Abstract
Although norovirus outbreaks periodically make headlines, it is unclear how much attention norovirus may receive otherwise. A better understanding of the burden could help determine how to prioritize norovirus prevention and control.
We developed a computational simulation model to quantify the clinical and economic burden of norovirus in the United States.
A symptomatic case generated $48 in direct medical costs, $416 in productivity losses ($464 total). The median yearly cost of outbreaks was $7.6 million (range across years, $7.5–$8.2 million) in direct medical costs, and $165.3 million ($161.1–$176.4 million) in productivity losses ($173.5 million total). Sporadic illnesses in the community (incidence, 10–150/1000 population) resulted in 14 118–211 705 hospitalizations, 8.2–122.9 million missed school/work days, $0.2–$2.3 billion in direct medical costs, and $1.4–$20.7 billion in productivity losses ($1.5–$23.1 billion total). The total cost was $10.6 billion based on the current incidence estimate (68.9/1000).
Our study quantified norovirus’ burden. Of the total burden, sporadic cases constituted >90% (thus, annual burden may vary depending on incidence) and productivity losses represented 89%. More than half the economic burden is in adults ≥45, more than half occurs in winter months, and >90% of outbreak costs are due to person-to-person transmission, offering insights into where and when prevention/control efforts may yield returns.
Although norovirus outbreaks in settings such as schools and cruise ships may periodically make headlines [1–3], it is not clear how much attention the virus may receive otherwise, even though it is the cause of one-fifth of all acute gastroenteritis cases worldwide each year [4]. Our group’s previous work found that norovirus cost society $60.3 billion worldwide each year and subdivided this estimate by World Health Organization region, with the Americas having the highest cost at $23.5 billion [5]. While these estimates may help provide an idea of the magnitude of the problem at a global or regional level, they may not fully inform decision making at a country or local level, and there are a limited number of country-level estimates.
Previous work has quantified portions of the burden for the United States. For example, the Economic Research Service estimated the cost of all foodborne norovirus outbreaks in 2011 at $2.3 billion [6], and a few epidemiologic studies have estimated the cost of specific outbreaks, primarily in hospitals, some of which were performed more than a decade ago [7–13]. These estimates focused on the cost of particular outbreaks and did not incorporate sporadic cases that occur throughout the community. Lopman et al [14] estimated the charges of norovirus-associated hospitalizations among adults to be $493 million. However, in the United States, norovirus causes 56 000–71 000 hospitalizations (of 19–21 million acute gastroenteritis cases) annually [15], and, according to a systematic review, only 0.5%–6% of older adults with norovirus end up requiring hospitalization [16].
Providing a country-level estimate of the burden and breaking it down by setting, transmission mode, age, geographic region, and season can help decision makers (eg, policy makers, funders, and product developers) better understand how and where to target efforts. Therefore, we developed a computational simulation model to determine and characterize the clinical and economic burden of norovirus in the United States.
METHODS
Model Description
We constructed a clinical and economic outcomes simulation model in Microsoft Excel (Microsoft) with the Crystal Ball add-in (Oracle Corporation) to estimate the burden of norovirus in the United States from the third-party payer and societal perspectives. For outbreaks occurring in hospitals, we estimated the cost from the hospital and societal perspectives. The Supplementary Material describes the model input parameters, values, and data sources and also shows how each symptomatic norovirus illness moves through the model. Each norovirus case had probabilities of seeking medical care (ie, outpatient or ambulatory care visits), hospitalization, and death. In addition, each case had a probability of missing work or school days for a given duration.
To estimate the burden of norovirus in the community, the model first determines the number of norovirus illnesses in each of 5 age groups (0–4, 5–17, 18–44, 45–64, and ≥65 years). To do this, we first determined the total number of cases in the population by applying an overall population incidence number (illnesses per 1000 person-years) to the total population to determine the total number of cases. Next, these cases were divided into age groups by applying the probability of a norovirus case being in a given age group. We calculated this distribution by normalizing the age-specific incidence rate (ie, dividing age-specific rates by total population rate) to determine the age distribution of norovirus illnesses.
The hospital perspective measured illness costs in lost bed days (ie, additional length of stay attributable to norovirus infection) derived from the attributable length of stay and cost per bed day, following the method described by Graves [17]. Third-party payer costs included all direct medical costs of illness (ie, outpatient visits and hospitalization). Societal costs included direct and indirect (ie, productivity losses due to absenteeism from work or school and mortality) costs. Daily wages served as a proxy for productivity losses. We assumed all symptomatic norovirus cases accrued productivity losses, regardless of age or employment status, because everyone is assumed to contribute to society. Death resulted in accruing the net present value of that person’s lifetime earnings, based on his or her age at death and remaining years of life based on life expectancy [18, 19]. All costs are in 2020 US dollars, converted using a 3% discount rate.
Simulation Experiments and Sensitivity Analyses
Experiments consisted of Monte Carlo simulations of 10 000 trials, varying each parameter across its range. Experiments evaluated the cost of community cases as well as outbreaks, and varied outbreak location. Sensitivity analyses varied the incidence of norovirus in the community (varied from 10 to 150 per 1000 population), the probability of death (from age-specific baseline values to twice baseline), the probability of missing school or work (from 50% to 100%), and the duration of missing school or work (from 2–3 days to 0–1 day). We varied the community norovirus incidence because estimates vary [20–24] and may change over time. However, when evaluating the burden by geographic region and season, we used the incidence in the United States reported by Grytdal et al [20]; their laboratory-based study of Kaiser Permanente members estimated age-specific community incidence rates of norovirus acute gastroenteritis, reporting an overall incidence of 68.9 per 1000 person-years (95% credible interval, 57.7–77.1).
