Influenza and Pneumococcal Vaccination Among Native American Elders in a Primary Care Practice FREE

MORE THAN 2 million Native Americans (hereafter referred to American Indians [AIs]) and Alaska Natives (ANs) live in the United States.¹ The number of urban dwellers has increased dramatically since World War II; only 25% of AIs/ANs reside on reservations and trust lands and 60% reside in cities.¹ Although vast strides have been made in recent years, AIs/ANs still lag behind the general population for major health status indicators.² Infectious diseases such as tuberculosis and gastroenteritis have been replaced as the leading causes of death by diabetes mellitus, cardiovascular disease, and hypertension.³ It is not surprising, then, that AI/AN life expectancy is 8 years less than nonnative peoples and is accompanied by higher age-specific mortality rates than the population at large.^4- 5 Yet, despite the emergence of chronic health conditions and increased mortality rates, virtually no information exists on the implementation of preventive measures among AIs/ANs.

American Indians/Alaskan Natives are especially susceptible to respiratory tract diseases, including pneumococcal infections.^6- 7 According to Indian Health Service statistics, influenza and pneumonia constitute the sixth leading cause of death among AIs of all ages and the fourth leading cause of death among American Native elders.⁸ Moreover, the mortality rate increases markedly with age: for persons 65 years or older, mortality due to influenza and pneumonia is greater for AIs/ANs than the general US population.^8- 9 However, although data specific to AIs/ANs are unavailable, national sources indicate that minority populations receive influenza and pneumococcal vaccines at substantially lower rates than white persons.⁹ Since minority adults use preventive services less, are initially seen for care at more advanced stages of disease, and have greater morbidity and mortality than higher-income individuals,^10- 11 AI/AN elders may be at great risk for not obtaining recommended interventions such as influenza and pneumococcal vaccinations.

In this retrospective study, we reviewed the medical records of 550 AIs/ANs who were at least 50 years of age seen in a primary care practice during a 1-year period. Our objectives were to evaluate how frequently influenza and pneumococcal vaccinations were given according to 2 age thresholds as well as published guidelines applicable to persons with chronic medical conditions. A secondary goal was to identify factors associated with nonadherence to immunization recommendations. In contrast with studies relying strictly on self-report, review of medical records allowed us to document compliance with vaccination guidelines.

According to the 1990 census, almost 17,800 AIs/ANs live in King County, Washington, with 7325 residing in Seattle. The Seattle Indian Health Board (SIHB), a multidisciplinary, community-based organization, is the major source of health care for the urban King County Native American population. It is considered one of the most comprehensive, off-reservation, primary care programs for AIs/ANs in the nation, providing diverse medical, mental health, dental, preventive, and educational services. The SIHB serves more than 6000 individuals who make more than 40,000 visits per year. Over half of the clinic clientele are at least 45 years old, almost half are unemployed, 58% have no form of health insurance, and 80% have incomes under the federally determined "poverty line." Owing to its urban location, SIHB patients are affiliated with approximately 250 tribes.

Because of increased morbidity and mortality, impairments associated with aging may occur 20 years earlier among AIs/ANs than in the general population.¹² Thus, AIs/ANs often require, at a younger age, services typically deemed the province of "the elderly." In addition, owing to the shorter life expectancy and high birth rate of AIs/ANs, elders also constitute a much smaller percentage of the total Native American population compared with other ethnic or racial minorities. This situation has led to lower age eligibility criteria for federally and state funded programs. For example, Title VI of the Older Americans Act has recognized the need to provide services to AIs living on reservations who are younger than 60 years by allowing each tribe to set its own age eligibility requirements. Finally, among AIs/ANs, "elder" status is not solely a function of chronological age, and may be conferred much earlier than in the majority culture. For these reasons, then, data were examined for 2 different definitions of elder. In one instance, elder was defined as a person at least 50 years of age, consistent with the foregoing observations, which relate as well to broader national concerns regarding race and health disparities. In the other instance, elder was defined in the more conventional fashion as a person aged 65 years or older.

A list was generated from the computerized visit records of SIHB with the names of all AIs/ANs aged 50 years or older who had been seen on at least 1 occasion between June 1, 1994, and June 30, 1995. A total of 550 patients were identified and their medical records retrieved. Of the 550 medical records eligible for review, all (100%) were located and abstracted, representing the entire eligible population.

