This application has a good purpose and the team has given some examples to support its motivation and help propose its targeted users. Progressive solutions, i.e. intermediate solution and proposed solution to achieve eventual significant aim were introduced expressively. Sysem Architecture diagram presented an interesting user interface. Demoed bar graphs conveyed an effective and attractive representation of data which expected users very likely want to see. PDF generator is also a considerate mechanism for satisfying users' needs. Future extention of this application and the possibility of its integration with other information systems has been considered. Overall, this application has completely considered necessary parts of all the important issues of its kind, including Security and Privacy. The presentation was well-organized and expressed each topic clearly. The goal and beneficiary of this application was featured all through the presentation along with the explanation of applied technology. This is a goal-directed application.
The presentation should have given more background knowledge to the audience to understand the problem this application attempts to solve. Detailed comparison to other systems in the market is needed to convince the sponsor the application's unique features and advantages. Security and Privacy issue was mentioned but implementation solution has not yet been provided. The application has chosen LAMP system architecture, which may not be portable in terms of its future integration with other information systems and may not be so secure compared to using some other language such as Java.
Overall, this demo was very impressive. The web site layout was very clean and easy to read. The use of the "tab" navigation bar was quite appropriate for a medical record system. The bar-graph graphics for displaying search results will be useful for users when they interpret the data. Generating PDF reports is also beneficial for distributing information to people outside the Healthy Data system. I liked the use of XML for marking up data, which will more easily allow integration with other hospital systems in the future. As well, the design of data-analysis searches was structured into the database in such a way so that new searches could be easily created and displayed on the web site for a client's use.
I had only minor suggestions for enhancing your current system. There seemed to be some redundancy of "edit" buttons for displayed records. Perhaps a user can pick a record first and then be given the option to edit that one record. Regarding the bar graph, the titles appearing along the horizontal axis were each displayed vertically, making them hard to read (without tilting your head sideways). These titles should be displayed horizontally. Finally, it appears that new SQL-searches can be added to the database and quickly displayed for a client to use. But the addition of new searches seems to be handled through the system's back end by an administrator only. It would be beneficial if users could create their own data-analysis queries based on available data choices.
I admire the group's effort on security and privacy concern. I agree that it is very important to avoid personal information storage and transfer as much as possible. Another part I like the project is the pdf creation. It is a very good idea since PDF is a very well-known cross-platform read-only file format. Moreover, the dynamic graph is also very impressive.
The group's project is pretty good overall beside the SQL query part using phpMyAdmin. First of all, the phpMyAdmin interface does not consistent with the project page as this is another party's product. Secondly, using SQL requires SQL syntax knowledge which the client user may not have. An suggestion to solve this is to implement your own query form so the appearance consistent with other part of the project, and the form should be more user friendly (fill-in the blank questions) instead of asking the user to type in SQL query.
In this project, they did a lot of research on the current issue. They provide a lot of information about the current situation. Their idea is very good, and with a clear architecture design. They use many technologies to develop their application. They make many considerations for deciding which technique to use. Also, they did show what each component does and how it works. In their application, they use many different technologies to generate report. Finally, they provide a lot of solutions for the security issues and compare each other.
My recommendation is that they can try to integrate the hospital system to their system. In this way, all information can be entered automatically. Also, they can try to combine the administration into their application. In this way, hospital can just add whatever function (e.g search) they want; instead of asking the administer to add the function for them. Distributing database is another consideration for handling large data. There is an issue for the data that hospital might not want to show all their data to the others. They might provide some wrong/not clear information to confuse others or hide their mistake from the other. Integrating the hospital system with their application might be able to solve this problem.
First of all, finding possible medical problems (such as adverse drug effects) in order to save lives is a very worthy goal. I really like the ability to generate bar graphs in order to visually spot potential drug problems. Especially useful (and neat) is the ability to perform drill-down in order to do a more thorough analysis of the problem area. The option to generate a pdf version of the bar graph is another nice touch. I found the security aspect of the system to be very well thought out. Firewalls, data encryption, user authentication, logging, etc. were all addressed. In addition, the 'future vision' of having data mining could be a useful way to find trends in medical problems.
