A retrospective study to assess the epidemiological and economic burden of pneumococcal diseases in adults aged 50 years and older in Taiwan

Abstract

Objective:

Invasive pneumococcal disease (IPD) and pneumococcal pneumonia cause substantial morbidity and mortality worldwide. This retrospective study was conducted to estimate the disease burden from pneumococcal disease in older adults in Taiwan from a health insurer’s perspective.

Methods:

Data for the years 2002–2009 from patients aged ≥50 years with insurance records indicating pneumococcal meningitis, pneumococcal bacteremia, or hospitalized or outpatient pneumonia were obtained from the National Health Insurance Research Database in Taiwan. Admission data for inpatients, visit data for outpatients, and associated costs were extracted from the database to estimate the incidence, case fatality rates, and direct and indirect costs of pneumococcal disease episodes. These data were applied to the estimated population of Taiwan in 2010 to provide an estimated disease burden for a single year from the payer perspective.

Results:

The average incidence per 100,000 person years was 2.4 for IPD, 278.8 for hospitalized pneumococcal pneumonia, and 1376.4 for outpatient pneumococcal pneumonia. The average case fatality rate was 12.3% for IPD and 10.0% for hospitalized pneumonia. Hospitalized pneumonia accounted for over 90% of direct medical costs. The incidence of hospitalized pneumococcal pneumonia per 100,000 person years was 84.4 for adults of 50–64 years, 313.1 for adults of 65–74 years, 820.3 for adults of 75–84 years, and 1650.9 for adults of 85+ year of age. In 2010, it was estimated there were over 113,000 episodes of pneumococcal disease, causing almost 2000 deaths, with direct medical costs of more than NT$3.4 billion annually.

Conclusions:

Pneumococcal disease is a significant cause of mortality and excess healthcare expense among the elderly in Taiwan. Disease burden in older adults increases with advancing age.

Keywords::

• Streptococcus pneumonia
• Incidence
• Cost
• Taiwan
• Invasive pneumococcal disease
• Pneumonia

Introduction

Diseases attributable to Streptococcus pneumoniae, including pneumococcal pneumonia and invasive pneumococcal disease (IPD), cause substantial morbidity and mortality worldwide, accounting for an estimated 1.6 million deaths in 2002, mostly in infants and the elderly1. Lower respiratory tract infections, including pneumonia, are the third leading cause of death worldwide, responsible for an estimated 3.8 million deaths in 20082. Globally, as many as 30–50% of cases of community-acquired pneumonia (CAP) are attributable to S. pneumoniae1,3. In developed countries, children and the elderly are at particularly high-risk for pneumococcal disease, and the risk of pneumococcal disease in adults increases with age and the presence of chronic diseases3,4.

In Taiwan, IPD is a notifiable disease monitored as part of national surveillance since October 2007. The incidence of IPD in October 2007 to September 2008 was highest in those aged ≥75 years (16.1 per 100,000 people) and children aged <5 years (15.0 per 100,000 people)5. Case fatality rates were greatest among the elderly, at 9% in persons aged ≥75 years compared with <2% in children aged <5 years5. Previous estimates in Taiwan have reported even higher IPD fatality rates in older adults, with case fatality rates of 42.5% in persons aged ≥65 years in 1998–1999, and 21.7% in elderly patients aged ≥70 years from the 2007 surveillance, with similar rates reported in other Asian countries6.

Pneumonia was the fourth leading overall cause of death in Taiwan in 20107. The rate of pneumonia deaths increases with age, and is highest among adults aged ≥65 years5. Patients with pneumonia, especially older patients, use considerable healthcare resources including hospitalizations, ambulatory care, and emergency room visits. Between 1997–2002 in Taiwan, males and females aged ≥65 years had an annual risk of being hospitalized for pneumonia of 3.64% and 1.87%, respectively8. The estimated cost of each hospital admission was US$3221. S pneumonia is the most common cause of hospitalized CAP in Taiwan, accounting for 24–26% of cases, with a higher prevalence in older adults9,10.

