibly due to its preferential inhibition of COX-2isoenzyme as compared with COX-1.(1,2) Despite being the largest NSAID trials conducted todate, these studies were not sufficiently large to conclude that the apparently improved GItolerability also extends to reduced risk of upper GI perforations, ulcers and bleeds.(1,2) Thus, itwas noted in the UK Drug and Therapeutics Bulletin(3) that “there is no convincing evidence thatthe risk of the severest adverse gastrointestinal events, namely peptic ulceration, perforation andbleeding, is lower with meloxicam than with other NSAIDs.” Post-marketing experience in theUK has been consistent with clinical trials in showing that the most common kinds of adverseevents most frequently reported with meloxicam are GI events.(4) For this reason, the labelingfor meloxicam was strengthened to alert health practitioners to the possible risk of GI events.It has been stated meloxicam “shares the common unwanted effects of other NSAIDs and thesame precautions apply to its use, especially in patients at high risk for peptic ulceration.”(3)Merely because meloxicam is associated with the same kinds of GI events as other NSAIDs,however, does not mean that meloxicam carries the same risk of these events as other NSAIDs.To make matters more complicated, in the post-marketing experience where conditions are notas well-controlled as they are in clinical trials, comparisons of disease frequency typically do notallow for valid inference. Indeed, even if meloxicam were safer than other NSAIDs in the generalpopulation, rates of GI events among patients receiving meloxicam could be greater than ratesof GI events among patients receiving other NSAIDs. For example, if meloxicam wasprescribed preferentially to patients who, at the time of receiving meloxicam, were already atincreased baseline risk of GI events, one would expect to observe high rates of GI eventsassociated with meloxicam but unrelated to any adverse drug effect.This study was conducted to evaluate whether or not patients receiving meloxicam soon after itsintroduction in the UK were at inherently greater baseline risk of GI events than patients receiving

There is widespread interest in assessing the clinical importance of a study result. This goal is impeded, however, by a lack of clarity about the biological interpretability of epidemiological effect measures, such as the relative risk. A relative risk is often interpreted merely as a measure of some vague statistical association, without a view toward a biological effect as an object of measurement. Not infrequently, if it is not statistically significant, the relative risk estimate is ignored completely. A key to biological interpretation is appreciating the theoretical framework stipulating that outcome rates derived from 2 comparison groups actually represent measures of different effects in the same population. For instance, by using a placebo group to estimate the number of background cases that occurred in the treatment group, an estimate of the number of excess cases that occurred as a result of treatment can be made. This kind of biological entity can be derived from a relative risk, and can be more easily evaluated as to its clinical importance than a statistical association or a statement about statistical significance. Interpretation then becomes a more directed task, with a focus on the validity of certain ancillary hypotheses upon which biological interpretability rests.

Objective To assess the effect on FEV 1 and clinical outcomes of adding ipratropium bromide to salbutamol in the treatment of acute asthma. Methods We conducted a pooled analysis of three randomized double-blinded clinical trials conducted in the United States, Canada, and New Zealand. The studies enrolled 1,064 patients aged 18 to 55 years who presented at the emergency department with acute asthma. Patients were randomized to treatment with a combination of nebulized 2.5 mg salbutamol plus 0.5 mg ipratropium bromide, or 2.5 mg salbutamol alone. Medications were administered at baseline and, in the US study, at 45 min. FEV 1 was measured at baseline, 45 min, and 90 min. Patients were followed up for 48 h after hospital discharge for occurrence of asthma exacerbation and hospitalization. Results Treatment groups were comparable at baseline. Of the 1,064 patients randomized, 1,015 patients (95%) remained in the study for measurement at 45 min, and 961 patients (90%) completed the final measurement at 90 min. Comparison of overall improvement in FEV 1 at 45 min indicated a better response for patients receiving combination therapy (mean difference=43 mL, 95% confidence interval [CI]=—20, 107). The distribution of change in FEV 1 was skewed by a small number of patients with extreme values (38 of 1,064=3.6%) that may have been due to unreliable lung function testing. Removing these outliers produced a larger and more precise estimate of effect (mean difference=55 mL, 95% CI=2,107). Because the distribution was skewed, we performed nonparametric analyses that showed evidence of a beneficial effect of combination therapy. The difference between median values at 45 min is 40 mL (Wilcoxon p value=0.03). In addition, 4.9% (95% CI=−1%, 11%) more patients in the combination group achieved at least 20% of their potential improvement, as measured by the difference between their baseline FEV 1 and their predicted FEV 1 . Patients receiving combination therapy had lower risk for each of three clinical outcomes: the need for additional treatment (relative risk [RR]=0.92, 95% CI=0.84, 1.0), risk of asthma exacerbation (RR=0.84, 95% CI=0.67, 1.04), and risk of hospitalization (RR=0.80, 95% CI=0.61, 1.06). Conclusion Adding ipratropium bromide to salbutamol in the treatment of acute asthma produces a small improvement in lung function, and reduces the risk of the need for additional treatment, subsequent asthma exacerbations, and hospitalizations. These apparent benefits of adding ipratropium bromide were independent of the amount of β-agonist that had been used earlier in the attack, and possibly related to a recent upper respiratory tract infection. Confirmatory studies are needed, especially for clinical outcomes.