With this system, the larger the user base, the more useful it is. Unfortunately, a 'catch-22' may occur where there's very few hospitals participating initially (which results in few data), and this makes it less appealing for other hospitals to participate. One way to ensure a large initial user base is to create a sample tutorial page. This can be effective in persuading some of the hospitals that the system is useful and easy-to-use. As a side note, there might be a more user-friendly approach in creating the SQL search queries. Right now, it requires the user to manually type in the SQL query. Perhaps a drag-n-drop approach might be easier to use (where users drags and drops the table attributes that they are interested in). Also, it seems like there's only an 'Edit' button, and no 'View' button when it comes to displaying information. It could be slightly dangerous since some users simply want to view, but may unintentionally change data. So, a 'View' button should be given.
The idea of collaborating life saving information between hospitals and making the results available on the internet seems so important yet so simple, it's surprising the idea has not already been implemented. A particularly strong point of the Healthy Data system is its portability. The use of a centralized database employing web based technologies and Adobe Acrobat files will allow the project to reach all those interested, while fueling and supporting its growth. The current use of Microsoft Excel and Access files are most likely only available to those within the hospital offices and not portable even within the Windows operating system (due to version changes, and availability of the programs). The Microsoft Office line of products is expensive, especially to outfit an entire hospital of computers. The cost of supporting the current system may prevent it from being widely accepted across hospitals. By allowing files to be exported in PDF format, graphical, standalone documents can easily be exchanged making the findings of the data more accessible and presentable. In addition to its portability, the project boasts impressive extendibility. Clearly a lot of time was spent in creating an extendible design for the project. Being able to add new queries, and compile new searches without any real understanding of the system will ensure that it will remain in use and can be continually improved with ease. Another design strategy for extendibility is the decision to incorporate an XML interface to the project. Since XML is a very simple way to markup data, applications which want to make use of the project's resources, such as data mining (which they mentioned), will have an easier time creating a layer to interface with the project. Even if the portability were not enough to maintain the projects popularity, surely the ease of extension to new applications would. In fact being a centrally located data store on the internet, leaves the possiblity of extending the system to hospitals all over the world, yeilding more data for analysis. Even without reaching the far corners of the earth, this system has clear advantages over the current one, which translates to better patient care for everyone.
Though the project is an improvement to existing solutions for medicating patients, there are a few issues not addressed in the proposal. As a database of statistics one wonders why these statistics are not exploited more. One area which would help the project gain more use, would be to create an expert system or neural network which would help associate problems and side effects with medications. By consulting one of these systems before administering drugs, doctors could help narrow down some possible problems without having to do an exhaustive search of the records, which in all like and analysis. The current format in Excel files is probably done specifically for this reason. It's one thing to look at a bunch of numbers and try to make sense of them, it's another to make use of statistics to help guide you. Another application of processing the data in some form, would allow hospital staff to compile better presentations for meetings. If a particular drug comes under review, staff could quickly put together a presentation to support or withdraw use of that drug. The histograms are a step in the right direction, but more use of inference techniques such as those mentioned above, are needed to make the project more than just a database. Finally the anonymity of the system seems appropriate to let the public have access to the information or some portion of it. There are many people, elderly in particular, who routinely take combinations of medications and could benefit from access to the same data as the hospital. Overall the project has made improvements that will encourage use of the system. However, to ensure that the project is more than just a database, some extension to data presentation and manipulation needs to be addressed.
The presentation of HEALTHY DATA is well organized and informative. From this presentation, we can see several attractive points of this project. First, this is a very good idea to provide a web based information system for organizations to share valuable data. Because it is easy to access web information system and no special software requirement on client side, the convenience provided by HEALTHY DATA is apparent. Second, the presentation paid enough attention on security and privacy part. To share sensitive data, protecting privacy and security is crucial. It is the main factor affects the will to share data. Third, the interface design is neat and seems easy to use, especially for the function of inputting and searching data.
However, I only see data input, search and bar graph display in this presentation. If this project can put some effort on designing prototype of how to integrate with other systems in a hospital and how to perform data mining, it will be more attractive. Moreover, to achieve better protection on security and privacy, I think the system can limit client's IP address or domain. My final thought is about the bar graph. By my understanding, the correlation between medicines use and causes of death is very subtle and complicated. A bar graph view is an excellent tool to present data, but, without good statistic models, the merit of bar graph will become drawback - misleading. Therefore, good statistic models should be introduced in creating bar graph function to avoid misleading.
It is a good idea not to have patient information on the web. The software may be useful for health official in keeping track of what combination of of symptoms and drugs are fatal. The ease of use is a good feature. The exyensive use of middle-ware made the system look quite proffesional.