RESULTS
Cost per Symptomatic Norovirus Case
As Figure 1A shows, the cost per illness and composition varies by age. For example, direct medical costs make up 6.4% of total costs for an 18–44-year-old and 23.6% for a person aged ≥65 years. For any given illness, median direct medical costs were $47 (95% uncertainty interval [UI], $39–$59) and productivity losses were $416 ($84–$994), thus amounting to $464 ($132–$1042) in total societal costs. The cost for a long-term care resident ($3915) was higher than an older adult ($751). A hospitalized case cost $10 040 from the hospital perspective ($832 in productivity losses).
Cost per norovirus illness by age group showing baseline probabilities for mortality and missing school and work (A), impact of varying the probability of missing school or work (B), and impact of varying the duration of missed school or work days (100% probability of missing) (C).
Increasing the probability of death (up to twice the base value) had little impact on the cost per case across all age groups, except for older adults, among whom costs increased to $1123 per case. Increasing the mortality rate for a long-term care resident increased the total cost per case to $10 426 when the mortality rate was twice as high. As Figure 1B and 1C show, productivity losses due to absenteeism were a large driver of cost. For any given norovirus illness, varying the probability of missing school or work from 100% to 25% resulted in a range of productivity losses from $416 to $138, and varying the duration of time missing school or work resulted in a range of costs from $416 to $110 (assuming that school or work was missed).
Cost of Norovirus Outbreaks
Annually, outbreaks cost a median of $173.5 million (range, $169.2–$185.2 million). Of this, direct medical costs were $7.6 million (range, $7.5–$8.2 million), and productivity losses were $165.3 million (range, $161.1–$176.4 million). Each year outbreaks in hospitals cost $10.8 million (range, $8.8–$17.1 million) from the hospital perspective. Productivity losses represented 95% of total outbreak costs. Annually, outbreaks resulted in a median of 2820 ambulatory care visits (range, 2521–2932), 1824 hospitalizations (range, 1708–2125), and 184 467 missed school and work days (range, 180 366–197 491). Table 1 shows the breakdown of total societal outbreak costs by geographic region and season.
Median Annual Total Societal Costs ($, in millions) and Cost per Person in Population of Norovirus Outbreaks and Sporadic Norovirus Illnesses in the Community by Geographic Region and Season
| Norovirus Outbreaks | Sporadic Illnesses in Communitya | |||
|---|---|---|---|---|
| Total Societal Cost | Per Person Cost | Total Societal Cost | Per Person Cost | |
| Region | ||||
| Northeast | 8.2 | 0.14 | 1,777.7 | 31.48 |
| Midwest | 11.8 | 0.17 | 2,145.4 | 31.47 |
| South | 8.9 | 0.07 | 3,994.8 | 32.31 |
| West | 5.2 | 0.07 | 2,434.9 | 31.45 |
| Seasonb | ||||
| Winter | 17.9 | … | 5,407.3 | … |
| Spring | 10.6 | … | 1,618.1 | … |
| Summer | 1.7 | … | 1,082.5 | … |
| Fall | 3.9 | … | 2,152.7 | … |
| Norovirus Outbreaks | Sporadic Illnesses in Communitya | |||
|---|---|---|---|---|
| Total Societal Cost | Per Person Cost | Total Societal Cost | Per Person Cost | |
| Region | ||||
| Northeast | 8.2 | 0.14 | 1,777.7 | 31.48 |
| Midwest | 11.8 | 0.17 | 2,145.4 | 31.47 |
| South | 8.9 | 0.07 | 3,994.8 | 32.31 |
| West | 5.2 | 0.07 | 2,434.9 | 31.45 |
| Seasonb | ||||
| Winter | 17.9 | … | 5,407.3 | … |
| Spring | 10.6 | … | 1,618.1 | … |
| Summer | 1.7 | … | 1,082.5 | … |
| Fall | 3.9 | … | 2,152.7 | … |
aBased on current US incidence estimate of 68.9 per 1000 population (estimated by Grytdal et al [20]).
bPer person cost not reported.
Median Annual Total Societal Costs ($, in millions) and Cost per Person in Population of Norovirus Outbreaks and Sporadic Norovirus Illnesses in the Community by Geographic Region and Season
| Norovirus Outbreaks | Sporadic Illnesses in Communitya | |||
|---|---|---|---|---|
| Total Societal Cost | Per Person Cost | Total Societal Cost | Per Person Cost | |
| Region | ||||
| Northeast | 8.2 | 0.14 | 1,777.7 | 31.48 |
| Midwest | 11.8 | 0.17 | 2,145.4 | 31.47 |
| South | 8.9 | 0.07 | 3,994.8 | 32.31 |
| West | 5.2 | 0.07 | 2,434.9 | 31.45 |
| Seasonb | ||||
| Winter | 17.9 | … | 5,407.3 | … |
| Spring | 10.6 | … | 1,618.1 | … |
| Summer | 1.7 | … | 1,082.5 | … |
| Fall | 3.9 | … | 2,152.7 | … |
| Norovirus Outbreaks | Sporadic Illnesses in Communitya | |||
|---|---|---|---|---|
| Total Societal Cost | Per Person Cost | Total Societal Cost | Per Person Cost | |
| Region | ||||
| Northeast | 8.2 | 0.14 | 1,777.7 | 31.48 |
| Midwest | 11.8 | 0.17 | 2,145.4 | 31.47 |
| South | 8.9 | 0.07 | 3,994.8 | 32.31 |
| West | 5.2 | 0.07 | 2,434.9 | 31.45 |
| Seasonb | ||||
| Winter | 17.9 | … | 5,407.3 | … |
| Spring | 10.6 | … | 1,618.1 | … |
| Summer | 1.7 | … | 1,082.5 | … |
| Fall | 3.9 | … | 2,152.7 | … |
aBased on current US incidence estimate of 68.9 per 1000 population (estimated by Grytdal et al [20]).
bPer person cost not reported.
