Outcomes for pembrolizumab stratified by pemetrexed maintenance post pembrolizumab-platinum-pemetrexed induction in metastatic non-small-cell lung cancer.

Aim: We assessed treatment patterns and outcomes in patients with metastatic nonsquamous non-small-cell lung cancer (mNSCLC) who initiated first-line pembrolizumab-platinum-pemetrexed (induction) in US community oncology settings. Methods: Patients initiating induction were retrospectively identified. Patients continuing pembrolizumab afterward underwent chart review. Clinical outcomes were described by maintenance pemetrexed exposure after inverse probability of treatment weighting (IPTW). Results: Median induction pembrolizumab and pemetrexed durations were 5.1 and 4.2 months. Among patients continuing pembrolizumab after induction, 64% received maintenance pemetrexed. Common discontinuation reasons for induction pemetrexed were completion of planned therapy (79%) and partial response (68%) and progressive disease (38%) and toxicity (29%) for maintenance pemetrexed. After IPTW, median overall survival and real-world progression-free survival were longer in patients continuing pembrolizumab with versus without maintenance pemetrexed (20.3 vs 12.0 months and 10.3 vs 5.8 months, respectively). Conclusion: Patient characteristics and planned treatment decisions affect maintenance pemetrexed utilization in the community oncology setting.

What is this summary about? Pembrolizumab is a drug that helps the lung cancer patient's immune system fight the cancer, even after the cancer has spread, or metastasized. After the patient gets better, the patient is treated with chemotherapy so the cancer will not come back. This is called 'maintenance treatment'. In KEYNOTE-189, a clinical trial, patients lived longer if they had pembrolizumab added to pemetrexed and platinum, which are chemotherapy drugs. If patients had maintenance treatment with pembrolizumab and pemetrexed, they also lived longer. However, do patients in community practices get those treatments? What were the results? We found that at cancer practices in the community instead of clinical trials, not all patients received pemetrexed in maintenance treatment. Many had finished their planned therapy and their tumors had shrunk. Also, some physicians chose not to give their patients pemetrexed. In addition, some women and some older and sicker patients did not get pemetrexed. Some patients had pemetrexed in maintenance but stopped because their cancer grew worse or because they had side effects. Those patients did not live as long as patients who did have maintenance pemetrexed. What do the results mean? Patients with metastatic non-small-cell lung cancer in the community practice do better on the treatments tested in clinical trials. However, certain patients do not get those treatments. The reasons need to be understood, to make sure that those patients get better treatments.

Patient-Reported Health-Related Quality of Life in KEYNOTE-604: Pembrolizumab or Placebo Added to Etoposide and Platinum as First-Line Therapy for Extensive-Stage SCLC.

Introduction: In the phase 3 KEYNOTE-604 study (NCT03066778), pembrolizumab plus etoposide and platinum chemotherapy (EP) significantly (p = 0.0023) improved progression-free survival versus placebo plus EP in previously untreated extensive-stage SCLC (ES-SCLC). We present health-related quality of life (HRQoL) results from KEYNOTE-604. Methods: Patients with stage IV SCLC were randomized 1:1 to pembrolizumab 200 mg or placebo every 3 weeks for 35 cycles plus four cycles of EP. Secondary end points included mean change from baseline to week 18 in the European Organisation for Research and Treatment of Cancer Quality-of-Life Questionnaire-Core 30 (QLQ-C30) global health status/quality of life (GHS/QoL) scale and time to deterioration in the composite outcome of cough, chest pain, or dyspnea from QLQ-C30 and QLQ-Lung Cancer Module 13. Two-sided, nominal p values are reported. Results: A total of 439 patients completed at least one QLQ-C30 and QLQ-Lung Cancer Module 13 assessment (pembrolizumab + EP, n = 221; placebo + EP, n = 218). GHS/QoL scores improved from baseline to week 18: least squares mean (95% confidence interval [CI]) changes were 8.7 (5.3-12.1) for pembrolizumab plus EP and 4.2 (0.9-7.5) for placebo plus EP. Between-group differences in least squares mean scores were improved for pembrolizumab plus EP (4.4 [95% CI: 0.2-8.7], p = 0.040]). Median time to deterioration for the composite end point was not reached and 8.7 (95% CI: 5.9-not reached) months, respectively (hazard ratio = 0.80 [95% CI: 0.56-1.14], p = 0.208). Conclusions: First-line pembrolizumab plus EP therapy maintained HRQoL in patients with ES-SCLC and may be associated with greater improvement than placebo plus EP. Together with the efficacy and safety findings in KEYNOTE-604, HRQoL data support the benefit of pembrolizumab in ES-SCLC.

