Our blog is moving

We have consolidated A Shot in the Arm blog with the SAS Voices blog.  All future posts intended for A Shot in the Arm will be published on SAS Voices. Please update your bookmarks and your RSS feeds with these links to make sure you don’t miss future posts:

Stay tuned for more exciting changes and content from the SAS blogs!

Post a Comment

Partnering for prevention: Dignity Health and SAS combat sepsis

There’s been an uptick recently in corporations working together to improve patient experiences and even saves lives. From the recent buzz about drug developers coming together to combat Alzheimer’s, to health systems opening data to researchers, companies realize that partnering just might produce the greatest impact.

As highlighted in Dignity Health demonstrates the power of advanced analytics at HIMSS16, Dignity Health, the fifth largest health system in the nation, and SAS, the world’s largest privately held software company, have teamed since 2014 to use analytics to improve health care delivery.

Read More »

Post a Comment

Reproducible research: Is my SAS code enough?

There was this very embarrassing day around year six of my career as a statistician working in clinical trials. I had a small group of interns working on a project that combined data from multiple clinical trials. The goal was to better understand sources of variation in the common control used in a series of trials. Pretty routine stuff.

What made this bad for me was when I was reviewing a block of code. I got very aggravated with trying to understand what it was doing and why “they” chose to do it this particular way. The very nice interns did not act defensive at all. I finally exclaimed: “Where did you get the idea for the awful approach, some crazy Google search?” One of the interns very hesitantly said they pulled it from the source code I had given them. It turns out this same block was used in each of the entire sequence of trials I had them working with. Who was the author of this block? Unfortunately, the answer was a younger and more careless version of me.

Is it just me?

It turns out this topic has a great name - reproducible research. This is not to be confused with “replicable research” where similar methods are applied to new data to verify similar outcomes. In reproducible research, the goal is for a new researcher to take the same raw data and code and exactly match the reported results without much effort. The key being the without much effort part. I’ll get to that in a minute.

The origin

This term became very popular after a 2006 study published in Nature Medicine was examined by Keith Baggerly at MD Anderson Cancer Center in an attempt to reproduce the results using the same data. In that case, the research was not only hard to reproduce but it had been biased by data manipulation that was uncovered in the years that followed.

The story was featured on the CBS news show 60 Minutes Deception at Duke: Fraud in Cancer Care?. There is a great overview of the story by Keith Baggerly himself.

Without much effort

Now back to the without much effort part of making my code reproducible. With every trial submission to the FDA, we provide data in a standard format (SDTM) and our code for analysis, tables, listings and figures. The known context of these data sources make our code easier to understand, but does it match the without–much-effort standard?

Back in 2006, I used my embarrassing example as motivation to do a few things better when I created code. First, liberal use of comments to say what I am doing, describe how I was doing it and to note any sources of information. Another practice I added was creating a readme.txt file in folders with easy-to-read notes about the analysis approach I was taking as well as any assumptions and interpretation that were important.

Finding a solution

As I moved further into managing groups of statisticians and programmers, it became more important to find ways for teams of people to adhere to these practices. With that also came the obligation to make it easy to do. I have continued to seek new ways to make this easy in my career at SAS. Below I will introduce three approaches that we will dig deeper into with future blog posts.

A platform approach

I highly recommend looking at is SAS life sciences analytics software recently covered in a blog post by my colleague Matt Becker. It has many features built-in that make reproducibility easier with minimal burden on the creator of content. It also enables collaboration in a unique way that facilitates reproducibility.


SAS® Studio

This is the most modern development environment for SAS coders and users. It is browser-based and resembles the layout many of us have used our whole careers with quick access to the code, log and listing. SAS Studio includes a number of features that are very helpful for reproducible research:

  • Legibility: Make code easier to read.
  • Reuse: Code snippet functionality.
  • Summary: Program summary with code, log, results and execution information.
  • Package: SAS program Package with code, log and results.
  • Custom Tasks: Users can interact with your code without programming.



