Racial disparities in medication use: imperatives for managed care pharmacy

Understanding contraceptive switching rationales from real world clinical notes using large language models

Understanding reasons for treatment switching is of significant medical interest, but these factors are often only found in unstructured clinical notes and can be difficult to extract. We evaluated the zero-shot abilities of GPT-4 and eight other open-source large language models (LLMs) to extract contraceptive switching information from 1964 clinical notes derived from the UCSF Information Commons dataset. GPT-4 extracted the contraceptives started and stopped at each switch with microF1 scores of 0.85 and 0.88, respectively, compared to 0.81 and 0.88 for the best open-source model. When evaluated by clinical experts, GPT-4 extracted reasons for switching with an accuracy of 91.4% (2.2% hallucination rate). Transformer-based topic modeling identified patient preference, adverse events, and insurance coverage as key reasons. These findings demonstrate the value of LLMs in identifying complex treatment factors and provide insights into reasons for contraceptive switching in real-world settings.

Disparities in Access to Valganciclovir Cytomegalovirus Prophylaxis in High-Risk African American Kidney Transplant Patients

BACKGROUND

While access and outcomes disparities for African American (AA) kidney transplant recipients are documented, there are limited studies assessing medication access disparities in transplantation. Cytomegalovirus (CMV) causes severe complications for transplant recipients, and we aimed to understand differences in access to CMV prophylaxis valganciclovir and its impact on CMV infection rates in AA transplant recipients.

METHODS

This single-center, retrospective longitudinal cohort study examined high-risk (CMV serostatus D+/R-) adult kidney transplant recipients between June 1, 2010, and May 31, 202, through EMR abstraction. Standard univariate comparative statistics were employed alongside binary logistic regression for multivariable modeling.

RESULTS

During the 10 year period, 418 kidney transplant recipients were included, with 179 (42.8%) identified as AA and 239 as non-AA. There were significant differences in mean age (p = 0.001) and private versus Medicaid insurance status (p < 0.001). AAs experienced higher death-censored graft loss rates (10.6% AA vs. 5.0% non-AA, p = 0.031). CMV infection rate, opportunistic infection rate, leukopenia incidence, and death did not differ significantly between AA and non-AA patients. AA patients were 42% less likely to receive valganciclovir out-of-pocket cost assistance compared to non-AA patients (OR 0.58, [0.379-0.892], p = 0.013). When incorporating age, Medicaid status, and donor marginality in a multivariable model, the impact of AA race on utilizing assistance programs became statistically non-significant (OR 0.70, [0.448-1.094], p = 0.118).

CONCLUSIONS

AAs were significantly less likely to leverage assistance programs or utilize personal resources to access valganciclovir. This disparity was partially explained by age, insurance status, and donor type. Despite this, CMV infection rates were similar between AA and non-AA cohorts.

Dynamic changes in medication burden leading to fall and hospital readmissions in older adults: Toward a strategy for improving risk and managing costs

The majority of a health plan's performance and designated Star Rating is related to medication-related behavior, eg, medication adherence, medication review, and reconciliation, that are intricately related to adverse drug events (ADEs). Altered pharmacodynamics and pharmacokinetics owing to aging make older adults more vulnerable to ADEs like falls, fractures, hospitalizations, and mortality. Prevention of avoidable risk factors such as medication burden can help maintain quality of life. Studies of multiple populations have established drug burden index (DBI), a dose-dependent measure of anticholinergic and sedative burden, to be strongly associated with worsening vertigo, dizziness, and balance, which all predicate falls. The mean difference in DBI greater than 0.1 provides greater predictive power for adverse events, such as falls and 30-day readmission rates. Inclusion of a DBI delta metric especially on an electronic medical record has the potential to reduce fall incidence and associated health outcomes such as hospitalizations and death; this presents an opportunity to improve Centers for Medicare & Medicaid Services Star Ratings by using meaningful tools to foster engagement among informed and active Medicare beneficiaries. We believe this information is extremely relevant in real-world decision-making for health care professionals, specifically when the changes are dynamic and happen very quickly. Moreover, managed care organizations are now dedicated to eliciting a deeper understanding and mitigation of social inequalities in medication use and consequences. Among the proposed solutions includes tailoring prescription utilization management tools with DBI to decrease avoidable incidences of complications and unintended costs. Understanding the dynamic relationship between medication exposures causing ADEs and associated health care utilization and costs to third-party payments remains vital because in the United States, approximately one-third of hospital admissions in older adults occur because of ADEs. This can be achieved by emphasizing equitable therapy and tailoring quality initiatives that minimize racial disparities and avoidable costs that affect the financial burden of these patients. Importantly, this approach becomes even more critical as health care systems increasingly emphasize star ratings, which reflect the quality of care delivered to patients. By prioritizing DBI metrics in these ratings, we can ensure that care is not only clinically effective but also equitable and focused on improving patients' overall well-being. Lastly, as the future directions, the timely application of advanced technologies like artificial intelligence and machine learning in analyzing DBI metrics could enhance our ability to predict the value of DBI adjustments and their correlation with falls and other unintended ADEs. These real-world technologies can process vast amounts of data quickly and accurately, identifying patterns and potential risks that might otherwise go unnoticed.

