Inpatient hospitals and ambulatory surgery centers have worked for decades to optimize approaches to treating patients within their four walls.
While there is always room to improve, changes in quality and cost tend to be incremental and differences between facilities relatively modest. Post-acute care is a different matter. The level of variability in cost and quality post-discharge is ridiculous. In fact, post-acute care is the largest driver of overall variation in Medicare spending, according to MedPAC.
Historically, acute care providers have given little thought to post-acute care. Where patients go next has often depended on discharge planner relationships, convenience, and marketing, with little formal consideration given to the impact on costs and quality. How long they stay has largely been driven by financial incentives -- skilled nursing facilities (SNFs) tend to hang on to patients until reimbursement runs out, rather than discharging patients when they are ready to return home. All of this is starting to change, as value based payment models grow, and readmission penalties begin to bite. Fortunately, big data and analytics solutions are emerging to assist in discharge and post-acute management.
Enter Analytics
The field of big data and analytics is the computerized application of principles that have been with us since the early days of the insurance industry. Big data and analytics may be thought of as solutions to the needle-in-the-haystack problem. In other words, how do you serve a patient in a way that’s directly relevant to that unique individual? There is enough data collected about patient service care patterns and variability, but until recently we have not had the machine power to interpret and analyze it. Big data and analytics is a sophisticated process of making sense out of seemingly meaningless streams of data.
Consider the case of an 80-year-old man presenting for outpatient cardiac catheterization. All the pertinent facts of his case can be analyzed and compared to other patients like him. In this way, discharge planners can identify risks that are specific to this person and design a discharge plan that reduces the risk of complications and need for hospital admission. Earlier tools relied on general averages, but by leveraging big data, new tools can make relevant recommendations based on a rich set of individualized medical, social, and geographic parameters. For example, analytics may determine that the safest discharge location for the patient is a skilled nursing facility. In the meantime, this patient’s experience and outcomes -- and those of many others -- is also being recorded and stored, so that the process of analysis is being continuously refined and improved.
Big data and analytics systems are far from perfect. They cannot determine which specific patients will suffer what complications, nor are these systems designed for such predictions. Rather, what predictive analytics does is provide statistical likelihoods of events occurring. They provide a map of the territory that we historically knew about but have not measured. But the insights generated by advanced techniques still need to be interpreted. They provide inputs for the clinician, but are not a replacement. It is up to healthcare planners to identify cutoff points at which to intervene for the patient’s benefit. For example, discharge managers may decide to prescribe a home health aide for an elderly patient who was predicted to have a 75 percent chance of having difficulties with activities of daily living after discharge.
Navigating Reams of Healthcare Data
Several analytics companies have stepped up to offer services that address the post-acute challenge. Las Vegas based Owned Outcomes specializes in applying artificial intelligence and machine learning to improve outcomes in post-acute care. The company digests large amounts of data into easily-actionable outputs, enabling healthcare providers to find the best, safest, and least-costly post-acute care settings. PointRight Inc., located in Cambridge, MA, is one of the more mature contributors to the predictive analytics market, having been founded in 1995. Radial Analytics, in Concord, MA, specializes in cloud-based solutions for discharge planners faced with the challenge of finding the ideal post-acute location for older patients. Newcomer Profility Inc, based in Boston, leverages proprietary predictive analytic tools to locate ideal discharge locations for post-acute care patients.
Challenges
Big data holds a great deal of promise for optimizing post-acute care and bending the cost curve, but substantial challenges remain in the path of full realization. One is the problem of interoperability. Analytics systems cannot work effectively if they do not all speak the same language. The healthcare industry has yet to develop standards for universal data transfer that all stakeholders can agree on. Healthcare would be wise to follow the example of banking and telecommunications, which have adopted such standards, to the mutual benefit of all stakeholders.
Though machines are great at crunching numbers, humans are still required to decide which outputs to act on and which to set aside. In addition, attention needs to be paid to the quality of data that enters the system. Good output depends to a large extent on good input. Patients and caregivers are important stakeholders whose input must be incorporated. After all, a nurse can present the best discharge plan predictive analytics can generate, but the patient and family still need to agree to it.
Taking Analytics to The Next Level
As the population continues to age, costs associated with post-acute care will continue to rise. Health plans, clinicians and caregivers should learn as much as possible about scientific tools to help bend the healthcare cost curve. Big data and analytics can provide effective pathways for selecting the best care for post-acute patients, including the best level of care, the optimal site for delivery of that care, and the appropriate duration of care. The tools are good today, and will only get better with use. An important feature of big data and analytics that stakeholders should understand is that the outputs generated are designed to improve continually over time. At the same time, we should never lose sight of the fact that our mission is focused on vulnerable human beings. Input from patients and families is essential to fostering a humane healthcare system, even as machines play a growing role.
The views, opinions and positions expressed within these guest posts are those of the author alone and do not represent those of Becker's Hospital Review/Becker's Healthcare. The accuracy, completeness and validity of any statements made within this article are not guaranteed. We accept no liability for any errors, omissions or representations. The copyright of this content belongs to the author and any liability with regards to infringement of intellectual property rights remains with them.