While I have not had direct experience with risk management or the scenarios you mentioned above, I think the examples in the lecture (using a scalpel for cutting) were great on depicting what kind of risk is worth the reward. Another great example where a risk is worth the reward would be with life saving surgeries where there are life threatening side affects and conditions. Many of these surgeries can result in mechanical failure of ventilation or even brain trauma, when you think about these risks they are the worst kind because they can result in death, however these surgeries are performed because the chance of success and saving a life are higher than the risk. On the contrary, a scenario where its better to end production rather than mitigate or accept the risk would be a medical drug that has great results in fighting cancerous cells but also causes issues with bone density loss and even internal bleeding. Due to the fact that there are other treatment options that are safer and longer proven, it would probably be best to end the production because the risk is higher than the "reward" in this situation. I hope these two examples really helped shed some light on the two types of scenarios and I hope to gain some experience of these in the future during my risk management industrial experience. 

As mentioned by @ms2768, the scalpel example from the lecture depicts what risks are acceptable and which ones must be mitigated. I believe that a simple way to differentiate the risks are to ask whether the risk is integral to the function of the product. If it is integral, then consider whether a modification can be made to the product to mitigate the risk without affecting the function. For example, a fan has a similar risk to the scalpel, with the blades being dangerous to people. However, unlike the scalpel the fan's risk can be mitigated by including a cage around the blades. This cannot be done with a scalpel because if anything prevents the blade from cutting people then it will also impede the scalpel from performing its purpose. 

I liked how @DevDesai compared the scalpel from the lecture to the dangerous fan and how the risks for the fan can be mitigated while the risks for the scalpel may not be. In this case the scalpel's risks are not worth the reward and should be cancelled. Another interesting realm that could be applied to this situation are experimental treatments. In cases where patients suffer from a severe disease there may only be an experimental option left for them that could be risky in terms of the success rate, however, without the procedure the patient may pass away. In this scenario, the reward would be greater than the risk as the only two options for that patient would be to get better through the procedure or pass away from the procedure OR pass away without the procedure. In either choice there are chances of passing away, however, by undergoing the procedure the patient has a chance of living. 

On the other hand, an example of when the risk is not worth the reward is when a company has a product that could treat patients with an illness, however, the illness is not life threatening and there are other ways to treat it. For example, if another variant of COVID arose the immediate response would be to tweak the current vaccine treatments to accommodate for the COVID variant. In this situation the current vaccines help mitigate the virus but not provide full immunity, however, the tweaked version of the vaccine provides full immunity but has a side effect of uncontrollable twitches. It would not make sense to take the risk of the tweaked formulation as there already is an effective mitigation vaccine on the market. The risks would outweigh the rewards in this case because it is not worth gaining twitches and full protectiveness when there already is a vaccine than can lighten the effects of the virus.

Figuring out the probability and severity of the product's risks (AKA Risk Evaluation) is another way to tell whether the risks are worth the benefits. Even if you have some risks that occur frequently and some risks that are life-threatening, which will be a given when dealing with medical devices, it should be manageable as long as the two do not overlap significantly. However, having many risks that are both high in probability and severity should be a red flag to the risk management team that the product is just too risky to justify the potential benefits. A Risk Matrix is a useful tool for determining how many risks have both high severity and probability.

If there is a situation where a product keeps breaking but doesn't cause any harm to the consumer.  Where maybe it doesn't last as long as thought, then I say you continue with the product.  I there is a defect on a product that at any instance could break and cause harm such as pain, permanent damage or death to the consumer, then this is a definite sign to end production.