Can a universal influenza immunization program reduce emergency department volume?

EM Advances

CJEM 2002;4(4):245-251

Abstract

Résumé

Introduction | Methods | Results | Discussion | Conclusions | References

Introduction

In the winter of 1999/2000 the province of Ontario suffered a severe outbreak of influenza, and news reports linked this outbreak to concurrent emergency department (ED) overcrowding.¹ In July 2000 the Ontario Ministry of Health and Long-Term Care announced a universal influenza immunization program for all Ontario residents.² Through this program all 11.5 million Ontarians were eligible to receive publicly funded influenza vaccine annually, starting in October of 2000. In October 2001, the MOHLTC announced a second year of the universal campaign, at a cost of $44 million.³ The program's 2 stated objectives are to decrease the impact of influenza on ED visits, and to decrease the number and severity of influenza cases in Ontario.

A MEDLINE search using the terms "influenza" and "ED volume," "ED overcrowding," "emergency room (ER) volume," "ER overcrowding" and the related articles, revealed no published research on the direct impact of influenza cases on ED volume, although research on other causes of ED overcrowding were found. Because reducing ED volume was one of the universal campaign's 2 objectives, we conducted a study to determine the relationship between population influenza rates and ED patient visits.

Our primary objective was to determine whether population influenza rates are associated with increasing ED volumes. Secondary objectives were to describe temporal patterns of ED utilization at the 3 study sites (5 hospitals), to estimate the number of cases of influenza and acute upper respiratory (AUR) disease that are diagnosed in the Kingston EDs and to determine the likely impact of influenza and AUR diseases on ED volume based on their numbers relative to all ED visits.

Methods

In this retrospective observational study, we obtained the total number of ED admissions from 5 hospitals in 3 Ontario cities between Nov. 1, 1996, and Mar. 31, 2001. ED visits were converted to age-standardized rates per 100 000 population and correlated with provincial influenza cases reported by Health Canada.

Data collection

ED data were obtained from the medical records departments of the Ottawa Hospital's Civic Campus in Ottawa, the London Health Sciences Centre's South St. and University campuses in London, and the Kingston General (KGH) and Hotel Dieu hospitals in Kingston, which are large, tertiary care centres representing eastern, western and southern Ontario. All sites provided the total number of ED visits, by month, between Nov. 1, 1996, and Mar. 31, 2001. The 2 Kingston hospitals also provided the number of influenza cases as a percentage of all respiratory diagnoses in adults (over age 18) seen in the ED. This site was chosen because it had the highest volume and was a convenient sample to obtain.

ED respiratory diagnoses

All patient visits were coded by ED discharge diagnoses (ICD-9 codes) and entered into a central database. Medical records staff provided the researchers with the ED diagnoses for all influenza season patient visits for the years 1996/97 to 2000/01. Visits within that time frame with the discharge diagnosis of AUR disease, pneumonia and influenza-like illness (ILI) or influenza (ICD-9 codes 460-466, 480-487) were identified by ICD-9 code, age and gender. We did not limit our focus to ICD-9 codes 487-487.8 (influenza and ILI) because this would have underestimated influenza rates. Many upper respiratory conditions have similar symptoms to influenza and ILI, and in the absence of confirmatory testing, emergency physicians may apply nonspecific respiratory diagnoses to patients with influenza. Consequently, this study and others^4,5 have used AUR illnesses and pneumonia, (ICD-9 codes 460-466,480-486) as a proxy measure for influenza and ILI.

Population influenza data

Provincial and national influenza data were obtained from Health Canada's Canada Communicable Disease Reports (CCDR),^5-9 which are based on 3 types of data collection. First, laboratories participating in the Canadian case-by-case surveillance program are required to report confirmed influenza cases to the Division of Disease Surveillance at the Centre for Infectious Diseases Prevention and Control (CIDPC). Second, ILI is reported by sentinel physicians on a weekly basis, and third, influenza activity levels are assessed by the provincial and territorial epidemiologists. These data are compiled and published in the CCDR.

Confirmed influenza is a reportable disease in Ontario, and the reporting definition is the presence of 3 symptoms -- fever greater than 39°C; cough or sore throat; and myalgia, malaise or prostration -- as well as at least 1 of the following: laboratory confirmation by detection or isolation of influenza virus in pharyngeal or nasal secretion OR a 4-fold increase in hemagglutination antibody titers to influenza between acute and convalescent sera. This study used laboratory-confirmed cases of influenza reported to the CIDPC and published in CCDR because these data were consistently available for the entire 5-year study period.