There is a security risk. This is because any one who logs into the system can make changes to the data. I am wondering if there is a similar system around.
Healthy Data is going to provide a much better solution for recording medical data of drugs compared to the current system in which they are using. From the presentation, I have learned that the current system uses a client application which is then transfered manually from the client application through FTP via an MS Access file, which is then appended into the ISMP server. Then the data sent to the server is then manually appended to the master MS Access file. Then the master MS Access file is converted into an MS Excel file to the hospital. No wonder the hospitals do not use this system. It's not user-friendly, especially for doctors and nurses how have other priorities then to mess around with a computer application. None of the data is actually networked and data may be not as accureate since the master file residing at the hospital may not be updated. I am surprised that they had even built a system which requires so many data conversions. I am interested to see the comparison of how many people would luse the current system versus how many people would use the proposed system.
The overall user interface was well done. Colors and the navigation of the system seems to be easy to use. All the pages are straight forward knowing where to enter data. The feature of providing graphs and charts is a great idea as "a picture is worth a thousand words" and it's good for comparisons.
The feature for administrators to be able to update search functions and the actual data tables/records is a great asset to the system. Administrators need not know too much programming to add/modify/remove the data tables and search functions available.
The main goal of HealthyData is to provide a centralized place for storing and accessing information regarding deaths of using certain drugs. Since the information deals with human lives, it is important that the stored data are as accurate as possible. The HealthyData system can provide better accuracy by not having medical staff enter the information manually as it is now. I like the idea of running the system on Apache in Linux because they tend to be more robust than other servers. Another appealing feature of the system is the ability to edit/add queries without changing the implementation code. The administrator simply goes into PHPmyAdmin and edit the query statement; This will definitely speed up the process of fullfilling requests by doctors. When it comes to security, I believe the group has done a good job in making certain that only authorized personnel are allowed to access the system. IP logging and allowing only IP addresses from hospitals are particularly interesting.
An interesting point I want to make is that since all the data in the database are anonymous, why is there a need for tight security? I think what would make this system more useful is to allow the public to have restricted access (eg: only view data) to the database. Since none of the information containts any patient or doctor names, there is no violation of privacy here. In the demo, when the presenter searched for a particular drug, each record was listed separtely. I think it would be better if all the records of the same drug are shown together on a single page. This allows the user to view all the records at once without having to "go back" to the previous page and clicking on the next record. My final comment is that HealthyData should be integrated with the PharmaNet system that is being used by all Pharmacists in BC. This way, before pharmacists fill a prescription they can check with the system to make sure that the prescribed drugs will not cause any complications to the patient.
The proposed Healthy Data web information system has explored its application domain thoroughly. The use of PDF and XML demonstrate the understanding that every hospital uses different platforms and applications. Healthy Data also recognizes the importance of patient confidentiality and takes considerable measures to ensure that their privacy is upheld. To guarantee privacy, Healthy Data has adopted preventative measures so it does not store or receive any information that can identify a patient. The ability for an administrator to create ad hoc queries is a valuable feature. This creates flexibility and allows hospitals to customize the application. The features of the Healthy Data system indicate a clear understanding and comprehensive analysis of the Hospital Information System domain.
The success of the system depends on the number of participants. Currently, the proposed system is only used for hospitals. Data could also be gathered from pharmacist databases. Pharmacists also keep track of patients and the medication they are taking. This faucet should be explored to enrich the quantity and quality of data in the Healthy Data Warehouse. The ISMP should try to bridge an alliance between hospitals and pharmacists. User access should also be broadened to pharmacists and doctors who work in a practice. The most current data on prescribing medication should be available to doctors and pharmacists to ensure that the patient gets optimal care.
This could be a good and useful system if a lot of hospitals are involved in this. It could be very useful for researchers because there was the feature of making a graph for easier analysis of data and the generation of pdf files so you could export the data. In addition, the privacy issue was dealt well. Not including hospital names or personal patient information was a reasonable idea, because obviously hospitals are not going to let others know about their mistakes. Leaving out this information could make hospitals more accepting of the notion of divulging information about their mistakes and failures. Plus, the point that was made about not having personal information leaked out in the event of the system being hacked was a good one to make. Another appealing aspect of the project was analysis of the present situation and having a intermediate and proposed solution because the present situation of tracking the errors looked to be too cumbersome and time-consuming. Entering, sending and viewing data through a web browser and having the information updated automatically is obviously much better than the current implementation.
