[MUSIC] Hello, to everyone and welcome back. We have covered a lot of topics in the past modules and lessons. I hope all of you have gained new insights about how to understand processes associated with healthcare and how these in turn can lead to data that can be extracted and transformed into actionable information. There are huge opportunities to apply analytics to big problems. This involve scientific medical innovations, public health interventions to prevent diseases, and analytical processes within healthcare organizations to make healthcare cheaper and safer. In this lesson, I will review a few topics that you have already heard in various forms within the past lessons. I think these ideas are so fundamental and critical that it makes the repetition worthwhile. After the lesson, you will be able to articulate what is meant by the general phrase, context matters when analyzing and interpreting healthcare data. Moreover, you will be able to communicate specific questions on ideas that will help you and others on your analytical team understand the meaning of your data. Let's drive in. Hopefully by now, you understand the essential point made about data. Just like words, data cannot always be taken literally. People must work hard to see the context of the data. In working to see the context and garner the meaning of the data, you need to ask yourself these important questions. What was the purpose of collecting or creating the data? Who created the data and what is their role in their organization? Are all the data creators similar? What types of tools or applications were used to collect the data? Did the data collection tools, such as forms, have different formats or versions that might have led to different data types? How were the data stored? Did the data storage process have Extract Transform Load or ETL processes that might have altered the original data? Were the data altered after being collected? Did the data collection processes change over time. What artifacts were left behind that might help answer some of these questions? Is it possible to talk to people who are or were part of the data collection process? Some data processes such as healthcare clinical trials might have relatively straightforward answers. This is because clinical trials have many government regulations to ensure the data are collected and stored in a rigorous manner. However, I found that even research projects with strict data collection protocols can have various ways in which the data ends up being collected. A nonstandard example is collection of EHR data in a typical hospital. Although, you might think many of the questions above are relatively simple, I have been shocked at how much variation I've seen. In sum, ask good questions to challenge your assumptions about the meaning of the data. You may have heard GIGO principle, G-I-G-O stands for garbage in, garbage out. If you don't understand or you can't document the data that is being captured, you may end up with garbage. Without good processes and data management standards, you're not going to be able to make any meaningful sense of the data. That does not mean that data cannot be processed, reports cannot be generated, and knowledge cannot be inferred from this data. It just means it's not going to be easy to transform data into actual information. And you might not be able to get reliable and valid information from the data. Moreover, data quality paired with the knowledge of the analyst gathering the data is what determines whether you are going to be able to gain valuable information. If not, you might simply be generating numbers, charts some pretty pictures, but don't tell you what you need to know. While highly sophisticated clinical analytical environments do exist at some institutions, these tend to be the exception and not the rule. If you start with sets of questions such as the ones I've provided as examples above, you'll remind yourself about how to obtain meaningful data. Here are a few common processes or behaviors that would be good to keep in mind. First, it is important to learn work flows and human processes that create data. As this been discussed in the course modules, clinical data are captured during clinical processes. Data are artifacts of how medical staff process information and work together to improve patient health. It is important to spend the effort required to understand the metadata that will make the raw data meaningful. Second, spent the effort on documentation and communication of complex concepts. Healthcare is an extremely complex topic and medicine has evolved into a huge number of concepts that often are only understood by specialist. Thus, we should not be surprised that healthcare data is so complex. Third, it is important to work in teams and learn from domain experts. You will not have to be an expert in all domains that you work with. But you will need to know enough, so that you can effectively know when you need to partner with domain experts. In addition, knowing some aspects of various domains will help you better communicate when working on your analytical projects. In general, people who can write and talk about these complex topics should be valued. Finally, it is important to understand the data change is often tied to ever-changing technology and associated social systems. It is worth remembering that technology is a product of cumulative cultural evolution. People build off previously created ideas and tools. There are rewards from creating better and more efficient technology. Therefore, companies, inventors, and others are always looking to create new technology, or more likely, build off preexisting technology. Thus, it is important to appreciate the technology changes quickly. This is especially true in a rapidly evolving area of medical informatics. Also, even if we remember to look for and appreciate technological changes, we may forget that social systems also evolve. Sometimes technology drives social change. For example, new medical technology might create new procedures and new ways to treat patients. These can then lead to new workflows and manage a processes to augment new procedures. At other times, social change may drive technology. This is when people form new preferences about how to live their lives and manage their companies. And this then drives people to come up with new technology to augment these new behaviors. In the medical world, the past generation of physicians who were historically more independent and autonomous are becoming more integrated into cooperative medical teams. As a result, this new cohort of doctors might be more likely to demand communication software to help them perform these more collaborative medical treatments. Overall, one of the most important factors of all might be to notice how technology is integrated into social systems. However, often this is the least appreciated topic. For example, even though we think about electronic health record systems as physical computers and software. That has interfaces to show us charts and notes. The most important part of the EHR's are the actual patients within the charts. Now, we can't forget the doctors and medical staff who painstakingly document notes and click buttons while they listen to patients explaining their conditions. So really, the EHR is a system that is very closely tied to people and people who are often in painful and challenging situations. With that example, you might agree that technology itself does not solve problems, people using technology in effective ways can improve health access, improve the quality of care, and of course, lower costs. Okay, excellent, this completes our module and our course. I sincerely hope that you've gained insights about healthcare data. And this will help you become better healthcare data analysts, thank you very much.
