Methodological framework to identify possible adverse drug reactions using population-based administrative data.

PURPOSE: We present a framework for detecting possible adverse drug reactions (ADRs) using the Utah Medicaid administrative data. We examined four classes of ADRs associated with treatment of dementia by acetylcholinesterase inhibitors (AChEIs): known reactions (gastrointestinal, psychological disturbances), potential reactions (respiratory disturbance), novel reactions (hepatic, hematological disturbances), and death. METHODS: Our cohort design linked drug utilization data to medical claims from Utah Medicaid recipients. We restricted the analysis to 50 years-old and older beneficiaries diagnosed with dementia-related diseases. We compared patients treated with AChEI to patients untreated with anti-dementia medication therapy. We attempted to remove confounding by establishing propensity-score-matched cohorts for each outcome investigated; we then evaluated the effects of drug treatment by conditional multivariable Cox-proportional-hazard regression. Acute and transient effects were evaluated by a crossover design using conditional logistic regression. RESULTS: Propensity-matched analysis of expected reactions revealed that AChEI treatment was associated with gastrointestinal episodes (Hazard Ratio [HR]: 2.02; 95%CI: 1.28-3.2), but not psychological episodes, respiratory disturbance, or death. Among the unexpected reactions, the risk of hematological episodes was higher (HR: 2.32; 95%CI: 1.47-3.6) in patients exposed to AChEI. AChEI exposure was not associated with an increase in hepatic episodes. We also noted a trend, identified in the case-crossover design, toward increase odds of experiencing acute hematological events during AChEI exposure (Odds Ratio: 3.0; 95% CI: 0.97 - 9.3). CONCLUSIONS: We observed an expected association between AChEIs treatment and gastrointestinal disturbances and detected a signal of possible hematological ADR after treatment with AChEIs in this pilot study. Using this analytic framework may raise awareness of potential ADEs and generate hypotheses for future investigations. Early findings, or signal detection, are considered hypothesis generating since confirmatory studies must be designed to determine if the signal represents a true drug safety problem.

Consumer-directed care for beneficiaries with mental illness: lessons from New Jersey's Cash and Counseling program.

OBJECTIVE: Previous research from the Cash and Counseling Demonstration and Evaluation (CCDE) programs in New Jersey, Arkansas, and Florida suggests that consumers' control over personal care greatly improves their satisfaction with care arrangements and their outlook on life. Still, some argue that consumer-directed care may not be appropriate for consumers with a diagnosis of mental illness. This study examined the effectiveness of the CCDE program for those with a diagnosis of mental illness. METHODS: This study examined nonelderly Medicaid beneficiaries in New Jersey with a diagnosis of mental illness and compared and contrasted the experiences of those in New Jersey's CCDE program (N=109) and those who received services provided by an agency (N=119). Logistic regression analyses were performed on baseline and nine-month follow-up data. RESULTS: By examining outcome measures--including satisfaction with care arrangements, consumers' perceptions of paid caregivers' attitudes, unmet needs, adverse events, and satisfaction with life--this study offers evidence that, from the perspective of consumers, the CCDE program is appropriate for participants with a mental illness diagnosis. For most outcome measures the CCDE program demonstrated a positive effect after baseline characteristics were controlled for. The analysis of measures of adverse events, health problems, and general health status did not yield statistically significant differences between the control group and the treatment group, indicating that CCDE care was at least as safe as agency-directed care. CONCLUSIONS: Considering the growing need for long-term care services and the limited resources available, a consumer-directed option can be a valuable alternative for persons with a diagnosis of mental illness.

Understanding medication compliance and persistence from an economics perspective.

OBJECTIVES: An increased understanding of the reasons for noncompliance and lack of persistence with prescribed medication is an important step to improve treatment effectiveness, and thus patient health. Explanations have been attempted from epidemiological, sociological, and psychological perspectives. Economic models (utility maximization, time preferences, health capital, bilateral bargaining, stated preference, and prospect theory) may contribute to the understanding of medication-taking behavior. METHODS: Economic models are applied to medication noncompliance. Traditional consumer choice models under a budget constraint do apply to medication-taking behavior in that increased prices cause decreased utilization. Nevertheless, empiric evidence suggests that budget constraints are not the only factor affecting consumer choice around medicines. Examination of time preference models suggests that the intuitive association between time preference and medication compliance has not been investigated extensively, and has not been proven empirically. The health capital model has theoretical relevance, but has not been applied to compliance. Bilateral bargaining may present an alternative model to concordance of the patient-prescriber relationship, taking account of game-playing by either party. Nevertheless, there is limited empiric evidence to test its usefulness. Stated preference methods have been applied most extensively to medicines use. RESULTS: Evidence suggests that patients' preferences are consistently affected by side effects, and that preferences change over time, with age and experience. Prospect theory attempts to explain how new information changes risk perceptions and associated behavior but has not been applied empirically to medication use. CONCLUSIONS: Economic models of behavior may contribute to the understanding of medication use, but more empiric work is needed to assess their applicability.

