A Survey : Study of Data Mining and Data Warehousing in Healthcare

Health care is one of the speedy growing areas. The Health care system contains large amount of medical data which should be mined from data warehouse. The mined data from data warehouse helps in finding the important information. Comprehensive amount of data in health care database need the growth of tools which can be used to access the data, analyze and analysis the data, discovery of knowledge, and versed use of the stored knowledge. The health care system has lot of data about the patient’s details, medications etc. In this paper we have studied different data mining and warehousing techniques used in healthcare areas.

Prediction of in-hospital mortality of Clostriodiodes difficile infection using critical care database: a big data-driven, machine learning approach

Research objectivesClostriodiodes difficile infection (CDI) is a major cause of healthcare-associated diarrhoea with high mortality. There is a lack of validated predictors for severe outcomes in CDI. The aim of this study is to derive and validate a clinical prediction tool for CDI in-hospital mortality using a large critical care database.MethodologyThe demographics, clinical parameters, laboratory results and mortality of CDI were extracted from the Medical Information Mart for Intensive Care-III (MIMIC-III) database. We subsequently trained three machine learning models: logistic regression (LR), random forest (RF) and gradient boosting machine (GBM) to predict in-hospital mortality. The individual performances of the models were compared against current severity scores (Clostridiodes difficile Associated Risk of Death Score (CARDS) and ATLAS (Age, Treatment with systemic antibiotics, leukocyte count, Albumin and Serum creatinine as a measure of renal function) by calculating area under receiver operating curve (AUROC). We identified factors associated with higher mortality risk in each model.Summary of resultsFrom 61 532 intensive care unit stays in the MIMIC-III database, there were 1315 CDI cases. The mortality rate for CDI in the study cohort was 18.33%. AUROC was 0.69 (95% CI, 0.60 to 0.76) for LR, 0.71 (95% CI, 0.62 to 0.77) for RF and 0.72 (95% CI, 0.64 to 0.78) for GBM, while previously AUROC was 0.57 (95% CI, 0.51 to 0.65) for CARDS and 0.63 (95% CI, 0.54 to 0.70) for ATLAS. Albumin, lactate and bicarbonate were significant mortality factors for all the models. Free calcium, potassium, white blood cell, urea, platelet and mean blood pressure were present in at least two of the three models.ConclusionOur machine learning derived CDI in-hospital mortality prediction model identified pertinent factors that can assist critical care clinicians in identifying patients at high risk of dying from CDI.

Identification of prodromal presentations of Parkinson’s disease among primary care outpatients in Germany

Background: This study aimed to identify clinical features that predate the diagnosis of PD in a primary care setting. Methods: This retrospective case-control study was based on data from the Disease Analyzer database (IQVIA) and included 17,702 patients with Parkinson’s disease and 17,702 non-PD controls matched for age, sex, and index year. We analyzed the prevalence of 15 defined diagnoses and symptoms documented within 2 years, ≥2 to <5, and ≥5 to <10 years prior to the index date in patients with and without PD. Logistic regression analyses were conducted to assess the association between PD and the predefined diagnoses. Results: The prevalence of motor, neuropsychiatric and autonomic features was higher in those with a later diagnosis of Parkinson’s disease than controls for all three periods except for rigidity in the ≥2 to <5 and ≥5 to <10-year periods and erectile dysfunction in the most recent period before diagnosis. The clinical presentation recorded in the greatest percentage of patients was depression, followed by dizziness, insomnia, and constipation, but these were also common in the control population. The odds ratios were highest for increase in tremor, followed by balance impairment and memory problems, particularly in the latest period before diagnosis, and by constipation particularly in the earliest period examined. Conclusion: The prodromal features of PD could be identified in this large primary care database in Germany with similar results to those found in previous database studies despite differences in methodologies and systems.

Increase in Achilles Tendon Rupture Surgery in Japan: Results From a Nationwide Health Care Database