The notebook approach to coding involves embedding code into a virtual notebook that enables the creation of literate programming. A popular open-source notebook that is enabled by kernels to many languages, including SAS, is Jupyter. This allows the author to create very descriptive text alongside code blocks that can also be executed in line to display results in blocks. I like to describe this as a combined code, log and listing in a single document flow. A great introduction can be found in this blog post: How to run SAS programs in Jupyter Notebook


Let’s talk!

At the upcoming DIA 2016 Annual Meeting in Philadelphia, we will be demonstrating the three approaches introduced above. We would love to get your feedback. Please stop by Booth 1825 to see these ideas in action.

Post a Comment

Three things you must do to create a central clinical platform

Collecting, managing, standardizing and analyzing clinical data during (and after) a clinical trial is crucial in the process for submission and regulatory approval of a new compound, biological, device or other therapy. A central clinical platform requires:

  1. Robust and auditable analytics to prove the result to the authorities and external stake holders such as patient communities and physicians.
  2. A controlled environment (Figure 1) that fosters collaboration between the different groups in clinical development such as data management, biostatisticians, clinical writers and scientific investigators.
  3. Enhancing your platform with new functions and integration capabilities to meet the needs of a clinical development group.

Figure 1. Controlled environment with groups, users and role-based access.


Beyond providing a reliable way to access data, programs and clinical data standards (Figure 2), you need a way to satisfy these three requirements. SAS® Life Science Analytics Framework is a modern services-based extensible platform that allows organizations to integrate and extend capabilities for CDISC metadata management and facilitates integration with another system such as an EDC system.


Figure 2. clinical data standard within SAS Life Science Analytics Framework.


In addition, SAS provides the solutions and the underlying technology to:

  • Standardize collected clinical data.
  • Easily combine and transform it into readily analyzable data sets.
  • Manage (in detail) the CDISC standards in the context of clinical studies (Figure 3).
  • Enable statisticians to analyze the data with the highest scientific rigor.

Figure 3. Study metadata from within SAS Life Science Analytics Framework


SAS provides the analytics to optimize the processes and resources required to move a new molecule or therapy from the laboratory to the clinic. SAS also hosts the integrated analytics environment (Figure 4) in a secure, private cloud and handles all management and maintenance of the complex IT requirements for a strongly audited and regulated system.


Figure 4. Fully interactive SAS environment.


Want to learn more? At the upcoming DIA 2016 Annual Meeting in Philadelphia, we will have experts available to demonstrate the full capabilities of LSAF. Please stop by Booth 1825 to discuss how you may be able to use SAS Life Science Analytics Framework in your organization.

Post a Comment

Let SAS® Studio do your clinical research ‘tasks’

SAS is widely used in clinical research activities including:

  • Managing and transforming data.
  • Generating tabular and graphical summaries.
  • Performing powerful statistical analyses such as safety and efficacy evaluations.

In addition, SAS provides a number of interfaces from which a user can select to work with the data. One of these newer interfaces, SAS® Studio, offers capabilities that can help busy users accomplish more in a day to ensure research studies are performed efficiently and that patients have access to the safest, most effective treatments possible.

SAS® Studio is a browser-based interface for writing and executing SAS programs. While visually similar to the familiar SAS windowing environment (aka “PC SAS”), the code can be executed by a SAS engine in the cloud, on your network or even on your PC with results returned to the browser.

SAS Studio ships with a variety of predefined “tasks” that provide point-and-click interfaces to perform common actions. Some of the tasks provided with SAS Studio include data tasks:

  • Performing random sampling or standardizing data.
  • Creating a variety of graph types.
  • Using a variety of statistical analyses (including correlations and regressions).

More importantly, SAS Studio allows users to create their own tasks by easily designing an input form to support a set of code.

Why would you want to create a SAS Studio task?

A task provides a mechanism for exposing standardized code to a broader audience. A prime candidate for creating a task is to add a user interface to an existing macro. This eliminates the need to remember the parameters required for the macro – a user simply makes choices from the form that has been designed. This interface can include checks for interdependencies and enforce requirements for the user selections before the code is submitted to execute. This combination of features allows experienced users to reuse algorithms without having to refer to documentation or risk copy/paste issues from previous work and allows users with less programming experience to easily leverage validated code.

