Impact of two different models on influenza and pneumococcal vaccination in hospitalized patients.

Objective: The Centers for Disease Control (CDC) recommends trivalent influenza (TIV) and pneumococcal (PPV) vaccination for eligible hospitalized patients. We conducted a retrospective study comparing two different methods of assessment and its impact on TIV and PPV vaccination in hospitalized patients.

Design: Two sequential models were compared in a community hospital for assessing patients' vaccination status and eligibility for TIV and PPV in an inpatient setting. In the first model (Model I), physicians were responsible for assessing eligibility and ordering TIV and/or PPV. In the second model (Model II), nurses were responsible for assessment and vaccination of eligible patients. Charts of hospitalized patients were randomly analyzed for completion rates of the assessment form, rate of vaccination, and documentation of the reason for not vaccinating eligible patients.

Results: A total of 138 charts were analyzed for Model I and 168 charts were analyzed for Model II. A significantly higher completion rate for assessment was noted for Model II compared with Model I (79.16% versus 34.78%, P < 0.001 for TIV; 72.02% versus 33.33%, P < 0.001 for PPV). Hospital vaccination rates were not significantly different between the two models for TIV (P = 0.625) and PPV (P = 0.689). A significant percentage of patients refused PPV in Model II [8.03% versus Model I at 3.2% (P = 0.04)].

Conclusion: A standing order protocol for assessing hospitalized patients' vaccination status by nursing staff and allowing them to vaccinate eligible patients without depending upon a physician order significantly improved the assessment compliance rate, but not the vaccination rates. Our findings suggest that a system-based method with nursing assessing the need for vaccination and a physician ordering the vaccination would probably have a higher potential for success.

Key Words: influenza, pneumonia, vaccination, hospitalized patients.

**********

Influenza and pneumococcal pneumonia are serious infections associated with considerable morbidity and mortality. Each year influenza accounts for 114,000 hospitalizations and 36,000 deaths and pneumococcal infection accounts for approximately 60,000 cases of invasive disease with 6,000 to 7,000 deaths in the United States. (1) The economic burden of these two vaccine-preventable diseases is staggering. Hospitalization represents an opportunity to identify and immunize people not vaccinated against both diseases to achieve the Centers for Disease Control's "Healthy People 2010" goal of achieving 90% compliance with influenza and pneumococcal vaccines by the year 2010. (2)

One of the reasons for poor vaccination rates is a lack of effective assessment methods of vaccination status among hospitalized patients. New York State promulgated a regulation empowering nursing staff to assess and administer vaccinations to hospitalized patients without a physician's order, known as the standing order for immunization. (3,4) This standing order was introduced in our hospital in the year 2003. Before the use of standing orders, physicians were required to assess patients' immunization status and write an order for an influenza and/or pneumococcal vaccination. After the implementation of standing orders in the year 2003, nurses were given the responsibility to assess and vaccinate eligible patients. This study compares the effectiveness of the physician-driven model versus the nurse-driven model on trivalent influenza vaccination (TIV) and pneumococcal vaccination (PPV) rates in hospitalized patients.

Methods

This study was done at Unity Health System, a community hospital in New York State, which serves a suburban population of over 100,000. The acute care inpatient population is predominantly over 50 years of age, admitted for various chronic illnesses. The institutional review board approved the study. A list of patients discharged from the hospital for the months of December 2002 and December 2003 were obtained. December was chosen because the vaccination rates are higher during this month in our hospital. The discharge list was used because the patients were usually vaccinated on the day of discharge. December 2002 represented the physician-driven model (Model I), where physicians were responsible for assessing the influenza and pneumococcal vaccination status of patients and writing orders based on their assessment, whereas December 2003 represented the nurse-driven model (Model II), where nurses were given the authority to complete an assessment form and vaccinate patients based on their assessment.