A simulation-based evaluation of methods to estimate the impact of an adverse event on hospital length of stay.

INTRODUCTION: We used agent-based simulation to examine the problem of time-varying confounding when estimating the effect of an adverse event on hospital length of stay. Conventional analytic methods were compared with inverse probability weighting (IPW). METHODS: A cohort of hospitalized patients, at risk for experiencing an adverse event, was simulated. Synthetic individuals were assigned a severity of illness score on admission. The score varied during hospitalization according to an autoregressive equation. A linear relationship between severity of illness and the logarithm of the discharge rate was assumed. Depending on the model conditions, adverse event status was influenced by prior severity of illness and, in turn, influenced subsequent severity. Conditions were varied to represent different levels of confounding and categories of effect. The simulation output was analyzed by Cox proportional hazards regression and by a weighted regression analysis, using the method of IPW. The magnitude of bias was calculated for each method of analysis. RESULTS: Estimates of the population causal hazard ratio based on IPW were consistently unbiased across a range of conditions. In contrast, hazard ratio estimates generated by Cox proportional hazards regression demonstrated substantial bias when severity of illness was both a time-varying confounder and intermediate variable. The direction and magnitude of bias depended on how severity of illness was incorporated into the Cox regression model. CONCLUSIONS: In this simulation study, IPW exhibited less bias than conventional regression methods when used to analyze the impact of adverse event status on hospital length of stay.

Out-of-pocket pharmacy expenditures for veterans under medicare part D.

OBJECTIVES: Because the VA pharmacy benefit is deemed equivalent coverage to Medicare Part D, veterans can use either or both of these Federal benefits. We sought to determine how these programs' different benefit structures and low-income assistance thresholds would affect pharmacy out-of-pocket expenses for veterans. METHODS: We reviewed income and asset tests performed at the Salt Lake City VA in fiscal year 2005, and estimated the number of individuals, age 65 and older, who meet eligibility for Part D low-income assistance. Using past VA pharmacy utilization data, we estimated the difference in pharmacy out-of-pocket expenditures for veterans eligible for assistance through Medicare but not through the VA. RESULTS: The income and asset thresholds for low-income assistance through Part D were reached by 4127 veterans. From this group, we identified 926 veterans who had used the VA pharmacy during the prior year, who are ineligible for VA copayment waivers, and who qualify for premium waiver under Part D. These veterans' estimated annual savings ranged from $6 to $714, with an average savings of $353 per year (or 2% of their average annual income) by using Part D. CONCLUSIONS: Although VA pharmacy coverage has been deemed to be, on average, equivalent to Part D, some veterans living near poverty can reduce out-of-pocket expenditures by using Medicare prescription coverage. Currently available data can identify veterans who are likely to achieve savings under Medicare.

Developing indicators of inpatient adverse drug events through nonlinear analysis using administrative data.

BACKGROUND: Because of uniform availability, hospital administrative data are appealing for surveillance of adverse drug events (ADEs). Expert-generated surveillance rules that rely on the presence of International Classification of Diseases, 9th Revision Clinical Modification (ICD-9-CM) codes have limited accuracy. Rules based on nonlinear associations among all types of available administrative data may be more accurate. OBJECTIVES: By applying hierarchically optimal classification tree analysis (HOCTA) to administrative data, derive and validate surveillance rules for bleeding/anticoagulation problems and delirium/psychosis. RESEARCH DESIGN: Retrospective cohort design. SUBJECTS: A random sample of 3987 admissions drawn from all 41 Utah acute-care hospitals in 2001 and 2003. MEASURES: Professional nurse reviewers identified ADEs using implicit chart review. Pharmacists assigned Medical Dictionary for Regulatory Activities codes to ADE descriptions for identification of clinical groups of events. Hospitals provided patient demographic, admission, and ICD9-CM data. RESULTS: Incidence proportions were 0.8% for drug-induced bleeding/anticoagulation problems and 1.0% for drug-induced delirium/psychosis. The model for bleeding had very good discrimination and sensitivity at 0.87 and 86% and fair positive predictive value (PPV) at 12%. The model for delirium had excellent sensitivity at 94%, good discrimination at 0.83, but low PPV at 3%. Poisoning and adverse event codes designed for the targeted ADEs had low sensitivities and, when forced in, degraded model accuracy. CONCLUSIONS: Hierarchically optimal classification tree analysis is a promising method for rapidly developing clinically meaningful surveillance rules for administrative data. The resultant model for drug-induced bleeding and anticoagulation problems may be useful for retrospective ADE screening and rate estimation.