Ethnic-racial disparities in poisoning cases: analysis of drugs of abuse, medicines and pesticides in Brazil

In Brazil, ethnic-racial inequalities exist in all fields, obstructing access to goods, services, and opportunities, including healthcare services. However, there are no apparent studies that assess, at a national level, ethnic-racial disparities in poisoning cases, emphasizing skin color as a determining factor. The study aimed to examine the relationship between race/ethnicity and general poisoning cases, by medications, pesticides, and drug of abuse in Brazilian states. Poisoning cases data were extracted for the years 2017, 2018, and 2019. Notification data for general poisoning cases and toxic agents were collected: medications, pesticides, and drugs of abuse. Data were categorized between whites and non-whites (blacks, browns, and indigenous) and without information on skin color/ethnicity. Rates of poisonings amongst ethnic-racial groups and cases of not declared skin color as well as relative risk (RR) of poisoning among non-whites were calculated. All states in the North, Northeast (states with the worst Human Development Index), Midwest, and 2 states in the Southeast exhibited higher rates of poisoning cases per 100,000 inhabitants among non-whites. The RR values for nonwhite individuals were higher in the North and Northeast regions for all types of poisonings. The type of poisoning cases that presented the highest RR for non-whites over the 3 years was drugs of abuse (2-2.44), when compared to other types of poisonings from pesticides (2-2.33) and medications (1.5-1.91). The spatial distribution of poisoning cases rates and RR of nonwhite population support public policies to reduce socioeconomic and environmental inequalities.

Total cost of care of Medicare Advantage beneficiaries participating in an appointment-based model in a national pharmacy chain

BACKGROUND

The appointment-based model (ABM) is a pharmacy service to improve medication-related health outcomes. ABM involves medication synchronization and medication review, plus other services such as medication reconciliation, medication therapy management, vaccine administration, and multimedication packaging. ABM can improve medication adherence, but the economic impact is unknown.

OBJECTIVE

To assess the effect of a national pharmacy chain's ABM program for Medicare Advantage beneficiaries on total cost of care (TCOC).

METHODS

This study analyzed administrative claims data from April 7, 2017, through February 29, 2020, for Medicare Advantage beneficiaries with Part D using a propensity score-matched cohort design. The national pharmacy chain provided a list of ABM participants. Eligibility criteria for the ABM and control (non-ABM) groups included age 65 years or older on the index date (initial participation, ABM; random fill date, control) and continuous enrollment from at least 6 months pre-index (baseline) date through at least 6 months post-index (follow-up) date. Medical inflation-adjusted (2020) TCOC was calculated as the sum of all health care spending from Medicare Advantage beneficiaries with Part D plan and patient paid amounts, standardized to per patient per month (PPPM), during the follow-up period. Secondary outcomes included medication adherence calculated across prevalent maintenance therapeutic classes using proportion of days covered (PDC).

