Article Text
PDF
Abstract
Background The objectives were to clarify the trend in the cause-specific mortality rate and changes in health and long-term-care use after the Great East Japan Earthquake in 2011.
Methods We obtained the following data from national sources: the number of deaths by cause, age and month; the amount of healthcare insurance expenditures by type of services, age and month; the amount of long-term-care insurance expenditures by type of services, age, care need and month. We estimated increase in standardised mortality rate postearthquake compared with pre-earthquake, and change in the standardised amount of health and long-term-care insurance expenditures post-earthquake compared with pre-earthquake in three severely affected prefectures, Iwate, Miyagi and Fukushima, by the adjustment for trends in the other prefectures.
Results The risk of indirect mortality increased in the month of the earthquake (relative risk (RR) with 95% CI 1.20 (1.13 to 1.28) for those 60–69 years of age, 1.25 (1.17 to 1.32) for 70–79 years, and 1.33 (1.27 to 1.38) for 80 years and older). The amount of health and long-term-care insurance expenditures decreased among elderly persons in the month of the earthquake, and recovered to 95% of usual level within 1–5 months. Among cities and towns hit by tsunami, higher percentage of households flooded was associated with higher risk of indirect mortality (p<0.001), lower expenditures for outpatient medical care (p<0.001), and lower expenditures for home-care services (p<0.001).
Conclusions This study showed transient increase in indirect mortality and recovery of health and long-term-care system after the earthquake.
- Disaster Relief
- Elderly
- Health Services
- Mortality
Statistics from Altmetric.com
Introduction
A 9.0 magnitude earthquake struck on 11 March 2011, at an epicentre of 130 km east of the northeast coast of Japan's main island. A tsunami triggered by the quake reached Iwate, Miyagi and Fukushima Prefectures within 15 min of the main shock; the maximum wave height was about 9 m or higher in these three prefectures.1 Tsunami flooded 602 000 households in six coastal prefectures; 85% of the households were located in the three prefectures.2 Strong aftershocks followed. Magnitudes of six aftershocks were greater than 7.0.1 The cabinet named the disasters related to the quake, aftershocks, and tsunami the Great East Japan Earthquake.
The total number of dead, missing and injured persons from the direct consequence of the earthquakes and tsunami was 15 884, 2636 and 6147, respectively; 99.6%, 99.8%, and 73.9%, respectively, of these were residents of the three prefectures.3
In addition to these direct deaths, increases in the incidence of communicable and non-communicable health problems were reported in these prefectures. The number of pneumonia hospitalisations increased 2 weeks after the earthquake in select hospitals in Miyagi Prefecture.4 ,5 Frequency of ambulance transports and number of admissions to hospitals of cardiovascular disease cases increased within 2 months of the earthquake in Miyagi Prefecture.6 ,7 The number of peptic ulcer cases increased within 1 month of the earthquake in Miyagi Prefecture.8 Patients with end-stage renal failure or stroke died during transportation from the evacuation area in Fukushima Prefecture.9 Metabolic profiles such as HbA1c deteriorated postquake (compared with prequake examination) among evacuees.10 The Reconstruction Agency reported that 1632 deaths, 95% of whom were 60 years of age or older, were accredited by municipalities as recipients of condolence money for deaths from deterioration of health problems due to the indirect effects of the earthquake, including physical and mental fatigue, deterioration of chronic diseases, inadequate healthcare, and so on.11
More than half of hospitals and clinics were damaged in Iwate, Miyagi and Fukushima Prefectures. Out of 380 hospitals and 4036 clinics in the three prefectures, 10 and 83 facilities were damaged completely, and 290 and 1176 partly, respectively.12 In 39 cities and towns hit by the tsunami in the three prefectures, the number of hospital beds and clinics decreased from 24 548 and 1284 in October 2010 to 22 314 and 1191 in October 2011, respectively, according to the Survey of Medical Institutions. Substantial damage to long-term-care facilities was also reported in these prefectures. For example, 30 and 291 facilities providing home-care services were damaged in the Iwate and Miyagi Prefectures, respectively, and functioning of the services deteriorated substantially in the disaster-affected areas.13
Availability of health and long-term-care services is key to preventing deaths that are likely to increase as living condition deteriorate. Based on the experience of the Great Hanshin-Awaji Earthquake in 1995, Disaster Medical Assistance Teams (DMAT) were established. From 11 March to 22 March 2011 about 340 DMATs consisting of 1500 members were dispatched to Iwate, Miyagi, Fukushima and Ibaraki Prefectures.14 Healthcare teams followed the DMAT, and residents of disaster-affected areas were able to receive medical care service without copayment. We focused on the duration of the impact of the earthquake on the use of health and long-term-care services, as well as health risks, notably among elderly persons.
