OUP user menu

Changes in COPD mortality rate after amendments to the Preventive Vaccination Law in Japan

Kosuke Kiyohara, Noriko Kojimahara, Yasuto Sato, Naohito Yamaguchi
DOI: http://dx.doi.org/10.1093/eurpub/ckr172 133-139 First published online: 12 December 2011

Abstract

Background: The Japanese government amended the Preventive Vaccination Law in November 2001 to specify elderly people aged ≥65 years as the target population for influenza vaccinations. The vaccine coverage among this age group rapidly increased thereafter. Our goal was to evaluate how this amendment affected the nationwide mortality rate of chronic obstructive pulmonary disease (COPD). Methods: The number of monthly COPD deaths by gender and age was obtained from the Monthly Vital Statistics Reports of the Ministry of Health, Labour and Welfare. Data between January 1995 and December 2009 were used for analyses. The COPD mortality rate for each month was calculated separately for the two age groups: age <65 years and age ≥65 years. Changes in the COPD mortality rates after amendment were evaluated each month using the Poisson regression analysis to calculate risk ratios (RRs) and to compute 95% confidence intervals (95% CIs) adjusting for gender, age, trend and seasonal variations. Results: After amendments to the law, a statistically significant reduction in the COPD mortality rates were observed in January (RR 0.84; 95% CI 0.81–0.88), February (RR 0.85; 95% CI 0.81–0.89) and March (RR 0.92; 95% CI 0.88–0.96) among the population aged ≥65 years. However, in the population aged <65 years, statistically no significant changes in the COPD mortality rate were found in any month after the amendments were made. Conclusion: A legal approach to improving influenza vaccine coverage for the elderly population would contribute to the risk reduction of COPD deaths during the influenza season.

Introduction

Chronic obstructive pulmonary disease (COPD) is a worldwide public health issue that is becoming both a substantial and increasing economic and social burden.1 COPD ranked sixth as a leading cause of death worldwide in 1990, and it is projected to become the third by 2020.2 Since the mortality and morbidity rate of patients with COPD are frequently related to its exacerbation, effective strategies to prevent such exacerbations are needed.1 COPD exacerbation is ∼50% more likely to occur in winter than in other seasons,3 and infection from influenza is one of the main causes of exacerbation and increased mortality rate.4 Influenza vaccines can reduce the risk for serious illness and death in COPD patients,59 therefore, an annual influenza vaccination is recommended almost universally in COPD guidelines.1,1014

The Japanese government made amendments to the Preventive Vaccination Law, specifying elderly people aged ≥65 years as the target population for influenza vaccinations in November 2001.15 This amendment aimed at keeping elderly people, many of whom have various underlying diseases, from infection and aggravation of influenza and to improve their prognosis. Municipalities are now legally required to offer vaccinations to the target groups. After the amendment, vaccine coverage among the population aged ≥65 years rapidly increased from 17.2% in the 2000–01 influenza season to 33.9% in 2001–02, to 39.1% in 2002–03, and to 48.1% in the 2003–04 season. The coverage has been stable at ∼50% since the 2004–05 season.16 Although there are no data available about the vaccine coverage before 2000, considering that the Japanese government amended the Preventive Vaccination Law to exclude influenza from the list of target diseases in June 1994 and the amount of influenza vaccines produced plummeted and remained scarce between 1995 and 2000,16 the vaccine coverage during the period must have been much lower.

Although the influenza vaccination is known to reduce the mortality rate among patients with COPD,5,8 the effects of such a legal approach to improving vaccine coverage have not been evaluated. The aim of this study was to assess the effects of the amendment to the Preventive Vaccination Law in 2001 to augment influenza vaccine coverage of elderly people on the nationwide COPD mortality rate.

Methods

Data source

The number of monthly COPD deaths by gender and age was obtained from the Monthly Vital Statistics Reports of the Ministry of Health, Labour and Welfare, targeting all Japanese living in Japan and released 5 months after the survey month.17 COPD is defined in terms of codes J40–44 in the International Statistical Classification of Diseases and Related Health Problems 10th version (ICD-10). Data between January 1995 and December 2009 were analysed. In addition, Japan's total population by gender and age was obtained from reports by the Population Estimation in the Statistics Bureau, Ministry of Internal Affairs and Communications18 for each month.

