Purpose Variations in the medication treatment received by racial and ethnic minorities and women and the negative health outcomes associated with these differences were examined.
Methods Studies published between January 1990 and June 2008 were identified via electronic searches of MEDLINE, PsychINFO, International Pharmaceutical Abstracts, PubMed, and CINAHL using search terms related to race, ethnicity, sex, drug treatment, and disparity or variation. Articles were excluded if they addressed only medical or surgical care or did not include a statistical analysis of differences in drug treatment based on race, ethnicity, or sex. Data regarding the frequency of reported race, ethnic, and sex differences in medication treatment, the types of treatment differences observed, and associated health outcomes were extracted.
Results A total of 311 research articles were identified that investigated whether race, ethnicity, or sex was associated with disparities in medication treatment. Seventy- seven percent (n = 240) of included articles revealed significant disparities in drug treatment across race, ethnicity, and sex (p < 0.05). The most frequent disparity, found in 73% of the articles studied, was differences in the receipt of prescription drugs; however, documented disparities occurred related to differences in the drugs prescribed, drug dosing or administration, and wait time to receipt of a drug. Documented outcomes associated with pharmacotherapeutic disparities included increased rates of hospitalization, decreased rates of therapeutic goal attainment (e.g., low- density-lipoprotein cholesterol, blood pressure goals), and decreased rates of survival.
Conclusion A literature review revealed significant disparities in the medication treatment received by racial and ethnic minorities and women.
The United States spends more than any other nation in the world on health care, but despite consistent spending increases shouldered by insurers, patients, employers, and taxpayers, significant health disparities persist. The term health disparities is not universally defined; yet, most definitions include an implied, if not explicit, provision that such disparities extend beyond mere mathematical or statistical differences in measurable health-related variables. The Institute of Medicine (IOM) has defined health disparities as “differences in the quality of health care that are not due to access-related factors or clinical needs, preferences, and appropriateness of intervention.”1 Similarly, the World Health Organization described disparities as “differences in health which are not only unnecessary and avoidable but, in addition, are considered unfair and unjust.”2 In addition to perpetuating social and historical inequities, health disparities are often reflected in negative health outcomes, including higher mortality rates, greater burdens of disease and disability, and a reduced quality of life for those populations that experience disparate medical care and treatment.2
It was not until 1985 that the U.S. Department of Health and Human Services (DHHS) established a separate Office of Minority Health.3 Consequently, research analyzing racial and ethnic disparities in health status, quality of care, and access proliferated and was summarized and evaluated in an IOM report published in 2003, entitled Unequal Treatment: Confronting Racial and Ethnic Disparities in Health Care.1 In January 2000, DHHS launched Healthy People 2010, a comprehensive, nationwide health promotion and disease prevention agenda calling for the elimination of all health disparities, including differences that occur due to ethnicity, sex, education, income, disability, geographic location, or sexual orientation.4 Eliminating health disparities represents a tremendous challenge in preventing and treating diseases and improving the health and quality of life of all people.
Despite a national research agenda targeted at reducing or eliminating health disparities, little research or attention has been placed specifically on the role of pharmacy and pharmacists in reducing health disparities, particularly identifying, quantifying, and rectifying disparities related to or caused by disparate treatments with pharmaceuticals—we, the authors, have termed such differences “pharmacotherapeutic disparities.” Pharmacotherapeutic disparities can encompass differences in the receipt or nonreceipt of drugs, drugs selected, dosage or intensity of treatment, method of drug delivery (e.g., oral versus injection), and wait time to drug administration. Existing literature reviews concerning racial, ethnic, and sex disparities in drug treatment have investigated disparities within a single disease or condition rather than examining overall disparities in medication treatment as the primary methodological and analytic focus. This article critically examines the research on racial, ethnic, or sex disparities in the medication management or treatment of diseases or medical conditions and introduces and applies a new term (pharmacotherapeutic disparities) for a potentially underrecognized and therefore unaddressed health care disparity.