We predicted that health status and the volume of health and long-term-care services provided in the disaster-affected areas were influenced by the earthquake. It is recommended to consider the amount of health and long-term-care insurance expenditures to capture the total amount of health services provided in Japan.15
The objectives of this study were to clarify the impact of the Great East Japan Earthquake on cause-specific mortality, and to describe recovery in terms of the use of health and long-term-care services after the earthquake.
Methods
We analysed ecological time series data about the number of deaths and the amount of health and long-term-care insurance expenditures. Study areas were Iwate, Miyagi and Fukushima Prefectures, three neighbouring coastal prefectures that were most widely affected by the tsunami during the Great East Japan Earthquake. For city-level and town-level analyses, we selected cities flooded by the tsunami in the three prefectures. Statistical data collected in this study were publicly available, and did not contain information to identify specific individuals.
Data collection
The numbers of deaths by cause, age of decedent (5-year group), month of occurrence (from January 2009 to December 2012), and prefecture, as well as the numbers of deaths by year of occurrence (2010 and 2011), and city and town, were collected from the Vital Statistics produced by the Ministry of Health, Labour and Welfare.
The number of deaths as direct consequences of the earthquakes and tsunami at the city and town level was collected from a portal site that reported damage due to the earthquake operated by each prefecture.16–18
The amounts of healthcare insurance expenditure for inpatient and outpatient medical care by age of person (less than 70 years of age, and 70 years of age and older), month of use (from April 2010 to March 2012), and prefecture were collected from the Survey on the Trend of Medical Care Expenditures by the Ministry of Health, Labour and Welfare. The data covered all types of healthcare insurance.
The amount of expenditures for National Health Insurance operated by municipalities for one fiscal year (FY) per insured person was estimated by the Ministry of Health, Labour and Welfare at the city and town level. This insurance system covered farmers, self-employed persons, retired persons and other persons who were not covered by other health insurance systems. This insurance system did not cover people 75 years of age and older, or those 65–74 years of age certified as having a specific disability.
The amount of long-term-care insurance benefit expenses by type of services (home-care services, community-oriented services, and facility services), level of need (requiring support 1 and 2, and requiring long-term-care levels 1 through 5), month of use (from April 2010 to March 2012), and prefecture was collected from the Survey of Long-term-Care Benefit Expenditures by the Ministry of Health, Labour and Welfare. The amount of long-term-care insurance benefit expenses by type of service, age of person (40 years through 64 years of age, and 65 years of age and older), level of need, FY of use (FY2010 and FY2011), and city and town, was collected from the Annual Report on the Status of the Long-term-Care Insurance Project by the Ministry of Health, Labour and Welfare. We also collected the number of insured persons 65 years of age and older by city and town as of April of each year from the report. Persons with long-term-care insurance were all 40 years of age and older.
The number of households flooded by the earthquakes and tsunami by city and town was estimated by the Statistics Bureau based on a flooded area map developed by the Geospatial Information Authority of Japan using remote sensing images, and on the number of households located in flooded areas according to the 2010 Census.
