Thank you for your time today. The purpose of this presentation is to discuss overall healthcare issues and specific disease management in today’s complex world and how predictive analytics solutions can reduce healthcare costs.
There are several challenges facing organizations that deal with healthcare from a macroeconomic standpoint.
So how can we move from patients being in hospitals and traditionally only having a 1-to-1 relationship with the doctor…..
To patients taking ownership of their lives and organizations leveraging the entire community to improve patient care and satisfaction. Doctors are not the only source of information to try to improve care and we should be leveraging that. An important component is the patient itself – being able to provide information and also be empowered to live their lives with the optimal treatment for them, as an individual.
Additionally, not every drug treatment works for everyone. Determining a treatment plan is still very much a balancing act between art and science and in most cases is very much trial-n-error.
Given the situation that we are in, organizations need to stop doing things the “same old ways” and take the opportunity to utilize technology to make better decisions and take action. Think about the current processes in your organization and ask yourself…We can now start to examine how predictive analytics can provide value…
SPSS Solutions for threat and risk control are designed to meet you where you are – and provide value in two ways. For many, providing insight for decision makers is the top priority and advanced analytics clears the picture of what is happening and why. For others, moving beyond insight to actually identifying the next best action in a mission critical process is the key to breaking away. We can help you craft a roadmap that adds value where and how you need it.
Our most successful clients approach their use of predictive analytics as something which impacts the entire organization. We provide value to senior management by giving them visibility into what will happen in their operations via key performance predictors, usually displayed in interactive dashboards, reports, and alerts. We help line managers and policy makers define how operations will work through highly accurate forecasts and business focused optimization. Decision makers see what will happen in the operations they manage.At the operational level, we help individual contributors take the next best action. It’s here where the Smarter Planet really shows its mettle. Predictive Analytics makes it’s biggest impact when it is deployed to help run the business.
So what’s different about predictive analytics, compared to other forms of analysis?CLICKIn traditional business intelligence and other types of conventional analysis, we’ re working with a snapshot of the present and the historical data leading up to that. Valuable insights and metrics – such as KPIs - can be obtained, but it’s fundamentally a “rearview mirror” approach. And if you see something of interest and want to explore what’s behind it, it’s up to you to take the initiative and drill down into the data.CLICKPredictive analytics works from the same historical data, CLICKbut here the algorithms themselves explore the data in he context of a business issue – such as how to tell which customer will stay loyal or defect, or which transactions are safe to accept – and automatically discover the relevant patterns and relationships. CLICKThese provide deep and valuable insights which can drive better decision making, CLICKbut the algorithms also “learn” from the historical data, building “predictive models” . CLICKThese models are applied to current or new cases. They use current data as well as any relevant historical information, CLICKand make robust, accurate predictions. These predictions can be delivered to drive better decisions and better outcomes in key business processes.CLICKThis is how we achieve he transformation in business decision making we described earlier: from “Sense & Respond” to “Predict & Act”.
During the Rare Disease Diagnosis project, scientific knowledge and technology became intertwined as never before and showed how ICT can support smarter diagnosis and in that way improve healthcare quality and cost-effectiveness.The rare disease diagnosis platform allows for an earlier, quicker and more accurate diagnosis by integrating both medical expertise and data mining tools. Rules are generated much faster and more accurately through a predictive model based on known patient data. In comparison with a pure rule based system, a combination with data mining tools provides both higher sensitivity and more specificity.The solution can serve as an intelligent and dynamic knowledgebase on rare diseases. The improvement in the quick diagnosis and treatment of rare diseases can mean the difference in the lives of patients.
Company name: Ciudad Real HospitalInstrumentedThe software is used to apply advanced, multivariate statistical processing to data contained in historical patient records. InterconnectedIntegrated with existing systems and software at the hospital, the SPSS Statistics application connects patient evaluation data with treatment outcomes information.IntelligentThe predictive analytics solution based on SPSS Statistics software helps identify which treatment aspects could influence patients’ clinical outcomes and yields new information not available through use of a less sophisticated statistical toolset. For the eating disorders unit at the hospital, applying the SPSS Statistics application provided knowledge pertaining to the relevance of a patient’s expectations at the beginning of treatment. This knowledge has, in turn, allowed more accurate initial patient evaluations and easier identification of patients for whom initial interventions should lead to favorable outcomes.
ClientHospital Santa Bárbara Business NeedHospitalSanta Bárbara, a research and treatment hospital in Spain, wanted to be able to analyze data from its research unit in combination with administrative patient data to more effectively isolate patient risk factors, perform diagnoses, administer treatment and improve treatment methods in the future. Solution ImplementationThe hospital implemented IBM SPSS Statistics software, enabling it to aggregate hospital and research data into one centralized database where employees can easily extract data and run several types of in-depth analysis to gain new insight on diseases and effective treatments. BenefitsSPSS Statistics software enabled the hospital to create new diagnostic models, isolate risk factors and improve treatment for diseases such as deep vein thrombosis, colon cancer and chronic obstructive pulmonary disease. The hospital expects that these new tools and insights will help it reduce mortality rates and improve patient outcomes for years to come. InstrumentedData from the research unit of the hospital and from patient records in the administrative database are gathered into one centralized database for easy extraction and analysis. InterconnectedBoth the administrative and research departments use statistics software to manage all of the information in a centralized database. Data from both sources informs more reliable diagnostic models and new methods of treatment.IntelligentBy conducting advanced statistical analyses and tests in near-real time, employees have been able to gain insights into patient risk factors and serious illnesses such as pulmonary disease and cancer that weren’t available with their previous methods of data gathering and analysis. Researchers and hospital staff can now glean this information and apply it to patient care in a fraction of the time that it would have taken with manual data extraction and analysis. Insights gained from statistical analysis have enabled staff to predict probability of illness, determine accurate prognoses and prescribe the best preventive measures and treatments for disease patients, with the expectation of saving the hospital money and saving countless lives.
