Big Data for Big Disease: Experiences from Applying Data-Driven Strategies to Combating the Opioid Epidemic
The views expressed are those of the author and do not necessarily reflect the views of ASPA as an organization.
By Daniel Xu
June 18, 2018
Opiate drugs have taken a toll in the country. In 2016 alone, drug overdose killed over 60,000 Americans, of which 66 percent involved opioids. In fact, drug overdose surpassed motor vehicle as the leading cause for accidental deaths. In October 2017, President Donald Trump declared opioid epidemic as a public health emergency and called for increased efforts for drug overdose prevention measures. Data collection and analysis in public health has helped detect the crisis and has provided the public with the knowledge on what has happened and where it happened under the auspice of the Center for Disease Control and Prevention (CDC). Drug overdose, or drug addiction, once stigmatized, has been generally seen as a disease after public health education programs. But that’s not enough. The public needs a solution and demands to know “why it happened,” “how it happened” and, more importantly, “how to stop it.” Big data, characterized by its sheer volume, velocity and variety, holds enormous potential in finding and implementing solutions. However, successful application of big data strategy is yet to overcome significant challenges.
Admittedly, led by CDC, a considerable amount of data reporting and data analysis has already been used in routine disease surveillance and decisionmaking in resource allocation for public health. What’s more, some states have already used the results of toxicology testing and the statistical analysis of opioid-related emergency department visits data for naloxone distribution among county emergency medical services. Meanwhile, the computing and modeling capacities required for processing big data have been developed and made more accessible to public organizations. For instance, SAS, Oracle and Microsoft have developed various types of powerful data analytics and data visualization tools that are applicable to public program management. However, there are still several major obstacles that prevent from formulating and implementing more effective and more responsive interventions through integrating and utilizing big data strategies.
The first obstacle is data quality. Nowadays, data is conveniently available in a digital age when all sorts of gadgets collect various kinds of data from citizens and businesses, ranging from payroll information to various social media posts, from prenatal care on birth certificates to causes of death on death certificates. Imagine the sheer amount of medical and financial data related to diagnostic procedures, drug information, insurance claims, when over 1 million Medicaid recipients in Alabama who interact with health providers, program administrators, payers and insurers, on a daily, weekly or monthly basis. So volume is not a major concern for big data in public health. However, the blessing of data volume becomes a deficit if the data are messy or replete with errors. The process of data collection/generating in the government system in many cases is a major concern due to human input errors or computer programming bugs. “Garbage in, garbage out” is the cliché. Investments have been made to ensure the quality of data, but there is still a significant gap between the automated data reporting and usable data in many areas.
The second barrier is fragmented data ownership and usage. To create a mega-sized data requires assembling and “linking” various data sources. However, data sharing is hindered by administrative and legal barriers. In an ideal world, all agencies and programs shared their data for collaborative projects or programs. But, just in the private sector, often times they operate in silos in data collection and analysis. In case of fighting opioid epidemic, the assessment, design and implementation of effective strategies require data from various departments at federal, state and local levels, including public health, mental health, Medicaid, law enforcement, courts and prisons, educational institutions, as well as non-governmental organizations such as hospitals, treatment and rehab services, medical and pharmacy associations, community-based and faith-based nonprofit organizations. In addition, the data from private health providers such as health insurance companies are also needed. Although there were some successes in sharing data through memorandums of understanding or other mechanisms among some agencies and organizations, it is yet to clear many more administrative and legal hurdles to obtain the necessary data from other agencies.
The third issue is data security and confidentiality. The Health Insurance Portability and Accountability Act (HIPPA) mandates the protection of patient privacy and data security to public and private health service agencies and organizations as a top priority, which makes the data linkage that is critical for many data sharing on some collaborative programs almost impossible, particularly when such linkage requires individual-level patient data. Even for some of the data collected by a single program, using personal identifiable information to track individual patients and create linkages may sometimes be prohibited in order to ensure equal and timely access to emergency care for drug overdose. A related concern is data security. The central question is “Are patient’s private information in good hands?” The increased sharing of data, especially data containing identifiable personal information, potentially pose high risk of data breaches, as evidenced by the recent data breaches in both the public and private sectors.
The big data and data analytics potentially will create more and better opportunities to improve the efficiency and effectiveness to the programs targeted on the opioid epidemic. However, not until these prerequisites are met will the full potential of big data and data analytics be realized. To make the data-driven preventive programs more efficient and effective, increased data sharing is a must. More efforts are required to remove the administrative impediments and improve data quality at various levels and among various stakeholders while ensuring data security. The lessons learned from combating the opioid epiademic can be a good reference for the initiatives in many other public programs that are experiencing similar challenges in applying big data strategies.
Author: H. Daniel Xu, PhD, MPA, is a Research Analyst at Alabama Department of Public Health. He has taught graduate programs and conducted research in public and healthcare administration and policy. In addition to his research in maternal and infant health, he is currently working on a CDC-funded data-driven preventive initiative for opioid and drug overdose. [email protected]
(1 votes, average: 4.00 out of 5)
Loading...