The medical records were reviewed by a single chart abstractor who was trained in data collection. To confirm the accuracy and reliability of the medical record abstraction, a sample of medical records was independently reviewed by a general internist. The abstracting process focused on determining the frequency of administration of influenza and pneumococcal immunizations, as well as demographic information including date of birth, sex, tribal affiliation, blood quantum series, educational level, employment status, marital status, number of household members, and type of insurance. Other data included whether patients smoked, used alcohol or other substances, and a listing all health problems (as documented in the medical record by a clinic provider, typically in notes or problem lists), and current medications.

Several sources were reviewed regarding the appropriate preventive guidelines for immunization in this population. Influenza and pneumococcal vaccinations are recommended by the US Preventive Services Task Force, the American College of Physicians, American Academy of Family Physicians, and the Centers for Disease Control and Prevention. Influenza vaccination is also recommended by the Canadian Task Force on the Periodic Health Examination, American Cancer Society, and the American College of Obstetrics and Gynecology (reviewed in reference 13). The US Preventive Services Task Force states that individuals aged 65 years or older should be vaccinated once against pneumococcus and annually against influenza.¹⁴ For persons at high risk, the task force also recommends annual influenza vaccination and a single administration of pneumococcal vaccine. High-risk populations should receive a second dose of pneumococcal vaccine 5 years after the first administration or when they reach the age of 65 years if 5 or more years have passed since the first dose. High-risk individuals are those who are alcoholics; who are in renal failure; or those who have cardiovascular, pulmonary, and immunocompromising diseases; diabetes mellitus; human immunodeficiency virus infection; and cancer.

Performance was examined for each measure using 2 age thresholds. The conventionally defined sample of older adults (those aged ≥65 years) was, thus, a subset of the more culturally appropriate one (those aged at least 50 years). We assessed adherence for influenza immunizations using 2 time frames: ever having received a vaccination and having received a vaccination in the year prior to the last primary care visit. For pneumococcal vaccination, we determined only if it had ever been administered. Group differences on discrete variables were examined with χ² analyses. Continuous variables were compared with t test. Univariate logistic models, controlled for age and sex, were also calculated for each potential predictor. Because the 2 univariate techniques yielded similar results, we only list the latter in Table 1. Owing to the number of statistical tests performed and to minimize the type 1 error rate, only differences of P≤.01 were considered statistically significant. Trend level differences (P≤.05) were noted for descriptive purposes only.

Three logistic regression analyses were performed to determine the best predictors of receiving pneumococcal vaccine at least once, influenza vaccine at least once, and influenza vaccine within the last year (dependent variables). To assess factors associated with administration, only baseline variables that were related to immunization at P≤.10 significance levels in relevant univariate logistic analyses were used in the models (independent variables included the following: age; aged ≥65 years; sex; marital status; insurance status; current alcohol use; other substance abuse; and smoking; purified protein derivative placement; depression; number of health problems; having cardiovascular disease, pulmonary disease, or diabetes mellitus; and the mean number of medications taken). For all regressions, educational level and the number of household members were excluded owing to missing data. All models were assessed for statistical outliers; these were removed and the models were refit. For regression analyses, P≤.05 were considered statistically significant.

The mean age of the 550 older urban Native American patients was 60.6 years; females composed 61%, and 28% were married. Overall, 45% had at least a high school education; 43% were employed; 49% were receiving Medicare or Medicaid insurance coverage. A total of 492 (89%) of the 550 patients had documented medical problems (mean=2.95) and 299 (54%) of the 550 patients had documented psychological problems (mean=0.7). Influenza vaccine had been administered to 270 patients ever (49%), and 163 patients (30%) in the last year. For pneumococcal immunization, 118 (21%) had been previously vaccinated. Characteristics of the study sample by vaccination status are given in Table 1. In general, patients who were younger (P≤.01-P≤.001) or current users of alcohol (P≤.05-P≤.001) were immunized less often while those with Medicare insurance (P≤.05), depression (P≤.01-P≤.001), a greater number of health problems (P≤.01), and taking more medications were vaccinated more frequently (P≤.001).