Statistical analysis

The COPD mortality rate for each month was calculated separately for the two age groups: age <65 years and age ≥65 years. Since the amendment to the Preventive Vaccination Law was enforced in November 2001, the COPD mortality rate was assessed by comparing the two periods: January 1995–November 2001 and December 2001–December 2009. Changes in the COPD mortality rate for each month after amendment were evaluated by using Poisson regression analysis to calculate risk ratios (RRs) and 95% confidence intervals (95% CIs) adjusting for gender, age, trend and seasonal variations. Seasonal variations, in this case, denote the repetitive movement around the trend line estimated for each month in one year. The following formula shows the Poisson regression model that was used. Here, the estimated number of COPD deaths is shown as D, the corresponding population as P, the category of gender and age as G&A, the category of months before/after amendment as M&L, and time-trend, treated as a continuous variable, as T. The constant term and partial regression coefficients are shown as α and β, respectively. Suffixes i, j and k denote the subcategory defined in the respective variables. We examined some intercept-change models as well and chose the model below following the consideration of goodness of fit. Embedded Image All analyses were conducted using IBM SPSS Statistics version 19 for statistical software.

Results

During the study period (January 1995–December 2009), a total of 202 338 COPD deaths were reported. The total number of COPD deaths for men in the age group of <65 years was 5260 and 146 505 in the age group of ≥65 years; for women, 1331 in the age group of <65 years and 49 242 in the age group of ≥65 years.

Overall, the crude COPD mortality rate was 0.88 per 100 000 person–months (age <65 years, 0.04; age ≥65 years, 4.66). After the amendment to the Preventive Vaccination Law, crude COPD mortality rates decreased for both age groups (age <65 years, 0.04–0.03; age ≥65 years, 4.96–4.46). Figure 1 shows the trend of the monthly age-adjusted COPD mortality rate during the study period for each age group. Age-adjusted COPD mortality rates in both age groups were obviously high every winter. In the population aged ≥65 years, the mortality rate decreased year by year, and its seasonal variation gradually attenuated throughout the study period. Among this age group, the age-adjusted mortality rate ranged from 3.08 (September 2000) to 8.37 (January 1995) prior to the law amendment, whereas it ranged from 2.87 (September 2008) to 5.32 (January 2003) after the amendment. However, in the population aged <65 years, no clear trends in age-adjusted mortality rates could be seen throughout the study period. The age-adjusted mortality rate ranged from 0.02 (September 1999) to 0.06 (January 2001) before the amendment, whereas it ranged from 0.01 (October 2009) to 0.04 (January 2009) after the amendment.

Figure 1

Monthly age-adjusted COPD mortality rates in Japan (January 1995 to December 2009)

The COPD mortality rate of each variable and the RRs estimated by the Poisson regression models are shown in table 1 (age <65 years) and in table 2 (age ≥65 years). The COPD mortality rate was lowest among the female population aged <40 years (0.003 per 100 000 person–months), and highest among the male population aged ≥90 years (45.545 per 100 000 person–months). Results of univariable and multivariable analyses showed that male gender and older age were tangible risk factors of COPD death. In multivariable analysis, results showed a statistically significant downward tendency in the trend in COPD mortality rate throughout the study period for both age groups (RR = 0.97 per 12 months).

View this table:
Table 1

COPD mortality rates among the Japanese population aged <65 years from January 1995 to December 2009 and risk ratios estimated by Poisson regression analysis

