Estimating the health economic impact of the universal varicella vaccination in Argentina from 2015 to 2019

ABSTRACT

Since the introduction of Universal Varicella Vaccination (UVV) in the Argentinean National Immunization Program in 2015, a significant decline in the incidence of varicella has been reported. This study aimed to estimate the economic impact of single-dose UVV in Argentina from 2015 to 2019. The economic impact was assessed based on the observed incidence of varicella in the post-UVV period and the number of cases avoided, obtained from a previously published study that used an Autoregressive Integrated Moving Average (ARIMA) model. The weighted average cost per case was calculated using local studies. The post-UVV cost reductions were calculated by multiplying the number of cases avoided from 2015 –2019 by the weighted average cost per case. Data were summarized yearly and by peak (September−November) periods for the target (1−4 years) and overall populations. We estimated avoided costs of United States dollars (USD) $65 million in the target population and $112 million in the overall population over 4 years following UVV introduction. We observed a trend toward greater reductions in costs over time, with substantial differences observed in peak periods. We estimated that the single-dose UVV program considerably reduced the economic burden of varicella in Argentina by avoiding direct and indirect costs associated with varicella management.

KEYWORDS:

• Varicella vaccination
• economic impact
• direct costs
• indirect costs
• time series
• ARIMA

Introduction

The implementation of Universal Varicella Vaccination (UVV) in Latin American countries has significantly reduced the clinical burden of varicella.¹ However, data on the economic impact of UVV are limited, presenting opportunities for new evidence to inform healthcare decision-making. Before the introduction of UVV in Argentina in July 2015,² the total cost for varicella management was estimated at United States dollars (USD) $40 million among children aged 1–14 years.³ Here, we estimated total cost reductions, based on varicella cases avoided since UVV was introduced in Argentina, in the 1–4 years age group (target population) and overall population.

Methods

A mid-term economic impact analysis⁴ of UVV in the 4 years following implementation was performed from a societal perspective in Argentina to generate costs associated with the incidence reductions attributed to UVV.

Epidemiological impact

Varicella incidence and cases avoided following UVV implementation were obtained from a previously published study that used an Autoregressive Integrated Moving Average (ARIMA) model to compare the observed incidence post-UVV to the predicted incidence from a hypothetical scenario without UVV,⁵ using national varicella surveillance data in Argentina between 2008 and 2015. Considering the strong seasonality of the pre-UVV data, a seasonal component was considered in the ARIMA model, thus resulting in predicted values with seasonality. The study estimated a reduction in 144,811 cases for the target population and 249,090 for the overall population from July 2015 to December 2019 (Table 1) by estimating base values and prediction intervals (PIs).

Table 1. Number of cases avoided and weighted cost estimates used for the economic analysis of universal varicella vaccination in Argentina, July 2015 to December 2019.

Input	Value	Source
Varicella cases avoided in target population (1–4 years of age)	144,811 (95% PI: −56,484 to 391,017)	Giglio et al.^Citation5
Varicella cases avoided in overall population	249,090 (95% PI: −186,013 to 784,111)	Giglio et al.^Citation5
Average cost per case, USD
Indirect, outpatient	$207	Giglio et al.^Citation3
Indirect, inpatient	$143.1	Giglio et al.^Citation3
Direct, outpatient	$322.7	Giglio et al.^Citation3
Direct, inpatient	$2,947.7	Giglio et al.^Citation3
Proportion of inpatient cases	4.9%	Neyro et al.^Citation6
Proportion of outpatient cases	95.1%	Neyro et al.^Citation6
Weighted average cost per case considering inpatient/outpatient proportion, USD
Indirect	$203.87
Direct	$247.46
Total	$451.33

PI: prediction interval; USD: United States dollar.

Costs per case

Direct medical and indirect (productivity loss) costs associated with outpatient and inpatient varicella cases were obtained from a retrospective study that evaluated healthcare resource utilization of 150 varicella patients from seven institutions across six cities in Argentina (Table 1).³ Then, the weighted average cost per case was calculated based on the proportion of inpatient cases (4.9%) as obtained by Neyro et al.⁶ The weighted costs used as inputs are summarized in Table 1. All costs reported were 2015 USD.

