“It’s just not easy to understand”: Health insurance literacy and insurance plan decision-making among cancer survivors

16

Background: Little is known about how cancer survivors choose health insurance plans and whether low health insurance literacy (HIL) may lead to inadequate coverage or high out-of-pocket (OOP) costs. Methods: This prospective, explanatory mixed methods study assessed health insurance decision-making in a sample of adult cancer survivors < 5 years from diagnosis. Quantitative eye-tracking data were collected from two simulated health insurance plan choice sets to gauge interest in plan benefits (measured by dwell time, or seconds of no eye movement; longer dwell time indicates more interest). Ranking of the most important and confusing plan benefit, HIL (HIL Measure; scored 20-84, higher scores indicate higher HIL), and sociodemographic data were also captured. Differences in benefit-specific dwell time by HIL were estimated using adjusted linear models. Qualitative, semi-structured interviews clarified a subset of survivors’ health insurance plan decision-making. Content analysis identified emergent patterns and themes. Results: With a median HIL score of 60 (IQR 50-70), survivors (N = 80) were 65% female, 59% non-Hispanic White, and a median age of 43 at diagnosis (IQR 34-52). Breast cancer diagnoses were most common (38%) and most survivors had received surgery (75%), radiation (55%), or systemic therapy (75%). In the choice set comparing a traditional and high-deductible health plan, survivors were most interested in prescription drug costs (median dwell time 58s, IQR 34-109), but no differences in interest were found comparing survivors with high and low HIL in adjusted models. In the choice set comparing a health maintenance organization and preferred provider organization plan, survivors were most interested in test and imaging costs (40s, IQR 14-67), with survivors with low vs. high HIL having more interest in deductible (β = 19s, 95% CI 2-38) and hospitalization costs (β = 14s, 95% CI 1-27) in adjusted models. Survivors with low vs. high HIL more often ranked low OOP maximums as the most important benefit (53% vs. 38%) and coinsurance as most confusing (68% vs. 53%). Interviews (n = 20) revealed survivors felt alone “ to do their own research” about insurance choices. OOP maximums were cited as the deciding factor since it’s “how much money is going to be taken out of my pocket.” Coinsurance was referred to “rather than a benefit, it's a hindrance.” Opinions about specialist gatekeeping were mixed; some survivors valued the accompanying lower OOP costs, while others worried about care delays. Survivors also reported issues with coverage and cost transparency post-insurance decision. Conclusions: Benefit considerations in health insurance decisions are varied and confusing for cancer survivors with both high and low HIL. Interventions to aid in health insurance understanding and choice are needed to optimize plan choice and prevent cancer-related financial hardship.

Associations of employment disruptions and financial hardship among individuals diagnosed with cancer in the United States

47

Background: Financial hardship is experienced by approximately half of working-age individuals diagnosed with cancer. Many working individuals diagnosed with cancer also experience disruptions with their employment. This study examines whether employment disruptions are associated with financial hardship among individuals diagnosed with cancer in the US. Methods: We utilized data from the 2016/2017 Medical Expenditure Panel Survey (MEPS) Experiences with Cancer self-administered questionnaires and identified individuals diagnosed with cancer at age >18 who worked for pay at the time of or following their cancer diagnosis. Employment disruption was defined as taking extended paid time off work; switching to a part time or less demanding job or to a flexible work schedule; and/or retiring early due to cancer, cancer treatment, or late effects. Financial hardship was defined in 3 domains: material (borrowing money, financial sacrifices, or being unable to cover medical costs); psychological (worrying about medical bills, financial stability, keeping job/income, or future earnings); and behavioral (delaying/forgoing medical care because of cost). Multivariable logistic regression analyses adjusting for the MEPS survey design were used to determine associations of employment disruption due to cancer with any financial hardship and with financial hardship intensity while controlling for patient demographic, health insurance, and clinical characteristics. Results: Among 732 individuals with a cancer history, 47.4% experienced employment disruptions and 55.9% experienced any financial hardship. Financial hardship was significantly more common among cancer survivors with employment disruption than without disruptions (68.7% vs. 44.5%; p-value of difference < 0.0001). Findings were consistent across multiple hardship measures and domains. Cancer survivors with employment disruptions were more likely to have greater (OR = 2.8; 95% CI 2.0, 3.9) financial hardship intensity. Individuals of race/ethnicity other than non-Hispanic White were also more likely to experience financial hardship while older individuals, women, and those with college education or who were married were less likely to experience financial hardship. Conclusions: Employment disruptions are common and significantly associated with multiple types of financial hardship among cancer survivors. Employer workplace accommodation and other policies to minimize disruptions among individuals diagnosed with cancer may reduce financial hardship in this vulnerable population.