RESULTS

Each group contained 5,225 patients with balanced characteristics after matching: 64% female, 73% White, mean age 75 years, mean Quan-Charlson comorbidity index score 0.9, and hypertension and dyslipidemia, each >65%. Median baseline all-cause PPPM health care costs in the ABM and control groups, respectively, were $517 and $548 ($221 and $234 medical, $135 and $164 pharmacy). Baseline PDC of at least 80% was 83% in the ABM group and, similarly, 84% in the control group. The mean (SD) follow-up was 604 (155) days for the ABM group and 598 (151) days for the control group. During the follow-up period, the median PPPM TCOC for the ABM group was $656 and was $723 for the control group (P = 0.011). Median pharmacy costs were also significantly less in the ABM group ($161 vs $193, P < 0.001), whereas median medical costs were $328 in the ABM group and $358 among controls (P = 0.254). More patients in the ABM group were adherent during follow-up, with 84% achieving PDC of at least 80% vs 82% among controls (P = 0.009).

CONCLUSIONS

The ABM program was associated with significantly lower follow-up median total costs (medical and pharmacy), driven primarily by pharmacy costs. More patients were adherent in the ABM program. Payers and pharmacies can use this evidence to assess ABM programs for their members.

Reducing disparities in medication use: Responding to managed care pharmacy's imperatives

Although disparities and inequities in health status and access to health care services have long existed in our nation, the COVID-19 pandemic cast a bright spotlight on them. Communities of color and socioeconomically disadvantaged populations were disproportionally affected by the pandemic. These same populations suffer from higher prevalences of chronic illnesses, which puts them at greater risk for poor outcomes associated with SARS-CoV-2. At long last, in the wake of the pandemic, the health care community began to acknowledge improving health equity as a public health imperative. In a November 2020 JMCP Viewpoints article, Dr Stephen Kogut of the University of Rhode Island College of Pharmacy presented an insightful analysis of disparities in medication use (DMU) and offered 4 suggestions on how the managed care pharmacy community can help eliminate DMU. This Viewpoints article assesses what progress has been made in addressing those imperatives and proposes further steps that should be taken. Although the managed care pharmacy community has broadly acknowledged the existence of DMU and taken steps to mitigate them, there is much work to do in examining and improving benefit design and coverage policies; collecting and reporting data on race and ethnicity and DMU; incorporating the perspectives of patients, including those representing minority populations, in benefit design and coverage policies; and addressing the challenges associated with traditional cost-sharing models. The entire managed care pharmacy community, including AMCP and other membership organizations, must remain steadfast in its efforts to improve health equity and eliminate DMU.

Racial and social inequities in medication use: A review of articles responding to the Journal of Managed Care + Specialty Pharmacy's Call to Action

This article provides a summary of Viewpoint and Research articles responding to the 2020 Journal of Managed Care + Specialty Pharmacy Call to Action to address racial and social inequities in medication use. We find great heterogeneity in terms of topic, clinical condition examined, and health disparity addressed. Common recommendations across Viewpoint articles include the need to increase racial and ethnic diversity in clinical trial participants, the need to address drug affordability and health insurance literacy, and the need to incentivize providers and plans to participate in diversity initiatives, such as the better capture of information on social determinants of health (SDOH) in claims data to be able to address SDOH needs. Across research articles, we also find a large range of approaches and study designs, spanning from randomized controlled trials to surveys to observational studies. These articles identify disparities in which minoritized beneficiaries are shown to be less likely to receive medications and vaccines, as well as less likely to be adherent to medications, across a variety of conditions. Finally, we discuss Healthy People 2030 as a potential framework for future health disparity researchers.

The Impact of Prescription Drug Coverage on Disparities in Adherence and Medication Use: A Systematic Review

Prescription drug cost-sharing is a barrier to medication adherence, particularly for low-income and minority populations. In this systematic review, we examined the impact of prescription drug cost-sharing and policies to reduce cost-sharing on racial/ethnic and income disparities in medication utilization. We screened 2,145 titles and abstracts and identified 19 peer-reviewed papers that examined the interaction between cost-sharing and racial/ethnic and income disparities in medication adherence or utilization. We found weak but inconsistent evidence that lower cost-sharing is associated with reduced disparities in adherence and utilization, but studies consistently found that significant disparities remained even after adjusting for differences in cost-sharing across individuals. Study designs varied in their ability to measure the causal effect of policy or cost-sharing changes on disparities, and a wide range of policies were examined across studies. Further research is needed to identify the types of policies that are best suited to reduce disparities in medication adherence.

