Association between introduction of 13-valent pneumococcal conjugate vaccine and burden of hospitalized adult pneumococcal disease in Hong Kong – A retrospective database study Authors: Qiuyan Yu, Celine Sze Ling Chui, See-Hwee Yeo, Dony Patel, Cindy Thiow Koon Lim, Xue Li, Vivien Kin Yi Chan, Eriko Yamada, Isaya Sukarom Published: 2025 ABSTRACT ABSTRACTThere is limited information on the association between burden of pneumococcal disease and vaccination. This study aimed to assess pneumococcal disease burden among adults in Hong Kong by estimating incidence rate (IR), case fatality rate (CFR), healthcare resource utilization and costs before and after the introduction of 13-valent pneumococcal conjugate vaccine (PCV13) into Hong Kong’s childhood immunization program (CIP) in 2011. Adults ≥18 years hospitalized with invasive pneumococcal disease (IPD) and pneumonia from 2005 to 2019 were identified from the Clinical Data Analysis and Reporting System (CDARS). Interrupted time series analysis (ITSA) was performed and incidence rate ratio (IRR) was computed to compare baseline and trend of IRs between pre-introduction (2005–2010) and post-introduction (2012–2019) of PCV13. Overall IPD and pneumonia IRs were 0.78 and 5.34 episodes per 100,000 person-years, respectively. ITSA showed no changes in baseline IPD IR (IRR = 0.82; P value = .447) and trend of IR (IRR = 0.98; P value = .112). There was no difference in baseline pneumonia IR (IRR = 0.81; P value = .075), but a marginal (small but statistically significant) reduction in trend was observed in the post-PCV13 period (IRR = 0.98; P value = .019). CFR was 24.12% for IPD and higher at 30.06% for pneumonia. The average total costs per pneumonia episode were higher in the post-PCV13 period (2,975 vs. 2,810 Hong Kong Dollar). IPD and pneumonia incidence fluctuated over the study period. After PCV13 introduction into Hong Kong’s CIP, there was only a marginal reduction in trend of pneumonia among adults. CFRs of IPD and pneumonia were high, and disease burden remained substantial.KEYWORDS: Pneumococcal disease, pneumonia, pneumococcal vaccine, Hong Kong, adults, immunization program Background Streptococcus pneumoniae (S. pneumoniae) causes a wide range of diseases including pneumonia, meningitis and bacteremia.1 Based on the United States Centers for Disease Control and Prevention, the overall incidence of invasive pneumococcal disease (IPD) in 2019 was reported as 5 and 28 cases per 100,000 population among adults 18–64 years and those ≥65 years, respectively.2 In Hong Kong, the incidence rate (IR) of IPD in adults ≥18 years was 2.2 episodes per 100,000 person-years from 2015 to 2019.3 Data from the Center for Health Protection also reported a range of 162–190 IPD cases annually in the same period.4,5 Meanwhile, the combined IR of bacteremic and non-bacteremic pneumococcal pneumonia was 6.2 episodes per 100 000 person-years in Hong Kong, with higher incidence in adults ≥65 years.3 Pneumococcal polysaccharide vaccine (PPSV) and pneumococcal conjugate vaccine (PCV) have been developed to confer immunity against infection by S. pneumoniae and are now widely available across countries.6–8 13-valent PCV (PCV13) was introduced into the childhood immunization program (CIP) in Hong Kong in 2011.9,10 For adults aged ≥65 years, free or subsidized 23-valent PPSV (PPSV23) had been available under the national immunization program since 2009.11 PCV13 was subsequently added into the national immunization program in 2017. A single dose of PCV13 was provided to high-risk elderly adults followed by one dose of PPSV23 a year after.12 PCV immunization in pediatrics is reported to have the potential of indirectly protecting vulnerable and unvaccinated adult populations from infection in a phenomenon known as herd protection.13,14 However, this protection may not be significant as only a minor herd effect was observed in some European countries.15,16 Possible reasons for the varying effectiveness of PCV across territories could be due to geographical differences in the prevalence of circulating pneumococcal serotypes and vaccine coverage rates.17 In Hong Kong, there is limited information on the burden of pneumococcal disease and the effectiveness of vaccination. Therefore, this study aimed to assess the association between PCV13 introduction and burden of pneumococcal disease among adults by estimating hospitalized IR, case fatality rate (CFR), healthcare resource utilization (HCRU) and costs related to IPD and pneumonia, before and after the introduction of PCV13 into Hong Kong’s CIP. Patients and methods Study design and population A retrospective population-based observational study was performed using the Clinical Data Analysis and Reporting System (CDARS), which covers 80% of hospital admissions and serves around 12 million patients in Hong Kong.18 Adults aged ≥18 years hospitalized