SEARCH

SEARCH BY CITATION

Keywords:

  • Economics;
  • hospital;
  • infection control;
  • interventions;
  • norovirus;
  • outbreak

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Transparency Declaration
  9. References

Clin Microbiol Infect 2011; 17: 640–646

Abstract

Although norovirus is a significant cause of nosocomial viral gastroenteritis, the economic value of hospital outbreak containment measures following identification of a norovirus case is currently unknown. We developed computer simulation models to determine the potential cost-savings from the hospital perspective of implementing the following norovirus outbreak control interventions: (i) increased hand hygiene measures, (ii) enhanced disinfection practices, (iii) patient isolation, (iv) use of protective apparel, (v) staff exclusion policies, and (vi) ward closure. Sensitivity analyses explored the impact of varying intervention efficacy, number of initial norovirus cases, the norovirus reproductive rate (R0), and room, ward size, and occupancy. Implementing increased hand hygiene, using protective apparel, staff exclusion policies or increased disinfection separately or in bundles provided net cost-savings, even when the intervention was only 10% effective in preventing further norovirus transmission. Patient isolation or ward closure was cost-saving only when transmission prevention efficacy was very high (≥90%), and their economic value decreased as the number of beds per room and the number of empty beds per ward increased. Increased hand hygiene, using protective apparel or increased disinfection practices separately or in bundles are the most cost-saving interventions for the control and containment of a norovirus outbreak.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Transparency Declaration
  9. References

Norovirus has continued to be a threat in the community and in health care settings [1–4]. Norovirus is highly infectious and can spread rapidly in health care settings, consuming resources and resulting in longer hospital stays [5–8]. The average cost of a microbiologically confirmed nosocomial infection in the United States is estimated to be over $15 000 [9]. A 2007 norovirus outbreak at Johns Hopkins Hospital, a 946-bed hospital, cost an estimated $650 000 [2]. A 2003 outbreak cost a Swiss hospital $40 675 [10]. Outbreaks in the United Kingdom have been estimated to cost $1 million per 1000 hospital beds in 2002–2003 and cost the National Health Service (NHS) an estimated £1 billion annually [3].

Promptly identifying and preventing the spread of a norovirus outbreak may be keys to minimizing its impact. Health care facility administrators and infection control specialists have several containment interventions at their disposal including: (i) enhanced hand hygiene measures, (ii) contact isolation with protective apparel, (iii) isolation or cohorting of infected patients and staff, (iv) modified staff policies to exclude staff from work and prohibit exposed staff from working in unexposed areas, (v) modified visitor policies, (vi) enhanced disinfection practices through increased cleaning of wards and bathrooms, (vii) education of health care workers regarding identification of norovirus enhanced outbreak control measures, and (viii) active surveillance of the outbreak. Each of these interventions have associated costs, such as an increase in hygiene, protective and disinfection materials, reduction in number of available beds, and loss of staff time and productivity.

Deciding whether to implement various norovirus detection and control measures depends on the balance between the costs of implementation and the potential cost-savings from each measure. To better understand this balance, we developed a computer simulation model that simulated the decision regarding whether to perform such strategies. Sensitivity analyses varied model parameters and allowed us to delineate how the cost-benefit of each strategy may vary by initial norovirus outbreak size, prevention strategy efficacy, and strategy cost. The results of our model may help guide policy making and the design of future clinical studies.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Transparency Declaration
  9. References