Figures 2 and 3 show how total costs varied by outbreak setting and by mode of transmission (eg, person-to-person or foodborne). Outbreaks occurring in long-term care facilities resulted in the highest overall costs (range, $150–$165 million, annually), with person-to-person transmission driving 93%–95% of the total burden. Nonhealthcare settings (ie, settings other than hospitals, long-term care/nursing homes, or other healthcare settings) resulted in total costs ranging from $13.5 million to $16.1 million, with 74%–78% resulting from person-to-person transmission. The cost of hospital outbreaks varied annually, ranging from $8.7 million (52 outbreaks) to $16.9 million (85 outbreaks) from the hospital perspective, with person-to-person transmission generating ≥86% of costs. Only 2 outbreaks on ships or boats were reported from 2013 to 2017, costing $2994 and $18 961.
Total annual cost of norovirus outbreaks (societal perspective) by mode of transmission for outbreaks occurring in a long-term care/nursing home setting (A) or any nonhealthcare setting (B).
Total annual cost of norovirus outbreaks (societal perspective) by mode of transmission for outbreaks occurring in a nonhealthcare setting, specifically in a school/university (A), daycare (B), office/business (C), or restaurant (D) setting.
Cost of Sporadic Norovirus in the Community
Figure 4 shows how the direct medical costs and productivity losses of sporadic cases in the community vary with incidence, mortality rate, and the probability and duration of missing school or work days. Table 2 shows the clinical and economic outcomes and how they vary with incidence, assuming the baseline probability of death and assuming that all persons with symptomatic illness miss school or work. Overall, productivity losses represent 90%–91% of the total cost, varying with incidence and mortality rate. Assuming the current US incidence estimate (68.9 per 1000 person-years [20]), the cost was highest in adults aged ≥45 years, generating 57.5% of the total burden. In addition, norovirus resulted in 1.9 million ambulatory care visits, 95 732 hospitalizations, and 55.1 million missed school and work days. The costs per person in the population were similar across geographic regions (Table 1).
Clinical and Economic Burden of Sporadic Norovirus Illnesses in the Community in the United States
| Median (95% Uncertainty Interval) | ||||||||
|---|---|---|---|---|---|---|---|---|
| Cost, $ (in Millions) | ||||||||
| Age Group, y | Illnesses, Estimated No. | Deaths, No. | Ambulatory Care Visits, No. | Hospitalizations, No. | Missed School or Work Days, No. (in Millions) | Total Direct Medical | Productivity Losses | Total Societal |
| Incidence: 10 per 1000 population | ||||||||
| 0–4 | 415 503 | 3 (2–3) | 69 869 (64 408–75 297) | 1814 (1673–1953) | 1.0 (0.8–1.2) | 22.2 (14.6–31.6) | 159.3 (29.7–374.1) | 182.2 (52.9–397.6) |
| 5–17 | 199 572 | 1 (1–1) | 33 534 (30 918–36 140) | 366 (338–394) | 0.5 (0.4–0.6) | 7.8 (4.2–12.2) | 75.8 (14.1–178.2) | 83.9 (22.2–186.6) |
| 18–44 | 935 699 | 4 (4–5) | 59 866 (55 197–64 526) | 2053 (1895–2212) | 2.3 (1.9–2.8) | 24.2 (17.7–32.1) | 354.7 (66.2–834.4) | 379.9 (91.8–859.3) |
| 45–64 | 1 171 259 | 5 (5–6) | 63 242 (58 373–68 175) | 2570 (2372–2768) | 2.9 (2.4–3.5) | 31.6 (24.2–40.5) | 442.4 (82.3–1040.0) | 474.8 (114.5–1073.2) |
| ≥65 | 549 641 | 184 (174–193) | 56 611 (52 241–61 008) | 7316 (6983–7647) | 1.4 (1.1–1.6) | 70.6 (63.4–79.0) | 328.8 (62.5–757.5) | 400.3 (133.9–828.7) |
| Total | 3 271 674 | 197 (187–206) | 283 068 (272 996–293 611) | 14 118 (13 607–14 632) | 8.2 (6.6–9.7) | 156.2 (125.6–193.8) | 1359.6 (256.3–3191.1) | 1518.4 (415.4–3340.6) |
| Incidence: 68.9 per 1000 population (estimated by Grytdal et al [20]) | ||||||||
| 0–4 | 2 826 440 (2 392 462–3 201 192) | 18 (15–21) | 472 090 (394 592–550 934) | 12 275 (10 245–14 329) | 7.0 (5.3–8.9) | 150.3 (96.8–220.7) | 1090.8 (201.4–2578.9) | 1243.1 (348.3–2740.9) |
| 5–17 | 1 357 581 (1 149 135–1 537 581) | 6 (5–7) | 227 045 (189 471–264 096) | 2480 (2077–2895) | 3.4 (2.5–4.3) | 52.7 (28.3–84.8) | 519.3 (95.9–1229.9) | 573.6 (147.8–1285.3) |
| 18–44 | 6 365 054 (5 387 748–7 208 984) | 29 (24–34) | 404 991 (338 347–471 984) | 13 895 (11 627–16 217) | 15.8 (11.9–20.1) | 163.4 (115.9–225.3) | 2429.8 (449.3–5755.7) | 2597.6 (610.3–5926.8) |
