I was heartened to see an article evaluating the Royal College of Psychiatrists’ patient information leaflets using quantitative and qualitative methods. ^{Reference Briscoe, Briscoe, Timms and Ramsay1} The provision of information is critical to my clinical practice and has often involved these very leaflets. I was also pleased that the authors acknowledged that ‘much patient information is written in complex language and is poorly presented’ as these are often barriers to patients accessing information. Disappointingly, however, they did not conduct any analysis of the language; one respondent had commented regarding one leaflet that ‘It has quite a high reading age’.

The complexity of language can be assessed using a range of readability measures such as Flesch Reading Ease (FRE; a document should have a score of greater than 60, the higher the score the easier it is to read) and Flesch-Kincaid Grade Level (FKGL; refers to US school grades, so lower scores indicate better readability - a 13 year old should understand a document scoring 7). These are widely available, contained within word processing packages, and have been used to evaluate patient information leaflets in other specialties ^{Reference Briscoe, Briscoe, Timms and Ramsay1} and standard appointment letters in child and adolescent mental health services. ^{Reference Payne, Large, Jarrett and Turner2} When these measures are applied to the College leaflets (Table 3 in the paper), the mean FRE is 7.81 (7.1-8.4) and mean FKGL 63.13 (58.7-69.8). This suggests the leaflets are readable as far as these computerised measures are concerned but their readability could be improved. When the top- and bottom-ranked leaflets (Table 3, which, curiously, has four highest ranked and three lowest ranked rather than four of each as described in the text) are compared, there is no statistical difference on either of the measures. This confirms that, although the language may be readable, the reader may not like the content.

I was confused by the quantitative method employed in the study. The original feedback was on a 5-point Likert scale ranging from ‘strongly agree’ to ‘strongly disagree’. These are ordinal variables (variables which represent categories of a feature with some inherent ordering ^{Reference Bennett and Gilchrist3} ); however, they were converted into continuous variables (one which can take any value within a range ^{Reference Bennett and Gilchrist3} ) and analysed as such. Unfortunately, one cannot convert discrete categories into a linear scale in this way. Given this conversion, the values could only range 1-5, and it is unsurprising that the authors found there was little variability in the feedback ‘scores’ assigned to each leaflet. It was also confusing to find that a correlation between modalities was included in the discussion but not presented in the results. My understanding of the analysis would have been aided to see the information presented in the original categories which those reading the leaflets had decided.

Despite these potential improvements and confusions, the conclusion remains undoubtedly true that ‘reader feedback provides invaluable guidance about the substance and presentation of our public mental health information.’ One can only hope that we continue to strive to produce information which is accessible to those who need it.