Skepticism is a double-edged sword. Without skepticism, IT managers might invest in software that is useless. Enough scepticism can allow IT departments to wait for enough evidence to prove that a particular platform can deliver good results.
Data analysis of medical industry reaches critical point
Large data analysis has now reached a tipping point in the healthcare industry. Some vendors promise to provide better quality of care while reducing spending, but there is evidence that these claims are controversial. Similarly, critics of the big data movement point out that the medical industry can extract information from small datasets before launching large projects.
For example, in a recent blog post, the consultant and Doctor David C. Kibbe and Vince Kuraitis argued that operators had better focus on using small data rather than succumbing to the temptation of large data analysis. In other words, focus on the clinically existing data stored in digital form, dedicated to the use of it tools that are directly applicable to nursing management.
On the other hand, large data analysis attempts to parse stacks of data from many different sources, trying to find patterns that are useful for solving problems. For example, researchers have used large data to study the performance of genetic and environmental factors in multiple sclerosis, thus finding individualized treatments.
Some research may bring exciting rewards, but it companies are impatient. As Kibbe and Kuraitis point out, technology companies are touting big data analysis as a health-care system, and doctors ' groups aim to be a trusted medical institution or, with taxpayers ' support, to establish similar organizations like ACO. As these ACO and medical institutions try to gain profits from contracts that share savings or financial risks, these big data advocates claim that big data can help them consolidate data, making quality better and reducing costs.
Some suppliers have already used large data in patient care. Many suppliers are turning to companies like Microsoft, SAS, IBM, Dell and Oracle to gain expertise in data mining, according to Business Week. Health care analysis is an evolving business. Frost&sullivan's study showed that half of the hospitals used advanced analytics software in 2016, and now 10%.
Have medical providers prepared for big data analysis? Are they satisfied with the fact that more limited data analysis functions are built into their electronic medical records system and the relational database that points to new policies and procedures?
"It's not a matter of choice," said David Bouloux Blumenthal, a former head of the National Health Coordination It office, in a recent interview with the Information weekly health care section. Large data is started from small data. We have more information on health, disease and care patterns that will give us useful insights into what works, what doesn't, and what is the natural history of disease. It enables us to do research faster and more efficiently ... But it will take a while for us to figure out how to use the data.
"We believe that science offers opportunities, and that most of the time, we are right," Blumethal said, hearing many big data critics. ”
From this point of view, the author summarizes the work of 7 companies and large medical centers, pointing out that they are already making similar attempts.
The Cleveland Clinic's product Explorys provides a cloud-based performance management platform that provides healthcare providers with clinical, financial, and operational data management that has never been the model before. Among its clients are Joseph Health Systems, MedStar and Catholic Health partners.
Unlike the old-fashioned analysis of relational databases, the company's services depend on Cloudera companies. This is a service company based on Hadoop software that can help engineers and information professionals deal with heavy work.
The Explorys platform allows vendors to do three things: search in patient populations and health-care settings to help identify trends in disease development; Coordinate rule-driven patient registration; Performance metrics-This is a key factor if an organization intends to meet ACO requirements.
Of course, all these efforts are meaningless if it does not produce strong data to show better medical quality and reduce costs.
"Because the company is relatively young, it has not yet produced these results," explains Anil Jain, chief medical information officer and Doctor of Medicine, Explorys. In other words, there is no evidence that it can reduce the number of people who have diabetes or who have died from myocardial infarction in patients with congenital heart disease.
But the data produced by Explorys show that it is approaching this goal. For example, in collaboration with the Cincinnati Catholic Health Partners, the analysis platform helped to increase the vaccination rate of 14% of the pneumonia vaccine, a 13% increase in breast cancer screening, and an increase of 3% glycosylated hemoglobin detection rates in diabetics-a measure of long-term glycemic control.
The Journal of the American Association of Medical Informatics (Harvard) recently outlined the Explorys project in a report that collects data from the electronic medical records of nearly 1 million patients in several different health care systems. Analysis can help doctors find the most dangerous blood clots in their limbs and lungs. The analysis takes only 125 hours and requires the least amount of manpower. Such an analytical project, using traditional research methods, usually takes several years to complete.
Humedica provides a cloud based demographic analysis system. It connects patients with a longitudinal view of the information produced by different medical settings-outpatient and inpatient as well as patient care during the time period. The company has data from nearly 250,000 patients in more than 30 countries, allowing personalised clients to compare their physical performance with a very large demographic data.