with IPD or pneumonia from 1 January 2005 to 31 December 2019 were included. Pneumococcal-specific International Classification of Disease (ICD) codes were used for case identification (ICD codes are provided in Supplementary Table S1). Patients with missing age, date of birth or date of discharge during the study period were excluded. The date of the first hospitalization was considered the start date of the episode. A time gap of ≥90 days between hospitalizations with the same diagnosis codes was used to define disease episodes.19 Any subsequent hospitalization with the same disease condition within this time gap was assumed to be related to the earlier hospitalization and both combined to form a single episode. The date of hospital discharge marked the end of one episode if there was no further hospital admission within 90 days. Outcome measurements Annual IR was analyzed as the number of new disease episodes divided by total person-years in a calendar year. CFR was estimated as the proportion of deaths among disease episodes. Disease-specific death was identified from the discharge status of each hospital admission based on ICD codes. Components of HCRU included total length of hospital stay, as well as number of inpatient, outpatient and emergency department visits. Costs were expressed as Hong Kong Dollar (HKD) and analyzed from the patients’ perspective. Subsidized costs were used for calculation, and they included inpatient, outpatient, outpatient pharmacy, and emergency department costs. Unit costs for medications and services in 2019 were applied to all other years to ensure comparability across the study period. Statistical analyses Outcomes including IR, CFR, and HCRU and costs were summarized using descriptive statistics. Continuous variables were summarized using means, standard deviations, medians and interquartile ranges. Student’s t-test or Wilcoxon rank-sum test was used to compare between groups. To assess the association between disease incidence and PCV13 introduction into Hong Kong’s CIP, interrupted time series analyses (ITSA) were performed on quarterly IR time trends based on a Poisson or negative binomial regression model for each disease. Incidence rate ratio (IRR) was calculated to compare the baseline levels and trends of IRs between time periods. The HCRU and costs incurred before and after the introduction of PCV13 were also compared. As PCV13 was introduced into Hong Kong’s CIP in 2011, study outcomes were compared between pre-PCV13 (2005–2010) and post-PCV13 (2012–2019) periods. The year 2011 was considered as the transition year, and thus excluded from analyses. Analyses were adjusted for seasonality using indicator variables for each quarter of the year and stratified by age groups (18 to <65 years and ≥65 years). All analyses were performed using the latest stable version of R (R Core Team (2017) Vienna, Austria). Ethics This study was approved by the Institutional Review Board of the University of Hong Kong and Hospital Authority Hong Kong West Cluster (UW 21–076). The requirement for the informed consent was waived by the Institutional Review Board of the University of Hong Kong and Hospital Authority Hong Kong West Cluster because of the retrospective nature of the study. All clinical data were anonymized and de-identified from the Hospital Authority. Results There were 4,447 hospital episodes identified from CDARS during the study period, comprising 568 IPD and 3,879 pneumonia episodes. IPD incidence The overall IPD IR was 0.78 episodes per 100,000 person-years, with annual IPD IRs fluctuating between 0.22 and 1.36 episodes per 100,000 person-years (Figure 1, numerical data further provided in Supplementary Table S2). IPD IR was 6.7 times higher among adults aged ≥65 years, compared with those aged 18 to <65 years (2.68 vs. 0.40 episodes per 100,000 person-years). ITSA results showed that there was no difference in baseline IRs among all adults between pre- and post-PCV13 periods (IRR = 0.82; P value = .447) (Figure 2(a) and Table 1). No change was also observed in trend of IR in the post-PCV13 period (IRR = 0.98; P value = .112). Similar results were observed when stratified by age groups (Figure 2(b,c) and Table 1). Baseline IRs were comparable between the time periods among adults aged 18 to <65 years (IRR = 0.77; P value = .513) and those aged ≥65 years (IRR = 0.86; P value = .642), while there were no changes in trend of IRs for adults aged 18 to <65 years and those aged ≥65 years in the post-PCV13 period (P values = .453 and 0.086, respectively). Trend in annual incidence of hospitalized invasive pneumococcal disease in adults from 2005–2019. Trend in quarterly incidence of hospitalized invasive pneumococcal disease in adults (a) ≥18 years, (b) 18 to <65 years, and (c) ≥65 years from 2005–2019. Table 1.IRRs of baseline IR and trend of IR in post-PCV13 period compared