Nationwide vaccination programs against pneumococcal disease have the potential to reduce the incidence of illness and hospitalization and their associated costs11. Although 23-valent pneumococcal polysaccharide vaccine (PPSV23) is available in many countries and is generally recommended for use in the elderly, its efficacy against CAP is limited12. Moreover, the immunogenicity conferred by PPSV23 may decline over 3–5 years13,14. PPSV23 has been available in Taiwan since January 2001, and cumulative coverage of people aged ≥65 years reached 41% by 200815. A 13-valent pneumococcal conjugate vaccine (PCV13) is approved in many countries for use in adults. Whereas the immune response induced by polysaccharide vaccines is T-cell–independent, the response induced by pneumococcal conjugate vaccines is T-cell–dependent, a qualitatively different response that is expected to impart immunological memory in vaccinated subjects16,17. A recent analysis in the US reported favorable cost-effectiveness with PCV13 vs PPSV23 in adults, based on the enhanced effectiveness of PCV13 in preventing non-bacteremic pneumococcal pneumonia, as observed with the use of the pneumococcal conjugate vaccine in children11.

While the burden of pneumococcal disease in children has been estimated and reported18, data on the burden of pneumococcal disease in adults has not been described. This retrospective study was undertaken using the largest electronic database in Taiwan to estimate the incidence, mortality, and economic impact of pneumococcal disease in adults aged ≥50 years in Taiwan from the perspective of the national insurer. Studies of disease incidence can help guide the development of recommendations and funding policies for the use of pneumococcal vaccines in older adults.

Methods

Study design

This study was a retrospective analysis using data extracted from the National Health Insurance Research Database (NHIRD), a population-level administration-claims database in which 22.6 million of Taiwan’s 22.96 million population are enrolled. The database contains patient identification numbers, sex, date of birth, diagnostic codes, prescription drugs dispensed, registration fields, and original claims data for reimbursement of medical and pharmaceutical expenditures by enrollees. Case data for the years 2002–2009, including admission data for inpatients, visit data for ambulatory patients, and associated expenditures for each were extracted to estimate the numerator for calculations of incidence and case fatality rates and the episode cost. National census data were used to estimate the denominator for calculations of incidence.

Potential episodes of pneumococcal disease were identified by the International Classification of Diseases, Ninth Revision (ICD-9) codes for pneumococcal meningitis (code 320.1), pneumococcal bacteremia (codes 038.2, 038.9, and 790.7; excluding codes 482, 486, and 510.9), and pneumonia (codes 481–486 and 487.0). Invasive diseases from other sites, such as pleural effusion (empyema), synovial fluid (bone/joint infection), and ascites (peritonitis), were not evaluated from the database as they are generally rare, but are included in sources applied to sensitivity analyses. Due to limitations in the identification of pneumococcal pneumonia, the epidemiologic and economic burdens of pneumococcal pneumonia were both estimated to be 23.8% of those of all-cause pneumonia in both the hospital and outpatient settings9. Claims for pneumococcal pneumonia (code 481) were also identified to apply to alternative cost and case fatality rates in sensitivity analysis.

Subjects were all adults aged ≥50 years enrolled in NHIRD, and were stratified into age groups (50–64, 65–74, 75–84, and ≥85 years).

Outcomes assessed

We assessed the incidence rates of pneumococcal meningitis, bacteremia, and pneumococcal pneumonia (hospitalized and non-hospitalized); case fatality rates of hospitalized pneumococcal disease; and cost per episode of pneumococcal disease. The average incidence, case fatality rates, and cost of disease from 2002–2009 were applied to the population of Taiwan in 2010 to estimate the total burden of pneumococcal disease in a single recent calendar year.

Analysis of incidence

The incidences of pneumococcal diseases were calculated as the number of episodes with a medical claim of interest by age within each calendar year, divided by the corresponding population denominator. Patients could have more than one event during the calendar year; each event was counted as a separate episode. Claims of interest were hospitalizations with a primary discharge diagnosis of pneumococcal meningitis, pneumococcal bacteremia, or all-cause pneumonia, and outpatient all-cause pneumonia. Census data were used to estimate the denominator for calculations of incidence rate by year. Pneumonia-related claims and dispensing of antibiotics for the same patient within 30 days of each other were considered to be the same episode. Incidence rates were expressed as episodes per 100,000 person years.