Delayed treatment initiation of oral anticoagulants among Medicare patients with atrial fibrillation

Study objective

This study aimed to identify factors associated with delayed oral anticoagulant (OAC) treatment initiation among atrial fibrillation (AF) patients in United States (US) clinical practice.

Participants

Medicare beneficiaries newly diagnosed with AF without moderate-to-severe mitral stenosis or a mechanical heart valve, were aged ≥65 years and prescribed OAC on or after 10/1/2015 through 2019 were included. Delayed and early OAC initiation were defined as >3 months and 0-3 months initiation from first AF diagnosis, respectively.

Main outcome measures

Association between delayed OAC initiation and patient demographics, clinical and index OAC coverage and formulary characteristics was examined using multivariable logistic regression.

Results

A total of 446,441 patients met the inclusion criteria; 30.0 % (N = 131,969) were identified as delayed and 70.0 % (N = 314,472) as early OAC initiation. Median age for both cohorts was 78 years. In the early and delayed OAC cohorts, 47.1 % and 47.6 % were male and 88.8 % and 86.6 %, were White, respectively. Factors associated with delayed OAC initiation (odds ratio; 95 % confidence interval) included Black race (1.29; 1.25 to 1.33), west region (1.29; 1.26 to 1.32), comorbidities such as dementia (1.27; 1.23 to 1.30), recent bleeding hospitalization (1.22; 1.18 to 1.27), prior authorization (1.69; 1.66 to 1.71), tier 4 formulary for index OAC at AF diagnosis (1.26; 1.22 to 1.30).

Conclusion

Our study revealed that nearly one-third of Medicare patients with AF experienced delayed OAC initiation. Key patient characteristics found to be associated with delayed OAC initiation included race and ethnicity, comorbidities, and formulary restrictions.

Cost sharing in managed care and the ethical question of business purpose

For-profit managed care organizations face decisions about cost sharing that can involve a tradeoff between the interests of investors and the interests of patients. No successful business can ignore the interests of its investors, but moral philosophy points to ethical reasons for managed care organizations to make patients' health, rather than investors' profit, their primary goal. One reason is the ethical obligation of all businesses to avoid wrongful exploitation of vulnerable customers. An insurance company's cost-sharing policy can exploit customers either by collecting an unfairly large amount of money from them or by unfairly deterring them from making claims for resources they medically need. Another reason stems from the fact that managed care organizations' profits derive in part from the existence of artificial barriers to access to medicine, notably including patents. Putting a fence around a water well in the desert is legitimate only if doing so facilitates a financial arrangement that maximizes people's access to water they need. Likewise, patents and other artificial barriers to access to medically necessary drugs are legitimate only if they are used to help finance access to medical resources people need. For these reasons, managed care organizations should make cost-sharing decisions that maximize the sustainable availability of effective drugs to patients who need them. DISCLOSURES: The thoughts and opinions expressed in this article are those of the author only and are not the thoughts and opinions of any current or former employer of the author. Nor is this publication made by, on behalf of, or endorsed or approved by any current or former employer of the author.

Patterns and disparities in prescribing of opioids and benzodiazepines for older adults in North Carolina

BACKGROUND

We characterized real-world prescribing patterns of opioids and benzodiazepines (BZDs) for older adults to explore potential disparities by race and sex and to characterize patterns of co-prescribing.

METHODS

A retrospective evaluation was conducted using electronic health data for adults ≥65 years old who presented to one of 15 primary care practices between 2019 and 2020 (n = 25,141). Chronic opioid and BZD users had ≥4 prescriptions in the year prior, with at least one in the last 90 or 180 days, respectively. We compared demographic characteristics between all older adults versus chronic opioid and BZD users. We used logistic regression to identify characteristics (age, sex, race, Medicaid use, fall history) associated with opioid and BZD co-prescribing.

