State-level policies and receipt of CDC-informed opioid thresholds among commercially insured new chronic opioid users.

OBJECTIVES: To evaluate the association of state-level policies on receipt of opioid regimens informed by Centers for Disease Control and Prevention (CDC) morphine milligram equivalent (MME)/day recommendations. DESIGN: A retrospective cohort study of new chronic opioid users (NCOUs). SETTING: Commercially insured plans across the United States using IQVIA PharMetrics® Plus for Academics database with new chronic use between January 2014 and March 2015. PARTICIPANTS: NCOUs with ≥60-day coverage of opioids within a 90-day period with ≥30-day opioid-free period prior to the date of the first qualifying opioid prescription. INTERVENTIONS: State-level policies including Prescription Drug Monitoring Program (PDMP) robustness and cannabis policies involving the presence of medical dispensaries and state-wide decriminalization. MAIN OUTCOME MEASURES: NCOUs were placed in three-tiered risk-based average MME/day thresholds: low (>0 to <50), medium (≥50 to <90), and high (≥90). Multinomial logistic regression was used to estimate the association of state-level policies with the thresholds while adjusting for relevant patient-specific factors. RESULTS: NCOUs in states with medium or high PDMP robustness had lower odds of receiving medium (adjusted odds ratio [AOR] 0.74; 95 percent confidence interval [CI]: 0.62-0.69) and high (AOR 0.74; 95 percent CI: 0.59-0.92) thresholds. With respect to cannabis policies, NCOUs in states with medical cannabis dispensaries had lower odds of receiving high (AOR 0.75; 95 percent CI: 0.60-0.93) thresholds, while cannabis decriminalization had higher odds of receiving high (AOR 1.24; 95 percent CI: 1.04-1.49) thresholds. CONCLUSION: States with highly robust PDMPs and medical cannabis dispensaries had lower odds of receiving higher opioid thresholds, while cannabis decriminalization correlated with higher odds of receiving high opioid thresholds.

Short-term healthcare resource utilization associated with receipt of CDC-informed opioid thresholds among commercially insured new chronic opioid users.

OBJECTIVE: To evaluate the impact of recent changes to the Centers for Disease Control and Prevention (CDC) morphine milligram equivalent (MME)/day threshold recommendations on healthcare utilization. DESIGN: A retrospective cohort study of new chronic opioid users (NCOUs). SETTING: Commercially insured plans across the United States using IQVIA PharMetrics® Plus for Academics database with new use between January 2014 and March 2015. PATIENTS: NCOUs with ≥60-day coverage of opioids within a 90-day period with ≥30-day opioid-free period prior to the date of the first qualifying opioid -prescription. INTERVENTIONS: NCOU categorized by the CDC three-tiered risk-based average MME/day thresholds: low (>0 to <50), medium (≥50 to <90), and high (≥90). MAIN OUTCOME MEASURES: Multivariable logistic regression was used to calculate adjusted odds of incurring an acute care encounter (ACE) (all-cause and opioid-related) between the thresholds (adjusted odds, 95 percent confidence interval). RESULTS: In adjusted analyses, when compared to low threshold, there was no difference in the odds of all-cause ACE across the medium (1.01, 0.94-1.28) and high (1.01, 0.84-1.22) thresholds. When compared to low threshold, a statistically insignificant increase was observed when evaluating opioid-related ACE among medium (1.86, 0.86-4.02) and high (1.51, 0.65-3.52) thresholds. CONCLUSIONS: There was no difference in odds of an all-cause or opioid-related ACE associated with the thresholds. Early-intervention programs and policies exploring reduction of MME/day among NCOUs may not result in short-term reduction in all-cause or opioid-related ACEs. Further assessment of potential long-term reduction in ACEs among this cohort may be insightful.

The DTC microbiome testing industry needs more regulation.

Tests lack analytical and clinical validity, requiring more federal oversight to prevent consumer harm.

Methodological Comparison of Mapping the Expanded Prostate Cancer Index Composite to EuroQoL-5D-3L Using Cross-Sectional and Longitudinal Data: Secondary Analysis of NRG/RTOG 0415.

