It would appear that in both studies, the categories involve the same mixture of treatments and treatment targets that is found in much more detail in the PBE studies. Hart et al93 have presented reliability

and validity data on operational definitions of learning-based treatment contents in TBI rehabilitation. They used a bottom-up process to develop a classification of skilled performance training, with a dividing line between treatments targeting function (more or less equivalent to the ICF concept of bodily function) and treatments aimed at altering ICF activity. In the terminology of DeJong,2 the PBE methodology and similar approaches to classification of rehabilitation interventions are an experience-driven, bottom-up, inductive method guided by front-line clinicians’ opinion and by scientific selleck screening library evidence, where such is available. A perusal of any rehabilitation journal will indicate that studies evaluating treatments are increasing in number but are still relatively scarce7, 8 and 94; the articles that are published tend to lack qualitative and quantitative specification of the ingredients of the treatment provided.9 Quantification of the amount of treatment, other than by gross indicators (eg, length of stay or number of sessions), is largely absent.7 Until recently, the most sophisticated studies used

hours of therapy provided by specific disciplines1, 12, 95, 96, 97 and 98 or number of visits.99 However, given that every rehabilitation discipline may deliver multiple interventions and that different disciplines may deliver the same Integrase inhibitor intervention, it is not surprising that

these studies have not been very effective at explaining differences in outcomes, either among therapists or among programs. For instance, analyses of the data of the inpatient stroke rehabilitation PBE study2 suggest that spending more time per day in PT and OT is not associated L-gulonolactone oxidase with better outcomes. However, when PT and OT are differentiated into specific treatment activities, there are significant improvements in outcome prediction.100 For instance, patient characteristics by themselves explained 40% of variance in discharge FIM motor scores for moderate stroke and 45% for severe strokes. When total PT and OT treatment time was added, this did not result in a significant increase in variance explained. However, when total time in specific OT and PT activities was added to the regression equation, the percent of variance explained increased to 52% and 68%, respectively.101 The PBE studies have taken advantage of the fact that therapists completed specially developed forms after every treatment session on which they noted not only what activities were delivered, but also how much time (in multiples of 5min) was dedicated to each. A more detailed view than in the older studies, which only had administrative data on hours by discipline, was possible and has been applied extensively.