A clinico-pathological study of geriatric anemias.

INTRODUCTION: Anemia in the older age (e.g., >60 years) is a major health problem in India and many parts of the world since it signifies an underlying disease and is associated with poor clinical outcome like increased morbidity and affects health-related quality of life. Since symptoms like fatigue or shortness of breath related to anemia could also be attributed to the aging process, anemia is often easily overlooked in the elderly. AIMS AND OBJECTIVES: Clinico-hematological patterns and morphological types of anemia in older age (e.g. >60 years) are manifold, hence this study was undertaken to determine them and to know more about associated disorders. MATERIALS AND METHODOLOGY: The present study was conducted on a sample size of 1257 patients who were 60 years and above and clinically diagnosed as anemic. Routine haematological investigations including peripheral blood smear examination and complete hemogram were done. Special investigations like bone-marrow examination and iron studies were done whenever required. RESULTS: Males (aged >60 years) were more affected than females (aged >60 years) and patients in the age group of 60-69 years were affected the most. The most common presenting symptom was generalized weakness. The most common morphological type was normocytic normochromic anemia, and chronic diseases were the commonest etiological factors. CONCLUSION: In spite of modern diagnostic advances, geriatric anemias still remain under-reported and inadequately investigated, necessitating evaluation of even mild anemias. Prompt diagnosis and definite categorization helps in appropriate management of anemias.

Multidisciplinary approach to L3/L4 lumbar disc prolapse masquerading as focal limb myositis-a radiological challenge.

Prolapsed intervertebral discs are commonly associated with back ache and sciatica. Management is often conservative with analgesia and physiotherapy. Nerve root injections and discectomy procedures are used where conservative measures fail. Majority of patients present with symptoms of pain and motor weakness; however, a few can present as focal myositis of lower limb muscles in the distribution of radiculopathy. MRI scans of limbs are rarely done in these cases but if done can confound the radiologist. Our case report emphasize the importance of multidisciplinary approach for a L3 nerve radiculopathy with confounding clinical presentation of focal lower limb myositis of unknown etiology.

Assessment of sensitivity of whole body CT for major trauma.

INTRODUCTION: Whole body computed tomography has become standard practice in many centres in the management of severely injured trauma patients, however, the evidence for it's diagnostic accuracy is limited. AIM: To assess the sensitivity of whole body CT in major trauma. METHOD: Retrospective review of all patients with injury severity score (ISS) > 15 presenting with blunt trauma to a UK Major Trauma Centre between May 2012 and April 2014. Injuries were classified as per ISS score-1 = head and neck 2 = face 3 = chest 4 = abdomen. The authors reviewed patient's electronic charts, radiological results; interventional procedure records, discharge letters and outpatient follow up documentation and referenced this with Trauma Audit and Research Network data. RESULTS: 407 patients with ISS > 15 presented to the Trauma centre during May 2012 and April 2014. Of these, 337 (82.8%) had a whole body CT scan. 246 pts were male, 91 were female. 74 (21.9%) were due to a fall from > 2 m, 41 (12.2%) due to a fall from < 2 m, 208 (61.7%) were due to motor vehicle crashes, 1 (0.3%) due to a blast injury, 5 (1.5%) due to blows, and 8 (2.4%) due to crush injuries. Sensitivity for Region 1 was 0.98, Region 2 = 0.98, Region 3 = 0.98 and Region 4 was 0.95. Overall sensitivity was 0.98. 15 injuries (2.4%) were not identified on initial CT (false -ve). These injuries were: colonic perforation = 1, splenic contusion = 1, pneumothorax = 1, liver laceration = 1, intracranial haemorrhage = 1, cerebral contusions = 1, spinal injuries = 7, canal haemorrhage = 1, maxilla fracture = 1. CONCLUSION: These results show that whole body CT in trauma has a high sensitivity and a low rate of missed injuries (2.4%). However, our study only evaluated a subgroup of patients with ISS > 15 and further work is required to assess the use of this investigation for all major trauma patients.

Patient attitudes to tonsillectomy.