Analysis of mortality

Discharge status on hospital claims was used to identify fatalities, when available. For each condition, estimates of case fatality were made by year and by age group. Case fatality rates were calculated as the number of inpatient deaths for the diagnosis of interest divided by the total number of episodes of that diagnosis and expressed as percentage mortality. No mortality was assumed for outpatient pneumonia.

Episode costs

The total cost per hospitalization, number of days of hospitalization, and average cost per hospital day were calculated by age and risk group for each study year. Medical claims for pneumococcal disease and charges for antibiotics within 30 days following hospitalization were counted as costs related to follow-up care. For outpatient CAP, all costs related to the episode were analyzed. Episode costs were limited to direct medical costs paid in order to represent the perspective of the national health insurer. All costs were expressed in 2009 Taiwan dollars based on annual change in CPI19.

As a secondary analysis, direct medical costs were supplemented with an estimate of the economic impact of productivity loss. Costs related to time lost from work were calculated by multiplying the average daily earnings in Taiwan in 2010 by the employment rate20, and by the number of work-days lost due to illness. We assumed that a consultation with a primary care physician would result in 0.5 days lost from work, that hospitalization would result in a work loss equal to the length of the hospital stay, and that pneumococcal disease risk is the same in employed and un-employed adult populations21.

Burden of illness

To estimate the total burden of illness due to pneumococcal disease in adults in 2010, the average age-stratified epidemiological data on incidence and case fatality rates, and the average per-episode outpatient and inpatient cost data from 2002–2009, were applied to the estimated 2010 population by age group.

Sensitivity analyses

To explore the impact of uncertainty on estimates of 2010 disease burden, one-way sensitivity analyses were performed on uncertain parameters such as total episodes, deaths, and costs due to pneumococcal disease, based on different alternative assumptions in order to examine outcome robustness. Parameters that were varied included the estimated percentage of pneumonia assumed to be pneumococcal (18.0% and 30.8%, reflecting a calculated confidence interval around the 23.8% estimate reported by Lauderdale et al.9), use of incidence rates from the lowest and highest year for each disease, the use of case fatality rates and cost data for pneumococcal pneumonia-specific claims (ICD-9 code 481), incidence rates of IPD from other published sources with more broadly defined IPD definitions5,15, and application of population demographic data from the years 2000 and 2020.

Results

Overall incidence, mortality, and cost

The population of Taiwan at 50+ years of age increased from 5,015,101 in 2002 to 6,619,981 in 2009 (Table 1). Reflecting the growth in the older population, the number of episodes of IPD, hospitalized pneumococcal pneumonia, and non-hospitalized pneumococcal pneumonia all increased over the study period. Over the study period the annual incidence of IPD was 2.4 episodes per 100,000 person years. The average incidence of hospitalized pneumococcal pneumonia was 278.8 episodes per 100,000 person years over the study period. The incidence rate rose to a peak of 306.9 per 100,000 person years in 2005, but then fell to 250.1 per 100,000 person years in 2009. The average annual incidence of non-hospitalized pneumococcal pneumonia was 1376.4 episodes per 100,000 person years, and it increased steadily over the study period from 1281.7 per 100,000 person years in 2002 to 1560.5 per 100,000 person years in 2009.

Table 1. Pneumococcal disease burden, 2002–2009, in adults aged >50 years.