Health Care Resource Use Among Patients with Advanced Non-Small Cell Lung Cancer in Japan, 2017-2019.

Background: First-line immune checkpoint inhibitor (ICI) monotherapy for advanced non-small cell lung cancer (NSCLC) was introduced in Japan in February 2017. Limited information is available since that time regarding health care resource use for NSCLC in Japan, where the hospitalization burden is high. Objective: We evaluated health care resource use from first- through third-line systemic anticancer therapy for patients with advanced NSCLC included in a multicenter, retrospective chart review study. Methods: Eligible patients were aged 20 years or older with unresectable locally advanced/metastatic NSCLC with no known actionable genomic alteration who initiated first-line systemic anticancer therapy from July 1, 2017, to December 20, 2018, at 23 Japanese hospitals. We calculated the percentage of patients with a record of each resource used, the total number of each resource, and the resource use per 100 patient-weeks of follow-up from initiation of first-, second-, and third-line therapy, overall and by the 3 most common regimen categories, namely, ICI monotherapy, platinum-doublet chemotherapy (without concomitant ICI), and nonplatinum cytotoxic regimens (nonplatinum). Study follow-up ended September 30, 2019. Results: Among 1208 patients (median age = 70 years; 975 [81%] men), 463 patients (38%) received ICI monotherapy, 647 (54%) received platinum-doublet chemotherapy, and 98 (8%) received nonplatinum regimens as first-line therapy. During the study, 621 (51%) patients initiated second-line, and 281 (23%) initiated third-line therapy. The majority of patients experienced ≥1 hospitalization (76%-94%) and ≥1 outpatient visit (85%-90%) during each therapy line. The number of hospitalizations increased from 6.5 per 100 patient-weeks in first-line to 8.0 per 100 patient-weeks in third-line. During first-line therapy, the number of hospitalizations per 100 patient-weeks were 4.8, 8.4, and 6.5 for patients receiving ICI monotherapy, platinum-doublet chemotherapy, and nonplatinum regimens, respectively, and the percentages of hospitalizations categorized as attributable to NSCLC treatment administration (no surgery, procedure, treatment of metastasis, or palliative lung radiation) were 64%, 77%, and 73%, respectively. The number of outpatient visits increased from 43.0 per 100 patient-weeks in first-line to 51.4 per 100 patient-weeks in third-line therapy. During first-line therapy, outpatient visits per 100 patient-weeks were 41.0, 46.7, and 33.0 for patients receiving ICI monotherapy, platinum-doublet chemotherapy, and nonplatinum regimens, respectively, and the percentages of outpatient visits for infusion therapy were 48%, 34%, and 36%, respectively. Conclusions: The results of this study, although solely descriptive, showed differing patterns of health care resource use during first-line therapy among the 3 common systemic anticancer therapy regimens for advanced NSCLC in Japan and suggest that further research is needed to investigate these apparent differences by treatment regimen.

First-Line Pembrolizumab Monotherapy for Advanced NSCLC With Programmed Death-Ligand 1 Expression Greater Than or Equal to 50%: Real-World Study Including Older Patients in Japan.

Introduction: Pembrolizumab became available in Japan in February 2017 for first-line monotherapy of unresectable advanced and metastatic NSCLC with programmed death-ligand 1 (PD-L1) tumor proportion score (TPS) greater than or equal to 50%. This retrospective chart review study aimed to describe real-world clinical outcomes of first-line pembrolizumab monotherapy, including for patients 75 years or older, who are under-represented in clinical trials. Methods: We identified patients (≥20 y old) at 23 sites initiating first-line pembrolizumab monotherapy from July 1, 2017, to December 20, 2018, for stages IIIB, IIIC, and IV NSCLC with PD-L1 TPS greater than or equal to 50% and Eastern Cooperative Oncology Group performance status of 0 to 2 or unknown. Patients with actionable genomic alterations (EGFR, ALK, ROS1, BRAF) and clinical trial participants were excluded. Time-to-event outcomes were estimated using Kaplan-Meier, with data cutoff on September 30, 2019. Results: Of 441 eligible patients (78% men), 303 (69%) were younger than 75 years and 138 (31%) were 75 years or older; median age was 70 years. With median follow-up of 13.5 months, median overall survival (OS) was not reached (NR); 12- and 24-month OS rates were 72% and 58%, respectively. For ages younger than 75 and 75 years or older, median OS was NR and 23.5 months (95% confidence interval: 16.2-NR), respectively; 12-month OS rates were 74% and 67% and 24-month OS rates were 62% and 48%, respectively. Median real-world progression-free survival was similar in the two age groups (10.1 and 9.5 mo, respectively), as was median real-world time on treatment with pembrolizumab (5.7 and 5.6 mo). Conclusions: These findings complement clinical trial results, adding real-world evidence supporting benefits of first-line pembrolizumab monotherapy for advanced NSCLC with PD-L1 TPS greater than or equal to 50%, including for patients 75 years or older.

