A couple of weeks ago, I asked ChatGPT to write an exercise programme for my client. I then compared it against what I have planned for the client for the session. As a chiropractor, I can say the result was actually not bad. Out of the five exercises I asked for, three were usable. However, there were other issues that made ChatGPT an unlikely replacement for our services at Square One Active Recovery.
This article will solely focus on ChatGPT’s ability to generate a meaningful exercise programme. For perspective on if ChatGPT can replace a musculoskeletal healthcare provider wholesale from a diagnosis to treatment plan, you might want to check what this physiotherapist had to say.
Writing good prompts is all the rage for ChatGPT
From what I have seen on TikTok and other social media platforms, the quality of prompts used will heavily influence the quality of results you receive. For the purpose of this experiment, I am going to keep it simple by asking ChatGPT to ‘write a 12-week rehabilitation programme for low back pain’.
Being able to write effective prompts will really come down to having subject matter expertise in the results you are after. Given most of you will not be experts in chronic pain or injury management, I have kept the prompt simple. My suggestion is for you to use my feedback later to tweak the results for your own case.
Personalisation is key for best outcomes
This is probably true even outside of a chatGPT situation. Even in real life setting, cookie cutter approaches yield the poorest outcomes. Therefore, being able to tweak the programme to your specific condition is going to be extremely beneficial.
The importance of individualised is acknowledged by the AI chatbot. It starts by telling users that the programme “may need to be modified to suit individual needs and circumstances”.
It also recommends that you should “consult with a qualified healthcare professional before beginning any new exercise program”.
True on both counts. We are off to a great start.
Making sense of the structure of the exercise programme
I am not a tech person so I cannot tell you exactly how the AI mechanism works. From what I have read and understood, ChatGPT combs online content and repackage them into direct responses to your query. I am going to assume this is how it derived the exercises as well.
The exercises are sorted into a three-part programme. The first part is for two weeks, the second for four, and the last for six weeks. Broadly speaking, that does things easier so you can definitely stick to the same structure for your own rehab.
In our private practice, we follow a test-retest process so our exercise programme changes based on how our clients respond. In other words, the programme change on a session-to-session basis. Quite often, we also modify the programme mid-way through a session because of what we could observe.
Given that you will not have the expertise to assess and evaluate yourself in real time, we suggest sticking to a generic programme structure. Again, ChatGPT’s 2-4-6 weeks structure is not a bad idea.
If you want to go for something more responsive (i.e., more changes in the programme), you can try a 2-2-4-4 format. This means changing up the exercises at the end of week two, week four, and week eight.
Goals and objectives are not required for each stage of recovery
What I personally like about the AI generated programme is that it states what are the goals for each stage of its programme. It gives perspective for users, and create meaningful for these exercises. However, there is no research to suggest that these mechanisms mediates the response we see in recovery.
To put that in other words, we don’t know how or why exercise works for chronic pain. But we do know that exercise is beneficial for your recovery.
A lot of clients find that hard to accept. We live in a society where we question why a lot and we also want to know why. The truth of the matter is that sometimes science and technology is not advanced enough to tell us why.
For example, we still don’t fully understand why planes can stay in the air! Despite this knowledge gap, we are still pretty damn good at keeping ginormous planes in the air. The moral of the story? We don’t need to know why it works to make something work!
Let’s talk about inflammation
One of the goals in the programme is to reduce inflammation. The “inflammation” word is thrown around a lot. I do also use it in clinical practice. Having said that, I don’t think it’s something that we should excessively focus on. Yes, it’s true there is some kind of inflammatory involvement for chronic pain and definitely for acute injuries. This is a normal physiological response.
There is no need to actively try to beat the inflammation. Inflammation is not always bad for you.
A paper published earlier this year found that the use of anti-inflammatories such as ibuprofen can prolong pain and inflammation. This effectively leads to a delayed recovery. In other words, it can cause more harm than good.
