Modelling and Forecasting Mortality Patterns

Thesis submitted in partial fulfillment of the requirements for the Degree of Doctor of Statistics and Information Management

The role of longevity risk in Hungary

In this paper, I analyze the role of longevity risk in Hungary in the public pension system and the life annuity segment of the life insurance market, which are two primary financial sectors of relevance to this special type of actuarial risk, using state-of-the- art econometric methodology. To this end, I present an overview and the mathematical background of several important current mortality forecasting techniques from the Lee–Carter model up to unifying paradigm of the Age–Period–Cohort family of models. After presenting the findings of a case study on the public pension system based on the paper of Bajk ́o, Maknics, T ́oth and V ́ekas, I conclude that longevity risk jeopardizes the sustainability of the Hungarian public pension system in the long run. In another case study, I present an analysis of the role of longevity risk in the pre- mium of private pension annuities, a relevant topic due to recent changes in a law on Hungarian voluntary pension funds, following an earlier analysis of M ́ajer and Kov ́acs. Based on the criterion on out-of-sample forecasting accuracy, I find that the Cairns–Blake– Dowd mortality forecasting model aimed specifically at modeling old-age mortality outperforms the Lee–Carter model applied by M ́ajer and Kov ́acs . Based on numerical results, I finally conclude that the role of longevity risk in the Hungarian life annuity mar- ket has increased significantly in the past decade and is likely to further increase in the future.

Improving the health forecasting alert system for cold weather and heat-waves in England: a proof-of-concept using temperature-mortality relationships

Objectives In this study a prototype of a new health forecasting alert system is developed, which is aligned to the approach used in the Met Office’s (MO) National Severe Weather Warning Service (NSWWS). This is in order to improve information available to responders in the health and social care system by linking temperatures more directly to risks of mortality, and developing a system more coherent with other weather alerts. The prototype is compared to the current system in the Cold Weather and Heatwave plans via a case-study approach to verify its potential advantages and shortcomings. Method The prototype health forecasting alert system introduces an “impact vs likelihood matrix” for the health impacts of hot and cold temperatures which is similar to those used operationally for other weather hazards as part of the NSWWS. The impact axis of this matrix is based on existing epidemiological evidence, which shows an increasing relative risk of death at extremes of outdoor temperature beyond a threshold which can be identified epidemiologically. The likelihood axis is based on a probability measure associated with the temperature forecast. The new method is tested for two case studies (one during summer 2013, one during winter 2013), and compared to the performance of the current alert system. Conclusions The prototype shows some clear improvements over the current alert system. It allows for a much greater degree of flexibility, provides more detailed regional information about the health risks associated with periods of extreme temperatures, and is more coherent with other weather alerts which may make it easier for front line responders to use. It will require validation and engagement with stakeholders before it can be considered for use.

The out-of-sample performance of stochastic methods in forecasting age-specific mortality rates /

"July 2008."

Coherent forecasting of multiple-decrement life tables: a test using Japanese cause of death data

Planners in public and private institutions would like coherent forecasts of the components of age-specic mortality, such as causes of death. This has been di cult toachieve because the relative values of the forecast components often fail to behave ina way that is coherent with historical experience. In addition, when the group forecasts are combined the result is often incompatible with an all-groups forecast. It hasbeen shown that cause-specic mortality forecasts are pessimistic when compared withall-cause forecasts (Wilmoth, 1995). This paper abandons the conventional approachof using log mortality rates and forecasts the density of deaths in the life table. Sincethese values obey a unit sum constraint for both conventional single-decrement life tables (only one absorbing state) and multiple-decrement tables (more than one absorbingstate), they are intrinsically relative rather than absolute values across decrements aswell as ages. Using the methods of Compositional Data Analysis pioneered by Aitchison(1986), death densities are transformed into the real space so that the full range of multivariate statistics can be applied, then back-transformed to positive values so that theunit sum constraint is honoured. The structure of the best-known, single-decrementmortality-rate forecasting model, devised by Lee and Carter (1992), is expressed incompositional form and the results from the two models are compared. The compositional model is extended to a multiple-decrement form and used to forecast mortalityby cause of death for Japan

Cancer mortality in Europe, 2000-2004, and an overview of trends since 1975.