PURPOSE: To compare the predictive ability of mapping algorithms derived using cross-sectional and longitudinal data. METHODS: This methodological assessment used data from a randomized controlled noninferiority trial of patients with low-risk prostate cancer, conducted by NRG Oncology (ClinicalTrials.gov identifier: NCT00331773), which examined the efficacy of conventional schedule versus hypofractionated radiation therapy (three-dimensional conformal external beam radiation therapy/IMRT). Health-related quality-of-life data were collected using the Expanded Prostate Cancer Index Composite (EPIC), and health utilities were obtained using EuroQOL-5D-3L (EQ-5D) at baseline and 6, 12, 24, and 60 months postintervention. Mapping algorithms were estimated using ordinary least squares regression models through five-fold cross-validation in baseline cross-sectional data and combined longitudinal data from all assessment periods; random effects specifications were also estimated in longitudinal data. Predictive performance was compared using root mean square error. Longitudinal predictive ability of models obtained using baseline data was examined using mean absolute differences in the reported and predicted utilities. RESULTS: A total of 267 (and 199) patients in the estimation sample had complete EQ-5D and EPIC domain (and subdomain) data at baseline and at all subsequent assessments. Ordinary least squares models using combined data showed better predictive ability (lowest root mean square error) in the validation phase for algorithms with EPIC domain/subdomain data alone, whereas models using baseline data outperformed other specifications in the validation phase when patient covariates were also modeled. The mean absolute differences were lower for models using EPIC subdomain data compared with EPIC domain data and generally decreased as the time of assessment increased. CONCLUSION: Overall, mapping algorithms obtained using baseline cross-sectional data showed the best predictive performance. Furthermore, these models demonstrated satisfactory longitudinal predictive ability.

Mapping expanded prostate cancer index composite to EQ5D utilities to inform economic evaluations in prostate cancer: Secondary analysis of NRG/RTOG 0415.

PURPOSE: The Expanded Prostate Cancer Index Composite (EPIC) is the most commonly used patient reported outcome (PRO) tool in prostate cancer (PC) clinical trials, but health utilities associated with the different health states assessed with this tool are unknown, limiting our ability to perform cost-utility analyses. This study aimed to map EPIC tool to EuroQoL-5D-3L (EQ5D) to generate EQ5D health utilities. METHODS AND MATERIALS: This is a secondary analysis of a prospective, randomized non-inferiority clinical trial, conducted between 04/2006 and 12/2009 at cancer centers across the United States, Canada, and Switzerland. Eligible patients included men >18 years with a known diagnosis of low-risk PC. Patient HRQoL data were collected using EPIC and health utilities were obtained using EQ5D. Data were divided into an estimation sample (n = 765, 70%) and a validation sample (n = 327, 30%). The mapping algorithms that capture the relationship between the instruments were estimated using ordinary least squares (OLS), Tobit, and two-part models. Five-fold cross-validation (in-sample) was used to compare the predictive performance of the estimated models. Final models were selected based on root mean square error (RMSE). RESULTS: A total of 565 patients in the estimation sample had complete information on both EPIC and EQ5D questionnaires at baseline. Mean observed EQ5D utility was 0.90±0.13 (range: 0.28-1) with 55% of patients in full health. OLS models outperformed their counterpart Tobit and two-part models for all pre-determined model specifications. The best model fit was: "EQ5D utility = 0.248541 + 0.000748*(Urinary Function) + 0.001134*(Urinary Bother) + 0.000968*(Hormonal Function) + 0.004404*(Hormonal Bother)- 0.376487*(Zubrod) + 0.003562*(Urinary Function*Zubrod)"; RMSE was 0.10462. CONCLUSIONS: This is the first study to identify a comprehensive set of mapping algorithms to generate EQ5D utilities from EPIC domain/ sub-domain scores. The study results will help estimate quality-adjusted life-years in PC economic evaluations.

Prescription drug monitoring programs: Assessing the association between "best practices" and opioid use in Medicare.