Reference URL: http://w3-01.ibm.com/sales/references/crdb/ibmref.nsf/allbydocid/0CRDD-8EZ377?opendocumentClientA pain management clinicBusiness needThis specialized medical facility in Australia is devoted to high-standard and medical interventional management of acute and chronic pain. Its practitioners work to relieve musculoskeletal, neuropathic and other complex pain, including headaches, back and neck pain, joint pain and pelvic pain. To manage patient information, the clinic used a standard paper-based patient information system commonly used by medical practitioners. The process worked efficiently, but it worked slowly. The clinic wanted a new, more robust information management solution that would give it a depth of information and insight that it never had before, and a competitive advantage over other clinics. Solution implementationTo address its needs, the clinic implemented IBM SPSS® Data Collection Web Interviews software, which it uses to collect important patient information online. Online data collection enabled the clinic to acquire a deeper level of information. For example, when a patient fills out a paper survey, she may be asked to rate pain on a scale of one to ten. The new solution features an intuitive GUI that enables patients to indicate the areas where they are experiencing pain and rate the pain levels through a series of custom-built Adobe Flash objects. To facilitate the online survey process, the clinic installed Internet hotspots in waiting areas for patients without Internet access at home. Once the data is captured, the clinic uses IBM SPSS Statistics software to analyze the data and build reports for future use. The data from questionnaires, when put through statistical analysis, can help doctors more effectively diagnose problems and recommend remedies. BenefitsBy implementing IBM SPSS Data Collection Web Interviews and IBM SPSS Statistics software, the clinic created a robust new patient information system that has helped it improve the quality of patient care it can deliver. As a result, patient satisfaction levels have increased. Further, because patient information is now collected online, the clinic reduced administration costs, including data entry, emailing patients and maintaining follow-ups, by 75 percent. What’s more, the clinic seamlessly integrated the solution into its existing infrastructure, helping to reduce the risk of faults and keep implementation costs down. With up-to-date patient information and real-time results of validated patient diagnostic and outcome indexes, the quality of patients’ care has not only improved but become consistent among all treating physicians at the clinic. Measuring patient outcomes gives future patients more realistic expectations of the treatments, enabling the clinic to continually improve upon patient treatment options and care. Instrumented - Rather than relying on manual data entry and paper surveys, the clinic uses online data collection to gather patient information. The new solution features an intuitive GUI that enables patients to indicate the areas where they are experiencing pain and rate the pain levels through a series of custom-built Adobe Flash objects. This data is then saved to an Extensible Markup Language (XML) file, which is captured by the online data collection software.Interconnected - Once the data is collected, it flows directly into statistics software for further analysis. This process is seamless and saves time because all the variables flow directly from the data collection software. Doctors can instantly access data online and easily transfer it to peer reviews, reports or articles for publication.Intelligent - Statistics software helps physicians find correlation between a new patient's symptoms and their previous medical history, other historical patient data, and external articles and information on pain management. This provides the doctor with a strong statistical prediction of what issues the patient may be encountering. In effect, it also allows the doctor to spend more time exploring treatment options with patients and less time hunting for information.
Client name: Intelligent: The solution enables decision-tree and regression analyses to create predictive models based on approximately 2,500 patients who had liver diseases and to define an additional 400 factors that affect treatment options. It allows doctors to objectively compare new patients’ profiles and determine which treatment options are most effective on a percentage basis. And the solution provides never-before-available insight into the practice of liver disease treatment, targeting care specifically to individual patients’ needs to yield better results. Instrumented: The solution captures highly detailed patient records, providing a database of patient profiles from which to mine data. Interconnected: The solution aggregates patient data cross-functionally across various departments of the hospital with 400 attributes per patient, allowing advanced analysis and other medical uses from one central source.
Reference URL http://w3-01.ibm.com/sales/references/crdb/ibmref.nsf/allbydocid/0CRDD-8LPQXX?opendocumentClientA cardiac medical research organizationBusiness needWith a myocardial infarction, also known as a heart attack, every minute counts. Studies have shown that the critical treatment time for a heart attack is within one to two hours of the first onset of symptoms. The longer the delay, the more damage is done. However, it can be tricky to catch that window of time when it is open. Many heart attacks are “silent” attacks, with symptoms presenting as pain in other parts of the body, tingling or shortness of breath. Delayed 911 calls, long ambulance rides and inconclusive diagnoses can mean that by the time a patient gets to the hospital door, time is already running out. One research company in the Netherlands wanted to develop an intelligent tool to help paramedics more effectively diagnose heart attacks and begin prehospital treatment. But it needed a statistical analysis tool robust enough to handle thousands of patient, doctor and hospital surveys as well as a wealth of other data. The tool also had to be sophisticated enough to perform complex analysis on subjective data, applying what-if scenarios and other predictive capabilities.Solution The research team captured all of this data in IBM SPSS® Statistics Base software, where it could analyze it with a variety of techniques to gain more insight into factors such as symptom patterns and average speed of treatment. IBM SPSS Advanced Statistics software enabled the team to verify and refine information with more sophisticated generalized linear modeling procedures, allowing them to analyze various medical possibilities and outcomes. IBM SPSS Regression software enabled researchers to apply regression analysis to the data to study the relationship of dependent variable changes to independent ones and more accurately predict and forecast results of treatment within a specific timeframe. IBM SPSS Custom Tables software gave researchers a fast way to organize and view their results in table and report form, making it easier for outside parties to comprehend the procedures and results of testing and survey analysis. With the IBM solution, the company developed an algorithm that paramedics can use to determine the probability of myocardial infarction while the patient is still in the ambulance. If the probability of heart attack is high, the process then directs the ambulance to the nearest hospital equipped to deal with the condition and prompts paramedics to treat the patient en route using thrombolysis or other procedures. BenefitsThe solution was recently implemented at the leading cardiac hospital in the Netherlands, and hospital staff members expect it to produce positive results for the hospital and its cardiac patients and, eventually, doctors and patients around the world. By diagnosing and treating heart attacks en route to the hospital, paramedics will help reduce the amount of damage myocardial infarction inflicts on the heart. This will result in a more successful short-term treatment and less long-term damage, improving patient survival rates and recovery times. The new prehospital triage procedures are also expected to reduce the average length of stay at the hospital, saving both hospital operating costs and patient expenses. The research company is extremely satisfied with the results it achieves with SPSS software and its relative ease of use compared with other statistical software platforms. It is comprehensive enough to help staff perform sophisticated analyses on complex data, yet easy enough that nonexpert medical staff can use it with minimal need for training.Instrumented - The solution incorporates data from thousands of physician, hospital and patient surveys, as well as records of procedures and treatment results, and analyzes them using a mixture of sophisticated statistical methods. Interconnected - The new system organizes and reports results in clear, easy-to-read tables that can be quickly distributed and easily read by expert researchers and nonexpert medical staff alike. Intelligent - The solution enables researchers to use regression analysis to predict any number of what-if scenarios to consider in determining a methodology. Researchers at the organization were able to not only determine a solid, consistent methodology for prehospital diagnosis and treatment, but were also able to incorporate an algorithm that would direct ambulance drivers to the nearest hospital offering full cardiac care services.
Business needThis leading Italian cancer institute is recognized as a scientific research and treatment institution in the field of pre-clinical and clinical oncology. The Institute’s special status as a research center enables it to transfer research results directly to clinical care. The institute’s vision is to allow and to accelerate information sharing for better disease management (e.g. epidemiology & statistics), for enabling clinical decision support solutions (e.g. definition of guidelines) and for providing advanced services to physician services (second opinion, teleconsultation, etc.).There is an urgent need to better manage pharmaceutical cancer treatments in order to maximize benefits and minimize risks, and to reduce unnecessary treatments. Scientific literature reports that 70% of cancer patients will survive even if they do not receive adjuvant treatment (i.e. supplemental treatment after surgery). In addition, post-operative treatment is given to about 80% of the patients, but about 60% of the patients are receiving unneeded treatments. Meanwhile, receiving the proper adjuvant therapy has proven very effective, providing a 30% reduction in the risk of distant metastases. The institute’s objective was to monitor the oncology care process (in particular for rare forms of cancer), assessing the level of clinical guideline adherence, clinical outcomes and resource/benefits ratios. Specifically, the institute wanted to identify deviations from clinical guidelines and assess clinical outcomes related to these deviations, eventually updating guidelines. The goal of such analysis was to improve patient care by tailoring treatment approaches for specific individuals. the institute needed a new decision support solution to personalize treatment based on automated interpretation of pathology guidelines and intelligence from a number of past clinical cases, documented in the hospital information system. Moreover it wanted to implement standard definitions and procedures for retrospectively assessing the quality and cost of delivered care.One of key factors for adopting smart control and management systems by the institute is the consideration that chemotherapy treatment causes suffering to patients and may even lead to irreparable damage, not to mention the waste of money due to these treatments.Solution implementationIBM Research scientists from Haifa, Israel, are collaborating with the institute on a unique new biomedical analytics platform which will be tested by the Institute’s physicians. The institute’s Clinical Genomics (Cli-G) solution integrates and analyzes all available clinical data, and correlates it with available patient data to create a specific course of treatment for each patient. Selecting the most effective treatment can depend on a vast number of characteristics including genetic profile, age, weight, family history, current state of the disease, or general health. Extracting evidence-based knowledge and patient’s data, Cli-G provides vital clinical decision support on how to treat different rare types of cancer for different genetic and clinical profiles. In addition, an aggregated view of patient care provided by the solution enables administrators to evaluate performance, streamline processes for maximum safety, and reduce costs, even as they provide more effective disease treatment. The institute will be able to retrospectively monitor care appropriateness by comparing performed procedures to clinical guidelines (guideline adherence, refinements of guidelines) and use this knowledge to streamline processes for maximum safety. For example, hospital administrators can drill down into the data to better understand what the guidelines were for recommendations, what succeeded, and whether treatment quality has improved.The IBM Cli-G solution represents the state-of-art application of analytics for healthcare through the integration of advanced statistical techniques with easy-to-use human-machine interface. Thanks to this, IBM delivers the best evidence-based medical insights at the patients’ bedside. Any patient data securely collected from hospitals and health organizations is ‘de-identified’ or made anonymous through the removal of personal identifying details. The IBM system does not have to know which individuals the information came from in order to draw conclusions. It works by identifying similar cases based on age, sex, symptoms, diagnosis, or other related factors. The Cli-G has been developed by the IBM Haifa Research lab with the support of IBM Italy Global Business Services. Based on the valuable background of structured information, IBM and the institute decided to work together to develop a clinical decision support system (CDSS), taking into account a specific pathology that leverages the Lombardy Oncology Network (ROL) information backbone.Project activities were focused in the deep understanding on how information technology and data analysis can better support clinicians in improving quality of care and decisions. In this context, the IBM Research is developing the innovative features of the (CDSS) together with ROL clients.As a first step the Cli-G solution focuses on a specific pathology: Sarcoma. Using state-of-the-art analytics for clinical evidence, and their derivative clinical guidelines, it analyzes vast amounts of clinical and genomic data within the context of the individual patient's profile. The analysis of a vast amount of patients' data through deep analytics and machine learning is done in the context of clinical trials, epidemiological studies and similar research activities. The analysis of patient data gathered at point of care and the enhancement of patient treatment represent the key factors. The innovative Cli-G project has been awarded with a FOAK Grant (first of a kind – meaning the first and innovative project where IBM covers this kind of topics and issues) of $1 Million that will allow the development of the solution for other cancer treatments.Benefit of the solution- Avoids unnecessary treatment (estimated to be up to 60% of all treatment) and delays in treatment delivery- Creates tailored and personalized treatments, increasing the chances of successful outcomes- Improves hospital performance, both clinical and operational by providing a “big picture” view of treatment delivery, helping streamline processes and lower costs- Optimizes patient routing among primary and secondary hospitals- Helps develop new business models for healthcare industry based on evidence-based medicine.The Cli-G solution is an example of how IBM is contributing to the transformation of the healthcare industry by creating analytics technologies that will simplify the complex interactions in the healthcare delivery process while optimizing outcomes, ultimately helping to make patient care smarter and lower costs.Instrumented:The Cli-G solution acquires data from several medical applications and makes it available within an integrated ecosystem. The solution is highly instrumented thanks to the adoption of IBM software, including:- WebSphere ILOG jRules- Content Analytics (Text mining analytics and entities recognizer)- Cognos Real Time Monitoring (Real time analytics)- IBM SPSS (Statistical