General model structure

Using TreeAge Pro Suite 2009 (TreeAge Software, Williamstown, MA, USA), which included Microsoft Excel (Microsoft Corporation, Redmond, WA, USA), we developed a stochastic, Monte Carlo decision analytical computer simulation model with dynamic transmission elements that simulated the decision regarding whether to implement a norovirus containment intervention. Fig. 1 outlines the model and the steps that follow the appearance of n primary norovirus cases (base case, 1) in a hospital ward. When no intervention was implemented, each infected primary case generated R0 additional secondary cases, with R0 being the reproductive rate (i.e. the expected number of new cases generated by a single infectious individual upon entering a fully susceptible population) [11]. Alternatively, implementing containment interventions reduced transmission (i.e. decreased R0) proportional to the intervention’s efficacy [effective reproductive rate Re = R0*(1 − intervention efficacy)], which reflected the combination of the inherent efficacy of the intervention and compliance with the intervention. For example, if R0 had a mean of 3.74 (range, 3.179–4.301), an intervention with an efficacy of 50% reduced R0 by 50% to 1.87 (range, 1.59–2.15).

image

Figure 1.  Containment intervention strategy diagram.

Download figure to PowerPoint

Each primary and secondary patient had a probability of being symptomatic or asymptomatic. Symptomatic patients experienced an increased length-of-stay (LOS), based on published studies (Table 1). This increased LOS resulted in occupied bed days that could have been used for other patients. A method described by Graves [12] translated these lost bed-days to opportunity costs. Asymptomatic patients did not experience increases in LOS but could transmit the virus. Each additional secondary case added cost based on their increased LOS. The model considered costs of only primary and secondary cases.

Table 1.   Data inputs for model variables
Description (units)MeanStandard deviation95% rangeSource
Costs ($US)
 Bed day1,742[19]
 Paper towels per bed day0.2770.136[20]
 Soap per bed day0.1060.077[20]
 Alcohol per bed day0.0510.046[20]
 Gloves, gown and mask1.56[21,22]
Wages ($US)
 Registered nurse per hour30.9321.50, 45.68[23]
 Custodian per hour10.627.63, 17.59[23]
Durations
 Increased length of stay (days)2.000.96, 13.05[6,7,24–30]
Numbers
 Reproductive rate (high)7.265.26, 9.25[16]
 Reproductive rate (low)3.743.179, 4.301[15]
 Patient contacts per day3825, 50[21]
 Patient per nurse ratio4.42.9[31]
Probabilities
 Asymptomatic infection0.337110.02833[6,14]

Interventions

Based on the Centers for Disease Control and Prevention’s (CDC) recently published guidelines [13], the interventions we modelled were: (i) increased hand hygiene with soap, water and/or alcohol, (ii) enhanced use of protective apparel, including gloves, gowns and masks with each patient contact, (iii) increased disinfection of the ward, (iv) staff exclusion policies, where ill staff were excluded from the workplace for an additional 2 days after symptoms resolved, increasing the nurse to patient ratio of the remaining staff, (v) patient isolation, where sick patients had a room to themselves, and (vi) ward closure, in which the ward halted new admissions.

To compare strategies, we compared the distribution of incremental cost of implementing the prevention strategy vs. not implementing the strategy:

  • image

Additional analyses examined how this incremental cost varied with the number of base cases, prevention efficacy, R0 and containment measure costs.

Data inputs

Table 1 shows the model input variables and distributions. All probabilities drew from beta distributions, all costs from gamma distributions, and all others from triangular distributions. The probability of a case being symptomatic came from challenge studies [6,14].

The cost of each intervention was as follows:

  • • 
    Increasing hand hygiene: the mean cost was $0.4341 per day (standard deviation 0.168) and included the cost of paper towels, soap and alcohol.
  • • 
    Enhanced protective apparel: the mean cost was $39.55 per patient contact and included the costs of gloves, gowns and masks.
  • • 
    Increased disinfection: the mean cost was $15.93 per day (range, $11.45–$26.39) accounting for custodial wages.
  • • 
    Staff exclusion policies: the mean cost was $674.84 and included the cost of nurse wages (average $30.93 per hour) for the duration of their illness plus 2 days (to account for viral shedding) and the increased nurse-patient ratio of remaining staff (average 0.227 per remaining staff).
  • • 
    Patient isolation: the mean cost was $3484 for one empty bed and was derived from the cost of a bed day ($1742) multiplied by the LOS, accounting for the size of a patient’s room (having one, two or three additional empty beds).
  • • 
    Ward closure: the mean cost was $3484 for one empty bed and was calculated by multiplying the number of empty beds in the closed ward by the cost per bed day and the patient’s LOS.