One issue is that some hospitals may not see the need to look at statistics of all hospitals to see the need to fix their own mistakes. They may not see the motivation to fix errors based on the fact that many hospitals see an error ocurring because if there are mistakes occurring in your own hospital, don't you need to correct the problem anyways, and not because it is happening a lot elsewhere? I remember if there was any other way to look at errors other than drug name, but maybe there should be other ways of searching such as searching by surgical mistakes (dirty surgical tools, wrong blood type organs, cutting the wrong area) or patient illness (diabetes, AIDs, HIV, cancer). Also, there may be problems in people entering the cause of the error. For example, a nurse may have given an overdose of a certain drug, but the real cause may have been the hospital over-working their staff leading to mistakes and slips due to lack of sleep. Also, I don't know if there was any way of accounting for multiple causes that led to some bad outcome. Is that going to be in a category all by itself? There could also be a way to enter a possible solution to fix a certain problem, so that way a hospital seeking a way to deal with a problem can have different perspectives on how to do it.
| Good Features | The tracing and monitoring the efficiency of certain drug is important, such that it provides feedback for medical researchers who are looking for ways to guard human health against various diseases. The program can be a very useful statistic generated tool. It can help scientists to collect data, observe error trend, and link the drag usage to certain disease. In other words, the program provides persuasive data proof in evaluating drugs usage. Moreover, it can also aid to direct medical research and reveal some unexpected side effects that have never been considered before. |
| Sugession | However,some issues are worth considered here are: how the data can be further generate meaningful information, and how different systems are integrated together. As we can see the causes of death are various. The factors affect human health conditions are including: disease, emotion, certain weather condition (as we might consider the temperature might have some effect relate to body condition), living condition (sometimes the chemical level of the water the patient use might have significant effect), etc. When combining all the factors together, the result will be more persuasive. As regarding to the system each individual medical organization using, there are reasons why the hospital applies one system but not the other. To integrate different systems together, the application designer have to worry how the integration among different platforms and how the languages can be used on top of different systems, since the issue may limit the numberof participants. The more the participants and the bigger the samples sizes, we will expect the more precise the result are. |
A well-organized construction of an information collection system that presents some useful functionality, this project is packed with great features. Its ability to generate a pdf format to print on the fly is a good feature for producing reports. Also, the correlation graphs between errors and possible causes makes for a graphical presentation that would be useful in detecting possible trends of causes as a result of erroneous prescription of drugs. The flexibility in allowing the user to change the database tables to include other information is added bonus. Every measure seems to be taken in ensuring a secure system with firewall, data encryption, user authentication to protect the patient's privacy.
The idea of this system is to improve communication between doctors and pharmacist and collecting information that may be useful to preventing errors. The security issues were covered with the use of passwords and logins, but I would have to question the integrity of users who make mistakes that will follow up by entering such information into a database. I perceive this system to be used by a smaller number of administrators who audit the work of the employees at a hospital. Also, the correlation of data perhaps should not be accessible to all users, since that may bias the participants entries into the database. Take for examle a case where people may see a specific correlation and are somehow convinced that this correlation is true. This self-fulfilling prophecy may lead to inaccurate information being collected, especially if all employees confirm each others beliefs and obscure the facts. Although, the information collected is meant to be transparent to all, this very fact may not contribute to collecting accurate information. I suggest that this information system be used for users to enter data, but not have access to viewing the data correlation and the anaylsis. Overall, the system has great potential.
This is a great presentation, since it first points out problem on the current implementation, as well as providing proposed implementation. Knowing that changing a computing system has never been easy, it also presents an intermediate implementation. It is very good that even in the intermediate implementation, security has already improved. A very insightful idea from the presentation is to enable external data mining. This should provide strong support to medical research and drug development.
Providing multi-stage changes to computing system has two sides. It allow early feedback, get to the final solution earlier, and providing a good base ground the system can fall back to in case of any critical issue. However, since people work in hospital are generally ordinary computer users with special medical domain knowledge only; making them to go through a intermediate system basically means forcing them to learn another new program. From the hospital point of view, the differences between the proposed design and the intermediate design do not seem to be significant enough. An alternative suggestion is to go straight from the current design to the final one directly. Even this action would cause longer development time before any improvement happens, but taking this extra time to provide better training and testing should overcome this lose. All medical system is life-critical system. Although this system does not directly cause life lose, but it can greatly affect the research result and medication decision made. The final words are the final web interface should look somewhat similar to existing interface. This can greatly reduce the number of errors the user made.
