Cost of illness of the vaccine-preventable diseases influenza, herpes zoster and pneumococcal disease in France

Abstract

Background

The incidence of certain vaccine-preventative diseases, such as influenza, herpes zoster and pneumococcal infection, continues to be high despite the availability of vaccines, resulting in a substantial health and economic burden on society, particularly among older adults aged ≥65 years.

Methods

A cost calculator was developed to assess the cost of illness of influenza, herpes zoster and pneumococcal disease in France. Direct medical costs related to diagnosis and treatment in the older adult population in both inpatient and outpatient settings were modelled over a 1-year time horizon. Scenario analyses were conducted to determine the impact of hospitalizations on the results by considering only influenza-attributed diagnoses.

Results

In France, influenza has the highest incidence, followed by herpes zoster and pneumococcal disease. Similarly, influenza poses the greatest cost burden among all older adults, while pneumococcal disease poses the greatest cost burden among those aged 65–74 years. When considering only influenza-attributed diagnoses, the number of inpatient visits and associated costs was reduced by 63% in the overall older adult population. In the low-incidence season, the number of inpatient visits and associated costs were reduced by 69%, while in the high-incidence season, the number of inpatient visits and associated costs increased by 63%.

Conclusion

Influenza remains a leading vaccine-preventable disease among older adults in France, resulting in a substantial economic burden that could be prevented by increasing vaccine uptake.

Introduction

Immunization is a successful global health intervention that has resulted in the prevention of 24 diseases.¹ The leading vaccine-preventable diseases (VPDs) include influenza, herpes zoster (HZ) and pneumococcal infection. While these infections can present mildly, they are also associated with severe complications, including encephalopathy, septicaemia and death.^2–5 Despite the availability of vaccines, some VPDs, particularly influenza, continue to have high incidence rates.⁶ Due to constant viral mutation, influenza outbreaks manifest annually. Worldwide, an estimated 1 billion cases of influenza occur each year, with 3–5 million severe cases.⁷ In the European Union, seasonal influenza causes up to 50 million symptomatic infections and 15 000–70 000 deaths annually.⁸ Each year, influenza causes 2–6 million infections and approximately 10 000 deaths in France.⁹

VPDs are associated with a substantial economic burden. In the USA, in 2015, 16.6 million cases of influenza, 1.1 million cases of HZ and 283 000 cases of pneumococcal disease were estimated to lead to $5.79 billion, $782 million and $1.86 billion in costs, respectively, among adults aged ≥19 years; of these costs, $7.1 billion (80%)¹⁰ were attributable to unvaccinated persons. Severe complications due to VPDs are more common in older adults aged ≥65 years, who are more likely to require hospitalization and face higher morbidity and mortality.¹¹ Among older adults aged ≥50 years, the estimated combined annual costs of four VPDs, including influenza, pneumococcal disease, HZ and pertussis, are $26.5 billion; of these costs, $15.3 billion were incurred by those aged 65 years.¹²

However, in France, vaccine coverage among older adults (defined as those aged ≥65 years¹³) is very low (5–20%),¹⁴ and the coverage of the influenza vaccine, which has the highest uptake rate among vaccines for VPDs, in 2020 was 59.9%.¹⁵ The coverage for pneumococcal disease in France is even lower (varying between 5% and 20%).¹⁶ While HZ data are collected by the French Sentinel Surveillance Reporting System based on consultations with general practitioners (GPs), national surveillance is lacking.¹⁷ Despite the low vaccination rates, systematic vaccination is not recommended for all older adults in France and is only recommended for those with compromised immune functioning or severe chronic illnesses, such as heart disease or diabetes.^14,16

Cost of illness (COI) studies aim to estimate and compare the economic burden of diseases on society by quantifying all costs related to an illness.¹⁸ In this study, a cost calculator was developed to evaluate the COI and economic burden of the three most prevalent VPDs, namely, influenza, HZ and pneumococcal disease, in the older adult population in France and determine the value and benefits of immunization according to the disease relative COI weight. The study aims to encourage decision-makers to implement effective healthcare policies and reimbursement strategies that mitigate the high COI related to VPD in France.

Methods

Model description

A cost calculator was developed in Microsoft Excel^® to assess the COI of influenza, HZ and pneumococcal disease in France. The model structure is fully transparent and has the flexibility to model each VPD in specified age groups. The model accounts for direct medical costs, including costs related to diagnosis and other treatment services linked to a particular diagnosis of VPD and costs related to outpatient (defined as not requiring hospitalization) and inpatient care in a 1-year time horizon. The time horizon was set to 1 year due to the acute nature of the VPDs considered in this study, and the model does not consider costs beyond the year of infection.