The company's service integration, standardization and validation of clinical data are carried out from different levels of care, not only for drug treatment, laboratory testing results, vital signs, demographics, hospitalization and outpatient, but also for physician notes and laboratory testing results, both structured and unstructured data are utilized. Its customer base comes from four major categories: the Integrated Delivery Network (IDN), the large academic medical Center, the multiple hospital medical system and the practice of large medical groups.
A typical example: the patient-centered healthcare provider in the Hudson Group has used Humedica's Minedshare Analysis service to measure its patient population and compare it with industry best practice services.
For example, the 125-doctor's approach is to extract data from diabetics to determine that the Hga1c readings in the last interrogation were at 7% (less than the best signs of glycemic control), and indicate which patient had not been seen in the last year. Because the system is used, hospitals are able to focus on these high-risk patients and ensure that about one-third of them receive a return visit at least once in the first 8 months of the project.
In this group, one-third of the patients achieved the goal of keeping the Hga1c figure below 8%, and the number of patients with 60%HGA1C more than 9% were trying to reduce blood sugar through frequent visits to primary care physicians.
Further evidence suggests that the Humedica is a target of ROI for the clinical indicators provided by Minedshare, and that the supplier now has a 3-level identification Mark issued by the National Quality Committee (NCQA).
InterSystems like to remind medical service providers that even large-scale data warehouses may not be enough to provide all the necessary intelligence to provide quality care and significantly save money. The intelligence needed to be responsible for the emergence of a performance model for health-care institutions and similar payments is now the most urgent need.
InterSystems's Healthshare Healthcare information platform, with its built-in active analysis components, can solve this problem. Like many other large data providers, collect, assemble, standardize, and present data from a variety of isolated island patients to help decision-makers provide clinical and financial support.
Rhode Island State is using healthshare throughout the state to facilitate the exchange of health information and to collect and analyze patient data. This makes the national medical Practice team able to do clinical summary Exchange to improve the quality of care, which is an important part of ACOs.
Gari Christensen, chief information officer of the Rhode Island State Quality Association, praised Healthshare on the InterSystems website, saying: "Healthshare provides a cost-saving and a certain level of quality care services to Riqi, This is the result of the analysis that doctors cannot get from their own records. In a recent interview, Christensen said: "His team used InterSystems's analytical tools to conclude that 8% to 12% of the tests carried out in Rhode Island State over One-fourth of the population were repetitive and medically unnecessary."
Sweden has also discovered the benefits of using Healthshare InterSystems, which creates a national electronic medical record system for 9 million people. The system is browser-based and can display patient demographics, drug lists, laboratory data, allergies, and related information.
Insurance fraud is one of the most intractable problems in the healthcare industry, taking up a lot of pervasive time and energy. Pervasive's datarush is a high-speed parallel data-processing multi-core computer and multiple computer-networked application frameworks and analysis engines that can help service providers collaborate with national institutions in detecting Medicaid fraud. In a case study, the company stressed on its website that pervasive could help restore medical compensation, which should have been paid by private insurers.
To detect fraud, some service providers use SQL Server to match insurance files, which is a tedious process for a long time. Datarush fast fuzzy Matching system searches two databases, one containing the name of Medicaid received from the state, the other is the name of a patient participating in a private program, and ultimately to find overlapping parts. According to the pervasive situation report, the end result is to reduce operating costs and achieve faster ROI.
Clinical inquiries may not have the same profit motive as big business data companies, but it is certain that it has its advantages in the untapped medical data.
Clinical enquiry is a medical information platform used at Beth Israel's Women's Deacon Medical Center to improve product quality while reducing costs. To achieve these two goals, doctors need to focus not only on the patients sitting in front of them, but also on people who have the same conditions or conditions-the so-called population health management. The data analysis tools needed for this task are more complex than ever before.
BIDMC's chief information officer, John Haramka, has "access to clinical inquiry" as a clinical trial or clinical research business Intelligence system. This is a search engine that is linked to a huge database of patient records, allowing hospital employees to test what causes a disease, such as testing drugs, diet or lifestyle variables that could lead to disease.
The library contains 2 million data points for 2.2 million patients, including drug taking, diagnostics and laboratory values. The query tool is able to browse 20,000 medical concepts that use Boolean expressions. All data is mapped to the standard medical language code. For example, diagnostics map to ICD-9, drugs are mapped to Rxnorm code, lab data maps to the logical observer identifier name and code (LOINC).
