What Ails the Healthcare Operations Executive’s Mind?

Avijit Datta

Allow artificial intelligence to be your remedy. it’s a HECK OF A LOT EASIER THAN YOU THINK.


Nothing I say here would surprise you except for the remedy. You have been at this for a while, know every trick in the book, seen virtually every situation play out and while you have solutions in place, you are always curious to find “what else can we do?” The answer is “quite a lot” given that new technology allows for insights and cures that were not available before.


Delivering efficient and safe services that satisfy patient needs is your credo. The attempt of this paper is to address a few operational issues that you typically face and illustrate how advanced technologies can make your life a whole lot simpler – and that too without having to hire expensive resources – and a whole lot more effective. Let’s take a few run-of- the-mill situations and apply advanced technology to them:


Outpatient no-shows – the bane of many operational efficiency attempts. If you notice these numbers are high, what do you do? Do you want to wait till the numbers are high enough to hurt or would you rather preempt them and avoid a revenue shortfall altogether?  You need a solution that can understand behavior – more importantly, potential changes in behavior – and then corral the downslide and have a remedial strategy that is most likely to succeed. You want to understand the contributing factors to the change in behavior and put together an action plan that neutralizes these factors. This is not a wish list. Rather, it is a reality with today’s advanced technology options. AI allows you to monitor patient behavior, through machine learning learn typical versus non-typical behavior, understand the factors that have the most bearing on patient behavior, tweak those factors to identify cause and effect and, in the process, present you with a plan that slashes no-shows to its bare minimum. And what’s more through machine learning it “self-learns” to adjust ideal factor levels without necessarily being prompted to do so.


“Door-to-doc” time in the emergency department. Common issue. Many moving parts to tweak to make an impact. The consequences, though, are significant. We know length of patients’ wait time has implications on staffing levels, patient safety, and patient and employee satisfaction. These metrics are available to the public on Medicare Hospital Compare, and patients will use this information to decide which hospital to visit. AI based systems today have the ability to get this complex issue resolved. Solutions that allow you to mix in a range of variables including external data and give the analyst the ability to build, change and compare their models themselves is where you will find the most effective strategies evolving. The ability to conduct sensitivity analysis by altering moving parts like staffing levels of skilled and unskilled personnel, schedule planned machine downtimes, understand seasonality and time sensitiveness of visits will allow you insights that probably was not possible and certainly not easily possible in the past. Artificial intelligence, predictive technologies, machine learning and self-learning systems put this concern to bed today. Astute analysis of past data and behavior by ailment type, by demographic, by region or even by season and the ability to use this to predict factors that are likely to cause delays, allow for preemptive measures that compress the time a patient remains unattended.


Medical Coding Errors are a source of frustration to many care givers. After having provided excellent service it is a shame if reimbursement does not reflect the actual work done. The revenue loss has the potential of being a source of considerable concern. This is money that has been earned but not claimed. Why not claim them without incurring additional expenses to do so. Where does advanced technology fit in here? Very snugly. AI algorithms allow for scrutiny of the provider’s diagnosis and compare it to the DRG codes submitted. Mismatches are either a case of “under-coding” or “over-coding” both of which have an adverse effect on the organization.


There are a plethora of such everyday cases that AI and other advanced technologies can get you to be on top of the situation. This must be expensive. Negative. This does not need to cost you an arm and a leg. Multiple deployment options and the ability to make your current non-technical staff proficient in conducting their own analysis, build their own models, use their own assumptions using these technologies is where we allow ourselves to rise up to the challenge with surprisingly reasonable expense.

Avijit Datta