- This topic has 15 replies, 12 voices, and was last updated 4 years, 8 months ago by
.
-
Posts
-
-
-
Managing unplanned large database is difficult to share data for use or data analysis. Management should start from a discussion of the specific data storage needed for reduce unnecessary information and going in the same direction follow a construction of the same data structure. Using big data to present data should not come from selecting specific information in order to reduce the problem of voluntary choosing the data, which may use the technique of data sampling to be fair in using information for presentation rather than using data from self-selection. There is a process for data validation. Check before recording data rather than using techniques for data analysis. There is a standardized data analysis guideline for everyone to use to analyze to reduce misinterpretation and bias problems or error in analysis. The data analyst should have knowledge and understanding of the subject to be analyzed and understand the purpose data structure details because each person’s understanding of information is different. Protecting the confidentiality and privacy of patient information is necessary for data storage. There will be a system for standardized protection, access to appropriate information and systematic user management to reduce the risk of malware, cyber attraction, virus, hacker and inadvertent disclosure.
-
Missing data: Data validation rules are one of the best ways to prevent data missing. E.g. Making compulsory field for data entry.
When the data becoming big, it is not easy to utilize. It needs training and technical support and tools.
In our country, we are transforming to e-Health but there are few people doing research or analysis from those data (struggling with the data entry).
There is a whole bunch of techniques dedicated to cleansing data. But first things first. Our big data needs to have a proper model. Only after creating that, we can go ahead and do other things, like:
• Compare data to the single point of truth (for instance, compare variants of addresses to their spellings in the postal system database).
• Match records and merge them, if they relate to the same entity.
But mind that big data is never 100% accurate. We must know it and deal with it.
Generally good levels of basic rights to information privacy exist globally. Specific health information privacy protection is not as widely present and is often contained in professional codes of conduct rather than law. Legislation specifically aimed at protecting privacy in EHRs is limited to countries where considerable deployment of EHRs already exists. -
-
Suggestion on coping with the challenges in health informatics includes:
1. Missing data: we should imply the solution in these three steps involve
1.1 Collection: The necessary information should be set as compulsory in the electronic medical record.
1.2 Consolidation: Algorithm to check required variables should be set up.
1.3 Analysis: Several analysis techniques could be applied such as imputation, mixed-effects regression model, generalized estimating equations, or inference.
2. Selection Bias: The training, monitor and evaluation should be implemented to reduce the difference between health care. Big data analytic technique included propensity score analysis, instrumental variable analysis and Mendelian randomization could address this systematic error.
3. Data analysis and training: Besides training, the health sector should give the job position and proper incentives to recruit the data analyst or data scientist.
4. Interpretation and translational applicability of results: Promoting the big data study includes different techniques and dataset could reassure the result.
5. Privacy and ethical issue: Rule and regulation should be address firmly to all health sector in order to make e-health safe and private. -
-
I think all of the challenges share some concepts in common. It is about law, regulation and coordination of benefit and interest among participated projects/countries.
In any health informatics projects, one way to gain trust from participants and be accepted by professions and academic community including to make the project’s results exchangeable, it is critical for the project owner to take one of the significant privacy compliance frameworks. Currently, there are frameworks that have been enacted, the GDPR in EU and the HIPAA in the U.S. GDPR is not only legislation applies to EU regions but also covering subsidiaries or business unit of organization in the EU. Apart from local law and regulation informatic health project should adopt either of the wide global frameworks to make the project be more acceptable, contributable and ethical concerning personal privacy and rights. Most to concerns about the framework, especially GDPR is that it does not go along with using broad consent. The law prefers participant to be clearly informed about the scope of use of collected data and its result. While at the same time, using dynamic consent may not help facilitate extremely growth in research needs and its possible information exchange. New approaches, combining the two type of consent is needed. The researcher may have to look for broad and possible scope on their project usage in future and how information from the project can be exchanged widely possible. So the researcher can coordinate with possible elaborated projects and organizations to draw a scope and make it written in the consent for future usage. For this vision, application of Entrepreneur Architecture and coordination of benefit and interest between organizations/projects is required for researcher and informatics health project owner. The EA can help speed up using EHR in countries where has less usage in order to join such projects. The EHR will help reduce missing data. Moreover, the coordination between projects and countries will help early assessment of different situation especially in the healthcare setting and society may effect variable required for analysis. The missing data and selection bias can be reduced and improved by training on EHR, data collection and storage, and dissemination among projects/countries. Coordination between projects/countries can help to share infrastructure, software, hardware for data analysis. This, in addition, offers cost containment for joined projects. They can develop core curriculum, procedure and requirement on such shared activities. In the end, integration of the outputs will not be a problem as used to be due to complexity and inadequate variable. The predefined and known variations and lacking some variables among projects help researchers planning an framing of statistical approaches and proper research methodology in advance. The coordination also helps each project/countries improve their project practices, improve health outcome and system in a timely manner.
