- This topic has 16 replies, 12 voices, and was last updated 1 year, 6 months ago by
.
-
Posts
-
-
-
– Steps in the data management workflow:
A data management workflow for a research project consists of six steps: planning and design, collect, generate and store, clean, analyze and visualize, manage, store and preserve, share and publish, discover, reuse and cite. We have done some of these steps in our household survey, such as planning and design, collect, generate and store, manage, store and preserve, share and publish. However, we have not done some other steps, such as clean the data systematically, analyze using statistical methods and visualize using Google Studio and we didn’t publish in a peer-reviewed journal because we could not convinced the ethnic health leaders about the importance of publishing.
– Steps that we should have done to improve our project: To improve our project, we should have done the following steps:
– Clean our data systematically by checking for errors, inconsistencies, and missing values, and correcting or removing them as appropriate as well as work with staff who collected the data to make sure data consistency. This would improve the accuracy and reliability of our data.
– Analyze and visualize our data by using appropriate statistical methods and tools to explore patterns, trends, relationships, and outliers in our data. This would help us answer our research questions and test our hypotheses. We only wanted to do a baseline study at that time and we didn’t have enough resources to properly analyzed the data.
– Discover, reuse and cite other relevant data sources that could complement or support our own data. This would help us contextualize our findings and compare them with existing knowledge, in other words, publish in a peer review journal would help us turning this data into a scientific evidence. -
Based on my experience in data collection/management, I have done many projects related to data analysis and database management that cover most of the steps in the lecture. However, each work did not include all of the processes due to the different specifications of the task. The steps which I never do is Data Quality Control and SAE Reconciliation since I mostly work with secondary data and never work with clinical data.
Talking about the step I have learned that could improve my project, I will raise an example of feedback data gathering described in the previous section. The project contains several steps in data management workflow including, protocol preparation, Data design, Data acquisition, CRF development, DMP, Data Entry, Data analysis, and Study Report. As the amount of feedback was on a small scale, the database-related process is not required. However, there were some problems that occurred after the information-gathering process where some information was not clear and unusable. I would say that the Quality Control step would be a great step to include in the work to cross-check the integrity of the information gathered during the interview and reduce potential errors.
-
-
From the discussion previous, I didn’t really have a plan on data management in my medicinal formulation research. Even though it didn’t involve personal data or interviews, a proper plan would have been better. If I could turn back the time, there would be a few steps which I could add to my data management to enhance work quality.
Firstly, better planning on data design and variables collected, part of the experiment was composed with different variables. Without a plan for variables, I sometimes figured the variables out in the morning before the experiment started which made the process slow and some variables were forgotten.
Secondly, I should have created a proper recording format. Overall, it was hard to accumulate fractional data and transfer this to the electronic system. The experiment would have been better if I had a standard format for the study.
Thirdly, Database Accessing – my lab partner and I didn’t have a database for sharing finished records. It was slow to wait for each other to edit and send data forward/backward. Moreover, if we wanted to investigate the result, data between researchers was fractional which made it hard to trace back the data. Therefore, a shared database should had been set and authorized for two researchers.
Lastly – data entry and quality control. During that time I didn’t complete checkins while doing the data entry between two researchers. It might have been more precise if we cross checked each others data entry to confirm the accuracy of recorded data.
-
Since the project used secondary data for collection and analysis. However, there are many steps that I haven’t done during data collection.
If I could go back, I would like to do data validation/ data quality control and data quality assurance were not done. SAE reconciliation, external data merging, database lock, or document archive (and most of processes) were not involved in the project since the project is a small scale and it is not a clinical trial.
The reasons that I focus on QC/ QA of data because an error from data collection could none or less impact the analysis and findings. If QC/ QA process had done, we would have a better accuracy, completeness, and reliability of both data collection, accuracy, and findings.
-
-
During my research, these are steps that I have done: protocol discussion, data design, data acquisition, database setup, data entry and processing, data validation, and data quality control. However, there are steps that I would like to take after finishing this course.
– CRF development. I should consider some issues that might occur during data entry. These issues must be written with a definition in the instruction.
– Database access control. As the project grows and there are many people involved in the project, access control with authentication must be enforced.
– Reconciliation of serious adverse events. I would set up a separate database that contains data on serious adverse events.
– External data merging. Since my previous research used data from the same electronic medical record, I did not need data merging. However, if I need data from different systems, I have to consider the data merging process.
– Database lock. After finishing the study, the database should be locked against further changes. -
-
Topic discussion 2
Based on your experience in data collection/management, which steps in the data management workflow that you have done and have not done?
