General Demographics
Of a total of 125 candidates, 74 students completed the survey (response rate 59.2%). There were 43 (58%) female participants, and mean age was 23.1 ±3.8 years. Seventy-one (95%) students were in their 3rd year of medical studies, while 2 (3%) were in their 4th, and 1 (2%) in his 5th year. Thirty-nine (53%) were in the CDL, 35 (47%) in the PFDL group. There were 38 surgical discharge letters (21 CDL, 17 PFDL) and 36 medical discharge letters (18 CDL, 18 PFDL, see Figure 1). There was no significant difference regarding gender, age, year of medical studies or the distribution of surgical/medical letters between the two study groups (all p > 0.05). Raw data of the survey results is provided in additional file 1.
Results of the survey
Overall, the PFDL group rated the discharge letters significantly better than the CDL group (median [IQR]: 1 [1-2] vs. 3 [2-3], RI = +2, p < 0.001). Separate comparison of subsections yielded similar results, which were median 1 (IQR 1-2) vs. 2 (1-2) for section ii (structure, Cohen’s d = 0.5, p = 0.005), 1 (1-2) vs. 3 (2-3) for section iii (content, d > 0.8, p < 0.001) and 2 (1-2) vs. 6 (5-6) for section iv (patient-friendliness, d > 0.8, p < 0.001) in the PFDL and CDL group, respectively. When questions were analysed one by one, the PFDL group rated their letters significantly better in 12 cases and never worse than the CDL group. In questions with significant difference, Cohen’s d indicated small ES in one case, moderate ES in four, and large ES in seven cases (for detailed results see table 1).
Additionally, questions with the largest ES were identified in each section. In section ii (structure), ES was greatest for question 2.3 (“Does the content justify the length of the document?”, d = 0.74, moderate ES). In section iii (content), ES was greatest for question 3.10 (“Is there information on possible allergies?”, d = 4.40, large ES). This is unsurprising, as information on possible allergies was entirely missing in both original discharge letters, but was given in the new letters. In section iv (patient-friendliness), question 4.2 (“Would the indication of the prescribed medication be comprehensible for a medical layperson?”) had the largest ES (d = 3.23, large effect).
Subgroup analysis
The subgroup analysis revealed that students given the surgical PFDL rated their versions significantly better than the surgical CDL group in 11 cases, whereas students given the medical PFDL rated their versions significantly better than the medical CDL group in nine cases. Section by section, participants with the surgical PFDL rated their letters significantly better than the surgical CDL group in all sections (ii: p = 0.005, d = 0.89; iii: p = 0.006, d = 1.01; iv: p < 0.001, d = 2.53) and overall (p < 0.001, d = 1.24), whereas the medical PFDL group rated their letters significantly better than the medical CDL group overall (p < 0.001, d = 1.50) and in sections iii (content, p < 0.001, d = 1.50) and iv (patient-friendliness, p < 0.001, d = 3.14), however not in section ii (structure, p = 0.988, d = 0.09) for detailed results see Table 2).
There was no significant difference regarding mean age, gender or year of study between the PFDL and CDL groups within the surgical and medical subgroup. Furthermore, there were also no significant differences between the surgical and medical subgroups regarding demographic parameters (see Table 3).
Results of open-answer question
CDL
In the CDL group, 9 of 39 participants (23%) gave written feedback concerning the properties of the discharge letters. Thereof feedback was positive in 1 case, neutral in 1 and negative in 7 cases. Thorough analysis of the given answers revealed that the liberal use of abbreviations was a major concern and cause for negative feedback in 5 cases. For instance, one candidate wrote the following:
“Personally, I believe that the term NSAID should be avoided in referral letters. