I've noticed a lot of people still doing repeat 400s on the track.  This type of interval session became popular half a centruy ago and unforntunately is still all to common.  To train for the mile it is one thing (~25% of the race distance) but for many of these people I see training for 5-10km on the roads, it seems a bad idea.  If not bad, at least sub-optimal training.  A common workout is 12 x 400m at or slightly faster than goal 5k race pace.  This is 4800m worth of hard running which is good but still less than race distance.  The problem is that the 5k is, for well-trained athletes, a VO2max event.  By that I mean, one spends a large portion of the race at VO2max.  Therefore, interval training at 5k pace is important for a number of reasons: (1) it teaches biomechanics and efficiency at race pace, (2) the athlete learns to run at that pace and mentally become accustomed to the stress of that intensity, and (3) training at this pace helps to optimize VO2max. 

So why do I suggest that 400m intervals are a path to nowhere, besides the literal fact that you end up in the same place you start?  These short intervals to meet the criteria to enhance efficiency at race pace (number 1 above) and the athlete is learning to run at race pace for short distances.  However, the athlete only needs to focus for a short period of time (65-90 seconds).  More importantly, compared with longer intervals the athlete spends less total time at VO2max.  For example, an athlete that runs 75 seconds for the 400m might only be at VO2max for on average 60 seconds each interval (less during the first intervals and more by the last intervals).  This would be a total of 12 minutes at VO2max while covering 4800m.  This might be done on 60 seconds jog recovery.  If the same athlete were to do 6 x 800m (same total of 4800m), he would be at VO2max for an average of 2:15 seconds during each interval (again on average) if he used a 2:00-2:15 recover jog.  This means that the athlete would spend 13 and a half minutes at VO2max over the same 4800m.

It is true that an athlete could spend more time at VO2max during the 12x400m workout by shortening the recovery between intervals and thus getting to VO2max more quickly during each interval.  However, I think that doing longer intervals still retains the advantage of forcing longer periods of mental focus that is critical in a 5k, especially on the roads where there aren't constant splits like on the track.  An additional risk of doing 400m intervals is the tendency to run too quickly (faster than 3k or 5k pace).  Jack Daniels states in Daniel's Running Formula, "Running faster than vVo2max cannot produce a greater aerobic involvement than does running at vVo2max.  The stress of shorter intervals comes from shortening the recoveries, not from running at faster speeds."  I am not saying there are not times when running faster than 5k pace is beneficial but the workout should be designed for a purpose and the pace shouldn't be faster simply because of a short interval and lack of control.

Running 400m intervals at goal 5k pace can be beneficial for an athlete that has not trained at VO2max in a while and is getting back to interval training.  It is a gentler way to ease back into this type of training that repeat 1200s.  However, I suggest these athletes to increase the interval length rather than the pace as they get in shape.  Obviously, an alternative would be to start with longer intervals at a slightly slower pace and then increase the pace as fitness level increases.  I think either way will work but running a bunch of 400m intervals at 5k pace or slightly faster is not an ideal way to improve 5k racing in the long term.