Here is an interesting analysis of cancer treatment for three common cancers - breast, bowel and lung. The dataset covers admissions to any hospital in England over a 7 year period. The results show a trend against people from deprived areas getting the latest techniques but to my mind (it's a matter of opinion) the differences are actually quite small.
The description of method, and discussion of the results, are worth a careful read - they are a tutorial in strengths and weaknesses of routine data. Don't worry too much about the statistical techniques - multilevel modelling was used. This is needed because patients are clustered by virtue of the hospital they are treated in. A hospital policy may affect all patients as a group. Hence multi-level - in this case two levels, namely individual (age, sex etc) and group (hospital). If this had been an international study we might have needed a third level (country). The mathematics of multilevel modelling are horrendous.
Note also a subtlety. Each individual is assigned his or her own deprivation score, so it counts here as an individual level variable. But the score is actually an area score - assigned accordingly to the characteristics of the area of residence. Everyone in that area gets the same score.