Why is Healthcare Stuck in Reporting?

Size: px
Start display at page:

Download "Why is Healthcare Stuck in Reporting?"

Transcription

1 Roundtable Discussion Why is Healthcare Stuck in Reporting? April 2013 Moderated by: Bryan Engle should be directed to

2 Overview Understanding and obtaining the benefits of advanced analytics is crucial for healthcare organizations to be able to drive better health and business outcomes. While the healthcare sector is ahead of the game using science in medicine, in many practical ways it has not ascended in analytics maturity as quickly as other industries like banking and retail. This roundtable conversation focused on overcoming the challenges blocking healthcare analytics teams from deploying advanced analytics techniques. Context This summarizes a phone discussion lead by IIA faculty member Bryan Engle and attended by members of the Healthcare Analytics Research Council. These calls are held each month, focusing on a unique topic. The objective was to discuss what analytical activities healthcare organizations were striving to achieve, potential barriers to overcome, and successful ways of addressing those barriers. The discussion identified areas where healthcare organizations were applying advanced analytics and how analytics professionals successfully used analytics. Key Insights The support for analytics is coming from the top, and healthcare is transforming to an analytics-driven industry. Healthcare executives are driving the use of analytics to transform their organizations and to deliver high quality care in a redefined environment. These organizations are facing major changes on multiple fronts of business and need rapid innovation. Staying the same is not an option. Healthcare organizations are specifically facing an array of changes in financial and account reporting, and shifting arrangements between payers and providers. To sustain these redefined relationships, analytics departments need to provide rich information exchanges between provider and insurance partners. Healthcare organizations are pushing analytics to evaluate the potential ROI of business decisions to ensure that all new initiatives improve quality and meanwhile reduce cost. Another driver of change is the increase in consumerism in the industry overall. As one participant summed it up, As with all change, there are a few companies that have the margin or vision to really get out front and innovate because it is interesting, or because it Roundtable Discussion Stuck in Reports 2013 p. 2 should be directed to members@.

3 might produce value. But for most of us who live on very thin margins, this kind of innovation and change happens when everyone gets that you have to do these things to stay in business. The integration of clinical and administrative data might have been on the cool frontier a few years ago, but now they are table stakes. Performing the right intervention at the right time for the right patients, being efficient and measuring and managing for high quality, these are no longer cool edgy things to do in your spare time. You will go out of business if you do not do them. Key Insights Physician support is critical for any analytical solution. Getting buy-in from clinicians is critical for any analytical solution, particularly initiatives that address the most efficient way to deliver care. Some organizations faced big challenges when analytical solutions changed the way clinicians delivered care, such as providing actionable data at the point of patient care or incorporating reports as part of the clinician work flow. At other times, the suggested solutions were embraced. For example, one organization was inserting cost information to help inform clinical choices by making providers aware of costs up front when a doctor has a choice of different items. Providers wanted to do the right thing, but since they did not previously know what the items cost it had not entered into their decision at all. Two factors were critical for adoption success: 1) Data need to be actionable and trustworthy; 2) organizations need to find and recruit physician champions from the beginning. Participants described an iterative process of testing solutions early in the development process to identify any potential pushback and then finding ways to address those areas. Another way to encourage clinical support is by setting organizational goals to include targets for some of the analytics initiatives. Other organizations already had a strong culture of shared accountabilities, including patient engagement and cost-effective medicine, and found that made it easier to work together on solutions. Key Insights Start with a simple prototype to get clinical support. Analytics projects had more success if they started out with a simple prototype and asked potential users what should be added; clinicians got frustrated with too much complexity up front. It is easier to create a prototype when the data are not real. That gives the chance to design a user interface that doctors can touch and feel before investing a lot of time in the real system. Giving clinical staff a sense of the project and asking for direction creates a sense of ownership, particularly with providers. Providers might otherwise be reluctant to have health plans tell Roundtable Discussion Stuck in Reports 2013 p. 3 should be directed to members@.

4 them about the cost and quality of what they do. After getting input on how to expand from the starting point, the next iteration needs to be complex, such as moving from summary level data to patient level detail. Key Insights Bad data can kill analytics projects. Establishing the trustworthiness of the data source is critical to a project s success. The typical project lifecycle involved multi-disciplinary groups of administrators, clinicians, and physicians involved from the beginning. If the dataset does not make sense to the audience, the project gets off to a bad start. The suggestion was to be on the lookout for bad data and set expectations based on the quality of the data. In other words, point out the shortcomings of the data up front. At the same time, mend problem areas using risk adjustments and other techniques to eliminate excuses for not trusting the data. Organizations need to support outcome metrics with clinical data. Billing or financial data alone is not enough to present outcomes without a sense of patient population at a doctor level and facility level. Different facilities can have very different patient demographics and the doctors expect any system to account for those differences if the data is to be considered reasonable and trustworthy. One participant described an analytical tool comparing doctor performance using the billing data. Physicians who were not performing well could dismiss the results because they could claim the billing data did not reflect overall patient levels of sickness. In response, they implemented risk adjustment analyses to demonstrate whether their panel of patients was sicker than others to eliminate that argument. Key Insights Bringing together financial and clinical systems offers tremendous opportunity. One key challenge for healthcare organizations is coordinating many disparate sources of data. Participants were at different points in their progress towards digitization and integration of clinical and administrative data across providers. Some had a big divide between financial systems and clinical data. Those organizations were in the midst of a big push to bring new systems online that consolidate previously siloed information: financial, clinical, community hospitals, and physician practice data. Once information is integrated, organizations envision a shift from ad-hoc reports to true predictive modeling. Those within an integrated healthcare delivery system, on the other hand, have more control over information and data are already available from multiple clinical and financial sources (i.e., Roundtable Discussion Stuck in Reports 2013 p. 4 should be directed to members@.

5 pharmacy, inpatient, and outpatient) with richer reporting ability. This structure enables business organizations to bring different datasets into a single data warehouse and relate financial information to insurance data. With the foundation of a comprehensive longitudinal health record and consolidated information between payers and providers, an economic model can help identify revenue patterns. Healthcare organizations at the forefront of the analytics curve will be building an economic model that has financial, clinical, and quality data incorporated into it. Key Insights Predictive analytics can model patient risks and consumer demand. Some of the biggest advances in predictive analytics were at the point of physician-patient interaction. Some healthcare organizations are building models that score patients on the probabilities of different health outcomes and risk of readmission, such as the prediction of whether a patient winds up in emergent care within 90 days of discharge. Newer technologies can be used to look at unstructured data, such as discharge summaries and physician notes, to build a more complete picture about a patient. Social media data is another potential wealth of unstructured data to inform organizations about patient perceptions of the business. In addition to building dashboards for clinicians and executives, predictive analytics can be used to build at-risk payments and reimbursement payment models. Conclusion Analytics will be critical for healthcare organizations to maintain quality care while adjusting the new business environment. To succeed in the changing healthcare market, organizations will need to integrate clinical and financial data and leverage advanced analytics to find the most efficient way to deliver care. These areas will continue to be explored by the Healthcare Council on the monthly calls. Roundtable Discussion Stuck in Reports 2013 p. 5 should be directed to members@.