Examination itemsCOPD mortality (per 100 000 person–month)Univariable analysis RR (95% CI)Multivariable analysis RR (95% CI)
Gender
Male
Age (years)
    <400.003ReferenceReference
    40–440.0102.94 (2.24–3.86)2.94 (2.20–3.91)
    45–490.0236.66 (5.42–8.18)6.72 (5.41–8.35)
    50–540.05315.46 (13.00–18.39)15.45 (12.88–18.53)
    55–600.14642.71 (36.53–49.93)42.16 (35.79–49.67)
    60–640.450131.46 (113.25–152.60)130.06 (111.23–152.09)
Gender
Female
Age (years)
    <400.0030.72 (0.58–0.91)0.72 (0.57–0.92)
    40–440.0051.47 (1.03–2.10)1.47 (1.01–2.14)
    45–490.0102.77 (2.12–3.63)2.80 (2.11–3.72)
    50–540.0164.76 (3.81–5.96)4.76 (3.77–6.03)
    55–600.03710.79 (8.97–12.98)10.66 (8.78–12.94)
    60–640.08524.89 (21.11–29.36)24.64 (20.73–29.30)
Trend per 12 months0.99 (0.97–1.02)0.97 (0.96–0.98)
Seasonal variation
Before amendment
    January0.052ReferenceReference
    February0.0440.85 (0.42–1.69)0.85 (0.73–0.98)
    March0.0410.79 (0.39–1.59)0.78 (0.67–0.91)
    April0.0380.74 (0.36–1.51)0.73 (0.63–0.86)
    May0.0340.66 (0.31–1.39)0.66 (0.56–0.77)
    June0.0300.57 (0.26–1.24)0.57 (0.48–0.67)
    July0.0320.61 (0.28–1.30)0.60 (0.51–0.71)
    August0.0300.59 (0.27–1.27)0.58 (0.49–0.69)
    September0.0290.55 (0.25–1.21)0.55 (0.46–0.65)
    October0.0300.58 (0.27–1.25)0.57 (0.48–0.68)
    November0.0340.66 (0.31–1.39)0.65 (0.55–0.77)
    December0.0400.77 (0.36–1.61)0.75 (0.64–0.88)
After amendment
    January0.0430.82 (0.42–1.62)0.90 (0.75–1.06)
    February0.0390.75 (0.37–1.51)0.82 (0.69–0.97)
    March0.0350.67 (0.32–1.37)0.72 (0.61–0.87)
    April0.0360.68 (0.33–1.40)0.74 (0.62–0.89)
    May0.0330.64 (0.31–1.32)0.69 (0.58–0.83)
    June0.0300.58 (0.28–1.24)0.63 (0.52–0.76)
    July0.0310.60 (0.29–1.27)0.65 (0.54–0.78)
    August0.0300.58 (0.27–1.23)0.62 (0.52–0.75)
    September0.0290.56 (0.26–1.20)0.60 (0.50–0.72)
    October0.0310.59 (0.28–1.24)0.63 (0.52–0.76)
    November0.0370.71 (0.35–1.44)0.76 (0.63–0.90)
    December0.0400.77 (0.40–1.51)0.82 (0.70–0.97)
  • Before amendment: January 1995–November 2001

  • After amendment: December 2001–December 2009

View this table:
Table 2

COPD mortality rates among the Japanese population aged ≥65 years from January 1995 to December 2009 and risk ratios estimated by Poisson regression analysis

Examination itemsCOPD mortality (per 100 000 person–month)Univariable analysis RR (95% CI)Multivariable analysis RR (95% CI)
Gender
Male
Age (years)
    65–691.339ReferenceReference
    70–743.8372.89 (2.79–3.00)2.87 (2.70–3.05)
    75–799.7217.43 (7.18–7.68)7.26 (6.86–7.69)
    80–8420.19815.39 (14.89–15.90)15.08 (14.26–15.95)
    85–8933.44225.34 (24.51–26.20)24.98 (23.58–26.45)
    ≥9045.54535.12 (33.87–36.41)34.01 (31.96–36.20)
Gender
Female
Age (years)
    65–690.2100.16 (0.14–0.17)0.16 (0.14–0.18)
    70–740.5840.44 (0.41–0.46)0.44 (0.40–0.48)
    75–791.4981.13 (1.08–1.18)1.12 (1.04–1.21)
    80–843.2262.44 (2.35–2.54)2.41 (2.25–2.58)
    85–896.2324.74 (4.56–4.92)4.65 (4.36–4.97)
    ≥9011.2988.77 (8.45–9.11)8.44 (7.91–9.01)
Trend per 12 months0.99 (0.97–1.00)0.97 (0.97–0.97)
Seasonal variation
Before amendment
    January6.913ReferenceReference
    February6.1270.89 (0.56–1.39)0.89 (0.85–0.92)
    March5.7810.84 (0.53–1.32)0.83 (0.80–0.87)
    April5.1080.74 (0.46–1.19)0.74 (0.71–0.77)
    May4.8270.70 (0.43–1.13)0.70 (0.67–0.73)
    June4.2430.61 (0.37–1.01)0.61 (0.58–0.64)
    July4.2690.62 (0.37–1.02)0.61 (0.59–0.64)
    August4.2610.62 (0.37–1.02)0.61 (0.59–0.64)
    September3.9500.57 (0.34–0.95)0.57 (0.54–0.59)
    October4.3660.63 (0.39–1.04)0.63 (0.60–0.65)
    November4.6000.67 (0.41–1.08)0.66 (0.63–0.69)
    December5.1970.75 (0.46–1.23)0.74 (0.71–0.77)
After amendment
    January5.3490.77 (0.51–1.18)0.84 (0.81–0.88)
    February4.7810.69 (0.45–1.07)0.75 (0.72–0.79)
    March4.8820.71 (0.46–1.09)0.77 (0.73–0.80)
    April4.5520.66 (0.42–1.02)0.71 (0.68–0.75)
    May4.4620.65 (0.41–1.01)0.70 (0.67–0.73)
    June4.0740.59 (0.37–0.93)0.64 (0.61–0.67)
    July4.0390.58 (0.37–0.92)0.63 (0.60–0.66)
    August4.0350.58 (0.37–0.92)0.63 (0.60–0.66)
    September3.8290.55 (0.35–0.88)0.60 (0.57–0.62)
    October4.1990.61 (0.39–0.95)0.65 (0.62–0.68)
    November4.3980.64 (0.41–0.99)0.68 (0.65–0.71)
    December4.8590.70 (0.46–1.07)0.75 (0.72–0.78)
  • Before amendment: January 1995–November 2001.

