Large language model aided automatic high-throughput drug screening using self-controlled cohort study

Background

Developing medicine from scratch to governmental authorization and detecting adverse drug reactions (ADR) have barely been economical, expeditious, and risk-averse investments. The availability of large-scale observational healthcare databases and the popularity of large language models offer an unparalleled opportunity to enable automatic high-throughput drug screening for both repurposing and pharmacovigilance.

Objectives

To demonstrate a general workflow for automatic high-throughput drug screening with the following advantages: (i) the association of various exposure on diseases can be estimated; (ii) both repurposing and pharmacovigilance are integrated; (iii) accurate exposure length for each prescription is parsed from clinical texts; (iv) intrinsic relationship between drugs and diseases are removed jointly by bioinformatic mapping and large language model - ChatGPT; (v) causal-wise interpretations for incidence rate contrasts are provided.

Methods

Using a self-controlled cohort study design where subjects serve as their own control group, we tested the intention-to-treat association between medications on the incidence of diseases. Exposure length for each prescription is determined by parsing common dosages in English free text into a structured format. Exposure period starts from initial prescription to treatment discontinuation. A same exposure length preceding initial treatment is the control period. Clinical outcomes and categories are identified using existing phenotyping algorithms. Incident rate ratios (IRR) are tested using uniformly most powerful (UMP) unbiased tests.

Results

We assessed 3,444 medications on 276 diseases on 6,613,198 patients from the Clinical Practice Research Datalink (CPRD), an UK primary care electronic health records (EHR) spanning from 1987 to 2018. Due to the built-in selection bias of self-controlled cohort studies, ingredients-disease pairs confounded by deterministic medical relationships are removed by existing map from RxNorm and nonexistent maps by calling ChatGPT. A total of 16,901 drug-disease pairs reveals significant risk reduction, which can be considered as candidates for repurposing, while a total of 11,089 pairs showed significant risk increase, where drug safety might be of a concern instead.

Conclusions

This work developed a data-driven, nonparametric, hypothesis generating, and automatic high-throughput workflow, which reveals the potential of natural language processing in pharmacoepidemiology. We demonstrate the paradigm to a large observational health dataset to help discover potential novel therapies and adverse drug effects. The framework of this study can be extended to other observational medical databases.

Causal inference in medical records and complementary systems pharmacology for metformin drug repurposing towards dementia

Metformin, a diabetes drug with anti-aging cellular responses, has complex actions that may alter dementia onset. Mixed results are emerging from prior observational studies. To address this complexity, we deploy a causal inference approach accounting for the competing risk of death in emulated clinical trials using two distinct electronic health record systems. In intention-to-treat analyses, metformin use associates with lower hazard of all-cause mortality and lower cause-specific hazard of dementia onset, after accounting for prolonged survival, relative to sulfonylureas. In parallel systems pharmacology studies, the expression of two AD-related proteins, APOE and SPP1, was suppressed by pharmacologic concentrations of metformin in differentiated human neural cells, relative to a sulfonylurea. Together, our findings suggest that metformin might reduce the risk of dementia in diabetes patients through mechanisms beyond glycemic control, and that SPP1 is a candidate biomarker for metformin's action in the brain.

Early analysis shows that endoscopic flexor hallucis longus transfer has a promising cost-effectiveness profile in the treatment of acute Achilles tendon ruptures

PURPOSE

Current options for treating an Achilles tendon rupture (ATR) include conservative and surgical approaches. Endoscopic flexor hallucis longus (FHL) transfer has been recently proposed to treat acute ruptures, but its cost-effectiveness potential remains to be evaluated. Therefore, the objective of this study was to perform an early cost-effectiveness analysis of endoscopic FHL transfer for acute ATRs, comparing the costs and benefits of current treatments from a societal perspective.

METHODS

A conceptual model was created, with a decision tree, to outline the main health events during the treatment of an acute ATR. The model was parameterized using secondary data. A systematic review of the literature was conducted to gather information on the outcomes of current treatments. Data related to outcomes of endoscopic FHL transfers in acute Achilles ruptures was obtained from a single prospective study. Analysis was limited to the two first years. The incremental cost-effectiveness ratio was the main outcome used to determine the preferred strategy. A willingness-to-pay threshold of $100,000 per quality-adjusted life-year was used. Sensitivity analyses were performed to determine whether changes in input parameters would cause significant deviation from the reference case results. Specifically, a probability sensitivity analysis was conducted using Monte Carlo simulations, and a one-way sensitivity analysis was conducted by sequentially varying each model parameter within a given range.