Sensitivity analyses

Because the efficacy of certain interventions has not been clearly established and may vary under different circumstances (e.g. compliance), sensitivity analyses explored the effects of ranging each intervention’s efficacy from 10% to 90%. The intervention’s efficacy is the proportion of norovirus transmission that the intervention reduces. The efficacy is a function of the intervention’s inherent ability to reduce transmission, implement intensity and staff compliance. Sensitivity analyses also varied the values of the following variables: number of primary cases (range, 1–5), individual room size (range, 2–4 beds), number of empty beds in a closed ward (range, 1–5 beds), and R0 (low R0 [15] range, 3.179–4.301; high R0 [16] range, 5.26–9.25). At a low R0, an infectious person could generate 3.179–4.301 additional norovirus cases. Furthermore, each simulation run consisted of probabilistic sensitivity analyses, which simultaneously varied all input parameters over the ranges listed in Table 1.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Transparency Declaration
  9. References

Each simulation run involved 1 000 000 realizations, each introducing n primary cases to a hospital ward to create an outbreak of primary and secondary cases. Table 2 shows the cost of each intervention strategy (compared with no intervention) after a single primary case and low R0 (3.179–4.301). All reported negative cost values imply cost-savings to the hospital and all positive cost values indicate a net expenditure.

Table 2.   Cost ($US, mean and standard deviation) of intervention strategies (individual and bundled) compared with no intervention for one base case at a low reproductive rate. Grey shading values are cost-savingsThumbnail image of

Unmitigated norovirus outbreak

Initial simulation runs determined the cost of a norovirus case (symptomatic or asymptomatic) to the hospital: mean $6237, standard deviation $3211. Costs arose from the increased LOS from symptomatic cases that translated to lost hospital bed days.

Increased hand hygiene

Table 2 shows that increased hand hygiene yielded net cost-savings for all scenarios, for example increasing hand hygiene after detecting one primary case yielded costs of −$2336 (10% efficacy) to −$21 394 (90% efficacy). Savings increased with the number of primary cases and intervention efficacy. In an outbreak with five primary cases, costs ranged from −$11 464 (10% efficacy) to −$104 273 (90% efficacy). With higher R0 (5.26–9.25), increasing hand hygiene after one primary case showed cost-savings ranging from −$4539 (10% intervention efficacy) to −$39 748 (90% efficacy).

Enhanced use of protective apparel

Enhanced protective apparel use was cost-saving for all scenarios. With more primary cases and increased intervention efficacy, costs decreased (−$103 248 at 90% efficacy and five initial cases). A higher R0 yielded costs between −$4134 (10% efficacy) and −$40 129 (90% efficacy) for one primary case.

Increased disinfection

Increased disinfection was cost-saving (i.e. negative costs) as long as efficacy was ≥10% when there was a single primary case. With five primary cases, increased disinfection cost −$11 085 (10% efficacy) to −$99 363 (90% efficacy). Increased disinfection manifested even greater cost-savings with a higher R0 (cost-savings were as large as −$40 040 at 90% efficacy for one primary case).

Staff exclusion policies

Table 2 demonstrates that staff exclusion policies also yielded cost-savings throughout almost every scenario. With one primary case and low R0 (3.179–4.301), staff exclusion policies with 20% efficacy cost −$2460. Cost-savings grew as the number of primary cases and staff exclusion efficacy increased. Also, increasing R0 lowered the efficacy threshold at which such policies were cost-saving; that is, with one primary case, costs of implementing staff exclusion ranged from −$2096 (10% efficacy) to −$38 410 (90% efficacy).