In the United States, all laboratory-confirmed cases of influenza are collated by the Centers for Disease Control and Prevention and reported in the Morbidity and Mortality Weekly Reports.^10-15 These data were collected for comparison purposes in this study.

Analysis

ED volumes were aggregated by city. The ED patient visits were then transformed into age-standardized rates per 100 000 population (using 1996 Statistics Canada census data) in order to compare them to provincial influenza rates and to each other. Descriptive statistics, including counts, percentages, means and frequencies, were used to describe ED diagnoses.

Pearson correlation coefficients were used to identify significant relationships between ED volume and influenza cases, and (in Kingston) relationships between ED volume and upper respiratory diagnoses. Multiple linear regression was used to determine how much of the variance in ED volume could be accounted for by influenza, pneumonia and other upper respiratory conditions, and to see whether influenza rates were significantly associated with ED volume, after controlling for other upper respiratory diseases. All analyses were completed using Excel spreadsheets and SPSS version 10.1 (Chicago, Ill.).

Ethical considerations

This protocol received expedited ethical approval by the Queen's University Health Sciences and Affiliated Teaching Hospitals Research Ethics Board because the data collected from the hospitals (ED volume) were often publicly reported. At no time were the investigators aware of the identity of specific patients involved.

Results

Figure 1 shows reported rates of influenza cases per 100 000 population per year in Canada, the USA, and in Ontario, from the influenza seasons of 1996/97 to 2000/01. Annual fluctuations, as well as the decrease in influenza rates from the 1999/2000 season to the 2000/01 season are apparent.

Fig. 1. Laboratory confirmed influenza cases in Ontario, Canada and the USA, rates per 100 000

Figure 2 illustrates total ED visits per 100 000 population for Kingston, Ottawa and London, between Nov. 1, 1996, and Mar. 31, 2001. Vertical lines indicate peak volume periods. This figure shows that ED volumes follow similar seasonal fluctuations in all of the study regions and that peak ED volumes occured during summer.

Fig. 2. Emergency department (ED) volume (per 100 000 population) from Nov. 1, 1996, to Mar. 31, 2001.

Figure 3 shows mean ED volumes during influenza season only, with the rate of laboratory-confirmed influenza cases for the province of Ontario displayed on the Y-axis. There was no significant correlation between influenza rates and ED volumes, with Pearson correlation coefficients (r) of 0.22 (p = 0.72), 0.33 (p = 0.59), and 0.27 (p = 0.66) at the Kingston, London and Ottawa study sites, respectively.

Fig. 3. Mean ED visit rates per 100 000 population during the influenza season (November 1st to March 31st)

During the 5 influenza seasons studied in 2 Kingston EDs, AUR diagnoses accounted for only 4.4% of all ED visits, and influenza for only 0.34% of visits. Table 1 shows that, based on ED discharge diagnosis, influenza accounted for 7% to 10% of all acute respiratory conditions seen. In Table 2, Pearson correlation coefficients reveal that there are no statistically significant correlations between Kingston hospital ED volume and ED diagnosis of influenza, pneumonia or other upper respiratory disease. However, the diagnosis of influenza correlated significantly with diagnosis of "other AUR," indicating that upper respiratory disease may be a good proxy for influenza.

Table 1. Acute respiratory diagnoses in Kingston emergency departments (1996–2000) Year Colds, % Other AUR* diseases, % Pneumonia, % Influenza, % 1996/97 13.0 53.0 27.0 7.0 1997/98 13.0 44.0 29.0 9.0 1998/99 12.7 44.6 32.9 9.8 1999/2000 12.1 52.5 27.3 9.1 *AUR = acute upper respiratory

Table 2. Correlation* between Kingston ED volume and acute upper respiratory (AUR) illnesses Variable ED volume Influenza Pneumonia Other AUR ED volume r = 1.0       Influenza r = –0.45; p = 0.45† r = 1.0     Pneumonia r = 0.27; p = 0.66 r = 0.54; p = 0.35 r = 1.0   Other AUR r = –0.13; p = 0.83 r = 0.90; p = 0.04 r = 0.57; p = 0.32 r = 1.0 * Correlation (r) is based on Pearson’s product-moment correlation coefficient. † All tests of significance are 2-sided unless otherwise specified, and considered significant at the 0.05 level. n = 5 for all comparisons, reflecting the 5 comparison periods (1996–2000)

Finally, multiple linear regression was used to determine how much of the variance in Kingston ED volumes could be accounted for by influenza, pneumonia and other AUR diagnoses. The regression model determined that all upper respiratory diagnoses combined accounted for 27% of the variance in ED visit volume (r² = 0.27), although this was not statistically significant (p = 0.53). Neither influenza, pneumonia, nor other AUR diseases were significant predictors of ED volume, with influenza showing an inverse (negative) relationship to total ED visit volume.