RESULTS

For the reference case, incremental cost-effectiveness ratios exceeded the willingness-to-pay threshold for all the surgical approaches. Overall, primary treatment was the main cost driver. Conservative treatment showed the highest direct costs related to the treatment of complications. In the probabilistic sensitivity analysis, at a willingness-to-pay threshold of $100,000, open surgery was cost-effective in 50.9%, minimally invasive surgery in 55.8%, and endoscopic FHL transfer in 72% of the iterations. The model was most sensitive to parameters related to treatment utilities, followed by the costs of primary treatments.

CONCLUSION

Surgical treatments have a moderate likelihood of being cost-effective at a willingness-to-pay threshold of $100,000, with endoscopic FHL transfer showing the highest likelihood. Following injury, interventions to improve health-related quality of life may be better suited for improved cost-effectiveness.

LEVEL OF EVIDENCE

Level III.

Economics of Beta-Cell Replacement Therapy

PURPOSE OF REVIEW

Type 1 diabetes impacts 1.3 million people in the USA with a total direct lifetime medical cost of $133.7 billion. Management requires a mix of daily exogenous insulin administration and frequent glucose monitoring. Decision-making by the individual can be burdensome.

RECENT FINDINGS

Beta-cell replacement, which involves devices protecting cells from autoimmunity and allo-rejection, aims at restoring physiological glucose regulation and improving clinical outcomes in patients. Given the significant burden of T1D in the healthcare systems, cost-effectiveness analyses can drive innovation and policymaking in the area. This review presents the health economics analyses performed for donor-derived islet transplantation and the possible outcomes of stem cell-derived beta cells. Long-term cost-effectiveness of islet transplantation depends on the engraftment of these transplants, and the expenses and thresholds assumed by healthcare systems in different countries. Early health technology assessment analyses for stem cell-derived beta-cell replacement suggest manufacturing optimization is necessary to reduce upfront costs.

Modeling biological and economic uncertainty on cell therapy manufacturing: the choice of culture media supplementation

AIM

To evaluate the cost-effectiveness of autologous cell therapy manufacturing in xeno-free conditions.

MATERIALS & METHODS

Published data on the isolation and expansion of mesenchymal stem/stromal cells introduced donor, multipassage and culture media variability on cell yields and process times on adherent culture flasks to drive cost simulation of a scale-out campaign of 1000 doses of 75 million cells each in a 400 square meter Good Manufacturing Practices facility.

RESULTS & CONCLUSION

Passage numbers in the expansion step are strongly associated with isolation cell yield and drive cost increases per donor of $1970 and 2802 for fetal bovine serum and human platelet lysate. Human platelet lysate decreases passage numbers and process costs in 94.5 and 97% of donors through lower facility and labor costs. Cost savings are maintained with full equipment depreciation and higher numbers of cells per dose, highlighting the number of cells per passage step as the key cost driver.

Nosocomial amplification of MERS-coronavirus in South Korea, 2015

Background

Nosocomial amplification resulted in nearly 200 cases of Middle East respiratory syndrome (MERS) during the 2015 South Korean MERS-coronavirus outbreak. It remains unclear whether certain types of cases were more likely to cause secondary infections than others, and if so, why.

Methods

Publicly available demographic and transmission network data for all cases were collected from the Ministry of Health and Welfare. Statistical analyses were conducted to determine the relationship between demographic characteristics and the likelihood of human-to-human transmission. Findings from the statistical analyses were used to inform a hypothesis-directed literature review, through which mechanistic explanations for nosocomial amplification were developed.

Results

Cases that failed to recover from MERS were more likely to cause secondary infections than those that did. Increased probability of direct, human-to-human transmission due to clinical manifestations associated with death, as well as indirect transmission via environmental contamination (e.g., fomites and indoor ventilation systems), may serve as mechanistic explanations for nosocomial amplification of MERS-coronavirus in South Korea.

Conclusions

In addition to closely monitoring contacts of MERS cases that fail to recover during future nosocomial outbreaks, potential fomites with which they may have had contact should be sanitized. Furthermore, indoor ventilation systems that minimize recirculation of pathogen-bearing droplets should be implemented whenever possible.