-
-
How we can cope with big health data challenges?
1. Missing data
IT officers should cooperate with the health care officer to well and properly organized design the EMR supporting the collection of health data and preventing missing data. Also, increasing awareness in registering data in health care officer is an important milestone. If the data collection system works properly, it will reduce the number of missing data significantly. In addition, the statistical methods (eg. imputation techniques, mixed-effects regression model, generalized estimating equations, inference ) are one of the solutions to cope with unplanned missing data.2.Selection bias
There are many different factors in health care in different countries. Officers using big data for analytic information should aware of these limitations. But, we can reduce these limitations by using statistical techniques including propensity score analysis, instrumental variable analysis, and Mendelian randomization.3.Data analysis and training
Increasing knowledge of researchers about data analytics and statistical method is one of the solutions for this problem. Another is having support from data scientists and statisticians by recruiting them to the organization.4.Privacy and ethical issue
Law enforcement about health data management considering both personal privacy and public benefits must be legislated to prevent misapplication of personal health data. -
My suggestions on coping with big data challenges
– Missing data: It depends on what type of data you want to use from EMR. If you want to do prospective research that sources of data come from EMR. You can solve this problem by setting the desired data to be a mandatory field for a physician to fill it in EMR. However, if you want to do retrospective research from EMR data, the missing data should be a big problem like other retrospective research. You have to downsize the sample size, use only an EMR with complete data of interests or use statistical analysis technic to cope with this problem.
-Selection bias: Doing a research with data from EMR is the observational study that has a bias more than RCT. the research methodology for the research on big data should be matched with the nature of data sources to maximized the value of observational data and minimized bias.
-Data analysis and training: There is a new frontier for working on big data and come out with useful information that can make a better health care system. We have to train a lot of data analysts of big data because this type of data is unlike other data that we familiar with. and we also may need machine learning to generate useful information.
-applicability of the results: For the implementation of results, traditionally the results of the research will apply to the same population as stated in research population. While research on the big data, like EMR, we are difficult to define what is the character of the research population. In my opinion, we can cope with this problem by specifically identified what is the inclusion criteria. The well-defined inclusion criteria of the research population will help people to confidently apply research results into clinical practice.
-Privacy and ethical issue always a concern. the rule and regulation should be applying for data security and privacy. -
-
How we can cope with big health data challenges?
Big EMR health data contain many data base from many source and variety format and very huge. Data analysis from big data can not do until get it to the same structure. Hospital and health care center must to defind minimum data set of data order to change to health information system, or defind interorpoablity to exchange EMR in the same protocal.-Missing data:
IT from health care center have to defind code and structure EMR and health information system. Training health care provider to understand how to record the complete data-Selection bias:
Training and reserch with EMR and health information system in ethical can minimize bias.-Data analysis and training:
increate training about EMR and health data system from big health data can help health provider knows how to get the best patient ‘care. Health information from huge data can improve the right treatment of disease.-Privacy and ethical issue:
Rule and regulation of EMR should be confidential, health care provider must management both of privacy and public benific. -
– Missing data; Using EMR can cause this problem. Some of the patients cannot answer the question on EMR form and some of the healthcare providers ignore to complete the EMR form. The leader of the organization should inform their officer about how important each information which needs to fill in the EMR form.
– Selection Bias is unavoidable, the researcher should design or find a way to adjust it and acknowledge it in the study report.
