Anastasia Kostaki

A simple digital image processing system to aid in melanoma diagnosis in an everyday melanocytic skin lesion unit: a preliminary report.

Background  For early melanoma diagnosis, experienced dermatologists have an accuracy of 64–80% using clinical diagnostic criteria, usually the ABCD rule, while automated melanoma diagnosis systems are still considered to be experimental and serve as adjuncts to the naked‐eye expert prediction. In an attempt to aid in early melanoma diagnosis, we developed an image processing program with the aim to discriminate melanoma from melanocytic nevi, establishing a mathematical model to come up with a melanoma probability.

Microvascular histopathology in head and neck oncology.

A histological study of both recipient and flap vessels was performed in 30 patients with head and neck cancer, and relevant preoperative risk factors were assessed. A total of 35 free flaps were transferred in 30 patients; 16 patients had preoperative radiotherapy, 13 were smokers, eight had hypertension and six had peripheral vascular disease. No significant venous pathology was found in either the flap or the neck veins. However, over two-thirds of the neck arteries and one-half of the flap arteries were found to have microscopic arterial pathology. The only pre-existing factor significantly influencing vessel pathology was hypertension (P=0.007). All flaps survived, although in two there was some loss of the skin paddle. This study reveals that the majority of patients undergoing microsurgery in the head and neck region have pre-existing arterial damage in both the flap and the recipient arteries, but this does not have a significant effect on the overall patency of the microvascular anastomoses.

Status and perspectives of hospital mortality in a public urban Hellenic hospital, based on a five-year review

BackgroundAnalysis of hospital mortality helps to assess the standards of health-care delivery.MethodsThis is a retrospective cohort study evaluating the causes of deaths which occurred during the years 1995–1999 in a single hospital. The causes of death were classified according to the International Statistical Classification of Diseases (ICD-10).ResultsOf the 149,896 patients who were discharged the 5836 (3.4%) died. Males constituted 55% and females 45%. The median age was 75.1 years (1 day – 100 years).The seven most common ICD-10 chapters IX, II, IV, XI, XX, X, XIV included 92% of the total 5836 deaths.The most common contributors of non-neoplasmatic causes of death were cerebrovascular diseases (I60–I69) at 15.8%, ischemic heart disease (I20–I25) at 10.3%, cardiac failure (I50.0–I50.9) at 7.9%, diseases of the digestive system (K00–K93) at 6.7%, diabetes mellitus (E10–E14) at 6.6%, external causes of morbidity and mortality (V01–Y98) at 6.2%, renal failure (N17–N19) at 4.5%, influenza and pneumonia (J10–J18) at 4.1% and certain infectious and parasitic diseases (A00–B99) at 3.2%, accounting for 65.3% of the total 5836 deaths.Neoplasms (C00–D48) caused 17.7% (n = 1027) of the total 5836 deaths, with leading forms being the malignant neoplasms of bronchus and lung (C34) at 3.5% and the malignant neoplasms of large intestine (C18–21.2) at 1.5%. The highest death rates occurred in the intensive care unit (23.3%), general medicine (10.7%), cardiology (6.5%) and nephrology (5.5%).Key problems related to certification of death were identified. Nearly half of the deaths (49.3%: n = 2879) occurred by the completion of the third day, which indicates the time limits for investigation and treatment. On the other hand, 6% (n = 356) died between the 29th and 262nd days after admission.Inadequacies of the emergency care service, infection control, medical oncology, rehabilitation, chronic and terminal care facilities, as well as lack of regional targets for reducing mortality related to diabetes, recruitment of organ donors, provision for the aging population and lack of prevention programs were substantiated.ConclusionSeveral important issues were raised. Disease specific characteristics, as well as functional and infrastructural inadequacies were identified and provided evidence for defining priorities and strategies for improving the standards of care. Effective transformation can promise better prospects.