Real-World Treatments and Clinical Outcomes in Advanced NSCLC without Actionable Mutations after Introduction of Immunotherapy in Japan.

The aims of this study were to describe systemic treatment patterns and clinical outcomes for unresectable advanced/metastatic non-small-cell lung cancer (NSCLC) by first-line regimen type in real-world clinical settings in Japan after the introduction of first-line immune checkpoint inhibitor (ICI) monotherapy in 2017. Using retrospective chart review at 23 study sites, we identified patients ≥20 years old initiating first-line systemic therapy from 1 July 2017 to 20 December 2018, for unresectable stage IIIB/C or IV NSCLC; the data cutoff was 30 September 2019. Eligible patients had recorded programmed death-ligand 1 (PD-L1) tumor proportion score (TPS) and no known actionable EGFR/ALK/ROS1/BRAF genomic alteration. Kaplan-Meier method was used to determine time-to-event endpoints. Of 1208 patients, 647 patients (54%) received platinum doublet, 463 (38%) received ICI monotherapy, and 98 (8%) received nonplatinum cytotoxic regimen as first-line therapy. PD-L1 TPS was ≥50%, 1−49% and <1% for 44%, 30%, and 25% of patients, respectively. Most patients with PD-L1 TPS ≥50% received ICI monotherapy (453/529; 86%). Excluding 26 patients with ECOG performance status of 3−4 from outcome analyses, the median patient follow-up was 11.3 months. With first-line platinum doublet, ICI monotherapy, and nonplatinum cytotoxic regimens, median overall survival (OS) was 16.3 months (95% CI, 14.0−20.1 months), not reached, and 14.4 months (95% CI, 10.3−21.2 months), respectively; 24-month OS was 40%, 58%, and 31%, respectively. Differences in OS relative to historical cohort data reported in Japan are consistent with improvement over time in real-world clinical outcomes for advanced NSCLC.

Assessing Indoor Environmental Control Practices by Race/Ethnicity Among Children With Asthma in 14 US States and Puerto Rico, 2013-2014.

INTRODUCTION: In the United States, children in Puerto Rico and non-Hispanic black children in the mainland US have a higher burden of asthma than non-Hispanic white children in the mainland US. We examined indoor environmental control (IEC) practices that reduce asthma triggers, by race/ethnicity among children in the mainland US and Puerto Rico. METHODS: We used 2013 and 2014 data from the Behavioral Risk Factor Surveillance System Asthma Call-back Survey Child Questionnaire from 14 states and Puerto Rico to measure the association between race/ethnicity and IEC practices, adjusting for sociodemographic covariates, among children identified as ever receiving an asthma diagnosis. Racial/ethnic groups were compared in 14 US states using aggregated data. Separate analyses compared IEC practices for children diagnosed with asthma in Puerto Rico with children of all races/ethnicities diagnosed with asthma in 14 states. RESULTS: Among households in 14 US states that had a child with asthma, non-Hispanic black children were more likely than non-Hispanic white children to use an air purifier (36.8% vs 25.2%; adjusted odds ratio [aOR] = 2.0; 95% confidence interval [CI], 1.3-3.2) and avoid pets in the bedroom (87.9% vs 58.3%; aOR = 4.5; 95% CI, 2.3-8.8). Children in Puerto Rico were more likely than children in 14 states to use dust mite-impermeable pillow covers (53.7% vs 36.4%; aOR = 3.6; 95% CI, 1.8-7.1) and mattress encasements (60.3% vs 30.3%; aOR = 2.4; 95% CI, 1.2-4.8). CONCLUSION: IEC practices such as using air purifiers, pillow covers, mattress encasements, and avoiding pets in the bedroom vary by race/ethnicity among children with asthma. These findings show that vulnerable populations are using IEC practices, but asthma prevention and control measures should continue to be assessed.

