Why longitudinal data is crucial to making better care decisions
Clinicians need to make better use of analytics and large datasets to determine the best interventions for treating patients’ diseases.
The challenges associated with transforming healthcare can be explained simply: The way that we make important healthcare decisions is fundamentally flawed. When facing a major life decision, even experienced physicians like myself will simply ask our spouse or a close friend for advice or recommendations, rather than conduct the type of rigorous research befitting a life-and-death decision.
The reason we do this is simple. When you begin to peel the onion, you quickly experience a data avalanche of disparate and disconnected data points, but nothing resembling “the truth” for your individual situation.
We lack a longitudinal patient dataset (data that track the same patients over the course of many years) that will guide us to make the best treatment decisions based on actual, real-life experience from patients similar to the individual needing care. For example, longitudinal data helps to answer the question, “I had my aortic valve replaced 15 years ago. Based on the experience of patients with medical histories similar to myself, what course of treatment should I follow when it begins to fail?”
Making data like this accessible in new ways has the potential to significantly impact the way healthcare is delivered. The power of a multivariate dataset, collected from hundreds or thousands of people with similar diseases or conditions over an extended period of time, enables hospitals and health systems to make high quality, cost-effective treatment decisions for their patients. In fact, the ultimate outcome of longitudinal data is to identify at-risk patients and intervene to preempt the disease from occurring in the first place.
Historically, hospitals and health systems have relied upon datasets that measure the incidence, but not the prevalence, of disease. Scientific studies are important; however, they're limited in that they're not reflective of the specific population at any one hospital, taking into account demographics such as age, income, education and environment of those particular patients.
To deliver the right interventions, you need to have a longitudinal dataset as your underlying foundation. Through pattern recognition, this type of patient data intelligence platform can perform a higher level of analysis that accounts for the complexity and reality of a chronic patient, and unlocks the ability to answer the real underlying questions. In short, the “truth.”
In a new report about blockchains in healthcare, in which the IBM Institute for Business Value collaborated with The Economist Intelligence Unit on a survey of 200 healthcare executives (both payers and providers in 16 countries), the report authors wrote, “How valuable would it be to have the full history of an individual’s health? What if every vital sign that has been recorded, all of the medicines taken, information associated with every doctor’s visit, illness, operation and more could be efficiently and accurately captured? The quality and coordination of care would be expected to rise, and the costs and risks likely to fall.”
The future of healthcare will continue to see the growth of scalable big data platforms and integrated, diverse datasets that will impact clinical organizations. We’re seeing the role of the provider shift from that of a diagnostician, informed by training and evidence-based practices, to that of a decision-maker, informed by training and practices as well as by real-time, patient-specific analytics that guide the clinician at the point of care.
These processes will help to change the healthcare system through improvements and efficiencies in areas such as appropriate clinical and risk indicators at the point of care. With accessible, easy-to-understand data, clinicians will be able to have a longitudinal patient record on hand during a visit, tailored to the specific genotypic, phenotypic, and social economic circumstances of that patient, as of that moment in time.
Some of the most critical tools in today’s virtual doctor bag are real-world data, experiential evidence, historical evidence and longitudinal data. However, it’s not just information about when a patient was hospitalized, discharged, and managed in the 30-day post-acute program—it’s also a much longer time horizon. When we can see the long-term impact, and have data to back it up, we really know we’re moving the needle. The heart of value-based care lies with investing in those things that really deliver the outcomes we want.
Longitudinal data is valuable because it takes into account information that is collected over more than one moment in time or incidence. To do good work and effect great outcomes – which means delivering high quality of care on a repetitive basis, and in a manner where there’s an arbitrage between what you get paid and what you must pay out–physicians need access to information that enables continuous decision-making in the background. This can only happen with longitudinal data.
Longitudinal databases also enable institutions to pick their peer groups and measure themselves against like organizations. This exercise may result in the ultimate decision to avoid, for example, antibiotic-infused cement in hip replacements because it doesn’t actually change the outcome of care. Alternatively, longitudinal data may show that appropriate use of pain medication early on, followed by suitable follow-up and transition to physical therapy, might be a better approach to managing that pain.