Oncologist consideration of patient health insurance coverage and out-of-pocket costs for genomic testing in treatment decision

6600

Background: Use of genomic testing, especially multi-marker tumor panels, is increasing in the United States. Not all tests and related treatments are covered by health insurance, which can result in substantial patient out-of-pocket (OOP) costs. Although most patients are concerned about OOP costs, little is known about oncologists’ treatment decisions with respect to patient health insurance coverage and OOP costs for genomic testing. Methods: We identified 1,049 oncologists who reported using multi-marker tumor panels from the 2017 National Survey of Precision Medicine in Cancer Treatment. Separate multivariable ordinal logistic regression analyses were used to assess the associations of oncologist, practice, and patient characteristics and the oncologist ratings of the importance of health insurance coverage and OOP cost for genomic testing as part of treatment decisions. Results: Among oncologists, 47.3%, 32.7% and 20.0% reported that patient insurance coverage for genomic testing was very important, somewhat important, and a little/not important, respectively, in treatment decisions. 56.9%, 28.0%, and 15.2% reported patient OOP costs for genomic testing were very, somewhat, or a little/not important in treatment decisions, respectively. In adjusted ordinal logistic regression analyses, oncologists who used next-generation gene sequencing tests were more likely to report patient health insurance and OOP costs for testing as important (odds ratio (OR) = 2.0; 95% confidence interval (CI): 1.2, 3.5) and (OR = 2.1; 95%CI: 1.2, 3.7), respectively) in treatment decisions. Oncologists with more years of experience, who treated solid tumors (rather than only hematological cancers), worked in practices without molecular tumor boards for genomic tests, and with higher percentages of patients insured by Medicaid or self-paid/uninsured also reported insurance coverage or OOP costs for testing were important in treatment decisions (all p < 0.05). Conclusions: Physician, practice, and patient characteristics were associated with oncologists’ ratings of the importance of patient health insurance and OOP costs in treatment decisions. Identifying factors that influence physicians’ priorities in treatment decisions may inform the development and targeting of interventions to support patient and physician discussions about oncology care.

Medication nonadherence and patient cost responsibility for rural and urban cancer survivors

4

Background: Little is known about the specific out-of-pocket costs which may lead to prescription nonadherence in older cancer survivors, and how patterns may differ for those living in rural areas. This study quantified patient costs overall and by residence for older cancer survivors who did and did not report cost-related prescription nonadherence. Methods: This retrospective cohort study used data from the Surveillance, Epidemiology, and End Results Program, Medicare claims, and the Consumer Assessment of Healthcare Providers and Systems survey linked data resource (SEER-CAHPS) from 2007-2015. Older cancer survivors self-reported cost-related prescription nonadherence in the prior six months. Patient cost responsibility (deductibles, coinsurance, copayments) was summed for all medical care received in the year prior to survey. Differences in patient cost responsibility by cost-related adherence was estimated using gamma generalized linear models adjusted for patient age, race, sex, education, dual Medicaid enrollment status, residence, comorbidity count, cancer type, stage, and phase of care. Models stratified by urban/rural residence as designated by Rural-Urban continuum codes assessed effect modification. Results: Of 11,829 older adult survivors of prostate (37%), breast (32%), colorectal (14%), gynecologic (10%), or lung (6%) cancer, 12% reported any cost-related prescription nonadherence in the prior year. Median age of survivors was 76 (interquartile range [IQR] 71-82), 15% had less than a high school degree, 59% had at least one non-cancer comorbidity, and 16% had ever been dual eligible. Prevalence of cost-related nonadherence was similar by patient characteristics. Median cost responsibility in the year prior to survey was $1,529 (IQR $744-$2,959) for patients reporting nonadherence and $1,123 (IQR $572-$2,362) for those reporting adherence. In adjusted models, patients reporting nonadherence had $656 higher patient cost responsibility in the year prior (95% CI $564-$760) compared to those reporting adherence. Approximately half of the difference in cost was outpatient spending (β = $277, 95% CI $210-$359). Differences in cost responsibility for patients reporting nonadherence compared to adherence were smaller for patients residing in rural areas (18% of respondents; β = $341, 95% CI $177-$564) compared to those residing in urban areas (82% of respondents; β = $715, 95% CI $613-$830). Conclusions: Compared to those reporting adherence, cost-related prescription nonadherence was associated with higher health care cost responsibility in cancer survivors. Furthermore, prescription adherence decisions may be more cost-sensitive for patients living in rural compared to urban areas. Interventions to address out-of-pocket health care costs, particularly for rural cancer survivors, could aid in increased prescription adherence and subsequent health outcomes.

Accessing Targeted Therapies: A Potential Roadblock to Implementing Precision Oncology?