OBJECTIVE: To estimate the impact of implementing prescription drug monitoring program (PDMP) best practices on prescription opioid use. DATA SOURCES: 2007-2012 Medicare claims for noncancer pain patients, and PDMP attributes from the Prescription Drug Abuse Policy System. STUDY DESIGN: We derived PDMP composite scores using the number of best practices adopted by states (range: 0-14), classifying states as either no PDMP, low strength (0 < score < median), or high strength (score ≥ median). Using generalized linear models, we quantified the association between the PDMP score category and opioid use measures-overall and stratified by disability/age. Sensitivity analyses assessed the general Medicare sample regardless of pain diagnoses, individual PDMP characteristics, and compared GEE model findings to models with state fixed effects. PRINCIPAL FINDINGS: Compared to non-PDMP states, strong PDMP states had lower opioid cumulative doses (-296 mg; 95% CI: -512, -132), days supplied (-7.84; 95% CI: -10.6, -5.04), prescription fill rates (0.97; 95% CI: 0.95, 0.98), and mean daily doses (-2.31 mg; 95% CI: -3.14, -1.48) but greater prevalence of high opioid doses in disabled adults, whereas there was little or no change in older adults. Findings in states with weak PDMPs were substantively similar to those of strong PDMPs. Results from sensitivity analyses were mostly consistent with main findings except there was a null relationship with mean daily doses and high doses in models with state fixed effects. CONCLUSIONS: Comprehensive or minimal adoption of PDMP best practices was associated with mostly comparable effects on Medicare beneficiaries' opioid use; however, these effects were concentrated among nonelderly disabled adults.

The Impact of Regulatory Policies on the Future of Fecal Microbiota Transplantation.

In this article, the authors explore the impact of a potential future regulatory decision by FDA whether or not to continue its enforcement discretion policy allowing physicians to perform, and stool banks to sell, stool product for fecal microbiota transplantation as a treatment for recurrent Clostridium Difficile infection without an Investigative New Drug (IND) application. The paper looks at the Agency's regulatory options in light of the current gut microbiota based products that are in the FDA pipeline for drug approval and the potential impact and repercussions of their approval on FDA action. In laying out FDA's options we consider the implications of market exclusivity and off-label use of newly approved drugs. Ultimately, we explore the potential impact of FDA's decision on patients, research, and innovation.

The authors reply.

This is a reply to the commentary of Ossorio and Zhou.

Assessing Physician and Patient Perceptions of Generic Drugs via Facebook: A Feasibility Study.

Background: Social media offer a novel avenue to engage with and recruit research participants. Facebook in particular is a promising option given its popularity and widespread use. Objective: To explore the feasibility of using Facebook to recruit physicians and patients to participate in a survey to assess their perceptions about generic venlafaxine extended release (ER) tablet indicated for depression. Methods: Web-based surveys were developed to gauge physicians' prescribing experiences with and patients' perceptions of generic venlafaxine ER tablet. The surveys included questions specific to venlafaxine ER tablets, such as perceived safety and efficacy of the drug and overall comfort level with either prescribing or taking the drug. Survey links were then posted and advertised on Facebook to recruit physicians and patients. Results: Advertisement for physicians reached 1898 Facebook users and advertisement for patients reached 1144 users during a 10-day advertising period. However, only 14 and 35 users clicked on the survey for physicians and patients, respectively. No physician completed the physician survey while 3 patients completed the patient survey. Conclusions: The findings of this study suggest that Facebook may not be an effective method to recruit physicians. Facebook holds promise to recruit patients, but additional recruitment efforts, such as incentives, are needed.

ISPOR Code of Ethics 2017 (4th Edition).

As the leading health economics and outcomes research (HEOR) professional society, ISPOR has a responsibility to establish a uniform, harmonized international code for ethical conduct. ISPOR has updated its 2008 Code of Ethics to reflect the current research environment. This code addresses what is acceptable and unacceptable in research, from inception to the dissemination of its results. There are nine chapters: 1 - Introduction; 2 - Ethical Principles respect, beneficence and justice with reference to a non-exhaustive compilation of international, regional, and country-specific guidelines and standards; 3 - Scope HEOR definitions and how HEOR and the Code relate to other research fields; 4 - Research Design Considerations primary and secondary data related issues, e.g., participant recruitment, population and research setting, sample size/site selection, incentive/honorarium, administration databases, registration of retrospective observational studies and modeling studies; 5 - Data Considerations privacy and data protection, combining, verification and transparency of research data, scientific misconduct, etc.; 6 - Sponsorship and Relationships with Others (roles of researchers, sponsors, key opinion leaders and advisory board members, research participants and institutional review boards (IRBs) / independent ethics committees (IECs) approval and responsibilities); 7 - Patient Centricity and Patient Engagement new addition, with explanation and guidance; 8 - Publication and Dissemination; and 9 - Conclusion and Limitations.

Impact of prescription drug monitoring programs (PDMPs) on opioid utilization among Medicare beneficiaries in 10 US States.

