One technical problem -- this couldn't be tested in MT4/MT5 Strategy Tester.  Strategy Tester only supports a single time frame.  If you don't care about Strategy Tester then that's not an issue -- but I suspect most of us use it as a "sanity test" to confirm the portfolio EA we just built from *.mq4/*.mq5 source code executes as expected.

Alternatively, you could "simulate" different time frames by fiddling with the 'Period' setting of certain indicators.  Personally I think that is too much trouble and have found that things don't necessarily scale as expected.  In your example the "base" timeframe would be M15 and then you would use 4x (for H1) and 16x (for H4) multiples of the 'Period' setting -- something like that.

All timeframes use the exact same tick data -- with one exception, the Open tick.  MT guarantees an Open tick at the beginning of a time frame regardless if the price has changed.  As a result, over the course of an hour an M15 strategy will receive 3 more ticks than an H1 strategy.  This may or may not affect performance -- just something to think about...