With the help of clinical inquiries, clinicians or researchers can search for records to find out how many breast cancer patients are also taking ACE inhibitors, a type of drug used to treat high blood pressure. If there is a strong correlation between the drug and the malignancy, the hospital can do more in-depth analysis and establish a formal research project to investigate.
The ultimate goal is to discover a new medical method to improve the survival of the entire breast cancer patient.
"The uniqueness of clinical enquiry is that it is a fully self-service service," Haramka said. I don't need to hire an analyst. Nor do I need special permission from our Institutional Review committee to use it. ”
Ibm/memorial sloan-kettering Cancer Center
At a recent digital health conference hosted by the New York Electronic Health Association, IBM Chief Medical Science and medicine Doctor Martin Cohn and Memorial Sloan-Caitlin Cancer Center (MSKCC), Chief information Officer Pat Skarulis in New York, proposed a common use of Watson's supercomputer's large data processing capabilities, To help oncologists provide better services to MSKCC patients.
Mr Cohen points out that Watson is not only a steroid's "search engine", but even a huge database. It relies on the parallel probability algorithm to analyze the millions of pages in the patient's medical history and medical literature for the unstructured text, finding the most relevant answers to diagnose and treat related problems.
90% of the world's data has been produced in the past two years, and 80% of the data is unstructured. Any clinician with an unread medical journal knows that much of the information collected is not included in the literature that people read.
Watson read the literature for these people and increased the speed.
With natural language processing (NLP), the computer not only translates the relevant terms and conditions to meet clinicians ' search criteria in the query, but also understands idioms and other special forms of expression in English. With the help of time, statistical interpretation and geo-spatial algorithms, it is possible to find meaningful links between clinicians ' problems and the large collection of medical facts and theories.
Skarulis said MSKCC decided to work with IBM to "build an intelligent engine to provide specific diagnostic testing and treatment recommendations". The two organizations will now combine the vast database of MSKCC, known as Darwin, with data from Watson's natural language processing capabilities.
IBM is using all of the structured patient data of the medical Center and its NLP tools to convert the free text consultation notes in the Medical Center to the available data. Skarulis hopes to launch a pilot project that will allow supercomputers to work for real medical conditions in the near future.
Oracle/Pittsburg University Medical Center
The Pittsburg University Medical Center (UPMC) has taken a step forward in response to the Big Data initiative, investing 100 million of dollars to establish a comprehensive data warehouse that brings together data from over 200 data sources across the UPMC, UPMC Health plan and other associated entities.
To collect, store, manage, and analyze information that remains in the Data Warehouse, UPMC will use the Oracle Exadata database machine, a high-performance database platform. Use IBM's Cognos for software business intelligence and financial management. At the same time, using the Informatica data integration platform and the Dbmotion interoperability platform, it integrates patient records from medical institutions and health information exchange. These tools will manage 3.2 gigabytes of data that flows through the business unit of UPMC.
Our goal is to help doctors enter a more intelligent era of electronic medical records. For example, depending on the results of laboratory tests, attention to subtle changes in indicators, awareness of the risk of renal failure, and, for example, according to the patient's genetic and clinical information, for breast cancer patients with the most effective, the least toxic treatment options. In the case of breast cancer, most of the work will be done through the analysis of the patient group, allowing researchers and physicians to focus on the patient's response to the effects of treatment and their health over time.
UPMC officials explained that they would start using new analytical tools, and they collected data from a group of 140 breast cancer patients previously studied. Researchers have already had the genomes and electronic medical records of these patients, giving researchers a good start in seeking to understand individual differences and their response to treatment.
"Innovation is important for both Oracle and UPMC because they are developing an enterprise medical analysis platform that integrates data from clinical, genomics, finance, management and operations across the organization," said Nier Kreschenzo, senior vice president and general manager of Oracle Health Sciences. All of these areas require their workflows to be more productive, as this is the challenge UPMC to meet the exponential growth in data.
To complete the challenge of sorting this data, UPMC will use Oracle tools extensively, including Oracle Enterprise Medical Analysis and Oracle Health Science Network. UPMC will also use Oracle Fusion analysis and multiple components, such as Oracle Hyperionprofitability and cost management in Oracle Fusion Middleware, to support cost accounting, Oracle Identity and access Management suite to enhance compliance and data protection.
(Responsible editor: The good of the Legacy)