The authors, with the assistance of health science librarians from the University of Arizona Health Sciences Library, developed a search strategy, formulated search terms, and conducted a literature review to examine research articles published between January 1990 and June 2008 that addressed the influence of race, ethnicity, or sex on patients’ drug treatment in the United States (including U.S. Territories). The following approach was used:
CINAHL, Ovid/MEDLINE, and PsycINFO were searched using the terms sex or gender or ethnicity or race (in the text search field), drug therapy (in the subject search field), and difference or disparity (in the abstract search field).
International Pharmaceutical Abstracts was searched using the terms Native American, Hispanic American, or African American AND drug therapy, and results were limited to journal articles.
PubMed was searched using the MESH terms Sex Factors; Indians, North American, African Americans; Hispanic Americans AND Drug Therapy OR “Chemotherapy, Adjuvant.”
This initial search returned 45,881 articles. Articles were then excluded if they did not examine health disparities, resulting in 44,669 exclusions. Of the 1,212 remaining articles, 901 were excluded because they were not in English, lacked a comparison group, addressed only medical or surgical care, did not include statistical analysis that analyzed differences in drug treatment provided to at least two defined subject groups (e.g., whites versus African Americans, whites versus Hispanics, men versus women), or did not report on original research that occurred within the United States or its territories; 311 articles remained.5–315 A disparity, for purposes of this literature review, was defined as a significant difference in medication treatment that favored males, persons of white race, or persons of non-Hispanic ethnicity and disfavored females, persons of African-American, Native American–Alaska Native, or Asian–Pacific Islander race, or persons of Hispanic ethnicity. Males, whites, and non-Hispanics were used as the reference groups because the perviously mentioned IOM report1 identified and summarized studies where medical treatment had disfavored women and racial and ethnic minorities, and our aim was to investigate this occurrence in medical care involving medication. Accordingly, for purposes of this literature review, studies in which drug treatment was found to be uniform among groups or favored women or racial and ethnic minorities were classified as not demonstrating a disparity. Based on the standards for classification of racial and ethnic data promulgated by the U.S. Office of Management and Budget,316 we classified studies that examined race as those that used the terms African American or Black, Native American or American Indian, white, Asian, or Pacific Islander to describe the racial demographics of their study population. Similarly, we classified those studies that examined ethnicity as those studies that used the terms Hispanic, Mexican American, Latino, and Puerto Rican to describe ethnic demographics. The a priori level of significance was 0.05.
Although review articles were excluded from our analysis, their references were reviewed to ensure concurrence and completeness. Included articles were analyzed to identify
The types of clinical conditions in which pharmacotherapeutic disparities were examined and the frequency in which those articles reported statistically significant disparities in medication treatment (the type of clinical content area was determined by the disease or condition treated by the drug investigated in the study),
The frequency that minority groups were subjects in studies investigating pharmacotherapeutic disparities,
The frequency that each of the following types of pharmacotherapeutic disparities were observed: differences in the receipt or nonreceipt of drugs, differences in the drugs selected, differing dosages or intensity of treatment or method of drug delivery, differences in wait time to drug administration, and other differences in drug treatment related to (1) the frequency in which medications were changed at the next consecutive prescription, (2) the use of duplicative medications for an identified condition, and (3) the upward adjustment of medication dosage, and
The outcome of the observed pharmacotherapeutic disparity on measurable health-related or therapeutic variables (e.g., survival rates, quality of life, target goal achievement).
The literature search, which included additional references obtained from excluded studies, ultimately identified 311 articles that investigated whether race, ethnicity, or sex was associated with disparities in medication treatment. These articles were then analyzed based on the four aims described above.
Clinical conditions and frequency of pharmacotherapeutic disparities
Included articles were sorted and classified based on the primary medical condition or clinical content area studied as determined by the medical condition treated by the drugs addressed. Table 1 summarizes features of articles in the nine clinical condition categories identified (asthma, cardiovascular disease prevention and treatment, diabetes mellitus, HIV infection, mental health, oncology, osteoporosis, pain control or palliative care, and Parkinson’s disease).
