Objective: To determine which office and patient factors affect adult influenza and pneumococcal vaccination rates.
Study Design: Patient interviews and self-administered surveys of office managers.
Patients and Methods: In a 2-stage random cluster sample, 22 practices in 4 strata (Veterans' Affairs, rural, urban/suburban, and inner city) and 15 patients per physician in each practice (n = 946) were selected. Office managers completed a questionnaire regarding office practices and logistics affecting immunizations. Data were examined using2 and regression analyses without and with patient factors in the models.
Results: Practice factors significantly related to influenza vaccination status were stratum (VA OR = 2.04; 95% CI = 1.18,3.53; P < .05 vs inner-city), time allotted for acute care visits (16-20 min vs 10-15 min OR = 2.49; 95% CI = 1.68, 3.09; P < .001), the practice not having a source of free vaccines (OR = .43; 95% CI = .3, .62; P < .001), and the interaction between being an urban/suburban practice and having a source of free flu vaccines (OR = 4.0; 95% CI = 2.63, 6.09; P < .001). Practice factors related to pneumococcal vaccination status were the number of immunization promotion activities (≥3 vs 0-2 OR = 1.97; 95% CI = 1.33, 2.94; P = .002) and the time allotted for acute care visits (16-20 min vs 10-15 min OR = 1.94; 95% CI = 1.18,3.19; P = .011). When practice and patient factors were combined in the analyses, patient factors were more important.
Conclusion: Although patient factors are more important than practice factors, practices that allot more time for acute care visits and use more immunization promotion activities have higher vaccination rates.
(Am J Manag Care. 2004;10:13-19)
Deaths from influenza and pneumonia are increasing, disproportionately affecting the growing elderly population. Cause-specific and underlying circulatory and respiratory disease mortality rates associated with influenza and pneumonia are significantly higher among those older than 65 years.1 During the past decade, 90% of influenza-related deaths occurred in persons 65 years and older,1 despite the fact that safe and effective vaccines against influenza and pneumococcus are readily available for adult patients. Moreover, Medicare covers the cost of both vaccines, and Health Plan Employer Data and Information Set measures, against which the National Committee for Quality Assurance evaluates health plans, include influenza and pneumococcal vaccinations for individuals older than 65 years. Yet, adult influenza and pneumococcal immunization rates have not reached Healthy People 2010 goals of 90%.2 Among adults 65 years and older, the national influenza vaccination rate was 63% to 65% in 1999 through 2001,3 and the pneumococcal vaccination rate was 55% in 2002. Rates among minority populations were lower still (48% among Hispanics and 50% among blacks for influenza vaccination and 25% among Hispanics and 35% among non-Hispanic blacks for pneumococcal vaccination).3
The Task Force on Community Preventive Services4 recommended reminder/recall systems for patients and physicians and system interventions to increase immunization rates, and Gyorkos et al5 stressed the importance of system- or practice-level interventions. However, different combinations of practice, physician, and patient factors create unique office environments that can result in a wide range of immunization rates across practices.6 Understanding the relative impact of various factors will assist in the development of targeted interventions to raise immunization rates. This study examines specific practice factors, as reported by office managers, which affect self-reported influenza and pneumococcal immunization rates among patients from practices in diverse geographic and socioeconomic settings. Further, patient and practice factors are combined to examine the relative strength of their associations with immunization rates.
Practices, office managers, and patients were selected from 4 strata to include a broad spectrum of older patients and vaccination policies. Detailed methods, including the sampling strategy for this and related studies, have been published previously.7
Office Managers and Practices.
The practice distribution from the 4 strata is as follows: 8 rural medical practices and 7 urban/suburban medical practices in a large network of nonacademic practices (n = 63) affiliated with the University of Pittsburgh Medical Center in western and central Pennsylvania, 3 outpatient clinics in Veterans Affairs (VA) health centers in western Pennsylvania, and 4 inner-city neighborhood health centers in Pittsburgh, Pa.
In the large rural and urban/suburban University of Pittsburgh Medical Center strata, random samples of practices were selected, and in the smaller, inner-city and VA strata, all of the practices were selected. The office manager for each practice was asked to complete the questionnaire.
A random sample of elderly patients for each clinician in each practice was selected using billing lists, with a target of 15 completed patient interviews per participating clinician. Patient inclusion criteria were age 66 years or older, an office visit after September 30, 1998, and the ability to personally complete a study interview by telephone. Patients who were homeless, nursing home residents, or not currently living in the region and those who were deaf, had severe psychosis, or had dementia were excluded.
Office Manager Questionnaire.
The self-administered office manager questionnaire focused on office logistics and adult immunization practices and economics. The questionnaire took approximately 20 minutes to complete and consisted of approximately 50 openand closed-ended questions. Responses to open-ended questions were grouped and coded.