Engine) Interconnected: The Cli-G solution extracts deep medical insights integrating several sources of information (clinical data, genomic profile, texts and doctors’ notes) transformed into a unique medical knowledge platform. Cli-G integrates data from discharge letters and additional supportive data into the developed disease model for generating a disease data mart. A flexible and adaptable data integration layer is provided for collecting patients' records, transforming them according to the longitudinal model definitions and generating the disease's data mart. The core component of this layer is an ETL (Extract, Transform, Load) tool. ETL has inherent data integration and transformation capabilities. The ETL acts as a 'slave' component within the proposed solution. Integration, transformation, cleansing and enrichment directives are provided through the disease model. Data Processing is executed on the integrated disease-specific data mart for generating the core clinical intelligence and decision support building blocks.The purpose of the data processing layer is to supply clinical decision makers with a holistic view of all aspects associated with the care process.Intelligent:This solution is just one example of how IBM researchers are helping transform the healthcare industry by creating analytics technologies that will simplify the complex interactions in the healthcare delivery process while optimizing outcomes, ultimately helping to make patient care smarter and lower costs. In addition to cancer and AIDS, this solution can be adapted to support evidence-based treatment for any disease with complex genetic variants and a variety of treatment options. The solution aims to improve the care process by adapting the available treatment to a specific individual, avoiding delays of treatment delivery, and improving outcomes for various diseases. IBM Clinical Genomics uses some of the same natural language processing and machine learning capabilities as IBM Watson, but the clinical genomics platform has unique capabilities that are complimentary to the deep question and answering abilities of IBM Watson for healthcare.The Institute’s Clinical Genomics (Cli-G) solution integrates and analyzes all available clinical data, and correlates it with available patient data to create a specific course of treatment for each patient. Selecting the most effective treatment can depend on a vast number of characteristics including genetic profile, age, weight, family history, current state of the disease, or general health. Extracting evidence-based knowledge and patient’s data, Cli-G provides vital clinical decision support on how to treat different rare types of cancer for different genetic and clinical profiles. IBM solution integrates and analyzes all available clinical knowledge and guidelines, and correlates them with available patient data to create evidence that supports a specific course of treatment for each patient. Developed by researchers in Haifa, Israel, with GBS Italy support, the new research prototype works by investigating the patient’s personal makeup and disease profile, and combines this with insight from the analysis of past cases and clinical guidelines. The solution provides physicians and administrators with a better picture of the patient-care process and reduces costs by helping clinicians choose more effective treatment options.In addition, the institute will be able to retrospectively monitor care appropriateness by comparing performed procedures to clinical guidelines (guideline adherence, refinements of guidelines) and use this knowledge to streamline processes for maximum safety. For example, hospital administrators can drill down into the data to better understand what the guidelines were for recommendations, what succeeded, and whether treatment quality has improved.The Cli-G solution integrates all the available medical knowledge and guidelines in the context of a specific disease (e.g., Sarcoma). It uses semantic maps to ease the formulation of clinical knowledge into a self descriptive machine-processable knowledge model. Clinical rules are anchored into the model and support the computation of deductive fields. Secondly, Cli-G provides a semantic interaction point and a collaboration tool among various users: the interaction is provided through a set of high level tools (i.e., concept maps, rule editor) with different perspectives and concepts for different users. The front-end differences in concepts are integrated, at the back-end, into a coherent and aligned set of artifacts (e.g., transformations, features, statistical outcomes), that are being used for generating the system's outcome and results.Data Processing is executed on the integrated disease-specific data mart for generating the core clinical intelligence and decision support building blocks. The purpose of the data processing layer is to supply clinical decision makers with a holistic view of all aspects associated with the care process.
ClientA government healthcare organization in EuropeBusiness needDiagnosis-related group (DRGs) have become a common method in many countries for classifying, coding and pricing healthcare products and services. In this central European country, the implementation of a DRG system for acute patient care funding has been in effect since 1995. DRG codes became an official part of the country’s hospital fee calculations in 2007.As the organization responsible for calculating and recommending healthcare payment prices, this government organization must maintain unparalleled knowledge of clinical treatment paths and trends in the healthcare sector. The organization collects and analyzes enormous amounts of hospital data, categorizing patient hospitalization cases with approximately 1,000 DRG codes according to diagnosis and cost. An essential part of administering the DRG codes is the calculation of relative weights—indices that describe the average cost of providing each treatment. These numbers are critical in determining a fair price for insurance companies to pay hospitals. However, the organization was managing them manually in spreadsheets, where it was difficult to follow, manage and modify the algorithm. The process sometimes lasted several days, and the organization had to repeat it for every new data source. Moreover, the organization was using a borrowed data model, one that yielded good results in other countries but was not tailored to the country’s unique healthcare system. The organization needed powerful statistical analysis capabilities to devise a model specifically for the country’s own healthcare sector.Solution implementationTo calculate relative weights and define hospitalization fees under the DRG system, the organization adopted IBM® SPSS® Modeler Desktop V14.2 software. With powerful data analysis capabilities, the solution enables the organization to test different algorithms and select the one best able to model the country’s healthcare pricing. Useful algorithms can be captured in the form of a data stream and reused—one of the features of the SPSS software’s visual programming. The calculation process itself has been automated, which eliminates hours of manual spreadsheet work. This significant time savings enables the organization to process data monthly and quarterly rather than just once per year. The frequent analysis results help the organization forecast target numbers, which can be particularly useful in strategic planning and decision making.When calculating relative weights, the organization can account for a wide range of factors that affect costs across hospitals and regions, and even compare the country’s success rate in DRG implementation with that of other countries. The solution also allows the organization to modify and fine-tune its calculations as new hospital data comes in—a critical, ongoing cultivation process that keeps DRG codes accurate and relevant. If analysis shows that costs are not homogeneous in any given DRG group, the organization must rethink the classifications and make recommendations on whether to split or merge codes. The SPSS solution contains scripts and macros that visualize the distribution of costs and duration of hospital stay, helping to identify outlying cases that may necessitate changes to the classification system. The visual programming also makes it easier to verify calculations and make corrections quickly.Hospitals and insurance companies also benefit from the sophisticated analysis of hospitalization data, which enables them to track trends in healthcare costs and treatment paths. For example, the numbers may show that patients being treated for heart disease tend to stay longer and cost more than patients with other acute conditions. The pattern is a red flag for hospitals, allowing them to dig more deeply into the causes and find ways to operate more efficiently, with the aim of improving patient care and increasing profitability. In fact, the organization plans to expand the solution by distributing preprogrammed data models to individual players in the healthcare sector, offering a single source of truth to everyone involved.BenefitsThe SPSS Modeler Desktop solution transformed the organization’s approach to DRG classification, creating a system of codes that matches the healthcare system more accurately. Among the most important benefits is the speed with which the organization can collect and analyze hospitalization data. Instead of taking three days, the process now takes just five minutes from beginning to end. In addition, whereas the analysis used to require the time of three medical experts, it now requires only half the time of one expert. This significant time savings enables the organization to shift its attention from manual data preparation and processing to in-depth analysis, spending 10 percent of its time on the former and 90 percent on the latter. As a result, the organization can focus on more scenario planning and cultivation of the DRG system. And it can run analyses every month rather than once yearly, creating a more complete picture of emerging healthcare trends.Instrumented - The solution automatically gathers accounting and acute patient hospitalization data from all hospitals and insurance companies in the Czech healthcare system, creating a single source of information for DRG codes and costs.Interconnected - By aggregating data from healthcare providers across the country, the organization can track macro-level trends and patterns in hospitalization costs, treatments and length of stay. The large-scale perspective offers insight into how effectively the DRG system is being implemented and what providers can do to increase efficiency.Intelligent - Complex statistical analysis on hospitalization data enables the organization to keep DRG classification codes in step with the actual healthcare system. It also enables hospitals to identify areas of inefficiency so that they can work to improve patient care and increase profitability.
Company name: SESCAMInstrumentedIn the research department, the solution helped achieve better statistical control over processes and strengthen the effectiveness in initiatives requiring significant monitoring and management of data and predictive capacity. This is important because in this area some of Spanish science’s reference lines of research are being conducted, which have ramifications for other projects at national and international levels.InterconnectedEffective management and better focused research. All the data from the organization is gathered in the a database that includes demographic data (age, sex, residence) and registers the main diagnosis for admission in hospital, risk factors and possible complications, some relevant diagnosis techniques and therapies. Each year it conducts 13 million medical appointments, more than a million in the pediatric sector, more than eight million nurse appointments, 180,000 hospitalizations 1.2 million hospital stays and 200,000 surgical operations. The administrative department uses Statistics solutions to manage all of the information contained in the CMDB. This enables them to identify and compare potential problems affecting the organization’s health care provision, facilitating their resolution and improving the quality of hospital careIntelligenceUsing the tools very quickly bore tangible results. In the health care management field, one of the most important aspects is identifying inefficiencies across hospital processes.In the research field, the National Hospital for Paraplegics required a robust statistical tool for the development of more ergonomic wheelchairs and other innovative devices to assist their patients in their daily lives. The objective of Hospital’s Nutritional Illness Unit is to improve the cure rate for illnesses such as anorexia and bulimia, which are estimated to affect three percent of the population. Only fiftypercent patients are cured. The goal is to monitor sufferers over the long term and record the possible variables so as to improve prognoses and treatment. The predictive analytics tool was of fundamental help to them in implementing this monitoring.For its part, Primary Care Management has a research group focused on studying the sensory capacity loss related to aging. To do so, it uses tools to develop broad image of this loss of auditory and visual functions covering a large number of patients, enabling a treatment to be prepared sufficient time and quality of life to be improved in old age.
InstrumentedBoth a web interface and scanners help researchers gather data on more than 500 patients a year. Data is then analyzed by using tabulation and other statistical techniques. The data management group benchmarks the resulting information against national and international data. InterconnectedClinicians work hand in hand with the data management team to help guide treatments and record feedback. The data management group also enjoys strong support from hospital and executive staff, who value the independent, unbiased data and information the group provides in support of enhanced patient outcomes. IntelligentRapid, unbiased analysis of huge volumes of data is having a direct positive impact on care quality and patient outcomes. By connecting doctors with reliable, unbiased information, the solution provides a clear view of patient information enabling doctors to make real-time decisions and act confidently to deliver better patient outcomes.