  • After amendment: December 2001–December 2009.

Figure 2 shows the risk of COPD death prior to and after the amendment to the law for each month, estimated by the Poisson regression models. After the amendment, a statistically significant risk reduction was observed in January (RR 0.84; 95% CI 0.81–0.88), February (RR 0.85; 95% CI 0.81–0.89) and March (RR 0.92; 95% CI 0.88–0.96) among the population aged ≥65 years. On the other hand, no statistically significant changes were found in the risk of COPD death among the population aged <65 years in any month after the amendment.

Figure 2

Risk of COPD death before and after the amendment to the Preventive Vaccination Law in Japan estimated from the multiple Poisson regression analysis

Discussion

In the current study, we investigated whether the 2001 amendment to the Preventive Vaccination Law affected the nationwide mortality rate of COPD, and if so, how. Our results showed 8–16% of estimated risk reduction in COPD death per month during the influenza season (January–March) in the population aged ≥65 years after the amendment. This amendment originally aimed to keep elderly people from infection and aggravation of influenza, and did not particularly target COPD patients. One plausible explanation for our finding is that the influenza vaccine coverage of elderly COPD patients would have increased after the amendment along with that of the general population. The Japanese government amended the Preventive Vaccination Law to exclude influenza from the list of target diseases in June 1994 and the amount of influenza vaccines produced plummeted.16 Therefore, it can be assumed that many COPD patients would not have been vaccinated until the amendment was made in November 2001, although the influenza vaccination was recommended for COPD patients according to COPD guidelines.1,1014 Since the amendment to the law in 2001, municipalities must provide vaccinations to the target groups with a subsidy of ∼80% of the influenza vaccination cost. The cost of providing the influenza vaccine, including not only the cost of the vaccine and the cost of its administration, but also a health consultation fee to determine whether it is indicated, is roughly 5000 Japanese Yen (¥); the municipality provides a subsidy (¥4000), and the individual contributes a self-payment (¥1000). However, the cost varies among municipalities and a fixed cost would depend on an agreement between the municipal government and the medical association of each community.15 As vaccine coverage among the general population aged ≥65 years rapidly increased since the 2001–02 influenza season,16 many COPD patients in this age group also would have been vaccinated and received its benefit.

Despite the fact that the population aged <65 years was not the target of the amendment in 2001, their vaccine coverage also increased from 3.9% in the 2000–01 influenza season to 27.7% in the 2008–09 season.16 Our results, however, show no statistically significant changes in the risk of COPD death after the amendment in any month among this age group. The plausible reasons for the lack of detectable change would be that the number of COPD deaths among this age group was too small to indicate any statistically significant differences, and/or that the effect of influenza vaccines on this age group was insufficient in reducing COPD deaths. As shown in previous studies,1921 aging is a strong predictive factor of COPD morbidity and mortality rates. In addition, since younger COPD patients are likely to be categorized as ‘mild’ or ‘moderate’22 according to the GOLD criteria,1 they would have a better prognosis than elderly patients. It is assumed that the influence of influenza infection on the prognosis of COPD would probably be small among the population aged <65 years, and therefore, the effect of increased vaccine coverage could not be detected in our results.

