Asssessing sensitivity analysis in patient satisfaction: a case study from "Luigi Sacco" Hospital

Despite of the recognized relevance of conceptual models as a theoretical reference in patient satisfaction studies, in some practical circumstances they are too complex to be applied. This is particularly true when patients are aged, so that higher risks of misunderstanding and consequent biased answers would suggest employing simpler data collecting procedures. These considerations suggest the importance of employing statistical methods capable to deal with data collected according to quality evaluation systems that are necessarily simplified. By starting from a survey carried out experimentally on a set of patients admitted in 2003 at Sacco Hospital in Milan, in this work we have relied on a combination of statistical methods for: 1) constructing patient typologies, by using some indicators of clinical picture complexity, effectiveness of given treatments and perceived quality, this latter in the binary form Satisfied/Not satisfied; 2) performing a sensitivity analysis, to assess if results derived from analysis are sufficiently stable, and then reliable; 3) carrying out a more thorough inspection on the underlying data structure by referring to the results derived from sensitivity analysis.