to pre-PCV13 period. Baseline IRTrend of IRConditionIRR95% CIP-valueIRR95% CIP-valueIPD≥18 years0.820.50–1.37.4470.980.94–1.01.11218 to <65 years0.770.36–1.73.5130.980.94–1.03.453≥65 years0.860.45–1.68.6420.960.93–1.00.086Pneumonia≥18 years0.810.64–1.02.0750.980.97–1.00.019*18 to <65 years0.710.50–1.02.0670.970.95–0.99.006*≥65 years0.770.59–1.00.0500.970.96–0.98<.001**P-value <.05.Abbreviations: CI, confidence interval; IPD, invasive pneumococcal disease; IR, incidence rate; IRR, incidence rate ratio; PCV13, 13-valent pneumococcal conjugate vaccine. IRRs of baseline IR and trend of IR in post-PCV13 period compared to pre-PCV13 period. *P-value <.05. Abbreviations: CI, confidence interval; IPD, invasive pneumococcal disease; IR, incidence rate; IRR, incidence rate ratio; PCV13, 13-valent pneumococcal conjugate vaccine. Pneumonia incidence The overall pneumonia IR was 5.34 episodes per 100,000 person-years, with annual IRs ranging between 3.33 and 6.54 episodes per 100,000 person-years (Figure 3, numerical data further provided in Supplementary Table S3). Pneumonia IR was nearly 12 times higher among adults aged ≥65 years, compared with those aged 18 to <65 years (22.25 vs. 1.90 episodes per 100,000 person-years). ITSA results showed that there was no difference in baseline IRs among all adults between pre- and post-PCV13 periods (IRR = 0.81; P value = .075), but a marginal reduction in trend of IRs was observed (IRR = 0.98; P value = .019) (Figure 4(a) and Table 1). The trend corresponds to 0.086 fewer pneumonia cases per 100,000 person-years annually, translating to approximately 5–6 preventable pneumonia cases per year in Hong Kong’s adult population of 6.5 million. Similarly, baseline IRs were comparable between the two time periods among adults aged 18 to <65 years (IRR = 0.71; P value = .067) and adults aged ≥65 years (IRR = 0.77; P value = .050), but there were decreases in trend of IRs for adults aged 18 to <65 years (IRR = 0.97; P value = .006) and adults aged ≥65 years (IRR = 0.97; P value <.001) in the post-PCV13 period (Figure 4(b,c) and Table 1). Trend in annual incidence of hospitalized pneumonia in adults from 2005–2019. Trend in quarterly incidence of hospitalized pneumonia in adults (a) ≥18 years, (b) 18 to <65 years, and (c) ≥65 years from 2005–2019. CFR Overall CFRs for hospitalized IPD and pneumonia were 24.12% and 30.06%, respectively (Table 2). The CFRs in adults aged ≥65 years were about 2–4 times higher, compared with those aged 18 to <65 years for IPD (32.52% vs. 12.55%) and pneumonia (38.12% vs. 10.89%). Compared with pre-PCV13 period, CFR for IPD among all adults decreased from 33.33% to 19.72% in the post-PCV13 period. Similarly, CFRs for IPD declined in the post-PCV13 period among adults aged 18 to <65 years (from 23.26% to 9.29%) and adults aged ≥65 years (from 40.68% to 27.57%). For pneumonia, lower CFR was observed among adults in the post-PCV13 period (35.86% vs. 25.71%). In addition, the respective CFRs of pneumonia in adults aged 18 to <65 years and adults aged ≥65 years were 9.31% and 33.77% in the post-PCV13 period, which were lower than the pre-PCV13 period (14.84% and 42.46%, respectively). Table 2.Case fatality rates by pneumococcal disease, age groups and PCV13 periods. OverallPre-PCV13 periodTransitionPost-PCV13 periodConditionCases (N)Deaths (n)CFR (%)Cases (N)Deaths (n)CFR (%)Cases (N)Deaths (n)CFR (%)Cases (N)Deaths (n)CFR (%)IPD ≥18 years56813724.121023433.33401947.504268419.7218 to <65 years2393012.55431023.2613323.08183179.29≥65 years32910732.52592440.68271659.262436727.57Pneumonia ≥18 years3879116630.06118542535.8631412941.08238061225.7118 to <65 years114812510.892834214.84811012.35784739.31≥65 years2731104138.1290238342.4623311951.07159653933.77Abbreviations: CFR, case fatality rate; IPD, invasive pneumococcal disease; PCV13, 13-valent pneumococcal conjugate vaccine. Case fatality rates by pneumococcal disease, age groups and PCV13 periods. Abbreviations: CFR, case fatality rate; IPD, invasive pneumococcal disease; PCV13, 13-valent pneumococcal conjugate vaccine. HCRU For IPD, the median length of hospitalization per inpatient visit among all adults was longer in the post-PCV13 period, compared with pre-PCV13 period (14 vs. 11 days; P value = .041) (Table 3). However, no differences were detected for adults aged 18 to <65 years (15 vs. 11 days; P value = .117), and adults aged ≥65 years (14 vs. 11 days; P value = .196). For pneumonia, there was no change in the median length of hospitalization per inpatient visit in the post-PCV13 period, compared with pre-PCV13 period (10 vs. 9 days; P value = .628) (Table 4). Similar results were observed in adults aged ≥65 years (both 11 days; P value = .980), but hospitalization was longer in the post-PCV13 period for those aged 18 to <65 years (8 vs. 6 days; P value = .020). Table 3.IPD-associated HCRU and costs during pre-PCV13 