Economic impact of UVV

We estimated the economic impact of UVV on direct, indirect, and total costs by multiplying the difference between monthly incident cases observed versus predicted (without UVV)⁵ by the weighted average cost per case. Data were summarized on a yearly basis and, considering the seasonality of varicella in Argentina,⁵ during the peak period (September–November). PIs from the epidemiological estimation were used to estimate the PIs for the economic impact. We also estimated the economic impact per eligible child and per vaccinated child by dividing the economic impact for each year by the eligible and vaccinated population, respectively. We used both the cost reductions specific to the target population and the reductions estimated for the overall population.

Results

Following the introduction of UVV in July 2015, we estimated cost reductions in the target population starting from 2016 with increasing reductions over time (from $10.5 million [m] in 2016 to $21.0 m in 2019) (Table 2, Figure S1a). The cumulative cost variation for the whole post-UVV period (July 2015 to December 2019) was estimated to be -$65.4 m (95% PI: -176.5 m to 25.5 m), including -$35.8 m (95% PI: -96.8 m to 14.0 m) in direct costs and -$29.5 m (95% PI: -79.7 m to 11.5 m) in indirect costs. Higher relative reductions were estimated in varicella peak periods (Table 3). The highest annual variation was estimated in 2019 (-$21.0 m [95% PI: –47.8 m to 0.2 m]) with 44.4% of change occurring in the peak period (-$9.3 m [95% PI: -16.1 m to −2.6 m]) (Table 3).

Table 2. Annual national estimated varicella costs after universal varicella vaccination implementation in Argentina, July 2015 to December 2019 (USD).

Target population
		Predicted			Observed vs Predicted
Year	Observed	Mean	LPI	UPI	Mean	%	LPI	%	UPI	%
2015*	$18,164,425	$17,418,309	$7,813,466	$27,023,153	$746,115	4.3%	$10,350,959	38.3%	-$8,858,729	−32.8%
2016	$15,479,963	$25,936,863	$6,852,944	$49,584,690	-$10,456,900	−40.3%	$8,627,019	17.4%	-$34,104,727	−68.8%
2017	$12,285,040	$26,613,382	$6,389,170	$51,759,054	-$14,328,342	−53.8%	$5,895,870	11.4%	-$39,474,014	−76.3%
2018	$6,385,784	$26,675,296	$5,917,930	$52,667,018	-$20,289,512	−76.1%	$467,854	0.9%	-$46,281,233	−87.9%
2019	$5,602,286	$26,630,613	$5,451,221	$53,359,490	-$21,028,327	−79.0%	$151,065	0.3%	-$47,757,204	−89.5%
Total	$57,917,498	$123,274,463	$32,424,730	$234,393,405	-$65,356,966	−53.0%	$25,492,768	78.6%	-$176,475,907	−75.3%
Overall population
		Predicted			Observed vs Predicted
Year	Observed	Mean	LPI	UPI	Mean	%	LPI	%	UPI	%
2015*	$46,430,059	$41,052,412	$20,954,403	$61,150,420	$5,377,648	13.1%	$25,475,656	41.7%	-$14,720,361	−24.1%
2016	$43,379,102	$65,472,977	$25,809,320	$113,640,764	-$22,093,875	−33.7%	$17,569,783	15.5%	-$70,261,662	−61.8%
2017	$37,926,645	$56,306,418	$15,410,244	$107,711,908	-$18,379,773	−32.6%	$22,516,401	20.9%	-$69,785,263	−64.8%
2018	$23,483,794	$60,849,502	$14,568,625	$118,039,236	-$37,365,708	−61.4%	$8,915,169	7.6%	-$94,555,443	−80.1%
2019	$23,463,935	$63,422,568	$13,988,468	$128,029,983	-$39,958,633	−63.0%	$9,475,468	7.4%	-$104,566,048	−81.7%
Total	$174,683,536	$287,103,877	$90,731,060	$528,572,312	-$112,420,341	−39.2%	$83,952,476	92.5%	-$353,888,776	−67.0%

*From July to December 2015.

LPI: lower prediction interval; UPI: upper prediction interval; USD: United States dollar.

Table 3. Estimated total costs of varicella management in the peak period (September to November) after universal varicella vaccination implementation in Argentina, July 2015 to December 2019 (USD).