Medical Students Who Pursue a Joint MD/MBA Degree: Who Are They and Where Are They Heading?

Increasingly, health care is being delivered in large, complex organizations, and physicians must learn to function effectively in them. As a result, several medical and business schools have developed joint programs to train physician leaders who receive both medical degree (MD) and master of business administration (MBA) degrees. We examined several themes in relation to these programs, revolving around concerns about who is attracted to them and whether exposure to the differing cultures of medicine and business have an impact on the professional identities of their graduates as manifested in their motivations, aspirations, and careers. We addressed these issues by studying students in the joint MD/MBA program at Harvard Medical School (HMS) and Harvard Business School (HBS). Our data came from several internal sources and a survey of all students enrolled in the joint program in spring 2013. We found relatively few differences between joint program students and equivalent cohorts of HMS students in terms of personal characteristics, preadmission performance, and performance at HMS and HBS. Contrary to the concerns that such programs may draw students away from medicine, the vast majority embraced careers involving extensive postgraduate medical training, with long-term plans that leveraged their new perspectives and skills to improve health care delivery.

Disease-based modeling to predict fluid response in intensive care units

OBJECTIVE

To compare general and disease-based modeling for fluid resuscitation and vasopressor use in intensive care units.

METHODS

Retrospective cohort study involving 2944 adult medical and surgical intensive care unit (ICU) patients receiving fluid resuscitation. Within this cohort there were two disease-based groups, 802 patients with a diagnosis of pneumonia, and 143 patients with a diagnosis of pancreatitis. Fluid resuscitation either progressing to subsequent vasopressor administration or not was used as the primary outcome variable to compare general and disease-based modeling.

RESULTS

Patients with pancreatitis, pneumonia and the general group all shared three common predictive features as core variables, arterial base excess, lactic acid and platelets. Patients with pneumonia also had non-invasive systolic blood pressure and white blood cells added to the core model, and pancreatitis patients additionally had temperature. Disease-based models had significantly higher values of AUC (p < 0.05) than the general group (0.82 ± 0.02 for pneumonia and 0.83 ± 0.03 for pancreatitis vs. 0.79 ± 0.02 for general patients).

CONCLUSIONS

Disease-based predictive modeling reveals a different set of predictive variables compared to general modeling and improved performance. Our findings add support to the growing body of evidence advantaging disease specific predictive modeling.

Reducing unnecessary lab testing in the ICU with artificial intelligence

OBJECTIVES

To reduce unnecessary lab testing by predicting when a proposed future lab test is likely to contribute information gain and thereby influence clinical management in patients with gastrointestinal bleeding. Recent studies have demonstrated that frequent laboratory testing does not necessarily relate to better outcomes.

DESIGN

Data preprocessing, feature selection, and classification were performed and an artificial intelligence tool, fuzzy modeling, was used to identify lab tests that do not contribute an information gain. There were 11 input variables in total. Ten of these were derived from bedside monitor trends heart rate, oxygen saturation, respiratory rate, temperature, blood pressure, and urine collections, as well as infusion products and transfusions. The final input variable was a previous value from one of the eight lab tests being predicted: calcium, PTT, hematocrit, fibrinogen, lactate, platelets, INR and hemoglobin. The outcome for each test was a binary framework defining whether a test result contributed information gain or not.

PATIENTS

Predictive modeling was applied to recognize unnecessary lab tests in a real world ICU database extract comprising 746 patients with gastrointestinal bleeding.

MAIN RESULTS

Classification accuracy of necessary and unnecessary lab tests of greater than 80% was achieved for all eight lab tests. Sensitivity and specificity were satisfactory for all the outcomes. An average reduction of 50% of the lab tests was obtained. This is an improvement from previously reported similar studies with average performance 37% by [1-3].

CONCLUSIONS

Reducing frequent lab testing and the potential clinical and financial implications are an important issue in intensive care. In this work we present an artificial intelligence method to predict the benefit of proposed future laboratory tests. Using ICU data from 746 patients with gastrointestinal bleeding, and eleven measurements, we demonstrate high accuracy in predicting the likely information to be gained from proposed future lab testing for eight common GI related lab tests. Future work will explore applications of this approach to a range of underlying medical conditions and laboratory tests.

