FormalPara Key Summary Points

Why carry out this study?

Most studies estimating the economic burden of pediatric pneumococcal disease focus solely on direct medical costs and do not account for indirect non-medical costs associated with pediatric pneumococcal disease.

We conducted a reanalysis of a previous study where non-medical costs associated with caregiving for a child < 5 years old with pneumococcal disease are considered, and we subsequently calculated the annual indirect, non-medical economic burden of pediatric pneumococcal disease outcomes in 13 countries.

What was learned from the study?

The 20-valent pneumococcal conjugate vaccine (PCV20) serotypes contributed to a substantial societal burden of indirect non-medical costs, amounting to $413.97 million across 13 countries.

The inclusion of non-medical costs nearly tripled the total economic burden compared with only including direct medical costs.

This study can help inform decision-makers on the broader economic societal burden associated with PCV serotypes and the need for higher-valent PCVs with broader serotype coverage.

Introduction

We recently published a study that estimated the clinical and economic burden of pneumococcal disease attributable to vaccine serotypes in children under 5 years of age across 13 countries. Aggregately across all included countries, 20-valent pneumococcal conjugate vaccine (PCV20) serotypes were estimated to cause a total of 1,234,000 pneumococcal disease cases, amounting to $213.5 million in annual direct medical costs in children under 5 years of age [1]. However, our analysis adopted a healthcare payer perspective and considered only the direct medical costs incurred to the healthcare system [2]. We conducted a reanalysis where non-medical costs associated with caregiving for a child with pneumococcal disease are considered, and subsequently calculated the annual indirect, non-medical economic burden of pediatric pneumococcal disease outcomes in the 13 countries of interest.

Methods

Compliance with ethics guidelines

This article is based on previously published conducted studies and does not contain any new studies with human participants or animals performed by any of the authors.

Model Structure

The previously published decision-analytic model was adapted to include indirect non-medical costs. An extensive description of the model structure has previously been provided [1]. In brief, Australia, Austria, Canada, Finland, France, Germany, Italy, the Netherlands, New Zealand, South Korea, Spain, Sweden, and the UK were selected for the analysis due to well-established PCV national immunization programs (NIPs) and public available data sources. The annual clinical burden of 13 selected countries was informed by their estimated annual pneumococcal disease cases [inclusive of invasive pneumococcal disease (IPD), inpatient/outpatient pneumonia, and acute otitis media (AOM)] and deaths reported in Wasserman et al. [1]. Age- and country-specific incidence rates of pneumococcal disease, epidemiologic input parameters, and indirect costs were used to calculate the remaining economic societal burden in each country (Fig. 1). This decision analytic model is informed by previously conducted studies and does not contain any new studies with human participants or animals performed by any of the authors.

Fig. 1
figure 1

Model structure for estimating the remaining economic societal burden associated with PCV serotypes in each country

Full size image

The model was updated to employ a human capital approach for estimating indirect non-medical costs for caregivers of children with pneumococcal disease. The monetary value of caregiver productivity loss was measured by its opportunity cost of losing daily paid income. Therefore, only missed workdays from paid employment were accounted for in the societal cost calculation and were valued at the average daily pretax wage as reported in each country. All indirect costs were inflated to 2021 values in US dollars (USD).

Model Input and Assumptions

Input parameters were derived from published literature or other publicly available sources. As shown in Fig. 1, the model inputs derived for caregivers in each country included: (1) missed workdays: the number of workdays missed per event of childhood IPD, AOM, inpatient pneumococcal pneumonia, and outpatient pneumococcal pneumonia, (2) proportion of caregivers missing work: the proportion of caregivers who missed workdays from paid employment, assuming only one caregiver per infected child would miss work, (3) serotype-specific incidence: the annual clinical burden of respective pneumococcal disease outcome attributable to serotype coverages of licensed and investigational PCVs (i.e., PCV10, PCV13, PCV15, and PCV20) reported in Wasserman et al. [1], and (4) indirect non-medical costs: the indirect cost per disease case, estimated by the average number of missed workdays multiplied by the unit costs of productivity loss (average daily wage).

Table 1 summarizes the estimates of the number of missed workdays for caregivers of children with respective pneumococcal disease outcome. For IPD, the number of missed workdays for caregivers of children with bacteremia or meningitis was estimated to be 6.47 and 13.2 days, respectively. The number of missed workdays reported for inpatient pneumonia was estimated to be 8.6 days. For outpatient pneumonia and AOM, the number of workdays missed was estimated to be 3.3 days. We provide details regarding the calculations and sources used to derive these estimates below.