| 45–64 | 7 967 445 (6 744 104–9 023 833) | 36 (30–42) | 428 064 (358 257–498 040) | 17 393 (14 553–20 300) | 19.7 (14.9–25.2) | 213.8 (156.5–284.1) | 3029.8 (561.1–7179.1) | 3246.6 (768.9–7404.7) |
| ≥65 | 3 738 913 (3 164 831–4 234 648) | 1248 (1050–1430) | 383 217 (320 671–446 242) | 49 574 (41 761–56 902) | 9.3 (7.0–11.8) | 478.4 (394.1–570.3) | 2251.9 (418.6–5211.9) | 2732.3 (889.1–5729.9) |
| Total | 22 206 315 (18 850 188– 25 196 379) | 1334 (1127–1532) | 1 919 498 (1 628 664–2 190 426) | 95 732 (80 903–109 313) | 55.1 (41.7–70.3) | 1059.0 (810.4–1367.2) | 9333.4 (1722.4–21 955.2) | 10 393.5 (2778.2–23 088.3) |
| Incidence: 150 per 1000 population | ||||||||
| 0–4 | 6 232 540 | 40 (37–43) | 1 047 229 (965 907–1 128 817) | 27 182 (25 112–29 304) | 15.6 (12.6–18.5) | 333.9 (219.9–470.6) | 2421.6 (476.3–5600.2) | 2761.8 (814.4–5953.3) |
| 5–17 | 2 993 582 | 14 (13–15) | 503 144 (463 129–542 323) | 5497 (5067–5920) | 7.5 (6.1–8.9) | 117.0 (63.6–182.2) | 1152.2 (226.6–2667.6) | 1272.1 (347.1–2792.6) |
| 18–44 | 14 035 483 | 64 (59–69) | 897 860 (828 217–967 586) | 30 782 (28 405–33 209) | 35.1 (28.4–41.7) | 362.9 (266.6–480.0) | 5391.5 (1060.7–12 477.8) | 5762.4 (1417.8–12 849.2) |
| 45–64 | 17 568 891 | 80 (74–86) | 948 059 (875 263–1 022 114) | 38 531 (35 556–41 569) | 44.0 (35.6–52.2) | 474.6 (362.7–606.2) | 6724.9 (1323.7–15 567.6) | 7203.3 (1787.5–16 049.8) |
| ≥65 | 8 244 619 | 2754 (2609–2898) | 848 981 (782 933–914 354) | 109 688 (104 674–114 805) | 20.6 (16.7–24.5) | 1059.9 (951.3–1182.6) | 5017.4 (966.7–11 359.7) | 6075.9 (2037.0–12 448.7) |
| Total | 49 075 115 | 2951 (2807–3093) | 4 245 083 (4 093 440–4 397 574) | 211 705 (203 860–219 568) | 122.9 (99.5–145.8) | 2346.1 (1891.0–2893.3) | 20 713.9 (4057.9–47 557.7) | 23 091.6 (6395.5–49 939.5) |
| Median (95% Uncertainty Interval) | ||||||||
|---|---|---|---|---|---|---|---|---|
| Cost, $ (in Millions) | ||||||||
| Age Group, y | Illnesses, Estimated No. | Deaths, No. | Ambulatory Care Visits, No. | Hospitalizations, No. | Missed School or Work Days, No. (in Millions) | Total Direct Medical | Productivity Losses | Total Societal |
| Incidence: 10 per 1000 population | ||||||||
| 0–4 | 415 503 | 3 (2–3) | 69 869 (64 408–75 297) | 1814 (1673–1953) | 1.0 (0.8–1.2) | 22.2 (14.6–31.6) | 159.3 (29.7–374.1) | 182.2 (52.9–397.6) |
| 5–17 | 199 572 | 1 (1–1) | 33 534 (30 918–36 140) | 366 (338–394) | 0.5 (0.4–0.6) | 7.8 (4.2–12.2) | 75.8 (14.1–178.2) | 83.9 (22.2–186.6) |
| 18–44 | 935 699 | 4 (4–5) | 59 866 (55 197–64 526) | 2053 (1895–2212) | 2.3 (1.9–2.8) | 24.2 (17.7–32.1) | 354.7 (66.2–834.4) | 379.9 (91.8–859.3) |
| 45–64 | 1 171 259 | 5 (5–6) | 63 242 (58 373–68 175) | 2570 (2372–2768) | 2.9 (2.4–3.5) | 31.6 (24.2–40.5) | 442.4 (82.3–1040.0) | 474.8 (114.5–1073.2) |
| ≥65 | 549 641 | 184 (174–193) | 56 611 (52 241–61 008) | 7316 (6983–7647) | 1.4 (1.1–1.6) | 70.6 (63.4–79.0) | 328.8 (62.5–757.5) | 400.3 (133.9–828.7) |
| Total | 3 271 674 | 197 (187–206) | 283 068 (272 996–293 611) | 14 118 (13 607–14 632) | 8.2 (6.6–9.7) | 156.2 (125.6–193.8) | 1359.6 (256.3–3191.1) | 1518.4 (415.4–3340.6) |
| Incidence: 68.9 per 1000 population (estimated by Grytdal et al [20]) | ||||||||
| 0–4 | 2 826 440 (2 392 462–3 201 192) | 18 (15–21) | 472 090 (394 592–550 934) | 12 275 (10 245–14 329) | 7.0 (5.3–8.9) | 150.3 (96.8–220.7) | 1090.8 (201.4–2578.9) | 1243.1 (348.3–2740.9) |
| 5–17 | 1 357 581 (1 149 135–1 537 581) | 6 (5–7) | 227 045 (189 471–264 096) | 2480 (2077–2895) | 3.4 (2.5–4.3) | 52.7 (28.3–84.8) | 519.3 (95.9–1229.9) | 573.6 (147.8–1285.3) |
| 18–44 | 6 365 054 (5 387 748–7 208 984) | 29 (24–34) | 404 991 (338 347–471 984) | 13 895 (11 627–16 217) | 15.8 (11.9–20.1) | 163.4 (115.9–225.3) | 2429.8 (449.3–5755.7) | 2597.6 (610.3–5926.8) |
| 45–64 | 7 967 445 (6 744 104–9 023 833) | 36 (30–42) | 428 064 (358 257–498 040) | 17 393 (14 553–20 300) | 19.7 (14.9–25.2) | 213.8 (156.5–284.1) | 3029.8 (561.1–7179.1) | 3246.6 (768.9–7404.7) |
| ≥65 | 3 738 913 (3 164 831–4 234 648) | 1248 (1050–1430) | 383 217 (320 671–446 242) | 49 574 (41 761–56 902) | 9.3 (7.0–11.8) | 478.4 (394.1–570.3) | 2251.9 (418.6–5211.9) | 2732.3 (889.1–5729.9) |