A total of 1400 and 1674 patients were discharged in December 2002 and December 2003, respectively. From this sample, every 10th chart was selected to ensure randomization. Information was obtained by reviewing the standardized immunization assessment form in the charts. Information collected for both models included the completion rate for vaccination status assessment (assessment rate), the percentage of patients eligible for vaccination (eligibility rate), the percentage of eligible patients who received vaccination (vaccination rate), and the reason for not administering the vaccine to eligible patients. We assumed that patients from long-term care facilities had been vaccinated; and these patients were documented as having been previously vaccinated. The data was collected and analyzed for both influenza vaccination and pneumococcal vaccination. The Centers for Disease Control (CDC) eligibility criteria for vaccination was used for both influenza and pneumococcal vaccinations. (5,6)

Statistical Analysis

Statistical analyses were performed using the Statistical Program for Social Sciences (version 12.0 SPSS Inc., Chicago, IL). A frequency analysis was done for analyzing the distribution of demographic variables in both models. A [chi square] test was used to detect if the models differed in categorical variables. (7) Fisher's exact test was used if the cell count was less than 5. A P value of <0.05 was considered to be statistically significant. With an estimated vaccination rate in the nursing model between 22 to 25% (8) and in the physician model of 5%, (8) we calculated an estimated sample size of 88 in each arm to detect a difference in vaccination rates between the two models with an [alpha] 0.05 and [beta] 0.02. (9)

Results

One hundred and thirty-eight patients were included for Model I and 168 patients were included for Model II. The demographic characteristics of the patient population are listed in Table 1. The elderly population ([greater than or equal to]65 yr of age) constituted 53.62% and 63.10% of the study population in models I and II respectively. The gender ratio in both groups was similar. The majority of patients were admitted under the medical service in both models (68.84% in model I and 81% in model II).

Influenza Vaccination (TIV)

Model I. The assessment form was completed in 48 patients (34.78%). Based on age and comorbidity, 9 patients were eligible to receive TIV. However, only 2 patients (22.2%) received TIV while in the hospital. The reason for not vaccinating the remaining patients was undocumented in 4 cases, (10%), and was due to patient refusal in 3 patients, (7.5%). Thirty-three patients (82.5%) had vaccination before hospitalization. (Table 2).

Model II. The assessment form was completed in 133 patients (79.16%). Twenty patients were eligible to receive TIV. However, only 4 patients (20%) were vaccinated in the hospital. The reason for not vaccinating the remaining 16 patients included patient refusal (11 patients, 9.2%) and undocumented (5 patients, 4.2%). One hundred and four patients were vaccinated with TIV before admission. The in-hospital vaccination rate and refusal to receive TIV in the hospital did not differ significantly in the two models (P = 0.625 and P = 0.373 respectively).

Pneumococcal Vaccination (PPV)

Model I. The assessment form was completed in 46 patients (33.33%). Ten patients (21.7%) were eligible to receive PPV; however, only 2 patients (20%) received PPV in the hospital. Twenty-three patients were vaccinated before hospitalization. The reason for not vaccinating the remaining patients included patient refusal (1 patient, 3.2%) and undocumented (7 patients, 22.6%) (Table 2).

Model II. The assessment form was completed in 121 patients (72.02%). Twenty-two patients (18.2%) were eligible to receive PPV; however, only 5 patients (18.2%) received PPV in the hospital. The reason for not vaccinating the remaining 17 patients included patient refusal (9 patients (8.4%) and undocumented (8 patients, 7.5%). Ninety patients had vaccination before hospitalization. In-hospital vaccination rates were not significantly different (P = 0.627) but the refusal rate was significantly higher in model II (P = 0.04).

Discussion

Vaccination is pivotal to limiting morbidity and mortality from influenza and pneumococcal infections. (10) The benefits of administering both influenza and pneumococcal vaccines during influenza seasons have been shown to be additive in elderly persons with chronic lung disease. (11) The cost-effectiveness of the influenza vaccination is largely unquestioned as it has been shown to reduce hospitalizations for pneumonia and influenza, other respiratory conditions, congestive heart failure and to reduce mortality from all causes. (12) The efficacy of the pneumococcal vaccination in the elderly has been a subject of controversy until recently. A retrospective cohort study of elderly patients with chronic lung disease showed the pneumococcal vaccine to be associated with a 43% reduction in the number of hospitalizations for pneumonia and a 29% reduction in the risk of death from all causes. (13) In addition, it is associated with decreased rates of bacteremia and cost savings. (12) Standing orders authorize nurses and pharmacists to administer vaccinations according to an institution and/or physician-approved protocol without requiring additional assessment by the physician. Some studies have shown increased vaccination rates in hospitalized patients and in the outpatient setting by using standing orders and organizational strategies. (8,13-22) In our study, the introduction of standing orders improved the assessment rate for both influenza and pneumococcal vaccine. Even though a higher assessment rate by nurses could be due to more education before the introduction of the protocol, the differences were almost twice that of the physician model, proving a significant difference.

