Analytics in Medical Practice: How to Get Physicians On Board and Engaged
Quickly digesting a great deal of information about a patient, determining a diagnosis, and creating a treatment plan is a critical part of the training physicians’ complete medical school. The attending physicians at my medical school often left me in awe with their seemingly uncanny ability to decipher important clinical clues and develop the appropriate treatment for patients.
The Affordable Care Act has changed the landscape of the healthcare industry from fee-for- service to pay-for-performance. Physicians are under pressure to provide high-quality, cost effective care in a safe manner while still treating 40-50 patients day and spending as little as 5-10 minutes with each of them.
Measuring the effectiveness of treatments via data and analytics has also been a challenge. Adoptions and use of analytical method by physicians, and the availability of meaningful data, is a recent development.
In addition, medical knowledge has exploded and the array of diagnostic tests and treatments has increased significantly. It has become more challenging for physicians to decide on an optimal diagnosis and treatment plan. Interpreting new medical information can be difficult, especially when studies are sometimes contradictory.
A well-trained, intelligent physician simply cannot keep pace with all the latest research or have all of the knowledge available to them for every medical event.
Descriptive, Comparative, and Predictive Analytics
Healthcare organizations are driving quality improvement using descriptive and comparative analytics. Descriptive analytics describe groups of patients. Comparative analytics examine the differences between groups of patients via risk stratification or treatment effectiveness. This is evidence-based medicine.
Quality and professional organizations are using descriptive analytics to develop evidence-based care guidelines and algorithms. Tools used in the clinical environment include, scorecards and alerts in the electronic health record ordering systems. The introduction of comparative analytics to compare different performance measures have proven critical in improving outcomes.
Driving tangible, long-term change requires healthcare organizations to move beyond descriptive or comparative analytics to predictive analytics. Predictive analytics takes known information from medical literature including best practice models, the latest scientific research, and existing databases, to develop a predictive model.
Predictive analytics allows physicians to continuously improve the effectiveness of their treatments. Adding a patient’s medical, genetic, or social information creates a more precise, predictive view for an individual, sometimes called precision medicine.
Empowering physicians with predictive analytics can continuously improve the effectiveness of their own practice. Tools can calculate the risk of a patient developing a venous thromboembolism after bariatric surgery or the 30-day outcome of a patient admitted with acute coronary syndrome.
These predictive tools should not replace physician judgement, but can enable physicians to practice more efficiently and effectively. Improving the predictability of these tools will require improved processes to determine which deidentified patient data should be developed and incorporated into the daily physician workflow.
Analytics in Medical Practice
Clinical research, insurance, or billing purposes, not quality improvement was the focus of data gathering. The Institute of Medicine’s publishing of To Err is Human and Crossing the Quality Chasm, has spurred physicians to utilize data to improve outcomes.
Adoption of data analytics by physicians to drive quality improvement has lagged behind the rest of the healthcare community. One common refrain? My patient is special or our organization is different. In reality, patient populations across the country are very similar.
Data is an essential tool for helping physicians understand exactly how their performance compares to industry benchmarks. Incorporating data into healthcare practice is an evolutionary journey. Health systems have been slow to digitize, and only recently has meaningful data been available to physicians.
Getting Physicians on Board
Self-Service versus Self-Sufficient
Physicians need self-service access to data. Waiting weeks for answers does not allow physicians to incorporate clinical best practices at the point of care. Physicians, even those excited to utilize the data, can become discouraged when compelled to use an inefficient, ineffective process. If this occurs repeatedly, physicians will likely lose patience with the process and disengage.
Relevant and timely data is exponentially more effective than scheduling a meeting to announce the implementation of a new protocol. Proactively delivering the information at the point of care is much more valuable to physicians and patients.
Healthcare leadership must also understand the difference between self-service and self-sufficient. Physicians are not analytics experts and need assistance in interpreting data and discovering the best ways to apply analytics to improve patient care.
Embedding data analysts in different departments of a hospital to serve as a resource to physicians is a smart strategy. Over time, these analysts become very familiar with the department’s clinicians, its data, and its processes enabling them to play an invaluable role in optimizing the adoption and application of analytics.
Physicians need transparency. They need to know that they can trust the data—where it came from and how accurate it is. At the same time, organizations need to help physicians understand that data can be “perfect enough” to identify areas for quality improvement. Utilizing readily available data sources and making incremental improvements enables physicians to have more confidence in the data and the process.