Discussion

The findings from this study suggest that influenza rates do not correlate significantly with ED volumes and that, although influenza vaccination programs may have substantial public health benefit, they are unlikely to lead to meaningful improvement in ED overcrowding. Our findings also show that respiratory diagnoses in general and influenza in particular account for a small proportion of ED visits, and that the influence of these conditions on ED volume is not statistically significant. We recognize that because of the need for confirmatory testing, influenza may underdiagnosed in the ED; therefore, we included pneumonia and other acute respiratory illnesses in our analysis. But even when all AUR diagnoses were examined, these accounted for only 4.4% of all ED visits during the influenza season, and influenza accounted for only 0.34%.

These findings echo results from the Hospital Report 2001,¹⁶ which states that in Ontario, injuries and symptoms account for 23% and 17% of ED visits respectively, and in adults over 18 years of age respiratory system diseases in their entirety account for approximately 10% of ED visits.¹⁶ Our regression analysis indicated that neither flu, pneumonia nor other AUR diagnoses were significant predictors of Kingston ED volumes.

Figure 3 illustrates the relationship between influenza rates and ED visit rates. Of note, the 2000/01 influenza season was the first season after Ontario's universal influenza vaccination program and the first since 1995/96 in which Influenza A (H3N2) viruses did not predominate.¹⁰ If population influenza rates are an important determinant of ED volumes, we would expect 2000/01 ED volumes to fall along with influenza rates. Figures 2 and 3 show that this did not occur.

Although influenza rates did not influence ED volumes significantly, the season did. In all 5 tertiary care hospitals studied, ED volumes were highest in the summer months (May to August) and lower during influenza season. This finding is supported by research from the Institute for Clinical Evaluative Sciences (ICES),¹⁷ which reported that "there are highly predictable peaks periods in emergency volume on public holidays and weekends. The week between Christmas and New Year's is the busiest week of the year, with volumes rising as much as one-third above average. Summer is the most demanding season in the ED."

Our study also revealed that ED volume has been increasing every year for the past 5 years at all hospital sites. Other researchers have found that from 1993 to 2000 Ontario ED volumes have decreased; however they clarify that these findings may be due to strict inclusion criteria whereby a portion of ED visits were not counted, resulting in numbers that are approximately 15% lower than those reported by hospitals.¹⁷ The same report states that ED volume in the USA rose 14% from 1992 to 1999, and that overcrowding and lower utilization are not necessarily mutually exclusive.¹⁷ The reasons for the difference in findings warrant further research.

In Canada, influenza and pneumonia (as a complication of influenza) are responsible for approximately 6700 deaths and 75 000 hospitalizations each year, with between 500 to 1500 deaths per year occurring in Ontario.¹⁸ Since 1989 Ontario has provided yearly influenza vaccinations for persons at high risk of influenza complications (i.e., people with chronic cardiac or pulmonary disorders, those over the age of 65, residents of long-term care facilities and health services workers). The cost-effectiveness of vaccinating low-risk people remains controversial,^19-30 and although numerous studies have looked at the causes of ED utilization and overcrowding there are no data suggesting influenza causes overcrowding.^31-36 However, although universal vaccination may have little influence on ED volumes, vaccinating individuals at high risk for complications arising from influenza provides an average direct cost savings of $117 per person, in addition to reducing the number of hospitalizations and deaths.^37-40

Limitations

The accuracy of ICD-9-coded ED discharge diagnoses and the validity of using administrative databases in health research are areas of debate.⁴¹ We acknowledge that these are potential limitations of the study but hope that the use of a single site and medical records system reduced variability and error. It is possible, however, that the number of ED visits for influenza and other respiratory diagnoses were underestimated.

The number of reported influenza and ILI cases underestimates the true number of cases in any given year. This is because many people with influenza do not seek medical attention and because, of those who do, many receive an alternate diagnosis (e.g., "viral illness"). Therefore, although the numbers presented in this study are not "actual" numbers of influenza cases in any given year, they represent a consistent proportion of the true number of cases because the method of data collection is constant from year to year. Furthermore, these are the official reporting numbers for the Canadian and US governments, and the numbers upon which health policy decisions are based.