Parallel Session 1.9 Using Health Analytics for Improved Outcomes
Healthcare | BAO | NetherlandsUniversitair Ziekenhuis Antwerpen (UZA) - (University Hospital of Antwerp)How can we improve the care and prognosis for patients diagnosed with rare diseases?For this University Hospital, a rare-disease diagnosis platform allows for an earlier, quicker and more accurate diagnosis by integratingboth medical expertise and data mining toolsThe Opportunity What Makes it SmarterOne of the biggest challenges in treating rare The University’s rare disease diagnosis platform allows for an earlier,diseases (those that affect fewer than five quicker and more accurate diagnosis by integratingpeople in 10,000) is the provision of an both medical expertise and data mining tools. Rules areearly diagnosis. Because of the generated much faster and more accurately through a predictive model basedrarity of their disease, patients may not be on known patient data. In comparison with a pure rule based system, adiagnosed early enough for the most combination with data mining tools provides both higher sensitivity and moreeffective treatments. UZA wanted to be able specificity. The solution can serve as an intelligent and dynamicto diagnose and treat these rare diseases knowledgebase on rare diseases. The improvement in the quick diagnosis and treatment of rare diseases can mean the difference in the lives of patients.earlier and more effectively. To do so, itneeded to be able to access and useinformation from many sources. Real Business Results • Rules are generated more quickly and accurately using a predictive model based on known patient data • The solution can serve as an intelligent and dynamic knowledgebase on rare diseases, improving the quick diagnosis and treatment of rare diseases • Compared with a pure rule based system, a combination of rules and data mining tools provides both higher sensitivity and more specificity ―The improvement in the quick diagnosis and treatment of rare diseases can mean the difference in the lives of patients.‖ – Geert Smits
Healthcare | Business Analytics & Optimization | Southern EuropeCiudad Real HospitalWhat if you could know in advance which patients would benefit most from initial interventions?A Spanish hospital is using predictive analytics to make significant improvements in the treatment of severe eating disorders.The Opportunity What Makes It SmarterThe hospital wanted to identify positive The ability to effectively handle and analyze data is essential to diagnosing illnesses earlier and speeding patients to recovery. Ciudad Real Hospital implemented a powerfuland negative prognosis factors in predictive analytics solution that enabled its practitioners to establish reliablelong-term monitoring of patients forecasting, control and early diagnosis variables for patients withbeing treated for serious eating severe eating disorders. The solution provides more accurate initialdisorders, such as anorexia and bulimia.Because such disorders affect almost 3% of patient evaluations, and has helped the clinical staff identify specific subgroupsSpains population, the goal was an urgent one. within the total patient population for whom initial interventions should lead to moreBut due to the high number of variables that successful treatment outcomes. The solution is also pointing the way towards new linespotentially factor into prognosis, the hospital had of research. For example, in supporting studies on the link between patient motivationbeen unable to execute the complex statistical and treatment effectiveness, the solution has discovered a direct link between patientsanalysis required to identify those that were expectations (e.g., feelings of despair) and poorer outcomes, even when other variablesmost important. A more powerful solution was previously identified as being able to predict treatment responses were controlled.needed. Real Business Results • Enabled 100% improvement in data handling for more accurate initial patient evaluations helping to develop more successful treatment outcomes • Uncovered specific links between patients’ expectations and treatment results • Helped identify new lines of research to be explored • An estimated 5-10% improvement of efficiency and effectiveness treatments of several chronic diseases ―In our opinion we were provided with enough support and it met our needs to incorporate this solution into the work we were beginning.‖ – Dr. Luis Beato Fernández, Ciudad Real Hospital
Healthcare | Information and Analytics | EuropeHospital Santa BárbaraWhat if you could find patients at high risk for serious disease by looking at analytic data?A hospital in Spain uses statistics and data analysis to identify key risk factors, improve diagnosis and treatment, use resources moreefficiently and effectively, and give patients a better quality of life.The Opportunity What Makes it SmarterLike many large research hospitals, Hospital Sometimes leg pain is just that. Sometimes it’s deep-vein thrombosis (DVT), aSanta Bárbara in Spain has amassed a large blood-clotting condition that may not be discovered until clots reach the lungs—often with fatal results. But researchers at Hospital Santa Bárbara have been able to useamount of research data and a wealth of statistics software to extract research data as well as patient records, and analyzeinformation on past and present patients. that data to more effectively target which patients are at risk for chronic, hard-to-However, the hospital was often at a loss detect diseases. By using statistical analysis, researchers came up with a morein terms of how to use that data to effective diagnostic model for DVT, pinpointed that 44 percent of colon cancer patients were between 75 and 79 years of age, and determined that chronicimprove processes and outcomes. obstructive pulmonary disease patients with a BODE rate of greater than 7 had anResearchers and hospital staff wanted to be 80 percent mortality rate within 48 months after diagnosis. With insight from in-depthable to extract critical data from various analysis, this hospital and others like it can more effectively screensources and analyze it to better diagnose patients for these serious conditions and others, advise preventive measures before disease takes hold, and even create new devices and treatments based onailments, refine treatments and innovate with both new research and past experience.new devices and procedures. Real Business Results • Established a new, reliable diagnostic model for DVT, expected to enable earlier diagnosis and treatment in high-risk patients • Helped researchers determine that age is the biggest risk factor in colon cancer patients, enabling staff to more effectively target tests to more high- risk patients • Reduced the cost of colon cancer diagnosis by 99 percent with targeted testing • Enabled researchers to isolate obesity as a key risk factor for chronic obstructive pulmonary disease, helping doctors get patients on the track to good health early ―Based on diagnostic models for chronic illnesses, we can provide clear evidence of risk factors and prescribe more effective treatment for individuals, resulting in better health outcomes.‖
Healthcare | Information and Analytics | AustraliaMetro Spinal ClinicWhat if your patient could show you what pain looks like?Metro Spinal Clinic, a pain management clinic in Australia implements an online patient data collection system that enables patients todescribe their pain symptoms more graphically and allows faster, more accurate diagnosis and treatment with statistical analysis.The Opportunity What Makes it SmarterMedical offices have gotten by with What causes chronic pain? Sometimes it’s obvious, but sometimes getting to the root of painpaper-based patient information systems for takes a little more digging and a lot of hindsight and research. One chronic pain managementyears. Done well, they can be quite efficient, clinic in Australia is diagnosing and treating pain with a new solution based on online databut a manual system will never match the collection and statistical analysis. Instead of filling out two-dimensional paper questionnaires, patients complete an online survey where they can describe their pain on a graphicalspeed and accuracy of an online data representation of the human body. This and other information gives physicians a more visualcollection process. One chronic pain clinic in look at patient pain. When combined with historical case data and peer discussions of painAustralia recognized that its paper-based management, staff can more accurately diagnose pain, refer treatment andpatient data process was taking too long, raise red flags when something isn’t right. Current patients benefit from fast treatment, andboth to input data and to find that data when future patients can benefit from the ever-growing database of information and analysis. The newdoctors needed it—and when patients are in system alleviates pain for the clinic as well, saving thousands of dollars in administrative costspain, even a few minutes is too long. The and reducing staff labor.clinic wanted a more efficient system, butalso wanted more analytical power to Real Business Resultstake that patient data and analyze it formore insight into diagnosis and • Reduced total administrative costs at the clinic by 75 percenttreatment. • Cut the cost per survey from USD10.65 to USD1.14, a 90 percent decrease • Increased post-treatment questionnaire follow-up rates to 85 to 100 percent • Enabled physicians to diagnose and treat pain more quickly and accurately with real-time access to data and visual representations of patient pain Measuring our own patient outcomes gives our future patients more realistic expectations of the treatments, and by benchmarking ourselves, we can continually improve upon patient treatment options and care.