Seventy-seven percent (240 out of 311) of included studies demonstrated significant pharmacotherapeutic disparities. Those articles with clinical content areas with a smaller sample (e.g., osteoporosis [4 articles], Parkinson’s disease [1 article]) all reported a significant pharmacotherapeutic disparity. Clinical content areas with larger sample sizes (e.g., HIV treatment [33 articles], mental health treatment [90 articles], oncology [23 articles], pain control or palliative care [51 articles]) had a substantial number of articles demonstrating a pharmacotherapeutic disparity.
Racial or ethnic minority groups and women as subjects
Table 1 shows the frequency in which racial, ethnic, or sex disparities in medication treatment were examined for each clinical content area; because race and ethnicity were considered separately, some studies examined both simultaneously. Across clinical content areas, potential racial differences were more frequently examined, representing 88% of all studies, than were ethnic or sex differences. Almost all studies included in this review examined racial differences in medication treatment across clinical content areas. Ethnic differences in drug treatment were most frequently assessed in mental health and pain control or palliative care and least frequently assessed in oncology. Sex differences in drug treatment were more frequently examined in HIV and pain control or palliative care treatment than in other clinical content areas.
Types of pharmacotherapeutic disparities
Racial, ethnic, or sex differences in the receipt and nonreceipt of drugs were the most frequently documented types of disparity, representing 73% of the total number of articles documenting a pharmacotherapeutic disparity. Twenty-three percent of the total number of articles documenting disparities addressed differences in the drugs prescribed, with even fewer (10%) addressing differences in drug dosing or administration.
Types of outcomes examined
Approximately 15% of the articles documenting a pharmacotherapeutic disparity (36 of 240) addressed or identified health-related outcomes associated with the disparity (Table 1). Health-related outcomes for selected clinical content areas are described below.
Four articles examining pharmacotherapeutic disparities in asthma treatment found a health-related outcome associated with the disparity; all four revealed an increase in hospitalizations associated with disparate drug treatment.18,19,25,38 Two studies compared asthma prevalence and treatment for Alaska Native and nonnative Medicaid recipients younger than 20 years.18,19 Gessner’s18 2003 study demonstrated that the Alaska Native race was significantly associated with increased odds of asthma-related hospitalization (relative risk [RR], 1.6; 95% confidence interval [CI], 1.2–2.3) and that Alaska Natives younger than 20 years who had been hospitalized for asthma were significantly less likely than their nonnative counterparts to have received a long-term asthma-control medication including inhaled corti-costeroids (RR, 0.54; 95% CI, 0.33–0.88). In a 2005 follow-up study, Gessner and Neeno19 demonstrated that the yearly hospitalization risk for Alaska Natives younger than 20 years decreased (p = 0.02) concurrently with an increase in yearly use of inhaled corticosteroids (p < 0.001).
Joseph et al.25 and Zoratti et al.38 compared emergency department utilization for asthma care by race for a managed care program’s patients. African-American children who were seen by a nonasthma specialist provider were significantly less likely to have a prescription filled for an antiinflammatory medication during a postindex visit observation period (RR, 0.50; p = 0.02).25 Moreover, during the subsequent 12-month observation period, African- American children were more likely to visit the emergency department (RR, 1.8; p < 0.01) and to be hospitalized for asthma compared with white children (RR, 10.2; p < 0.01). Similarly, African-American patients with asthma age 15–45 years filled fewer prescriptions for inhaled or oral corticosteroids (p = 0.038), were more likely to visit the emergency department for asthma-related complaints (p < 0.001), and were more likely to be hospitalized for asthma complaints (p = 0.002).38 This discrepancy persisted even after the investigators adjusted for socioeconomic status.