This study does not discuss vaccination coverage rates, which are not systematically collected and thus unavailable. Prior studies have documented coverage rates of 13.8% for low-risk Canadian adults and 44.8% for people over age 65.⁴² The target of the universal campaign was 90% for high-risk groups and 60% for low-risk groups. The Ontario Ministry of Health and Long-Term Care, the Public Health Research Education and Development programs of provincial health units, the Institute for Clinical Evaluative Sciences, and Aventis Pasteur recently completed a telephone survey to assess the success of the universal immunization campaign, but these results have not yet been released. If the campaign increased vaccine coverage rates, but ED volumes failed to decline, this would be direct evidence that influenza vaccination does not reduce ED utilization; however, if vaccine coverage rates did not increase, then the program itself needs to be re-examined.

Another limitation is that, at the time of this study, only 6 months of data were available after the universal immunization campaign, in contrast to 5 years of precampaign data. It is conceivable that, as vaccination coverage increases, future data will show a beneficial effect on ED utilization; however, this seems unlikely given the relatively small proportion of influenza- and respiratory-related ED visits during influenza season and the predictable peaks in the summer months.

Finally, this study was carried out at tertiary care centres only and therefore should not be generalized outside of this population. Furthermore, the actual admission diagnoses were collected from 1 centre only and should not be generalized beyond that centre. However, the primary finding that AUR diagnoses account for less than 10% of diagnoses is supported by other research.¹⁶

Conclusions

In this study, influenza accounted for only 0.34% of ED visits. ED volumes did not correlate with population influenza rates or ED influenza diagnosis rates. During the influenza season after a universal immunization campaign, ED visits increased. While universal influenza immunization may have health benefits for a population, it is unlikely to reduce ED volumes or overcrowding.