BACKGROUND AND AIMS: Prescription Drug Monitoring Programs (PDMPs) are a principal strategy used in the United States to address prescription drug abuse. We (1) compared opioid use pre- and post-PDMP implementation and (2) estimated differences of PDMP impact by reason for Medicare eligibility and plan type. DESIGN: Analysis of opioid prescription claims in US states that implemented PDMPs relative to non-PDMP states during 2007-12. SETTING: Florida, Louisiana, Nebraska, New Jersey, Vermont, Georgia, Wisconsin, Maryland, New Hampshire and Arkansas, USA. PARTICIPANTS: A total of 310 105 disabled and older adult Medicare enrolees. MEASUREMENTS: Primary outcomes were monthly total opioid volume, mean daily morphine milligram equivalent (MME) dose per prescription and number of opioid prescriptions dispensed. The key predictors were PDMP status and time. Tests for moderation examined PDMP impact by Medicare eligibility (disability versus age) and drug plan [privately provided Medicare Advantage (MAPD) versus fee-for-service (PDP)]. FINDINGS: Overall, PDMP implementation was associated with reduced opioid volume [-2.36 kg/month, 95% confidence interval (CI) = -3.44, -1.28] and no changes in mean MMEs or opioid prescriptions 12 months after implementation compared with non-PDMP states. We found evidence of strong moderation effects. In PDMP states, estimated monthly opioid volumes decreased 1.67 kg (95% CI = -2.38, -0.96) and 0.75 kg (95% CI = -1.32, -0.18) among disabled and older adults, respectively, and 1.2 kg, regardless of plan type. MME reductions were 3.73 mg/prescription (95% CI = -6.22, -1.24) in disabled and 3.02 mg/prescription (95% CI = -3.86, -2.18) in MAPD beneficiaries, but there were no changes in older adults and PDP beneficiaries. Dispensed prescriptions increased 259/month (95% CI = 39, 479) among the disabled and decreased 610/month (95% CI = -953, -257) among MAPD beneficiaries. CONCLUSIONS: Prescription drug monitoring programs (PDMPs) are associated with reductions in opioid use, measured by volume, among disabled and older adult Medicare beneficiaries in the United States compared with states that do not have PDMPs. PDMP impact on daily doses and daily prescriptions varied by reason for eligibility and plan type. These findings cannot be generalized beyond the 10 US states studied.

A proposed definition of microbiota transplantation for regulatory purposes.

The advent of fecal microbiota transplantation (FMT) and the prospect of other types of microbiota transplants (MT), e.g. vaginal, skin, oral and nasal, are challenging regulatory agencies. Although FDA is regulating FMT (as a biologic), there is currently no widely accepted or agreed upon scientific or legal definition of FMT or MT. The authors report on discussions regarding a definition of MT that took place among a working group of stakeholders convened under a National Institutes for Allergies and Infectious Diseases grant to address the regulation of MT. In arriving at a definition, the group considered the 1) nature of the material being transplanted; 2) degree of manipulation of the transferred materials prior to implantation; 3) ability to characterize the transplanted product using external techniques; and 4) origin of the stool product (single vs multiple donors).

Navigating Through a Complex and Inconsistent Regulatory Framework: Section 503B of the Federal Food Drug and Cosmetic Act Outsourcing Facilities Engaged in Clinical Investigation.

The integrity of unbiased clinical data is essential to the future of the health care system by facilitating the discovery of lifesaving medicines and ensuring investigational drugs are safe and effective. Since 2002, the US pharmaceutical industry has invested over $500 billion, which is the largest research and development investment of any sector of the US economy. As a consequence of this significant investment, pharmacy compounders and other stakeholders must be acutely aware of the consequences of noncompliance. Pharmacy compounders are required to navigate through a complex and ever-changing regulatory landscape governed by US federal and state authorities competing for oversight and enforcement authority. In particular, pharmacy compounders participating in clinical investigation are faced with inconsistent federal and state drug labeling regulations, which can lead to enforcement for violating acceptable standards for clinical investigation and informed consent. As FDA registrants, Outsourcing Facilities are able to produce large volumes of clinical supplies without the need for prescriptions for individually named patients. Some states, however, may have prescription drug labeling laws that could thwart the ability to provide these clinical supplies. Accordingly, researchers should be aware of the implications of federal and state laws, including any inconsistencies, prior to engaging in clinical investigation.

Perceptions of patient provider agreements.