However, the differences in the assessment rate did not translate into an increased vaccination rate of eligible hospitalized patients. Several factors may have contributed to this outcome. The vaccination status of patients was assessed at the time of admission, but vaccination was deferred until the day of discharge so that the patients were well enough to receive the vaccine. Because of this practice, a nurse who was unaware of the vaccination recommendation may have discharged the patient without providing the vaccination. Moreover, during busy periods when there is high hospital turnover and increased pressure to discharge patients, some of the eligible patients may have been discharged without being vaccinated. In some cases, a delay at the pharmacy may have resulted in a patient leaving the hospital before vaccination.

Another significant finding of this study was a higher rate of refusal among eligible hospitalized patients in model II. One can argue that the physician's involvement in model I led to more patient satisfaction and discouraged refusal. Active physician involvement in recommending vaccination has the potential to be effective. (10) Physician involvement in patient education may help improve the patient's understanding about vaccination, and may reduce the number of refusals. In our study, a higher percentage of patients were vaccinated when a physician was involved (although the differences were not statistically significant). We also noted that a higher percentage of patients were vaccinated before coming to the hospital in model II. This increased percentage reached statistical significance for PPV, which reflects increasing awareness among healthcare professionals about PPV.

Reason for not vaccinating was not documented in a significant number of those eligible for vaccination during hospitalization; these patients probably represent missed opportunities to vaccinate hospitalized patients. The need for ongoing education of healthcare professionals is crucial to the success of all vaccination programs. It is worth mentioning that inadvertent revaccination with pneumococcal vaccine three or more years following initial vaccination has been found to be safe and not associated with greater rates of serious local or systemic adverse reactions. (23)

Our study had certain limitations. We assessed patients' eligibility based on the standard immunization form. We may have found more eligible patients if we had looked for the discharge diagnosis and medical history in the charts, especially for patients with high-risk conditions. Secondly, we assumed that patients from the long-term care facilities had been vaccinated; we documented such patients as being previously vaccinated. This may have caused an inflated rate of patients assumed to be vaccinated. However, when the data was analyzed, including those patients who had already been vaccinated, the results were the same. Thirdly, we did not assess the impact of readmitted patients in either model.

Furthermore, our study was retrospective, involving a collection of data from two different years for each model. A prospective study design would have been the ideal method to compare the two groups during the same time period, as this would reduce differences between the groups. There is increased awareness about vaccination needs secondary to campaigns by the CDC and Centers for Medicare and Medicaid Services (CMMS). (24) CMMS, along with the CDC and other outside organizations, including the National Coalition on Adult Immunization, works to promote influenza and pneumococcal immunizations to Medicare beneficiaries. In addition to its national immunization campaign, CMMS has implemented quality improvement efforts to improve rates in a variety of healthcare settings, including hospitals and nursing homes. Some of these campaigns might have led to increased awareness in the nurse's model. Another potential limitation is that only 4.2% of the overall sample was vaccinated. The differences between the two models were derived from a small outcome, and therefore should be interpreted with caution.

In conclusion, a standing order empowering nurses increased the assessment rate for vaccination status among hospitalized patients, but did not translate into an increased rate of vaccination. On the other hand, when a physician was involved, a higher percentage of patients were vaccinated. A system-based solution involving both nurses and physicians with nurses completing the assessment and physicians ordering the vaccination has a better potential for higher vaccination rates among hospitalized patients.

Acknowledgments

The we thank Ms. Katie O'Leary, RN, BSN, CPHQ, Quality Supervisor, Unity Health System for her help with the study.