Inpatient care in rural hospitals at the beginning of the 21st century.

CONTEXT: National data documenting the role that rural hospitals play in providing inpatient care to patients both younger than 65 and 65 years and older has previously been unavailable. PURPOSE: To present descriptive nationally representative data on the numbers and types of inpatients, and the care they received, in rural hospitals. METHODS: This study includes inpatient data from the 2001 National Hospital Discharge Survey, a nationally representative survey of short-stay, nonfederal hospitals in the United States. Inpatients in rural hospitals were compared to those in urban hospitals in terms of demographic and clinical characteristics and patterns of utilization. Among the variables examined were age, number and type of diagnoses, avoidable hospitalizations, comorbidity, procedures received, source of payment, average length of stay, and discharge disposition. FINDINGS: Seventeen percent (5.7 million) of hospitalizations were in rural hospitals in 2001 and a similar percent of the US population lived in rural areas. Rural hospitals provided 23 million days of inpatient care and 4.7 million inpatient procedures. Despite the emphasis placed on Medicare's role in supporting rural hospitals, half of rural hospital inpatients were younger than 65 years. Rural hospital inpatients had shorter average stays and received fewer procedures on average. Seven percent of rural hospital inpatients were transferred to other short-stay hospitals. CONCLUSIONS: National data on the broad scope of patients served and inpatient services provided by rural hospitals illustrate one important role these hospitals play in serving rural communities.

Medication use among Medicaid users of home and community-based services.

This study examines factors associated with access to medications for older persons and younger adults with physical disabilities who use Medicaid home and community-based services (HCBS). Using data from a six-State survey, we find the large majority of this population receives the medications it needs and has needed assistance in taking these drugs. However, 6.3 percent of this population reports unmet medication needs with financial access listed as the primary reason. Lack of transportation increases the probability of unmet need, while increased satisfaction with paid help lowers this probability. This study highlights the need to study the interaction of Medicaid, social services programs, and financial barriers for this disabled population.

Ambulatory care visits for treating adverse drug effects in the United States, 1995-2001.

BACKGROUND: Adverse d[rug events (ADEs) are a well-recognized patient safety 4concern, but their magnitude is unknown. Ambulatory viisits for treating adverse drug effects (VADEs) as recordeed in national surveys offer an alternative way to estimatte the national prevalence of ADEs because each VA]DE indicates that an ADE occurred and was seriousenough to require care. METHODS: A nationallyrepresentative sample of visits to physician offices, hospital outpatient departments, and emergency departments was analyzed. VADEs were identified as tthe first-listed cause of injury. RESULTS: In 2001, there Awere 4.3 million VADEs in the United States, averaging 15 visits per 1,000 population. VADE rates at physicianoffices, hospital outpatient departments, and hospittal emergency departments were at 3.7, 3.4, and 7.3 lper 1,000 visits, respectively. There was an upward tr'end in the total number of VADEs from 1995 to 2001 ((p < .05), but the increases in VADEs per 1000 visits an.d per 1,000 population were not statistically significant. VADEs were lower in children younger than 15 and higher in the elderly aged 65-74 than in adults aged 225-44 (p < .01) and were more frequent in females than irn males (p < .05). DISCUSSION: Although methodologically conservative, the study suggests that ADEs are a significant threat to patient safety in the United States.

Physicians' prescribing responses to a restricted formulary: the impact of Medicaid preferred drug lists in Illinois and Louisiana.

