To say that the industry that is most likely to obtain information from different data sources and benefit from analysis, medical service is undoubtedly a well-deserved winner. But it is not enough to have such will. Practitioners need to be prepared for possible difficulties and obstacles.
With the continuous rise of electronic health record system, imaging system, electronic prescription software, medical claims, public health reports, related applications and mobile medical equipment, the medical service industry has become the primary beneficiary of big data with the most analysis and development potential. Massive data from patients, medical records and even related institutions are waiting for the in-depth exploration of big data analysis tools.
Data analysis has made many beautiful promises to the development of medical industry: cutting costs, improving execution efficiency and bringing better nursing experience. Except for a few industry leaders who take care of big data tools, it is still difficult for most medical institutions to obtain all kinds of data from proprietary systems in complex ways.
Data is data, regardless of source.
Generally speaking, the data of most medical service institutions still come from clinical, financial or commercial applications. In itself, each type of data has a specific purpose. According to the report entitled "Analysis: Nervous System of IT Medical Treatment" by Health Technology Transformation Association (iHT2 for short), clinical data can improve the quality of nursing and simplify the health management of the population. Financial data can help hospitals conduct cost analysis at the grass-roots level, check infrastructure operation data and manage resource utilization.
Putting all factors together, institutions can begin to evaluate macro issues, such as meeting the needs of employees, improving work efficiency and nursing quality. Laura Madsen is a business intelligence advocate of Lancet software and a pioneer of medical services, so she doesn't think it is necessary to strictly divide different types of data sources. "Data is data," she pointed out. "In the final analysis, data is composed of bits and bytes ... As excellent data experts, we should integrate clinical data with business data."
On the one hand, the government's planning and entrustment put pressure on the medical service industry, on the other hand, it also prompted the latter to start to examine and analyze the work with a serious and rigorous attitude. The ideal use scheme will push managers to choose electronic medical record system through economic leverage, the mode of responsible nursing institution (ACO) needs to coordinate patient care, and the patient-centered family bed will push the attention to nursing quality to the extreme-all of these need to be based on more complex medical service data analysis functions.
A large number of unstructured data increase the difficulty of analysis.
Of course, medical institutions must complete data collection before they can start data analysis. In the field of medical services, iHT2 summarizes the following main implementation difficulties for us. First of all, up to 80% of medical data are unstructured, including paper grids and other non-electronic free forms, which need to be manually summarized by technicians. In addition, even structured data (such as information from the process of medical information exchange (HIE)) is not enough to support the analysis work. The report concludes that medical service providers often need to make a macro evaluation of themselves from the claims data of insurance companies.
In the field of medical business intelligence, the scale of data is very important, and Madsen has repeatedly emphasized this point. As the largest medical service provider in the world, this organization named Inter Mountain Healthcare and Kaiser Permanente has been working hard for a long time, but for small suppliers, the gap between ideal and reality is still "huge". Madsen added that most organizations have clearly realized the value of business intelligence, but they have not found a clear answer to the following question-"What should we do?"
Most schemes focus on business intelligence to meet the needs of regulatory reporting. This is understandable because every hospital needs to submit thousands of reports to government agencies every year. However, despite the amazing number and variety of reports, it is often difficult for medical service organizations to use the data in them to guide their operations, improve their efficiency or make improvements in other aspects, Madsen explained.
Fortunately, iHT2 made a series of suggestions in the report. First of all, service organizations can assess the medical needs of patients, which can help organizations to formulate appropriate service delivery methods, determine the subtle differences between different individual care needs, and even predict which patients may develop into severe patients. In addition, institutions need to evaluate their own resource reserves to promote the improvement of service quality and find out the reasons for the fluctuation of nursing level.
In addition, I want to sound the alarm and list some practices that sound reasonable but harmful. The purpose of medical insurance * * * sharing saving plan and ACO model is to reduce medical costs and finally realize expenditure saving. According to iHT2 survey, this will lead to the invalidation of income cycle management tools. In addition, today's cost calculation system often cannot accurately evaluate the overall nursing cost of institutions. To do this, the system needs to consider the problem from a macro and long-term perspective, and offset the money saved by the lack of hospital facilities in the early stage with the income loss caused by it in the future. IHT2 believes that in order to find a reasonable total cost summary mechanism, everyone needs to "take the causal relationship of income as the context and adopt a precise calculation system based on actual conditions".
Medical services and data analysis-it will take effect whenever and wherever we meet.
Not all analytical systems in the medical service industry have to be very complex. Chunshan Memorial Hospital in Mobile, Alabama has just introduced an automatic dispensing system called Pandora Clinic. The addition of this analytical tool set greatly reduces the time spent in transporting drugs in hospitals.