| Total | 22 206 315 (18 850 188– 25 196 379) | 1334 (1127–1532) | 1 919 498 (1 628 664–2 190 426) | 95 732 (80 903–109 313) | 55.1 (41.7–70.3) | 1059.0 (810.4–1367.2) | 9333.4 (1722.4–21 955.2) | 10 393.5 (2778.2–23 088.3) |
| Incidence: 150 per 1000 population | ||||||||
| 0–4 | 6 232 540 | 40 (37–43) | 1 047 229 (965 907–1 128 817) | 27 182 (25 112–29 304) | 15.6 (12.6–18.5) | 333.9 (219.9–470.6) | 2421.6 (476.3–5600.2) | 2761.8 (814.4–5953.3) |
| 5–17 | 2 993 582 | 14 (13–15) | 503 144 (463 129–542 323) | 5497 (5067–5920) | 7.5 (6.1–8.9) | 117.0 (63.6–182.2) | 1152.2 (226.6–2667.6) | 1272.1 (347.1–2792.6) |
| 18–44 | 14 035 483 | 64 (59–69) | 897 860 (828 217–967 586) | 30 782 (28 405–33 209) | 35.1 (28.4–41.7) | 362.9 (266.6–480.0) | 5391.5 (1060.7–12 477.8) | 5762.4 (1417.8–12 849.2) |
| 45–64 | 17 568 891 | 80 (74–86) | 948 059 (875 263–1 022 114) | 38 531 (35 556–41 569) | 44.0 (35.6–52.2) | 474.6 (362.7–606.2) | 6724.9 (1323.7–15 567.6) | 7203.3 (1787.5–16 049.8) |
| ≥65 | 8 244 619 | 2754 (2609–2898) | 848 981 (782 933–914 354) | 109 688 (104 674–114 805) | 20.6 (16.7–24.5) | 1059.9 (951.3–1182.6) | 5017.4 (966.7–11 359.7) | 6075.9 (2037.0–12 448.7) |
| Total | 49 075 115 | 2951 (2807–3093) | 4 245 083 (4 093 440–4 397 574) | 211 705 (203 860–219 568) | 122.9 (99.5–145.8) | 2346.1 (1891.0–2893.3) | 20 713.9 (4057.9–47 557.7) | 23 091.6 (6395.5–49 939.5) |
Clinical and Economic Burden of Sporadic Norovirus Illnesses in the Community in the United States
| Median (95% Uncertainty Interval) | ||||||||
|---|---|---|---|---|---|---|---|---|
| Cost, $ (in Millions) | ||||||||
| Age Group, y | Illnesses, Estimated No. | Deaths, No. | Ambulatory Care Visits, No. | Hospitalizations, No. | Missed School or Work Days, No. (in Millions) | Total Direct Medical | Productivity Losses | Total Societal |
| Incidence: 10 per 1000 population | ||||||||
| 0–4 | 415 503 | 3 (2–3) | 69 869 (64 408–75 297) | 1814 (1673–1953) | 1.0 (0.8–1.2) | 22.2 (14.6–31.6) | 159.3 (29.7–374.1) | 182.2 (52.9–397.6) |
| 5–17 | 199 572 | 1 (1–1) | 33 534 (30 918–36 140) | 366 (338–394) | 0.5 (0.4–0.6) | 7.8 (4.2–12.2) | 75.8 (14.1–178.2) | 83.9 (22.2–186.6) |
| 18–44 | 935 699 | 4 (4–5) | 59 866 (55 197–64 526) | 2053 (1895–2212) | 2.3 (1.9–2.8) | 24.2 (17.7–32.1) | 354.7 (66.2–834.4) | 379.9 (91.8–859.3) |
| 45–64 | 1 171 259 | 5 (5–6) | 63 242 (58 373–68 175) | 2570 (2372–2768) | 2.9 (2.4–3.5) | 31.6 (24.2–40.5) | 442.4 (82.3–1040.0) | 474.8 (114.5–1073.2) |
| ≥65 | 549 641 | 184 (174–193) | 56 611 (52 241–61 008) | 7316 (6983–7647) | 1.4 (1.1–1.6) | 70.6 (63.4–79.0) | 328.8 (62.5–757.5) | 400.3 (133.9–828.7) |
| Total | 3 271 674 | 197 (187–206) | 283 068 (272 996–293 611) | 14 118 (13 607–14 632) | 8.2 (6.6–9.7) | 156.2 (125.6–193.8) | 1359.6 (256.3–3191.1) | 1518.4 (415.4–3340.6) |
| Incidence: 68.9 per 1000 population (estimated by Grytdal et al [20]) | ||||||||
| 0–4 | 2 826 440 (2 392 462–3 201 192) | 18 (15–21) | 472 090 (394 592–550 934) | 12 275 (10 245–14 329) | 7.0 (5.3–8.9) | 150.3 (96.8–220.7) | 1090.8 (201.4–2578.9) | 1243.1 (348.3–2740.9) |
| 5–17 | 1 357 581 (1 149 135–1 537 581) | 6 (5–7) | 227 045 (189 471–264 096) | 2480 (2077–2895) | 3.4 (2.5–4.3) | 52.7 (28.3–84.8) | 519.3 (95.9–1229.9) | 573.6 (147.8–1285.3) |
| 18–44 | 6 365 054 (5 387 748–7 208 984) | 29 (24–34) | 404 991 (338 347–471 984) | 13 895 (11 627–16 217) | 15.8 (11.9–20.1) | 163.4 (115.9–225.3) | 2429.8 (449.3–5755.7) | 2597.6 (610.3–5926.8) |
| 45–64 | 7 967 445 (6 744 104–9 023 833) | 36 (30–42) | 428 064 (358 257–498 040) | 17 393 (14 553–20 300) | 19.7 (14.9–25.2) | 213.8 (156.5–284.1) | 3029.8 (561.1–7179.1) | 3246.6 (768.9–7404.7) |
| ≥65 | 3 738 913 (3 164 831–4 234 648) | 1248 (1050–1430) | 383 217 (320 671–446 242) | 49 574 (41 761–56 902) | 9.3 (7.0–11.8) | 478.4 (394.1–570.3) | 2251.9 (418.6–5211.9) | 2732.3 (889.1–5729.9) |
| Total | 22 206 315 (18 850 188– 25 196 379) | 1334 (1127–1532) | 1 919 498 (1 628 664–2 190 426) | 95 732 (80 903–109 313) | 55.1 (41.7–70.3) | 1059.0 (810.4–1367.2) | 9333.4 (1722.4–21 955.2) | 10 393.5 (2778.2–23 088.3) |
| Incidence: 150 per 1000 population | ||||||||
| 0–4 | 6 232 540 | 40 (37–43) | 1 047 229 (965 907–1 128 817) | 27 182 (25 112–29 304) | 15.6 (12.6–18.5) | 333.9 (219.9–470.6) | 2421.6 (476.3–5600.2) | 2761.8 (814.4–5953.3) |