Data analysis
A mortality rate ratio comparing the Iwate, Miyagi and Fukushima Prefectures with the other 44 prefectures in Japan was calculated for each month between January and September 2011. The age and cause-specific numbers of deaths were summed for Iwate, Miyagi and Fukushima Prefectures, and for the other prefectures for each month between January 2009 and December 2012. The combined population in all other prefectures as of October 2010 was defined as a standard population. Then, age-standardised mortality rates were calculated for the combined population of Iwate, Miyagi and Fukushima Prefectures, and that of all other prefectures. Age and cause-specific background mortality rate ratios comparing these two combined populations were obtained dividing the average mortality rate between January 2009 and December 2010 for the combined population of Iwate, Miyagi and Fukushima Prefectures by that for all other prefectures. An age and cause-specific mortality rate ratio for each month between January and September in 2011 was calculated as a ratio of the standardised mortality rate ratio comparing the two combined populations divided by the background mortality ratio. The mortality rate ratio estimated in this method can be interpreted as the ratio comparing before and after the earthquake for Iwate, Miyagi and Fukushima Prefectures after adjusting for general trend in the mortality rate estimated from all other prefectures. The methods to estimate population by age (5-year group) and sex at national or prefecture level for each month between January 2009 and December 2012 are described in the online supplementary file.
Ratios of the amount of health and long-term-care insurance expenditure per insured person for Iwate, Miyagi and Fukushima Prefectures to the average amount for all other 44 prefectures were estimated for each month between April 2010 and March 2012. The amount of healthcare insurance expenditures for persons 70 years of age and older, and that of long-term-care insurance expenditures for those 40 years of age and older was divided by the estimated population size. Background ratios were calculated comparing the average amount of expenditure between April 2010 and February 2011 for Iwate, Miyagi and Fukushima Prefectures, and that for all other prefectures. Ratios of the amount of expenditure for Iwate, Miyagi and Fukushima Prefectures to that for all other prefectures were calculated adjusting for the background ratios.
The mortality rate from any causes excluding the direct consequences of the earthquakes and tsunami in 2011 was compared with the predicted value estimated by extrapolation from data between 2008 and 2010 for each city or town hit by tsunami. The number of deaths from the direct consequences of the earthquakes and tsunami was subtracted from the total number of deaths in 2011. The denominator of mortality rate was population on 1 October in each year.
Changes in the amount of health and long-term-care insurance expenditures per insured person were estimated for each city or town hit by tsunami. The amount of benefit expenses from the National Health Insurance operated by municipalities, per insured person less than 75 years of age in FY2011, was compared with the predicted value estimated by extrapolation from data between FY2008 and FY2010. The amount of long-term-care insurance expenditure for primary insured persons (those 65 years of age and older) in FY2011 was divided by the number of primary insured persons in April of the year, and was compared with the predicted value estimated by extrapolation from data between FY2008 and FY2010.
Associations between change in mortality rate in 2011 from the trend since 2008, and the percentage of households flooded, and that between the changes in the amount of health and long-term-care insurance expenditures per insured person in 2011 from the trend since 2008, and the percentage of households flooded, at the city and town level were assessed using interrupted time series analysis.19
Results
Table 1 shows trend in the mortality rate ratio comparing Iwate, Miyagi and Fukushima Prefectures with all other prefectures in Japan. The mortality rate ratio from all causes of death excluding the direct consequences of the earthquakes and tsunami increased significantly in March, and returned to the standard level within 3 months. Mortality rate from malignant neoplasms, endocrine, nutritional and metabolic diseases, vascular or unspecified dementia, heart diseases, cerebrovascular diseases, pneumonia, chronic obstructive pulmonary disease and senility, increased significantly for at least one of the age groups in March. An increase in the mortality rate from heart diseases and senility continued until June among persons 80 years of age and over. All the transient increases in the cause-specific mortality rate ended by July. Increase in the mortality rate from suicide was observed in April among persons 70 years of age and older.
Mortality rate ratio from selected causes of deaths comparing Iwate, Miyagi and Fukushima Prefectures with all other prefectures in Japan from January to September 2011
Figure 1 shows the trend in the ratio of the amount of healthcare insurance expenditure per insured person 70 years of age and older for Iwate, Miyagi and Fukushima Prefectures to that for all other prefectures. The ratio for inpatient medical care decreased in March to 0.82 and 0.81 in Miyagi and Fukushima Prefectures, respectively, and recovered to reach 0.90 or higher by April in both prefectures, and 0.95 or higher by May, and July in the respective prefectures. The ratio for outpatient medical care decreased in March to 0.85, 0.74 and 0.79 in Iwate, Miyagi and Fukushima Prefecture, respectively, and recovered to reach 0.90 or higher by April in all prefectures, and 0.95 or higher by April, June and June in the respective prefectures.