Target Population
		Predicted			Observed vs Predicted
Year	Observed	Mean	LPI	UPI	Mean	%	LPI	%	UPI	%
2015*	$10,690,503	$10,535,542	$5,468,109	$15,602,976	$154,960	1.5%	$5,222,394	33.5%	-$4,912,473	−31.5%
2016	$6,326,203	$11,087,637	$4,959,434	$17,215,841	-$4,761,434	−42.9%	$1,366,770	7.9%	-$10,889,637	−63.3%
2017	$4,692,412	$11,158,183	$4,766,974	$17,549,392	-$6,465,771	−57.9%	-$74,562	−0.4%	-$12,856,980	−73.3%
2018	$2,155,973	$11,150,954	$4,562,766	$17,739,143	-$8,994,981	−80.7%	-$2,406,793	−13.6%	-$15,583,170	−87.8%
2019	$1,793,560	$11,127,138	$4,359,080	$17,895,196	-$9,333,578	−83.9%	-$2,565,520	−14.3%	-$16,101,636	−90.0%
Total	$25,658,651	$55,059,455	$24,116,363	$86,002,548	-$29,400,804	−53.4%	$1,542,289	6.4%	-$60,343,897	−70.2%
Overall Population
		Predicted			Observed vs Predicted
Year	Observed	Mean	LPI	UPI	Mean	%	LPI	%	UPI	%
2015*	$28,003,814	$24,487,664	$14,018,865	$34,956,462	$3,516,150	14.4%	$13,984,949	40.0%	-$6,952,649	−19.9%
2016	$18,822,058	$29,025,094	$16,679,210	$41,370,978	-$10,203,036	−35.2%	$2,142,848	5.2%	-$22,548,920	−54.5%
2017	$15,716,490	$23,945,989	$10,783,529	$37,108,449	-$8,229,498	−34.4%	$4,932,962	13.3%	-$21,391,959	−57.6%
2018	$8,182,522	$26,265,852	$10,971,387	$41,560,317	-$18,083,330	−68.8%	-$2,788,865	−6.7%	-$33,377,795	−80.3%
2019	$8,545,839	$27,677,178	$11,187,300	$44,167,057	-$19,131,339	−69.1%	-$2,641,461	−6.0%	-$35,621,218	−80.7%
Total	$79,270,723	$131,401,777	$63,640,290	$199,163,263	-$52,131,054	−39.7%	$15,630,433	24.6%	-$119,892,540	−60.2%

*From July to December 2015.

LPI: lower prediction interval; UPI: upper prediction interval; USD: United States dollar.

In the overall population, we also estimated cost reductions starting from 2016 with increasing reductions over time (from $22.1 m in 2016 to $40.0 m in 2019) (Table 2, Figure S1b). The total cumulative cost variation was -$112.4 m (95% PI: -353.9 m to 84.0 m) from July 2015 to December 2019. Direct costs accounted for -$61.6 m (95% PI: -194.0 m to 46.0 m) and indirect costs for -$50.8 m (95% PI: –159.9 m to 37.9 m). Higher relative cost reductions were again observed in the peak varicella period (Table 3). Differences between the observed and predicted values were highest in 2019, a variation of -$40.0 m (95% PI: –104.6 m to 9.5 m) in the entire year (63% reduction), with a 69% reduction in the peak period (Table 3).

Yearly cost reduction per eligible child and per vaccinated child are presented in Table S1. These values showed a similar progression to the overall cost differences with higher cost reduction per child over time. When considering cost reductions specific to the target population, the cost reduction per eligible child was $28.21 and per vaccinated child was $36.31. In the overall population, the cost reduction per eligible child was $53.61 and per vaccinated child was $69.00.

Discussion

We estimated avoided costs of $65.4 m (target population) and $112.4 m (overall population) over a four-year period following UVV implementation in Argentina, which was primarily attributed to cases avoided, especially in the target population to be vaccinated. The economic impact on the overall population may be attributed to herd immunity, whereby indirect benefits of childhood varicella vaccination are observed in the population who are not vaccinated. A substantial proportion of the overall UVV impact was observed during peak periods, due to varicella seasonality in Argentina. This may be relevant as a reduction in the burden of varicella during peak periods could free up potential healthcare resources to address other seasonal diseases. The benefit of UVV will primarily be observed during peak periods, with a smaller effect over the rest of the year.

An increasing trend in avoided costs was observed over time in the post-UVV period due to higher estimated reductions in incident cases. This may be attributed to more cohorts becoming eligible for vaccination along with increases in vaccination coverage rates over time.