To estimate the total annual costs associated with each VPD, the core model concept was derived from a simple multiplication rule of $A=B\times C$, where A is the total annual cost associated with a given VPD, B is the estimated annual number of a given VPD and C is the estimated cost per case of a given VPD. Parameters B and C were derived as follows:

$$\begin{array}{c}B=\left(\mathrm{age}\mathrm{ }\mathrm{specific}\right)\mathrm{ }\mathrm{est}.\mathrm{ }\mathrm{annual}\mathrm{ }\mathrm{incidence}\mathrm{ }\mathrm{rate}\mathrm{ }\\ \times \hspace{0.17em}\left(\mathrm{age}\mathrm{ }\mathrm{specific}\right)\mathrm{ }\mathrm{population}\mathrm{ }\mathrm{size}\\ C=\left(\mathrm{age}\mathrm{ }\mathrm{specific}\right)\mathrm{ }\mathrm{est}.\mathrm{direct}\mathrm{ }\mathrm{cost}\mathrm{ }\mathrm{per}\mathrm{ }\mathrm{case}\hfill \\ +\hspace{0.17em}\left(\mathrm{age}\mathrm{ }\mathrm{specific}\right)\mathrm{ }\mathrm{est}.\mathrm{indirect}\mathrm{ }\mathrm{cost}\mathrm{ }\mathrm{per}\mathrm{ }\mathrm{case}\end{array}$$

Due to limitations of the available data, parameters B and C were derived from disaggregated inputs, and the micro-costing method as described in the ‘Model Inputs’ section.

The analysis was conducted from a payer perspective by considering the direct costs per disease case of influenza and HZ.

Target population

The target population in this analysis included people aged ≥65 years in France. The population in the model was further stratified into those aged 65–74 years and those aged ≥75 years. The model does not further segregate the population into subgroups based on gender, severity of disease, comorbidities, immunocompetence or other characteristics.

Model inputs

Due to limitations of the currently published sources containing case numbers (incidence) and relevant costs per case that could be used to estimate the overall economic burden, the model uses a micro-costing approach in which disaggregated inputs from multiple sources were collected and used to estimate the outcome. Micro-costing allows for a more precise assessment of the economic costs and is particularly effective in estimating the costs of new interventions, the costs of interventions with large cost variations among providers, and accurate costs for healthcare systems and society.¹⁹ As a part of the micro-costing approach, a targeted literature review was conducted to obtain relevant inputs, including the age-specific population size, employment percentages, age- and VPD-specific incidence, and VPD-specific direct and indirect costs in the outpatient setting and by hospitalizations, for each of the specified VPDs. The literature review was conducted in 2019, and the search was restricted to articles focusing on international surveillance datasets and national surveillance systems. Additional details of the search terms and results of the targeted literature review are provided in Supplementary table S1, and the abstract and full-text screening results are shown in Supplementary table S2.

Population size

The population size by age group in France was informed by the Institut National de la Statistique et des Etudes Economiques.²⁰

Incidence rates

The incidence rates in the model reflect those of healthcare utilization rather than those at the disease level. We purposefully omitted asymptomatic cases that did not require medical attention based on the assumption that these cases did not incur costs to the healthcare system. Therefore, due to the disaggregation of cases by type, the model does not provide a total number of cases as this may cause double counting. Not all incidence values could be sourced for each age category; thus, appropriate calculations based on weighted averages were used for these missing values based on sourced data. Furthermore, in the literature, the incidence rates are reported at heterogeneous time points, rendering a comparison of disease attack rates in a constant reporting year challenging.

Published evidence was used to inform the incidence of influenza in France in the inpatient setting²¹ and the outpatient setting.^22–24 The base-case analysis uses respiratory-cause excess hospitalization data.

Regarding the incidence of pneumococcal disease in France, the incidence of meningitis,²⁵ bacteraemia²⁵ and non-invasive pneumonia²⁶ in the inpatient setting and non-invasive pneumonia^26,27 in the outpatient setting were considered.

Regarding the incidence of HZ in France, the inpatient incidence rate was not available in the literature. Therefore, the following incidence data were considered: outpatient incidence²⁸ and percentage of HZ cases developing post-herpetic neuralgia (%PHN/HZ).²⁹

Costs

The model considers direct medical costs (i.e. costs related to diagnosis and treatment) and costs related to outpatient and inpatient care. In most cases, an overall cost per case per disease was not available; therefore, the costs were disaggregated by inpatient and outpatient status using a micro-costing approach. All cost inputs were homogenously inflated to 2020 values and are expressed in Euros.