Trend in ratio of the amount of healthcare insurance expenditures per insured person 70 years of age and older for Iwate, Miyagi and Fukushima Prefectures, to that for all other prefectures in Japan. The 95% CIs for the ratios estimated by linear regression analyses were shown from March 2011 to March 2012. Ratios were adjusted for the average amount of expenditures between April 2010 and February 2011.
Figure 2 shows the trend in the ratio of the amount of long-term-care insurance expenditures per insured person for Iwate, Miyagi and Fukushima Prefectures to that for all other prefectures after the earthquake. The ratio for home-care services decreased in March to 0.72, 0.57, and 0.68 in Iwate, Miyagi and Fukushima Prefectures, respectively, and recovered to reach 0.90 or higher by April, June and May, and 0.95 or higher by May, August and June in the respective prefecture. The ratio for community-oriented services decreased in March to 0.85, and 0.88 in Miyagi and Fukushima Prefecture, respectively, and recovered to reach 0.90 or higher by April, and 0.95 or higher by May in both prefectures.
Trend in ratio of the amount of long-term-care insurance expenditures per insured person 40 years of age and older in Iwate, Miyagi and Fukushima Prefectures to all other prefectures in Japan. The amount of long-term-care insurance expenditures was a sum of expenditures for levels of required need for long-term-care levels 2 through 5. Home-care services included outpatient daily long-term-care, home-visit long-term-care, outpatient rehabilitation service, short-term admission for daily life long-term-care, daily life care for elderly in specified facility, etc.; community-oriented services included communal daily long-term-care for a dementia patient, multifunctional small group home, outpatient long-term-care for dementia patient, etc.; and facility services included services provided in facilities covered by public aid providing long-term-care to the elderly, long-term-care health facilities, etc. The 95% CIs for the ratios estimated by linear regression analyses were shown from March 2011 to March 2012. Ratios were adjusted for the average amount of expenditures between April 2010 and February 2011.
Table 2 shows increase in the mortality rate from any causes excluding the direct consequences of the earthquakes and tsunami in 2011, and the change in the amount of health and long-term-care insurance expenditures per insured person in FY2011 at the city and town level. The mortality rate increased significantly in 5 out of 37 cities and towns. The amount of home-care services per insured person 65 years of age and older decreased significantly in 8 out of 29 cities and towns, and the amount of community-oriented services increased in 24 out of 29 of cities and towns.
Difference between observed and predicted value of the mortality rate, and the amount of health and long-term-care insurance expenditures per insured person in 2011 in flooded cities and towns in Iwate, Miyagi and Fukushima Prefectures
Table 3 shows the association between increase in mortality rate and percentage of households flooded by the tsunami, and that between change in the amount of health and long-term-care insurance expenditures per insured person and percentage of households flooded by the tsunami. As more proportion of households were flooded, the mortality rate increased (β=14.9 (95% CI 10.6 to 19.3) per 100 000 population), the amount of outpatient medical care decreased more (β=−335.6 (95% CI −526.7 to −144.4)), and that of home-care services decreased more (β=−474.5 (95% CI −683.6 to −265.4)).
Regression coefficient of the absolute deviation in the mortality rate, and the amount of health and long-term-care insurance expenditures per insured person in 2011 from the trend estimated from 2008 to 2010, on the percentage of households flooded among cities and towns hit by the tsunami in Iwate, Miyagi and Fukushima Prefectures
Discussion
Impact on the number of deaths
The morality rate from a wide range of causes increased temporarily among elderly persons. After the earthquake, electricity, gas and water were stopped for a few days to several months, public transportation and roads were damaged, and it became harder to find food and fuel in the study area. Families of long-term-care patients could not contact care providers or city offices due to interruption of telephone service, destruction of roads, or lack of fuel.20 More than 20% of hospitalised patients in a clinic in Iwate Prefecture died in a short period because of lack of water and the cold temperatures.21 More than 75 000 were displaced to community centres and crowded school gymnasiums in the three prefectures.22 Immediately after the earthquake, 205 and 191 hospitals either limited or refused outpatient and inpatient admissions, respectively.12 These changes in the circumstances are thought to have increased the risk of death.