Cardiovascular disease prevention and treatment: Dyslipidemia
Clark et al.,46 Goff et al.,57 Massing et al.,75 and Persell et al.83 investigated racial, ethnic, and sex differences in the detection, treatment, and control of dyslipidemia and in lipid goal attainment. Massing et al.75 compared African-American and white patients with coronary artery disease and found that lipid-lowering drugs were prescribed for fewer than half of African-American men and women compared with 60% and 55% of white men and women, respectively (African-American men: odds ratio [OR], 0.59; 95% CI, 0.45–0.78; African-American women: OR, 0.62; 95% CI, 0.46–0.83). Less than 20% of the African-American patients (19% of women and 18% of men), including those for whom lipid-lowering drugs were prescribed and those for whom such drugs were not prescribed, achieved the recommended low-density-lipoprotein (LDL)-cholesterol goal, defined as an LDL-cholesterol concentration of <100 mg/dL (women: OR, 0.55; 95% CI, 0.36–0.82; men: OR, 0.47; 95% CI, 0.30–0.74). Among white patients, 31% of men and 25% of women achieved the LDL-cholesterol goal. In another study, during the initial index visit, Persell et al.83 discovered that women with atherosclerosis were significantly less likely to be taking a cholesterol-lowering medication compared with their male counterparts (p = 0.04). Further, over the one-year follow-up period, cholesterol management (defined as medication intensifica-tion or LDL-cholesterol monitoring) occurred significantly less frequently for women (p = 0.01), and women were more likely to have high cholesterol (defined as an LDL-cholesterol concentration of >130 mg/dL) after one year (p = 0.003). The observed difference in cholesterol control between men and women was not significant after adjusting for the differences in cholesterol management during the year.
In two separate multicenter cohort studies, African-American and Hispanic patients were less likely to report lipid-lowering drug treatment than were white patients.46,57 Goff et al.57 reported that African-American and Hispanic patients with dyslipidemia were less likely than white patients to have their dyslipidemia controlled (African Americans: adjusted OR, 0.72; 95% CI, 0.60–0.86; Hispanics: adjusted OR, 0.74; 95% CI, 0.59–0.94) and were less likely to use lipid-lowering drugs, including high-efficacy statins (African Americans: adjusted OR, 0.86; 95% CI, 0.73–0.99; Hispanics: adjusted OR, 0.8; 95% CI, 0.66–0.98). Similarly, Clark et al.46 reported that African-American patients with dyslipidemia were less likely to be using lipid-lowering drugs (p = 0.03) or high-efficacy statin therapy (p = 0.007) compared with white patients. Moreover, African-American patients were less likely to achieve LDL-cholesterol treatment goals (p < 0.001) than were their white counterparts.
Cardiovascular disease prevention and treatment: Hypertension and myocardial infarction
Two studies examined racial, ethnic, and sex differences in the treatment of hypertension and found lower rates of blood pressure control as a result of pharmacotherapeutic disparities.68,73 Kramer et al.68 found that among African-American patients with hypertension, the rate of β-blocker use was significantly lower (p < 0.0001) and the rate of calcium-channel blocker or diuretic use was significantly higher than in white patients (p < 0.0001). In addition, these African-American patients had a greater likelihood of having treated but uncontrolled hypertension (adjusted OR, 1.49; 95% CI, 1.19–1.84), defined as systolic blood pressure of ≥140 mm Hg or diastolic blood pressure of ≥90 mm Hg. Similarly, a study comparing racial disparities across a national sample of women with hypertension found that Hispanic women had the lowest rates of treatment with antihypertensive medication (p < 0.01) and the lowest rates of hypertension control (p < 0.001), defined as systolic blood pressure of ≥140 mm Hg or diastolic blood pressure of ≥90 mm Hg.73
Vaccarino et al.97 investigated sex and racial differences in the treatment of patients hospitalized with acute myocardial infarction and the inhospital mortality rates for these patients. Although all women were less likely to receive aspirin (p < 0.001) and β-blockers (p < 0.001), African-American women were the least likely of any group to receive these medications and had the highest rates of inhospital mortality (p < 0.001; OR, 1.11).
Wendel et al.113 examined daily insulin doses and resulting glycemic control across races and ethnicities among veterans and found that white patients received significantly more daily units of insulin than both African-American and Hispanic patients (p < 0.01) and that African-American and Hispanic patients had poorer glycemic control, exhibited by a higher mean glycosylated hemoglobin value (p = 0.05). Similarly, Okosun and Dever110 found that the rates of diabetic treatment among diagnosed patients were higher in white than in African-American or Hispanic patients (p < 0.05), and Hispanic or African-American patients were less likely to achieve glycemic control (p < 0.05).