	Year
	2002	2003	2004	2005	2006	2007	2008	2009
Population aged ≥50 years	5,015,101	5,220,135	5,436,303	5,666,749	5,914,757	6,133,944	6,372,514	6,619,981
Average age	63.0	62.9	62.9	62.8	62.7	62.7	62.7	62.7
Incidence rate per 100,000 person years
Invasive pneumococcal disease	2.3	1.6	2.6	2.4	2.6	2.5	2.6	2.4
Hospitalized pneumococcal pneumonia	275.5	263.3	287.2	306.9	280.8	301.8	264.3	250.1
Outpatient pneumococcal pneumonia	1281.7	1220.3	1229.3	1455.7	1326.7	1484.2	1453.0	1560.5
Episodes, n
Invasive pneumococcal disease	113	85	140	136	153	152	166	157
Hospitalized pneumococcal pneumonia	13,819	13,747	15,615	17,389	16,606	18,511	16,845	16,554
Outpatient pneumococcal pneumonia	64,281	63,699	66,826	82,490	78,474	91,037	92,591	103,302
Deaths, n
Invasive pneumococcal disease	17	8	17	17	18	19	19	20
Hospitalized pneumococcal pneumonia	1192	1218	1414	1641	1670	1927	1899	1900
Total cost, 2009 NT$ 1000s
Invasive pneumococcal disease	18,590	11,199	19,852	17,742	21,207	21,086	20,402	20,231
Hospitalized pneumococcal pneumonia	2,043,306	2,465,977	2,947,552	3,083,643	2,998,764	3,207,380	2,688,670	2,561,949
Outpatient pneumococcal pneumonia	118,973	121,976	122,488	155,372	160,312	189,428	200,213	224,186

The average case fatality rate over the period was 12.3% for IPD and 10.0% for hospitalized pneumonia. The number of deaths from hospitalized pneumonia increased steadily from 2002 to 2007 and then plateaued, whereas the number of deaths due to IPD was similar throughout the period. Total costs associated with IPD, hospitalized pneumococcal pneumonia, and non-hospitalized pneumococcal pneumonia all increased during the study period.

Incidence, fatality, and cost by age

The average incidence rates of pneumococcal bacteremia, hospitalized pneumococcal pneumonia, and outpatient pneumococcal CAP were greater in each age category, whereas there was a low incidence of pneumococcal meningitis in all age groups (Table 2). Case fatality rates for pneumococcal bacteremia and hospitalized pneumococcal pneumonia also increased by age group. Direct costs per case of hospitalized pneumococcal pneumonia and outpatient pneumococcal CAP was greater in older age groups. Indirect costs for all four diagnoses were highest in the group aged 50–64 years, reflecting the higher proportion of employed individuals.

Table 2. Average incidence, fatality, and cost of pneumococcal diseases by age, 2002–2009.

	Age groups, years
	50–64	65–74	75–84	≥85
Incidence rate as episodes per 100,000 person years
Pneumococcal meningitis	0.1	0.1	0.1	0.1
Pneumococcal bacteremia	0.9	2.5	6.1	12.3
Hospitalized pneumococcal pneumonia	84.4	313.1	820.3	1650.9
Outpatient pneumococcal pneumonia	716.0	1645.5	3198.2	5305.8
Case fatality rate, %
Pneumococcal meningitis	5.3	–	12.5	–
Pneumococcal bacteremia	7.8	8.2	16.1	18.7
Hospitalized pneumococcal pneumonia	7.6	7.8	10.9	13.6
Direct cost per case, NT$
Pneumococcal meningitis	273,293	193,528	344,621	203,359
Pneumococcal bacteremia	120,597	133,753	139,662	126,994
Hospitalized pneumococcal pneumonia	152,568	169,419	176,382	177,723
Outpatient pneumococcal pneumonia	1614	1983	2281	2481
Indirect cost per case, NT$
Pneumococcal meningitis	15,960	5901	4496	8771
Pneumococcal bacteremia	11,704	3372	3653	3372
Hospitalized pneumococcal pneumonia	11,704	3372	3653	3934
Outpatient pneumococcal pneumonia	532	141	141	141