Regarding influenza-associated costs in France, the outpatient costs were informed by Uhart et al.,³⁰ while the inpatient costs were informed by Fouad et al.²¹

Regarding the pneumococcal disease-associated costs in France, the costs associated with meningitis,³¹ bacteraemia³¹ and non-invasive pneumonia³² in the inpatient setting and those associated with non-invasive pneumonia in the outpatient setting²⁷ were considered.

The direct costs associated with HZ in France were informed by Belchior et al.³³ A summary of the incidence and cost inputs is provided in Supplementary table S3.

Scenario analyses

Three scenario analyses of influenza were considered in this study. To understand the impact of hospitalizations on the results, in the first scenario analysis, the burden of influenza was estimated when the incidence rate estimates among inpatients were restricted to directly influenza-attributed diagnoses (International Classification of Diseases [ICD] codes J09-J11). Since the incidence of influenza greatly varies across seasons, in the second and third scenario analyses, the epidemiological and financial burden of influenza were estimated during a low-incidence influenza season (i.e. 2013–14) and a high-incidence influenza season (i.e. 2016–17), respectively.

Model validation

Three scenario analyses of influenza were considered in this study. To understand the impact of hospitalizations on the results, in the first scenario analysis, the burden of influenza was estimated when the incidence rate estimates among inpatients were restricted to directly influenza-attributed diagnoses (ICD codes J09-J11). Since the incidence of influenza greatly varies across seasons, in the second and third scenario analyses, the epidemiological and financial burden of influenza were estimated during a low-incidence influenza season (i.e. 2013–14) and a high-incidence influenza season (i.e. 2016–17), respectively.

Results

Base-case results

The total incidences of medically attended influenza, pneumococcal and HZ episodes in France are shown by different age groups in figure 1. The pattern is consistent in all age groups. Influenza has the highest incidence, followed by HZ and pneumococcal disease. Tables 1–3 report the incidences and direct costs of each disease by inpatient and outpatient visits. The incidence of HZ is based on outpatient visits, and thus, no direct costs associated with inpatient visits are reported.

Figure 1.

Total incidence of medically attended influenza, pneumococcal, and HZ episodes by age group in France

Table 1.

Incidence and costs of influenza

Concept category	Age category
	65+	65–74	75+
Annual incidence of inpatient visits	26 438	6266	18 903
Annual incidence of outpatient visits	135 911	72 915	62 996
Total annual incidence	162 349	79 181	81 899
Direct costs: inpatients	€130 606 528	€29 726 781	€97 667 313
Direct costs: outpatients	€3 291 662	€1 765 943	€1 525 720
Total inpatient costs	€130 606 528	€29 726 781	€97 667 313
Total outpatient costs	€3 291 662	€1 765 943	€1 525 720
Total costs	€133 898 190	€31 492 723	€99 193 033

Table 2.

Incidence and costs of pneumococcal disease

Concept category	Age category
	65+	65–74	75+
Annual incidence of inpatient visits	12 350	4472	10 019
Annual incidence of outpatient visits	29 564	15 861	13 703
Total annual incidence	41 914	20 333	23 722
Direct costs: inpatients	€94 112 146	€33 975 041	€76 551 455
Direct costs: outpatients	€3 730 307	€2 001 271	€1 729 036
Total inpatient costs	€94 112 146	€33 975 041	€76 551 455
Total outpatient costs	€3 730 307	€2 001 271	€1 729 036
Total costs	€97 842 453	€35 976 312	€78 280 491

Table 3.

Incidence and costs of HZ

Concept category	Age category
	65+	65–74	75+
Annual incidence of inpatient visits	0	0	0
Annual incidence of outpatient visits	129 893	58 436	71 457
Total annual incidence	129 893	58 436	71 457
Direct costs: inpatients	€0	€0	€0
Direct costs: outpatients	€37 716 629	€13 136 036	€26 172 185
Total outpatient costs	€37 716 629	€13 136 036	€26 172 185
Total costs	€37 716 629	€13 136 036	€26 172 185

In France, the average annual cost of influenza, pneumococcal disease and HZ per person aged ≥65 years is €9.74, €7.12 and €2.74, respectively. Influenza has the highest cost burden in the overall older adult population and those aged ≥75 years. Within the 65- to 74-year age group, the highest cost burden is associated with pneumococcal disease by a narrow margin over influenza. This shift is due to the much higher cost per case of pneumococcal disease (both inpatient and outpatient) vs. influenza outweighing the lower incidence. In all age groups, HZ has the lowest cost burden, which was expected due to the assumption that the management of cases is only within the outpatient setting.