Objective. To describe the frequency and costs of medical services for patients with osteoarthritis (OA) or rheumatoid arthritis (RA) in a managed care setting. Methods. Individual utilization records of medical and pharmacy services for OA and RA patients were obtained from a group-model health maintenance organization (HMO). Estimates were made for costs of drugs and medical services for arthritis from July 1, 1993 to June 30, 1994 using Medicare reimbursement schedules and average wholesale drug prices. Calculated rates for each population were expressed as counts of events or as dollars per person-year. Results. The average individual cost rate of arthritis-related care for 365 RA patients was $2,162 per year, and the total cost of RA care to the HMO was $703,053. Prescription medications accounted for 62% ($436,440) of the total cost of RA care, while ambulatory care accounted for 21% ($150,938), and hospital visits accounted for 16% ($115,674). With regard to 10,101 OA patients, the average individual cost rate was $543 per year, and total cost to the HMO was $4,728,425. Hospital care accounted for 46% ($2,170,890) of the total cost of OA care, medications accounted for 32% ($1,509,637), and ambulatory care accounted for 22% ($1,047,898). Conclusion. RA care, in the setting of this study, was characterized by intensive treatment, especially frequent use of medications that were delivered to most patients. Although the cost of RA care per patient was high, cost to the managed care provider was relatively low, owing to the rarity of RA. OA care tended to be infrequent, and the largest component of cost was hospital care for a small proportion of patients (5%). Owing to the greater prevalence of OA, care of OA was nearly 7 times more costly to the managed care provider than was care of RA.

To evaluate the hypothesis that fenoterol or all inhaled beta-agonists caused an epidemic of asthma mortality in New Zealand from the late 1970s to the mid-1980s, we examined trends from 1970 to 1992 in per capita sales of inhaled fenoterol, inhaled beta-agonists, and asthma mortality in New Zealand and nine other countries that marketed fenoterol. During the last two decades, there has been a large and widespread increase in sales of inhaled beta-agonists, including fenoterol. Asthma mortality in most countries, however, has been relatively stable. Only New Zealand experienced an epidemic of asthma mortality. In addition, sales rates of fenoterol similar in magnitude to those in New Zealand near the peak of the epidemic also occurred in Belgium, Austria, and Germany, while asthma mortality in these countries remained low. Also, sales rates of all beta-agonists in Australia were similar to those in New Zealand, but no epidemic of asthma mortality occurred in Australia. Therefore, the difference between asthma mortality rates in New Zealand and other countries is not explained by differences in per capita sales of fenoterol or all beta-agonists. Within New Zealand, the beginning and end of the epidemic correlated with a rise and fall in sales of all beta-agonists, including fenoterol. From 1980 to 1989, however, sales of fenoterol and all beta-agonists doubled in New Zealand while asthma mortality declined by 40%. International data on medication sales and asthma mortality, therefore, do not point to a relation between asthma mortality and beta-agonists in general nor fenoterol in particular.