A Narrative Review of Intensive Group Tobacco Treatment: Clinical, Research, and US Policy Recommendations.

INTRODUCTION: Clinical practice guidelines recommend comprehensive treatment for tobacco dependence including pharmacotherapies and behavioral interventions. Group counseling may deliver unique treatment aspects not available with other modalities. This manuscript provides a narrative review of group treatment outcomes from real-world practice settings and complements recent meta-analyses of randomized controlled trials (RCTs). Our primary goals were to determine whether group treatments delivered in these settings have yielded similar quit rates compared to individual treatment and to provide recommendations for best practices and policy. METHODS: Group treatment was defined as occurring in a clinical or workplace setting (ie, not provided as part of a research study), led by a professionally trained clinician, and offered weekly over several weeks. English language PubMed articles from January 2000 to July 2017 were searched to identify studies that included outcomes from both group and individual treatment offered in real-world settings. Additional data sources meeting our criteria were also included. Reports not using pharmacotherapy and research studies (eg, RCTs) were excluded. The primary outcome was short-term, carbon monoxide (CO)-validated point prevalence abstinence (4-week postquit date). RESULTS: The review included data from 11 observational studies. In all cases, group treatment(s) had higher 4-week CO-validated quit rates (range: 35.5%-67.3%) than individual treatment(s) (range: 18.6%-53.3%). CONCLUSIONS: Best practice group treatments for tobacco dependence are generalizable from research to clinical settings and likely to be at least as effective as intensive individual treatment. The added advantages of efficiency and cost-effectiveness can be significant. Group treatment is feasible in various settings with good results. IMPLICATIONS: A major barrier to achieving high rates of tobacco abstinence is under-utilization of evidence-based treatment interventions. This review demonstrates the effectiveness and utility of group treatment for tobacco dependence. Based on the available data described in this narrative review in conjunction with existing RCT data, group treatment for tobacco dependence should be established and available in all behavioral health and medical settings. Group tobacco treatment is now one of the mandated reimbursable tobacco treatment formats within the US health care system, creating enormous opportunities for widespread clinical reach. Finally, comprehensive worksite group programs can further extend impact.

Impact of Preexisting Mental Illness on All-Cause and Breast Cancer-Specific Mortality in Elderly Patients With Breast Cancer.

Purpose Limited data are available on the survival of patients with breast cancer with preexisting mental illness, and elderly women are of special interest because they experience the highest incidence of breast cancer. Therefore, we compared all-cause and breast cancer-specific mortality for elderly patients with breast cancer with and without mental illness. Methods A retrospective cohort study was conducted by using SEER-Medicare data, including 19,028 women ≥ 68 years of age who were diagnosed with stage I to IIIa breast cancer in the United States from 2005 to 2007. Patients were classified as having severe mental illness if an International Classification of Diseases, Ninth Edition, Clinical Modification code for bipolar disorder, schizophrenia, or other psychotic disorder was recorded on at least one inpatient or two outpatient claims during the 3 years before breast cancer diagnosis. Patients were followed for up to 5 years after breast cancer diagnosis to assess survival outcomes, which were then compared with those of patients without mental illness. Results Nearly 3% of patients had preexisting severe mental illness. We observed a two-fold increase in the all-cause mortality hazard between patients with severe mental illness compared with those without mental illness after adjusting for age, income, race, ethnicity, geographic location, and marital status (adjusted hazard ratio, 2.19; 95% CI, 1.84 to 2.60). A 20% increase in breast cancer-specific mortality hazard was observed, but the association was not significant (adjusted hazard ratio, 1.20; 95% CI, 0.82 to 1.74). Patients with severe mental illness were more likely to be diagnosed with advanced breast cancer and aggressive tumor characteristics. They also had increased tobacco use and more comorbidities. Conclusion Patients with severe mental illness may need assistance with coordinating medical services.

Participation in a Diabetes Self-Management Class Among Adults With Diabetes, New Jersey 2013-2015.

Identifying patient groups with low participation in diabetes self-management education can inform efforts to improve its use. Data from the 2013-2015 Behavioral Risk Factor Surveillance System were used to assess variation in participation in a diabetes self-management class in New Jersey. Nonparticipation varied significantly by race/ethnicity (P < .001), education (P < .001), health care coverage (P = .04), county (P < .001), years since diagnosis (P < .001), and whether a diabetes provider visit occurred in the past year (P = .002). Attention is warranted in identifying participation barriers among patients who live in certain counties, have less education, are without health care coverage, have been diagnosed with diabetes more recently, visit a provider less often, or belong to certain racial/ethnic minority groups.