Background: Nationwide epidemiologic studies in Scandinavian countries have shown that the incidence of Achilles tendon ruptures (ATRs) has increased, and the rate of surgical treatment has declined markedly in the past decade. However, there is a lack of national-level data on the trend of ATRs and surgical procedures in other regions. Purpose: To clarify the trend in the incidence of ATRs and the proportion of surgery using the nationwide health care database in Japan. Study Design: Descriptive epidemiology study. Methods: Age- and sex-stratified data on the annual number of ATRs and surgical procedures between 2010 and 2017 were obtained from the Japanese national health care database, which includes almost all inpatient and outpatient medical claims nationwide. The Japanese population data were also obtained from the population census. The change in the annual incidence of ATRs per 100,000 people was assessed using a Poisson regression analysis. The trend in the annual proportion of surgeries relative to the occurrence of tendon ruptures was determined using a linear regression analysis. Results: A total of 112,601 ATRs, with men accounting for 67%, were identified over 8 years. Patients aged ≥60 years accounted for 27,106 (24%), while those aged 20 to 39 years and 40 to 59 years accounted for 36,164 (32%) and 49,331 (44%), respectively. The annual incidence of ATR ranged from 12.8/100,000 to 13.9/100,000 (women, 8.2-8.9/100,000; men, 17.2-19.5/100,000), which did not change over the study period ( P = .82). Moreover, the annual incidences did not change across sexes and age categories. The annual proportion of surgery increased significantly, from 67% in 2010 to 72% in 2017 ( P = .003). The annual proportions increased across sexes and age categories except for women aged 40 to 59 years. Conclusion: The incidence of ATR did not change between 2010 and 2017, according to the Japanese nationwide health care database. Furthermore, the proportion of surgical treatment increased during the study period. Overall, 70% of patients underwent surgical treatment. This study suggested that the trend in ATR and surgery differed across regions.

Prevalence and Incidence of Osteoarthritis: A Population-Based Retrospective Cohort Study

While trends data of osteoarthritis (OA) are accumulating, primarily from Western Europe and the US, a gap persists in the knowledge of OA epidemiology in Middle Eastern populations. This study aimed to explore the prevalence, incidence, correlations, and temporal trends of OA in Israel during 2013–2018, using a nationally representative primary care database. On 31 December 2018, a total of 180,126 OA patients were identified, representing a point prevalence of 115.3 per 1000 persons (95% CI, 114.8–115.8 per 1000 persons). Geographically, OA prevalence was not uniformly distributed, with the Southern and Northern peripheral districts having a higher prevalence than the rest of the Israeli regions. OA incidence increased over time from 7.36 per 1000 persons (95% CI 6.21–7.50 per 1000 persons) in 2013 to 8.23 per 1000 persons (95% CI 8.09–8.38 per 1000 persons) in 2017 (p-value for trend = 0.02). The incidence was lowest in patients under 60 years (in both sexes) and peaked at 60–70 years. In older ages, the incidence leveled off in men and declined in women. The growing risk of OA warrants a greater attention to timely preventive and therapeutic interventions. Further population-based studies in the Middle East are needed to identify modifiable risk factors for timely preventive and therapeutic interventions.

Trends, Diagnoses, and Hospitalization Costs of Child Abuse and Neglect in the United States of America

We conducted a secondary analysis of the National Inpatient Sample (NIS) to examine child abuse and neglect hospitalization from 1998–2016. The NIS is the largest all-payer, inpatient care database in the United States and is maintained by the Health Care Utilization Project. Participants were youth 18 years and younger with discharged diagnoses of child abuse and neglect from hospitals. The rate of child abuse or neglect hospitalizations did not vary significantly over the study period (1998–2016), which on average was 6.9 per 100,000 children annually. Males (53.0%), infants (age < 1; 47.3%), and young children (age 1–3; 24.2%) comprised most of the child maltreatment cases. Physical abuse was the most frequent type of maltreatment leading to hospitalization. Government insurance was the most common payer source, accounting for 77.3% of all child maltreatment hospitalizations and costing 1.4 billion dollars from 2001–2016. Hospitalizations due to child abuse and neglect remain steady and are costly, averaging over $116 million per year. The burden on government sources suggests a high potential for return on investment in effective child abuse prevention strategies.

Disease burden of diabetes, diabetic retinopathy and their future projections in the UK: cross-sectional analyses of a primary care database

ObjectivesTo estimate the current disease burden, trends and future projections for diabetes mellitus (DM) and diabetic retinopathy (DR) in the IQVIA Medical Research Data (IMRD).Participants/design/settingWe performed a cross-sectional study of patients aged 12 and above to determine the prevalence of DM and DR from the IMRD database (primary care database) in January 2017, involving a total population of 1 80 824 patients with DM. We also carried out a series of cross-sectional studies to investigate prevalence trends, and then applied a double exponential smoothing model to forecast the future burden of DM and DR in the UK.ResultsThe crude DM prevalence in 2017 was 5.2%. The DR, sight-threatening retinopathy (STR) and diabetic maculopathy prevalence figures in 2017 were 33.78%, 12.28% and 7.86%, respectively, in our IMRD cross-sectional study. There were upward trends in the prevalence of DM, DR and STR, most marked and accelerating in STR in type 1 DM but slowing in type 2 DM, and in the overall prevalence of DR.ConclusionOur results suggest differential rising trends in the prevalence of DM and DR. Preventive strategies, as well as treatment services planning, can be based on these projected prevalence estimates. Improvements that are necessary for the optimisation of care pathways, and preparations to meet demand and capacity challenges, can also be based on this information. The limitations of the study can be overcome by a future collaborative study linking DR screening and hospital eye services data.