and post-PCV13 periods, by age groups. ≥18 years18 to <65 years≥65 yearsPre-PCV13 (n = 102)Post-PCV13 (n = 363)P-valuePre-PCV13 (n = 43)Post-PCV13 (n = 153)P-valuePre-PCV13 (n = 59)Post-PCV13 (n = 210)P-valueHealthcare resource utilization Inpatient Number of visitsa, n148555-g60249–88306– Length of hospitalization per visit (days) Mean ± SD26.71 ± 53.6421.94 ± 23.48.38433.53 ± 73.0722.11 ± 23.63.31821.73 ± 32.9321.82 ± 23.43.984 Median (Q1, Q3)11 (5, 23)14 (7, 28).041*11 (5, 23)15 (8, 26).11711 (6, 23)14 (7, 28).196 Total length of hospitalizationb27247966–14423383–12824583– Per episode Average number of visits1.451.53–1.401.63–1.491.46– Average length of hospitalization (days)26.7121.94–33.5322.11–21.7321.82– Outpatient Number of visitsc, n1741043–105486–69557– Average number of visits per episode1.712.87–2.443.18–1.172.65– ED Number of visitsd, n53225–2499–29126– Average number of visits per episode0.520.62–0.560.65–0.490.60–Costs (in Hong Kong dollars) Inpatient Per visit Mean ± SD2283.65 ± 4919.491797.38 ± 2212.52.2432959.00 ± 6967.461705.36 ± 2118.67.1731823.18 ± 2727.871872.25 ± 2286.76.878 Median (Q1, Q3)1035 (435, 2025)1155 (555, 2115).1801035 (435, 2055)1155 (435, 2115).7381035 (435, 2025)1155 (555, 2115).147 Total inpatient costs337 980997 545–177 540424 635–160 440572 910– Average inpatient costs per episode3313.532748.06–4128.842775.39–2719.322728.14– Outpatient Per visit (excluding pharmacy costs) Mean ± SD77.95 ± 60.7970.10 ± 40.97.10286.99 ± 63.3273.90 ± 43.94.046*64.20 ± 54.3366.80 ± 37.92.701 Median (Q1, Q3)80 (19, 80)80 (60, 80).14080 (50, 135)80 (60, 80).025*80 (19, 80)80 (50, 80).364 Total outpatient visit costs13 56473 118–913435 913–443037 205– Pharmacy costs per visit Mean ± SD15.87 ± 3.5216.47 ± 5.35.035*17.91 ± 5.9816.23 ± 4.79.030*15.00 ± 0.0016.68 ± 5.78<.001* Median (Q1, Q3)15 (15, 15)15 (15, 15).20415 (15, 15)15 (15, 15).001*15 (15, 15)15 (15, 15)<.001* Total outpatient pharmacy costs355518 810–12008535–235510 275– Total outpatient costse17 11991 928–10 33444 448–678547 480– Per episode Average outpatient visit costs132.98201.43–212.42234.73–75.08177.17– Average pharmacy costs34.8551.82–27.9155.78–39.9248.93– Average outpatient costs167.83253.25–240.33290.51–115.00226.10– ED Per visit Mean ± SD180 ± 0180 ± 0–180 ± 0180 ± 0–180 ± 0180 ± 0– Median (Q1, Q3)180 (180, 180)180 (180, 180)–180 (180, 180)180 (180, 180)–180 (180, 180)180 (180, 180)– Total ED costs954040 500–432017 820–522022 680– Average ED costs per episode93.53111.57–100.47116.47–88.47108.00–Total costsf364 6391 129 973–192 194486 903–172 445643 070–Average total costs per episode3574.893112.87–4469.633182.37–2922.803062.24–*P-value <.05.Abbreviations: ED, emergency department; IPD, invasive pneumococcal disease; PCV13, 13-valent pneumococcal conjugate vaccine; Q1, 25th percentile; Q3, 75th percentile; SD, standard deviation.aNumber of inpatient visits is defined as the number of IPD-associated hospital admissions.bTotal inpatient hospitalization is defined as the total duration of stay (in days) for all IPD-associated hospitalizations.cNumber of outpatient visits is defined as the number of hospital outpatient visits that occur on separate dates. Multiple records on the same date will be considered as a single visit.dNumber of ED visits are defined as the number of IPD-associated visits to the hospital ED that occur on separate dates.eTotal outpatient costs is defined by the sum of total outpatient visit costs and total outpatient pharmacy costs.fTotal costs is defined by the sum of total inpatient costs, total outpatient costs and total ED costs.gP-value not computed for average costs per episode as they were calculated using the total costs divided by the number of episodes. IPD-associated HCRU and costs during pre-PCV13 and post-PCV13 periods, by age groups. *P-value <.05. Abbreviations: ED, emergency department; IPD, invasive pneumococcal disease; PCV13, 13-valent pneumococcal conjugate vaccine; Q1, 25th percentile; Q3, 75th percentile; SD, standard deviation. aNumber of inpatient visits is defined as the number of IPD-associated hospital admissions. bTotal inpatient hospitalization is defined as the total duration of stay (in days) for all IPD-associated hospitalizations. cNumber of outpatient visits is defined as the number of hospital outpatient visits that occur on separate dates. Multiple records on the same date will be considered as a single visit. dNumber of ED visits are defined as the number of IPD-associated visits to the hospital ED that occur on separate dates. eTotal outpatient costs is defined by the sum of total outpatient visit costs and total outpatient pharmacy costs. fTotal costs is defined by the sum of total inpatient costs, total outpatient costs and total ED costs. gP-value not computed for average costs per episode as they were calculated using the total costs divided by the number of episodes. Table 