Vaccine availability in the United States during the 2009 H1N1 outbreak

OBJECTIVE

After initial flu cases are reported, months elapse before vaccine becomes available. The authors report the experience of US states during the fall of 2009 on H1N1 vaccine availability in relation to the occurrence of disease.

DESIGN

The authors used data from the Centers for Disease Control and prevention and state health departments to approximate second wave H1N1 epidemic curves. The authors compared these curves to two sources of vaccine distribution data-shipment and administration.

RESULTS

Ten states received their first shipments of vaccine after the epidemic peaked, four states during the week of the peak, and 10 states only 1 week prior to the peak. In nearly half of all states, the epidemic had already begun to decline before any individuals could have been protected.

CONCLUSIONS

A sensible approach would be to highlight the importance of diligent hygienic behavior and to reduce the rate of human-to-human contacts before vaccine is available.

Trends in the use and role of biomarkers in phase I oncology trials

PURPOSE

There has been interest in using biomarkers that aid the evaluation of new anticancer agents. We evaluated trends in the use of biomarkers and their contribution to the main goals of phase I trials.

EXPERIMENTAL DESIGN

We did a systematic review of abstracts submitted to the American Society of Clinical Oncology annual meeting from 1991 to 2002 and the publications related to these abstracts. We analyzed the use of biomarkers and their contribution to published phase I trials.

RESULTS

Twenty percent of American Society of Clinical Oncology phase I abstracts (503 of 2458) from 1991 to 2002 included biomarkers. This proportion increased over time (14% in 1991 compared with 26% in 2002; P < 0.02). Independent predictors of the use of biomarkers included National Cancer Institute sponsorship, submission in the time period of 1999 to 2002, adult population, and drug family (biological agents). Biomarkers supported dose selection for phase II studies in 11 of 87 of the trials (13%) emanating from these abstracts. However, the primary determinants of phase II dose and schedule were toxicity and/or efficacy in all but one of these 87 trials (1%). Biomarker studies provided evidence supporting the proposed mechanism of action in 34 of 87 of the published trials (39%).

CONCLUSIONS

The use of biomarkers in phase I trials has increased over the period from 1991 to 2002. To date, biomarker utilization has made a limited and primarily supportive contribution to dose selection, the primary end point of phase I studies. Additional studies are needed to determine what type of biomarker information is most valuable to evaluate in phase I trials.

A framework to evaluate the economic impact of pharmacogenomics

INTRODUCTION

Pharmacogenomics and personalized medicine promise to improve healthcare by increasing drug efficacy and minimizing side effects. There may also be substantial savings realized by eliminating costs associated with failed treatment. This paper describes a framework using health claims data for analyzing the potential value of pharmacogenomic testing in clinical practice.

METHODS

We evaluated a model of alternate clinical strategies using asthma patients' data from a retrospective health claims database to determine a potential cost offset. We estimated the likely cost impact of using a hypothetical pharmacogenomic test to determine a preferred initial therapy. We compared the annualized per patient costs distributions under two clinical strategies: testing all patients for a nonresponse genotype prior to treating and testing none.

RESULTS

In the Test All strategy, more patients fall into lower cost ranges of the distribution. In our base case (15% phenotype prevalence, 200 US dollars test, 74% overall first-line treatment efficacy and 60% second-line therapy efficacy) the cost savings per patient for a typical run of the testing strategy simulation ranged from 200 US dollars to 767 US dollars (5th and 95th percentile). Genetic variant prevalence, test cost and the cost of choosing the wrong treatment are key parameters in the economic viability of pharmacogenomics in clinical practice.

CONCLUSIONS

A general tool for predicting the impact of pharmacogenomic-based diagnostic tests on healthcare costs in asthma patients suggests that upfront testing costs are likely offset by avoided nonresponse costs. We suggest that similar analyses for decision making could be undertaken using claims data in which a population can be stratified by response to a drug.

Off-label prescribing among office-based physicians

BACKGROUND

Unlike medicines prescribed for Food and Drug Administration-approved indications, off-label uses may lack rigorous scientific scrutiny. Despite concerns about patient safety and costs to the health care system, little is known about the frequency of off-label drug use or the degree of scientific evidence supporting this practice.