– Data Analysis and Training; The leader of the project or organization should support all staff to get trained until they can perform proper work. If they can perform their work fluently, the data that they collect will be useful.
– Privacy and Ethical Issue is one of the most important things. Healthcare providers should manage both of them to maintain both healthcare providers and patient’s privacy, confidential and ethical issues.
-
As Dr.Piya said, there are more data from digitization of health care and health services, and the challanges are how to use the information for value creation of health services and improving health system performance. Moreover, missing data, selection bias, data analysis and training, interpretation and translational applicability of results, as well as privacy and rthical issue are should be concerned. For coping all of the challenges, we should have a plan cover every approach: collecting data, quality control, analysis, systemically storing, etc. Good plan and preparation lead to a good performances. In addition, there are few clinicians and researchers received a formal training in informatics coding, data analysis, or other increaingly relevant skills to handle very large information. Therefore, it is very important to plan to training more staffs. Furthermore, when we implement follow the plan we should do the risk assessment and always improve the plan and mornitor the situation. As a consequence of this, if we do follow the plan, I hope we will have a high quality results with “least impact”.
-
Missing Data
– We need to learn and educate on what can Big Data do, and cannot; instead of either relying on wild assumptions or mindlessly collecting everything. Good education will help health personnel cooperate.Selection Bias
– Trends toward higher quality Big Data will help mitigate this.Data Analysis and Training
– This is the real problem here, as we might also need more research into Big Data. Even so, relying on training is not enough. Explanation of requirements in Plain English is as important. I saw ELI5 everywhere on the internet.Privacy and Ethical Issue
– Another real problem is whether we can have “Big Data” in the first place, if we cannot ethically collect the data. -
According to “BIG HEALTH DATA CHALLENGES” in this paper, there are different types of challenges facing the implementation. For the possible method of handling these issue as below:
1) Missing Data: This is one of the most common problems which can reduce the statistical power of a study and can produce biased estimates, leading to invalid conclusions. To prevent the problem by well-planning the study and collecting the data carefully. The following are suggested to minimize the amount of missing data in clinical research:
1.1) Should limit the collection of data to those who are participating in the study, collecting only the essential information at each visit.
1.2) Before the beginning of the clinical research, a detailed documentation of the study should be developed in the form of the manual of operations, which includes the methods to screen the participants, methods to communicate between the investigators or between the investigators and participants, implementation of the treatment, and procedure to collect, enter, and edit data.
1.3) Before the start of the participant enrollment, training should be conducted to instruct all personnel related to the study on all aspects of the study.
1.4) Should set priority targets for the unacceptable level of missing data. With these targets in mind, the data collection at each site should be monitored and reported in as close to real-time as possible during the course of the study.
1.5) If a patient decides to withdraw from the follow-up, the reasons for the withdrawal should be recorded for the subsequent analysis in the interpretation of the results.2) Selection Bisa: It is a kind of error that occurs when the researcher decides who is going to be studied. It is usually associated with research where the selection of participants isn’t random.
2.1 Try to make a study representative by including as many people as possible.
2.2. Using random methods when selecting subgroups from populations.
2.3 Ensuring that the subgroups selected are equivalent to the population at large in terms of their key characteristics.3) Data Analysis and Training:
3.1 Prepare dataset what suitable for study.
3.2 Train to upskill and knowledge for the researcher, for example, data science, statistics, program usage, database management etc.4) Interpretation and Translational Applicability of Results:
4.1 Prepare the adequate description of the dataset variables and associated metadata.
4.2 Prepare and control the quality of data which use in research.5) Privacy and Ethical Issue:
5.1 For personal data, it must be collected in a lawful and fairway for a purpose directly related to a function/activity of the data user.
5.2 Information on a patient should be released to others only with the patient’s permission or allowed by law.
5.3. Information shared as a result of clinical interaction is considered confidential and must be protected.
5.4 To train research staff about privacy and confidentiality issues that may arise when collecting data. The research staff should not open drawers or closets without permission and should collect samples only in pre-approved, designated areas.
-
You must be logged in to reply to this topic. Login here