The patient interview was designed by a multidisciplinary team using an iterative process. It was based on the Triandis model8 for consumer decision making from the Theory of Reasoned Action. This model includes facilitating conditions (eg, ease of getting to a location for an influenza shot) and behavioral intention, consisting of attitude about the activity (eg, getting an influenza shot is wise), social influences (eg, physician or family member recommends influenza shot), and the value of the consequences of the activity (eg, the influenza shot prevents influenza). The model predicts a variety of behaviors well,8-11 including exercise11 and birth control/fertility10 behavior; has been used in different cultural and economic situations10; and, as used for influenza immunization, has been shown to be internally consistent and externally valid (Cronbach Â¦Ã = .79-.91).9 The final questionnaire contained approximately 100 questions, depending on the skip pattern, including multiple- choice items and Likert scale items, and took approximately 30 minutes to complete. Both questionnaires used in this study can be viewed at http://www. pitt.edu/~familymd/immunization/.
The self-administered questionnaire and a stamped return envelope were given to office managers at the time of a visit by study staff. Surveys were completed and returned by mail between August 11, 1999, and May 15, 2000. Office managers received a $10 honorarium for participation.
A personalized introductory letter and a letter from the sites endorsing the project and encouraging participation were sent to each of the sampled patients. A $10 honorarium was offered to encourage participation in the survey.
Trained interviewers conducted the surveys using computer-assisted telephone interviewing,6,12 which facilitated management of the sample interviewed, allowed for direct data entry during the interviews, directed the sequence of questioning, prevented skipped questions through automated skip patterns, and blocked illogical or out-of-range values. A detailed description of the methods used in this study has been published previously.7 The study was approved by the institutional review board of the University of Pittsburgh and by the human subjects use subcommittee of the institutional review board of the VA Healthcare System of Pittsburgh, and signed informed consent was obtained from participants.
SUDAAN software (RTI, Research Triangle, NC) was used for analyses to account for the cluster-correlated nature of the data, ie, patients are clustered within a provider. Analyses were weighted by the estimated fraction of eligible elderly patients for each physician. Closed-ended questions included some that required rating responses on Likert scales. Due to low frequencies for some response alternatives, responses were collapsed into categories as appropriate. Either 2 or Fisher exact tests were used to compare patients who received influenza and pneumococcal vaccinations with those who did not with regard to practice logistics variables.
Logistic regression analysis using SUDAAN was performed to determine significant correlates of patient receipt of influenza vaccination in the 1999-2000 season and pneumococcal vaccination ever. All variables with < .10 in bivariate analyses with the outcome variable were included in a stepwise forward-selection procedure in 2 types of models. The first model included only office practice factors and immunization status. The second model combined significant office practice factors, patient factors previously found to be associated with self-reported immunization status,6,12 and interaction terms. Interaction terms were determined by selecting all factor pairs correlated at 0.5. Statistical significance was set at < .05 for all outcome analyses.
Twenty-two office managers and 946 patients participated in the study. Office manager demographics were not collected, as these data were not relevant to office logistics and other office characteristics. Patient demographics have been described previously.6 Overall vaccination rates for influenza (78.4%) and pneumococcus (67.8%) have been reported previously. 6,12
Staffing patterns, that is, the number of variously trained staff and their patient care responsibilities, had varying effects on immunization status for influenza and pneumococcal vaccines (Table 1). An increased number of licensed practical nurses and clerical staff was associated with higher vaccination rates for influenza and pneumococcal vaccines; however, the number of physicians, registered nurses, and medical assistants did not affect vaccination rates. When physicians or office managers provided triage for the practice, influenza immunization rates were lower. In practices in which medical assistants were not responsible for assessing vital signs and giving shots, influenza vaccination rates were higher. Pneumococcal vaccination rates were higher in practices in which physicians or office managers were not responsible for assessing vital signs.
Longer office visits, both acute and chronic care, were associated with higher influenza vaccination rates but did not affect pneumococcal vaccination rates (Table 1). Offices that had implemented immunization promotion strategies, such as designated vaccination clinics, computerized immunization tracking systems, prompts on patient charts, and a source of free or reduced-cost vaccinations, had higher pneumococcal vaccination rates.
In logistic regression analyses (Table 2), influenza vaccination rates were significantly associated with stratum and the length of time allotted for acute care visits and inversely related to having a source of free vaccinations. Among those intercorrelated factors entered into the model, only the interaction between urban/suburban stratum and having a source of free vaccinations was significantly related to higher influenza vaccination rates. Pneumococcal vaccination rates were also significantly associated with having more time allotted for adult acute care visits, as well as with the number of office systems in place for promoting immunizations.