METHODS

We used nationally representative data from the 2001 IMS Health National Disease and Therapeutic Index (NDTI) to define prescribing patterns by diagnosis for 160 commonly prescribed drugs. Each reported drug-diagnosis combination was identified as Food and Drug Administration-approved, off-label with strong scientific support, or off-label with limited or no scientific support. Outcome measures included (1) the proportion of uses that were off-label and (2) the proportion of off-label uses supported by strong scientific evidence. Multivariate analyses were used to identify drug-specific characteristics predictive of increased off-label use.

RESULTS

In 2001, there were an estimated 150 million (95% confidence interval, 127-173 million) off-label mentions (21% of overall use) among the sampled medications. Off-label use was most common among cardiac medications (46%, excluding antihyperlipidemic and antihypertensive agents) and anticonvulsants (46%), whereas gabapentin (83%) and amitriptyline hydrochloride (81%) had the greatest proportion of off-label use among specific medications. Most off-label drug mentions (73%; 95% confidence interval, 61%-84%) had little or no scientific support. Although several functional classes were associated with increased off-label use (P<.05), few other drug characteristics predicted off-label prescription.

CONCLUSIONS

Off-label medication use is common in outpatient care, and most occurs without scientific support. Efforts should be made to scrutinize underevaluated off-label prescribing that compromises patient safety or represents wasteful medication use.

The economic burden of anemia in cancer patients receiving chemotherapy

BACKGROUND

Anemia is one of the most common hematologic complications of cancer and cytotoxic treatment. The economic burden associated with anemia in patients with malignancy has not yet been extensively studied.

METHODS

Patients receiving chemotherapy within 6 months of initial cancer diagnosis were identified in a database of commercial health-care service claims and encounters. Patients with anemia were identified through a coded diagnosis of anemia, transfusion, or erythropoietin treatment. Exponential conditional mean models and a decomposition analysis were used to analyze mean 6-month health-care expenditures.

RESULTS

Twenty-six percent (26%) of 2760 cancer patients with recently diagnosed invasive cancer treated with chemotherapy had anemia. Mean (SD) 6-month unadjusted total expenditures were 62,499 dollars (78,016 dollars) for anemic patients and 36,871 dollars (52,308 dollars) for nonanemic patients (P < 0.0001), with inpatient services representing the largest cost differential between the groups. The adjusted mean 6-month expenditure for the average anemic patient receiving chemotherapy was 57,209 dollars. If anemic patients had the same average health status as nonanemic patients, their predicted 6-month expenditures would have been 19% lower (46,237 dollars). Alternatively, if anemic patients had the same expenditure structure or parameter estimates as nonanemic patients, their predicted expenditures would have been 51% lower (27,847 dollars). Thus, for any given health status, treating a patient who is anemic is associated with considerably higher expenditures.

CONCLUSIONS

Anemia among cancer patients receiving chemotherapy is associated with a substantial burden in terms of direct medical costs. Implications for the treatment of anemia are suggested by this research and should be confirmed in prospective studies.

Benefit plan design and prescription drug utilization among asthmatics: do patient copayments matter?

The ratio of controller-to-reliever medication use has been proposed as a measure of treatment quality for asthma patients. In this study we examine the effects of plan-level mean out-of-pocket asthma medication patient copayments and other features of benefit plan design on the use of controller medications alone, controller and reliever medications (combination therapy), and reliever medications alone. The 1995--2000 MarketScan claims data were used to construct plan-level out-of-pocket copayment and physician/practice prescriber preference variables for asthma medications. Separate multinomial logit models were estimated for patients in fee-for-service (FFS) and non-FFS plans relating benefit plan design features, physician/practice prescribing preferences, patient demographics, patient comorbidities, and county-level income variables to patient-level asthma treatment patterns. We find that the controller-to-reliever ratio rose steadily over 1995--2000, along with out-of-pocket payments for asthma medications, which rose more for controllers than for relievers. After controlling for other variables, however, plan-level mean out-of-pocket copayments were not found to have a statistically significant influence on patient-level asthma treatment patterns. On the other hand, physician/practice prescribing patterns strongly influenced patient-level treatment patterns. There is no strong statistical evidence that higher levels of out-of-pocket copayments for prescription drugs influence asthma treatment patterns. However, physician/practice prescribing preferences influence patient treatment.

