Objective The objective of this study was to use machine learning and health standards to address the problem of clinical data interoperability across healthcare institutions. Addressing this problem has the potential to make clinical data comparable, searchable and exchangeable between healthcare providers. Data sources Structured and unstructured data has been used to conduct the experiments in this study. The data was collected from two disparate data sources namely MIMIC-III and NHanes. The MIMIC-III database stored data from two electronic health record systems which are CareVue and MetaVision. The data stored in these systems was not recorded with the same standards; therefore, it was not comparable because some values were conflicting, while one system would store an abbreviation of a clinical concept, the other would store the full concept name and some of the attributes contained missing information. These few issues that have been identified make this form of data a good candidate for this study. From the identified data sources, laboratory, physical examination, vital signs, and behavioural data were used for this study. Methods This research employed a CRISP-DM framework as a guideline for all the stages of data mining. Two sets of classification experiments were conducted, one for the classification of structured data, and the other for unstructured data. For the first experiment, Edit distance, TFIDF and JaroWinkler were used to calculate the similarity weights between two datasets, one coded with the LOINC terminology standard and another not coded. Similar sets of data were classified as matches while dissimilar sets were classified as non-matching. Then soundex indexing method was used to reduce the number of potential comparisons. Thereafter, three classification algorithms were trained and tested, and the performance of each was evaluated through the ROC curve. Alternatively the second experiment was aimed at extracting patient’s smoking status information from a clinical corpus. A sequence-oriented classification algorithm called CRF was used for learning related concepts from the given clinical corpus. Hence, word embedding, random indexing, and word shape features were used for understanding the meaning in the corpus. Results Having optimized all the model’s parameters through the v-fold cross validation on a sampled training set of structured data ( ), out of 24 features, only ( 8) were selected for a classification task. RapidMiner was used to train and test all the classification algorithms. On the final run of classification process, the last contenders were SVM and the decision tree classifier. SVM yielded an accuracy of 92.5% when the and parameters were set to and . These results were obtained after more relevant features were identified, having observed that the classifiers were biased on the initial data. On the other side, unstructured data was annotated via the UIMA Ruta scripting language, then trained through the CRFSuite which comes with the CLAMP toolkit. The CRF classifier obtained an F-measure of 94.8% for “nonsmoker” class, 83.0% for “currentsmoker”, and 65.7% for “pastsmoker”. It was observed that as more relevant data was added, the performance of the classifier improved. The results show that there is a need for the use of FHIR resources for exchanging clinical data between healthcare institutions. FHIR is free, it uses: profiles to extend coding standards; RESTFul API to exchange messages; and JSON, XML and turtle for representing messages. Data could be stored as JSON format on a NoSQL database such as CouchDB, which makes it available for further post extraction exploration. Conclusion This study has provided a method for learning a clinical coding standard by a computer algorithm, then applying that learned standard to unstandardized data so that unstandardized data could be easily exchangeable, comparable and searchable and ultimately achieve data interoperability. Even though this study was applied on a limited scale, in future, the study would explore the standardization of patient’s long-lived data from multiple sources using the SHARPn open-sourced tools and data scaling platforms / Information Science / M. Sc. (Computing)

The accomplishment of pilot implementation of the HIVQUAL-T model, an innovative HIV care quality management tool, has led to an authoritative decision to scale up the use of the model nationwide in Thailand. However, the level of implementing this model varies across target hospitals. Some hospitals have fully adopted the model by conducting quality improvement (QI) activities following performance measurement (PM) results while others have partially adopted only PM or have not used this model at all. The differential level of implementation could be a contributing factor accounting for discrepancies in the quality of care across different HIV ambulatory care facilities. A cross-sectional study was conducted by using two main datasets, including Thailand's national HIV care performance results and an online survey of all public hospitals nationwide. A total of 382 hospitals responded to the survey, accounting for a response rate of 50%. A confirmatory factor analysis (CFA) method was performed to examine the validity of latent constructs developed from the diffusion of innovation theory. Structural equation modeling (SEM) approach was employed to investigate the relationship between the determinants of organizational decision-making and their contribution to organizational outcomes, under the context-design-performance framework. Furthermore, using a panel model of hospitals that reported performance results across a 3-year period, the improvement in HIV ambulatory services performance among the adopters was examined. The results indicated that two innovation attributes--relative advantage and simplicity perceived by HIV care practitioners in hospitals--were found to be positively associated with the level of the HIVQUAL-T model implementation. Two structural characteristics--interconnectedness and organizational slack--appeared to be positively associated with the level of model implementation, while rate of adoption in the region also had significant positive contribution. Ultimately, the extensiveness of the HIVQUAL-T model implementation demonstrated a proportionate impact on the variation in hospitals' HIV ambulatory services performance. It was noted that the implementers considerably improved their performance within two years of implementing the model. The study findings imply that adoption is more likely when individual practitioners assess the innovation and find it to be easy to comprehend and operate and also worthwhile to implement. Furthermore, hospitals' decision making is likely influenced by their relations to external environment. The findings suggest more emphasis on individual and hospital-level capacity building for meaningful use of this quality management initiative, accompanied by an adjustment of performance measurement software with valid, reliable, and interpretable indicators.

Needs assessment--Cost analysis--Modeling--Pilot studies--Sample programs--Training programs--Communication.

