April 22, 2016
Walk into any health system board meeting and you are bound to hear the conversation turn to “data.” There is no bigger trend in healthcare right now. Massive investment is flowing into the collection and storage of data with databases of information to be mined.
Yet while population health and value-based care steal the attention, data has enormous potential to improve healthcare operations. With only 30% of physicians working at “top-of-degree,” there is no more critical time to streamline your operations. Here are five things your data should tell you:
How to reduce no-shows
Your data should highlight when an appointment is likely to be missed, enabling you to double book or avoid scheduling the patient all together. Some relatively straightforward models can predict the likelihood of no-shows with reasonable accuracy using appointment, clinical and patient details]. Furthermore, no-show visits should be identified and action taken within 24 hours to ensure better care. This helps avoid “failure-to-diagnose” suits, which can be levied even if a patient has missed a specialist appointment .
The referring provider has rarely had access to quality and timely show information in the past, not to mention evidence when a referral has been completed. Yet for healthcare providers, no-shows lead to direct cost losses of roughly $210 per event and can cost up to 14% of anticipated daily revenue for a typical outpatient clinic . The use of care coordination software can substantially increase show rates through automated messaging and reduced lead times.
How to eliminate long lead times
Your data should identify when an appointment is scheduled too far away from the actual date of the appointment. Long lead times delay treatment, and a delay in receiving a critical diagnosis can be truly devastating both from the perspective of cost to the paying entity and to the patient. In a study of delays in the treatment of lung cancer from primary diagnosis to treatment, the median time between diagnostic and specialist CT scans was 54 days, with a median increase in tumor size of 19% . The same study found 6 of 29 patients became incurable as a result of treatment waiting time.
Your data should provide a real-time view of three key measures: percent of appointments with lead times exceeding N days by specialty, (2) the median and average lead time and (3) the overall distribution of lead time. The first measure allows a manager to identify exceptions, or times when waiting for a physician can be dangerous and costly. Knowing the median and average lead time helps distribution get a quick feel for what the true wait is for a specific specialty.
How to optimize physician capacity
Your data should tell you the utilization of your physicians. Behind long lead times and low show rates is often poor physician/service capacity utilization. Not unlike other operational processes, health appointments are subject to Little’s Law : as more appointments are made to a specific physician with the amount of time to service each patient unchanged, the queue waiting for an appointment will increase in length. Overbooking to a physician means that patients must wait days and sometimes weeks before an appointment is made while opportunities may exist to book to providers performing similar functions.
There is substantial variance between physicians in the total time spent seeing patients, the total time spent on administrative work and the total number of patients seen per day . Yet ~75% of physicians claim to be at or over capacity , implying opportunities for operational efficiency and a reduced workload on some providers and an expanded one for others. Even if a hard standard for patient-hours worked cannot be established, physicians can at least be benchmarked by specialization and region.
How to book an alternative when capacity is limited
Your data should be driving referral patterns by helping schedulers (and patients) identify acceptable alternatives for a referral or appointment. The benefits of analyzing physician capacity are fully realized with an automated way of identifying similar providers who can accomplish the same procedures. This enables rerouting patients from over-capacity to underutilized physicians. An individual making an appointment, whether that is a referral coordinator in a call center or the patient, typically cannot decide whether two providers in the same specialty can truly provide the same service.
This enables optimization for lead times and capacity while driving improvement in show rates. Using a combination of public and private data sources, it is possible identify providers who performed >50% of the same services using a simple distance algorithm. Optimized networks mean better care, faster and more capacity utilization.
How to measure results
To connect it all, your data should let you identify whether an action you’ve taken is working. Assuming the effectiveness of a business decision happens all too frequently across a variety of businesses, and healthcare operations are no different. Typically, this type of work is reserved for high cost consultants and internal strategy teams. However, the advent of more user-friendly data analysis tools and an increased focus on data collection has made measuring the results of an action far easier.
When taking an action, such as choosing to double book appointments when the first patient to book is likely to no-show, one should always ask, “How can I measure such an action’s results?” In this example, measuring the number of patients actually showing up, and the wait time increase as a result would be critical. Another key is to ensure you have a standard for comparison. Much like running a clinical trial, ensure you have a control or holdout group to compare your actions against a baseline, or observe a metric over time before and after a change.
1: Characteristics and Direct Cost of Academic Pediatric Subspecialty Outpatient No-Show Events
Read the original article at Becker’s Health IT & CIO Review