This software, created by Omnicell, can track the circulation process of drugs from the pharmacy cabinet as the starting point. It will generate monthly reports to help hospital managers accurately classify drugs and match them with other drugs. JoeAdkins, a clinical pharmacist, said that although in the worst case, hospital employees may take drugs for themselves or even resell drugs, on the positive side, nurses or clinicians can help patients eradicate pain faster with the help of this mechanism. He told us that this software can't deepen the relationship between posts, but it is the first tool to guide and promote the interaction and cooperation between posts.
More importantly, adkins added, Pandora Clinic has little impact on the whole workflow. The report will be automatically sent to the employee's mailbox on time, and the information will be represented by an intuitive bar chart (instead of a lengthy written explanation). In short, people who read the report no longer need to look for results or make comparisons through mental arithmetic: "You don't have to think much, you can see all the conclusions, which is the ideal report form."
On the other hand, for patients who are payers, the goal of the product should be to improve the customer experience and let patients know the situation easily and intuitively, said Bob Dutcher, vice president of business marketing for InsightsOne predictive analysis.
In the pilot activity of 20 12 (which is still going on), the company cooperated with Independent Blue Cross (IBC for short) to help insurance companies in Philadelphia determine which patients may be dissatisfied with the service experience and hope to eliminate serious problems before they happen—sometimes as high as three months ago, Dacher pointed out. (The company also helps IBC identify potential new customers and evaluate whether existing customers can benefit from unused services. )
In order to achieve this, IBC collects data through its call center, aiming to know which patients need to be asked frequently and regard them as a group that may need extra attention. They also collect data from medical service member institutions, hoping to know what follow-up inquiries patients generally need and why individual patients need additional treatment. This conclusion can help IBC quickly locate those "patients with high probability of negative results" and use it as a breakthrough to send information related to preventive health care (avoiding repeated hospitalization due to the same symptoms) or long-term family health care services (if patients are about to usher in a long physical recovery period).
Dacher pointed out that this positive method improves the overall experience of patients, and also helps family members bid farewell to unnecessary or repetitive diagnosis processes, saving a lot of money in disguise. InsightOne called this analysis result "predictive intelligence", and said that its function is to guide analysts to draw conclusions according to the actual situation of a single patient.
Analysis needs talents and data warehouse; At present, the supply of both is still in short supply.
Generally speaking, it is easier for insurance companies to invest in advanced analysis projects, but it is relatively more difficult to urge medical service institutions to invest. However, Cynthia Burghard, head of IT strategic research on nursing responsibility in IDC's medical analysis department, said that two major reasons have begun to reverse the negative concept of medical institutions.
The first argument is that patients are more willing to participate in fitness programs than health rehabilitation programs recommended by doctors (although the latter may be more suitable for patients).
Another way of saying this is that the medical reform efforts in the 1990s failed due to lack of data support. "Not only is the available information too limited, but doctors don't have enough historical records and correct data formats to help them understand the difference between current performance and treatment goals," Berghard pointed out in a recent report entitled "Business strategy: analytical ability becomes a priority investment point for medical responsibility". "Discussions and even disputes between payers and suppliers are often triggered by the accuracy of data and the timeliness of discussions."
The emerging ACO model has brought a revolutionary wave to the medical service field, that is, it has evolved from a centralized charging service model to a decentralized collaborative nursing model, with more emphasis on analysis function and data warehouse technology. In order to achieve this goal, we should first make a specific nursing plan for patients, so as to further strengthen management and improve the nursing effect of patients, and finally integrate this nursing mechanism into the workflow of clinical medical staff, Burghard concluded.
She added that with the popularization of ACO model and the spread of the concept of coordinated care, medical service institutions will increasingly cross with unstructured data, market sentiment analysis and other data sources, which may include the combination of predictive intelligence, clinical intelligence and business intelligence, so as to check patients' visits and support operational decisions through clinical manifestations.
Burghard pointed out that such an advanced analysis mechanism will bring a lot of technical challenges. In his view, medical service providers will "need to hire highly skilled personnel to be responsible for data warehouse management, and only in this way can their investment be rewarded", and those employees who can neither create data warehouse systems nor take over management will be forced to merge or be assigned to large-scale integrated delivery networks.
This behavior of shuffling the existing technical groups will undoubtedly bring turmoil to the whole industry, but in the final analysis, it is a major development at the expense of sacrifice. "There is a saying that' we can't manage things that we don't know accurately'. This sentence also applies to the field of nursing responsibility, "Berghard wrote in her report. "In the 1990s, medical institutions lacked understanding of patients with chronic diseases and did not know how to manage them according to their characteristics. But times are different. Now that we have more comprehensive information in our hands, we are willing to invest money to enjoy data more smoothly with payers, medical staff and patients. "