| 5–17 | 2 993 582 | 14 (13–15) | 503 144 (463 129–542 323) | 5497 (5067–5920) | 7.5 (6.1–8.9) | 117.0 (63.6–182.2) | 1152.2 (226.6–2667.6) | 1272.1 (347.1–2792.6) |
| 18–44 | 14 035 483 | 64 (59–69) | 897 860 (828 217–967 586) | 30 782 (28 405–33 209) | 35.1 (28.4–41.7) | 362.9 (266.6–480.0) | 5391.5 (1060.7–12 477.8) | 5762.4 (1417.8–12 849.2) |
| 45–64 | 17 568 891 | 80 (74–86) | 948 059 (875 263–1 022 114) | 38 531 (35 556–41 569) | 44.0 (35.6–52.2) | 474.6 (362.7–606.2) | 6724.9 (1323.7–15 567.6) | 7203.3 (1787.5–16 049.8) |
| ≥65 | 8 244 619 | 2754 (2609–2898) | 848 981 (782 933–914 354) | 109 688 (104 674–114 805) | 20.6 (16.7–24.5) | 1059.9 (951.3–1182.6) | 5017.4 (966.7–11 359.7) | 6075.9 (2037.0–12 448.7) |
| Total | 49 075 115 | 2951 (2807–3093) | 4 245 083 (4 093 440–4 397 574) | 211 705 (203 860–219 568) | 122.9 (99.5–145.8) | 2346.1 (1891.0–2893.3) | 20 713.9 (4057.9–47 557.7) | 23 091.6 (6395.5–49 939.5) |
| Median (95% Uncertainty Interval) | ||||||||
|---|---|---|---|---|---|---|---|---|
| Cost, $ (in Millions) | ||||||||
| Age Group, y | Illnesses, Estimated No. | Deaths, No. | Ambulatory Care Visits, No. | Hospitalizations, No. | Missed School or Work Days, No. (in Millions) | Total Direct Medical | Productivity Losses | Total Societal |
| Incidence: 10 per 1000 population | ||||||||
| 0–4 | 415 503 | 3 (2–3) | 69 869 (64 408–75 297) | 1814 (1673–1953) | 1.0 (0.8–1.2) | 22.2 (14.6–31.6) | 159.3 (29.7–374.1) | 182.2 (52.9–397.6) |
| 5–17 | 199 572 | 1 (1–1) | 33 534 (30 918–36 140) | 366 (338–394) | 0.5 (0.4–0.6) | 7.8 (4.2–12.2) | 75.8 (14.1–178.2) | 83.9 (22.2–186.6) |
| 18–44 | 935 699 | 4 (4–5) | 59 866 (55 197–64 526) | 2053 (1895–2212) | 2.3 (1.9–2.8) | 24.2 (17.7–32.1) | 354.7 (66.2–834.4) | 379.9 (91.8–859.3) |
| 45–64 | 1 171 259 | 5 (5–6) | 63 242 (58 373–68 175) | 2570 (2372–2768) | 2.9 (2.4–3.5) | 31.6 (24.2–40.5) | 442.4 (82.3–1040.0) | 474.8 (114.5–1073.2) |
| ≥65 | 549 641 | 184 (174–193) | 56 611 (52 241–61 008) | 7316 (6983–7647) | 1.4 (1.1–1.6) | 70.6 (63.4–79.0) | 328.8 (62.5–757.5) | 400.3 (133.9–828.7) |
| Total | 3 271 674 | 197 (187–206) | 283 068 (272 996–293 611) | 14 118 (13 607–14 632) | 8.2 (6.6–9.7) | 156.2 (125.6–193.8) | 1359.6 (256.3–3191.1) | 1518.4 (415.4–3340.6) |
| Incidence: 68.9 per 1000 population (estimated by Grytdal et al [20]) | ||||||||
| 0–4 | 2 826 440 (2 392 462–3 201 192) | 18 (15–21) | 472 090 (394 592–550 934) | 12 275 (10 245–14 329) | 7.0 (5.3–8.9) | 150.3 (96.8–220.7) | 1090.8 (201.4–2578.9) | 1243.1 (348.3–2740.9) |
| 5–17 | 1 357 581 (1 149 135–1 537 581) | 6 (5–7) | 227 045 (189 471–264 096) | 2480 (2077–2895) | 3.4 (2.5–4.3) | 52.7 (28.3–84.8) | 519.3 (95.9–1229.9) | 573.6 (147.8–1285.3) |
| 18–44 | 6 365 054 (5 387 748–7 208 984) | 29 (24–34) | 404 991 (338 347–471 984) | 13 895 (11 627–16 217) | 15.8 (11.9–20.1) | 163.4 (115.9–225.3) | 2429.8 (449.3–5755.7) | 2597.6 (610.3–5926.8) |
| 45–64 | 7 967 445 (6 744 104–9 023 833) | 36 (30–42) | 428 064 (358 257–498 040) | 17 393 (14 553–20 300) | 19.7 (14.9–25.2) | 213.8 (156.5–284.1) | 3029.8 (561.1–7179.1) | 3246.6 (768.9–7404.7) |
| ≥65 | 3 738 913 (3 164 831–4 234 648) | 1248 (1050–1430) | 383 217 (320 671–446 242) | 49 574 (41 761–56 902) | 9.3 (7.0–11.8) | 478.4 (394.1–570.3) | 2251.9 (418.6–5211.9) | 2732.3 (889.1–5729.9) |
| Total | 22 206 315 (18 850 188– 25 196 379) | 1334 (1127–1532) | 1 919 498 (1 628 664–2 190 426) | 95 732 (80 903–109 313) | 55.1 (41.7–70.3) | 1059.0 (810.4–1367.2) | 9333.4 (1722.4–21 955.2) | 10 393.5 (2778.2–23 088.3) |
| Incidence: 150 per 1000 population | ||||||||
| 0–4 | 6 232 540 | 40 (37–43) | 1 047 229 (965 907–1 128 817) | 27 182 (25 112–29 304) | 15.6 (12.6–18.5) | 333.9 (219.9–470.6) | 2421.6 (476.3–5600.2) | 2761.8 (814.4–5953.3) |
| 5–17 | 2 993 582 | 14 (13–15) | 503 144 (463 129–542 323) | 5497 (5067–5920) | 7.5 (6.1–8.9) | 117.0 (63.6–182.2) | 1152.2 (226.6–2667.6) | 1272.1 (347.1–2792.6) |
| 18–44 | 14 035 483 | 64 (59–69) | 897 860 (828 217–967 586) | 30 782 (28 405–33 209) | 35.1 (28.4–41.7) | 362.9 (266.6–480.0) | 5391.5 (1060.7–12 477.8) | 5762.4 (1417.8–12 849.2) |
| 45–64 | 17 568 891 | 80 (74–86) | 948 059 (875 263–1 022 114) | 38 531 (35 556–41 569) | 44.0 (35.6–52.2) | 474.6 (362.7–606.2) | 6724.9 (1323.7–15 567.6) | 7203.3 (1787.5–16 049.8) |