I’ve lived my life mired in data. Yet, when I injured my shoulder and learned I needed surgery for multiple rotator cuff repairs, did I rely on my medical experience and knowledge to find the right surgeon? I’ve been practicing at Mount Sinai Hospital for over 25 years and know a lot of orthopedic surgeons— but my first step was to ask my wife, who has a doctorate in physical therapy, for her advice. She recommended a surgeon that her friend had used. I did nothing more. I was like everyone else. With all the data available to me, I still turned to my wife for insights. That’s because there are very few easily accessible, longitudinal databases available that include historical information for patients to access when making healthcare decisions.
One exception is the decades-old Framingham Heart Study. Over the years, careful monitoring of the study’s population has led to the identification of the major CVD risk factors—high blood pressure, high blood cholesterol, smoking, obesity, diabetes, and physical inactivity—and a great deal of valuable information on the effects of related factors such as blood triglyceride and HDL cholesterol levels, age, gender, and psychosocial issues. Although the Framingham cohort is primarily Caucasian, the importance of the major CVD risk factors identified in this group have been shown in other studies to apply almost universally among racial and ethnic groups, even though the patterns of distribution may vary from group to group.
Another is the Dartmouth Atlas Project, which has been documenting medical resource distribution and usage variations in the United States for over 20 years. As the website notes, “the project uses Medicare data to provide comprehensive information and analysis about national, regional, and local markets, as well as individual hospitals and their affiliated physicians.”
This project has helped stakeholders including policymakers, the media, and health care analysts improve their understanding of the efficiency and effectiveness of our health care system, contributing to ongoing efforts to improve health and health systems nationwide. Amass enough information, line it up longitudinally, and it becomes easier to make a valuation/evaluation of care over time to look at trends, variations, and personal interactions necessary to change behavior.
Prevalence makes the difference. I believe if you look at healthcare and say, “I want to find some place where I can make an impact,” you should look at our treatment failures—at hospitals. A hospitalization is an outpatient treatment failure, unless it’s an elective visit, like plastic surgery. From a population standpoint, if we are providing appropriate care across the continuum, in theory, hospitalizations should be limited to trauma and life events. Most chronic conditions that drive healthcare today occur in a way that clinicians can address them in a non-acute environment. Access to longitudinal data can help.
As we focus on harnessing the power of longitudinal data, I began to realize it’s not wise for hospitals and health systems to make decisions that are consistently, episodically, incidences. Rather, they need to look at prevalence, to assess over time, what the outcome of various interventions has been. This approach will only make us all, as an industry, better at delivering care.
The reason we do this is simple. When you begin to peel the onion, you quickly experience a data avalanche of disparate and disconnected data points, but nothing resembling “the truth” for your individual situation.
We lack a longitudinal patient dataset (data that track the same patients over the course of many years) that will guide us to make the best treatment decisions based on actual, real-life experience from patients similar to the individual needing care. For example, longitudinal data helps to answer the question, “I had my aortic valve replaced 15 years ago. Based on the experience of patients with medical histories similar to myself, what course of treatment should I follow when it begins to fail?”
Making data like this accessible in new ways has the potential to significantly impact the way healthcare is delivered. The power of a multivariate dataset, collected from hundreds or thousands of people with similar diseases or conditions over an extended period of time, enables hospitals and health systems to make high quality, cost-effective treatment decisions for their patients. In fact, the ultimate outcome of longitudinal data is to identify at-risk patients and intervene to preempt the disease from occurring in the first place.Historically, hospitals and health systems have relied upon datasets that measure the incidence, but not the prevalence, of disease. Scientific studies are important; however, they're limited in that they're not reflective of the specific population at any one hospital, taking into account demographics such as age, income, education and environment of those particular patients.
To deliver the right interventions, you need to have a longitudinal dataset as your underlying foundation. Through pattern recognition, this type of patient data intelligence platform can perform a higher level of analysis that accounts for the complexity and reality of a chronic patient, and unlocks the ability to answer the real underlying questions. In short, the “truth.”
In a new report about blockchains in healthcare, in which the IBM Institute for Business Value collaborated with The Economist Intelligence Unit on a survey of 200 healthcare executives (both payers and providers in 16 countries), the report authors wrote, “How valuable would it be to have the full history of an individual’s health? What if every vital sign that has been recorded, all of the medicines taken, information associated with every doctor’s visit, illness, operation and more could be efficiently and accurately captured? The quality and coordination of care would be expected to rise, and the costs and risks likely to fall.”