The total number of deaths was elevated significantly until May 2011. According to the Reconstruction Agency, 78% of deaths identified via recipients of condolence money, occurred before June 11th, 2011; 53% of deaths were due to physical and mental fatigue caused by transportation to evacuation facilities or inconvenient daily life at the evacuation facilities, and 19% were due to deterioration of health problems or inadequate provision of healthcare because of reduced functioning of medical facilities.11 It was thought that the increase in the total number of deaths quantified in the present analysis included these deaths.
The mortality rate from pneumonia continued to be high after March. It was suggested that community outbreaks of pneumonia occurred in the study area after the earthquake.23 The fact that most patients hospitalised for pneumonia came from shelters, homes, other hospitals or nursing homes suggested that their pneumonia was not ‘tsunami lung’ but ‘shelter pneumonia’.24
The number of suicide deaths decreased. This fact contradicts the general prediction that the incidence of mental disorder and distress or suicide was likely to increase after the earthquake.25 ,26 Reduction in the suicide rate was also reported after the Great Hanshin-Awaji Earthquake in 1995.27
Impact on health and long-term-care insurance use
The earthquake had an impact on the amount of medical and long-term-care services use among elderly persons. The disturbance was transient, and services usage recovered almost to baseline within a year of the earthquake; however, the amount of home-care services expenditures in Miyagi Prefecture remained below 90% of baseline for 3 months. The amount of facility services tended to be at a higher than usual level through the fifth month after the earthquake, likely a result of the reduction in home-care services use for a relatively longer period.
Strengths and limitations
This study used data about the total population of three prefectures, and provided more comprehensive findings than a facility-based report. It should be noted that data about health and long-term-care insurance use was aggregated at each person's registered residential area. Because of this system, we could follow health and long-term-care use of residents even after evacuation to other places. The analysis at city and town level did not account for seasonal variations in mortality and insurance expenditures that may bias the estimates.
Cost of medical care provided through DMAT and healthcare teams was not included in the healthcare insurance record. The number of healthcare teams decreased during May 2011, and new clinics started to operate in the area.28
Conclusions
Increase in the number of deaths from a wide range of causes continued for a maximum of 3 months. The amount of health and long-term-care insurance expenditures decreased among the elderly population after the earthquake, and recovered to almost standard levels a few months after the earthquake.
What is already known on this subject
-
Health facilities were destroyed, and an increase in the incidence of communicable and non-communicable diseases were reported after the 2011 Great East Japan Earthquake. Medical teams and volunteer health staff provided free medical and long-term-care services to the victims soon after the earthquake. The magnitude of potential excess mortality and the status of health and long-term-care services use should be clarified for the evaluation of ongoing activities relative to health and welfare and for the efficient provision of community health services.
What this study adds
-
The mortality rate from various causes increased immediately after the earthquake; however, the rate returned to the normal level within 3 months. The amounts of health and long-term-care use decreased immediately after the earthquake, and returned to baseline within a year of the earthquake. This study clarified the disturbances due to the earthquake in health and long-term-care systems.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Files in this Data Supplement:
- Data supplement 1 - Online supplement
Footnotes
-
MU and MK contributed equally.
-
Contributors MU and MK conceptualised and designed the study, collected data, performed statistical analysis, and the writing of the manuscript. MU and MK contributed equally. TT conceptualised the study, interpreted the results of analysis and critically revised the manuscript. AM and KS participated in data analysis and interpretation, and critically revised the manuscript.
-
Competing interest None.
-
Provenance and peer review Not commissioned; externally peer reviewed.