Performance Decrements Resulting From Illness in the Workplace: The Effect of Headaches

OBJECTIVES

Studies on the impact of illness on work productivity are important to rationally allocate healthcare resources and to design programs to mitigate these effects. This investigation was conducted to develop and apply daily measures of illness episodes, and to collect subjective and objective data on work performance impacts. Medical bill reviewers completed daily responses to a questionnaire about headache manifestations, severity, and speed of work using interactive voice response (IVR). Of 134 eligible enrolled subjects, 117 (86%) provided at least 30 daily reports over 3 months. Their responses were matched to difficulty-adjusted objective measures: daily output, time on the system, and productivity. Respondents were clinically classified as migraineurs (n = 56), other headache disorders (n = 47), or having no headache disorder (n = 14). Each headache episode was classified as a migraine or nonmigraine headache based on reported manifestations.

RESULTS

The three groups were similar in a variety of demographic factors, and mean subject-specific measures of speed, output, and productivity. In a multivariate model using general estimating equations, only episode severity (not type of headache or person-specific diagnosis) was found to be associated with a significant decrement in speed or productivity. The self-reported decrement in speed (approximately 20%) was much greater than the actual measured effect on productivity (approximately 8%). Intensive daily diary collection by IVR on symptoms and work performance is feasible. However, analysis of detailed daily objective productivity data can be complex, with significant unmeasured sources of variance. Severity may be a more important determinant of headache effect on work performance than specific diagnosis. Future studies on illness episodes and work performance should measure informal accommodations that may enable employees to compensate for episodic illnesses.

Trends in the risks and benefits to patients with cancer participating in phase 1 clinical trials

CONTEXT

In the past, cancer patients entering phase 1 studies confronted the prospects of high risk and unlikely benefit. Over the last decade, cancer drugs under development have become more targeted, and the clinical research environment has become more scrutinized. The impact of these changes on the risks and benefits to patients who participate in phase 1 cancer trials is unknown.

OBJECTIVE

To determine trends in the rates of treatment-related (toxic) death, objective response, and serious toxicity and to identify factors associated with these outcomes.

DATA SOURCES

We searched abstracts and journal articles reporting the results of phase 1 cancer treatment trials originally submitted to annual meetings of the American Society of Clinical Oncology (ASCO) from 1991 through 2002.

STUDY SELECTION

We focused on published single-agent trials that enrolled patients with advanced solid tumors and excluded studies testing agents already approved by the US Food and Drug Administration at the time of the ASCO presentation.

DATA EXTRACTION

Multiple observers independently extracted information on trial design, location, sponsorship, types of tumors treated, drug class, route of administration, and clinical outcomes.

DATA SYNTHESIS

The overall toxic death rate for 213 studies (involving 6474 cancer patients) published in peer-reviewed journals was 0.54%, while the overall objective response rate was 3.8%. Toxic death rates decreased over the study period, from 1.1% over the first 4 years of the study (1991-1994) to 0.06% over the most recent 4-year period (1999-2002) (P<.01). Response rates also decreased but by proportionally much less. After adjusting for characteristics of the experimental trials and the investigational agents, the odds of a patient dying from an experimental treatment while participating in a trial submitted during the most recent 4-year period were less than one tenth those of a patient participating in a trial submitted during the first 4-year period (odds ratio, 0.09; 95% confidence interval, 0.01-0.67; P = .009). In comparison, the adjusted odds of a patient experiencing an objective response over the same time periods decreased by approximately half (odds ratio, 0.46; 95% confidence interval, 0.32-0.66; P<.001).

CONCLUSIONS

The level of risk experienced by cancer patients who participate in phase 1 treatment trials appears to have improved over the 12-year period from 1991 through 2002. Because toxic death rates have decreased more quickly than have objective response rates, the ratio of risk to benefit may have also improved. These changes relate in part to the targeted and less-toxic nature of newer cancer drugs and are coincident with the increased attention that has been paid to the safety of clinical research over the time period we analyzed.

The use of disease-modifying new drugs for multiple sclerosis treatment in private-sector health plans

OBJECTIVES

The aims of this study were to estimate the effects of demographics, location, severity of multiple sclerosis (MS), comorbidities, plan type, coinsurance levels, and time of entry into the sample on the use of disease-modifying agents.