In the example below, we created a single SAS Studio task to make use of two popular published macros. The first macro performs a 1:N case-control propensity score matching and the second macro reports on the balance of the results from the first macro. In observational studies, it is common to match cases and controls based on similar propensity scores calculated using logistic regression of baseline covariates. This approach can reduce bias between treated and non-treated groups to more closely resemble a randomized clinical trial. A comparison of the distribution of baseline covariates between the resulting groups is needed to ensure the propensity score has been adequately specified.

first macro

This tab created for our task allows the user to specify details about the logistic regression used to calculate the propensity score. There are additional tabs for specifying the details for the case-control matching and the diagnostics output.


second macro

These are the results generated by the diagnostic macro. Note the buttons that allow the user to save the output as HTML, Adobe® PDF or a Microsoft® Word document.

The same process can be applied to any other SAS code you may be using. At the upcoming DIA 2016 Annual Meeting in Philadelphia, we will present examples of SAS Studio Tasks for clinical data management, analytic data preparation, analytic methods and reporting including pushing results to other tools such as SAS Visual Analytics. Please stop by Booth 1825 to discuss how you may be able to use SAS Studio tasks in your organization.

For more details on SAS Studio and SAS Studio tasks, see the official documentation or refer to the links below:

For more details on propensity score matching, see the SAS Knowledge Base Usage Note 30971, Computing and matching observations based on propensity scores.

Post a Comment

Hospitals, data and analytics (part 2): Optimism and progress

In my previous post, I wrote about the Quick Pulse survey of leaders of health care institutions and the vital role data and analytics has in patient care and engagement. In this post, I'll tell you what the survey had to say about the progress these leaders believe they are making in integrating data and analytics into their organizations.

When asked whether they consider their own organizations to be industry leaders in integrating clinical and financial data, 17 percent of survey respondents indicated agreement. However, 72 percent perceive their organizations to be growing significantly and making progress with the integration of clinical and financial data. This is encouraging news and conveys a sense of optimism and progress in achieving the data and analytics objectives.


Using data scientists and analysts

Finally, survey respondents indicated that only 15 percent of health care organizations employ data scientists that are centralized and accessible to all areas of the organization. In most cases, access is limited to specific areas or locations.

Data scientist and analysts

Key takeaways

This survey helps confirm general observations across the U.S. health care industry that most hospital-based organizations are in the process of retooling to get better at using data for making real-world decisions. They are actively looking for ways to minimize data siloes, provide actionable data at the point of service and learning to work together across their organizations to encourage data-driven performance improvement.

This means that if:

  •  Your hospital or care delivery system is actively implementing data governance and a unified analytics capability, rest assured that you are not a bunch of geeks. Your leaders are getting it and you need to support those efforts.
  • If your organization is not doing this, you may want to think hard about your ability to effectively deal with the emerging environment of value-based care. Remember: If you can’t measure it, you can’t manage it.
  • Make sure you are building a culture that uses data to drive clinical and operational decisions. This is not just for researchers – we have to work smarter in health care.
  • Make it a priority to develop citizen data scientists or more simply, help your clinicians and business professionals get comfortable working with data and simple statistical test, then using the results to make decisions. You don’t need to turn them into statisticians or programmers, just teach them fundamentals and help them develop a healthy respect for information-driven practice.

To that end, SAS is the ideal partner. No other organization excels in data management, enabling front-line data exploration and providing flexibility in being able to customize and share visualizations of what the data are communicating, in order to create real solutions. Now that's a toolkit for real progress!

Interested in learning more about how health care and life sciences organization use SAS analytics? Tune into our free, on-demand virtual forum  in which industry experts share their strategies, best practices and challenges.

Post a Comment

One more ‘giant leap’: Four decades of cardiovascular research data now shareable

In 1969, a human walked on the moon. I remember my parents waking me in the wee hours to watch (albeit bleary eyed) the grainy images on our black and white television as Neil Armstrong set foot on the lunar surface. Heady stuff then and now.