4.Pneumonia-associated HCRU and costs during pre-PCV13 and post-PCV13 periods, by age groups. ≥18 years18 to <65 years≥65 yearsPre-PCV13 (n = 1185)Post-PCV13 (n = 2049)P-valuePre-PCV13 (n = 283)Post-PCV13 (n = 673)P-valuePre-PCV13 (n = 902)Post-PCV13 (n = 1376)P-valueHealthcare resource utilization Inpatient Number of visits a, n17472994-g389981–13582013– Length of hospitalization per visit (days) Mean ± SD17.96 ± 29.5218.84 ± 34.10.44316.68 ± 41.8017.31 ± 33.19.82118.37 ± 24.4519.59 ± 34.53.324 Median (Q1, Q3)9 (5, 20)10 (5, 22).6286 (4, 13)8 (4, 18).020*11 (6, 23)11 (5, 23).980 Total length of hospitalizationb21 28738 601–4,72011 650 16 56726 951– Per episode Average number of visits1.471.46–1.371.46–1.511.46– Average length of hospitalization (days)17.9618.84–16.6817.31–18.3719.59– Outpatient Number of visits c, n44178399–8872286–35306,113– Average number of visits per episode3.734.10–3.133.40–3.914.44– ED Number of visits d, n16322670–333758–12991912– Average number of visits per episode1.381.30–1.181.13–1.441.39–Costs (in Hong Kong dollars) Inpatient Per visit Mean ± SD1523.19 ± 2298.751622.13 ± 3173.05.2161468.17 ± 3075.641500.08 ± 3141.72.8631538.95 ± 2023.161681.62 ± 3187.28.112 Median (Q1, Q3)915 (555, 1635)915 (555, 1755).837675 (435, 1395)795 (315, 1635).7971035 (555, 1755)1035 (555, 1875).203 Total inpatient costs2 661 0104 856 670–571 1201 471 575–2 089 8903 385 095– Average inpatient costs per episode2245.582370.26–2018.092186.59–2316.952460.10– Outpatient Per visit (excluding pharmacy costs) Mean ± SD62.93 ± 50.2068.62 ± 44.34<.001*76.77 ± 45.8074.88 ± 46.30.29759.45 ± 50.6666.29 ± 43.36<.001* Median (Q1, Q3)80 (0, 80)80 (50, 80)<.001*80 (60, 80)80 (50, 80).09180 (0, 80)80 (50, 80)<.001* Total outpatient visit costs277 967576 377 68 099171 176–209 868405 201– Pharmacy costs per visit Mean ± SD15.68 ± 3.2116.65 ± 4.95<.001*15.61 ± 3.2316.37 ± 4.38<.001*15.69 ± 3.2116.75 ± 5.14<.001* Median (Q1, Q3)15 (15, 15)15 (15, 15)<.001*15 (15, 15)15 (15, 15)<.001*15 (15, 15)15 (15, 15)<.001* Total outpatient pharmacy costs97 530182 910–17 52046 695–80 010136 215– Total outpatient costse375 497759 287–85 619217 871–289 878541 416– Per episode Average outpatient visit costs234.57281.30–240.63254.35–232.67294.48– Average pharmacy costs82.3089.27–61.9169.38–88.7098.99– Average outpatient costs316.88370.56–302.54323.73–321.37393.47–ED Per visit Mean ± SD180 ± 0180 ± 0–180 ± 0180 ± 0–180 ± 0180 ± 0– Median (Q1, Q3)180 (180, 180)180 (180, 180)–180 (180, 180)180 (180, 180)–180 (180, 180)180 (180, 180)– Total ED costs293 760480 600–59 940136 440–233 820344 160– Average ED costs per episode247.90234.55–211.80202.73–259.22250.12–Total costsf3 330 2676 096 557–716 6791 825 886–2 613 5884 270 671–Average total costs per episode2810.352975.38–2532.432713.05–2897.553103.69–*P-value < .05.Abbreviations: ED, emergency department; PCV, pneumococcal conjugate vaccine; Q1, 25th percentile; Q3, 75th percentile; SD, standard deviation.aNumber of inpatient visits is defined as the number of pneumonia-associated hospital admissions.bTotal inpatient hospitalization is defined as the total duration of stay (in days) for all pneumonia-associated hospitalizations.cNumber of outpatient visits is defined as the number of hospital outpatient visits that occur on separate dates. Multiple records on the same date will be considered as a single visit.dNumber of ED visits are defined as the number of pneumonia-associated visits to the hospital ED that occur on separate dates.eTotal outpatient costs is defined by the sum of total outpatient visit costs and total outpatient pharmacy costs.fTotal costs is defined by the sum of total inpatient costs, total outpatient costs and total ED costs.gP-value not computed for average costs per episode as they were calculated using the total costs divided by the number of episodes. Pneumonia-associated HCRU and costs during pre-PCV13 and post-PCV13 periods, by age groups. *P-value < .05. Abbreviations: ED, emergency department; PCV, pneumococcal conjugate vaccine; Q1, 25th percentile; Q3, 75th percentile; SD, standard deviation. aNumber of inpatient visits is defined as the number of pneumonia-associated hospital admissions. bTotal inpatient hospitalization is defined as the total duration of stay (in days) for all pneumonia-associated hospitalizations. cNumber of outpatient visits is defined as the number of hospital outpatient visits that occur on separate dates. Multiple records on the same date will be considered as a single visit. dNumber of ED visits are defined as the number of pneumonia-associated visits to the hospital ED that occur on separate dates. eTotal outpatient costs is defined by the sum of total outpatient visit costs and total outpatient pharmacy costs. fTotal costs is defined by the sum of total inpatient costs, total outpatient costs and total ED costs. gP-value not computed for