Patient isolation

Patient isolation resulted in net hospital cost-savings under certain conditions. Assuming two beds per room and one primary case, patient isolation yielded cost-savings for all low and high R0 scenarios when isolation efficacy was ≥50%. However, as the numbers of primary cases or beds per room increased, so did net costs. Isolation with three beds per room had a net cost (i.e. expenditure) to the hospital as long as efficacy was <90% regardless of the number of primary cases. With four beds per room, patient isolation was never cost-saving, with costs ranging from $26 724 (10% efficacy) to $8568 (90% efficacy). Increasing R0 lowered net costs and the thresholds at which patient isolation became cost-saving, for example with two beds per room for all tested base cases, patient isolation with 30% efficacy cost −$4083. With three beds per room, patient isolation became cost-saving at 50% efficacy, and with four beds per room it became cost-saving at an efficacy ≥70%. Fig. 2 shows the costs of implementing this intervention after detecting one case. Each line represents the costs for a specific room size at low and high R0s and highlights how costs change as beds per room and isolation efficacy change. Net hospital costs persist until patient isolation becomes ≥20% efficacious, at which point patient isolation in a two-bed room and low R0 is cost-saving.

image

Figure 2.  Cost of patient isolation with one initial case at low and high reproductive rates.

Download figure to PowerPoint

Ward closure

Ward closure generated net costs for a majority of scenarios explored. Because each empty bed in a closed ward represents opportunity cost to the hospital, ward closure cost increased as the number of empty beds increased. Fig. 3 shows how the value of ward closure increased as efficacy increased. The bands show how ward closure’s economic value varied with R0. The differently shaded bands indicate how the cost increased with number of empty beds. For example, with low R0, one empty bed per ward, and ward closure initiated as soon as a single case appeared, ward closure was only cost-saving when efficacy exceeded 50%. This efficacy threshold decreased somewhat when R0 increased. Increasing the number of empty beds per ward to three increased hospital costs by as much as $25 592 (10% ward closure efficacy). In general, ward closure was cost-saving only when there were no more than three empty beds.

image

Figure 3.  Cost of ward closure with one initial case.

Download figure to PowerPoint

Combined interventions

Table 2 also shows the economic effects of different combinations (i.e. bundles) of interventions and their variation with bundle efficacy (i.e. efficacy of the entire bundle together; individual strategy efficacies within the bundle can vary). The bundles that did not include patient isolation or ward closure were all cost-saving. Patient isolation bundles (two beds per room) were not cost-saving at 40% efficacy but became cost-saving when intervention efficacy was ≥50%. Patient isolation bundles (four beds per room) were not cost-saving at any efficacy. Ward closure plus increased disinfection with one empty bed only became cost-saving at ≥50% efficacy. All other bundles with ward closure were not cost-saving.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Transparency Declaration
  9. References

Our results indicate that increasing hand hygiene, using protective apparel, instituting staff exclusion policies, and increasing disinfection practices are all cost-saving nosocomial norovirus outbreak control measures. In other words, implementing one of these strategies may actually save hospitals money, even when intervention efficacy is not very high. This is important because baseline efficacies for different interventions may vary by institution. For example, baseline rates for hand hygiene between 1977 and 2008 have been found to range from 6.3% to 62% [17].

By contrast, ward closure and patient isolation incurred net costs in many situations. Both interventions appear to be widely used; a systemic review found the ward closure rate for norovirus to be 44.1% [18]. Closing a ward leads to opportunity costs (i.e. lost potential revenue), which grow as the overall ward size and, in turn, number of empty beds increases. Zingg et al. [10] also reported that ward closures had the greatest impact on hospital resources. Ward closure would only be cost-saving at much higher R0s. Ward closure cost may be prohibitive for a hospital, unless it occurs promptly after a single case is detected.

With different norovirus control measures available, decision makers may need to better understand the economic trade-offs. Because intervention efficacy and compliance may vary by circumstances and location, the goal of our study was not to dictate which policy to use but to show how the economic impact of each intervention may vary by different circumstances (e.g. efficacy/compliance, norovirus transmissibility, ward size and bed availability).