References

1. Ontario invests $38 million to ease emergency room pressures with universal vaccination program [press release]. Toronto (ON): Govt of Ontario; 2000 July 25. Available: www1.newswire.ca/government/ontario/english/releases/july2000/25/ (accessed 2002 June 4).
2. Persistent flu plagued Ontario [television program]. "CBC Front Page News." Toronto (ON): Canadian Broadcasting Corporation; 2000 Jan 6.
3. Ontario launches free flu shot program [press release]. Toronto (ON): Govt of Ontario; 2001 Oct 5. Available: www1.newswire.ca/government/ontario/english/releases/October2001/05/ (accessed 2002 June 4).
4. National Advisory Committee on Immunization. Statement on influenza vaccination for the 1997-1998 season. Can Commun Dis Rep 1997;23:July 1.
5. National Advisory Committee on Immunization. Statement on influenza vaccination for the 1996-1997 season. Can Commun Dis Rep 1996;22:June 15.
6. National Advisory Committee on Immunization. Statement on influenza vaccination for the 2001-2002 season. Can Commun Dis Rep 2001;27:Aug 1.
7. National Advisory Committee on Immunization. Statement on influenza vaccination for the 1999-2000 season. Can Commun Dis Rep 1999;25:June 1.
8. National Advisory Committee on Immunization. Statement on influenza vaccination for the 1998-1999 season. Can Commun Dis Rep 1998;24:July 1.
9. Influenza in Canada: 1995-1996 season. Can Commun Dis Rep 1996;22-23:Dec 1.
10. World Health Organization. Influenza activity: United States and worldwide, 1999-2000 season, and composition of the 2000-01 influenza vaccine. Morbid Mortal Wkly Rep 2000;49:375-81.
11. World Health Organization. Influenza activity: United States, 1999-2000 season. Morbid Mortal Wkly Rep 2000;49:53-7.
12. World Health Organization. Influenza activity: United States, 1998-1999 season. Morbid Mortal Wkly Rep 1999;48:177-81.
13. World Health Organization. Influenza activity: United States and worldwide, 1997-98 season, and composition of the 1998-99 influenza vaccine. Morbid Mortal Wkly Rep 1998;47:280-4.
14. World Health Organization. Influenza activity: United States, 1996-97 season. Morbid Mortal Wkly Rep 1997;46:325-30.
15. World Health Organization. Influenza activity: United States, 1995-96 season. Morbid Mortal Wkly Rep 1996;45:326-30.
16. Hospital Report Research Collaborative. Hospital report 2001: emergency department care. Toronto (ON): Ont Hospital Association and the Govt of Ontario, editors; Dec 2001. Available: www.hospitalreport.ca/HospitalReport2001EDReport.htm
    (accessed 2002 June 4).
17. Chan BTB, Schull MJ, Schultz SE. Atlas Report. Emergency department services in Ontario: 1993-2000. Toronto (ON): Institute for Clinical Evaluative Sciences; 2001. Available: www.ices.on.ca/PDFs/AtlasReports/EDreport.pdf (accessed 2002 June 4).
18. Russell ML, Maxwell CJ. The prevalence and correlates of influenza vaccination among a home care population. Can J Public Health 2000;91:441-4.
19. Buxton Bridges C, Thompson WW, Meltzer MI. Effectiveness and cost-benefit of influenza vaccination of healthy working adults: a randomized controlled trial. JAMA 2000; 284:1655-63.
20. Monto AS. Preventing influenza in healthy adults -- the evolving story. JAMA 2000;284:1699-700.
21. Patriarca PA, Strikas RA. Influenza vaccine for healthy adults? N Engl J Med 1995;333:933-4.
22. Schabas RE. Mass vaccination in Ontario: a sensible move. CMAJ 2001;164:36-7.
23. Demicheli V. Mass influenza vaccination in Ontario: Is it worthwhile? CMAJ 2001;164:38-9.
24. Demicheli V, Jefferson T, Rivetti D, Deeks J. Prevention and early treatment of influenza in healthy adults. Vaccine 2000;18:957-1030.
25. Nichol KL. Cost-benefit analysis of a strategy to vaccinate healthy working adults against influenza. Arch Intern Med 2001;161:749-59.
26. Kumpulainen V, Makela M. Influenza vaccination among healthy employees: a cost-benefit analysis. Scand J Infect Dis 1997;29:181-5.
27. Hurwitz ES, Haber M, Chang A, Shope T, Teo S, Ginsberg M, et al. Effectiveness of influenza vaccination of day care children in reducing influenza-related morbidity among household contacts. JAMA 2000;284:1677-82.
28. Reichert TA, Sugaya N, Fedson DS, Glezen WP, Simonsen L, Tashiro M. The Japanese experience with vaccinating schoolchildren against influenza. N Engl J Med 2001;344:889-96.
29. Wright P. Influenza in the family. N Engl J Med 2000;343:1331-2.
30. McIntosh K, Lieu T. Is it time to give influenza vaccine to healthy infants? N Engl J Med 2000;342:275-6.
31. Sokhandan M, McFadden ER Jr, Huang YT, Mazanec MB. The contribution of respiratory viruses to severe exacerbations of asthma in adults. Chest 1995;107:1570-4.
32. Kelen GD, Scheulen JJ, Hill PM. Effect of an emergency department (ED) managed acute care unit on ED overcrowding and emergency medical services diversion. Acad Emerg Med 2001;8:1095-100.
33. Velianoff GD. Overcrowding and diversion in the emergency department: the health care safety net unravels. Nurs Clin North Am 2002;37:59-66.
34. Baer RB, Pasternack JS, Zwemer FL Jr. Recently discharged inpatients as a source of emergency department overcrowding. Acad Emerg Med 2001;8:1091-4.
35. Schneider S, Zwemer F, Doniger A, Dick R, Czapranski T, Davis E. Rochester, New York: a decade of emergency department overcrowding. Acad Emerg Med 2001;8:1044-50.
36. Schull MJ, Szalai JP, Schwartz B, Redelmeier DA. Emergency department overcrowding following systematic hospital restructuring: trends at twenty hospitals over ten years. Acad Emerg Med 2001;8:1037-43.
37. Foster DA, Talsma A, Furumoto-Dawson A, Ohmit SE, Margulies JR, Arden NH, et al. Influenza vaccine effectiveness in preventing hospitalization for pneumonia in the elderly. Am J Epidemiol 1992;136:296-307.
38. Gross PA, Hermogenes AW, Sacks HS, Lau J, Levandowski RA. The efficacy of influenza vaccine in elderly persons: a meta-analysis and review of the literature. Ann Intern Med 1995;123:518-27.
39. Govaert TM, Thijs CT, Masurel N, Sprenger MJ, Dinant GJ, Knottnerus JA. The efficacy on influenza vaccination in elderly individuals: a randomized double-blind placebo-controlled trial. JAMA 1994;272:1661-5.
40. Nichol KL, Lind A, Margolis KL, Von Sternberg T. The efficacy and cost effectiveness of vaccination against influenza among elderly persons living in the community. N Engl J Med 1994;331:778-84.
41. Iezzoni LI. Risk adjustment for measuring healthcare outcomes. 2nd ed. Chicago (IL): Health Administration Press; 1997.
42. Duclos P, Hatcher J. Epidemiology of influenza vaccination in Canada. Can J Public Health 1993;84:311-5.