Fool’s Gold, Goodhart’s Law, and other limitations #physicaltherapy
If you are new reader, the last few posts have been on observations at our profession’s largest conference and in particular #physicaltherapy student recruitment. We are going to stay on this related theme for a bit.
Next week, at this year’s Therapy Con I am giving a keynote on all things #Physicaltherapy First and participating in a panel discussion on Culture, Recruiting, and Training Staff to Improve Retention with some very impressive colleagues and moderated by Nick Hedges the CEO of Raintree. One area that will definitely be addressed is retention.
Let’s level set a few definitions first. There are two related HR data points. The first is retention rate that relates to our discussion.
Retention rate is typically calculated as the number of retained therapists (or whatever you are studying-we will use customers) divided by the total number of therapists at the start of the period, multiplied by 100 to get a percentage. The formula is:
This means you retained 80% of your PT’s from the start of the period.
The beauty of this formula is its simplicity, the limitation is the same. What if you add locations? What if you really need 90 PT’s? Is retention rate for PT’s that have been retained vs. new hires? Those are the right questions and it would take more than simple arithmetic to get anything close to a meaningful answer.
Another way to look at it in reverse and far more valuable in my opinion is turnover. Turnover rate is typically calculated as the number of employees who leave a company during a specific period divided by the average number of employees during that period, multiplied by 100 to get a percentage. The formula is:
Whether you calculate the turnover rate or the retention rate of your therapists or employees is a matter of preference, and many calculate both. It you get caught up in over analysis or overtly focused on these data points, it can cause problems. Data always has limitations and trade offs.
The information does make for good discussion but can be difficult to compare across time periods and in particular across distinct companies in an industry as there are no real standards or for that matter agreed upon best practices to really derive meaning from singular data. The best you can do is compare intra or within your own clinic/company over time and even then, you have to answer separately the more highly variable factors- causality. The biggest mistake would be to compare industry-wide or via some data you find on the internet at some obscure HR website.
Let me digress a bit. One of the best practices of any company is to track patient no-shows or cancellations. The idea is that you have done the heavy lifting, you have gotten a new patient in the door-now you have to keep the patient through the episode of care. Like turnover and retention, the formulas can be a bit variable and are highly dependent on time and convention. Is it measured daily, weekly, or monthly? What if they reschedule in the same week-does that count? What if they call and cancel and reschedule immediately-does that count? Lots of variables and at the end of the day we don’t know much except that the best companies have an internal calculation that they stick with and compare across their own locations over time vs. any published standards which frankly don’t exist. We do know that clinics that have a high pediatric population generally can have a 14–16% no show rate (somewhat similar to geriatric) and that general orthopedic population range from 8–12%. What are the means to improve? Plenty to include hi technology like reminder calls, texts, emails. Others that are more high touch like personal calls. Anecdotally, my experience shows that once an on-going patient misses, the best in class have the PT call the patient and express how much they were missed and the importance of making the appointment as that is where accountability lies but realize this is hard to execute. The overall point though is even with executing the best of the best practices, you are going to improve the percent to a point (let’s say out of convenience 8–10%) but at some point the cost, effort, and inconvenience does not result in any incremental improvement. A statistical number of patients simply don’t show.
Back to retention and turnover. The analogy with no show/cancellations is sound. You have done the heavy lifting and hired a PT, now you have to retain past the current average 8 month tenure. You can improve through high tech and high touch just like no show/cancellations. This can include pulse checks or scheduled check ins, “stay interviews”, professional development, structured mentoring, creating a world class culture, and fantastic benefits-but, at some point, you aren’t going to get that turnover down to 0% or retention to 100%. At some point, the effort to improve it provides no return on investment and becomes a waste of money and resources. In practice, I see constantly these two extremes: the false illusion that any one specific technique is somehow a panacea for retention and relatedly what we refer to as Fool’s Gold, the belief that adding certain expensive undertakings: complete day off for employee appreciation, Starbucks cards, schwag, and other tokens somehow convinces people to stay. This doesn’t mean you shouldn’t do these types of activities at all; just realize if you are justifying for retention purposes, you should re-think your strategies, including paying people more rather than the occasional coffee card. These are all tradeoffs to be heavily debated and manifest ultimately in an overall program that has to have cultural elements and traditions embedded.
The other related phenomenon is that retention and turnover (and in particular, no-show/cancellation rates) are subject to Goodhart’s law if taken too seriously or worse used for financial incentives. Goodhart’s Law is an adage in economics and social science that states: “When a measure becomes a target, it ceases to be a good measure.” This implies that once a metric is used as a target for decision-making, its reliability as an indicator can be compromised because people may manipulate their behavior to achieve the target, often leading to unintended consequences. It is very easy to manipulate the data and more easily to be captivated by it. Thus, the retention rate, which was originally intended as a measure of keeping your therapists, becomes distorted and loses its value as a reliable indicator and even worse becomes the target rather than the real importance of keeping your therapists.
An additional criticism of the data measurement is that both retention and turnover are lag data vs. leading indicators. The data only tells you were you have been and measuring it per se has no impact on changing it. It’s like a bathroom scale and weight, you now know how much you weigh but getting on the scale has no direct influence on losing weight. By comparison, a lead indicator is a measurement of that provides direct correlation to influencing an outcome. On retention, one leading indicator might be the percentage of PT’s at 90 day hiring point that you re-recruited. Re-recruitment is an excellent technique for retention. It can mean you formally do something for them at that point (e.g. celebrate first 90 days with a Starbuck card!), a period of time where they give their supervisor and company a review of how well they onboarded, and perhaps a check in from their supervisor on their development plan and what support they need going forward. As mentioned in last post, several of these are documented ways to support employee engagement. This technique has both high touch and high technology as most HRIS type of systems can trigger a reminder to the supervisor that the 90 day period has been reached and might even be able to auto send a congratulations to the PT on their 90 day anniversary. In my experience, it has to be coupled with high touch to be effective.
And finally, retention as a system while necessary does not solve or mask a bigger issue-proper hiring, something we will explore in future posts.
Thoughts?
What retention strategies do you find helpful?
Please consider subscribing to my Substack, Medium, or view on EIM’s blog where prior posts for years can be accessed.
larry