Web-based system for medication error tracking has great potentials. Data mining tools can be applied to the aggregate data from different sources to find explanations and patterns for medication errors. Health care providers can establish better practices based on these findings. Policy makers can make use of these findings when introducing new drug product standards. These findings are also invaluable to pharmaceutical companies in their product development. The proposed system has substantial improvement over the existing system described in the presentation. In the proposed system, data are collected through a web browser directly into database. This allows data collecting organization (ISMP) to produce timely special alerts that go to all hospitals to prevent recurrence of the same errors. Using XML as transport medium between client and server is a good decision. It allows data further exported to external data mining centers that may use different technologies. The designers of the system stress the importance of system security, and grant hierarchical access privileges to different users. This indicates that the designers are fully aware of the special nature of the system.
There are several important issues that are not addressed by the system. ISMP is a voluntary medication error reporting system. To provide information that leads to improved medication safety should be the mission of the organization. Because holding health care provider accountable is not the purpose of the system, the system should provide the option for anonymous reporting of medication errors. This feature is not provided by the proposed system, since users must login to report medication errors. The designers do not address legal issues. Can the data submitted be used as evidence in legal proceeding? Seeking legal protection from the government is an important issue for the system. If this issue is not resolved, it's difficult to promote the system to health care providers. Last but not least, in order to collect adequate information to analyze medication errors, care should be taken to design the fields included in the error reporting form. If there are existing standards to categorize medication errors, the standards should be followed. Otherwise this should be done with the help of hospital associations.
What I found most appealing about this project is it envisions providing a medium for hospitals to cooperate and aggregate all their data. This is like having a patient being diagnosed by all the doctors across Canada. I like how by using data mining techniques and generating statistical charts, information can be derived that can serve immediate practicality in making life-saving decisions. I also approve of the effort of the team in making this product as customizable as possible. It has the power to allow staff at the hospital to generate their own custom SQL queries, thus extending the usefulness of this system.
In addition to their fast-paced work and with the amount of information generated by this new system, one has to wonder whether this can overwhelm medical professionals. Though this system intends to aggregate data into a central repository, the doctor still needs to treat each patient on a case-by-case basis. Each person is unique and the professional judgment of the doctor is still needed to ensure the patient is receiving the best possible care. Despite these issues, this project serves as an excellent analysis tool and will be well received in the medical community.
Medical Error Database
In terms of the big picture, Health Data tries to automate the integration of hospital information transmitted from Analyze-Err into the ISMP server while allowing authorized personnel to analyze and view those aggregate medical data. By sharing aggregate data, hospital personnel may extract knowledge and information generated from other hospitals. This would achieve the pinnacle of information sharing in the medical field. Examples of benefits include avoidances of medical errors and exposures of diagnosis made by other hospitals. All transmitted data is encoded in XML format. This ensures that the server would be able to read the data regardless the applications involved. Also, transmitting data in XML provides the flexibility of integrating with other systems which would be greatly desirable. In terms of implementation, the concept of drill down supports complemented with the easily understood graph bars are great ways that help users to perform searches effectively and efficiently. For example, there is a graph that shows number of errors vs. possible causes. By looking at the graph, users can immediately detect the most common types of errors or possible causes. Hospital personnel can start to draw closer attentions to avoidable errors that persist incessantly. The system also allows users to convert graphs into PDF reports which is a convenient feature for medical personnel.
Despite of many of the great features found in the system, users may confront the problem of information overloading when using this system. When a query is performed, the system just outputs all relevant results in a long list. This is not a problem if the output list is short. However, if the list is long, which can be expected for medical information, users may be overwhelmed. This problem can be solved by categorizing outputs such that users can narrow down the query results and draw better analysis and make better decisions. Also, security issues must be addressed; the confidential medical data must be protected. The developing group suggested several approaches in addressing this matter: firewall, data encryption, user authentication logging, and privacy of patient identity, etc. However, these methods can only protect the data from unauthorized accesses or interceptions of data; they don't prevent authorized users from using the data illegitimately. In order to prevent such cases, the system should leave an inerasable trial for each access.