To move further towards value-based care, organizations need to make greater use of data analytics. To improve outcomes for patients at lower cost with higher patient satisfaction, data needs to be collected and analyzed. There are many ways to do this and I’ll point out a few, including some that I use.
Healthcare is awash in data. There is billing data, including diagnostic and procedure codes. There is data based on claims from payers. There is much more detailed data in EHRs, such as test results. These data are clinical data. Financial data is also important. Much of this data can be found in financial software like QuickBooks and patient management software.
To make effective use of this data, it is necessary to organize it, graph it, and investigate significant changes. Each of these requires the use of statistics, either at an elementary level to spot trends or at an advanced level to see if there are any real, significant changes that indicate the provider is moving toward providing better care at lower cost. The examination and analysis of data collected from patients and grouped by patient characteristics is called health management at the population level.
One approach to using analytics to improve patient health is to examine and track key indicators for chronically ill patients, typically the patients who incur the highest cost in a practice. For example, a practice may collect and analyze the A1c levels of all patients diagnosed with diabetes. This data can be collected each month and a mean and standard deviation determined. A month-by-month chart using an Excel spreadsheet will help to visually indicate any trends that are occurring. If over a period of several months there is an upward trend, steps can be taken to break the trend. This could be making use of a nurse coordinator to help patients better manage their diabetes. Control charts of engineering statistics could be used to better analyze whether the trends are real or due to random fluctuations that are normal in any data collection over a period of time. One could also use t statistical tests to determine if the changes are really significant or not.
It can be very helpful to graph the data collected, such as A1c levels, plot the means over time as above, and present it in a Dashboard with brief descriptions and analysis. These dashboards can be shared in a practice to drive improvement. This can be very effective in making improvements with a care team. I recently listened to an NPR podcast of ‘The Hidden Brain’ describing how a hospital in Pittsburgh improved hand washing by caregivers before entering patient rooms. The rate hovered around 10% for a long time despite repeated education of caregivers. The hospital then began displaying monthly handwashing statistics on a dashboard that everyone could access and see. Management focused the caretakers’ attention on the boards. Handwashing rates quickly improved to 90% and stayed there. The images had a significant impact on vendors’ handwashing awareness.
Analytics can also be used to improve patient satisfaction scores. The Medical Group Management Association (MGMA) provides a very good patient satisfaction survey for its members. I have adapted it to different providers depending on their demands. The survey covers 36 basic questions and ends with “Would you recommend the provider to others?”, a very good final question. I also add demographic questions. Providers can use the survey with patients and track performance in five areas: appointment making, quality of front desk staff and billing, ease of communication, doctor visits, and facility condition. The goal is to have as high a composite score as possible for each of these areas. With advanced analytics you can discover more to improve satisfaction. It is possible to identify which of the questions has the greatest impact on the last question. It has been found that patients who are most likely to refer a practice to friends and family are often the most satisfied. So finding out which of the many questions have the biggest impact on this can help identify which areas need improvement. Analysis of which questions have the greatest impact should be carried out on a regular basis, as those with the greatest impact may change over time.
Of course, a dashboard should be created to inform staff of the survey results every month or so. This will help incentivize staff to perform even better. If management likes, they can break down the results by staffing area or by vendor to help identify where people can improve. Individual coaching can be used to help staff make improvements. However, boards should never be used to scold staff. Whether individual dashboards are shared with other staff depends on how well the staff work as a team, how supportive of each other.
The MGMA collects a large amount of data from its members through surveys. It then provides data for benchmarking to its members, some for free and some for cost. Vendors who participate in your surveys can often get the results for free. In a recent article in your monthly publication Access provided a data dashboard on the most profitable independent practices. It found, contrary to common sense, that those organizations with the highest average costs per FTE physician were also the most profitable. Some of the data on the dashboard are as follows:
Median Total Medical Revenue per FTE Physician = $1,169,542
Median Total Operating Cost per FTE Physician = $630,680
Physician total median compensation and benefit = $462,722
Median total support staff per FTE physician = 5.12
Article, Designing the practice of the future, found in the March 2017 issue of the journal, also provided the baseline data for all multispecialty practices. This data, along with other item data and item analytics, can help providers develop long-term strategies to improve the profitability of their practice.
As you can see, there are many ways to use data and data analytics to improve results at your care site. It is important to identify what data to collect and analyze to achieve the most significant impact. Like I said, care providers are inundated with data and trying to analyze it would be a waste of time and energy. Those just starting out with advanced analytics should start with a few data projects and expand as time goes on. The intelligent use of data and analytics can have a significant impact on the care you provide and your results.