To determine which office factors were significant while accounting for patient factors previously determined to be significantly associated with immunization status, we performed stepwise forward logistic regression analyses for each vaccine using all significant factors ( < .1) from the office manager and patient surveys. For influenza vaccination, stratum was the only office factor that remained, along with several patient factors, including marital status, awareness of the recommendation to receive the influenza shot, willingness to obtain the influenza and pneumococcal vaccinations simultaneously, belief in the vaccine's efficacy, and the physician's recommendation of the vaccine (Table 3). For pneumococcal vaccination, no office factors remained in the model, and only 3 patient factors - physician's recommendation, belief in the value of pneumococcal vaccine, and having gotten the influenza vaccine during the most recent influenza season - were significantly associated with ever receiving pneumococcal vaccine (Table 4).
Certain practice factors, including larger numbers of licensed practical nurses and clerical staff, longer amounts of time allotted for visits, obtaining free vaccines, and a greater number of immunization promotion activities, were associated with higher adult immunization rates. Larger support staffs that perform routine tasks and can be trained to assess immunization status and administer immunizations reduce the need for physicians to perform these tasks and allow physicians to concentrate their time and energy on screening, diagnosis, and treatment. The VA has achieved high immunization rates through such interventions, including freestanding immunization clinics and standing orders for nursing staff to vaccinate eligible individuals.13,14
Previous research has shown that productivity incentives decrease the provision of preventive services.15 Increases in productivity are often associated with decreases in visit time; therefore, it follows that physicians need to focus on the most pressing concerns in the time allowed for either an acute or a chronic care visit. Longer visit times afford the physician more time to address preventive services such as immunizations, as we found. The cost of longer visit times can be offset by understanding that influenza and pneumococcal polysaccharide vaccinations have been shown to result in significant cost savings due to prevention of life-threatening illness, decreased hospitalizations, and increased years of quality life. For example, among individuals older than 65 years, influenza vaccination saves approximately $42 per person vaccinated, whereas pneumococcal vaccination saves approximately $20 per person vaccinated.16,17 An outcomes basis for productivity measures is justified in this instance.
Dickey and Kamerow18 reported that the number of office resources, such as flow sheets, patient educational materials, and patient and provider reminders, used by physicians' offices was positively related to the number of preventive services, including immunizations, that were provided. Gyorkos et al5 reported on the interventions with the strongest evidence for success in increasing immunization rates. Typically, interventions that are multimodal, ie, that address patient, physician, and system factors, are the most successful. The Task Force on Community Preventive Services4 similarly encourages interventions that combine intervention modalities. Our data, indicating that promoting immunization through 3 or more activities is related to higher pneumococcal vaccination rates, support this concept.
In the analyses that combined factors derived from the patient survey and factors derived from the office manager survey, patients' attitudes and knowledge about vaccines were more important than were any specific office factors identified through previous analyses. However, this does not mean that the office environment has no role to play in affecting adult immunization rates. Knowledge that Medicare covers the cost of influenza vaccination, awareness of the recommendation to get an annual influenza vaccination, and understanding that without the influenza shot an individual is more susceptible to influenza are all examples of information that can be disseminated in the physician's office via educational posters, fliers, or brochures or a well-informed staff. Patients' perceptions that getting either vaccine is not more trouble than it is worth can be fostered in the practice by offering walk-in vaccination clinics, offering the pneumococcal vaccine at the same time as the influenza vaccine, and offering immunizations at acute care visits. Finally, patients' perceptions that physicians and staff alike recommend influenza and pneumococcal vaccines is easily communicated by training all staff with patient care duties to discuss vaccinations at every appropriate visit.
Strengths and Limitations
The strengths of this study are the wide variety of physician practice types and settings, the demographic characteristics of physicians and patients, and data collection from several sources, ie, practices and patients.
A possible limitation is the use of self-report of patient immunization status, although self-report of vaccination status has been used in national surveys such as the Behavioral Risk Factor Surveillance System.19 The sensitivity and specificity of self-report compared with chart audit range from 92% to 100% and from 71% to 98%, respectively, and kappa ranges from 0.72 to 0.92.20-22
Although the sampling technique attempted to capture diverse practice types and settings and patient demographics, we cannot assume that they represent the entire United States. Furthermore, vaccination rates in this population were higher than those found nationally. However, our purpose was to understand which office and patient factors are truly related to immunization status, not how common these factors are across the country. Hypothesis testing does not require generalizability, as Rothman notes.23
Although patient factors are more important than practice factors in determining immunization rates, practices can affect patients by establishing an environment that promotes adult immunization through patient education, provider recommendation, and easy access to vaccination (immunization clinics).
From the Department of Family Medicine and Clinical Epidemiology, University of Pittsburgh School of Medicine (MPN, IJB, RKZ, SS), and the Department of Behavioral and Community Health Sciences, University of Pittsburgh Graduate School of Public Health (RKZ), Pittsburgh, Pa.
