Generalist and Subspecialist Physicians' Knowledge, Attitudes, and Practices Regarding Influenza and Pneumococcal Vaccinations for Elderly and Other High-Risk Patients:  A Nationwide Survey FREE

INFLUENZA AND pneumococcal diseases cause hundreds of thousands of hospitalizations and tens of thousands of deaths each year.^1- 2 In fact, these 2 vaccine-preventable diseases are responsible for more deaths than all other vaccine-preventable diseases combined, with more than 90% of these deaths occurring among the elderly and others with high-risk medical conditions.³

Influenza and pneumococcal vaccinations are safe, cost-effective, and, for the elderly, even cost-saving means for preventing these diseases.^4- 5 Vaccination rates of 90% for all elderly persons in this country and 60% for high-risk persons younger than 65 years have been established for the year 2010 national public health objectives.⁶ Despite long-standing recommendations for their routine use among elderly persons and other high-risk groups, however, these vaccines continue to be underused. In 1999, approximately 67% of persons 65 years and older had received an influenza vaccination, and 55% reported ever having received a pneumococcal vaccination.⁷ Immunization rates for high-risk persons younger than 65 years are even lower, especially for racial minority groups.⁸ Vaccine delivery efforts must make dramatic improvements if the 2010 influenza and pneumococcal vaccination goals for elderly persons and other high-risk adults are to be met.

Providers' recommendations to their patients for vaccination^9- 10 and the implementation of organized vaccination programs¹¹ are important facilitators of vaccine delivery. Successful programs often have incorporated multiple components, such as target group identification and recalls/reminders, standing orders, special clinics, and provider feedback.^12- 13 Little information is available regarding practitioners' knowledge, attitudes, and behaviors regarding these facilitators of vaccine delivery. This cross-sectional survey of a nationwide sample of practicing physicians was conducted to assess generalist and medical subspecialist physicians' knowledge and attitudes regarding influenza and pneumococcal vaccinations for adults and their use of various strategies to promote delivery of these vaccines in their clinical practice setting.

The target group was practicing physicians most likely to be involved with the provision of adult vaccinations. Hence, family physicians, general internists, and internal medicine subspecialists were selected for stratified random sampling from the American Medical Association's physician master file, which contains information on more than 700 000 active US physicians (nonmembers are included). From each of these groups, 2000 physicians were randomly selected. The sampling frame excluded residents, retired physicians, physicians practicing in US territories, and federal physicians. The sampling frame otherwise included all other doctors of medicine listed in each of the indicated groups.

The questionnaire was developed by experts in internal medicine, family medicine, preventive medicine, and public health using an iterative process. Questions addressed multiple domains, including perceptions of disease severity, vaccine safety and efficacy, cost-effectiveness of vaccination, programmatic interventions to raise rates, and liability. The survey instrument contained 23 questions. The study was approved by the human studies subcommittee of the Minneapolis Veterans Affairs Medical Center, Minneapolis, Minn.

Self-administered, anonymous questionnaires were mailed to these 6000 physicians between January 1, 1999, and May 31, 1999. A total of 3 mailings were conducted. Data were entered into a computerized database using Epi Info 6 software (version 6.02; Centers for Disease Control and Prevention, Atlanta, Ga).

χ² and t tests for independent samples were used to compare responses for categorical and continuous variables between respondents who were generalist physicians (ie, family medicine or general internal medicine physicians) and those who were medical subspecialists. Family medicine and general internal medicine physicians typically serve as primary care providers; to simplify the analysis, we grouped them together in one category as generalists. Stepwise multivariate logistic regression (SPSS 8.0; SPSS Inc, Chicago, Ill) was used to identify variables that were independently associated with practitioners strongly recommending these vaccinations to their patients. Variables were retained in the models if they had P≤.05. Goodness-of-fit was assessed using the Hosmer-Lemeshow test.

Of the 6000 physicians initially identified for the survey sample, 64 were listed with incorrect addresses, 47 indicated that they were retired, 27 indicated that they did not give any immunizations or did not see patients, and 4 refused to participate. This left a total of 5858 eligible survey recipients, of whom 1874 (32%) responded after the 3 mailings. The characteristics of the respondents are shown in Table 1. Approximately two thirds of respondents in each group practiced in a single specialty group. Generalist physicians were more likely to be women and spent more time in direct patient care than did medical subspecialists. Generalists were also more likely to report having been vaccinated against influenza during the preceding season. A higher proportion of patients cared for by subspecialists were adult and elderly patients. Most practitioners in both groups had cared for at least 1 patient in the previous 2 years who had been hospitalized for pneumonia.