Healthcare | Information & Analytics | JapanA Large Japanese HospitalWhat if predictive analytics could treat liver disease?A Japanese hospital uses regression and decision-tree analyses of patient records to predict the effectiveness of specific treatments foreach individual patient.The Opportunity The Solution What Makes it SmarterDetermining why one treatment The solution captures detailed patient By analyzing more than 400 factors perworks for one patient’s liver records and aggregates them into a central patient and by cross-functionallydisease and not another’s is greatly database, which provides a wealth of data in aggregating that data, doctors are able to which to run decision-tree and regression identify the specific treatments that will yieldenhanced by building predictive analyses. Through the creation of predictive the best result for each for patient.models based on more than 400 models, based on the records of patientsfactors, such as age, sex, race, blood who have had liver disease, doctors aretype, blood sugar content, body build, able to determine which treatmentmedical history and lifestyle. The models options would be the most effectiveenable the hospital to more accurately for each individual patient.assess the percentage of curerates for specific treatments. Real Business Results • Improved accuracy of virus removal by approximately 43 percent • Enables patients to avoid unnecessary, expensive and painful treatments if they are deemed inappropriate by the model Insight Smarter healthcare is using predictive analysis to help fight infectious disease and improve individual patient care.
Healthcare | Information and Analytics | Northeast EuropeA cardiac medical research organizationWhat if the doctor’s job were already halfway done by the time a heart attack patient checks into the hospital?A cardiac medical research firm in the Netherlands helps paramedics diagnose and treat heart attack patients on the way to thehospital when it applies advanced statistical analysis and predictive analytics to research data.The Opportunity What Makes it SmarterIn medical research, behind every discovery is The critical window for treating a heart attack is within one to two hours after the initiala mountain of data. And much of this data is attack occurs. After that, patients require more invasive treatment and longer recovery, andcomplex; any number of variables can be a the heart may sustain more damage in the long run. That often means administeringfactor in medical treatment. No case is the treatment before a doctor even enters the picture. How does a paramedic recognize a heart attack and know treatment options? Using advanced statistical analysis and predictivesame, and every case must be considered. analytics software, one medical research company created an algorithm that tellsOne Dutch research organization doing work paramedics the probability that a patient is having a heart attack based onon prehospital treatment for heart attacks had symptoms, patient history and other factors. The program will then suggestthousands of doctor, hospital and patient the most appropriate prehospital treatment and also direct the ambulance to thesurveys to consider as well as treatment nearest hospital with full cardiac care facilities. The solution saves precious time and moneyoutcomes and ambulance records. To make and helps ensure that hearts keep beating strong for years to come.sense of this data and help create aprotocol for prehospital treatment, the Real Business Resultsfirm needed a way to perform advancedanalytics on complex human data, including • Expects to improve patient survival rates and recovery times by treating heartpredictive analytics and in-depth regression attacks sooner after onsetanalysis. • Stands to improve treatment accuracy with a proven methodology and algorithm for heart attack diagnosis and prehospital treatment • Expects to save hospital operating costs and patient expenditures by reducing the average length of stay We now have a world-class analytics platform that matches our world-class reputation as a research organization. We will continue to use SPSS to support our research and hope to make further breakthroughs that enhance patient care and improve outcomes across the whole spectrum of cardiology.
Healthcare | Business Analytics and Optimization | ItalyA leading Italian cancer research instituteWhat if a cancer research center could create individualized cancer treatments that would reduce the number ofunnecessary treatments while improving therapeutic outcomes?This Italian medical institute pioneers the use of advanced analytics to analyze insights from clinical data, combined with patientinformation, to create personalized treatment plans for its patients, helping it treat cancer and other diseases more effectively.The Opportunity What Makes it SmarterThe fact that one-size-fits all cancer treatment may Until now, ―personalized‖ treatments in cancer and other disease treatment have generallyresult in more than half of all patients receiving been based on clinical trial results, a doctor’s subjective memory of past cases, and evenunnecessary treatment helps bring the goal of intuition. Finally, true evidence-based, personalized medicine is being implemented atfinding the right treatment plan for each the institute, where an in-depth analysis of a patient’s personal makeup and diseasepatient into sharper focus. This leading Italian profile, combined with insight gained from the analysis of past cases and clinicalcancer treatment and research center wanted to guidelines, enables doctors to provide an optimal treatment plan for each patient. Theimprove patient care by tailoring treatment solution proactively shows the physician statistics on similar clinical cases, possibleapproaches to specific individuals. The institute alternative treatments and predicted outcomes for each, allowing the doctor to makeneeded the ability to analyze past treatments and a truly informed decision. One insight from the solution showed that, statistically,cases, and combine that information with the physicians tend to give more aggressive medical therapy to women who are sick aspatient’s personal statistics and disease profile, to opposed to men with the same problem. Knowing this, physicians can guard against suchcreate a fact-based treatment plan for over-treatment, ensuring that patients receive only the medicine and procedures they need.each patient. In addition, being able to analyze Real Business Resultsoverall outcome data would help the institute • Avoids unnecessary treatment (estimated to be up to 60% of all treatment) andprovide more cost-effective, efficient care delays in treatment deliveryfor its patients. • Creates tailored and personalized treatments, increasing the chances of successful outcomes • Improves hospital performance, both clinical and operational, by providing a ―big picture‖ view of treatment delivery, helping streamline processes and lower costs By providing physicians with vital input on what worked best for patients with similar clinical characteristics, the institute can help improve treatment effectiveness and the final patient outcome.