OBJECTIVES: Use of patient provider agreements (PPAs) is increasing, yet there is limited evidence on the effectiveness of PPAs to prevent prescription opioid misuse and diversion, and few guidelines for providers. We conducted eight focus groups to understand patient and prescriber perceptions of PPAs. METHODS: We recruited 40 patients who had been asked to sign a PPA and 40 prescribers who had administered at least one PPA. We developed topic guides for the two groups based on prior literature. Focus groups were audio-recorded and transcribed verbatim. Two investigators independently performed the content analysis of the transcripts and reached consensus on recurring themes. KEY FINDINGS: PPA use varied according to physician specialty. General practitioners used PPAs the least but reported increasing pressure from liability insurers to use them. Many patients reported signing a PPA in the emergency room of a hospital. Prescribers and patients reported a lack of understanding among patients concerning the purpose and content of the PPA. Prescribers questioned the legal status of the PPA, while patients believed that the PPA was a legal document intended to protect prescribers. Patients and prescribers valued PPA content items differently, although both groups agreed that signing a PPA would not prevent opioid misuse. CONCLUSIONS: We identified several themes concerning the administration, content, effectiveness and utility of PPAs that highlight areas of research to improve PPAs. We also describe trends requiring further investigation. Understanding content of importance to patients will facilitate the development of a patient-centred PPA.

Relationships between antiparkinson medication nonadherence, regimen modifications, and healthcare utilization and expenditures.

OBJECTIVES: To examine 1) the effect of prior antiparkinson drug (APD) nonadherence on subsequent APD regimen modifications; and 2) the influence of modifications on healthcare utilization and costs by patients with Parkinson's disease (PD). METHODS: This retrospective cohort study included 7052 PD patients with ≥2 APD prescriptions who initiated a modification of APD regimens in 2007. Modification was assessed as changing from one APD to another and/or adding a new APD to an existing regimen. Nonadherence was measured using Medication Possession Ratio <0.8. Discrete-time survival analyses were used to estimate the effect of prior nonadherent behavior on initiating APD modifications. Generalized linear models were used to estimate the effect of initiating medication modifications on subsequent 3-month medical use and costs. RESULTS: Initiation of APD modifications in any given month was higher among patients who were nonadherent to APDs in the preceding month (adjusted hazard ratio [HR] = 1.23), compared to their adherent counterparts. Modifications significantly predicted higher risk of all-cause and PD-related hospitalizations (adjusted relative risk [RR] = 1.22 and 1.83, respectively), home health agency utilization (RR = 1.18 and 1.52), and use of physician services (RR = 1.14 and 1.41), as well as higher total all-cause healthcare expenditures (mean = $1064) in any given 3-month interval. CONCLUSIONS: Prior nonadherence to APDs might influence initiation of APD modification. APD modifications were associated with increased health care utilization and expenditures, with the caveats that indications of modifications and disease severity may still play roles. Prescribers should consider patients' medication adherence when changing APD regimens to lower the costs of medical services.

Probiotics: achieving a better regulatory fit.

The development and marketing of new probiotic products, substances containing live microorganisms that have a beneficial effect on the human body, have dramatically increased over the last few years. This article examines how the Food and Drug Administration and Federal Trade Commission currently regulate probiotics and makes recommendations as to changes that might be made to ensure that probiotic products are made available to the general public in a way that is both safe and effective.

Antiparkinson drug adherence and its association with health care utilization and economic outcomes in a Medicare Part D population.

OBJECTIVES: We examine the associations of adherence to antiparkinson drugs (APDs) with health care utilization and economic outcomes among patients with Parkinson's disease (PD). METHODS: By using 2006-2007 Medicare administrative data, we examined 7583 beneficiaries with PD who filled two or more APD prescriptions during 19 months (June 1, 2006, to December 31, 2007) in the Part D program. Two adherence measures--duration of therapy (DOT) and medication possession ratio (MPR)--were assessed. Negative binomial and gamma generalized linear models were used to estimate the rate ratios (RRs) of all-cause health care utilization and expenditures, respectively, conditional upon adherence, adjusting for survival risk, sample selection, and health-seeking behavior. RESULTS: Approximately one-fourth of patients with PD had low adherence (MPR < 0.80, 28.7%) or had a short DOT (≤ 400 days, 23.9%). Increasing adherence to APD therapy was associated with decreased health care utilization and expenditures. For example, compared with patients with low adherence, those with high adherence (MPR = 0.90-1.00) had significantly lower rates of hospitalization (RR = 0.86), emergency room visits (RR = 0.91), skilled nursing facility episodes (RR = 0.67), home health agency episodes (RR = 0.83), physician visits (RR = 0.93), as well as lower total health care expenditures (-$2242), measured over 19 months. Similarly, lower total expenditure (-$6308) was observed in patients with a long DOT versus those with a short DOT. CONCLUSIONS: In this nationally representative sample, higher adherence to APDs and longer duration of use of APDs were associated with lower all-cause health care utilization and total health care expenditures. Our findings suggest the need for improving medication-taking behaviors among patients with PD to reduce the use of and expenditures for medical resources.