Growing evidence suggests that Alzheimer’s disease and other types of dementia are underdiagnosed and poorly documented. In our study, we describe patterns of dementia coding and treatment in the Veteran’s Administration New England Healthcare System. We conducted a retrospective cohort study with new outpatient ICD-9 codes for several types of dementia between 2002 and 2009. We examined healthcare utilization, medication use, initial dementia diagnoses, and changes in diagnoses over time by provider type. 8,999 veterans received new dementia diagnoses during the study period. Only 18.3% received a code for cognitive impairment other than dementia, most often “memory loss” (65.2%) prior to dementia diagnosis. Two-thirds of patients received their initial code from a PCP. The etiology of dementia was often never specified by ICD-9 code, even by specialists. Patients followed up exclusively by PCPs had lower rates of neuroimaging and were less likely to receive dementia medication. Emergency room visits and hospitalizations were frequent in all patients but highest in those seen by dementia specialists. Dementia medications are commonly used off-label. Our results suggest that, for the majority the patients, no prodrome of the dementia syndrome is documented with diagnostic code, and patients who do not see dementia specialists have less extensive diagnostic assessment and treatment.

There is evidence that earlier initiation of HIV antiretroviral therapy (ART) is associated with better outcomes, including lower morbidity and mortality. Based on recent studies indicating that Medicaid enrollees are more likely to have suboptimal access to medical care, we hypothesized that HIV severity at time of ART initiation is worse for Medicaid patients than patients with other health care coverage. We conducted a US retrospective analysis of GE Centricity Outpatient Electronic Medical Records spanning 1 January 1997 through 30 September 2009. Subjects included all adult HIV patients initiating first-line ART who had CD4+ results within 90 days pre-initiation. HIV stage was defined using CD4 ranges:500 (n=520), 351-500 (n=379), 201-350 (n=580), or ≤200 (n=406) cells/mm(3), with lower CD4 count being indicative of increased disease severity. Payer type was defined as the patient's primary payer: Medicaid, Medicare, commercial insurance, self-pay or other/unknown. After controlling for demographic and clinical covariates, cumulative logit models assessed the effect of payer type on HIV stage at ART initiation. The study included 1885 subjects with the primary payer being Medicaid (n=218), Medicare (n=330), commercial insurance (n=538), self-pay (n=159) or other/unknown (n=640). Final logit models demonstrated that, compared to patients on Medicaid, the odds of initiating ART at a higher CD4 range were significantly greater for those commercially insured (odds ratio [OR]=1.53; P=0.005), self-paying (OR=1.56; P=0.023) and other/unknown (OR=1.79; P0.001) and similar for patients enrolled in Medicare (OR=1.11; P=0.521). Medicaid patients initiated ART at a more advanced stage of HIV than patients who were commercially insured, self-paying, or had other/unknown coverage. With HIV treatment guidelines now supporting ART initiation in patients with higher CD4 counts, these findings underscore the need for mitigating barriers, particularly in the Medicaid population, that may delay treatment initiation.

Ultrasound guidance enables visualization of the needle insertion site for thoracentesis and paracentesis. The improved accuracy of needle placement using ultrasound may reduce risk of complications and their costs associated with these procedures. Using claims data from the Premier Perspective hospital database from January 1, 2007, through December 31, 2008, we conducted an observational cohort study examining the effect of ultrasound guidance on risk of pneumothorax among patients undergoing thoracentesis and on risk of bleeding complications after paracentesis. Patients at elevated risk of these outcomes for reasons beyond the procedure of interest were excluded. Adjusted risk of events was assessed using multivariate logistic regression controlling for patient and hospitalization characteristics. Hospitalization cost and length of stay (LOS) were estimated using multivariate ordinary least squares regression of log-transformed values. We analyzed 61,261 thoracentesis and 69,859 paracentesis patient records. Approximately 45% of these procedures were ultrasound guided. Pneumothorax occurred in 2.7% (n = 1,670) of patients undergoing thoracentesis. Of patients undergoing paracentesis, 0.8% (n = 565) experienced bleeding complications. After adjustment, ultrasound guidance reduced the risk of pneumothorax after thoracentesis by 19% (OR, 0.81; 95% CI, 0.74-0.90) and by 68% for bleeding complications after paracentesis (OR, 0.32; 95% CI, 0.25-0.41). Pneumothorax increased the total cost of hospitalization by $2,801 (Plt; .001) and LOS by 1.5 days (Plt; .001). Bleeding complications increased cost by $19,066 (Plt; .0001) and LOS by 4.3 days (Plt; .0001). The data indicate that ultrasound guidance is associated with decreased risk of pneumothorax with thoracentesis and of bleeding complications with paracentesis. These complications resulted in measurable increases in hospitalization costs and LOS.