This project/publication was funded by grant HS09874-01A1 from the Agency for Healthcare Research and Quality, Rockville, Md.
Address correspondence to: Mary Patricia Nowalk, PhD, Department of Family Medicine and Clinical Epidemiology, University of Pittsburgh School of Medicine, 3815 Fifth Ave, Pittsburgh, PA 15261. E-mail: email@example.com.
1. Thompson W, Shay D, Weintraub E, Brammer L, Anderson L, Fukuda K. Mortality associated with influenza and respiratory syncytial virus in the United States. . 2003;289:179-186.
Healthy People 2010: Understanding and Improving Health and Objectives for Improving Health.
2. US Department of Health and Human Services. 2nd ed. Washington, DC: US Government Printing Office; 2000.
Early Release of Selected Estimates Based on Data From the January-September National Health Interview Survey (NHIS).
3. National Center for Health Statistics. Atlanta, Ga: Centers for Disease Control and Prevention; 2003.
MMWR Morb Mortal Wkly Rep.
4. Centers for Disease Control and Prevention. Vaccine-preventable diseases: improving vaccination coverage in children, adolescents, and adults: a report on recommendations of the Task Force on Community Preventive Services. 1999;48:1-15.
Can J Public Health.
5. Gyorkos TW, Tannenbaum TN, Abrahamowicz M, et al. Evaluation of the effectiveness of immunization delivery methods. 1994;85(suppl):S14-S30.
Am J Med
6. Zimmerman RK, Santibanez TA, Janosky JE, et al. What affects older patientsÂ¡Â¡Â¡Â¡Ã¬Â¬Ë† influenza vaccination behavior? an analysis from inner-city, suburban, rural, and Veterans Affairs practices. . 2003;114:31-38.
J Fam Pract.
7. Zimmerman RK, Silverman M, Janosky JE, et al. A comprehensive investigation of barriers to adult immunization: a methods paper. 2001;50:703.
J Soc Psychol.
8. Landis D, Triandis HC, Adamopoulos J. Habit and behavioral intentions as predictors of social behavior. 1978;106:227-237.
9. Montano DE. Predicting and understanding influenza vaccination behavior: alternatives to the health belief model. 1986;24:438-453.
Int J Psychol.
10. Davidson AR, Jaccard JJ, Triandis HC, Morales ML, Diaz-Guerrero R. Cross-cultural model testing: toward a solution of the etic-emic dilemma. 1976;11:1-13.
J Behav Med
11. Valois P, Desharnais R, Godin G. A comparison of the Fishbein and Ajzen and the Triandis attitudinal models for the prediction of exercise intention and behavior. . 1988;11: 459-472.
12. Zimmerman RK, Santibanez TA, Fine MJ, et al. Barriers and facilitators of pneumococcal vaccination among the elderly. 2003;21:1510-1517.
Am J Med.
13. Nichol KL. Ten-year durability and success of an organized program to increase influenza and pneumococcal vaccination rates among high-risk adults. 1998;105:385-392.
Am J Med.
14. Nichol KL, Korn JE, Margolis KL, Poland GA, Petzel RA, Lofgren RP. Achieving the national health objective for influenza immunization: success of an institution-wide vaccination program. 1990;89:156-160.
Am J Med.
15. Wee CC, Phillips RS, Burstin HR, et al. Influence of financial productivity incentives on the use of preventive care. 2001;110:181-187.
16. Nichol K, Goodman M. Cost effectiveness of influenza vaccination for healthy persons between ages 65 and 74 years. 2002;20:S21-S24.
17. Sisk JE, Moskowitz AJ, Whang, W, et al. Cost effectiveness of vaccination against pneumococcal bacteremia among elderly people. 1997;278:1333-1339.
Arch Fam Med.
18. Dickey LL, Kamerow DB. Primary care physicians' use of office resources in the provision of preventive care. 1996;5:399-404.
19. Centers for Disease Control and Prevention. Behavioral risk factor surveillance survey. Available at: http://www.cdc.gov/ BRFSS. Accessed October 27, 2003.
Am J Prev Med.
20. MacDonald R, Baken L, Nelson A, Nichol KL. Validation of self-report of influenza and pneumococcal vaccination status in elderly outpatients. 1999;16:173-177.
Am J Prev Med.
21. Nichol KL, Korn JE, Baum P. Estimation of outpatient risk characteristics and influenza vaccination status: validation of a self-administered questionnaire. 1991;7:199-203.
Can J Public Health.
22. Hutchison BG. Measurement of influenza vaccination status of the elderly by mailed questionnaire: response rate, validity, and cost. 1989;80:271-275.
23. Rothman KJ. Objectives of epidemiologic study design. In: Boston, Mass: Little Brown & Co Inc; 1986:77-97.