In another world-altering advancement, Duke Clinical Research Institute announced that the world’s oldest (and largest) cardiovascular database will be made 487041685available to researchers. These records span from the year of the moon landing to 2013. This database is a particularly important one because heart disease is still the leading cause of death in the US.

More than 50,000 patient records have been anonymized and data from more than 100,000 procedures are included. SAS will provide data management and analytics tools to help in exploration and analysis of the data. DCRI is implementing a data governance plan to make sure that researchers have reliable data sets and that access requests are processed as quickly as possible. Researchers should contact SOAR, DCRI’s project website, for access.

One of the goals of this collaboration is to encourage a new era of transparency and openness for the public good, according to DCRI Executive Director Eric Peterson, MD, MPH.

“SAS provides the environment and analytics to spur advances in cancer research through the Project Data Sphere initiative. The company also promotes new medical research through its work with the pharmaceutical industry to share clinical trial data from nearly 600 studies with researchers around the world. This new collaboration with the DCRI will foster more clinical research and data sharing, with the aim of improving people’s health today and tomorrow,” according to Matt Gross, Director of the SAS Health Care and Life Sciences Global Practice.

About a year or so after Apollo 11 made its historic voyage, my grandfather, a relatively young man at the time, had a heart attack (and I seem to recall that he was later treated at Duke). He was lucky to live a full and fairly healthy life, and my hope is that this data will allow researchers find treatments that help reverse the trend of cardiovascular diseases in our lifetime.

I’m proud that SAS is helping DCRI lead the way in data-driven health care.

Post a Comment

Hospitals, data and analytics: What the leaders think

Everyone in health care is talking about data and analytics these days and the vendor world has shifted into overdrive. So what is really going on in the trenches, and what does this mean for the future? Does this mean that we will finally find the holy grail of information-driven care?

Health and life sciences experts at SAS wanted to better understand how health care organizations are using and sharing their data. With the assistance of IDG Research Services, we conducted a Quick Pulse survey on health care data and analytics practices during March 2016.

About the survey

In order to participate in this six-question, online survey, respondents must be at a director-level or above and work for a health care institution with one or more hospitals and beds for acute inpatient care and at least 500 employees. The survey explored:

Areas or departments which have access to/use different types of data.

  • Progress with data integration.
  • Important data and analytics objectives.
  • Use of data scientists within their organizations.

The charts below show distribution of job titles and organization size, for the 54 respondents:

  cudney chart onecudney chart two

The importance of data and analytics for patient care and engagement

In several ways, the survey results tell us that a lot of work related to data and analytics is underway. On the other hand, the results tell us that they’re not there yet.

The table below compares objectives versus practices for data and analytics in patient care and engagement. The majority of respondents highly rated the importance of each objective, which underscored the priority placed on data and analytics. However, in some objectives there is a significant gap in the importance ratings versus actual practices.

The survey revealed a dichotomy. Interdepartmental collaboration to use data for improvement was noted as Critical or Very important by 93 percent of respondents. However, only 74 percent believe this is Always or Usually True in their organizations. Similarly, 91 percent of respondents said the ability to personalize patient care and engagement using analytics is Critical or Very important. Yet only 76 percent believes this is Always or Usually True in their organizations.

Access to and use of data

Respondents emphasized that even though personalized care is a top objective, clinical teams are the least likely to have access to different data types outside of their specific area.cudney chart three

Similarly, respondents indicated that clinical staff are usually able to personalize patient care. However, these efforts may be based on incomplete data because organizations are inconsistent in ensuring that insights gleaned from data across all populations is easily accessible to patient-facing staff.

This highlights the continuing challenges with providing analytics-ready data at the point of service, where front-line professionals make clinical and operational decisions that affect lives and outcomes of patients and their families.

In Part 2 of this series, I’ll tell you about the progress that hospitals are making and sense of optimism about analytics leading the future.