For example, one hospital system created an enterprise data warehouse (EDW) with visualization capabilities enabling physicians to have near real-time answers to their clinical quality improvement questions. The physicians learned how their decisions affected length of stay (LOS) and how careful, considered changes to the clinical processes would improve LOS. By accessing the data, it was easier to convince the physicians to make the needed changes. The bottom line? Physicians who feel they are a part of the change the more likely they are to engage.
Everyone in the organization must understand and be committed to quality improvement and using data and analytics to achieve success. Leadership must commit to creating a culture of continuous improvement and support the use of data and analytics across the organization. Commitment must also be demonstrated by ensuring physicians have all the right tools and resources to incorporate analytics in everyday practice.
Physicians also need to believe leadership is willing to stand behind them to inspire and drive change. No matter how well intentioned, team members that lack the commitment and engagement of the executive team, may find it impossible to advance quality improvement initiatives and improve outcomes.
But, We’re Different
Organizations can expect physicians to push back against analytics. By engaging physicians with data, organizations can counter many physician objections. For example, a common pushback from physicians is the following: My patient is special or our organization is different.
The power of data to change attitudes and misperceptions was made very real to me early in my quality improvement journey as we kicked off a program reduce elective deliveries before 39 weeks. My team and I made presentations to our OBGYNs about the dangers of early deliveries. Their response was, “This is fascinating information, but our patients are different. The data you are presenting comes from patients in different demographics than ours. Our population is healthier and our outcomes are superior.”
We could not deny that they had a point: The data that we showed were from published studies in other institutions and did not necessarily reflect their patient population or their particular practice. We took their concerns seriously and looked at our system’s data for our own population.
Armed with this data, we regrouped with the physicians and showed them babies born before 39 weeks in our own system had markedly worse outcomes. Demonstrating the harm to our own patients, our OBGYNs quickly committed to the improvement program. Being able to show physicians what the impact of their medical practices are to patient outcomes is a powerful motivator, as all physicians want to improve the care that they deliver.
Playing the Blame Game
The culture of blame is entrenched in healthcare. It will take time, commitment, and open, honest communication to rid it from the industry. My team had a patient experience and survive a cardiac event. Following the incident, I called a meeting with the entire staff, who were understandably nervous, and began the conversation by telling them I was not seeking to place blame on anyone.
We discussed how the cardiac episode and how it could have been prevented or handled differently. The clinicians were skeptical at first; however, by the end of the discussion they felt safe enough to share their impressions and recommendations. As a direct result of the team’s feedback, we changed several procedures to streamline the process for accessing and directing rapid response teams to our postpartum unit and operating suite in Labor and Delivery.
We already had team debriefs following every surgery, but we extended them for any significant event or near miss. The teams ask three questions: What went right? What didn’t? What can we do to do better? A process such as this should become an integral part of the healthcare culture.
What Success Looks Like
Engaged physicians, as part of the quality improvement process can become proponents of analytics and quality improvement. I have witnessed this repeatedly.
For example, I presented a new best-practice process to a group of physicians. I was afraid they would balk at the requested changes. After reviewing the data, they recommended the process be even stricter as the data showed that a stricter process would lead to better outcomes for the patients. The level of engagement was remarkable.
The 2002 Olympic Winter Games, held in Utah from February 8 to March 16, 2002, provides another example. In the wake of the terrorist attacks on September 11, 2001, and the anthrax release in October 2001, bioterrorism surveillance was paramount during the Games. A team of informaticists and public health specialists implemented the Real-time Outbreak and Disease Surveillance (RODS) system for the Games. The system monitored 114,000 acute care encounters between February 8 and March 31, 2002. No outbreaks of were detected.
The system remains operational today to assist with the prediction of seasonal respiratory syncytial virus (RSV) epidemics in Utah that can result in significant pediatric morbidity and increased health care costs.
RSV epidemics generally occur between October and April, and can result in significant morbidity and increased healthcare costs.
Size and timing vary from year-to-year, making accuracy in predicting seasonal epidemic characteristics, including times of high activity and total size, and support efficient management of resources.
Health care facilities are using the data to forecast requirements for beds, staffing, testing, treatment, and other resources needed to care for sick children. For greatest effectiveness, these predictions are made early in the RSV season, so the predictions would be useful within the first month of the observed start of the RSV seasonal epidemic.
Empowered and engaged physicians, using actionable analytics, can improve outcomes and close the gap for getting best practices translated into clinical practice. The ultimate benefit? Using data to improve the overall health of individual and populations of patients.