Limitations

Every computer model is a simplification of real life. No model can fully represent every event and outcome that may ensue from norovirus illness or exposure. For example, hospitalized patients with cardiovascular, renal or autoimmune disorders may have an increased illness duration. Also, our model focused on primary and secondary cases and did not include tertiary cases. All of these simplifications could underestimate the cost-savings of norovirus prevention strategies.

Conclusions and future directions

Implementing increased hand hygiene, using protective apparel, increased disinfection practices or staff exclusion policies for the control and containment of a norovirus outbreak may provide cost-savings to hospitals. Using these strategies in conjunction with each other could maximize the effects of controlling an outbreak. Patient isolation and ward closure may be more costly, especially when not implemented early. Future studies may better elucidate the efficacy of these interventions. Decision makers, including policy makers, hospital administrators and infection control professionals, can then compare these efficacies with the benchmarks from our study in order to implement the optimal interventions for their local circumstances.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Transparency Declaration
  9. References

This study was supported by the National Institute of General Medical Sciences Models of Infectious Disease Agent Study (MIDAS) through grant 1U54GM088491-0109. The funders had no role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Transparency Declaration

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Transparency Declaration
  9. References

All authors declare that they have no conflicts of interest.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Transparency Declaration
  9. References
  • 1
    Atmar RL, Estes MK. The epidemiologic and clinical importance of norovirus infection. Gastroenterol Clin North Am 2006; 85: 275290.
  • 2
    Johnston CP, Qiu H, Ticehurst JR et al. Outbreak management and implications of a nosocomial norovirus outbreak. Clin Infect Dis 2007; 45: 534540.
  • 3
    Lopman BA, Reacher MH, Vipond IB et al. Epidemiology and cost of nosocomial gastroenteritis, Avon, England, 2002–2003. Emerg Infect Dis 2004; 10: 18271834.
  • 4
    Said MA, Perl TM, Sears CL. Healthcare epidemiology: gastrointestinal flu: norovirus in health care and long-term care facilities. Clin Infect Dis 2008; 47: 12021208.
  • 5
    Glass RI, Parashar UE, Estes MK. Norovirus gastroenteritis. N Engl J Med 2009; 361: 17761785.
  • 6
    Graham DY, Jiang X, Tanaka T, Opekun AR, Madore HP, Estes MK. Norwalk virus infection of volunteers: new insights based on improved assays. J Infect Dis 1994; 170: 3443.
  • 7
    Lopman BA, Reacher MH, Vipond IB, Sarangi J, Brown DW. Clinical manifestation of norovirus gastroenteritis in health care settings. Clin Infect Dis 2004; 39: 318324.
  • 8
    Patel MM, Hall AJ, Vinje J, Parashar UD. Noroviruses: a comprehensive review. J Clin Virol 2009; 44: 18.
  • 9
    Roberts RR, Scott RD 2nd, Cordell R et al. The use of economic modeling to determine the hospital costs associated with nosocomial infections. Clin Infect Dis 2003; 36: 14241432.
  • 10
    Zingg W, Colombo C, Jucker T, Bossart W, Ruef C. Impact of an outbreak of norovirus infection on hospital resources. Infect Control Hosp Epidemiol 2005; 26: 263267.
  • 11
    Anderson RM, May RM, Anderson B. Infectious diseases of humans. Dynamics and Control: Oxford University Press, 1992.
  • 12
    Graves N. Economics and preventing hospital-aquired infection. Emerg Infect Dis 2004; 10: 561566.
  • 13
