If we consider the scenario of Mouse Utopia - i.e. a massive reduction in human fitness due to mutation accumulation due to the removal of the main mechanism of generation-by-generation, mutation-purging natural selection (mainly the reduction of average child mortality from probably more than fifty percent to one or a few percent) - then there seems little or no grounds for realistic hope of avoiding a mega-death outcome.
(...Indeed, not merely mega- but a giga-death outcome, involving not just millions but billions of humans - because there are currently about seven billion humans compared to the recent historical global population of about one billion - and the Mouse Utopia scenario predicts that human fitness will drop below historical levels; presumably implying a lower than historical population. )
One particularly worrisome aspect is that the mutation driven incremental decline of human capability (due to an increase in human pathology) creates a positive feedback situation in which humans presumably would get less-and-less-capable of solving the multiple damaging consequences of mutation accumulation; including the problem of mutation accumulations itself.
This, at least, seems almost inescapable from the perspective of individual level selection; but if group selection mechanisms are real and applicable (as I believe they are), then it may be that behaviours will emerge that tend to counteract mutation accumulation, and avert the mutational meltdown positive feedback cycle.
But group selection is rather poorly understood, and perhaps works by multiple pathways. At any rate, by comparison with individual-level selection, group selection would appear to be ('as if) goal-directed, cognitive in nature, altruistic, and with foresight to look beyond short term individual reproductive disadvantage to long term group genetic benefit.
(Group selection may actually, or may not, actually be these things - teleological, cognitive, altruistic, predictive - but at any rate, this is how it would appear to us.)
So group selection might lead to a variety of quasi-purposive mutation-purging, fitness enhancing outcomes. These would not, of course, necessarily be conscious - indeed would likely be more effective if unconscious.
Possible examples might be an instinctive self-elimination of heavily mutated individuals from reproduction - varieties of genetic suicide. So, ordinary individual selection might tend to reduce reproduction of heavily mutated individuals due to them suffering from the effects of pathology - but in addition there might be some kind of system of internal monitoring of mutation load which made heavily mutated individuals 'give-up', lose interest in reproduction, avoid reproduction - perhaps at extremes allow themselves to die from starvation, infection or some other cause?
This would amount to a kind of apoptosis at the organism level (apoptosis is the biological process of cell-suicide, which may be triggered in genetically-damaged cells - for example to reduce the incidence of cancers).
Alternatively, group selection may lead to the emergence of a higher incidence of the kind of human individuals who - by their dispositions and behaviours - have the best chance of solving the mutation accumulation problem in some way. For instance, it might lead to an increased number of problem-solving
WD Hamilton believed-in and wrote about such mechanisms of positive self-elimination, and argued that their existence is theoretically-predicted as well as consistent with some observation - so we should not regard them as implausible - even though they not well understood, and hard to discriminate from the passive consequences of mutational damage.
(Of course, group selection is not - contrary to some views - nicer than individual level selection; because group selection involves in some way the sacrifice of the individual for the genetic benefit the group.)
Nonetheless, if group selection is active, a range of such possibilities may already be at work, or may become more powerful - and this may affect the development and outcome of the Mouse Utopia scenario to a greater or lesser extent.
Note: Some of these ideas were developed in conversation with Michael A Woodley of Menie, to whom acknowledgement and credit is due.