Antiparkinson drug use and adherence in medicare part D beneficiaries with Parkinson's disease.

BACKGROUND: Antiparkinson drugs (APDs) are the mainstay of managing Parkinson's disease (PD). However, there is paucity of evidence documenting patterns of APD use and examining factors associated with adherence to APDs. OBJECTIVES: Our goal was to provide updated, comprehensive population-based data on APD use and adherence and to examine characteristics associated with adherence behaviors. METHODS: We analyzed data from 7583 beneficiaries with PD who had ≥ 2 APD prescription fills and were continuously enrolled in Medicare Parts A, B, and D for up to 19 months (from June 1, 2006, through December 31, 2007) or until death in 2007. We assessed 5 patterns of APD use: (1) concurrent use of ≥ 2 APD classes for ≥ 30 days; (2) switching of APDs from 1 to another; (3) augmentation of the existing regimen with a new APD; (4) duration of therapy, defined as days of APD treatment; and (5) adherence measured by using the medication possession ratio (MPR). We corrected for sample selection bias inherent in patients' self-selection into either a Part D plan or a Medicare Advantage Prescription Plan by using Heckman's 2-stage procedures. RESULTS: APD users were pre-dominantly aged ≥ 65 years (93.6%), female (59.9%), and white (89.3%). Almost one half (43.2%) of APD users concurrently used ≥ 2 APD classes. One in 4 APD users experienced changes in their APD regimen, with 16.4% switching medications and 21.1% augmenting their current regimen. Three quarters of APD users had therapy lasting ≥ 436 days (75.3%) and an MPR ≥ 0.8 (72.7%). Multivariate analyses revealed that patients aged ≥ 65 years, of non-white race, non-low-income subsidy recipients, late Part D enrollees, cognitively impaired, highly comorbid, and who experienced multiple changes in APD therapy were less likely to adhere to APD therapy. We were able to generalize our findings to all Part D enrollees by correcting for sample selection bias using the Heckman approach. These population-level, generalizable observations provide better understanding of APD use and adherence and assist in the design of interventions for poor adherence. Limitations include cross-sectional study design and constraints in administrative data that preclude measurement of other potential factors related to adherence. CONCLUSIONS: A substantial proportion of these Medicare beneficiaries with PD used multiple APDs concurrently, experienced switching and/or augmentation of APDs, and had poor adherence to APDs. Patient characteristics and clinical and drug-related factors were important predictors of APD adherence.

Racial/ethnic and age disparities in chemotherapy selection for colorectal cancer.

OBJECTIVES: To test the hypotheses that African American patients and older patients with stage IV colorectal cancer were less likely to receive newer chemotherapy agents. STUDY DESIGN: Retrospective cohort design. METHODS: Among 5068 Surveillance, Epidemiology, and End Results-Medicare patients diagnosed as having stage IV colorectal cancer between 2000 and 2002, a total of 2466 received chemotherapy and were included in the analysis. Irinotecan hydrochloride was the first of the "newer" chemotherapy agents and was marketed in 2000 as a first-line add-on agent. Descriptive statistics were generated, and a multivariable logistic regression was run to estimate the odds of receiving irinotecan among African American patients and older patients and within 2 months of chemotherapy initiation. RESULTS: African American patients had lower odds of initiating treatment with a newer chemotherapy than white patients (adjusted odds ratio, 0.641; 95% confidence interval, 0.453-0.907). An age disparity was also found, with all older age groups being significantly less likely to initiate treatment with a newer chemotherapy than the youngest age group: the adjusted odds of receiving newer chemotherapy agents (relative to patients aged 66-70 years) were lower and significant among patients aged 71 to 75, 76 to 80, and older than 80 years (odds ratios, 0.708, 0.527, and 0.213, respectively). CONCLUSIONS: Disparities in chemotherapy selection exist among patients receiving chemotherapy for stage IV colorectal cancer. On initiating chemotherapy, African American patients and older patients were less likely to receive a newer agent.