Disparities in Influenza Treatment Among Disabled Medicaid Patients in Georgia

Abstract

Objectives. We explored possible disparities in seasonal influenza treatment in Georgia's disabled Medicaid population. We sought to determine whether racial/ethnic, geographic, or gender disparities existed in antiviral drugs usage in the treatment of influenza.

Methods. Medicaid claims were analyzed from 69 556 clients with disabilities enrolled in a Georgia Medicaid disease management program.

Results. There were 519 patients who met inclusion criteria (i.e., adults aged 18–64 years with an influenza diagnosis on a 2006 or 2007 Medicaid claim). Roughly one third (36.2%) of patients were classified as African American, 44.5% as White, and 19.3% as “other.” Most patients had 2 or more comorbid chronic diseases. Antivirals were used in only 14.5% of patients diagnosed with influenza. Treatment rates were nearly 3 times higher for White patients (19.5%) than for African American patients (6.9%).

Conclusions. Our analysis suggests limited use of antiviral treatment of influenza overall, as well as significant racial disparities in treatment. Additional studies are needed to further explore this finding and its implications for care of racial/ethnic minority populations during seasonal influenza and for effective pandemic influenza planning for racial/ethnic minority populations.

Influenza is a highly contagious viral disease of the respiratory system. It can lead to serious complications and can result in death.¹ According to the US Centers for Disease Control and Prevention (CDC), 5% to 20% of the population of the United States acquires influenza each year.² Additionally, 200 000 are hospitalized and 36 000 die as a result of the complications from influenza.² Populations vulnerable to complications include children, the elderly, racial and ethnic minority populations, persons with disabilities, and persons with chronic diseases.^3–5

Influenza can be prevented with annual immunizations, and it can be effectively treated with disease-modifying antiviral drugs if diagnosed within 48 hours of initial symptoms.^1,6,7 Although prevention is the best option, many individuals do not receive annual immunizations. Rates of influenza immunization vary across racial and ethnic populations. Data from the National Immunization Survey (NIS)–Adult 2007 (for the 2006–2007 season) show influenza immunization coverage for those aged 18 to 49 years among 40.9% of Whites, compared with 30.9% of African Americans and 25.6% of Hispanics.⁸ National Health Interview Survey (NHIS) Early Release data from 2008 indicate that among those 65 years and older, only 49.3% of African Americans and 55.4% of Hispanics had received influenza immunizations, compared with 71% of Whites.⁹

Factors contributing to lower rates of immunizations among different races and ethnicities include limited access to health care, personal beliefs, and anticipated negative side effects.¹⁰ One study examined the disparity between African Americans and Whites, in which African Americans were less likely to receive an annual influenza immunization.¹¹ Reasons for the disparity were in part attributed to low perceptions of susceptibility to influenza on the part of American Americans and the subsequent view that an influenza immunization is not needed. Additional reasons include feared side effects of the immunization, including the possible acquisition of influenza as a result of the vaccination. Those African Americans who were more receptive to receiving an annual influenza immunization were of a higher income and had a greater perception of susceptibility to influenza.¹¹

The widespread existence of racial and ethnic health disparities has been well documented and remains a national public health concern.^12–14 Health disparities are not limited to persons without access to health care; they have been documented in persons with access to both Medicare and Medicaid benefits.^15,16 Disparities in infectious diseases have been widely documented and persist both domestically and globally.^17,18 Disparities in influenza immunization among racial and ethnic populations raise questions about whether disparities might also exist in influenza treatment.

Treatment of influenza within the first 48 hours reduces both the recovery time and the severity of the illness.⁶ In an influenza pandemic, it is anticipated that the development of an effective vaccine to counter the pandemic strain of influenza will take 6 to 8 months.¹⁹ As a result, the treatment of influenza with the antiviral drugs oseltamivir and zanamivir will be a first line of defense for patients with influenza.^1,7 However, the rates of prescribing antivirals for treatment of influenza in the United States are not known. One national study found that physicians prescribed antivirals to 19% of patients diagnosed with influenza.²⁰ Linder et al. noted that “it is impossible to be certain whether this represents underprescribing, overprescribing, or appropriate prescribing.”^20(p534) Examining patients' receptivity to antiviral medications, one study of patient knowledge and attitudes about antiviral medication found patients to be ill informed about antiviral medication and its benefits, with medication cost a potential barrier to treatment.²¹

With the threat of an influenza pandemic occurring at any time, it is a public health priority to assess potential treatment disparities, especially in those most susceptible to contracting influenza.^1,22 Examining the treatment of influenza in the Medicaid population, specifically persons with disabilities, provides a vehicle to assess such potential disparities. Studying Medicaid claims data is useful because patients under Medicaid all have the same health insurance, as well as the same prescription coverage for medication.²³ Although health disparities in the receipt of influenza immunizations have been noted, very little research has been conducted to explore disparities in the use of pharmacological treatments for influenza. We undertook to examine possible racial/ethnic, geographic, and gender disparities in the treatment of influenza among Medicaid patients.