Two studies examined zidovudine treatment for HIV infection across races and ethnicities and its relationship to survival.131,137 Both found that African-American patients with HIV infection were less likely to receive zidovudine treatment than their white or Hispanic counterparts. One study found that zidovudine as well as Pneumocystis jirovicii pneumonia prophylaxis independently predicted improved mortality rates; however, African-American patients were less likely to receive such therapy than white patients with HIV infection (RR, 0.76; p < 0.05).131 The RR of death associated with zidovudine treatment was 36% less than that in untreated patients. The second study addressed the prevalence of zidovudine treatment among childbearing women with HIV.137 Blood specimens collected from infants were anonymously tested for HIV antibody; HIV-positive specimens were tested for the presence of zidovudine. Of the HIV-positive specimens, 37% were from African-American infants and 35% were from Hispanic infants; moreover, 24% of African-American women and 29% of Hispanic women were not treated with zidovudine. Pediatric AIDS cases diagnosed within the ensuing three years were disproportionately comprised of African-American and Hispanic children, representing 80% of all children diagnosed with AIDS.
Two articles addressed differences in the survival rate of oncology patients based on racial or ethnic pharmacotherapeutic disparities in chemotherapy.257,258 African-American race was associated with a delay in initiation of chemotherapy for patients with ovarian cancer, and this delay in treatment was associated with a 13% increase in overall mortality compared with that for white patients.257 Moreover, white women with localized breast cancer were more likely to receive hormonal therapy and multiagent chemotherapy than were African-American women (p < 0.0001).258 After controlling for age, the risk of all-cause mortality was 2.35 times higher for African-American women compared with their white counterparts (95% CI, 1.75–3.16). The risk of death for African-American women dropped to 1.83 (95% CI, 1.38–2.53) when treatment included surgery, radiation, hormone therapy, and chemotherapy.258
Pharmacotherapy is the predominant treatment modality used during outpatient medical care visits.317 In 2006, 7 of every 10 ambulatory care visits resulted in at least one medication provided, prescribed, or continued.318 This, coupled with the significant increases in medications prescribed in all health care settings (including physician offices, hospital outpatient departments, and hospital emergency departments),317 further highlights the need to address this gap in care, as medication disparities will undoubtedly grow without adequate attention and resources to reduce and eliminate them. This literature review is the first of its kind, demonstrating in one article the existence of significant pharmacotherapeutic disparities across clinical conditions and underscoring the need for greater attention and resources to address pharmacotherapeutic disparities. It takes an important step toward reducing or eliminating pharmacotherapeutic disparities, naming and characterizing a previously underreported and undefined health care disparity, and identifying the health outcomes associated with such disparities. Although racial and ethnic differences in health care utilization rates and access to medical care are well documented, 319 the reasons for these disparities are varied, complex, and, in general, still poorly understood even after two decades of data collection. In fact, according to Mayberry et al.,320 our understanding of health care disparities by race and ethnicity has advanced very little since the 1985 release of the DHHS Task Force Report on Black and Minority Health.
Although we do not know why access and utilization disparities exist, it is clear from the entire body of literature that disparities in health care and health status are not adequately explained by differences in insurance, income, or other measures of socioeconomic status, comorbidities, severity of disease at diagnosis, availability of services, or patient preferences.319,320 Naturally, the same factors that affect health care access and utilization rates are clearly at play in producing or at least contributing to observed pharmacotherapeutic disparities.
Specific pharmacy factors may be partly responsible for certain pharmacotherapeutic disparities. Two cross-sectional surveys of pharmacies conducted in New York City and Michigan addressed the availability of opioids and found that pharmacies in areas with fewer white residents were less likely to be adequately stocked with opioid analgesics.321,322 In contrast, Mayer et al.323 found no rural–urban or socioeconomic disparity in the opioid analgesic inventory of Washington state pharmacies. Accordingly, more research must be conducted to determine the relationship between health care system factors and pharmacotherapeutic disparities.