RESULTS

We identified 833 (3.3%) chronic opioid and 959 chronic BZD users (3.8%) among all older adults seen in these practices. Chronic opioid users were less likely to be Black (12.7% vs. 14.3%) or other non-White race (1.4% vs. 4.3%), but more likely to be women (66.8% vs. 61.3%). A similar trend was observed for BZD users, with less prescribing among Black (5.4% vs. 14.3%) and other races (2.2% vs. 4.3%) older adults and greater prescribing among women (73.6% vs. 61.3%). Co-prescribing was observed among 15% of opioid users and 13% of BZD users. Co-prescribing was largely driven by the presence of relevant co-morbid conditions including chronic pain, anxiety, and insomnia rather than demographic characteristics.

CONCLUSIONS

We observed notable disparities in opioid and BZD prescribing by sex and race among older adults in primary care. Future research should explore if such patterns reflect appropriate prescribing or are due to disparities in prescribing driven by biases related to perceived risks for misuse.

Theory Integration for Examining Health Care Discrimination among Minoritized Older Adults with Chronic Illness

Prevalence of chronic illnesses, including type 2 diabetes (T2DM), is increasing disproportionately among Latinx adults in the United States. Health care inequities such as health care discrimination contribute to the disparities in this population. Academic and clinical nurses must address health care discrimination from a strong theoretical framework. In this article, we integrate the minority stress theory and ecosocial theory of disease distribution to offer a whole-person model that identifies the concepts most relevant to Latinx older adults who function at multiple levels of intersectionality. This paper uses T2DM as an exemplar of chronic illness. The integrated model depicts possible pathways of physiological and psychological embodiment of lived experiences of minoritized older persons managing chronic illness who are living in a society deeply embedded with structural racism and oppression. This model may guide future research aimed at elucidating the social and structural determinants that impact health-related outcomes among Latinx older adults.

The role of integrated medical and prescription drug plans in addressing racial and ethnic disparities in medication adherence

Medication nonadherence in the United States contributes to 125,000 deaths and 10% of hospitalizations annually. The pain of preventable deaths and the personal costs of nonadherence are borne disproportionately by Black, Latino, and other minority groups because nonadherence is higher in these groups due to a variety of factors. These factors include socioeconomic challenges, issues with prescription affordability and convenience of filling and refilling them, lack of access to pharmacies and primary care services, difficulty taking advantage of patient engagement opportunities, health literacy limitations, and lack of trust due to historical and structural discrimination outside of and within the medical system. Solutions to address the drivers of lower medication adherence, specifically in minority populations, are needed to improve population outcomes and reduce inequities. While various solutions have shown some traction, these solutions have tended to be challenging to scale for wider impact. We propose that integrated medical and pharmacy plans are well positioned to address racial and ethnic health disparities related to medication adherence. DISCLOSURES: This study was not supported by any funding sources other than employment of all authors by Humana Inc. Humana products and programs are referred to in this article.

Emerging treatments for lupus nephritis: health equity considerations in clinical research and coverage

DISCLOSURES: No funding was received for the writing of this commentary. The authors report no conflicts of interest.

AMCP Partnership Forum: Racial health disparities-a closer look at benefit design

As part of its stand against racial injustice and its commitment to action, AMCP dedicated a partnership forum to discussion of potential sources of racial health disparities and inequities in benefit design because these are primary drivers of medication use. This forum, held virtually March 23-24, 2021, convened more than 40 experts representing key stakeholders from managed care settings. Key principles that emerged from the forum discussion as means to mitigate racial health disparities were to acknowledge that structural racism exists and that it can impact the provision of health care, including but not limited to formulary development and benefit design processes; to integrate proactive strategies to improve equity, beginning with education and training, into all aspects of health care; and to view patients holistically with an understanding of the compounding effect of social determinants of health on their personal wellness. With these principles in mind, participants highlighted several priority considerations including improving existing gaps in data, addressing diversity and equity as they relate to formulary development, evaluating systems such as benefit offerings through a lens of increasing equity, recognizing cost-related factors that affect equity, considering patients' interactions with and their ability to access the system, and committing to patient-centered care. Participants also suggested areas for policy-related focus and noted the need to develop and deploy specific education and training. DISCLOSURES: This forum was sponsored by Amgen, the National Pharmaceutical Council, Pfizer, PhRMA, Sandoz, and Takeda. Their representatives joined the forum as panelists or participants. These proceedings were prepared as a summary of the forum to represent common themes; they are not necessarily endorsed by all attendees nor should they be construed as reflecting group consensus.