AI Predicting and Explaining the ‘Why’ in Disease

PHI focus: Corresponds to Informatics in Disease Prevention and Epidemiology

Summary:
Identifying the cause of disease(s) outside the method of randomized controlled trials is rare. Have you ever asked yourself “…why do diseases occur?”. An emerging technology in the field of Artificial Intelligence (AI) presents an opportunity for implementation of predictive analytics for effective disease prevention by investigating the cause behind cases of disease diagnoses by leveraging the power of data to unravel the complexities of disease. AI has the potential to improve clinical decisions and improving patient outcomes but that means first understanding the causal drivers of disease(s), and this area of emerging technology is called causal AI. Causal algorithms that reveal the “why?” behind disease can be a difficult task to infer causal relationships from data, but it’s a major necessity requiring organizations to invest in building data infrastructure for it. Observational data and data from practice simulations that, for example, increase treatment for potential outcome(s), can be applied to various frameworks with AI methods. Methods such as Bayesian networks, structural equation models, and potential outcome frameworks can be used to discover mechanisms of disease, treatment optimization, and social determinants of health. Social Determinants of health use causal machine learning techniques that have the capability of streamlining investments. Building more hospitals is a common example of investments that never made a difference in factors surrounding patient impact for access to health facilities and interventions that yield successful results, especially when the real issue could be arranging safe and reliable transportation in rural populations around the world. Rate of survival from certain cancers, for example colorectal cancer, can be better understood if the underlying mechanisms of that particular cancer is also understood. Targeting the right treatment for the right patient can be facilitated by identification of causal drivers and molecular drivers, drivers that also serve as biomarkers for survival.

Assessment: Surveillance Officer of Population Health and Biostatistics at World Health Organization

To know the “why?” behind disease would open many avenues for coordinating and implementing technical activities directed towards surveillance of outbreaks and health emergencies, scale up the implementation of evidence-based interventions, , analysis and quality throughout the full cycle of the disease incident(s). Causal AI algorithms would facilitate comprehensive treatment optimizations because knowing the “why?” could provide technical oversight for the implementation, monitoring and evaluation of public health policies and programs pertaining to surveillance of emerging and re-emerging infectious diseases particularly those with epidemic and pandemic potentials. Treatment optimization is facilitated by predictive models made available as decision support to care providers. The predictive models that simulate “what-if” scenarios are used to generate data about disease outcomes and disease progression under variable action sequences and health care interventions. Such models are the vehicle for individualized treatment for patients. Decision support care can prevent recommendations made for patients that have failed to take into account certain patient factors that put them at risk for receiving less care. Getting adequate data can be challenging because causal AI can only provide trustworthy conclusions about disease if the data is representative and accurate. Specifications for accuracy are met when the models can be trained on quality data representative of the right populations and then merged with other stat sets, then compared to a good control group. Causal AI algorithms could decrease unnecessary spending for healthcare initiative and decrease patient recovery times by improving predictive treatment plans for those suffering at the hand of disease(s).

 EHR Data Display Could Hinder Children at Play

PHI Focus: Corresponds to Privacy, Confidentiality, and Security; Ethics and Information Technology

Summary:
EHR usability challenges contribute to patient safety threats and events that have threated the patient. The system that contains the electronic health record containing pediatric patient information does not have to undergo testing after being customized and implemented. Glitches and EHR mistakes are acknowledged to occur with the volume of information contains in health informational technology systems that are used by hospitals and clinicians including system feedback, visual display data entry and workflow support. It is also acknowledged there are 12 major avenues of patient safety events that occur more often for pediatric patients. The purpose of an EHR is to display an array of patient health information and therein lies the first mistake when basic patient information cannot be accessed. Comments and instructions written by physicians, medication administration information and drug history in the past and present are to be stored. Poor information display conflicts automatic EHR functions that can contribute to a clinician incorrectly scheduling medical treatment and administration, even increasing the chance to seeing the wrong patient information for the wrong patient being seen. When that information conflicted there have been mistakes where nurses have administered drugs that can put the child at risk, the epitome of poor information display mistakes that lead to a myriad of medical errors. Rigorous test are needed to asses safety-related EHR protocol and usability features utilizing consistent checkpoints of the EHR lifecycle. This can prevent future incidents and can be facilitated by ONC certification testing drafting new rules as outlined in the 21st Century Cures Act.

Assessment: Senior Director / Chief Health Information and Exchange Officer
EHR functionality should require testing for the entire HER lifecycle because if so usability testing can be extended to development, implementation and customization. Rigorous testing to assess safety-related EHR protocol is an imperative factor in patient care assessment and communication amongst clinicians. The newly drafted voluntary rules for use in pediatric care seek to reduce protentional threats to patient safety are not only needed immediately but will also serve pediatric patients the advocacy and medical care that they deserve when admitted to our hospitals here in the state of Ohio. If more rigorous testing is implemented processes that identify and prioritize the needs of clinicians using the software. Age-based care can be appropriately administered with surety, in turn lessening the chance of EHR mistakes is suboptimal usability that contributes to such errors. If pediatric EHR systems are better optimized with the new rules the potential for EHR-associated patient harm could be reduced. The potential impact would be facilitated by the voluntary certification program that the Act calls upon the ONC to not only create but make more rigorous because the program is tailored to the specific needs of pediatric care. Documentation requirements and design changes could ensure that the EHR systems meet not only the needs of the clinicians but also the be reflectively accurate of the patient and patient diagnosis. EHR vendors and health care organizations can work to together to boost transparency concerning system usability rectifying the gap by decreasing the likelihood of patient harm associated with EHRs.