Introduction. Recent changes to primary care trusts' Procedures of Limited Clinical Value (PLCV) policy mean that otolaryngologists must now follow policy rather than exercising clinical judgment when listing patients for tonsillectomy. Objectives. To gauge perception within the general public of when tonsillectomy is acceptable and to compare this to the current policy. Method. All patients or their parents attending the adult and paediatric outpatient ENT departments were asked to anonymously complete questionnaires. Results. One hundred and twenty-five completed questionnaires were collected. Thirty-one percent of respondents thought tonsillectomy should be offered solely on patient request, 19% after one to three bouts, and 35% after four to six bouts of tonsillitis. Only 9% thought the current guidelines were reasonable. Patients who had suffered recurrent tonsillitis or had undergone previous tonsillectomy generally thought tonsillectomy advisable after more bouts of tonsillitis than those who had not. Fourteen patients fulfilled the SIGN guidelines for tonsillectomy for recurrent tonsillitis. Of these, 13 (93%) felt that suffering 4-6 bouts of tonsillitis was reasonable before tonsillectomy. Conclusion. All patients we surveyed who meet the current PLCV and SIGN guidelines regarding the appropriateness of tonsillectomy for recurrent tonsillitis perceive that they are excessive, believing that 4-6 bouts of recurrent tonsillitis are adequate to justify tonsillectomy.

Classifiability-based omnivariate decision trees.

Top-down induction of decision trees is a simple and powerful method of pattern classification. In a decision tree, each node partitions the available patterns into two or more sets. New nodes are created to handle each of the resulting partitions and the process continues. A node is considered terminal if it satisfies some stopping criteria (for example, purity, i.e., all patterns at the node are from a single class). Decision trees may be univariate, linear multivariate, or nonlinear multivariate depending on whether a single attribute, a linear function of all the attributes, or a nonlinear function of all the attributes is used for the partitioning at each node of the decision tree. Though nonlinear multivariate decision trees are the most powerful, they are more susceptible to the risks of overfitting. In this paper, we propose to perform model selection at each decision node to build omnivariate decision trees. The model selection is done using a novel classifiability measure that captures the possible sources of misclassification with relative ease and is able to accurately reflect the complexity of the subproblem at each node. The proposed approach is fast and does not suffer from as high a computational burden as that incurred by typical model selection algorithms. Empirical results over 26 data sets indicate that our approach is faster and achieves better classification accuracy compared to statistical model select algorithms.

A classification paradigm for distributed vertically partitioned data.

In general, pattern classification algorithms assume that all the features are available during the construction of a classifier and its subsequent use. In many practical situations, data are recorded in different servers that are geographically apart, and each server observes features of local interest. The underlying infrastructure and other logistics (such as access control) in many cases do not permit continual synchronization. Each server thus has a partial view of the data in the sense that feature subsets (not necessarily disjoint) are available at each server. In this article, we present a classification algorithm for this distributed vertically partitioned data. We assume that local classifiers can be constructed based on the local partial views of the data available at each server. These local classifiers can be any one of the many standard classifiers (e.g., neural networks, decision tree, k nearest neighbor). Often these local classifiers are constructed to support decision making at each location, and our focus is not on these individual local classifiers. Rather, our focus is constructing a classifier that can use these local classifiers to achieve an error rate that is as close as possible to that of a classifier having access to the entire feature set. We empirically demonstrate the efficacy of the proposed algorithm and also provide theoretical results quantifying the loss that results as compared to the situation where the entire feature set is available to any single classifier.

Phase transition in a swarm algorithm for self-organized construction.

This paper reports on a system where very simple, noncommunicating mobile agents in a cellular (lattice) environment use purely local rules to construct connected structures from initially randomly distributed building blocks. We study the effect of block density on the final structure, demonstrating a percolationlike phase transition: Low block densities lead to the formation of small, disconnected structures but a single connected structure emerges abruptly beyond a critical density. The empirical study of the structure at the transition point shows scaling behavior, providing strong evidence for criticality. We also demonstrate that a simple change of rules can completely change the phase-transition effect. The results have implications for the self-organized construction of complex structures by swarms.