BACKGROUND: To update the pattern of cancer mortality in Europe. Materials and methods: We analysed cancer mortality in 34 European countries during 2000-2004, with an overview of trends in 1975-2004 using data from the World Health Organization. RESULTS: From 1990-1994 to 2000-2004, overall cancer mortality in the European Union declined from 185.2 to 168.0/100 000 (world standard, -9%) in men and from 104.8 to 96.9 (-8%) in women, with larger falls in middle age. Total cancer mortality trends were favourable, though to a variable degree, in all major European countries, including Russia, but not in Romania. The major determinants of these favourable trends were the decline of lung (-16%) and other tobacco-related cancers in men, together with the persistent falls in gastric cancer, and the recent appreciable falls in colorectal cancer. In women, relevant contributions came from the persistent decline in cervical cancer and the recent falls in breast cancer mortality, particularly in northern and western Europe. Favourable trends were also observed for testicular cancer, Hodgkin lymphomas, leukaemias, and other neoplasms amenable to treatment, though the reductions were still appreciably smaller in eastern Europe. CONCLUSION: This updated analysis of cancer mortality in Europe showed a persistent favourable trend over the last years.

Continuing declines in cancer mortality in the European Union

BACKGROUND: From 1988 to 1997 age-standardised total cancer mortality rates in the European Union (EU) fell by around 9% in both sexes. Available cancer mortality data in Europe up to 2002 allow a first check of the forecast of further declines in cancer mortality. PATIENTS AND METHODS: We considered trends in age-standardised mortality from major cancer sites in the EU during the period 1980-2002. RESULTS: For men, total cancer mortality, after a peak of 191.1/100,000 in 1987 declined to 177.8 in 1997 (-7%), and to 166.5 in 2002. Corresponding figures for females were 107.9/100,000, 100.5 and 95.2, corresponding to falls of 7% from 1987 to 1997, and to 5% from 1997 to 2002. Over the last 5 years, lung cancer declined by 1.9% per year in men, to reach 44.4/100,000, but increased by 1.7% in women, to reach 11.4. In 2002, for the first year, lung cancer mortality in women was higher than that for intestinal cancer (11.1/100,000), and lung cancer became the second site of cancer deaths in women after breast (17.9/100,000). From 1997 to 2002, appreciable declines were observed in mortality from intestinal cancer in men (-1.6% per year, to reach 18.8/100,000), and in women (-2.5%), as well as for breast (-1.7% per year) and prostate cancer (-1.4%). CONCLUSIONS: Despite the persisting rises in female lung cancer, the recent trends in cancer mortality in the EU are encouraging and indicate that an 11% reduction in total cancer mortality from 2000 to 2015 is realistic and possible.

Future development of gastrointestinal cancer incidence and mortality rates in Switzerland: a tumour registry- and population-based projection up to 2030.

QUESTIONS UNDER STUDY: Since tumour burden consumes substantial healthcare resources, precise cancer incidence estimations are pivotal to define future needs of national healthcare. This study aimed to estimate incidence and mortality rates of oesophageal, gastric, pancreatic, hepatic and colorectal cancers up to 2030 in Switzerland. METHODS: Swiss Statistics provides national incidences and mortality rates of various cancers, and models of future developments of the Swiss population. Cancer incidences and mortality rates from 1985 to 2009 were analysed to estimate trends and to predict incidence and mortality rates up to 2029. Linear regressions and Joinpoint analyses were performed to estimate the future trends of incidences and mortality rates. RESULTS: Crude incidences of oesophageal, pancreas, liver and colorectal cancers have steadily increased since 1985, and will continue to increase. Gastric cancer incidence and mortality rates reveal an ongoing decrease. Pancreatic and liver cancer crude mortality rates will keep increasing, whereas colorectal cancer mortality on the contrary will fall. Mortality from oesophageal cancer will plateau or minimally increase. If we consider European population-standardised incidence rates, oesophageal, pancreatic and colorectal cancer incidences are steady. Gastric cancers are diminishing and liver cancers will follow an increasing trend. Standardised mortality rates show a diminution for all but liver cancer. CONCLUSIONS: The oncological burden of gastrointestinal cancer will significantly increase in Switzerland during the next two decades. The crude mortality rates globally show an ongoing increase except for gastric and colorectal cancers. Enlarged healthcare resources to take care of these complex patient groups properly will be needed.