The future of healthcare will continue to see the growth of scalable big data platforms and integrated, diverse datasets that will impact clinical organizations. We’re seeing the role of the provider shift from that of a diagnostician, informed by training and evidence-based practices, to that of a decision-maker, informed by training and practices as well as by real-time, patient-specific analytics that guide the clinician at the point of care.
These processes will help to change the healthcare system through improvements and efficiencies in areas such as appropriate clinical and risk indicators at the point of care. With accessible, easy-to-understand data, clinicians will be able to have a longitudinal patient record on hand during a visit, tailored to the specific genotypic, phenotypic, and social economic circumstances of that patient, as of that moment in time.
Some of the most critical tools in today’s virtual doctor bag are real-world data, experiential evidence, historical evidence and longitudinal data. However, it’s not just information about when a patient was hospitalized, discharged, and managed in the 30-day post-acute program—it’s also a much longer time horizon. When we can see the long-term impact, and have data to back it up, we really know we’re moving the needle. The heart of value-based care lies with investing in those things that really deliver the outcomes we want.
Longitudinal data is valuable because it takes into account information that is collected over more than one moment in time or incidence. To do good work and effect great outcomes – which means delivering high quality of care on a repetitive basis, and in a manner where there’s an arbitrage between what you get paid and what you must pay out–physicians need access to information that enables continuous decision-making in the background. This can only happen with longitudinal data.
Longitudinal databases also enable institutions to pick their peer groups and measure themselves against like organizations. This exercise may result in the ultimate decision to avoid, for example, antibiotic-infused cement in hip replacements because it doesn’t actually change the outcome of care. Alternatively, longitudinal data may show that appropriate use of pain medication early on, followed by suitable follow-up and transition to physical therapy, might be a better approach to managing that pain.
I’ve lived my life mired in data. Yet, when I injured my shoulder and learned I needed surgery for multiple rotator cuff repairs, did I rely on my medical experience and knowledge to find the right surgeon? I’ve been practicing at Mount Sinai Hospital for over 25 years and know a lot of orthopedic surgeons— but my first step was to ask my wife, who has a doctorate in physical therapy, for her advice. She recommended a surgeon that her friend had used. I did nothing more. I was like everyone else. With all the data available to me, I still turned to my wife for insights. That’s because there are very few easily accessible, longitudinal databases available that include historical information for patients to access when making healthcare decisions.
One exception is the decades-old Framingham Heart Study. Over the years, careful monitoring of the study’s population has led to the identification of the major CVD risk factors—high blood pressure, high blood cholesterol, smoking, obesity, diabetes, and physical inactivity—and a great deal of valuable information on the effects of related factors such as blood triglyceride and HDL cholesterol levels, age, gender, and psychosocial issues. Although the Framingham cohort is primarily Caucasian, the importance of the major CVD risk factors identified in this group have been shown in other studies to apply almost universally among racial and ethnic groups, even though the patterns of distribution may vary from group to group.
Another is the Dartmouth Atlas Project, which has been documenting medical resource distribution and usage variations in the United States for over 20 years. As the website notes, “the project uses Medicare data to provide comprehensive information and analysis about national, regional, and local markets, as well as individual hospitals and their affiliated physicians.”
This project has helped stakeholders including policymakers, the media, and health care analysts improve their understanding of the efficiency and effectiveness of our health care system, contributing to ongoing efforts to improve health and health systems nationwide. Amass enough information, line it up longitudinally, and it becomes easier to make a valuation/evaluation of care over time to look at trends, variations, and personal interactions necessary to change behavior.
Prevalence makes the difference. I believe if you look at healthcare and say, “I want to find some place where I can make an impact,” you should look at our treatment failures—at hospitals. A hospitalization is an outpatient treatment failure, unless it’s an elective visit, like plastic surgery. From a population standpoint, if we are providing appropriate care across the continuum, in theory, hospitalizations should be limited to trauma and life events. Most chronic conditions that drive healthcare today occur in a way that clinicians can address them in a non-acute environment. Access to longitudinal data can help.
As we focus on harnessing the power of longitudinal data, I began to realize it’s not wise for hospitals and health systems to make decisions that are consistently, episodically, incidences. Rather, they need to look at prevalence, to assess over time, what the outcome of various interventions has been. This approach will only make us all, as an industry, better at delivering care.
More for you
Loading data for hdm_tax_topic #better-outcomes...