A previous study estimated that, in 2015, varicella was associated with overall costs of $40 m in children ≤14 years of age before UVV introduction.³ This number is consistent with our estimated 2016 costs without UVV: $25.9 m for children aged 1–4 years and $65.5 m for the overall population (all age groups). This supports the accuracy of the cost predictions from our model which are consistent with data reported in the published literature.

In this study, the estimated costs associated with varicella were based on several assumptions, therefore it is important to consider limitations and uncertainties related to the estimated costs. We extrapolated the cost per case based on a study in patients aged between 0 and 15 years to the adult population.³ Nonetheless, the total economic burden is likely underestimated since varicella is usually more severe, and associated with higher rates of complications in adults compared to children.⁷ This study did not consider costs associated with acquisition and administration of vaccines and immunization infrastructure; these costs are often integrated into primary care services, thus difficult to estimate. As UVV has been an established policy in Argentina since 2015, the outcomes of this study could inform the potential benefits derived from UVV. All clinical inputs were derived from a previously published UVV impact study.⁵ Thereby, all strengths and limitations of the statistical approach from that study are inherited for the present estimations. First, the data sources only provided varicella cases that were diagnosed and reported in the official systems, hence there is potential for under-reporting of mild cases. The ARIMA model employed an ecological design which implies a few limitations. We did not have individual-level confounder variables that we could control for. The estimations reported must also be interpreted as population-level effects and not as individual effects. PIs for economic analysis are based upon varicella incidence obtained from the ARIMA model. The incidences in the non-peak periods were much closer to 0, hence the PIs could fall to negative as a statistical artifact. Of note, this study was conducted in the pre-COVID-19 period and provided a baseline for future evaluations. Intensification of vaccination campaigns will be critical to avoid increasing numbers of susceptible individuals due to vaccination disruptions during 2020–2022 COVID-19 waves.⁸

Our study estimated that UVV reduced the economic burden of varicella in the target and overall populations in Argentina by avoiding direct and indirect costs associated with varicella management, driven by cases avoided after UVV implementation. It is expected that increasing trends in health and economic benefits of UVV will be observed with long-term follow-up.

Author contribution’s

Norberto Giglio, Manjiri Pawaskar, Cintia I. Parellada, Virginia V. Babic, and Silvina E. Neyro all contributed to Conceptualization, Methodology, Writing – Review & Editing. Pieralessandro Lasalvia and Diego Rosselli contributed to Methodology, Software, Validation, Formal Analysis, Data Curation, Writing – Review & Editing. All authors approved the final article for publication.

Ethics approval statement

Not applicable as all data were extracted from public and structured databases containing anonymized and aggregated data.

Acknowledgments

We acknowledge Shikha Surati, MPH, and Claudia Beltran, MD, for providing administrative support, and Adelphi Values PROVE for providing editorial support.

Disclosure statement

Norberto Giglio is a medical scientific consultant and received honoraria from Merck Sharp & Dohme LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ, USA. Manjiri Pawaskar, Virginia Verdaguer Babic, and Cintia Irene Parellada are employees of Merck Sharp & Dohme LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ, USA, who may own stock and/or hold stock options in Merck & Co., Inc., Rahway, NJ, USA. Silvina Neyro is an employee of Dirección de Control de Enfermedades Inmunoprevenibles (DiCEI), Ministerio de Salud de la Nación, Argentina. Pieralessandro Lasalvia and Diego Rosselli are employees of Neuroeconomix who was contracted by Merck Sharp & Dohme LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ, USA to conduct the study.

Data availability statement

Third Party Data. Restrictions apply to the availability of these data. Data were obtained from the Ministerio de Salud de Argentina (Sistema Integrado de Información Sanitaria Argentino [SISA] and Sistema Nacional de Vigilancia en Salud [SNVS]) and Ministerio de Salud de la Nación (Dirección de Estadísticas e Información de la Salud [DEIS]) and are available at https://sisa.msal.gov.ar/sisa/and https://www.argentina.gob.ar/salud/deis/datos with the permission of the above parties.

Supplementary data

Supplemental data for this article can be accessed on the publisher’s website at https://doi.org/10.1080/21645515.2023.2245703.

Additional information

Funding

Funding for this research was provided by Merck Sharp & Dohme LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ, USA. Medical writing and editorial assistance were provided by Amy Sears and Ben Rousseau of Adelphi Values PROVE. This assistance was funded by Merck Sharp & Dohme LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ, USA.