METHODS

The study was a project of the National Center for Primary Care at Morehouse School of Medicine. The study used Medicaid claims data for non-Medicare (nondual; i.e., not receiving both Medicaid and Medicare) aged, blind, or disabled clients enrolled in the Medicaid disease management program known as Georgia Enhanced Care (GEC) for the calendar years 2006 and 2007, which would include most influenza activity for the 2005–2006 and 2006–2007 influenza seasons. The National Center for Primary Care at Morehouse School of Medicine works with this disease management program to identify opportunities to improve care and outcomes for this vulnerable segment of the Medicaid population. For these and similar analyses, personal identifiers are stripped from the data and stored in a secure and confidential manner.

These Medicaid data include both personal demographic information as well as records of each individual claim submitted to the state Medicaid program. The nondual (non-Medicare) segment of the adult aged, blind, or disabled population is automatically enrolled in disease management programs unless the client specifically opts out. Race/ethnicity is self-reported in 5 categories (non-Hispanic White, non-Hispanic Black, Asian/Pacific Islander, American Indian/Alaska Native, and Hispanic). At the time of the study, 2006 to 2007 data files were the most recent data available. The sample did not include those covered by Medicare or those who were in a nursing home or hospice care.

The study sample was drawn from the Georgia Enhanced Care program, the only disease management program serving this non-Medicare, noninstitutionalized, aged, blind, or disabled segment of the Georgia Medicaid population residing in the northern third of Georgia. It is important to note that these patients have high rates of comorbid conditions and are therefore highly vulnerable to the complications of influenza. For example, according to Valdez et al., individuals with diabetes are more likely to develop complications from influenza and are 2 to 3 times more likely to die from influenza compared with those without diabetes.²⁴ In an epidemic setting, this disparity is amplified.²⁴ For people with disabilities, a number of factors could increase risks of exposure, complications, and death in an influenza pandemic.⁴

The study sample was limited to GEC individuals aged 18 to 64 years who had any diagnosis of influenza using International Classification of Diseases, Ninth Revision, Clinical Modification, code 487.²⁵ (Other studies have deemed the ICD-9 code–based diagnosis of influenza to be valid, demonstrating sufficient sensitivity and predictive value.²⁰) Only those patients who were diagnosed in the emergency department, a doctor's office, or in a health care clinic were included. National Drug Codes were used to identify the main outcome variable, which was the prescription of a disease-modifying antiviral drug.²⁶ The drugs of interest included both the old-generation drugs rimantidine and amantadine and the new-generation drugs oseltimivir and zanamavir. (Of note, beginning with the 2005–2006 influenza season, the CDC recommended not using rimantidine and amantadine, because of the high levels of resistance to these drugs among circulating influenza viruses.²⁷) In addition to use with an influenza diagnosis, antivirals are also indicated for postexposure prophylaxis among uninfected persons; the study was not designed to include persons receiving antivirals for this purpose.

It should be noted that the Georgia Medicaid program did not cover influenza vaccinations for a substantial part of the study period. We were therefore unable to determine whether those who were diagnosed with influenza had received an influenza immunization.

Bivariate statistical analyses using the χ² test were conducted to calculate percentages and 95% confidence intervals (CIs) to assess the relationship between receiving a disease-modifying antiviral drug, and gender, race and geography. Race was entered as a categorical variable, whereas gender and geography were entered as dichotomous variables. Geography was defined using the 2003 Metropolitan Statistical Areas (MSAs) for Georgia.²⁸ Urban areas were defined as MSA, whereas rural areas were defined as non-MSAs. Percentages were calculated to assess the relationship between race/ethnicity and geography. Finally, multivariate adjusted odds ratios for receiving disease-modifying antiviral drugs were calculated. Age, gender, race/ethnicity and the rural–urban variable were all included in the model simultaneously. All analyses were conducted using SPSS version 15.0 (SPSS Inc, Chicago, IL) and SAS version 9.2 (SAS Institute, Inc, Cary, NC).

RESULTS

For calendar years 2006 and 2007, claims data were filed for 69 556 patients. Of these patients, 51 202 were aged between 18 and 64 years. From these patients, we focused on those diagnosed with influenza; for a total of 542 patients (345 patients in 2006 and 197 patients in 2007). In total, 519 unique patients met the inclusion criteria; some patients were diagnosed with influenza both years. Roughly one third (36.2%) of patients were classified as African American (n = 188), 44.5% as White (n = 231), and 19.3% as “other” (n = 100). Unfortunately, the number of patients in groups other than non-Hispanic White and non-Hispanic Black was too small for statistical analysis, which meant we were unable to analyze the data by ethnicity.