In terms of strict adherence to published guidelines, the data in Table 2 demonstrate that being 65 years or older (P≤.01), having cardiovascular disease (P≤.01), pulmonary disease (P≤.05), or diabetes mellitus (P≤.001) were correlated with receipt of influenza vaccine in the last year. Similarly, age (P≤.001) and having cardiovascular disease (P≤.01), pulmonary disease (P≤.01), or diabetes mellitus (P≤.01) were associated with ever receiving a pneumococcal vaccination. Overall, among Native American elders with any indication, only 31% had received influenza in the previous year and 21% pneumococcal immunizations in accord with widely accepted recommendations. Among patients not considered at high risk, 16% of all patients and 40% of those at least 65 years old had been vaccinated against influenza in the last year. All patients who received a pneumococcal immunization had a medical indication.

Odds ratios and 95% confidence intervals for receipt of immunizations are given in Table 3. The logistic regression revealed that the only factor significantly associated with receiving an influenza vaccine in the last year was the presence of diabetes mellitus. The factors associated with ever receiving an influenza vaccination included greater mean numbers of health problems and current medications, having had a purified protein derivative placed, and type of insurance, particularly Medicare (P≤.01). Similarly, predictive factors for lifetime receipt of pneumococcal vaccine were greater number of health problems, the presence of diabetes mellitus, and type of insurance, again most importantly Medicare (P≤.01).

Physicians agree that prevention and early disease detection are crucial. Furthermore, increasing public awareness, decreasing costs of screening, and the demonstrated cost-effectiveness of preventive services have encouraged their implementation.¹⁵ Yet, despite the proven benefit of many preventive measures, recent studies in clinical practices have shown that use of preventive services is low and only a minority of eligible patients actually receive them. Furthermore, little is known about the use of most preventive and screening measures among older adults. At least for some screening procedures, the elderly, as well as those who are poor, less educated, and nonwhite, are the least likely to receive them.^10,16- 17 This is unfortunate since screening and preventative guidelines, typically implemented in the general medical setting, may have the greatest benefit among older persons.¹⁸ These observations suggest that elderly AIs/ANs may be an especially neglected population.

Among the elderly, influenza and invasive pneumococcal infections may be the most vaccine-preventable infectious diseases.¹⁴ In the United States, approximately 120,000 older adults are hospitalized for influenza and 200,000 for pneumococcal pneumonia annually.¹⁹ Together, these 2 infections account for 50,000 to 80,000 deaths per year—almost 85% in the elderly.¹⁹ Yet, despite this substantial morbidity and mortality, adherence to immunization guidelines has varied widely. For example, statistics compiled using the 1997 Behavioral Risk Factor Surveillance System revealed that among persons at least 65 years of age, 66% reported receipt of influenza vaccine in the past year while 45% reported ever receiving pneumococcal vaccine.²⁰ In contrast, some authorities have estimated that only 10% to 20% of target populations have been vaccinated against influenza or pneumococcus despite adequate Medicare reimbursement.^19,21 Similarly, in a recent primary care study, influenza and pneumococcal vaccines were administered to just 5% and 21% of the eligible adults, respectively.²² Other investigators have observed higher, though still inadequate, rates of pneumococcal (32%) and influenza (57%) immunization among patients 70 years or older.²³ Thus, the rates of immunization at SIHB, although inadequate particularly for the high-risk groups, are similar to those published for other populations.

Respiratory diseases, most prominently pneumonia, are an important cause of disability and death among Native American peoples. For example, pneumonia accounts for 5% of hospitalizations among AIs/ANs.⁷ By the age of 75 years, the cumulative incidence (7%) and mortality (1%) of pneumococcal infections among ANs is 4 times higher than the rates observed in nonnatives.⁶ American Indians/Alaskan Natives also seem to have more recurrent disease than nonnatives. In this regard, an older study reporting the Navajo experience observed recurrences of pneumonia in 7.6% of adults.²⁴ Yet, despite their vulnerability, even those AIs/ANs at high risk are infrequently immunized—in another investigation only 17% of AN adults with predisposing conditions and invasive pneumococcal infections had previously received pneumococcal vaccine.⁶ Moreover, national statistics have consistently documented that older minority individuals (there are no data specific to AIs/ANs) are less likely than whites to receive influenza and pneumococcal vaccine.^20,25 The public health implications of these missed opportunities for immunization are obvious.