Pneumococcal disease burden in 2010

Applying the average estimates from the years 2002–2009 to the population of Taiwan in 2010 yielded an estimated pneumococcal disease burden for the year (Table 3). The total number of episodes of IPD and pneumococcal pneumonia was estimated at 113,582; the majority of these were outpatient pneumococcal pneumonia. The highest number of episodes of hospitalized pneumococcal pneumonia and outpatient pneumococcal pneumonia was projected to occur in the groups aged 75–84 years and 50–64 years, respectively. Most of the estimated deaths were attributed to hospitalized pneumococcal pneumonia. The number of deaths was highest in the group aged 75–84 years. The total annual direct cost of IPD and pneumococcal pneumonia was estimated to be more than NT$3.4 billion, of which >90% was due to hospitalized pneumococcal pneumonia. Consideration of indirect costs would add NT$125.7 million to the projected annual cost burden. Estimated total direct medical costs were highest in the group aged 75–84 years, whereas estimated total indirect costs were highest in the group aged 50–64 years.

Table 3. Estimated pneumococcal disease burden in adults in 2010.

	Pneumococcal meningitis	Pneumococcal bacteremia	Hospitalized pneumococcal pneumonia	Outpatient pneumococcal pneumonia	Total
Estimated pneumococcal episodes, n
Age group, years
50–64	3	39	3697	31,361	35,100
65–74	2	35	4339	22,802	27,178
75–84	1	53	7066	27,546	34,666
≥85	0	29	3941	12,667	16,637
Total	6	156	19,043	94,376	113,581
Estimated pneumococcal deaths, n
Age group, years
50–64	0	3	281	0	284
65–74	0	3	340	0	343
75–84	0	9	769	0	778
≥85	0	5	536	0	541
Total	0	20	1926	0	1946
Estimated total direct medical cost, NT$1000s
Age group, years
50–64	800	4740	564,077	50,620	620,238
65–74	331	4693	735,178	45,219	785,421
75–84	391	7395	1,246,249	62,828	1,316,864
≥85	36	3724	700,454	31,429	735,643
Total	1558	20,553	3,245,958	190,096	3,458,166
Estimated total indirect cost, NT$1000s
Age group, years
50–64	47	460	43,272	16,684	60,463
65–74	10	118	14,632	3204	17,964
75–84	5	193	25,811	3870	29,879
≥85	2	99	15,505	1780	17,386
Total	64	870	99,220	25,538	125,692
Estimated total cost, NT$1000s
Age group, years
50–64	847	5200	607,349	67,305	680,701
65–74	341	4812	749,810	48,423	803,386
75–84	396	7589	1,272,060	66,698	1,346,744
≥85	37	3823	715,959	33,208	753,028
Total	1622	21,424	3,345,179	215,634	3,583,858

Sensitivity analysis

Changes in assumptions regarding the etiology of pneumonia had a large impact on total episodes, deaths, and total cost. Varying the assumed percentage of hospitalized pneumonia that was pneumococcal from 18.0–30.8% had the greatest impact on estimated pneumococcal burden (Table 4). Similarly, use of incidence rates from the lowest and highest years of the study period varied estimates by roughly ±10%. The average case fatality rate and cost per case of pneumonia claims coded as pneumococcal (ICD-9) was lower than for the all-cause pneumonia claims, resulting in a lower total mortality and cost burden when these values were substituted. Use of IPD incidence and case fatality rates from surveillance, which are higher than those from insurance claims and expected to be more accurate, had little influence on total pneumococcal burden of disease. Finally, extrapolation of results to a 2020 population shows >40% growth in episodes, mortality, and cost, due to aging of the population and survival into older age groups, when incidence rates and case fatality rates are higher.

Table 4. Sensitivity of burden to change in assumptions.

	Episodes	Deaths	Direct medical cost (NT$ MM)
Base case	113,582	1946	3458
Percentage of pneumonia as pneumococcal
Low (18.0%)	85,942	1477	2621
High (30.8%)	260,198	4419	4469
Incidence rates
Lowest year	99,888	1759	3102
Highest year	127,973	2139	3816
Mortality and cost of pneumococcal pneumonia specific claims*
Coded as pneumococcal  CFR and cost	113,582	1064	2286
Alternative IPD estimates
CDC IPD, 2008–20095	113,845	1958	3494
Liao et al. IPD, 2000–200815	113,717	2025	3476
Population demographics and projected size
2000	76,457	1247	2295
2020	160,797	2863	5022

*Using CFR and cost from claims coded as pneumococcal pneumonia (ICD-9 code 481).