Of those with a diagnosis of influenza, only 75 (14.45%) received a disease-modifying antiviral drug. Relationships between the treatment with a disease-modifying drug and both race/ethnicity and geography were statistically significant (P = .001 and P = .016, respectively). Whereas 19.48% of White patients received treatment, only 6.91% of their African American counterparts received treatment. Additionally, whereas 20% of patients living in rural areas received treatment, only 12% of those living in urban areas received treatment. Neither age nor gender were associated with treatment with a disease-modifying drug (P = .756 and P = .698, respectively; Table 1).

TABLE 1.

Characteristics of Patients Diagnosed With Influenza and Receiving Antiviral Drugs: 2006–2007

	Diagnosed With Influenza, No. (%)	Diagnosed With Influenza And Received Antiviral Drugs, No. (%)
Total	519 (100)	75 (100)
Race/Ethnicity
Black	188 (36.2)	13 (6.9)
White	231 (44.5)	45 (19.5)
Other	100 (19.3)	17 (17.0)
Geography
Rural	160 (30.8)	32 (20.0)
Urban	359 (69.2)	43* (12.0)
Age, y
18–33	134 (25.8)	23 (17.2)
34–45	125 (24.1)	16 (12.80)
46–55	134 (25.8)	18 (13.4)
56–64	126 (24.3)	18** (14.3)
Gender
Men	163 (31.4)	25 (15.3)
Women	356 (68.6)	50 (14.0)

^*P < .05; **P < .01.

African Americans accounted for about one half of the urban population (46.24%), and Whites accounted for more than one half of the rural population (72.40%). After adjusting for age, gender, race/ethnicity and geography, we found that race/ethnicity continued to be a significant factor in the receipt of disease-modifying antiviral drugs (Table 2). Compared with African Americans, Whites were 2.90 times (95% CI = 1.45, 5.80) more likely to receive disease-modifying antiviral drugs. Other racial/ethnic groups were 2.82 times (95% CI = 1.29, 6.17) more likely to receive disease-modifying antiviral drugs compared with African Americans.

TABLE 2.

Multivariate Adjusted Odds Ratios (AORs) of Receiving Antiviral Drugs: 2006–2007

	AOR (95% CI)
Race/Ethnicity
Black (Ref)	1.00
White	2.90** (1.45, 5.80)
Other	2.82** (1.29, 6.17)
Geography
Rural (Ref)	1.00
Urban	1.42 (0.82, 2.45)
Age, y
18–33 (Ref)	1.00
34–45	0.67 (0.33, 1.35)
46–55	0.75 (0.39, 1.50)
56–64	0.68 (0.34, 1.36)
Gender
Men (Ref)	1.00
Women	0.92 (0.54, 1.57)

Note. CI = confidence interval. All variables were included in the model simultaneously.

^**P < .01

DISCUSSION

Our results suggest 2 important findings. First, the prescribing of influenza antiviral medications was limited overall. Over 85% of patients with influenza had no prescription claim for a disease-modifying antiviral drug. Reasons for this remain unclear. The national study mentioned earlier revealed a slightly higher rate (19%) of patients diagnosed with influenza who were prescribed antiviral medication; that study also suggests rates were lower among “non-Whites” than Whites.^20(p533)

Possible reasons for underutilization in general include poor drug supply, lack of provider knowledge about the availability and use of influenza antiviral medications, provider practices, medication costs, delay in seeking care by patients, and limited access to health care services for acute illnesses. As mentioned earlier, influenza must be treated within the first 48 hours of experiencing symptoms. Although the first 48 hours are a critical time, many patients do not visit a doctor until later, if at all. One important reason for this is that many US residents do not have a primary source of health care. Instead, these individuals only have access to the emergency department, which often results in visits that may take hours before being seen.^7,12 Alternatively, many patients diagnose themselves and self-medicate until they are feeling better. For the patients in our study, there may exist a racial disparity in the propensity to seek early care after illness onset, which could be due to a suspicion of the health care system, lack of knowledge of treatment availability, or other factors such as transportation.^13,29,30

Racial Disparity

Second, the data reveal a nearly 3-fold racial disparity in the treatment of influenza with disease-modifying antiviral drugs among disabled Medicaid patients. The authors deem this a disparity as defined by Healthy People 2010: “In the context of public health, a disparity is the quantity that separates a group from a reference point on a particular measure of health that is expressed in terms of a rate, proportion, mean, or some other quantitative measure.”^31(p7)