Table 1 Productivity loss model estimates for respective pneumococcal disease outcomes
Full size table

Estimates of Missed Workdays

IPD

For estimating indirect non-medical costs of IPD, 100% of IPD cases were assumed to be treated in a hospital setting and the caregiver was assumed to miss work for the entirety of the child’s length of stay (LOS) in hospital to provide care. Therefore, estimates of mean or median pediatric pneumococcal bacteremia or meningitis LOS were sourced, respectively, to estimate the total number of missed workdays for caregivers of children with IPD. Shiri et al. provided a systematic review on LOS for all pneumococcal diseases [3]. However, due to a lack of data measuring child hospitalization for bacteremia cases, the LOS for a bacteremia case was sourced from an Italian study (6.47 days) and was used as a proxy for all 13 countries. For meningitis, caregiver missed workdays associated with pediatric pneumococcal meningitis were calculated by taking the average LOS reported from five studies in the UK [4], Canada [5], Spain [6], Italy [7], and Austria [8] resulting in 13.2 days. The proportion of pneumococcal bacteremia and meningitis cases resulting in IPD for each country was taken from Wasserman et al. [1].

Pneumococcal AOM

For estimating indirect non-medical costs of pneumococcal AOM, all AOM cases were conservatively assumed to be outpatient and weighted by the proportion of caregivers reporting missed workdays for an AOM case. Barber et al. reported the mean number of missed workdays due to caregivers staying at home from a paid job and the proportion of caregivers reporting absent from work for a child’s AOM infection [9]. The mean missed workday estimate of 3.33 days was calculated using a weighted average of the workdays lost reported by the survey respondents [9]. Meanwhile, the study surveyed that 52% of the caregivers reported missing work for an AOM episode. This study included survey data from 6 countries included in our analysis—Germany, Canada, Australia, New Zealand, Spain, and the UK. The estimate was used as a proxy for all 13 countries.

Outpatient Pneumococcal Pneumonia

Due to a lack of existing literature on estimated indirect non-medical costs or productivity loss related to outpatient pneumococcal pneumonia, estimates of missed workdays and the proportion of caregivers reported to miss work for outpatient pneumonia were assumed to be the same as those for AOM, given both conditions are treated in a comparable setting. Therefore, as measured by Barber et al., a mean proportion of 52% of caregivers was assumed to miss a mean estimate of 3.33 workdays to care for children with outpatient pneumonia [9]. The proportion of pneumonia cases resulting in outpatient care for each country was taken from Wasserman et al. [1].

Inpatient Pneumococcal Pneumonia

For estimating indirect non-medical costs for inpatient pneumococcal pneumonia, the caregiver was assumed to miss work for the entirety of the child’s LOS to provide care. Estimates for the mean LOS in hospital for pediatric inpatient pneumonia were sourced from Zhang et al., which included data from Ireland, Germany, Australia, and Spain [10]. The mean LOS estimate of 8.6 days was calculated based on a weighted average of the LOS reported in included articles, after excluding data from the USA. The proportion of pneumonia cases resulting in inpatient care for each country was taken from Wasserman et al. [1]

Economic Inputs

The average pretax annual income of each country was sourced from the Organization of Economic Co-operation and Development (OECD) Statistics and inflated to 2021 USD [11]. The average daily wage was calculated by applying 260 working days per year to annual income from the source reference. The indirect non-medical cost per respective pneumococcal disease outcome was calculated using daily wage rates specific to each country, the estimated number of workdays missed for each disease manifestation, and the proportion of caregivers reporting missed workdays to care for the infected children.

Sensitivity Analysis

A deterministic sensitivity analysis was performed to evaluate the impact of assumptions around key input parameters on model outcomes. The number of caregiver missed workdays, the proportion of caregivers absent from work to care for a child with inpatient pneumococcal disease, the proportion of children hospitalized for inpatient pneumococcal conditions, and cost data (average daily wages) were varied by ± 20%. Additionally, to assess the caregiver productivity loss estimates on outpatient pneumonia, we also explored the scenario where a lower estimate of the proportion of caregivers missing work and missed workdays sourced from Wolleswinkel-van den Bosch et al. (2010) were assessed in the sensitivity analysis [12].

Results

Of all the pneumococcal disease-related indirect non-medical costs, pneumococcal meningitis accounted for the highest indirect cost per case (Table 2), with an average cost of $2525 across 13 included countries. The average indirect cost per pneumococcal bacteremia and inpatient pneumonia case was $1238 and $1653, respectively, whereas costs associated with pneumococcal outpatient pneumonia and AOM were relatively lower, ranging from $366 to $789 per disease case in different countries.

Table 2 Indirect non-medical cost per case (USD) by respective pneumococcal disease outcome in children < 5 years in 13 selected countries
Full size table

