I don't know what approach they seek, but my approach (potentially going beyond year 9 or 10 Maths*) is as follows:
Rearranging the given equation to make 

 the subject, I get that

and sketching this gives a rectangular hyperbola with asymptotes at 

 and 

.
View attachment 32730
From this graph, a few things are immediately clear:
- as 
 increases, 
 decreases 
- we can only have both 
 and 
 for 
 and hence the minimum possible value for 
 is 3. 
- as 
 increases, 
 approaches the value 1.5, and so I can find an upper bound for 
 by examining the case 
: 
 \qquad \text{provided $x \ne \cfrac{5}{2}$, so we don't multiply by 0} \\ 3x + 1 &= 4x - 10 \\ 1 + 10 &= 4x - 3x \\ x &= 11 \end{align*})
 
- So, if 
 then 
 (as 
 decreases when 
 increases), and since 
 is bounded below by 1.5 for positive 
, 
 cannot be an integer if 
.  In other words, all solutions must satisfy 
. 
- We already have one solution, 
 and 
, for which 
 
We could now find all solutions by trial-and-error by sequentially testing 

.  This gives us that 

 and that 

, and thus

So, there are two solutions:

If the requirement that 

 and 

 be positive is removed, but they are still required to be integers, two other solutions appear:

but it does not add any extra possible values for 

, which can only be -7 or 9.
* A note on "potentially going beyond year 9 or 10 Maths"...  I originally defined a function 
 = x - y= x - \cfrac{3x+1}{2x-5})
 which has an oblique asymptote of 

 and then used the fact that 
)
 must be an integer to constrain the possible domain of 

... once 
)
 is sufficiently close to the asymptote, any integer value of 

 will give a non-integer value of 
)
.  As I was typing this up, however, I realised that I could simplify the method by looking at just the above hyperbola, which makes the level of the Maths involved suited to junior high school so long as hyperbolas have been covered.