OBJECTIVE: To determine the extent to which the strictness of the criteria used to control utilization of drugs not on a Medicaid preferred drug list (PDL) changes physician prescribing habits for non-Medicaid patients. STUDY DESIGN: Quasi-experimental design based on different timing of states' enactment of PDLs for cardiovascular medications. METHODS: A regression model was developed to analyze the effect that PDL implementation had on off-PDL product market share across 3 patient types (Medicaid, third-party insurance, and cash paying). The models included data from 2 states (Illinois and Louisiana) with different PDL prior-authorization criteria. The data allowed examination of different physician responses according to the proportion of Medicaid patients treated by the physician. The analysis also followed prescribing patterns longitudinally to determine whether the PDL-induced prescribing behaviors changed over time. RESULTS: There was a decrease of 9 percentage points (67.7%) and 6.2 percentage points (40.5%) in the Medicaid prescription share of restricted cardiovascular drugs post-PDL for Illinois and Louisiana, respectively. In the third-party insurance market, prescription shares of off-PDL drugs decreased 0.9 percentage points (6.8%) in Illinois and 1.3 percentage points (8.6%) in Louisiana. For physicians with a high percentage of prescriptions paid for by Medicaid, the share loss for off-PDL drugs was estimated to be more than 37% for the non-Medicaid portion of the practice. CONCLUSION: The effects of a Medicaid PDL on prescribing behavior extend beyond the Medicaid population. The health outcomes and economic consequences of these "spillover" effects are poorly understood and warrant further research.

Gastric bypass as treatment for obesity: trends, characteristics, and complications.

OBJECTIVE: This paper describes national trends in gastric bypass procedures from 1998 through 2003 and explores the demographic and health profile of those who receive this procedure. Short-term outcomes such as length of stay and in-hospital complication rates are also examined. RESEARCH METHODS AND PROCEDURES: Data on obese hospital inpatients who had gastric bypass were obtained from the 1998 to 2003 National Hospital Discharge Survey. Gastric bypass was reported for an estimated 288,000 discharges during the 6-year study period. Trends within the 6-year period were tested using weighted regression. Characteristics of gastric bypass patients were compared with those of other inpatients using a chi(2) test of independence and the two-sided t test. RESULTS: The estimated number of hospital discharges with gastric bypass increased significantly, from 14,000 in 1998 to 108,000 in 2003. During this period, the average length of stay declined by 56% from 7.2 to 3.2 days. Gastric bypass patients were primarily women (84%), 25 to 54 years of age (82%), and privately insured (76%). A 1 in 10 complication rate was found for discharges with gastric bypass. DISCUSSION: Gastric bypass procedures in the United States have increased rapidly since 1998, whereas the average hospital stay has decreased. The decreasing length of stay needs to be evaluated in conjunction with potential complication rates and the permanent change in anatomy and lifestyle that must accompany this procedure. Monitoring trends in use of this procedure is important, especially if reimbursement policies change and the epidemic of obesity continues.

The relationship between HAART use and employment for HIV-positive individuals: an empirical analysis and policy outlook.

OBJECTIVE: This study jointly analyzes the determinants of the use of highly active antiretroviral therapy (HAART) along with the determinants of employment status for human immunodeficiency virus (HIV)-positive individuals. METHODS: Data from the HIV Cost and Services Utilization Study (HCSUS) were analyzed using a bivariate probit model. HSCUS collected data on 2864 individuals in 180 clinics, hospitals and private practices in 28 urban areas and 24 clusters of rural counties in the United States. We used data from the baseline interview to categorize individuals as employed or unemployed and for the determination of HAART use. RESULTS: The empirical results suggest that individuals taking HAART have an increased likelihood of working and that individuals with private health insurance are more likely to use HAART compared to individuals with public health insurance coverage or no coverage. CONCLUSIONS: Due to the fact that employment of HIV-positive is directly related to HAART use, policymakers need to look to the private health insurance industry and public program to increase access to HAART. Suggested models for consideration are mandating insurance benefits and ADAP expansion.

Effects of an increase in prescription copayment on utilization of low-sedating antihistamines and nasal steroids.