An evaluation of the performance of kernel estimators for graduating mortality data

In the graduation of the age-specific mortality pattern, recent emphasis has been placed on the use of kernel regression estimators. Three such estimators are the Nadaraya-Watson, Gasser-Muller and kernel weighted local linear estimators. This paper discusses the theoretical background of each estimator and evaluates their accuracy in graduating age-specific mortality using data for France, Japan and Sweden. For comparison, we also fit the Heligman-Pollard model and its nine-parameter variant by Kostaki. The Gasser-Muller estimator is found to be superior to the two other kernel estimators in that it is both more stable and not influenced by boundary effects. Furthermore, compared with the two parametrric models, the Gasser-Muller estimator provides a more satisfactory graduation, especially at older adult ages, in terms both of smoothness and of fidelity between the observed and graduated rates.

A relational technique for estimating the age-specific mortality pattern from grouped data

This paper presents a new technique of expanding an abridged life table. This technique is a non‐parametric one, which relates the probabilities of dying of the abridged table to those of a standard complete table. In order to evaluate the accuracy of the new technique we use it, as well as two other techniques, for expanding empirical abridged data sets. According to the results of our calculations the new method proves very efficient in producing complete life tables from grouped data, in many cases producing more accurate results than the other two methods in spite of its simplicity.

Degrouping mortality data for the elderly

Proposed in this paper is a technique for estimating, from coarsly grouped empirical death data, the age‐specific numbers of deaths for the elderly population. This question is primarily of interest in countries where the empirical data are available only in a grouped form, given usually in quinquennial age groups and in a large open‐ended interval for the ages 85 and over. The main reason that the official data are given in such a form in some countries of Southern Europe and in the Third World is the existence of heaping in the empirical data, i.e. misstatements in age recording, usually rounding to the nearest integer divisible with five. Our evaluation of the method on Swedish mortality data shows that the technique proposed can be efficiently applied to period mortality data.

Bootstrap techniques for mortality models

Biostatisticians, actuaries and demographers are interested in accurately finding the age specific mortality pattern of a population. Certain different approaches have been proposed in the literature for representing the mortality of a population. Among them laws of mortality assuming a specific functional parametrization for the mortality rates have become popular in recent years mainly because of the existence of computers to handle the vast number of computations needed. The uncertainty of such a functional representation has been overlooked. The researcher is interested in both the uncertainty on the parameters and the uncertainty on the curve itself. The former provides information for specific parts of the curve that directly correspond to certain parameters while the latter allows for comparisons over time or space. Manipulation of the uncertainty can be very helpful for prediction purposes. A bootstrap approach is described, as an alternative to standard inferential methods based on asymptotic standard error theory. Such an approach can provide standard errors for both the parameters and the curve as well as it can be used for direct comparison of different curves over time or space. Application to empirical data from Sweden is also provided.

A parametric model for estimating nuptiality patterns in modern populations

Nuptiality is a phenomenon closely related to fertility. The age-specific marriage distribution has a typical shape common in all human populations. In order to estimate this pattern, alternative parametric models have been proposed. However recent evidence suggests that mixture models are required to estimate nuptiality patterns. In this paper, a flexible parametric model is proposed in three versions, appropriate to describe the age pattern of first marriage rates. For evaluation purposes the models as well as the alternative existing models are fitted to a variety of empirical datasets.

Comparing Smoothing Techniques on Fertility Data

This paper compares different smoothing techniques for graduating fertility rates. In particular we focus on some well-known parametric models, standard non-parametric statistical methods such as kernels and splines, and Support Vector Machines (SVM). In this work, we apply these techniques to empirical age-specific fertility rates from a variety of populations and time periods.

Nonparametric graduation techniques as a common framework for the description of demographic patterns

The graduation of age-specific demographic rates is a subject of special interest in many dis-ciplines as demography, biostatistics, actuarial practice, and social planning. For estimating the unknown age-specific probabilities of the various demographic phenomena, some graduation technique must be applied to the corresponding empirical rates, under the assumption that the true probabilities follow a smooth pattern through age. The classical way for graduating demographic rates is parametric modelling. However, for graduation purposes, nonparametric techniques can also be adapted. This work provides an adaptation, and an evaluation of kernels and Support Vector Machines (SVM) in the context of graduation of demographic rates.