average costs per episode as they were calculated using the total costs divided by the number of episodes. The total costs, including outpatient, inpatient and emergency department costs, were higher in the post-PCV13 period than pre-PCV13 period for both IPD (HKD 1,129,973 vs. 364,639) (Table 3) and pneumonia (HKD 6,096,557 vs. 3,330,267) (Table 4). Compared with the pre-PCV13 period, average total costs per IPD episode were marginally lower in the post-PCV13 period (HKD 3,113 vs. 3,575). Although lower average total costs per IPD episode were observed in those aged 18 to <65 years in the post-PCV13 period (HKD 3,182 vs. 4,470), it was slightly higher among adults ≥65 years (HKD 3,062 vs. 2,923). For pneumonia, average total costs per episode were consistently higher in the post-PCV13 period than pre-PCV13 period for adults aged ≥18 years (HKD 2,975 vs. 2,810), adults aged 18 to <65 years (HKD 2,713 vs. 2,532), and adults aged ≥65 years (HKD 3,104 vs. 2,898). Discussion This retrospective database study estimated the clinical and economic burden of pneumococcal disease in Hong Kong from 2005 to 2019, using nationwide data from CDARS. A total of 568 IPD and 3,879 pneumonia hospital episodes were identified. Overall IRs were at 0.78 and 5.34 episodes per 100,000 person-years for IPD and pneumonia, respectively. CFRs for IPD and pneumonia remained high over the study period. There were considerable HCRU and costs associated with each pneumococcal disease. In our study, IPD IR was found to be nearly 7 times higher in older adults aged ≥65 years, compared with those aged 18 to <65 years. This finding is consistent with local data published by the Hong Kong Centre for Health Protection.9 It was reported that IPD annual incidence fluctuated between 1.7 and 2.9 per 100,000 from 2007 to 2015, with incidence similarly higher in the elderly aged ≥65 years.9 In addition, ITSA results from our study did not show a reduction in IPD IRs among adults aged 18 to <65 years and adults aged ≥65 years during the post-PCV13 period. Similar to IPD, overall baseline IRs for pneumonia did not decrease among adults in the post-PCV13 period of our study, compared with pre-PCV13 period. These observations may be attributed to the following reasons. In Hong Kong, the uptake of direct pneumococcal vaccination among the elderly aged ≥65 years remained limited, with cumulative proportion ranging approximately from 33.9% to 43.9% between 2015 and 2019.20 This is despite the government’s efforts to provide free or subsidized vaccination for eligible elderly persons since 2009.11,21 At the time of writing this manuscript, a single dose of PCV13 or PPSV23 was recommended for the elderly without any high-risk conditions.21 The low uptake of direct vaccination may have contributed to no substantial decline of IPD and pneumonia IRs among older adults in our study. Education on the benefits of vaccination, as well as regular screening and encouragement by healthcare providers, may help to improve the vaccination rate.22 Although nationwide childhood vaccination could have indirectly conferred a certain degree of herd protection in adult populations, this may also be disrupted by serotype replacement.23,24 One study reported that serotype 3 has emerged as the primary cause of IPD in Hong Kong adults, after displacing serotype 19F as the most prevalent serotype.25 Among pneumococcal isolates collected during 2015 to 2017, serotype 3 was identified in 20.5% to 36.4% of them.25 Previous research conducted in children has shown that the protection conferred by PCV immunization may be short-lived, particularly against serotype 3.26 This is because the level of serum IgG concentration needed for protection against serotype 3 may be much higher than other PCV7 serotypes, and it is presently difficult for such concentration to be sustained after vaccination.27 While PCV13 includes serotype 3 in its formulation, a study reported that PCV13’s efficacy against serotype 3 was only 26% (95% CI: −69 to 68) in children, which may explain the persistence in transmission despite vaccination in the pediatric population.27 Therefore, advances in vaccine technology to improve immunogenicity against serotype 3 will be important to further reduce the incidence of pneumococcal diseases. Furthermore, our study results were likely influenced to some extent by changes in disease surveillance in Hong Kong. IPD was made a nationally notifiable disease in 2015, requiring all cases of IPD to be reported to the government.10 This could result in surveillance bias, due to a higher frequency of testing compared with the period prior to 2015. Lastly, the development of new diagnostic tools may have enhanced the accuracy of pneumococcal disease diagnosis in more recent years (i.e., post-PCV13 period). The use of these modern diagnostic techniques could improve the sensitivity and specificity