CDC, Taiwan Centers for Disease Control; CFR, case fatality rate; ICD-9, International Classification of Diseases, Ninth Revision; IPD, invasive pneumococcal disease; MM, millions.

Discussion

We have analyzed epidemiologic and economic data using NHIRD on the burden of illness due to pneumococcal disease in adults aged 50+ years in Taiwan for the years 2002–2009, and applied data from those years to develop an estimated pneumococcal disease burden in that age group for 2010. During the study period, the incidences of pneumococcal bacteremia, hospitalized pneumococcal pneumonia, and outpatient pneumococcal CAP all increase in older age groups, as did the case fatality rate for pneumococcal bacteremia and hospitalized pneumococcal pneumonia. Direct costs for both hospitalized and outpatient pneumonia also increased with increasing age, whereas those for IPD were variable, likely due to small sample sizes. The indirect costs per case were highest in the group aged 50–64 years because they are more likely to be employed and, thus, have the greatest impact on lost time from work.

Applying these findings to the population of Taiwan in 2010 yields an estimate of 113,582 episodes of pneumococcal disease occurring in persons aged ≥50 years, the great majority of which are episodes of pneumococcal pneumonia. A total of 1946 deaths are expected, 98.9% of which are due to hospitalized pneumococcal pneumonia. The estimated total annual cost for these pneumococcal diseases in 2010 was greater than NT$3.6 billion, of which more than NT$3.4 billion was attributable to direct medical cost. Direct medical costs were expected to be highest in the group aged 75–84 years.

In other countries, widespread use of 7-valent pneumococcal conjugate vaccine (PCV7) has been shown to reduce the incidence of pneumococcal disease among vaccinated and unvaccinated populations22. In Taiwan, prior to 2013, universal infant vaccination with pneumococcal conjugate vaccines was not covered by the National Vaccine Fund. Use during the study period was low, and, in 2007, only 15.9% of children aged <5 years received PCV723. In a recent study on the cost-effectiveness of pneumococcal conjugate vaccine use in Taiwan using a transmission-dynamic model, a minimal vaccination threshold of ∼60% was predicted to be required to generate herd immunity21. Therefore, it is reasonable to assume that vaccination coverage has not achieved the threshold necessary to elicit a herd effect in the unvaccinated population. Thus, no evidence of a decline in cases of pneumococcal disease among adults was expected in the current study, which was consistent with the projection found in the simulation study21. Therefore, our study provides epidemiologic and economic data that can be used to inform public policy decision-making, and a baseline from which to monitor the potential impact of pneumococcal vaccination programs in the future.

Hospitalized pneumococcal pneumonia accounted for a large majority of the cost associated with pneumococcal disease in this study. Although the incidence of hospitalized pneumonia among older adults showed a slight decline during the study period, the number of episodes increased by 19.8%, corresponding with growth in the size of the adult population. As the proportion of older adults in Taiwan continues to increase, the number of episodes of hospitalized pneumonia would also be expected to increase.

Our results confirm and extend those of other studies of the burden of pneumococcal illness conducted in Taiwan and the surrounding region. In Singapore, data from the Ministry of Health for 2002–2006 estimated an all-cause pneumonia incidence of 1939 per 100,000 person years in adults aged >60 years, with a case fatality rate of 26%24. Similarly, the incidence of pneumococcal bacteremia in Singapore has been reported at 25 per 100,000 in adults aged >60 years, with a case fatality rate of 33.1%25. Lee et al.²⁶ estimated the annual number of adult hospitalizations from the Hong Kong Hospital Authority based on 2005 data, and estimated 4498 cases of pneumococcal pneumonia causing 743 deaths in adults aged ≥65 years. This yields an incidence of ∼514 per 100,000 persons assuming an estimated population of 875,000 from the US Census Bureau International Database, with a case fatality rate of 16.5%. In the US, where there is widespread use of pneumococcal conjugate vaccine in children and pneumococcal polysaccharide vaccine in adults, there were lower absolute rates of pneumococcal pneumonia by age, but a similar relative increase in incidence with increasing age. Non-bacteremic pneumococcal pneumonia in the US accounted for the majority (82%) of direct costs²⁷.