On that note, no clinical guideline in the world recommends NSAID (or any form of medication) as a first-line treatment for chronic pain. Hard to believe, I know. But yeah, you don’t really need to be popping by pills to feel better.
If anything, not popping pills can actually work to your advantage!
Introduction to exercising for low back pain
There are four exercises listed in the first stage. They are pelvic tilts, cat cow stretches, bridging, and side planks. The recommended dose is 10-15 repetitions for two or three times per day.
I am not going to go through the exercises because you can easily look it up on YouTube.
Four exercises is a pretty good number for a start. Typically we give about two to three exercises for people to do at home so four is pretty close. We try to limit to three because the more you give, the less likely people will do them. Yes, adherence is also an important consideration.
10-15 repetitions is kinda okay but honestly not optimal at all. The overall volume is on the low side? And asking someone to complete 15 repetitions per day may be okay for the first couple of days, but may flare things up after one or two weeks. This is where individualisation is important. How much you should be doing is dependent on how much you can do. To determine how much you can do, you will probably need a professional assessment.
So, instead of 10-15 reps, I would prefer to break it down to three sets of six repetitions. With this format, you can get a rest after six repetitions so the irritation to the spine, if any, is minimal. Furthermore, you can get more repetitions in per day (i.e., 18 instead of 15 repetitions).
When we talk about sets x repetitions, we have to discuss rest. Rest can be anything from a couple of seconds to two to three minutes. This is pretty much the standard in the exercise scene.
For recovery, rest can be as long as two to three hours or even six hours. How much rest you should be taking will depend on how much rest you need. In order to ascertain that, a professional assessment is ideal.
No pain doesn’t mean you should do more. Pain doesn’t mean it’s bad for you.
This is a good time for a segue to talk about pain. A lot of “concepts” or narratives pertaining to pain is inaccurate. I don’t just mean inaccurate on social media but also among exercise professionals and even healthcare providers.
Pain science is quite a new field and very poorly understood. One of the major research-supported tenets that is often by social media and pain sufferers aside, is that pain doesn’t equal to damage. Similarly, damage doesn’t equate to pain.
For this reason, it is very difficult for a layperson to be able to determine what is an appropriate exercise dose and rest for their recovery. Having said that, without a comprehensive assessment and working history (i.e., historical data), it is also very difficult for healthcare providers to make meaningful recommendations. This is why having a working relationship with your healthcare provider is important even for exercise for chronic pain.
Load management can make you or break you
This is technically the second segue but also very important so we will go for it. Load management can be thought of as how much stress you are applying to your body in a given period as well as how much rest or recovery you are having in the same time. Because, again, pain is not an indicator of whether you are applying enough or too much load, it is a fairly challenging concept for most people to grasp.
Because of the poor or inaccurate understanding, we often see poor exercise outcomes. This can range from overtraining injuries, recurrent pain, poor health or fitness improvements (i.e., exercising but not getting any benefits from it).
When we talk about load and stressors, we are not only talking about how much weight or how much volume (i.e., how many repetitions performed) alone. It can also include fatigue, poor motor control, anxiety, emotional state, past injuries, etc. Similarly, recovery is not just about how much rest you have. It can also be your sleep quality, nutrition, hydration, and a whole bunch of other factors.
It’s very important to take a big picture view when it comes to exercising. Having an appreciate that your body is not a car or a machine is extremely important. We are not always the same person today as we are yesterday or the day before. While listening to your body alone is bad advice (e.g., no pain doesn’t mean you can pushing), it is still important to take a holistic approach when trying to determine how much exercising you should be doing for that day or week.
This post is hitting 2000 words so I will leave it here. In the next instalment, I will be discussing ChatGPT’s choice of exercises as well as my evaluation for stage two and three of their recovery programme.
WANT TO GET STARTED IMMEDIATELY?
Frustrated by the lack of results-driven and ethical chiropractic clinics in Singapore, Chiropractor Jesse Cai found Square One Active Recovery to deliver meaningful and sustainable pain solutions.
Our goal? To make our own services redundant to you.