of diagnosis for pneumococcal pneumonia, thereby allowing more cases to be detected.28 Taken altogether, these aforementioned factors may have increased the number of IPD and pneumonia episodes identified in the post-PCV13 period of our study. CFRs in our study generally increased with age and were highest in adults aged ≥65 years, compared with those aged 18 to <65 years. Advanced age and underlying chronic diseases have been reported to be predictors of mortality in IPD and pneumonia.29,30 Elderly patients usually present with more severe disease, and they often have other medical conditions which may adversely affect their clinical outcomes.31 A few Hong Kong studies have also examined mortality in adults with pneumococcal disease.3,25 The published fatality rates vary due to differences in study methodology, time period and calculation. In one study, the 30-day mortality rates were 5.8% and 20.4% for patients with noninvasive and invasive pneumococcal disease, respectively.25 A more recent study conducted using data from CDARS reported the overall mortality rates to be 12.1% in IPD and 13.1% in pneumococcal pneumonia between 2015 and 2019.3 These rates similarly increased with age, but were lower than the CFRs estimated in our study. This may be explained by the longer study period in our study (2005–2019) and different approach used for disease identification. IPD and pneumococcal pneumonia could involve lengthy hospital admissions and expensive medical care.3 A study by Chan et al. found that the median length of hospitalization in adults from 2015 to 2019 were 12 days and 9 days for IPD and pneumococcal pneumonia, respectively.3 These findings were slightly shorter, relative to 14 days for IPD and 10 days for pneumococcal pneumonia in the post-PCV13 period of our study. In another retrospective cohort study assessing the burden of pneumococcal disease in Hong Kong from 2011 to 2018, the median length of hospitalization was 8 days among IPD patients with median age of 72.5 years.32 This length of hospitalization was much shorter, compared with 14 days for IPD among those aged ≥65 years in our study (post-PCV13 period). Variation in results could be explained by differences in study design, as well as inclusion and exclusion criteria. The retrospective cohort study was performed in a single hospital using electronic medical records, and only those with positive culture of S. pneumoniae were included.32 The small study population could also limit the representativeness of its findings. In our study, it was observed that length of hospitalization generally increased with age. One exception was for IPD episodes, where the median length of hospitalization was lower in older patients aged ≥65 years (14 days), compared with the younger adults (15 days). As IPD is a more severe condition and could be life-threatening for the elderly, it is plausible that older patients with IPD are more likely to experience early death after hospitalization.30,33,34 This was similarly reported in a separate study, with median length of hospitalization shorter in patients aged ≥65 years, compared with other age groups (18–49 years and 50–64 years).3 The economic burden of pneumococcal diseases was found to be substantial in our study. Compared with pre-PCV13 period, total costs were generally higher in the post-PCV13 period, due to more IPD and pneumonia episodes being identified in the post-PCV13 period of our study. There were also considerable costs among the elderly aged ≥65 years. The higher likelihood for more severe pneumococcal disease, exacerbation of underlying conditions and development of complications could lead to additional medical costs incurred in the older populations.35,36 Furthermore, as inpatient costs contributed substantially to total costs, average total costs per episode was observed to have a positive correlation with length of hospitalization in our study. However, it should be noted that the economic burden reported in this study only consisted of direct costs, and indirect societal costs are excluded. The inclusion of indirect societal costs would be much greater than the reported economic burden in this study. As there are several limitations in our study, result should be interpreted with caution. Firstly, CDARS covers about 80% of all hospitalizations in Hong Kong. Hence, it does not have information on 20% of the population who opt for private healthcare. This may affect the generalizability of our study results. Secondly, the association between pneumococcal disease burden and introduction of PCV13 into CIP may be influenced by other confounders such as vaccination status, changes in disease surveillance, and serotype replacement. In particular, the absence of serotype-level data in this study limits our ability to determine whether serotype replacement and/or serotype persistence are the primary