The infectious etiology of pneumonia is often unknown, and etiology studies generally fail to find a cause for most cases. In our study, we were limited to estimates of pneumonia cases through ICD-9-coded medical claims. Use of pneumococcal pneumonia-specific ICD-9 codes would yield an under-estimation of pneumococcal pneumonia because the etiology of pneumonia is often unknown, and use of ICD-9 codes has poor sensitivity under these conditions²⁸. The most common ICD-9 code for pneumonia, pneumonia due to unspecified organism (code 486.0), is sensitive but with poor specificity, and may include unconfirmed cases, nosocomial pneumonias, re-admissions, or hospitalizations due to other causes. We used broad coding to collect all-cause pneumonia hospitalizations, then assumed 23.8% were pneumococcal in origin based on a high-quality etiology study in adults with CAP in Taiwan9. The possible inclusion of cases that may not be CAP could over-estimate pneumococcal disease, and uncertainty around the true bacterial etiology of pneumonia may lead to an over- or under-estimation of cases. We attempted to develop a range of results by performing a sensitivity analysis; however, we note the significant importance of this assumption on results and encourage more research into local studies on the etiology of pneumonia and the identification of CAP with all-cause pneumonia coding.

IPD is often under-represented by ICD-9 coding; however, IPD is classified as a category IV infectious disease by the Taiwan Centers for Disease Control, and is a notifiable disease that is monitored under national surveillance. Data from national surveillance support a higher incidence of IPD in Taiwan than reported in the NHIRD database5. Furthermore, as the surveillance is a count of cultured specimens, it likely represents a minimum incidence of disease. We incorporated incidence rates from this surveillance into our sensitivity analysis, which increased the total burden of IPD to 425 episodes, 32 deaths, and NT$58.0 million estimated for 2010. However, due to the low incidence of IPD relative to pneumococcal pneumonia, this has very little impact on our overall pneumococcal disease burden estimate.

Conclusion

Pneumococcal disease causes significant burden of illness among older adults in Taiwan and accounts for an estimated NT$3.46 billion in direct costs annually. The risk of disease increases with age. Hospitalized pneumococcal pneumonia accounts for the great majority of fatalities and costs associated with pneumococcal disease.

Transparency

Declaration of funding

This study was sponsored by Pfizer Inc. Pfizer was involved in the study design, data collection, data analysis, data interpretation, and writing of the manuscript.

Declaration of financial/other relationships

CSR is an employee and shareholder of Pfizer Inc. CHF is an employee of Pfizer Inc. CJC, DW, and LC received a research grant from Pfizer, 2010–2011, to conduct this study. YCH has no conflicts of interest to declare. CMRO Peer Reviewers on this manuscript have no relevant financial or other relationships to disclose.

Acknowledgments

This study is based in part on data from the National Health Insurance Research Database provided by the Bureau of National Health Insurance, Department of Health, and managed by National Health Research Institutes. The interpretation and conclusions contained herein do not represent those of the Bureau of National Health Insurance, Department of Health or National Health Research Institutes. This study is supported by two grants: National Science Council (Grant Number: NSC97-2320-B-182-004-MY3) and the Chang Gung Medical Research Program grant (Grant Number: CMRPD1A0282). Medical writing support was provided by Nancy Price, PhD, at Excerpta Medica and was funded by Pfizer Inc. Authors’ contributions: DW participated in the design of the study and the analysis of the data; CSR participated in the design of the study and the analysis of the data; and YCH participated in the acquisition of the data. LC participated in the design of the study, and the analysis and acquisition of the data. CHF analyzed the data. CJC participated in the design of the study, and the analysis and acquisition of the data.