Provider factors may also contribute to pharmacotherapeutic disparities. Health care providers, like all human beings, are influenced by social context and cues. A sizable body of literature focuses on how physicians and other health care providers process information during patient encounters, including how prejudice, stereotyping, and uncertainty can affect assessments of patients and decisions about their treatment.324 Provider variations have been documented in assorted aspects of clinical decisions, including diagnosis, ordering tests, asking follow-up questions, giving lifestyle advice, and making referrals to specialists. Drug treatment, selecting drugs, and prescribing drugs are likewise subject to similar variations in treatment, as evidenced by this review. In the clinical content area of pain treatment, the impact of race and ethnicity has been addressed in the context of objective-versus-subjective pain complaints. Quazi et al.299 and Tamayo-Sarver et al.309 compared racial and ethnic pharmaceutical treatment rates for subjective-versus-objective diagnoses. Both studies determined that disparities in analgesic treatment were less pronounced in patient complaints that were objective (e.g., long-bone fractures) versus subjective (e.g., headaches). Accordingly, variability in provider decision-making could lead or contribute to pharmacotherapeutic disparities. Additional research into factors that systematically influence provider decision-making as it relates to pharmacotherapies is necessary. Also, additional research is necessary regarding the effect of provider race or ethnicity on variations in pharmaceutical treatment. Evidence is available demonstrating that pharmacotherapeutic disparities are reduced when racial minorities are treated by a provider of the same race.129,172
Patient-related variables may also contribute to pharmacotherapeutic disparities; such variables include (1) identification as a member of a racial, ethnic, or sex minority, (2) differing attitudes across race, ethnicity, or sex, (3) language differences, and (4) cultural differences. For example, patients of differing racial or ethnic backgrounds may communicate with their health care providers differently and may physically or verbally express differing cultural philosophies when seeking care. As explained by Juarez et al.325 and Anderson et al.,326 stoicism is an expected and accepted response to pain for Hispanic and African-American patients. The need to hide pain or endure pain is related to the acceptance of pain as part of life, and enduring pain is seen as a reflection of personal strength and pride, especially for male patients.325,326 This cultural difference may affect the way these patients communicate with their providers and may ultimately result in differing treatment with pain medications. Cultural differences are also manifested in communication barriers that may exist for racial or ethnic minority patients who are not fluent in written or spoken English. Low proficiency in English is associated with lower rates of preventive services and delayed use of physician services for needed care.327,328 Use of medical interpreters is currently inadequate,329 and failure to use interpreters may be particularly problematic in discussions of sensitive topics such as end-of-life care, mental health care, and treatment of HIV infection.
Finally, some study limitations may also contribute to misunderstanding pharmacotherapeutic and other disparities. Many included studies, regardless of reporting a significant pharmacotherapeutic disparity, did not adequately or consistently define racial or ethnic groups. Some studies used the terms race and ethnicity interchangeably, and others failed to explicitly define the term minority group when used. A problem with using the terms race and ethnicity interchangeably is that numerous distinct ethnicities can exist within a single race category.330 Moreover, several included articles failed to use standardized categories to capture race and ethnicity. Some articles included the term Hispanic as a racial category when it is considered an ethnic category in the standards for classification of federal data on race and ethnicity promulgated by the U.S. Office of Management and Budget.316,331 Consequently, additional research must accurately and consistently define racial and ethnic categories when investigating pharmacotherapeutic and other health disparities.
A literature review revealed significant disparities in the medication treatment received by racial and ethnic minorities and women.
References for this article are available at www.ajhp.org.
An audio interview with, which supplements the information in this article, is available on AJHP’s website at www.ajhp.org/site/misc/podcasts.xhtml.
Jennifer Martin and Sandy Kramer are acknowledged for their assistance with the literature review.
The authors have declared no potential conflicts of interest.
- Copyright © 2010 by the American Society of Health-System Pharmacists, Inc. All rights reserved.