DB, I don't think "below 1,000 feet amsl" comes into the practical equation. The 1,000 foot rule is a legality to ensure en route terrain and obstacle clearance under IFR.
The problem here was that a transition from hovering using external visual cues (whatever they were, most probably very limited), to an instrument climb, wasn't successfully completed.
As I often repeat: Hovering isn't too difficult. Instrument flying isn't too difficult. It's the bit in between that catches people out, time and time again. Both departing from and arriving at terra firma, as recent tragic accidents and recorded incidents have proven.
Strangely enough (or perhaps not), it's also the bit helicopter pilots aren't well trained for, at least, not in the civilian environment, once away from the "luxuries" of an airfield.
This often gets confused amongst the argument for and against Class 1 takeoff performance. Some seem to miss the obvious, i.e. if Class 1 is required, it's obvious that sufficient visual cues must be available in order to allow a Class 1 departure to be made. This requires sufficient visibility to assess an adequate clear departure path and a sufficiently high cloudbase. At TDP the pilot must be able to see the ground and the reject area ahead, in case an engine quits at the most difficult moment (i.e. just before TDP).
In marginal weather conditions, Class 1 performance for departure (and landing!) may have to be sacrificed altogether. But in truth, full Class 1 is seldom available at remote landing sites, at least, at ad hoc ones. And that, unfortunately, is the core business of the onshore corporate helicopter market.