METHODS

A retrospective analysis of medical claims data from 1996 through 2000 was conducted with a sample of MS patients covered by self-insured, employer-sponsored health plans. Proportional hazard analysis with the SAS procedure for proportional hazards regression was used to estimate the impact of the factors of interest on the use of disease-modifying agents. A simulation was conducted to assess the impact of changing drug copayments on the use of disease-modifying agents for MS.

RESULTS

The sample included 1807 patients. Patients were followed for as long as possible, but most were observed for <3 years; the mean (SD) follow-up time was 972.88 (440.59) days. Most factors associated with the use of disease-modifying agents were immutable. They included the following: high severity of illness (only marginally related; P = NS); history of seizures (P = 0.03), depression (P < 0.01), or heart disease (P = 0.01); census region of location (P < 0.01); union membership or association with a union member (P < 0.01); drug copayment requirements (P < 0.05); and year of entry into the sample (P < 0.01). In the simulation, a 50% reduction in drug copayments was associated with an increase of the proportion of patients treated with disease-modifying drugs from 41.2% to 54.7%. Patients' and physicians' preferences for treatment could not be measured directly. The true onset of MS may be unknown for many patients, but this would be the case even if medical records or other data were used for this study.

CONCLUSIONS

Our analyses showed an association between copayments and the use of disease-modifying drugs for MS. Insurance policies can be tailored to influence the use of disease-modifying drugs, enhancing the quality of care for MS patients and reducing price-related barriers to beneficial treatment. Future research should test whether reducing copayments for MS treatment would reduce the use of other health care services (via better MS treatment that modifies the course of illness), or whether the use of disease-modifying drugs would increase total costs to the plan, resulting in slightly higher premiums.

Costs of antidepressant medications associated with inadequate treatment

OBJECTIVE

To determine the costs of antidepressant medications used during inadequate treatment.

STUDY DESIGN

Retrospective database analysis of pharmacy claims made by patients who were treated under routine clinical conditions from July 1, 1999, through September 30, 2002.

PATIENTS AND METHODS

Our participants included 21,632 patients enrolled in a commercial HMO who had a primary care physician associated with our healthcare system. Patients never receiving at least a minimum likely effective antidepressant dose for at least 90 days were defined as having inadequate treatment. This study calculated the costs of antidepressants involved with inadequate treatment at the level of the patient and the medication trial.

RESULTS

A majority of patients (51%) received inadequate treatment. Of overall antidepressant costs, 16% were incurred during trials for patients never adequately treated. The majority of inadequate trials were short and unlikely to have been effective. Most patients (64%) had only a single trial of antidepressants. Venlafaxine, fluoxetine, and sertraline had significantly lower first-trial inadequacy rates compared with the most commonly prescribed agent, citalopram.

CONCLUSIONS

Improved patient care quality and lower antidepressant costs could result if clinicians and healthcare systems focus on reducing short trial rates. Initiating treatment with agents least likely to be discontinued prematurely may be helpful.

Prescriber intent, off-label usage, and early discontinuation of antidepressants: a retrospective physician survey and data analysis

BACKGROUND

Many patients discontinue antidepressant therapy long before the 6-month minimum duration recommended for the treatment of major depression and many other diagnoses. We explore various possibilities, including prescriber intent and patient diagnosis, to explain some of this early discontinuation.

METHOD

Patients from a single health maintenance organization who filled at least 1 prescription for an antidepressant during the first 4 months of 2001 and who did not fill an antidepressant prescription in the 6 months prior were identified retrospectively. Prescribers of those patients' antidepressants were surveyed for patient diagnosis and length of intended treatment with antidepressant medication. Actual length of treatment was then obtained from pharmacy data and correlated with survey data and other variables.

RESULTS

Prescriber surveys were returned for 51% (485/951) of the patients identified. Surveys indicated that for 34% of initial antidepressant prescriptions, < 6 months of treatment was intended. Important determinants of the length of antidepressant therapy included prescriber specialty area, number of prescribers, prescriber intent, diagnosis, specific antidepressant used, and concomitant benzodiazepine use.

CONCLUSIONS

Prescriber intention to treat many patients with short courses of antidepressants, often for off-label, non-mental health indications, was correlated with early discontinuation and needs further study of both its rationale and efficacy. Although less prevalent, short-term treatment of mental health disorders, including depression, was also intended by psychiatrists and other prescribers. The widespread practice of intended short-term treatment with antidepressants needs to be understood better, since it results in guideline-incompatible, early antidepressant discontinuation.