Make-or-buy? Implementing patient demand and nurse staffing predictive analytics

Published by Hiro Kawashima on July 09, 2019

As in all industries, the healthcare industry has recently seen an explosion of predictive analytics solutions that promise to glean clinical and operational insights from data. However, as hospitals face relentless pressure to cut expenses and improve their return on investments, leaders often grapple with the decision to make or buy these technologies. Hospitals must balance end user requirements with their ability to maintain the ever-expanding list of application technologies. Hospitals still have a tendency to make applications because there is an entrenched belief that making applications leads to long term savings, satisfied end users, and a higher return on investment. But are these assumptions accurate and is there a better way to acquire these technologies?

Staffing has been an area lacking in widely applicable predictive analytics. Traditionally, hospitals rely on basic historical analyses, an arbitrary budget number, or intuition to staff and schedule nurses. Although some staffing and scheduling vendors offer built-in analytics, most still lack accurate and comprehensive predictions for patient demand and nurse staffing need. This leaves hospitals to consider building in-house or outsourcing to third party vendors.

 

Criteria for make-or-buy

An investment in a new technology can have long lasting effects financially and on end users. Hospitals have to be diligent when considering the make-or-buy decision. Although the criteria for make-or-buy decisions are unique to each hospital, there are general topics to consider:

make-or-buy-1.png

 

Our experience making a predictive analytics platform for staffing

We spent more than 12 months building StaffRight, a SaaS platform that helps hospitals forecast patient demand and clinical staffing need. This was in addition to two years of research and engineering to construct and hone the underlying algorithms that power StaffRight’s predictions. A project on the scale of StaffRight required a product manager, a data scientist, and a software engineer. Based on average salary data from Built In Chicago, a hospital today would have to spend more than $325,000 annually just in wages alone to develop and maintain the platform. Additional expenses such as technology infrastructure costs (hosting, physical hardware, etc.) and non-wage employee costs (insurance, training, taxes, office space, etc.) can significantly increase the required budget for such a project.

make-or-buy-2.png

Data from Built In Chicago (https://www.builtinchicago.org/salaries/data-analytics/data-scientist/chicago)

If a hospital were to build something similar to StaffRight, it would take significant effort, money, and time before any return on investment.

 

Conclusion

The make-or-buy decision is unique to each hospital and requires careful examination of business and end user requirements, market research, and introspection. It is also important to realize that even though use cases and end user requirements seem specific or unique, a majority of uses cases such as predicting nursing need can be generalized across hospitals. With the estimated cost to build an in-house patient demand and nurse staffing predictive analytics solution exceeding $325,000 annually, it may be prudent to buy an existing solution from a trusted vendor at far less cost and faster return on investment timeframe.

Next post: Prescience Health on AHA's Advancing Health Podcast

Take the guesswork out of staffing

Schedule a demo and learn more about how predictive analytics can help you staff better
SCHEDULE A DEMO