Practitioners' responses to questions about their attitudes toward preventive health care services and vaccinations are also summarized in Table 1. Generalists were more likely than subspecialists to report that the delivery of preventive health care services was very important in their practices, that it was very important for their high-risk patients to be up-to-date on their immunizations, and that it was very important for health care workers to receive influenza vaccinations.

Practitioners' knowledge of the safety and cost-effectiveness of influenza and pneumococcal vaccinations is summarized in Table 2. Most (70%-80%) of practitioners in both groups estimated that local symptoms after vaccination occurred in fewer than 10% of vaccine recipients, a profile similar to that reported for the occurrence of systemic symptoms, although generalist physicians tended to estimate somewhat lower rates of adverse effects than did subspecialists, particularly for pneumococcal vaccination. Both groups also tended to have similar responses to questions regarding the cost-effectiveness of influenza and pneumococcal vaccinations for elderly persons. In each case, more than 45% of respondents indicated that these vaccinations were cost saving, and more than 35% of respondents indicated that the vaccinations were not cost saving but nevertheless cost effective.

Generalist physicians were more likely than subspecialists to use each of the vaccination strategies identified in the survey (Table 3), including the major strategies that have been shown to be most effective for increasing vaccination rates (ie, patient reminders, medical chart reminders for practitioners, special clinics, and standing orders) (Figure 1 and Figure 2). The mean ± SD number of major strategies used by generalists vs subspecialists was 1.18 ± 1.09 vs 0.70 ± 0.94 for influenza vaccinations and 0.79 ± 0.94 vs 0.51 ± 0.82 for pneumococcal vaccinations (P<.001 for both). In addition, generalists reported more often that vaccination rates were monitored as part of ongoing quality assurance activities, although fewer than 40% of all respondents indicated that vaccination rates were monitored (Table 3).

Generalist physicians were more likely than subspecialists to very strongly recommend influenza and pneumococcal vaccinations to their elderly and high-risk patients (Table 3 and Figure 1 and Figure 2). Although most respondents indicated that they very strongly recommend these vaccinations to their at-risk patients, a nontrivial proportion indicated that they did not.

Factors that were very important in affecting decisions to strongly recommend vaccination to high-risk patients in bivariate analyses are listed in Table 3. More than 50% of the respondents identified a patient's risk for illness, vaccine effectiveness, recommendations from expert groups, ease of targeting high-risk patients, and remembering to offer the vaccination as being among these very important factors. Fewer than 30% of respondents indicated that cost/reimbursement and liability issues were very important.

The results of the stepwise logistic regression analysis are shown in Table 4. The factors most highly correlated with practitioners strongly recommending influenza and pneumococcal vaccinations to their elderly patients were similar for both vaccinations: the practitioner having received an influenza vaccination himself or herself, belief that it is very important for health care workers to receive an influenza vaccination, belief that the vaccinations are cost saving, sex, a patient's risk for disease, vaccine effectiveness, and concern about liability issues. In addition, for influenza vaccination, ease of targeting high-risk patients was associated with the likelihood of vaccination. For pneumococcal vaccination, concerns about drug resistance, having sufficient time to counsel patients, and recommendations of expert groups were also associated with decisions to vaccinate.

The results of this survey demonstrate that generalist and medical subspecialist physicians may miss many opportunities to provide influenza and pneumococcal vaccinations to their elderly and other high-risk patients. Although physicians generally report that the provision of preventive health care services is important in their practices and that it is very important for their high-risk patients to be up-to-date on immunizations, a significant proportion of physicians nevertheless indicate that they do not strongly recommend influenza or pneumococcal vaccinations to their patients. Furthermore, most practitioners do not use strategies in their practice settings that have been demonstrated to successfully increase vaccination rates.