Regrading to my project, it was a big data collection which consist of almost of workflow step such as Protocol dicussion, Data desing, Data acquisition (paper base), Data collection development, Data managment plan, Database access control, ect. Although I had experiance in all of steps, I had done only some part of step because there were many member jioned to work. For exsample, database setup that I also attented meeting but the one who did it was a programmer. So, I maybe can say that I have not done the edit checks programming, data entry and processing, study data archive and data base lock/security steps. My reseach had not done the data standard managment, SAE reconciliation and external data merging.
If you have a chance to go back, which steps that you should have done it to improve your project?
If I can improve the project, I will chang the CRF formate because it still doesn’t really work in all of steps or areas of study, resulting in other steps were diffical to manage. -
I recently started a new job as a clinical data associate. My main responsibility is to support the lead data manager in handling all data-related tasks of clinical research studies. While I have some experiences in the early stage/initiation process, I have yet to gain much experience during the conduct or study end. Instead of going back to improve previous projects, I will use all the steps to integrate standardized data coding and data management into my future work. To ensure that the data we work with is of the highest quality, one of the important process is the data quality assurance, which includes:
– Reviewing the trial protocol and ensuring that the trial is carried out according to plan (no PD/PV identification).
– Identifying trends in subject sampling and data.
– Reviewing measurements and staff compliance according to regulations to identify any areas of improvement.I believe that following these processes will help me perform my role effectively and efficiently and lead to better outcomes. Thank you very much.
-
According to data management workflow, 60-70% were done in my project, the steps that have not done or incomplete and need improvement;
For the project initiation;
CRF/DMP development/ Edit checks programming; we don’t have a full completion of instruction manual and official DMP. It’s just a final draft use among investigators. We don’t have edit check programming. We should write a complete instruction manual and for each research study involved in joined project should have DMP for their projects. For the large project and sample size, edit check program will improve the data quality.For study conduct;
Data entry and QC; for questionnaire part from teacher and parents, we only written the instruction at the head of the survey document. There are some missing surveys that we try to track back to the school. The QC steps is quite unprofessionally done, it’s investigator-dependent. We might need an in-person meeting separately with the teacher and talk about the questionnaire in details and we should have focus more on data QC steps by steps, especially the large study project.For study closure;
Data standards; We applied the standard criteria for risk definition based on particular subjects such as criteria for ADHD, for joint hyperlaxity. For research study like clinical trial, we should select and follow proper standards coding.
SAP: again we didn’t create the official SAP ahead, we actually think about it in details like table used after we get the data. We should think about SAP in details, so that the analysis part will be more precise to the question. -
Based on your experience in data collection/management, which steps in the data management workflow that you have done and have not done?
If you have a chance to go back, which steps that you should have done it to improve your project?Our research back then, it was our first research which we designed from the beginning.
In the study initiation phase, what we have done were
(1) protocol discussion with our group and our advisor, using previous research paper as guideline.
(2) data collection design (based on existing, validated questionnaire)
(3) data acquisition (paper-based CRF)However, without prior knowledge on how to systematically conduct and comply with the data management process, we did not explicitly declared on data management plan. We only collected paper-based CRF, enter them into online collaborative spreadsheets and into SPSS software. We also did not plan thoroughly about where to keep them. It was a mess back then when we don’t know which was the latest version of data, etc.
Luckily, because the scope of our research was not large, it was handled without any problems.(4) Database access control -> Like I have mentioned earlier, our data was digitised from paper-based into shared spreadsheets. We used a built-in tool which help us define who has the designated role to view, access or edit our data. Edit Check Type was handled manually.
Moving on to the next process, we had done both data entry/processing and data validation steps 🙂
In the project closure workflow, regarding our study protocol, SAE Reconciliation and external data merging were not mandate. After we finished entering data into digital spreadsheets, we downloaded them and use those data locally with our statistic software. The original CRFs collected were photographed, archived locally and disposed accordingly.
-
-
Based on your experience in data collection/management, which steps in the data management workflow that you have done and have not done?
In many projects, I was involved in Data Design & Acquisition, Data Collection/CRF development, Database access control, Data entry screen processing, CRF completion training, and Quality Assurance (QA), DBL.
If you have a chance to go back, which steps that you should have done it to improve your project?
If I got a chance to go back, to improve the project, we should have invested more time in data validation and Edit checks programming. During the study, we found many issues due to different data structures in each site and we had to amend the eCRF completion guide to add more clarification on the required data and additional notes to investigators.
-
You must be logged in to reply to this topic. Login here