Scenario analyses

Three scenario analyses affecting the incidence of influenza in France were conducted. The scenarios do not affect the outpatient outcomes of influenza or the outcomes of HZ or pneumococcal disease; therefore, these results are the same as those summarized in the base-case analysis.

The first scenario considers the burden of seasonal influenza between 2010 and 2018 when the incidence rate estimates of inpatients are restricted to those with directly influenza-attributed diagnoses (ICD codes J09-J11). Compared to the base-case analysis, the number of inpatient visits and associated costs are reduced by 63% in those aged ≥65 years. The model-generated estimates of the cases and associated costs are presented in Supplementary tables S4 and S5.

The second scenario considers the epidemiological and financial burden of a low-incidence influenza season (i.e. 2013–14). The expected incidence and associated costs of influenza in France by age group obtained using data from the year with the lowest influenza incidence are presented in Supplementary tables S6 and S7. Compared to the base case, the number of inpatient visits and associated costs for those aged ≥65 years are reduced by 69%.

The third scenario considers the epidemiological and financial burden of a high-incidence influenza season (i.e. 2016–17). The expected incidence and associated costs of influenza in France by age group obtained using data from the lowest influenza year are presented in Supplementary tables S8 and S9. The results indicate that compared to the base case, the number of inpatient visits and associated costs for those aged ≥65 years increased by 63%. Comparing the second and third scenarios, the high-incidence season resulted in an economic burden 3.8 times higher than that in the low-incidence season. This finding demonstrates that the influenza burden could be much more important during a high-incidence season compared to the mean burden evaluated in the base case and illustrates the value of an effective and well-matched vaccine with high uptake rates.

Discussion

Summary of the results

The current analysis evaluated the economic impact of influenza, HZ and pneumococcal disease and highlighted the significant economic burden associated with these VPDs among older adults. Here, the results of the base-case analysis are discussed, followed by a discussion of the results of the scenario analyses. Our conservative estimates demonstrate that influenza is the leading cause of VPDs in older adults. The cost burden of influenza is higher than that of pneumococcal disease and HZ in all age groups examined, except for the 65- to 74-years age group in which the cost burden associated with pneumococcal disease was slightly higher than that associated with influenza. In France, inpatient costs constituted most of the total costs associated with both influenza and pneumococcal disease in all age groups. Costs for HZ were only accrued in the outpatient setting. The greatest epidemiological and economic burden was associated with influenza, while the associated total economic burden of pneumococcal disease was lower due to the significantly lower cost per case. However, the clinical burden associated with HZ was higher than the clinical burden associated with pneumococcal disease.

The results of this model are qualitatively consistent with those reported by Ozawa et al.¹⁰ and McLaughlin et al.,¹² who noted that among older adult patients in the United States, influenza resulted in the most significant burden, followed by pneumococcal disease and HZ when considering both the direct medical and societal perspectives. Another similar study involving older adult patients in the USA³⁴ revealed that pneumococcal disease led to the greatest burden by a narrow margin; these results were driven by estimates of non-invasive pneumococcal disease incidence rates an order of magnitude higher than those used by McLaughlin and in our study.

When specifically considering influenza, our results are consistent with the global burden review of direct costs by Peasah et al.³⁵; however, it is impossible to directly compare our results with the results of this study due to differences in the populations considered (general vs. older adults).

No recent studies could be found in the literature to comprehensively assess the external validity of the modelling results of pneumococcal disease and HZ in France. Regarding pneumococcal disease, in France, the model estimates that the direct cost burden of non-invasive pneumonia in older adults is approximately €85 million. Personne et al.²⁷ estimate that the burden of community-acquired pneumonia in all ages is approximately €97 million (€103 million in €2020). While the proportion of the overall burden assumed by older adults estimated by the model could potentially seem high, the analysis by Personne only considered those initially diagnosed by a GP, and emergency department admissions and direct hospitalizations were not considered; therefore, as noted in their discussion, their data likely reflect underestimations, especially among older adults.

Regarding HZ, no studies utilizing recent data were found for comparison. In 2010, Mick et al.³⁶ reported that the estimated annual cost of the management of HZ and post-herpetic neuralgia in France in those aged ≥50 years was €170 million (€192.83 million in €2020), of which €61.0 million (€69.19 million in €2020) were covered by the national health insurance. This finding suggests that our estimates of €37.72 million in direct payer costs among those aged ≥65 years may be conservative.