References

1. Norwalk MP, Middleton DB, Zimmerman RK, et al. Increasing pneumococcal vaccination rates among hospitalized patients. Infect Control Hosp Epidemiol 2003;24:526-531.

2. Healthy People 2010. Leading health Indicators: United States selected years 1990-2002 and 2010 target. Available at: http://www.cdc.gov/nchs/data/hus/tables/2003/03hus051.pdf. Accessed December 15, 2006.

3. Non-Patient Specific Standing Order and Protocol Guidelines. New York State Education Department. Available at: http://www.op.nysed.gov/immunguide.htm. Accessed December 4, 2006.

4. Use of Standing Orders Programs to Increase Adult Vaccination Rates. Recommendations of the Advisory Committee on Immunization Practices. Available at: http://www.cdc.gov/mmwr/PDF/RR/RR4901.pdf. Accessed December 15, 2006.

5. Centers for Disease Control and Prevention. Prevention and control of Influenza. Recommendations of the Advisory Committee on Immunization Practices. Morbidity and Mortality Weekly Report April 12 2002;51(RR-03):1-31.

6. Centers for Disease Control and Prevention. Prevention of Pneumococcal Disease. Recommendations of the Advisory Committee on Immunization Practices. Available at: http://www.cdc.gov/mmwr/preview/mmwrhtml/00047135.htm. Accessed December 15, 2006.

7. Agresti A. Statistical Methods for the Social Sciences. 3rd ed. New York, Prentice Hall, 1997, pp 196-212.

8. Rhew DC, Glassman PA, Goetz MB. Improving pneumococcal vaccine rates: nurse protocols versus clinical reminders. J Gen Intern Med 1999;14:351-356.

9. Hulley SB, Cummings SR, Browner WS, et al. Designing Clinical Research, 2nd edition. Philadelphia, Lippincott, Williams, and Wilkins, 2001, pp 65-91.

10. Fedson DS, Houck P, Bratzler D. Hospital-based influenza and pneumococcal vaccination: Sutton's law applied to prevention. Infect Control Hosp Epidemiol 2000;21:692-699.

11. Nichol KL. The additive benefits of influenza and pneumococcal vaccinations during influenza seasons among elderly persons with chronic lung disease. Vaccine 1991;17(Suppl):S91-4, S93.

12. Sisk JE, Moskowitz AJ, Whang W, et al. Cost effectiveness of vaccination against pneumococcal bacteremia among elderly people. JAMA 1997;278:1333-1339.

13. Nichol KL, Baken L, Wourenma J, et al. The health and economic benefit associated with pneumococcal vaccination of elderly persons with chronic lung disease. Arch Intern Med 1999;159:2437-2442.

14. Nichol KL. Tenyear durability and success of an organized program to increase influenza and pneumococcal vaccination rates among high-risk adults. Am J Med 1998;105:385-392.

15. Klein RS, Adachi N. An effective hospital based pneumococcal immunization program. Arch Intern Med 1986;146:327-329.

16. Landis S, Scarbrough ML. Using a vaccine manager to enhance in-hospital vaccine administration. J Fam Pract 1995;41:364-369.

17. Slobodkin D, Kitlas JL, Zielske PG. A test of the feasibility of pneumococcal vaccination in the emergency department. Acad Emerg Med 1999;6:724-727.

18. Herman CJ, Speroff T, Cebul RD. Improving compliance with immunization in the older adult: results of a randomized cohort study. J Am Geriatr Soc 1994;42:1154-1159.

19. Hoey JR, McCallum HP, Lepage EM. Expanding the nurse's role to improve preventive service in an outpatient clinic. Can Med Assoc J 1982;127:27-28.

20. Christy C, McConnochie KM, Zernik N, et al. Impact of an algorithm-guided nurse intervention on the use of immunization opportunities. Arch Pediatr Adolesc Med 1997;151:384-391.

21. Lawson F, Baker V, Au D, et al. Standing orders for influenza vaccination increased vaccination rates in inpatient settings compared with community rates. J Gerontol A Biol Sci Med Sci 2000;55:M522-M526.

22. Dexter PR, Perkins SM, Maharry KS, et al. Inpatient computer-based standing orders versus physician reminders to increase influenza and pneumococcal vaccination rates: a randomized trial. JAMA 2004;292:2366-2371.