Needs assessment--Cost analysis--Modeling--Pilot studies--Sample programs--Training programs--Communication.

published_or_final_version / Community Medicine / Master / Master of Public Health

A Markov model was used to determine the optimal number of Medical Service Corps Health Care Administrator officers the Navy must access each year in order to maintain a desired end-strength. The Markov model identified the flow of Ensigns through Lieutenant Commanders using sixteen-year data. Five scenarios were analyzed to determine the most qualified method in determining accession levels. Optimization was achieved by changing with distribution of accessions sources and recruiting ranks. The solver scenario provided the alternative with the least amount of underage and overage when comparing the force structure to the predicted values. A four year historical review identified that if the current business practices will not allow for rank steady states to be reached. A few significant characteristics were determined to influence retention at seven and ten year periods. The characteristic of primary concern, commissioning source, was determined to be significant. A survival analysis identified that the In-Service Procurement Program has a different survival function than other sources. Increased variations between the current force structure plan and the predicted Markov model outcomes suggest that greater efficiency could be obtained in future years. This Markov model can be used as a tool for accessioning to improving extended forecasts.

No description available.

Typescript (photocopy). / Digitized by Kansas Correctional Industries

A study to determine the effects of the new prospective reimbursement method on the role of discharge planners in the acute care hospitals and their impressions of its impact on elderly patients. A mail survey questionnaire examined the task responsibilities of personnel providing discharge planning services to the elderly. The components of successful discharge plans and the frequency of obstacles to optimal discharge plans were also assessed. Findings, based on a 75.3 percent response rate from hospitals (hospital N = 58, individual respondent N = 235), revealed that social work appeared to be the predominant discipline conducting discharge planning activities in New York City hospitals. The prospective payment system has changed discharge planning task frequency and the time allocation of tasks, perceptions of inpatient populations, the need for case management and readmission rates. Family support was the strongest predictor for successful discharge plans cited by discharge planners while lack of interdisciplinary communication was a prime reason why discharge plans go awry. Discharge planners do not generally view their role as responsible for advising patients of the appeals mechanism. Professional discipline and the prospective payment system were the single most important independent variables to explain the variance among the areas of inquiry. Hospital auspice and bed size, and respondent's years of experience and employment tenure at their hospital were not very useful to explain differences. The comparison of nurses and social workers on performing parallel discharge planning functions showed that nurses tend to be more collaborative than social workers. One-third of the nurse and social worker respondents reported that their professional preparation was inadequate for discharge planning responsibilities and have recommendations for the curriculum of their respective professional schools.

This dissertation reports the findings of an evaluation of the Five Hospital Homebound Elderly Program in Chicago. The Program is a model long term home care demonstration project which is community-based and provides a unique combination of medical and social supportive services to chronically impaired homebound elderly. The Program differs from most Medicare certified and reimbursed programs and is a potential model for expanded benefit programs now being considered in Congress. The evaluation of the Program encompassed both impact and process issues, utilizing a pre/post test quasi-experimental design with a non-equivalent control group. Within this design, the comprehensive functional status of 122 consecutively admitted Five Hospital clients and of 123 controls was assessed at time of acceptance to service and nine months later, using the Duke/OARS Multidimensional Functional Assessment Questionnaire. The control group consisted of recipients of OAA Title III home-delivered meals who live in an area that is geographically contiguous to the Five Hospital Program's catchment area and who were similarly elderly and impaired. Rates of hospitalization, institutionalization and mortality were also obtained in addition to rates of home care services utilization for both groups. Analysis of pretest data, which was obtained on 75% of newly accepted FHHEP clients and 74% of control group clients, revealed that, overall, the two groups were remarkably similar on measured variables with the exception of. the fact that Five Hospital clients were significantly older (x̄ age 80.4) and more impaired in Instrumental Activities of Daily Living than controls but also enjoyed better social supports. These differences were of interest because advanced age, poor ADL and poor social supports are variables which have been reported to be associated with admission to long term care institutions in the literature. At post test, functional status data were obtained on 83% of experimentals and 81% of controls, with mortality being the major cause of attrition in both groups (12% and 11%, respectively). Major findings derived from the analysis of outcome measures revealed no difference between the two groups with respect to mortality rates. However, bivariate analysis revealed interesting differences between the two groups with respect to the characteristics of subjects who died. Variables which appeared to be similarly and differentially associated with increased incidence of mortality in the two groups were therefore identified and discussed. However, the number of subjects who experienced this outcome in both groups was too small to permit any conclusions to be drawn. The fact that mortality rates were almost identical for the two groups, given the greater age and ADL impairment of experimental subjects at pretest, is interpreted as suggestive of a possible beneficial effect of treatment which might warrant follow-up in a study with a larger sample size. Effects of treatment on functional status included a trend towards perception of improved social supports. mental health and physical health, as well as a significant decrease in unmet needs among experimental subjects (p < .05). Somewhat paradoxically, these beneficial changes were accompanied by a significant (p < .01) deterioration in Physical Activities of Daily Living. It is suggested that this change in PADL status may be confounded with the reception of the treatment which consisted of regular assistance with ADL tasks. Comparison of pre and post test hospitalization rates revealed no differences between the two groups, with the exception of the fact that fewer Five Hospital subjects were repeat users at pre and post test. A highly significant difference was observed in the rates of institutionalization for the two groups with experimental subjects experiencing this outcome less frequently than controls (p < .01). Multivariate analysis which attempted to control for selection differences and other explanatory variables failed to degrade this effect which remained significant. Furthermore. when the above outcome measures were analyzed according to level of treatment received, the same relationships described above were observed. The above findings were interpreted to indicate that, overall, the Program had beneficial effects on its clientele and that this model of long term care services has important implications for the development of national long term care policy.