The Vital Statistics Report we applied for this study gives high-quality mortality rate statistics covering the entire Japanese population and is the only available data in Japan that is applicable to our population-based analysis. However, the report of COPD deaths must be interpreted cautiously because the accuracy of the COPD mortality data is affected by its terminology, recognition and awareness of the disease.1 In the Vital Statistics Report, underlying cause of death, which is defined as the disease or condition that initiated the train of morbid events leading directly to death, is adopted in death registration. Therefore, the reliability of this report would largely depend on the notability of COPD and the diffusion of inspection methods in Japan. In addition, COPD is likely to be listed as a contributory cause of death or omitted entirely from the death certificate and the death attributed to another condition. In fact, the main causes of death reported among COPD patients were cancer and cardiovascular-related diseases.23,24 Thus, death registration data may not accurately depict the COPD mortality rate. As for COPD terminology, since we applied the data after the introduction of ICD-10 in Japan, the problem of labeling would not affect the results. However, since spirometry, which is essential for detecting early stage COPD, has not been widely used in Japan, the prevalence of COPD has been much underreported. One recent study22 revealed that not <8.6% of the general population aged ≥40 years, i.e. ∼5.3 million people, was estimated to have airflow limitation consistent with COPD diagnosis in Japan. However, only 9.4% of them actually reported a previous diagnosis of COPD. Since the cause of death for undiagnosed COPD patients is seldom listed as COPD in death certificates,25 those deaths were not included in our analyses. Considering that many patients without COPD diagnosis would have mild or moderate COPD,22 they would have a better prognosis than those with COPD diagnosis. Therefore, the effects of the influenza vaccination in patients with undiagnosed COPD would not be equivalent to those with diagnosed COPD. However, since >90% of patients with COPD remain without diagnosis, not a few undiagnosed patients would have been vaccinated after the amendment in 2001 and received the benefit as well, whereas it cannot be evaluated from the Vital Statistics Report.

Since our study was limited to some socio-demographic factors including the amendment to the Preventive Vaccination Law, the observed reduction of the risk of COPD death might have occurred due to changes in other risk factors during the study period. We did not consider the important prognostic factors of COPD, such as tobacco smoking and long-term oxygen therapy (LTOT) or home oxygen therapy (HOT).1,13,26 Tobacco smoking is the most well-known risk factor of COPD morbidity and progression.2729 Smoking cessation can prevent or delay the development of airflow limitation or reduce its progression,30 and can have a substantial effect on subsequent mortality rates.23 Smoking prevalence in Japan has been much higher in men than in women,31 and it would be a plausible explanation for the statistically significant gender difference in COPD mortality rates, as shown in our results. Recently, smoking prevalence among men has been consistently decreasing and the gender difference has attenuated.31 However, we assume that these gradual changes in smoking prevalence would have affected the year-by-year trend in mortality rate but this does not explain the distinct winter-specific change as seen in our results. In addition, although LTOT/HOT is known to improve the prognosis of COPD patients with severe chronic respiratory failure,32,33 no exceptional changes were made in their application criteria during the study period. Thus, to our knowledge, there are no plausible factors that could have caused the winter-specific risk reduction of COPD deaths among elderly people after the 2001–02 influenza season, other than the amendment to the Preventive Vaccination Law. We believe that the changes in influenza vaccine coverage due to the amendment to the law should be the single explanation for our findings.

In summary, the results of this study suggest that the amendment to the Preventive Vaccination Law in November 2001 contributes to the reduction of COPD mortality rates among the elderly population during the influenza season. To our knowledge, this is the first report examining the nationwide effect of legal intervention to improve influenza vaccine coverage. Considering that in Japan, the influenza vaccine coverage for the population aged ≥65 years is ∼50%, and is not yet adequate even after the amendment, additional interventions to improve vaccine coverage could lead to further reduction of COPD deaths. Our findings should be helpful for policy makers who deal with influenza and COPD.

Funding

A research grant for emerging and re-emerging infectious diseases from the Ministry of Health, Labour and Welfare of Japan.

Conflicts of interest: None declared.

Key points

  • We set out to determine whether the amendment to the Preventive Vaccination Law in November 2001 affected the nationwide mortality rate of COPD.

  • The results of this study suggest that the amendment to the Preventive Vaccination Law contributes to the reduction of COPD mortality rates among the elderly population during the influenza season.

  • This is the first report examining the nationwide effect of legal intervention to improve influenza vaccine coverage.

References

View Abstract