Coherent forecasting of multiple-decrement life tables: a test using Japanese cause of death data

Planners in public and private institutions would like coherent forecasts of the components of age-specic mortality, such as causes of death. This has been di cult to achieve because the relative values of the forecast components often fail to behave in a way that is coherent with historical experience. In addition, when the group forecasts are combined the result is often incompatible with an all-groups forecast. It has been shown that cause-specic mortality forecasts are pessimistic when compared with all-cause forecasts (Wilmoth, 1995). This paper abandons the conventional approach of using log mortality rates and forecasts the density of deaths in the life table. Since these values obey a unit sum constraint for both conventional single-decrement life tables (only one absorbing state) and multiple-decrement tables (more than one absorbing state), they are intrinsically relative rather than absolute values across decrements as well as ages. Using the methods of Compositional Data Analysis pioneered by Aitchison (1986), death densities are transformed into the real space so that the full range of multivariate statistics can be applied, then back-transformed to positive values so that the unit sum constraint is honoured. The structure of the best-known, single-decrement mortality-rate forecasting model, devised by Lee and Carter (1992), is expressed in compositional form and the results from the two models are compared. The compositional model is extended to a multiple-decrement form and used to forecast mortality by cause of death for Japan

[impact Variables On The Decline In Infant Mortality In The State Of São Paulo, Brazil: 1998-2008].

This is an ecological, analytical and retrospective study comprising the 645 municipalities in the State of São Paulo, the scope of which was to determine the relationship between socioeconomic, demographic variables and the model of care in relation to infant mortality rates in the period from 1998 to 2008. The ratio of average annual change for each indicator per stratum coverage was calculated. Infant mortality was analyzed according to the model for repeated measures over time, adjusted for the following correction variables: the city's population, proportion of Family Health Programs (PSFs) deployed, proportion of Growth Acceleration Programs (PACs) deployed, per capita GDP and SPSRI (São Paulo social responsibility index). The analysis was performed by generalized linear models, considering the gamma distribution. Multiple comparisons were performed with the likelihood ratio with chi-square approximate distribution, considering a significance level of 5%. There was a decrease in infant mortality over the years (p < 0.05), with no significant difference from 2004 to 2008 (p > 0.05). The proportion of PSFs deployed (p < 0.0001) and per capita GDP (p < 0.0001) were significant in the model. The decline of infant mortality in this period was influenced by the growth of per capita GDP and PSFs.

Hypothermia And Early Neonatal Mortality In Preterm Infants.

To evaluate intervention practices associated with hypothermia at both 5 minutes after birth and at neonatal intensive care unit (NICU) admission and to determine whether hypothermia at NICU admission is associated with early neonatal death in preterm infants. This prospective cohort included 1764 inborn neonates of 22-33 weeks without malformations admitted to 9 university NICUs from August 2010 through April 2012. All centers followed neonatal International Liaison Committee on Resuscitation recommendations for the stabilization and resuscitation in the delivery room (DR). Variables associated with hypothermia (axillary temperature <36.0 °C) 5 minutes after birth and at NICU admission, as well as those associated with early death, were analyzed by logistic regression. Hypothermia 5 minutes after birth and at NICU admission was noted in 44% and 51%, respectively, with 6% of early neonatal deaths. Adjusted for confounding variables, practices associated with hypothermia at 5 minutes after birth were DR temperature <25 °C (OR 2.13, 95% CI 1.67-2.28), maternal temperature at delivery <36.0 °C (OR 1.93, 95% CI 1.49-2.51), and use of plastic bag/wrap (OR 0.53, 95% CI 0.40-0.70). The variables associated with hypothermia at NICU admission were DR temperature <25 °C (OR 1.44, 95% CI 1.10-1.88), respiratory support with cold air in the DR (OR 1.40, 95% CI 1.03-1.88) and during transport to NICU (OR 1.51, 95% CI 1.08-2.13), and cap use (OR 0.55, 95% CI 0.39-0.78). Hypothermia at NICU admission increased the chance of early neonatal death by 1.64-fold (95% CI 1.03-2.61). Simple interventions, such as maintaining DR temperature >25 °C, reducing maternal hypothermia prior to delivery, providing plastic bags/wraps and caps for the newly born infants, and using warm resuscitation gases, may decrease hypothermia at NICU admission and improve early neonatal survival.