    MacCannell T, Umscheid CA, Agarwal RK, Lee I, Kuntz G, Stevenson KB. Draft: guideline for prevention and management of norovirus outbreaks in healthcare settings. Atlanta, GA: Centers for Disease Control and Prevention, 2010.
  • 14
    Gray JJ, Cunliffe C, Ball J, Graham DY, Desselberger U, Estes MK. Detection of immunoglobulin M (IgM), IgA, and IgG Norwalk virus-specific antibodies by indirect enzyme-linked immunosorbent assay with baculovirus-expressed Norwalk virus capsid antigen in adult volunteers challenged with Norwalk virus. J Clin Microbiol 1994; 32: 30593063.
  • 15
    Vanderpas J, Louis J, Reynders M, Mascart G, Vandenberg O. Mathematical model for the control of nosocomial norovirus. J Hosp Infect 2009; 71: 214222.
  • 16
    Heijne JC, Teunis P, Morroy G et al. Enhanced hygiene measures and norovirus transmission during an outbreak. Emerg Infect Dis 2009; 15: 2430.
  • 17
    Allegranzi B, Pittet D. Role of hand hygiene in healthcare-associated infection prevention. J Hosp Infect 2009; 73: 305315.
  • 18
    Hansen S, Stamm-Balderjahn S, Zuschneid I et al. Closure of medical departments during nosomial outbreaks: data from a systematic analysis of the literature. J Hosp Infect 2007; 65: 348353.
  • 19
    United States Department of Health & Human Services. HCUP facts and figures: statistics on hospital-based care in the United States. Rockville, MD: AHRQ: Agency for Healthcare Research and Quality, 2007. Available at: http://www.hcup-us.ahrq.gov/reports.jsp (cited 15 September 2009).
  • 20
    Stone PW, Hasan S, Quiros D, Larson EL. Effect of guideline implementation on costs of hand hygiene. Nurs Econ 2007; 25: 279284.
  • 21
    Jernigan JA, Clemence MA, Stott GA et al. Control of methicillin-resistant Staphylococcus aureus at a university hospital: one decade later. Infect Control Hosp Epidemiol 1995; 16: 686696.
  • 22
    PDR. Red book. Montvale, NJ: Thompson Healthcare, Inc., 2008.
  • 23
    Bureau of Labor Statistics. Occupational employment statistics: May 2008 national occupational employment and wage estimates, United States. Washington, D.C.: U.S. Bureau of Labor Statistics Division of Occupational Employment Statistics, 2009. Available at: http://stat.bls.gov/oes/2008/may/oes_nat.htm#b00-0000 (cited 15 Septmember 2009).
  • 24
    Atmar RL, Opekun AR, Gilger MA et al. Norwalk virus shedding after experimental human infection. Emerg Infect Dis 2008; 14: 15531557.
  • 25
    Chadwick PR, McCann R. Transmission of a small round structured virus by vomiting during a hospital outbreak of gastroenteritis. J Hosp Infect 1994; 26: 251259.
  • 26
    Fretz R, Svoboda P, Schorr D, Tanner M, Baumgartner A. Risk factors for infections with Norovirus gastrointestinal illness in Switzerland. Eur J Clin Microbiol Infect Dis 2005; 24: 256261.
  • 27
    Kaplan JE, Feldman R, Campbell DS, Lookabaugh C, Gary GW. The frequency of a Norwalk-like pattern of illness in outbreaks of acute gastroenteritis. Am J Public Health 1982; 72: 13291332.
  • 28
    Lynn S, Toop J, Hanger C, Millar N. Norovirus outbreaks in a hospital setting: the role of infection control. N Z Med J 2004; 117: U771.
  • 29
    Murata T, Katsushima N, Mizuta K, Muraki Y, Hongo S, Matsuzaki Y. Prolonged norovirus shedding in infants ≤6 months of age with gastroenteritis. Pediatr Infect Dis J 2007; 26: 4649.
  • 30
    Rockx B, De Wit M, Vennema H et al. Natural history of human calicivirus infection: a prospective cohort study. Clin Infect Dis 2002; 35: 246253.
  • 31
    Kane RL, Shamliyan T, Mueller C, Duval S, Wilt TJ. Nurse staffing and quality of patient care. Evidence report/technology assessment No. 151. Rockville, MD: Agency for Healthcare Research and Quality, U.S. Department of Health & Human Services, 2007. Contract No.: 07-E005.