Effects of a 2014 Statewide Policy Change on Cash-Value Voucher Redemptions for Fruits/Vegetables Among Participants in the Supplemental Nutrition Program for Women, Infants, and Children (WIC).

Purpose In 2014, the New Jersey Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) began requiring WIC-authorized stores to stock at least two fresh fruits and two fresh vegetables. We aimed to evaluate the effect of this policy change on fruit and vegetable purchases among WIC-participating households and to assess variation by household access to a healthy food store such as a supermarket or large grocery store. Description Households with continuous WIC enrollment from June 2013 to May 2015 were included (n = 16,415). Participants receive monthly cash-value vouchers (CVVs) to purchase fruits and vegetables. For each household, the CVV redemption proportion was calculated for the period before and after the policy by dividing the total dollar amount redeemed by the total dollar amount issued. Complete redemption was defined as a proportion ≥90% and the change in complete redemption odds was assessed after adjusting for Supplemental Nutrition Assistance Program participation. Assessment We observed a small increase following the policy change [odds ratio (OR) 1.10, 95% confidence interval (CI) 1.04-1.17]; however, the effect varied by healthy food access (p = 0.03). The odds increased for households with access to at least one healthy food store (OR 1.13, 95% CI 1.06-1.20) while no effect was observed for households without such access (OR 0.91, 95% CI 0.76-1.10). Conclusion Policy change was associated with a small increase in purchasing, but only among households with healthy food access. The state is addressing this gap through technical assistance interventions targeting WIC-authorized small stores in communities with limited access.

Diagnosis and treatment delays among elderly breast cancer patients with pre-existing mental illness.

PURPOSE: This study aimed to compare diagnosis and treatment delays in elderly breast cancer patients with and without pre-existing mental illness. METHODS: A retrospective cohort study was conducted using the Surveillance, Epidemiology and End Results-Medicare data including 16,636 women 68+ years, who were diagnosed with stage I-IIIa breast cancer in the United States from 2005 to 2007. Mental illness was identified using International Classification of Diseases, Ninth Revision, Clinical Modification codes recorded on inpatient and outpatient claims during the 3 years prior to breast cancer diagnosis. Patients were classified as having no mental illness, anxiety, depression, anxiety and depression, or severe mental illness (bipolar disorder, schizophrenia, and other psychotic disorder). Multivariable binomial regression was used to assess the association between mental illness and delays of ≥60 and ≥90 days after adjustment for confounders. RESULTS: Patients with comorbid anxiety and depression had an increased risk for diagnosis delay of ≥90 days from symptom recognition (RR 1.11; 95% CI 1.00, 1.23), and those with severe mental illness had an increased risk for initial treatment delay of ≥60 days from diagnosis (RR 1.36; 95% CI 1.06, 1.74). Patients with any mental illness experienced an increased risk for adjuvant chemotherapy delay of ≥90 days from last operation (RR 1.13; 95% CI 1.01, 1.26) and each category of mental illness, except depression, showed a non-significant trend for this association. CONCLUSION: Breast cancer patients with mental illness should be closely managed by a cross-functional care team, including a psychiatrist, a primary care physician, and an oncologist, to ensure adequate care is received within an appropriate timeframe.

Racial differences in the effects of comorbidity on breast cancer-specific survival.

PURPOSE: In an effort to explain racial disparities in breast cancer survival, this study aimed to investigate how comorbidity affects breast cancer-specific mortality by race. METHODS: A retrospective cohort study was conducted using the Surveillance, Epidemiology and End Results-Medicare linked data including 68,090 women 66+ years, who were diagnosed with stage I-III breast cancer in the United States from 1994 to 2004. Hospital and outpatient claims from the year prior to breast cancer diagnosis were used to identify comorbid conditions and patients were followed for survival through 2010. RESULTS: Competing risk survival analysis failed to demonstrate any negative comorbidity effects on breast cancer-specific survival for black women. An increased breast cancer-specific mortality hazard was observed for white women who had diabetes without complication relative to white women without this condition after adjusting for age and year of diagnosis (hazard ratio: 1.22, 95% confidence interval 1.13, 1.30). The Cochran-Armitage Test showed diabetes was associated with a later stage of diagnosis (p < 0.01) and a more aggressive tumor grade (p < 0.01) among white women in the study population. CONCLUSION: Race specific comorbidity effects do not explain breast cancer-specific survival disparities. However, the relationship between diabetes and breast cancer, including the role of aggressive tumor characteristics, warrants special attention.

