Ok, so let's compare this to a comparable power.
The Human Heroic Effort power.
For this we'll take two Rangers, at L1 since that's the best time to compare racial powers and build them basically exactly the same. To make this completely fair, we'll burn the Human's extra feat on something worthless so that it doesn't color our findings (in reality, it likely means the human has better weapons at L1).
So we'll take the Heavy Blade Ranger, weilding dual long swords with a 20 str, taking the Heavy Blade Expertise feat and a 20 str to start.
At L1 our to hit and damage are as follows:
To Hit: 3 + 5 + 1 = 9
Damage on single TS is: 1d8
Quarry Damage: 1d6
This means both characters, without using their racial powers have the following average damage:
.7 * 8 + .05 * 14 + .7 * 4.5 + .7 * (.7 * 3.5 + .05 * 6) + .05 * 8 = 11.78
Now, let's introduce the racial powers into this. The shifter gets the following average damage for probably half the combat:
.7 * 10 + .05 * 16 + .7 * 6.5 + .7 * (.7 * 3.5 + .05 * 6) + .05 * 10 = 14.78
With this information we can determine who gets the better end of the damage stick. Let's assume that a combat goes 5 rounds. On average our rangers get 10 attacks in those 5 rounds (5 uses of twin strike, we'll ignore action points for now).
We know from our accuracy numbers that our Rangers hit 75% of the time (average L1 monster AC is 15). So we can expect to hit 7.5 times. Since our fighter is going to get an extra boost from heroic effort, he's going to hit an extra time so 8.5 times. Let's run our average damage calc using 85% instead of 75%:
.8 * 8 + .05 * 14 + .8 * 4.5 + .8 * (.8 * 3.5 + .05 * 6) + .05 * 8 = 13.58
We'll also assume our Shifter gets bloodied after his turn in the second round. So He will get 6 attacks after his minor action has been used. So he'll get 3 rounds of his top level average damage and 2 of his normal for an actual average damage comp of:
.6 * 14.78 + .4 * 11.78 = 13.58
So there's that. The average damage lines up *perfectly. There is literally no difference between getting +2 to hit and a free hit turned into a miss over the course of a 5 round fight with a .75 hit chance. (This is mostly a statistical anomaly based on the length of combat I chose, obviously with more attacks, the +2 becomes more significant and a single attack changed to a hit becomes much less significant. The fact that I chose 5 rounds is not an accident though, that's about how long combats tack in my experience)
Now, what about the regen? Again, that's literally 6 hit points, and when you compare it to Heroic Effort's flexibility (it's true purpose is to keep you from failing death saves), it seems like a pretty legit compromise.
For kicks and giggles, let's assume that a combat takes 10 rounds. We'll still allow our Ranger friend to get his encounter power off during the third round of combat. So now he gets 16 attacks at the higher damage margin with four at the lower. The Human ranger gets one attack in 20 changed to a hit, so a 5% increase in accuracy. The two equations become:
.75 * 8 + .05 * 14 + .8 * 4.5 + .75 * (.75 * 3.5 + .05 * 6) + .05 * 8 = 12.89
For the Human and
.8 * 14.78 + .2 * 11.78 = 14.18
for the Shifter. In other words, a DPR difference of about 1. That's 10 damage difference (and at most 16 HP) between Heroic effort and Longtooth Shifting over a 10 round combat. That's not even a quarter of a monster's HP at this level and certainly not enough to convince me that the power needs rebalancing (For even longer combats, the DPR difference will approach 3 asymptotically).
What does 16 HP matter in a combat. The average hit from a L7 standard monster using MM3 math is as follows (Assuming PC AC of 21, which is right for a Ranger with starting Dex of 18):
.45 * 15 + .05 * 20 = 7.75
That means if you get the entire full effect of the regen you're going to stand up to 2 hits from a standard monster. Put another way, this is about your surge value at L7, so you'll save yourself 1 healing surge. That's a solid savings, but going back to our comparison to Heroic Effort, I certainly wouldn't say it's imbalanced (you also have to take the entire race's feat support and other features into account, you can balance a stronger race against a weaker one with a strong racial encounter (for instance, we assumed our Human ranger spent his feat on something that didn't help, but he could have spent it on an expertise feat and gotten an accuracy boost that would even the odds even more, he also could have taken a third at-will instead of Heroic effort which would make this entire exercise pointless, this is offset to some degree by slightly lower stats, but in other ways not so), but that analysis would be quite long and is out of scope here).