| ≥65 | 8 244 619 | 2754 (2609–2898) | 848 981 (782 933–914 354) | 109 688 (104 674–114 805) | 20.6 (16.7–24.5) | 1059.9 (951.3–1182.6) | 5017.4 (966.7–11 359.7) | 6075.9 (2037.0–12 448.7) |
| Total | 49 075 115 | 2951 (2807–3093) | 4 245 083 (4 093 440–4 397 574) | 211 705 (203 860–219 568) | 122.9 (99.5–145.8) | 2346.1 (1891.0–2893.3) | 20 713.9 (4057.9–47 557.7) | 23 091.6 (6395.5–49 939.5) |
Annual costs of sporadic norovirus illnesses in the community varying with incidence (per 1000 population), broken down by direct medical costs (A), and productivity losses varying with norovirus-associated mortality risk (B), the probability of missing school or work (C), and the probability and duration of missing school or work (D).
As Figure 4B shows, mortality rate had little impact on the total cost (up to twice its baseline value), because the vast majority of productivity losses are due to absenteeism. Varying the probability of missing school or work from 100% to 25% decreased productivity losses from $9.3 billion to $3.0 billion (Figure 4C), while varying the number of missed days from 2–3 days to 0–1 day decreased productivity losses from $9.3 billion to $2.5 billion (Figure 4D), assuming the incidence estimated by Grytdal et al [20].
Cost of Sporadic and Outbreak Norovirus Cases
Including both sporadic and outbreak cases, norovirus causes 3.3 to 49.1 million symptomatic illnesses each year, varying with the incidence in the community (assuming everyone misses school or work for the full illness duration). These illnesses resulted in 500 to 3254 deaths, 15 942 to 213 529 hospitalizations, and 8.4 to 123.1 million missed school and work days and cost $163.8 million ($129.6–$207.9 million) to $2.4 billion ($1.9–$2.9 billion) in direct medical costs and $1.5 billion ($0.4–$3.4 billion) to $20.9 billion ($4.2–$47.7 billion) in productivity losses. Sporadic cases generated 90%–99% of the total annual burden. Using the current US incidence estimates [20], norovirus cost $10.6 billion to society annually and among key demographic groups, cost >$1.2 billion in children aged <5 years and >$2.9 billion in adults aged ≥65 years. However, these are likely to be underestimates, because they do not include cases from outbreaks occurring outside daycare and long-term care settings.
Assuming that only 25% of persons with illness miss work or school for 0–1 day, these 3.3 to 49.1 million cases resulted in 456 401 to 6.6 million missed school or work days, costing $0.3 billion ($0.06–$0.9 billion) to $3.2 billion ($0.6–$8.0 billion) in productivity losses. When the likelihood and duration of missing school and work days were reduced, productivity losses represented 58%–67% of the total annual burden.
DISCUSSION
Our results suggest that the economic burden of norovirus is $10.6 billion based on the latest US incidence numbers, which estimate a median incidence of 68.9 per 1000 [20]. Varying the incidence from 10 to 150 per 1000 population and missed school and work days resulted in a range of $0.5 to $23.3 billion each year. The economic burden may vary from year to year, depending on the disease burden of sporadic illness; if this is stable, the economic burden would be relatively stable, because the year-to-year variation in costs from outbreaks was low. The burden of norovirus is higher than other diseases in the United States that have received far more attention. For example, pneumonia in older adults (aged ≥50 years) cost an estimated $7.9 billion annually [25], and rotavirus in children <5 years cost $1.5 billion annually to society before vaccine introduction [26]; West Nile virus cost an estimated $71.5 million annually [27]; and the annual societal cost of influenza ranges from $13.0 billion to $146.2 billion (all converted to 2020 values) [28, 29]. The cost of norovirus is likely higher owing to its ubiquitous nature, affecting persons of all ages, with potential for multiple episodes throughout a person’s lifetime. In fact, it has been estimated that, on average, an individual in the United States experiences 5 norovirus episodes in his or her lifetime, given the diversity of circulating noroviruses and lack of persistent cross-protective immunity [15].