Big Data for Big Disease: Experiences from Applying Data-Driven Strategies to Combating the Opioid Epidemic
The views expressed are those of the author and do not necessarily reflect the views of ASPA as an organization.
By Daniel Xu
June 18, 2018
Opiate drugs have taken a toll in the country. In 2016 alone, drug overdose killed over 60,000 Americans, of which 66 percent involved opioids. In fact, drug overdose surpassed motor vehicle as the leading cause for accidental deaths. In October 2017, President Donald Trump declared opioid epidemic as a public health emergency and called for increased efforts for drug overdose prevention measures. Data collection and analysis in public health has helped detect the crisis and has provided the public with the knowledge on what has happened and where it happened under the auspice of the Center for Disease Control and Prevention (CDC). Drug overdose, or drug addiction, once stigmatized, has been generally seen as a disease after public health education programs. But that’s not enough. The public needs a solution and demands to know “why it happened,” “how it happened” and, more importantly, “how to stop it.” Big data, characterized by its sheer volume, velocity and variety, holds enormous potential in finding and implementing solutions. However, successful application of big data strategy is yet to overcome significant challenges.
Admittedly, led by CDC, a considerable amount of data reporting and data analysis has already been used in routine disease surveillance and decisionmaking in resource allocation for public health. What’s more, some states have already used the results of toxicology testing and the statistical analysis of opioid-related emergency department visits data for naloxone distribution among county emergency medical services. Meanwhile, the computing and modeling capacities required for processing big data have been developed and made more accessible to public organizations. For instance, SAS, Oracle and Microsoft have developed various types of powerful data analytics and data visualization tools that are applicable to public program management. However, there are still several major obstacles that prevent from formulating and implementing more effective and more responsive interventions through integrating and utilizing big data strategies.
The first obstacle is data quality. Nowadays, data is conveniently available in a digital age when all sorts of gadgets collect various kinds of data from citizens and businesses, ranging from payroll information to various social media posts, from prenatal care on birth certificates to causes of death on death certificates. Imagine the sheer amount of medical and financial data related to diagnostic procedures, drug information, insurance claims, when over 1 million Medicaid recipients in Alabama who interact with health providers, program administrators, payers and insurers, on a daily, weekly or monthly basis. So volume is not a major concern for big data in public health. However, the blessing of data volume becomes a deficit if the data are messy or replete with errors. The process of data collection/generating in the government system in many cases is a major concern due to human input errors or computer programming bugs. “Garbage in, garbage out” is the cliché. Investments have been made to ensure the quality of data, but there is still a significant gap between the automated data reporting and usable data in many areas.
The second barrier is fragmented data ownership and usage. To create a mega-sized data requires assembling and “linking” various data sources. However, data sharing is hindered by administrative and legal barriers. In an ideal world, all agencies and programs shared their data for collaborative projects or programs. But, just in the private sector, often times they operate in silos in data collection and analysis. In case of fighting opioid epidemic, the assessment, design and implementation of effective strategies require data from various departments at federal, state and local levels, including public health, mental health, Medicaid, law enforcement, courts and prisons, educational institutions, as well as non-governmental organizations such as hospitals, treatment and rehab services, medical and pharmacy associations, community-based and faith-based nonprofit organizations. In addition, the data from private health providers such as health insurance companies are also needed. Although there were some successes in sharing data through memorandums of understanding or other mechanisms among some agencies and organizations, it is yet to clear many more administrative and legal hurdles to obtain the necessary data from other agencies.
The third issue is data security and confidentiality. The Health Insurance Portability and Accountability Act (HIPPA) mandates the protection of patient privacy and data security to public and private health service agencies and organizations as a top priority, which makes the data linkage that is critical for many data sharing on some collaborative programs almost impossible, particularly when such linkage requires individual-level patient data. Even for some of the data collected by a single program, using personal identifiable information to track individual patients and create linkages may sometimes be prohibited in order to ensure equal and timely access to emergency care for drug overdose. A related concern is data security. The central question is “Are patient’s private information in good hands?” The increased sharing of data, especially data containing identifiable personal information, potentially pose high risk of data breaches, as evidenced by the recent data breaches in both the public and private sectors.
The big data and data analytics potentially will create more and better opportunities to improve the efficiency and effectiveness to the programs targeted on the opioid epidemic. However, not until these prerequisites are met will the full potential of big data and data analytics be realized. To make the data-driven preventive programs more efficient and effective, increased data sharing is a must. More efforts are required to remove the administrative impediments and improve data quality at various levels and among various stakeholders while ensuring data security. The lessons learned from combating the opioid epiademic can be a good reference for the initiatives in many other public programs that are experiencing similar challenges in applying big data strategies.
Author: H. Daniel Xu, PhD, MPA, is a Research Analyst at Alabama Department of Public Health. He has taught graduate programs and conducted research in public and healthcare administration and policy. In addition to his research in maternal and infant health, he is currently working on a CDC-funded data-driven preventive initiative for opioid and drug overdose. [email protected]
Follow Us!