Reasons for this disparity are unclear and merit further study to determine possible contributing factors. Blumenshine et al. suggest that minority populations, including African Americans and Hispanics, are more likely to be exposed to influenza, are more susceptible to contracting influenza, and have less access to quality health care.⁷ Increased exposure may occur as a result of household crowding, the use of public transportation, and the use of emergency departments and public clinics for routine health care (emergency departments and public clinics have been documented as sites of airborne transmission of infection).^7,12 Being socially disadvantaged and having a low income is associated with contracting infectious diseases; these conditions may more often be endemic in African American and Hispanic populations, compared with the White population.^17,32 Quality health care for minority populations can be limited by language and cultural barriers as well as the lack of a usual source of health care.^7,33-35 Finally, not receiving regular annual influenza immunizations leaves these groups without the benefit of the single best measure for preventing influenza.^7–9

The results of this study cannot completely be interpreted without examining the economic context of the patients' lives. It is important to note that the eligibility threshold for the program is set by Georgia Medicaid. For the aged, blind and disabled sector, the individual income limit is $317 per month ($3804 per year) and the limit for a couple is $375 per month ($4500 per year). These Georgia Medicaid income-eligibility requirements apply to both African Americans and Whites, as well as all other Georgia residents. The antiviral medications were lost-cost: Georgia Medicaid was charging 50 cents per prescription for generic drugs and the same (50 cents) for brand-name drugs on the preferred drug formulary, and no more than $3.00 for nonpreferred brand-name drugs.³⁶ Given these circumstances, we conclude that the disparities observed in the receipt of antivirals for influenza treatment cannot be attributed to differences in economic status.

The disparity we found raises important questions for both seasonal influenza management and effective pandemic influenza planning. Given the heavy toll of seasonal influenza on the nation's health,^1–3 ensuring equitable treatment is of paramount importance. If disparities in influenza treatment can be present during the relative calm of seasonal influenza, heightened problems may reasonably be expected to occur in a pandemic. Pandemic planning needs to take this into account and anticipate possible challenges in providing antiviral treatment in an equitable fashion. Those responsible for planning the response to an influenza pandemic cannot assume equal access to treatment modalities among all populations and must therefore plan accordingly.

Limitations

Our results should be interpreted in the context of the study's limitations. First, we did not have access to medical records or to direct personal interviews. Diagnosis of influenza is often a presumptive diagnosis not confirmed by viral culture at the clinical level. This diagnosis may or may not be used on the billing claim, and so episodes of influenza coded as fever or nonspecific viral illness or upper respiratory infection would not be captured in these data. Similarly, if patients received drug samples for antiviral treatment, this would not be captured in the pharmacy claims data. These limitations would usually apply to persons of all racial groups who had the same Medicaid coverage, drug formulary, etc. On the other hand, because Medicaid claims are closely audited, the drug data that are included are considered to among the highest-quality data available on drug exposure.³⁷ We also did not have sufficient data to assess use of preventive measures—i.e., influenza vaccination. Although the receipt of an annual influenza immunization may decrease an individual's chances of acquiring influenza, the study population did not have access to Medicaid-covered influenza immunizations during the entire study period, as noted above.

A second limitation is the lack of data on the percentage of those diagnosed with influenza who presented within 48 hours of symptoms. As noted earlier, a late diagnosis of influenza precludes effective treatment with antivirals. It is not possible to know how the disparity finding might be altered if the analysis were able to account for differences by race/ethnicity in the proportions of those with an influenza-coded medical visit who present within 48 hours of onset of symptoms.

A third limitation is that although the GEC Medicaid claims data include information for non-Hispanic White, non-Hispanic Black, Asian/Pacific Islander, American Indian/Alaska Native, and Hispanic racial and ethnic populations, the number of patients in groups other than non-Hispanic White and non-Hispanic Black was too small for statistical analysis, and analyzing the data as “other” is not helpful. In addition, the data lack information about specific racial and ethnic subpopulations, such as Mexicans or Puerto Ricans within the Hispanic category, a problem that has been noted to contribute to knowledge gaps about health disparities.³⁸ Even with larger numbers, grouping all subpopulations within a single racial/ethnic category does not allow for the assessment of any potential disparities within subgroups.

In summary, this study suggests racial disparities in the use of antiviral drugs in the treatment of influenza among disabled Medicaid patients in Georgia. The etiology and degree of the disparities found remain unclear and merit further investigation. These findings suggest that in both seasonal influenza management and effective influenza pandemic planning there is a need to pay careful attention to access to antiviral drugs and to their use in minority populations.