Delving into the details

 

Editor’s note: The Society of Hospital Medicine’s (SHM’s) Physician in Training Committee launched a scholarship program in 2015 for medical students to help transform health care and revolutionize patient care. The program has been expanded for the 2017-2018 year, offering two options for students to receive funding and engage in scholarly work during their first, second, and third years of medical school. As a part of the longitudinal (18-month) program, recipients are required to write about their experience on a monthly basis.

For my research project, we are looking to develop a tool that would use data from within 24 hours of a patient’s admission to the hospital to predict whether they will require post-acute care placement after discharge. While I have often been summarizing my project with this broad one-liner, in the last two weeks I have been delving more into the details of what exactly we mean by “data from within 24 hours of a patient’s admission.”

We have access to a large set of de-identified patient data from our institution, from which we are going to construct this model. However, it contains vast amounts of information about every patient’s hospital stay, and we only need a subset of that information. Making detailed decisions about which lab values, vital signs, and other information is most relevant will take some careful parsing. For example, for some lab values, we are looking to get the highest, lowest, and the median value to make sure we have a picture of the patient’s status in the first 24 hours that would be much more informative than any value alone. Others may not have enough data points to often collect three times in the first 24 hours, and so first and last may be more appropriate. Others still may not be recorded correctly in the database we have often enough to be a reliable piece of information to use in the analysis.

We are going through each of the variables systematically to take into account prior literature on how they were treated in other studies, as well as the practical limitations imposed by the data-gathering within our own system to choose how these values will be selected for each admission. My mentor Dr. Eduard Vasilevskis is helping me with making these decisions, based on the prototype model that was the inspiration for this project. Once we have identified all of the details of each variable we want to track, Dr. Jesse Ehrenfeld will be facilitating our use of the database.

Certainly this project has helped illuminate not only research-specific hurdles, but also underscores the fundamental difficulty of clinical decision-making in the first 24 hours of a patient’s admission. With data changing rapidly and sometimes incomplete data, clinicians need to quickly make care decisions that can impact a lot more than the patient’s post-discharge destination.

We anticipate that once we’ve made these choices, there will be further choices to make about how to treat these variables in the analysis. We hope to have the assistance of an experienced statistician to help guide us in making those decisions.

Monisha Bhatia, a native of Nashville, Tenn., is a fourth-year medical student at Vanderbilt University in Nashville. She is hoping to pursue either a residency in internal medicine or a combined internal medicine/emergency medicine program. Prior to medical school, she completed a JD/MPH program at Boston University, and she hopes to use her legal training in working with regulatory authorities to improve access to health care for all Americans.

 

Editor’s note: The Society of Hospital Medicine’s (SHM’s) Physician in Training Committee launched a scholarship program in 2015 for medical students to help transform health care and revolutionize patient care. The program has been expanded for the 2017-2018 year, offering two options for students to receive funding and engage in scholarly work during their first, second, and third years of medical school. As a part of the longitudinal (18-month) program, recipients are required to write about their experience on a monthly basis.

For my research project, we are looking to develop a tool that would use data from within 24 hours of a patient’s admission to the hospital to predict whether they will require post-acute care placement after discharge. While I have often been summarizing my project with this broad one-liner, in the last two weeks I have been delving more into the details of what exactly we mean by “data from within 24 hours of a patient’s admission.”

We have access to a large set of de-identified patient data from our institution, from which we are going to construct this model. However, it contains vast amounts of information about every patient’s hospital stay, and we only need a subset of that information. Making detailed decisions about which lab values, vital signs, and other information is most relevant will take some careful parsing. For example, for some lab values, we are looking to get the highest, lowest, and the median value to make sure we have a picture of the patient’s status in the first 24 hours that would be much more informative than any value alone. Others may not have enough data points to often collect three times in the first 24 hours, and so first and last may be more appropriate. Others still may not be recorded correctly in the database we have often enough to be a reliable piece of information to use in the analysis.