drivers of the observed trends. Thirdly, it is possible for misclassification of pneumococcal disease due to errors in coding or missing diagnosis codes. Lastly, an arbitrary 90-day time gap was used to define each new disease episode and determine whether an episode was new or related to an earlier episode. As such information is typically not reflected in diagnosis codes, bias could be introduced when unrelated hospitalizations were combined into a single episode. Conclusion This study provided an updated assessment of the clinical and economic burden of pneumococcal disease in Hong Kong from 2005 to 2019. A marginal reduction in trend of pneumonia IR was observed after PCV13 introduction into Hong Kong’s CIP. However, CFRs of hospitalized IPD and pneumonia were high, particularly for the elderly and the economic burden of pneumococcal diseases remained substantial. As the long-term protective effects of PCV13 may be threatened by serotype replacement and persistence, use of vaccines with wider coverage or longer-lasting immunogenicity will be critical to ensure that populations remain protected. Future studies to examine the effectiveness of PCV immunization in reducing disease burden among adults should be conducted. In addition, continuous surveillance in Hong Kong to generate local epidemiological information will be important. Supplementary Material Supplementary File.docx Acknowledgments The study authors would like to thank Shiau-Han Chen for providing medical writing support. We would also like to acknowledge the presentation of the study results at a conference in Asian Conference on Pharmacoepidemiology in 2022.37 Biographies Eriko Yamada is an Associate Director in Global Medical and Scientific Affairs at MSD Singapore. She is responsible for the development and execution of medical affairs plans as a therapeutic area lead focusing on vaccines and infectious diseases across Singapore, Malaysia, and Brunei. Eriko holds a PhD in Immunology from University College London. Isaya Sukarom is a Regional Director, Outcomes Research, Asia Pacific, at an MSD subsidiary of Merck & Co., Inc., Rahway, NJ, USA. In this role, she is responsible for outcomes research and health economics activities related to vaccines conducted in Asia Pacific countries. The author of multiple publications on cost-effectiveness analysis, systematic literature reviews, and burden of disease in vaccine and infectious disease areas, Dr Sukarom received her PhD in pharmaceutical socioeconomics from the University of Iowa. Funding Statement This work was supported by Merck Sharp & Dohme LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ, USA. IQVIA was contracted to carry out study activities. The sponsor was also involved in the study design, data interpretation and manuscript review. Disclosure statement CSLC has received grants from the Food and Health Bureau of the Hong Kong Government, Hong Kong Research Grant Council, Hong Kong Innovation and Technology Commission, Pfizer, IQVIA, Merck Sharp & Dohme LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ, USA., and Amgen, as well as personal fees from PrimeVigilance outside the submitted work. XL has received grants from Research Fund Secretariat of the Health Bureau, Health and Medical Research Fund (HMRF, HKSAR), Research Grants Council Early Career Scheme (RGC/ECS, HKSAR), Research Grants Council Research Impact Fund (RGC/RIF, HKSAR), educational and investigator initiate research fund from Janssen and Pfizer; internal funding from the University of Hong Kong as well as consultancy fees from Merck Sharp & Dohme LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ, USA outside the submitted work. SHY, DP, and CTKL are employees of IQVIA and paid consultants to Merck Sharp & Dohme LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ, USA in connection with this study. IS and EY are employees of Merck Sharp & Dohme LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ, USA and shareholders in Merck & Co., Inc., Rahway, NJ, USA. The other authors declare no competing interests. Data availability statement The datasets generated and/or analyzed during the current study are available from the Hong Kong Hospital Authority and cannot be shared publicly as stipulated in the Institutional Review Board application protocol. Inquiries for data access can be made at enquiry@ha.org.hk. Supplementary material Supplemental data for this article can be accessed online at https://doi.org/10.1080/21645515.2025.2583614 References Associated Data Supplementary Materials Supplementary File.docx Data Availability Statement The datasets generated and/or analyzed during the current study are available from the Hong Kong Hospital Authority and cannot be shared publicly as stipulated in the Institutional Review Board application protocol. Inquiries for data access can be made at enquiry@ha.org.hk.