Post a Comment

Dignity Health demonstrates the power of advanced analytics at HIMSS16

According to Lloyd Dean, president and CEO, "At Dignity Health, we are committed to developing partnerships and opportunities that harness the tremendous potential of technology, from improving the patient experience to providing caregivers with tools that will support their day-to-day care decisions."

Dignity Health, one of the largest health systems in the US, continues to demonstrate how its partnership with SAS and the use of advanced analytics can propel improvements in the delivery of health care. It recently presented several breakout sessions during HIMSS16 in Las Vegas.One of these sessions provides an excellent example of what Dean was talking about. Alyson D’Andrea and Ken Ferrell, representing the Dignity Health Insights (DHI) team, presented, Sepsis Biosurveillance at Dignity Health: Our Improvement Story.

The high cost of sepsis

clinicalSepsis is the body’s overwhelming and sometimes life-threatening response to infection which can lead to tissue damage, organ failure and death. Patients who develop sepsis have an increased risk of complications and death and face higher health care costs and longer treatment.

The sepsis biosurveillance program represents the vision of the DHI team:

  • Use big data and the world's best analytic tools in a secured cloud-environment to improve health outcomes and reduce costs.
  • Use data across the continuum of care to improve outcomes of patients, increase knowledge, worker productivity, enable self-service access and markedly increase calculation speed and insights to decision makers.

During this session, D'Andrea and Ferrell explained the process for creating an internal dashboard for monitoring performance on several measures related to severe sepsis. Specifically, the tool and associated workflows implemented at 16 facilities (so far) helped reduce sepsis mortality rates. According to Dignity, an average of 69,000 people per month were monitored during this phase of the program.

For more than two decades, researchers and professional societies, including the American College of Chest Physicians (ACCP) and the Society of Critical Care Medicine (SCCM) have driven improved practice and consensus around severe sepsis.

Clinical leaders at DHI understood the importance of this initiative and looked for a way to use data from the electronic medical record with their SAS analytics platform to create a dashboard that shows the status on key performance indicators, allowing rapid intervention to save lives and improve the clinical outcomes of patients at risk for sepsis. This also supports generating evidence from actual practice to inform research and further refinement of evidence-based practice. This performance dashboard uses Hadoop technology, which significantly decreases storage cost, aggregates a variety of data sources, as well as enables enhanced security and flexibility in using the data.

Alerts and analytics

The DHI project strategy included identifying key performance indicators, implementing the St. John Sepsis Agent tool within the Cerner EHR system and creating a feedback loop to enable stakeholders to improve response time to the alert. The DHI team statistically validated the sepsis dashboard for accuracy and functionality. This included a "silent mode" evaluation, whereby manual chart review helped to correlate the St. John's Agent activity with patient discharge diagnosis. Subsequently, the team developed clinical alerts, carefully tweaking the sensitivity, specificity and relative risk to achieve just the right balance. Using Cerner’s alert, the team validated their findings and proposed design with internal stakeholders to ensure the result would achieve agreed goals for the project. A major focus of the DHI team was to direct the smarter alert to the person responsible for the patient at the time of the alert. This was defined as the primary care nurse, charge nurse, triage nurse and attending provider.

The team also worked to ensure the alerts fired at the right time, which was when the presence of sepsis in a patient was identified and 24-48 hours after the previous alert fired. The team also worked to ensure that the alerts filtered out patients, who were known to have sepsis at the start of their hospitalization. Alert filtering also included those who had cardiac surgery or extracorporeal membrane oxygenation (ECMO) within the previous 48 hours.

Better response, lower mortality

With initial implementation of the sepsis dashboard, the team soon began to see that higher nurse communication rates (nurses responding to alerts) correlated with lower mortality rates. They also were pleased to see fewer patients progressing to severe stages of sepsis. As with many improvement projects, they also discovered other opportunities for improvement such as achieving greater accuracy of clinical coding and clinician compliance with sepsis bundle orders.

As these programs continue, the DHI time will continue to focus on governance and goal ownership three or four meaningful key performance indicators, optimizing alerts around workflow and content and providing feedback to stakeholders in an actionable format.