We found that sporadic norovirus cases in the community constitute >90% of the total annual cost, which suggests a need to better track and prevent such cases. Just detecting and controlling outbreaks is not enough and these numbers cannot be relied on to estimate what is really going on. However, community illnesses are harder to detect because ill persons must interact with the healthcare system and have a stool sample collected for diagnosis in order to be identified, and few cases require medical attention. Given the need for interaction, claims data and electronic health record data, both based on healthcare use, will not capture the full burden. Better surveillance can help overcome these issues and provide a more robust estimate of the burden of norovirus disease in the community.
Studies reporting community incidence vary widely (eg, 4.5 per 100 person-years in England and Wales [23], 65 per 1000 person-years [21] and 68.9 per 1000 person-years in the United States [20], and a total incidence of 14 906 per 100 000 population in the World Health Organization region of the Americas [24]). A review of community-based norovirus surveillance systems worldwide found estimates varied from 0.024 to 60 cases per 1000 person-years [22]. Currently, norovirus is not reportable [30]. In many situations persons with acute gastroenteritis are not tested for norovirus, and many studies focus on diarrheal symptoms and do not include vomiting-only episodes, which may represent 13%–27% of cases [31, 32]. Thus, dialogue regarding norovirus needs to shift away from outbreaks and toward sporadic cases. This is apparent by the fact that norovirus seems to make the news only during outbreaks and rarely when not pertaining to outbreaks [33, 34]. In addition, community surveillance may mean a combination of more awareness (eg, recognition of norovirus symptoms, clinical awareness for testing) and increased use of diagnostics in clinical settings.
As our results show, productivity losses represented 89% of total costs, and those losses due to absenteeism were the largest driver of annual cost. Thus, estimates that do not capture productivity losses (eg, from claims or electronic health record data) greatly underestimate the burden of norovirus. This makes it important for employers and businesses to have appropriate sick leave policies and encourage and use proper infection prevention and control (eg, hand hygiene, increase the frequency of cleaning and disinfecting high-traffic areas, and vaccination), which can reduce the infection risk. Employers could also work toward avoiding a culture of presenteeism, encouraging personnel to take the necessary time off when sick, instead of coming to work ill. Given that adults spend a large amount of time at their workplace, employee health is important for workplace productivity, and subsequently local, national, and global economies. To note, we used wages as a proxy for productivity losses, but these do not account for secondary costs that may occur if an employee is sick (eg, operational costs).
Our study showed that more than half the economic burden is concentrated in adults aged ≥45 years, more than half occurs during the winter months, and >90% of outbreaks costs are due to person-to-person transmission, offering insights into where and when prevention and control efforts may yield the greatest returns. During winter, 53% of all costs are incurred, which suggests the importance of timing interventions (eg, a vaccine). Similarly, because costs vary by age group, interventions may want to target or prioritize adults ≥45 years old. As another example, outbreak-related costs were predominantly driven by person-to-person transmission; thus, places (eg, daycare facilities, schools/universities, and offices/businesses) may want to consider expanding sick leave policies to prevent further spread, and locations such as long-term care facilities may want to consider policies for interventions that consider the long duration of viral shedding (eg, extended use of contact precautions). We also found that the burden is fairly ubiquitous across the United States, with similar costs per person in the population in all geographic regions, suggesting that norovirus is spread out in the United States.
Our study had some limitations. All models are simplifications of real life and cannot account for every event or outcome [35]. Our input data came from various sources of various quality and may change as more data becomes available. For example, inputs for seeking care were derived from a study of self-reported care seeking for diarrhea. Our results include the outbreaks reported to the National Outbreak Reporting System, which include enteric outbreaks reported by local, state, and territorial health departments to the Centers for Disease Control and Prevention; however, not all outbreaks may be reported. We did not include additional costs for immunocompromised persons, which may exacerbate symptoms and outcomes. For persons incurring productivity losses, we only considered losses for the duration of illness; in reality, a person may miss work or school beyond the duration of symptoms or hospitalization or have decreased productivity while recovering. In addition, we did not include the use or cost of self-treatment or home care (eg, over-the-counter medicines, diapers, or oral rehydration) because these can be highly variable.
In conclusion, our results suggest that the economic burden of norovirus is $10.6 billion, based on the latest US incidence numbers (estimated median, 68.9 per 1000 population [20]). Sporadic cases in the community constitute >90% of the total annual cost, suggesting a need to better track and prevent such cases. Productivity losses, mostly due to absenteeism, represented 89% of total costs and were the largest driver of annual cost, suggesting that employers and businesses are bearing the brunt of the burden. More than half the economic burden is concentrated in adults aged ≥45 years, more than half occurs during the winter months, and >90% of outbreak costs are due to person-to-person transmission, offering insights into where and when prevention and control efforts may yield the greatest returns.
Supplementary Data
Supplementary materials are available at The Journal of Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.
Notes
Disclaimer. The study sponsors did not have any role in the study design, collection, analysis and interpretation of data, writing of the report, or the decision to submit the report for publication.
Financial support. This work was supported by Vaxart, the Agency for Healthcare Research and Quality (grant R01HS023317), the Eunice Kennedy Shriver National Institute of Child Health and Human Development Office of Behavioral and Social Sciences Research (grant U54HD070725), and the United States Agency for International Development (agreement AID-OAA-A-15-00064).
Potential conflicts of interest. All authors: No reported conflicts. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.