Evolution Of The Deaths Registry System In Brazil: Associations With Changes In The Mortality Profile, Under-registration Of Death Counts, And Ill-defined Causes Of Death.

This paper examines the spatial pattern of ill-defined causes of death across Brazilian regions, and its relationship with the evolution of completeness of the deaths registry and changes in the mortality age profile. We make use of the Brazilian Health Informatics Department mortality database and population censuses from 1980 to 2010. We applied demographic methods to evaluate the quality of mortality data for 137 small areas and correct for under-registration of death counts when necessary. The second part of the analysis uses linear regression models to investigate the relationship between, on the one hand, changes in death counts coverage and age profile of mortality, and on the other, changes in the reporting of ill-defined causes of death. The completeness of death counts coverage increases from about 80% in 1980-1991 to over 95% in 2000-2010 at the same time the percentage of ill-defined causes of deaths reduced about 53% in the country. The analysis suggests that the government's efforts to improve data quality are proving successful, and they will allow for a better understanding of the dynamics of health and the mortality transition.

Insect mortality in Spathodea campanulata Beauv. (Bignoniaceae) flowers

537

Association of HDL cholesterol and triglycerides with mortality in patients with heart failure

It has been demonstrated that there is an association between serum lipoproteins and survival rate in patients with ischemic cardiomyopathy, as well as in patients with non-ischemic causes of heart failure. We tested the hypothesis of an association between serum lipoprotein levels and prognosis in a cohort of outpatients with heart failure, including Chagas' heart disease. The lipid profile of 833 outpatients with heart failure in functional classes III and IV of the New York Heart Association, with a mean age of 46.9 ± 10.6 years, 655 (78.6%) men and 178 (21.4%) women, was studied from April 1991 to June 2003. The survival rate was estimated by the Kaplan-Meyer's method and the Cox proportional hazards models. Etiology of heart failure was ischemic cardiomyopathy in 171 (21%) patients, Chagas' heart disease in 144 (17%), hypertensive cardiomyopathy in 136 (16%), and other etiologies in 83 (10%). In 299 (36%) patients, heart failure was ascribed to idiopathic dilated cardiomyopathy. Variables significantly associated with mortality were age (hazard ratio, HR = 1.02; 95%CI = 1.01-1.03; P = 0.0074), male gender (HR = 1.77; 95%CI = 1.2-2.62; P = 0.004), idiopathic dilated cardiomyopathy (HR = 1.81; 95%CI = 1.16-2.82; P = 0.0085), serum triglycerides (HR = 0.97; 95%CI = 0.96-0.98; P < 0.0001), and HDL cholesterol (HR = 0.99; 95%CI = 0.99-1.0; P = 0.0280). Therefore, higher serum HDL cholesterol and higher serum triglycerides were associated with lower mortality in this cohort of outpatients with heart failure.

Growth parameters and mortality rate of the Scomber japonicus peruanus (Jordán & Hubb, 1925) along the Peruvian coast, South Pacific

The growth parameters and the mortality rates of the Scomber japonicus peruanus (Chub mackerel) were studied based on monthly data of frequency of fork length classes obtained from commercial landings off the Peruvian coast from 1996 to 1998. The asymptotic body length and growth rate values obtained by the ELEFAN I (Electronic Length Frequency Analysis) ranged from 40.20 cm to 42.20 cm and from 0.38 to 0.39, respectively. The oscillation amplitude was 0.60; the Winter point values varied from 0.50 to 0.60 and the performance index from 2.79 to 2.84. The total mortality rate of the Chub mackerel obtained by the linearized catch curve oscillated between 1.68 and 3.35. The rate of fishing mortality varied from 1.16 to 2.78 and the exploitation rate from 0.68 to 0.84. The annual rate of natural mortality estimated by the Pauly's method ranged from 0.52 to 0.53. The results obtained allow us to conclude that the longevity of the Chub mackerel was slightly over seven years.