PURPOSE

Advances in genomic techniques have led to increased use of next-generation sequencing (NGS). We evaluated the extent to which these tests guide treatment decisions.

METHODS

We developed and distributed a survey assessing NGS use and outcomes to a survey pool of ASCO members. Comparisons between groups were performed with Wilcoxon two-sample, chi-square, and Fisher's exact tests.

RESULTS

Among 178 respondents, 62% were male, 54% White, and 67% affiliated with academic centers. More than half (56%) indicated that NGS provided actionable information to a moderate or great extent. Use was highest (median ≥ 70% of cases) for lung and gastric cancer, and lowest (median < 25% of cases) in head and neck and genitourinary cancers. Approximately one third of respondents reported that, despite identification of an actionable molecular variant, patients were sometimes or often unable to access the relevant US Food and Drug Administration-approved therapy. When NGS did not provide actionable results, individuals reporting great or moderate guidance overall from NGS in treatment recommendations were more likely to request the compassionate use of an unapproved drug (P < .001), enroll on a clinical trial (P < .01), or treat off-label with a drug approved for another indication (P = .02).

CONCLUSION

When NGS identifies an actionable result, a substantial proportion of clinicians reported encountering challenges obtaining approved therapies on the basis of these results. Perceived overall impact of NGS appears associated with clinical behavior unrelated to actionable NGS test results, including pursuing off-label or compassionate use of unapproved therapies or referring to a clinical trial.

NGS testing use and results: A survey of U.S. oncologists

e19145

Background: Scientific advances in genomics have ushered in new cancer tests and therapies. Many researchers, clinicians and advocates wonder how breakthroughs are implemented in practice. Methods: We examined Next Generation Sequencing (NGS) use, results, actions and outcomes among US oncologists (oncs) in treatment of their patients (pts) with advanced cancer (AC). 178 physician members of ASCO who used NGS for AC pts in past 3 months completed an online survey (response rate=30%). Results: Table shows NGS use, outcomes, and perceptions of treatment response among oncs who treat AC pts. Results from questions that were NOT asked by cancer type: 19% of respondents reported that NGS results guided treatment recommendations to a great extent, and 37% to a moderate extent, 65% reported tests often provided no actionable information; among oncs reporting NGS results provided actionable information, 31% reported they sometimes/often tried but couldn’t obtain the FDA-approved drug; respondents used many approaches to “decide what to do” with results. 83% explored trials, 68% reviewed scientific literature, 64% explored feasibility of FDA-indicated drug, 52% used NGS report recommendations, 47% talked with colleague, and 23% accessed molecular tumor board most (>60%) of the time. Conclusions: Oncologists reported high use of NGS results to guide treatment decisions for pts with select ACs and an associated benefit in some pts. Two-thirds of oncs often experienced lack of actionable information across a broad range of tumor types. Even when NGS yielded actionable information, it was common for oncs to experience problems obtaining FDA-approved drugs. The large number of approaches oncs used to make treatment decisions may suggest an important opportunity to improve decision-making efficiency in healthcare delivery. [Table: see text]

Detection of 'Candidatus Xenohaliotis californiensis' (Rickettsiales-like prokaryote) inclusions in tissue squashes of abalone (Haliotis spp.) gastrointestinal epithelium using a nucleic acid fluorochrome

Rickettsiales-like prokaryotes appear to be etiologic agents of a number of newly described diseases of fish and shellfish. 'Candidatus Xenohaliotis californiensis' is a Rickettsiales-like prokaryote responsible for withering syndrome, a fatal disease of wild and farmed Eastern Pacific abalone, Haliotis spp. The bacterium proliferates in gastrointestinal epithelial cells, forming large intracytoplasmic inclusions. We describe a method of rapidly detecting and assessing the intensity of 'Candidatus Xenohaliotis californiensis' infections in abalone gastrointestinal tissue using the nucleic acid-specific fluorochrome Hoechst 33258. In excised tissue pieces dried onto slides, rehydrated in the Hoechst stain and viewed with ultraviolet light, the large bacterial inclusions were strongly fluorescent and could be easily distinguished from smaller host cell nuclei. This provided a rapid, inexpensive alternative to paraffin section microscopy or molecular techniques, allowing detection of the pathogen within minutes of tissue excision. Comparison of the fluorochrome method with conventional histological analysis for the ability to detect inclusions in 109 samples was 90% accurate, with discrepancies due to false negative diagnosis of low-level infections. An alternative nucleic acid-specific fluorochrome, propidium iodide, showed a staining pattern identical to that of Hoechst 33258. These methods should prove useful for the rapid detection of inclusion-forming Rickettsiales-like prokaryotes in tissues from many host species.