Using its partnership with SAS, the DHI team has demonstrated a powerful use case for analytics in clinical care, and a strong argument for self-service visual analytics. Dignity Health has effectively enabled providers to better manage patient risk, decrease patient mortality, shorten length of stay in the intensive care unit and achieve overall reduction in cost of care. To learn more about the work at Dignity Health, please visit its website.

Post a Comment

Here comes MACRA: Analytics can help you measure and manage

Medicare payment changes are coming. The Centers for Medicare and Medicaid Services (CMS) has announced the intention of increasing the proportion of payments to providers based on outcomes and changes in health status, as opposed to delivery of services.

At the January 11th, 2016 J.P. Morgan Annual Health Care Conference, Jan. 11, 2016, CMS Acting Administrator, Andy Slavitt commented that Meaningful Use criteria will soon be replaced by outcomes-based measures. Specifically, he stated, "Now that we effectively have technology into virtually every place care is med_computer_MG_7104provided, we are now in the process of ending Meaningful Use and moving to a new regime culminating with the MACRA implementation." MACRA (Medicare Access and CHIP Reauthorization Act of 2015) makes significant changes to the way Medicare pays physicians, accelerating Medicare’s shift toward value-based payments.

Measures based on MACRA include: quality, cost, use of technology and practice improvement. MACRA also replaces the traditional sustainable growth rate model for reimbursing for serviced delivered by physicians. In addition, this new focus will "move away from rewarding providers for the use of technology and towards the outcome they achieve with their patients," according to Mr. Slavitt.

This is consistent with the January 2015 announcement of CMS to set an aggressive target for transitioning to Medicare fee-for-service payments that are linked to quality, as well as those based on alternative payment models, such as accountable care organizations (ACOs). This aggressive target means that in 2018, 90 percent of the reimbursements CMS pays for health care would be through those arrangements.

This unraveling of the longstanding link between delivery of services and payment for those services (fee-for-service health care) is forcing providers and care delivery organizations to think differently about their roles and assume a level of risk that few are prepared to handle.

Under the old way of thinking, hospital-based organizations would be less likely to promote ACOs because it would likely mean a decrease in demand for acute care as patient health improves. So, care delivery organizations are facing a transition that could impact their financial viability if it is not managed effectively.

In this situation, the old adage, "if you can't measure it, you can't manage it," still rings true. We seem to be drowning in a sea of digital data. Implementation of electronic health records is almost ubiquitous. Data seems to be everywhere but the ability to quickly harness the data, understand and analyze it, then turn it into actionable information is elusive.

Analysts and programmers can help, but creating meaningful reports takes time and resources, which are usually in short supply. The ability to understand and analyze the changes in value, will define the ability to manage risk and maintain margin in this emerging environment.

The power of information must be available to health care leaders and front-line professionals, using the power of their data, in real time, to explore diagnoses, interventions, responses and changes in health status.

Practitioners and leaders in care delivery organizations can no longer afford to be restricted to submitting requests for static reports and then receive those reports without the corresponding ability to explore the data and understand what is really going on.

The SAS Analytics Framework for Health Care incorporates unparalleled expertise in data management into powerful software that encourages discovery and promotes clear, understandable and versatile deployment of that information. Users can rapidly identify relationships between variables, test assumptions, develop hypotheses and create predictive models. These capabilities used to fall exclusively within the realm of programmers and data scientists but now are readily available to those who are in the trenches.

In short, health care professionals must know what works and what doesn't . . . now. The SAS Analytics Framework for Health Care uses the latest in analytic technology to give health care professionals the insights to make more effective decisions – for the patient and the business.

Be sure to visit the SAS booth at HIMSS16 to learn more about these new capabilities in health care analytics.

Post a Comment
  • About this blog

    Welcome to the SAS Health and Life Sciences blog. We explore how the health care ecosystem – providers, payers, pharmaceutical firms, regulators and consumers – can collaboratively use information and analytics to transform health quality, cost and outcomes.
  • Subscribe to this blog

    Enter your email address:

    Other subscription options

  • Archives