23. Jackson LA, Benson P, Sneller VP, et al. Safety of revaccination with pneumococcal polysaccharide vaccine. JAMA 1999;281:243-248.

24. CMS Initiatives. Available at: http://www.cms.hhs.gov/Adultlmmunizations/05_CMS%20Initiatives.asp. Accessed December 4, 2006.

Mobolaji Bakare, MD, Rakesh Shrivastava, MD, Vinodh Jeevanantham, MD, MPH, and Sankar D. Navaneethan, MD, MPH

From the Department of Medicine, Unity Health System, Rochester, NY.

Reprint requests to Sankar D. Navaneethan, MD, MPH, Department of Medicine, Unity Health System, 1555 Long Pond Road, Rochester, NY 14626. Email: [e-mail]

This study was presented at the Southern Medical Association's annual meeting in New Orleans in 2004 by Dr. Mobolaji Bakare.

Conception and design of the study: Rakesh Shrivastava, Mobolaji Bakare, Sankar D. Navaneethan; Data Collection: Sankar D. Navaneethan, Mobolaji Bakare; Statistical Analysis: Vinodh Jeevanantham; Writing the Manuscript: Rakesh Shrivastava, Mobolaji Bakare, Vinodh Jeevanatham, Sankar D. Navaneethan.

Accepted August 10, 2006.

RELATED ARTICLE: Key Points

* Nursing assessment rates on the need for influenza and pneumococcal vaccination is better than physician assessment rates in hospitalized patients.

* Physician involvement in the ordering process would increase vaccination rates.

* A system-based method with nursing assessing the need for vaccination and physicians ordering the vaccination would increase both the assessment and vaccination rates for influenza and pneumococcal vaccinations.

Table 1. Basic demographic characteristics of the patients included in
the study

                                Model I (n = 138)    Model II (n = 168)
Variables                       No. of patients (%)  No. of patients (%)

Age groups
  <50                           38 (27.5%)            27 (16.0%)
  50-64                         26 (18.8%)            35 (20.8%)
  [greater than or equal to]65  74 (53.6%)           106 (63.1%)
Gender
  Male                          67 (48.5%)            80 (47.6%)
  Female                        71 (51.4%)            88 (52.3%)
Service
  Medical                       95 (68.8%)           135 (81%)
  Nonmedical                    43 (31.1%)            33 (19%)

Table 2. Results of comparison of two models on the assessment form
completion rate and reasons for not administering the influenza and
pneumococcal vaccine

                                 Pneumococcal vaccine (PPV)
                                 Model I       Model II
                                 (n = 138)     (n = 168)
                                 No. of        No. of
                                 patients (%)  patients (%)  P value

Assessment form completed        46 (33.3%)    121 (72.02%)  <0.001 (a)
Eligible for vaccination         10 (21.7%)     22 (18.2%)    0.174 (b)
In-hospital vaccination rates     2 (20%)        5 (18.2%)    0.627 (b)
Reason for not administering
  PPV/TIV
  Previously vaccinated          23 (74.2%)     90 (84.1%)    0.04 (a)
  Patients refusing vaccination   1 (3.2%)       9 (8.4%)     0.04 (a)
  Undocumented                    7 (22.6%)      8 (7.5%)     0.04 (a)

                                 Influenza vaccine (TIV)
                                 Model I       Model II
                                 (n = 138)     (n = 168)
                                 No. of        No. of
                                 patients (%)  patients (%)  P value

Assessment form completed        48 (34.7%)    133 (79.2%)   <0.001 (a)
Eligible for vaccination          9 (18.7%)     20 (15%)      0.74 (b)
In-hospital vaccination rates     2 (22.2%)      4 (20%)      0.625 (b)
Reason for not administering
  PPV/TIV
  Previously vaccinated          33 (82.5%)    104 (86.7%)    0.373 (a)
  Patients refusing vaccination   3 (7.5%)      11 (9.2%)     0.373 (a)
  Undocumented                    4 (10%)        5 (4.1%)     0.373 (a)

(a) P value derived from Chi-square test statistic.
(b) P value derived from Fisher exact test.