We are going through each of the variables systematically to take into account prior literature on how they were treated in other studies, as well as the practical limitations imposed by the data-gathering within our own system to choose how these values will be selected for each admission. My mentor Dr. Eduard Vasilevskis is helping me with making these decisions, based on the prototype model that was the inspiration for this project. Once we have identified all of the details of each variable we want to track, Dr. Jesse Ehrenfeld will be facilitating our use of the database.

Certainly this project has helped illuminate not only research-specific hurdles, but also underscores the fundamental difficulty of clinical decision-making in the first 24 hours of a patient’s admission. With data changing rapidly and sometimes incomplete data, clinicians need to quickly make care decisions that can impact a lot more than the patient’s post-discharge destination.

We anticipate that once we’ve made these choices, there will be further choices to make about how to treat these variables in the analysis. We hope to have the assistance of an experienced statistician to help guide us in making those decisions.

Monisha Bhatia, a native of Nashville, Tenn., is a fourth-year medical student at Vanderbilt University in Nashville. She is hoping to pursue either a residency in internal medicine or a combined internal medicine/emergency medicine program. Prior to medical school, she completed a JD/MPH program at Boston University, and she hopes to use her legal training in working with regulatory authorities to improve access to health care for all Americans.

 

Editor’s note: The Society of Hospital Medicine’s (SHM’s) Physician in Training Committee launched a scholarship program in 2015 for medical students to help transform health care and revolutionize patient care. The program has been expanded for the 2017-2018 year, offering two options for students to receive funding and engage in scholarly work during their first, second, and third years of medical school. As a part of the longitudinal (18-month) program, recipients are required to write about their experience on a monthly basis.

For my research project, we are looking to develop a tool that would use data from within 24 hours of a patient’s admission to the hospital to predict whether they will require post-acute care placement after discharge. While I have often been summarizing my project with this broad one-liner, in the last two weeks I have been delving more into the details of what exactly we mean by “data from within 24 hours of a patient’s admission.”

We have access to a large set of de-identified patient data from our institution, from which we are going to construct this model. However, it contains vast amounts of information about every patient’s hospital stay, and we only need a subset of that information. Making detailed decisions about which lab values, vital signs, and other information is most relevant will take some careful parsing. For example, for some lab values, we are looking to get the highest, lowest, and the median value to make sure we have a picture of the patient’s status in the first 24 hours that would be much more informative than any value alone. Others may not have enough data points to often collect three times in the first 24 hours, and so first and last may be more appropriate. Others still may not be recorded correctly in the database we have often enough to be a reliable piece of information to use in the analysis.

We are going through each of the variables systematically to take into account prior literature on how they were treated in other studies, as well as the practical limitations imposed by the data-gathering within our own system to choose how these values will be selected for each admission. My mentor Dr. Eduard Vasilevskis is helping me with making these decisions, based on the prototype model that was the inspiration for this project. Once we have identified all of the details of each variable we want to track, Dr. Jesse Ehrenfeld will be facilitating our use of the database.

Certainly this project has helped illuminate not only research-specific hurdles, but also underscores the fundamental difficulty of clinical decision-making in the first 24 hours of a patient’s admission. With data changing rapidly and sometimes incomplete data, clinicians need to quickly make care decisions that can impact a lot more than the patient’s post-discharge destination.

We anticipate that once we’ve made these choices, there will be further choices to make about how to treat these variables in the analysis. We hope to have the assistance of an experienced statistician to help guide us in making those decisions.

Monisha Bhatia, a native of Nashville, Tenn., is a fourth-year medical student at Vanderbilt University in Nashville. She is hoping to pursue either a residency in internal medicine or a combined internal medicine/emergency medicine program. Prior to medical school, she completed a JD/MPH program at Boston University, and she hopes to use her legal training in working with regulatory authorities to improve access to health care for all Americans.