Evaluating Community Measures of Healthy Food Access.

Several community level measures of healthy food access exist, but evaluation efforts have been limited leaving uncertainty about how to prioritize communities for intervention. This study aimed to assess several existing measures to inform statewide public health planning efforts in New Jersey, USA. We assessed agreement between community measures of healthy food access and then evaluated the predictive validity of each measure by describing its association with complete fruit and vegetable cash-value voucher redemption (proportion redeemed ≥70, ≥80, ≥90%) among 30,078 low-income households participating in the New Jersey Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) during 2013-2014. The United States Department of Agriculture's (USDA) food desert measure agreed with the Centers for Disease Control and Prevention's (CDC) no healthier food retailers (NHFR) measure for 76.5% of New Jersey census tracts, but the Kappa statistic was only 0.10. For urban households, the NHFR measure was negatively associated with complete redemption after adjusting for demographic factors and Supplemental Nutrition Assistance Program participation (≥70% odds ratio (OR) 0.68, 95% confidence interval (CI) 0.61-0.75; ≥80% OR 0.67, 95% CI 0.62-0.73; ≥90% OR 0.72, 95% CI 0.66-0.77). For rural households, a negative association was observed for the USDA's low-income/low-vehicle access measure (≥70% OR 0.48, 95% CI 0.26-0.90). The CDC's NHFR measure is more appropriate for prioritizing urban areas while the USDA's low-income/low-vehicle access measure may be better for rural areas.

Hormonal therapy for breast cancer and diabetes incidence among postmenopausal women.

PURPOSE: The purpose of this study was to evaluate associations between hormonal therapy for breast cancer and subsequent diabetes incidence. METHODS: The Surveillance, Epidemiology and End Results-Medicare linked data were used. Stage I-III breast cancer patients 65 years or older who filled at least two prescriptions for an aromatase inhibitor (AI) or tamoxifen by the end of 2008, and within 12 months of breast cancer diagnosis, were selected. Women without cancer from a 5% random sample of Medicare beneficiaries were frequency matched to patients by age group, and new onset diabetes was monitored for 24 months postbaseline in both groups of women. RESULTS: Cox-proportional hazards analysis failed to show an association between AI use and subsequent diabetes onset after adjusting for age, race, and comorbidity (hazard ratio: 0.99; 95% confidence interval: 0.84-1.18). This study also failed to show an association between tamoxifen use and diabetes onset (hazard ratio: 0.79; 95% confidence interval: 0.54-1.17). CONCLUSIONS: Study findings provide evidence that postmenopausal AI and tamoxifen users do not experience an increased risk of diabetes in the 2 years after treatment initiation. Whether these findings will hold with longer duration follow-up deserves a closer look.

Effects of breast cancer on chronic disease medication adherence among older women.

PURPOSE: The purpose of this study was to determine the effects of breast cancer on chronic disease medication adherence among older women. METHODS: The Surveillance, Epidemiology, and End Results (SEER)-Medicare-linked data and a 5% random sample of Medicare enrollees were used. Stage I-III breast cancer patients diagnosed in 2008 and women without cancer were eligible. Three cohorts of medication users 66+ years were identified using diagnosis codes and prescription fill records: diabetes, hypertension, and lipid disorders. For each cohort, breast cancer patients were frequency matched to comparison women by age and geographic area. Medication adherence was measured by the proportion of days covered and medication persistence. RESULTS: During the post-baseline period, the percentage of breast cancer patients who were non-adherent was 26.2% for diabetes medication, 28.9% for lipid-lowering medication, and 14.2% for hypertension medication. Breast cancer patients experienced an increased odds of diabetes medication non-adherence [odds ratio (OR) = 1.44; 95% confidence interval (CI) = 1.07 to 1.95] and were more likely to be non-persistent with diabetes medication (hazard ratio = 1.31; 95%CI: 1.04 to 1.66) relative to women without cancer. The study failed to show a difference between breast cancer and comparison women in the odds of non-adherence to hypertensive (OR = 0.87; 95%CI: 0.71 to 1.05) or lipid-lowering medication (OR = 0. 91; 95%CI: 0.73 to 1.13) with a proportion of days covered threshold of 80%. CONCLUSION: Special attention should be given to the coordination of primary care for older breast cancer patients with diabetes. Copyright © 2016 John Wiley & Sons, Ltd.