Although this study does not directly address barriers to care, many patient- and physician-based factors have been associated with poor implementation of screening guidelines, including lack of health insurance, culturally influenced beliefs, the type of clinical setting in which medical care is provided, and physician factors.^10,26- 30 For example, the Survey of American Indians and Alaska Natives, a 1987 study by the Agency for Health Care Policy and Research,³¹ revealed that AIs were much less likely to have private insurance coverage than any other segment of the US population (28% vs 81% whites, 53% blacks, 50% Hispanics). Even excluding Indian Health Service coverage, AI households had significantly higher rates of public coverage than the total US population (17% vs 10%). Both findings relate directly, albeit in inverse ways, to employment and job opportunities. However, these data also indicate that AI people who work and have health coverage as an option are less likely to enjoy private coverage than their nonnative counterparts (eg, 36% vs 75%). Overall, 55% of AI families are either uninsured or entirely reliant on the Indian Health Service for health care. This set of circumstances is especially disheartening given the association between adherence to immunization guidelines and type of insurance observed in the present study.

Provider- and facility-based factors for low immunization rates may be of equal importance to economic ones. In this regard, physicians whose practices consisted primarily of minorities saw more patients, spent less time with each patient, and incorporated fewer preventive practices into their encounters than did providers who had more affluent, predominantly white patients.²⁷ Another study of primary care providers found that although 88% agreed that guidelines improved care for the elderly, physicians' attitudes often did not correlate with their use of preventive care recommendations.²⁸ These studies are consistent with others that have found lack of time was the reason for failure to comply with guidelines reported most often by physicians.^29- 30 Together, the issues cited above are important to the health of elderly AIs/ANs, as well as the attainment of the Healthy People 2000 objective to increase to 40% the proportion of older adults who receive all screening and immunization services and at least 1 age-and sex-appropriate counseling service as recommended by the US Preventive Services Task Force.³²

This study has several limitations. First, complicating studies of vaccination is the absence of universally accepted recommendations. Second, owing to the nature of the sample, our findings cannot be generalized to other, particularly nonurban, clinics serving AIs/ANs. This drawback is somewhat attenuated by the fact that most Native American people now live significant portions of their lives in cities. Third, because the SIHB is a model program for the health care of urban AIs/ANs, the proportion of patients receiving preventive procedures might exceed that in other metropolitan settings. Even so, since this investigation was retrospective and based exclusively on medical record reviews, it precluded the consideration of outside sources of care. Thus, it is likely that we underestimated the frequency of immunization since some patients may have been vaccinated in other settings (eg, Public Health Clinics, seniors centers) without this information being recorded in their SIHB medical record. Finally, studies such as the present one can not directly ascertain the barriers to care experienced by specific individuals or even the group as a whole.

If validated by further research, these data have important clinical implications and suggest strategies for optimizing immunization by improving physician performance and increasing patient awareness. For example, a nurse-initiated prompting system, flow sheets, and computerized reminders have all been shown to increase the delivery of preventive services^33- 36 and may be well advised here. Among urban minorities, compliance with preventive measures has been enhanced by an intense, personal intervention designed to remove barriers to care,³⁷ and by computerized patient reminders.³⁸ Other interventions such as mobile units, videotapes, culturally relevant educational material,³⁹ or site of service modification³⁴ also may be useful in caring for AI/AN populations. Tribal health care centers and organizations interacting with older AIs/ANs will need to develop effective means of educating both providers and patients about the importance of respiratory disease prevention. We believe our findings underscore the need to identify patient-based and primary care physician-based obstacles to immunization, and to conduct evidence-based prospective and intervention studies in clinic- and community-based Native populations.

Accepted for publication October 20, 1999.

This investigation was supported in part by grants 1P30AG/NE15292 from the National Institute of Aging and the National Institute of Nursing Research (Drs Buchwald and Manson), and 90-AM-0757 from the Administration on Aging (Dr Manson) Rockville, Md.

We thank the Seattle Indian Health Board and Executive Director Ralph Forquera, MPH, for facilitating the conduct of this work.

Corresponding author: Dedra Buchwald, MD, Harborview Medical Center, 325 Ninth Ave, Box 359780, Seattle, WA 98104 (e-mail: dedra@u.washinton.edu).