Is the bystander effect a myth?

I would say no, it is not a myth - Meta-analysis

The present work (a) integrates the bystander literature from the 1960s to 2010, (b) provides statistical tests of potential moderators, and (c) presents new theoretical and empirical perspectives on the novel finding of non-negative bystander effects in certain dangerous emergencies as well as situations where bystanders are a source of physical support for the potentially intervening individual. In a fixed effects model, data from over 7,700 participants and 105 independent effect sizes revealed an overall effect size of g = –0.35. The bystander effect was attenuated when situations were perceived as dangerous (compared with non-dangerous), perpetrators were present (compared with non-present), and the costs of intervention were physical (compared with non-physical). This pattern of findings is consistent with the arousal-cost-reward model, which proposes that dangerous emergencies are recognized faster and more clearly as real emergencies, thereby inducing higher levels of arousal and hence more helping. We also identified situations where bystanders provide welcome physical support for the potentially intervening individual and thus reduce the bystander effect, such as when the bystanders were exclusively male, when they were naive rather than passive confederates or only virtually present persons, and when the bystanders were not strangers.

Link to Article - PDF Warning!

While Philpot et al (2019) can say they 'disproved' the bystander effect, their research doesn't really paint that picture. Their conclusion:

We argue that it is time for psychology to change the narrative away from an absence of help and toward a new understanding of what makes intervention successful or unsuccessful.

should be noted that psychology does in fact study what makes intervention successful or not by studying the moderators of the bystander effect, noted in the meta analysis linked above:

Moderator Analyses on Variables Related to Emergency Danger

In sum, the bystander effect is reduced when the situation is perceived as dangerous, when a perpetrator is present, or when the focal bystander faces a physical cost of intervention. These meta-analytic findings support recent results reported by Fischer et al. (2006) and others (e.g., Harari et al., 1985). Fischer et al. argued that dangerous emergencies are most likely to induce arousal in the bystander. This argument, which is derived from the arousal: cost–reward model (Dovidio et al., 1991, 2006; Piliavin et al., 1981; Schroeder et al., 1995), suggests that bystanders intervene in part because they seek to reduce their own arousal. The present results suggest that arousal sources related to the perceived danger to oneself in case of intervention seem to explain more variance of the bystander effect than the perceived danger to the victim. Yet, the present findings are only partially consistent with the traditional empathy hypothesis (e.g., Batson, 1998), which assumes that empathy and empathy-induced helping increase mostly with the level of perceived danger to the victim. Finally, the finding that dangerous emergencies lead to increased helping may also be viewed from a social cognitive perspective. That is, danger might first raise the expectation that others will help, which in turn facilitates one’s own decision to help.

Moderator Analyses on Variables Related to“Bystanders as Potential Physical Support”

In sum, this set of analyses shows that additional bystanders can provide social, physical, and psychological support when focal individuals have to decide whether to intervene, and this is especially true for dangerous emergencies. For example, the effect is attenuated in dangerous situations when bystanders know one another, when they are real instead of passive confederates, when they can provide more physical support (i.e., when they are male instead of female), and when their presence is physically real instead of implied. Additional combined moderator analyses showed that these effects are stronger for dangerous (emergency) rather than non-dangerous situations. The latter point in particular supports our assumptions about bystanders as potential physical support; when others are actually there, and not just implied (e.g., via intercom systems), the focal helper can expect assistance from other bystanders in dangerous emergencies. The present metaanalysis as well as studies by Horowitz (1971), Fischer et al. (2006), and van den Bos et al. (2009) provide direct support for this reasoning. The effect of bystanders is not always as negative as traditional research suggests. Additional combined moderator analyses across different numbers of bystanders and different levels of emergency danger revealed support for Latane´’s (1981) social impact theory, which proposes that the bystander effect increases with the number of bystanders. This was found for both dangerous and nondangerous situations.

And finally, in the conclusion:

The present meta-analysis provided clear support for the assumption that passive bystanders in critical situations reduce helping responses (g = – 0.35). Consistent with our theoretical approach, however, bystander inhibition is less pronounced in dangerous than in non-dangerous situations. We argued that in dangerous emergencies, the focal individual may perceive additional bystanders as positive resources for helping. Three psychological processes may contribute to this perception. First, dangerous situations are more likely to be construed as clear-cut emergencies where someone needs help, which increases experienced arousal and thus helping responses. Second, bystanders can be seen as providers of physical support and thus reduce fear of intervention. Third, some dangerous emergencies can only be effectively resolved by cooperation and coordination among several bystanders. Rather counter-intuitively, but consistent with this line of argument, the bystander effect was attenuated when situational attributes referred to increased danger, that is, (a) when focal individuals expected increased physical costs for helping (instead of time or financial costs), (b) when a perpetrator was present (vs. absent), and (c) when the critical situation was recognized as a clear-cut emergency (compared with a non-emergency). Less counter-intuitively but consistent with the hypothesis, (d) the effect was marginally reduced when no male bystanders were present to provide physical support.

So yes, there are some situations where the bystander effect is reduced but it is not disproven.