The total annual indirect economic burdens associated with pediatric pneumococcal diseases attributable to PCV10, PCV13, PCV15, and PCV20 serotypes were $46.5 million, $159.0 million, $223.0 million, and $414.0 million, respectively (Table 3). Across all 13 markets, PCV20 serotypes contributed to a substantial societal burden of indirect non-medical costs, with PCV20-unique serotypes targeting an additional 9.0–69.2% ($0.4–71.5 million annually) of the total indirect burden incremental to other PCV serotypes. Five countries that use PCV10 in their NIPs bear a greater societal burden associated with PCV13 serotypes. The remaining total indirect non-medical costs due to PCV10-unique and PCV15-unique serotypes ranged from 1.5% to 14.4% and from 4.5% to 15.0%, respectively, whereas an additional 30.0% to 85.0% of the total indirect costs were attributable to PCV13-unique serotypes. In eight PCV13 NIP countries, the residual indirect burden was primarily attributable to non-PCV13 serotypes. PCV10-unique and PCV13-unique serotypes accounted for a comparatively lower proportion of the total indirect costs, estimated to be 3.0–18.9% and 11.2–28.8%, respectively. Australia had a higher distribution of PCV13-unique serotypes at 54.4%. An additional 1.9–33.0% of the indirect burden was due to PCV15-unique serotypes, while the most remaining indirect non-medical cost burden of 32.1–69.2% were primarily attributable to the additional PCV20-unique serotype.

Table 3 Annual estimated indirect non-medical costs due to pneumococcal disease attributable to PCV serotypes in children < 5 years in 13 selected countries
Full size table

Across pneumococcal disease outcomes, the total annual non-medical burden associated with remaining PCV20-type disease was $1.9 million, $49.1 million, and $363.0 million for IPD, pneumonia, and AOM, respectively (Table 3). Although IPD has the highest average indirect cost per case, IPD contributed to a relatively low proportion of the total societal burden (0.47%) compared with other pneumococcal disease outcomes. The greatest share of the total non-medical burden across all PCV serotypes was attributable to AOM (87.68%), due to high incidence rates among the pediatric population.

A deterministic sensitivity analysis was undertaken to assess the impact of assumption around key model parameters. A 20% increase or decrease in total non-medical costs was observed when varying the country-specific caregiver’s average daily wage or the number of caregiver missed workdays by 20%. When a 20% decrease was applied to the proportion of children hospitalized for pneumococcal meningitis, bacteremia, and inpatient pneumonia, the non-medical burden associated with PCV serotypes decreased by approximately 1.5%. Additionally, when the proportion of caregivers absent from work for an outpatient pneumococcal episode varied with a lower estimate of 21%, sourced from Wolleswinkel-van den Bosch et al., total non-medical costs decreased to $92.6 million for PCV15 serotypes and $173.2 million for PCV20 serotypes, respectively. Results of the sensitivity analyses for the total indirect costs and respective PCV serotypes are presented in Table 4.

Table 4 Deterministic sensitivity analyses results (in million USD)
Full size table

Discussion

We conducted an analysis to estimate the annual indirect non-medical costs associated with PCV10, PCV13, PCV15, and PCV20 serotypes in children under 5 years of age across 13 countries. Despite the success of PCV programs globally, our analysis demonstrates that there remains a substantial indirect non-medical burden of $414.0 million due to caregiver productivity loss resulting from pediatric pneumococcal disease attributable to PCV20 serotypes. Estimations of the economic burden from a healthcare payer’s perspective may greatly undervalue the burden associated with PCV serotypes. Wasserman et al. previously showed that PCV20 serotypes amount to $213.5 million in annual direct medical costs [1]. After implementing a societal perspective, the total estimated annual economic burden (direct medical and indirect non-medical costs) increased to $627.5 million for PCV20 serotypes and to $341.4 million for PCV15 serotypes.

Like many estimates of indirect costs, our analysis had some limitations. First, the scope of our indirect non-medical burden estimation is limited to include only the costs of caregivers missing workdays from paid employment. Not capturing labor market productivity gains, non-market productivity and leisure gains, caregiver leisure gains, and macroeconomic gains significantly underestimates the non-medical burden associated with pneumococcal disease [13, 14]. Second, few studies directly measured caregiver productivity loss associated with pediatric pneumococcal diseases. Therefore, we used the mean LOS for hospitalization associated with pediatric pneumococcal meningitis, bacteremia, and inpatient pneumonia as a proxy for estimating caregiver missed workdays. This estimation may not be an accurate assessment of productivity loss considering caregivers may continue to take time off from work once their child is discharged from hospital. Additionally, a lack of country-specific data may have resulted in caregiver missed workday estimates for pneumococcal bacteremia and outpatient pneumonia outcomes to be over- or underestimated for each country in the analysis. Finally, this study provides a conservative estimate because only one caregiver per infected child is assumed to miss work for providing care.

Conclusions

This study was a reanalysis of Wasserman et al. [1], where we expanded the analysis to include a societal perspective, demonstrating that including indirect non-medical costs nearly tripled the economic burden attributable to PCV20 serotypes in 13 countries. Inclusion of indirect costs in economic analyses of PCVs is necessary to ensure decision-makers are fully informed regarding the true economic and societal burden associated with PCV serotypes. Despite the success of the PCV13 pediatric NIPs, substantial burden of non-vaccine type pneumococcal disease remains. Higher-valent PCVs, which target serotypes known to frequently cause pneumococcal disease, have the potential to avert a substantial proportion of the total annual direct and indirect economic burden compared with lower-valent PCVs.