Government, Healthcare | Smarter Analytics | Central EuropeA government healthcare organization in Central EuropeWhat could you do if you had up-to-date healthcare cost statistics for an entire country at your fingertips?A government healthcare organization uses sophisticated statistical analysis to calculate fair prices for healthcare services under adiagnosis-related group (DRG) system while helping hospitals root out inefficiencies.The Opportunity What Makes it SmarterAs the organization responsible for Around the world, countries are working to standardize healthcare fees using diagnosis-relatedcalculating and recommending healthcare groups (DRGs). Though DRG codes help to create fair payment systems, they arefees in a Central European country, this complicated and dynamic, built on extremely complex calculations. This governmentgovernment organization must maintain organization has reined in the complexity by applying sophisticated statistical analysisunparalleled knowledge of clinical to vast amounts of data collected from every hospital in the country. The organizationtreatment paths and trends in the tested multiple data models to find an algorithm that effectively mirrors the country’shealthcare sector. The organization healthcare system. The continuing influx of data creates a feedback loop that refines thecollects and analyzes enormous amounts accuracy of the algorithm over time. With these powerful analysis capabilities, theof hospital data. Manual data collection organization can calculate relative weights to account for variations in hospital costs, monitorand spreadsheet calculations, however, macro patterns in treatment paths and identify true cost outliers that might signal a need formade it difficult to navigate the complex change in DRG codes. Hospitals also benefit from the insights, using the data to benchmarkalgorithm required to understand costs their costs and spot inefficiencies. For example, if a hospital sees that its costs for treatingand define fees. To increase accuracy heart disease patients far exceed the norm, it can take steps to find and fix inefficiencies.and speed, the organization neededpowerful statistical analysis Real Business Resultscapabilities. • Reduced the time to perform complex calculations of relative weights from 3 days to 5 minutes—a more than 99 percent improvement • Increased the frequency of analysis by 1,100 percent with one-sixth of the manual labor requirement • Shifted focus from data collection and processing to in-depth statistical analysis, yielding a clearer view of the healthcare system and more accurate DRG codes Achieving this level of accuracy and foresight in modeling and understanding healthcare costs is unthinkable when you’re limited to the calculations of a spreadsheet. The dynamic statistical analysis we have now enables us to do so much more when calculating DRG fees and relative weights.
Healthcare | Business Analytics | Western EuropeSESCAMWhat if a tool for scientific research also could help you bettermanage efficiencies and improve services?SESCAM is using a powerful analytics and optimization solution to more effectively manage and better focus itsclinical research while at the same time significantly enhancing the quality of the care it delivers to more than 2million people.The Opportunity What Makes It SmarterSESCAM has a large number of clinical research Detecting inefficiencies quickly across a large healthcare network is key to better resourcecenters with projects focusing on improving management and service delivery – but doing so effectively entails managing, and analyzing,diagnosis and treatment. The organization needed enormous quantities of data. Implementing a sophisticated analysis and asset management solution hasappropriate statistical and analytical tools to support vastly improved this health services ability to manage operations across a vast web of 6,000 doctors, 31these projects that could, at the same time, pharmacists, 80 dentists, and 7,000 nurses. In the health sector, the solution is helping the organizationcontribute to the more effective management of easily manage all this information, identifying potential problems and inefficiencies across hospitalall of its hospitals and health centers by processes and facilitating their resolution to improve quality of care. In the research field, the new solutiondetecting inefficiencies and enhancing the level is helping investigators in developing innovative devices to assist patients in their daily lives and toof service provided to clients. improve the cure rate of illnesses. For example, clinical researchers in the hospitals nutritional illness unit hope to improve the cure rate of individuals suffering from anorexia and bulimia, illnesses affecting an estimated 3% of the population. The solution’s predictive analytics tool has been of fundamental importance in enabling researchers to implement a long-term monitoring program of these individuals in which possible variables relating to these diseases are recorded and analyzed, with the goal of improving prognoses and treatment. Real Business Results • Improved the efficiency of managing the organizations vast network of healthcare facilities • Enhanced the quality of care by identifying and eliminating inefficiencies in hospital processes • Provided researchers a powerful statistical tool for the development of more ergonomic wheelchairs for paraplegics • Aided researchers in studying sensory capacity losses related to aging ―We are now in a better position to optimize patient care across our large network – one that conducts more than 13 million medical appointments and 200,000 surgical procedures per year, as well as managing 1.2 million hospital stays annually."
Healthcare | Business Analytics | AustraliaWesley Research InstituteWhat if you could use analytics to dramatically improve patient care and quality of life?Wesley Research Institute is deploying a powerful analytics tool that enables doctors to make better-informed decisions that resultin improved patient outcomes. What Makes It SmarterThe OpportunityHaving access to enormous volumes of patient Effective hospitals manage, integrate and analyze clinical and research data to tackledata, this medical research institute wanted to find complex problems and improve patient care. This institute implemented a statistics-baseda better way to leverage all this basic analytics solution that successfully collects huge amounts of data – including patientinformation to enhance patient outcomes. A demographics and data on discrete procedures, surgical complications, risk factors and post-core element of the institute’s research program is operative tracking – and then quickly analyzes it to establish clinical benchmarks, identifythe collection and analysis of large amounts of risk factors and improve treatment results. The solution uses powerful statistical analysesdata from multiple sources and in various formats. to evaluate the range of factors influencing successful medical treatments. This enables theThe data need to be carefully and accurately institute’s doctors to proactively modify protocols to improve quality of care: for example,managed and interconnected in order to create an doctors who are about to perform high-risk surgeries can view statistics on the risk ofaccurate picture of patient outcomes. Without mortality, based on patient demographics, and make better, more well-informed decisions oncomprehensive data collection and more behalf of their patients.sophisticated data analysis, the institute wasstruggling to gain deeper insights across patient Real Business Resultsgroups. • Improved patient care across the hospital – for example, by identifying trauma (bruising, bleeding and hematomas) caused by catheter use, which the hospital immediately addressed • Improved clinician productivity by providing doctors with access to accurate data across patient groups • Decreased time spent in the production and delivery of reports to clinicians • Provided clinical staff with a valuable resource for ad hoc queries ―By quickly and accurately analyzing large volumes of data, we are helping to enhance quality of care and improve patient outcomes. The solution is making an enormous contribution, not only to us but also to the larger healthcare community.‖