In the scenario analysis in which the incidence of influenza was restricted to inpatients with influenza-attributable diagnoses (ICD codes J09-J11), a 63% reduction was observed in the number of inpatient visits and associated costs among older adults aged ≥65 years in France. These results illustrate the need to consider influenza-related respiratory inpatient cases rather than simply hospitalization due to influenza ICD codes to properly represent the disease burden of influenza.

Considering the variability in influenza incidence by season, in the scenario analyses, the influenza burden was assessed during a high-incidence season and a low-incidence season. During the low-incidence season, a 69% reduction was observed in the number of inpatient visits and associated costs, while a 63% increase was observed in the high-incidence season. These results illustrate that differing from other VPDs, the burden of influenza can significantly vary, and during a high-incidence season, such burden can be 3.8 times higher than that in a low-incidence season.

Limitations

Despite the conservative modelling approaches applied in this study, several limitations are worthy of consideration. First, input data for the same age groups as those used in the model were not always available; for example, the HZ incidence in France was available for those aged 60–69 years and 70–79 years, requiring an estimation of the incidence in the 65- to 75-year age group in the model by averaging. As the distribution within these age categories is unknown, there is a risk that the model underestimated the results in the 65- to 70-year age group and overestimated the results in the 70- to 75-year age group. Second, the results of older adults are reported at the country level and not per capita. Differences in the population size and distribution by age directly impact the results. Third, from the societal perspective, data from the literature regarding the indirect cost per disease case of influenza and HZ in France were not available. Therefore, the economic burden of VPDs from a societal perspective could not be directly estimated based on the currently available data. However, the model functionality allows for the inclusion of indirect costs in France when such data become available. This shortcoming could be overcome by the utilization of other estimates or conservative assumptions. For example, absenteeism can be estimated based on the length of hospital stay (as reported in previous studies^9,37) and then multiplied by lost wages after adjusting for the employment rate and working days. Furthermore, since employment is reduced in the older adult population, adopting only a public payer perspective may be considered. Fourth, the incidence rates reported in the literature were obtained at heterogenous time points, rendering a comparison of disease attack rates at a constant reporting year challenging. Fifth, the scenario analyses were only performed in France; therefore, generalization of the results to other countries in Europe should be performed with caution. Sixth, since the influenza burden was evaluated by influenza-related respiratory hospitalizations, pneumonia may have been counted twice as influenza cases and pneumococcal disease cases. Seventh, the interplay resulting from successive viral and bacterial infections plays an important role in the epidemiological outcomes of many respiratory pathogens. For example, the influenza virus has been linked to the pathogenesis of various bacterial pathogens typically associated with pneumonia.³⁸ Such interactions were not considered in this study, potentially resulting in an underestimation of the disease burden. Eighth, this study sought to demonstrate the value of disease prevention, and thus, the cost of vaccination was not considered in the analysis. To expand upon this study, future studies could consider incorporating vaccination costs and conducting cost-effectiveness analyses. Finally, following the coronavirus disease 2019 (COVID-19) outbreak, influenza infection trends have exhibited significant variability. Given the measures adopted worldwide to limit the transmission of SARS-CoV-2, such as social distancing, social closures, and teleworking, influenza activity has significantly declined.³⁹ To avoid such variability, pre-COVID-19 data were used in our analysis.

Conclusion

This research highlights the significant economic burden associated with VPDs in older adults. Despite accounting for confirmed cases across influenza-like medical visits, our conservative estimates demonstrate that influenza is the leading cause of VPDs in older adults, causing a substantial economic burden that could be prevented by an increase in immunization in France, where vaccine uptake is suboptimal compared to that in Ireland and Denmark (75.4% and 78%, respectively¹⁵).

Supplementary data

Supplementary data are available at EURPUB online.

Funding

This study was funded by Sanofi.

Conflicts of interest

J.B.-H. and A.P. are employees and stakeholders of Sanofi. L.B., K.S., C.B. and J.R. received consultancy fees for the development of this study by Sanofi.

Key points.

• The incidence of vaccine-preventable diseases continues to be high despite the availability of vaccines.

• A cost calculator was developed to assess the cost of illness of influenza, herpes zoster, and pneumococcal disease in France among older adults aged ≥65 years.

• Influenza is the leading cause of vaccine-preventable diseases and is associated with the highest economic burden among older adults.

• Increasing vaccine uptake could alleviate the substantial economic burden caused by influenza.

Data availability

Aggregate data generated or analyzed during this study are available upon request.