BACKGROUND: Health plans are using 3-tier copayment designs and other methods to control utilization that shifts drug costs to plan members. There is a need to determine the effects of increased member cost sharing on drug utilization and drug costs. OBJECTIVE: To assess the impact of a 10 US dollars increase in prescription copayment in a public employer health plan for 2 classes of drugs used for allergic rhinitis. METHODS: Changes in the number of prescriptions dispensed for 2 therapeutic classes.low-sedating antihistamines (LSAs) and nasal steroids (NSs).were examined 1 year prior to and 1 year after copayment increase. Relative price effects were measured as arc price elasticity, the ratio of the percent change in prescription utilization over the percent change in price, an indicator of how responsive patients are to the copayment increase. RESULTS: Of 8,643 continuously enrolled health plan beneficiaries, 2,150 patients (24.8%) received at least 1 NS or LSA during the 2-year period of the study, from January 1, 1998, through December 31, 1999. An average 10 US dollars increase in copayment per prescription was associated with no statistically significant change in utilization of combined LSA and NS prescriptions, 2.89 per patient in 1998 and 2.94 in 1999 (P = 0.597). Health plan costs for study drugs, unadjusted for inflation, decreased by 16.3% from 86.86 US dollars per patient in 1998 to 72.68 US dollars in 1999 (P = 0.004). Health plan costs per patient per month (PPPM) for all drugs for the 2,150 allergic rhinitis patients decreased by 13% from 41.33 US dollars PPPM in 1998 to 35.93 US dollars in 1999 (P<0.001), and health plan drug costs for all 8,643 members decreased by 13% from 14.93 US dollars per member per month (PMPM) in 1998 to 12.99 US dollars in 1999 (P<0.001). The actual average copayment increase was 7.23 US dollars (a 41% increase) for LSAs, which was associated with a 14.8% increase in utilization of LSAs and an 11.8% increase in the number of patients using LSAs; the number of LSA prescriptions per patient per year was unchanged at 2.68 in 1999 versus 2.61 in 1998 (P = 0.429). The actual average copayment increase was 10.98 US dollars (71%) for NSs, which was associated with an 11.3% decrease in utilization of NSs and a 10.2% decrease in the number of users of nasals steroids in 1999; the number of nasal steroid prescriptions per patient per year was unchanged at 2.05 in 1999 versus 2.07 in 1998 (P =.842). The combined utilization of LSA and NS prescriptions increased by 8.9% following the increase in copayments for these 2 therapeutically interchangeable drugs for allergic rhinitis. LSA prescriptions were less elastic, with an unadjusted arc elasticity of 0.39, while nasal steroid prescriptions were more responsive to the copayment change, with an unadjusted arc elasticity of.0.22. CONCLUSIONS: An average 10 US dollars increase in patient cost sharing per prescription (46.9% copayment increase) was associated with an increase in combined utilization of 2 drug classes used for allergic rhinitis (LSAs and NSs) but no change in the number of prescriptions per patient. Health plan costs decreased significantly for allergic rhinitis drugs, all drugs used by allergic rhinitis patients, and all drugs used by continuously enrolled health plan members. NSs exhibited a greater arc price elasticity compared with low-sedating oral antihistamines.

Obesity and physical inactivity in rural America.

CONTEXT AND PURPOSE: Obesity and physical inactivity are common in the United States, but few studies examine this issue within rural populations. The present study uses nationally representative data to study obesity and physical inactivity in rural populations. METHODS: Data came from the 1998 National Health Interview Survey Sample Adult and Adult Prevention Module. Self-reported height and weight were used to calculate body mass index. Physical inactivity was defined using self-reported leisure-time physical activity. Analyses included descriptive statistics, chi 2 tests, and logistic regression. FINDINGS: Obesity was more common among rural (20.4%, 95% CI 19.2%-21.6%) than urban adults (17.8%, 95% CI 17.2%-18.4%). Rural residents of every racial/ethnic group were at higher risk of obesity than urban whites, other factors held equal. Other predictors of obesity included being male, age 25-74, lacking a high school diploma, having physical limitations, fair to poor health, and a history of smoking. Proportionately more rural adults were physically inactive than their urban peers (62.8% versus 59.3%). Among rural residents, minorities were not significantly more likely to be inactive than whites. Males and younger adults were less likely to be inactive. Rural adults who were from the Midwest and South, had less than a high school education, had fair to poor health, and currently smoked were more likely to be inactive compared to their respective referent group. CONCLUSIONS: The high prevalence of obesity and inactive lifestyles among rural populations call for research into effective rural interventions.

Return to work and claim duration for workers with long-term mental disabilities: impacts of mental health coverage, fringe benefits, and disability management.

This paper examines the relationship of mental health benefits provided by 116 employers, to return to work and duration of disability claims for 407 of their employees who were on long-term disability (LTD) leave for mental disorders. Mental health benefits data were coded from summary plan description booklets provided by employers. Information on other fringe benefits and employers' disability management practices were obtained from a survey of the employers. Relationships between mental health benefit features, other fringe benefit and disability management factors, and our outcomes were estimated via logistic regression and survival analysis. Results indicated that three mental health benefit plan features were negatively related to the return-to-work probability:(1) a high deductible (> $600), (2) longer preexisting condition exclusion periods, and (3) having a carve-out. This suggests that cost saved by access restrictions may be partially offset by higher turnover costs for employees with disabilities due to mental disorders. Carve-outs were also predictive of shorter claims duration.