A practitioner's recommendation for vaccination has been shown previously to be a strong predictor of the vaccination behavior of high-risk patients,^9- 10 and lack of a practitioner recommendation is often cited by unvaccinated patients as a reason for their not being immunized.¹⁴ Even when patients otherwise report having negative attitudes toward vaccination, if a practitioner has recommended the immunization, vaccination rates in excess of 80% or more can be achieved.^9- 10 Thus, the finding that significant numbers of practitioners do not strongly recommend influenza or pneumococcal vaccinations for their elderly and other high-risk patients is concerning. The multivariate analysis suggested that attitudes toward prevention, knowledge of vaccine effectiveness, cost-effectiveness, patients' risk for illness, and system issues, such as having sufficient time or ease of targeting patients, may affect decisions to strongly recommend vaccinations. These results suggest that practitioner education and enhanced attention to processes that identify patients needing vaccinations and facilitate efficient interactions with patients may increase the likelihood of practitioners recommending these vaccinations to their patients. Patients should also be educated about the importance of vaccinations and should be encouraged to ask their practitioner about vaccinations.³

Effective implementation of organized programs to promote vaccination that include components such as patient identification and recall systems, standing orders, and expanded access through special clinics can be critically important for improving vaccination and sustaining high vaccination rates.^12- 13 Respondents in this study reported that these kinds of strategies were used relatively infrequently in their clinical practice settings. Practitioners should be educated about the importance of such strategies and how to implement them, as has been done in the Teaching Immunization for Medical Education Project.¹⁵ The Centers for Disease Control and Prevention has published materials on strategies for increasing adult vaccination rates that address advantages and disadvantages of specific strategies and provide examples and suggestions for effective implementation.¹⁶ Other initiatives, such as the National Pneumonia Project,¹⁷ a Medicare quality improvement project that involves peer review organizations and local hospitals and health care providers in every US state and territory, are also providing help and support to providers in these areas.

Fewer than 30% of responding physicians indicated that cost or reimbursement levels were very important in their decision to offer vaccination to their patients. Although achieving adequate reimbursement levels is important,³ other issues, such as ease of targeting high-risk patients, remembering to offer the vaccination, and having sufficient time and personnel, were more often identified as being very important to practitioners. These types of findings reinforce the importance of organized systems and processes that maximize efficiency and minimize workload burden on physicians and their office setting.

Having objective measures of performance is essential for effective quality improvement activities, both for the identification of areas in need of improvement and for monitoring the success of interventions.¹⁸ Such measurements might include the monitoring of pharmacy data on numbers of vaccine doses used, patient surveys, or medical record audits. As demonstrated in this study, however, measurement of vaccination rates is frequently not incorporated into ongoing evaluation efforts within many practice settings. Unless these types of measures are used by practitioners, it is unlikely that the 2010 goals for influenza and pneumococcal vaccinations will be met.

Both specialty (generalist vs subspecialist) and sex were associated with significant differences in behaviors regarding recommending vaccinations and using specific vaccination strategies in the practice setting. Female physicians were more likely than their male counterparts to recommend both of these vaccinations to their patients. Female sex has previously been associated with higher delivery rates of preventive health care services.^19- 20 This may be due to either differences in patient mix according to the sex of the physician or differences in attitudes and practices between female and male physicians. This phenomenon deserves further attention and study.

Differences in practice behaviors between generalist and subspecialist physicians have also been observed across a variety of types of services.²¹ In some circumstances, generalists provide more appropriate care, whereas in others it seems as if subspecialist physicians provide more appropriate care. Regarding preventive health care services, generalist physicians have been observed to outperform subspecialists in some instances.²¹ In addition to the Advisory Committee on Immunization Practices of the US Public Health Service^1- 2 and the US Preventive Services Task Force,²² the American Academy of Family Physicians²³ and the American College of Physicians²⁴ have also endorsed the delivery of influenza and pneumococcal vaccinations to elderly and other high-risk patients by physicians providing care to these patients. Because almost all of the patients who many medical subspecialists provide care for would fall into 1 or more groups targeted for influenza and pneumococcal vaccination, subspecialists seem to be missing many opportunities to vaccinate these high-risk patients. The provision of needed immunizations should be considered part of the spectrum of care provided by generalists and subspecialists. Routine administration of influenza and pneumococcal immunizations to high-priority groups has been identified as an important practice standard by the Infectious Diseases Society of America.²⁵ The American Diabetes Association²⁶ has published a position statement highlighting the importance of influenza and pneumococcal vaccinations for patients with diabetes mellitus. Other professional organizations have also encouraged vaccination of high-risk patients²⁷ and have endorsed the importance of prevention within their fields.²⁸

