From potential to reality: Is there a real ROI of predictive analytics in healthcare?
Some time ago, we conducted a survey on predictive algorithms in healthcare . The purpose of the survey was to gather the opinions of representatives from companies and organizations developing medical technologies, and to take a closer look at the opportunities and challenges they see in implementing such algorithms in the solutions their organizations bring to market. Today, we would like to invite you to review the detailed results of this survey:
1. Predictive algorithm survey – what do our respondents think?
2. Is predictive analytics new hit-or-miss? How the survey results might be combined with global trends + infographic
3. Predictive algorithms delivering measurable ROI in healthcare: real-world benefits that predictive analytics brought to our client’s product
Predictive algorithm survey – what do our respondents think?
In the first question, we asked about the areas in which our respondents are developing or co-developing medical applications. Their apps focus on disease diagnosis (35%) and treatment recommendations (25%). This is followed by mental health (12%), patient monitoring (15%), and wellbeing (11%).
Based on these primary areas and data required by applications our respondents suggested the data for training and validating predictive algorithms in their health apps should be based on various resources. 19% of respondents pointed to the electronic health records (EHRs), 43% to medical imaging 43%, and 27% reported patient data. Wearable devices were also specified, but to a lesser extent (7%). This suggests how important it is to the technology industry to increasingly recognize the value of a variety of data sources to improve the accuracy and reliability of predictive algorithms.
The majority of respondents (50%) expressed a neutral stance about the accuracy and reliability of predictive algorithm predictions. This suggests that there is still some skepticism about the capabilities of these algorithms, although there is also a significant minority who are confident in their potential.
64% of answerers are concerned about the possible regulatory challenges or compliance issues when using predictive algorithms in their environment. This is likely due to the fact that these algorithms can have a significant impact on patient care and decision-making, and it is important to ensure that they are used responsibly and ethically.
Among the barriers and obstacles, the most indicated challenges to implementing predictive algorithms in health apps are user acceptance (43%), cost and resources (19%), concerns about potential regulatory challenges or compliance issues (18%) and data quality (11%), and eventually integration complexity (9%). These challenges highlight how complex the integration of predictive algorithms into existing software systems might be and how demanding the ensuring that they are accepted by end-users is.
On the other hand, respondents expect a range of business benefits from incorporating predictive algorithms into their applications. The most anticipated benefits are increased revenue (24%) and cost reduction (33%). Predictive algorithms can also help companies differentiate themselves in the market (15%), monetize data (17%), and improve decision-making (11%) – respondents suggested.
Finally, a vast majority (83%) see the integration of predictive algorithms in their apps as a feasible option in the future. This suggests that despite the challenges, medical companies hold optimism about the potential of these algorithms to transform healthcare.
Is predictive analytics new hit-or-miss?
The survey concluded that predictive algorithms may be a major force in healthcare and personalized medicine. However, it is important to address the challenges that need to be overcome before they can be widely adopted.
The results of the survey also show a continuing trend of interest in predictive analytics in the field of medical technology, and they are consistent with the trends and analyses conducted by market consulting and analysis giants. An example is a Deloitte report , which identifies the following as the main opportunities associated with the implementation of predictive algorithms:
- Improving efficiencies for operational management of health care business operations
- Accuracy of diagnosis and treatment in personal medicine
- Increased insights to enhance cohort treatment
However, it is clear that the technological environment is far from being overly optimistic and rather approaches this trend with some reservation. Would the situation look different if predictive algorithms ceased to be potential and became measurable ROI and savings values?
Predictive analytics delivering measurable ROI in healthcare
One of the projects we conducted focused on developing a model to help predict and stratify the risk of postoperative delirium (POD) in patients after surgery. The project was successful, and the client was able to demonstrate the real ROI that healthcare can achieve by implementing such a predictive algorithm in their network. According to their case study  describing a pilot, two-month deployment conducted on a group of 862 patients over the age of 60, the hospital achieved monthly savings of >CHF 470,000 due to better management of patient length of stay in the hospital (reducing the average length of stay (LOS) by 0.8 days). Additionally, there were additional revenues (CHF 30,900) due to improved management of postoperative screenings and better diagnosis.
And can you think of similar, real-world benefits that predictive analytics could bring to your organization, your clients or end users?