Objectives To measure the association between asthma drugs and death or ICU admission due to asthma (severe life-threating attack of asthma [SLTA]), and to assess the possibility that these associations may not be causal but due to the prescription of these drugs to patients with more severe disease (confounding). Design Retrospective cohort study of 655 asthmatics who attended an emergency department in 1986 to 1987 followed till death or May 1989. Methods Outcome events were death or ICU admission due to asthma (SLTA). All hospital attendances were identified and patients classified at each according to drug exposure and a wide variety of measures of asthma severity. Incidence rates were computed as total outcome events divided by person-time contributed for each subject classified according to drug use and asthma severity. Rate ratio (RR) estimates for severe asthma outcomes associated with use as compared to nonuse of asthma drugs were calculated. Severity markers were identified and used to adjust the crude RR estimates. Results One hundred five SLTAs (15 deaths, 90 ICU admissions) occurred in 66 patients. Like inhaled fenoterol, oral β-agonists, theophylline, cromolyn, inhaled steroids, and oral steroids were all associated with an increased risk of SLTA. When adjusted progressively for measures of severity, these increased risks became insignificant except for cromolyn. Conclusion Unadjusted RR estimates for severe asthma events comparing exposure to a particular drug with nonuse are overestimates due to confounding. Control with two severity markers (hospital admission in the last year, use of oral corticosteroid at the time of previous admission) removes some confounding but control for additional severity markers not available in previous studies reduces the effect estimates further. These results suggest that the problem of confounding is substantial in nonrandomized epidemiologic studies of asthma drugs. Previous studies reporting RR estimates are likely to be confounded.

In order to characterise asthma management in a managed care setting, we identified 10,301 patients who were diagnosed with asthma between 1 January 1988 and 31 December 1991 at a group model health maintenance organisation in central Massachusetts, US. We obtained for these patients automated utilisation files containing data on medications, hospitalisations, emergency room visits, office visits, and estimated costs of these services. The medication dispensed to the greatest proportion of patients was beta 2 agonists either by inhalation (56%) or orally (21%). Theophylline was dispensed to 23% of the patients. Maintenance therapy was inhaled anti-inflammatory medication was uncommon, as inhaled corticosteroids (17%) and sodium cromoglycate (cromolyn sodium) [8%] were dispensed to fewer patients than other asthma medications. Among patients who had been hospitalised in the previous year, 36% were presently receiving inhaled corticosteroids, and among patients who used at least one beta 2 agonist metered-dose inhaler per month, 49% were presently receiving inhaled corticosteroids. Economic analyses showed that only 8% of the patients had either a hospital admission or an emergency room visit, but hospital costs among these patients accounted for 25% of the total costs of asthma care. In addition, the top 10% most expensive patients accounted for 42% of the total cost of asthma care. We conclude that a substantial proportion of patients at increased risk of a severe attack, by virtue of having a recent hospitalisation, do not receive maintenance anti-inflammatory therapy, and that hospitalisations among a relatively small proportion of asthma patients contribute significantly to the cost of asthma care.

The growing understanding of the use of biomarkers in Alzheimer's disease (AD) may enable physicians to make more accurate and timely diagnoses. Florbetaben, a beta-amyloid tracer used with positron emission tomography (PET), is one of these diagnostic biomarkers. This analysis was undertaken to explore the potential value of florbetaben PET in the diagnosis of AD among patients with suspected dementia and to identify key data that are needed to further substantiate its value. A discrete event simulation was developed to conduct exploratory analyses from both US payer and societal perspectives. The model simulates the lifetime course of disease progression for individuals, evaluating the impact of their patient management from initial diagnostic work-up to final diagnosis. Model inputs were obtained from specific analyses of a large longitudinal dataset from the New England Veterans Healthcare System and supplemented with data from public data sources and assumptions. The analyses indicate that florbetaben PET has the potential to improve patient outcomes and reduce costs under certain scenarios. Key data on the use of florbetaben PET, such as its influence on time to confirmation of final diagnosis, treatment uptake, and treatment persistency, are unavailable and would be required to confirm its value.