Generalists and medical subspecialists also differed in their personal behaviors regarding receipt of influenza vaccination, with generalists being more likely than subspecialists to report having received an influenza vaccination during the previous vaccination season. However, even among generalists, nearly 20% had not been immunized. Health care workers are among the groups targeted for annual influenza vaccination because they have been implicated as potential sources or vectors for the transmission of influenza to high-risk patients.¹ Thus, vaccination of health care workers is important for the protection of the patients they treat. The findings from this study suggest that the personal vaccination behaviors of health care providers may also affect the likelihood of their recommending influenza and pneumococcal vaccinations to their patients. Health care workers should be strongly encouraged to receive annual influenza vaccinations.

Despite the fact that physicians tended to overestimate the occurrence of systemic adverse effects after vaccination, concern about adverse effects was not an independent predictor of the likelihood of recommending vaccinations to high-risk patients. Nevertheless, it is important that physicians and their patients have accurate information regarding the safety of vaccinations. Recent clinical trials have established that systemic symptoms do not occur at higher rates after influenza vaccination of elderly^29- 30 or healthy, younger³¹ adults than after placebo injections. Similarly, studies^32- 33 suggest that the rates of systemic symptoms after pneumococcal vaccination are low and may be similar to the rates experienced at baseline.

The findings from this study should be interpreted with some caution. The response rate to our survey was 32%, similar to the 20% to 45% response rates reported in other recently published physician surveys^34- 38 but lower than the average response rate of 54% calculated from a review³⁹ of studies published in 1991. We were unable to collect information on nonrespondents in our study. Because the potential exists for significant respondent-nonrespondent bias, the study findings may have limited applicability to other generalist and medical subspecialist physicians. However, an analysis⁴⁰ of 5 studies reporting on the demographic characteristics of early (ie, respondents to first mailing) vs late (ie, nonrespondents to first mailing) respondents found that the demographic characteristics of early and late respondents (considered proxies for nonrespondents) did not differ significantly. Thus, the nature of nonrespondent bias in physician surveys is unclear and may be modest. For the analysis, we grouped general internists and family practice physicians together as generalists. It is possible that by so doing we missed important differences between general internists and family practice physicians. A limited subgroup analysis of the data, however, showed that general internists and family practice physicians did not differ significantly with regard to their opinions about the importance of patients being up-to-date on influenza and pneumococcal vaccinations, how often they strongly recommend the vaccinations to high-risk patients, and their opinions regarding the importance of health care workers being immunized (data not shown). Such findings suggest that this grouping was appropriate. Finally, as a cross-sectional survey, this study provides evidence for associations between variables and outcomes but does not establish actual cause-and-effect relationships.

Influenza and pneumococcal diseases cause hundreds of thousands of hospitalizations and tens of thousands of deaths each year. Despite the availability of safe and effective vaccines for the prevention of these diseases, these immunizations are underused. Many opportunities for the prevention of these diseases are still being missed. Although people may increasingly be receiving their influenza vaccinations in nontraditional settings or from nontraditional providers such as pharmacists, more than 60% of elderly persons receive their influenza vaccinations at their physician's office or health maintenance organization.⁴¹ Physicians continue to play a critical role in the delivery of influenza and pneumococcal vaccinations to their elderly and high-risk patients. Health care providers should make every effort to recommend these vaccinations to their patients. They should also incorporate other strategies into their clinical practices to ensure effective provision of these cost-reducing and lifesaving vaccinations to their elderly and otherwise high-risk patients.

Accepted for publication July 31, 2001.

This study was supported in part by a grant from Aventis Pasteur, Swiftwater, Pa.

We thank Roderick MacDonald, MS, for assistance in conducting the survey.

Corresponding author and reprints: Kristin L. Nichol, MD, MPH, MBA, VA Medical Center (111), One Veterans Drive, Minneapolis, MN 55417 (e-mail: nicho014@tc.umn.edu).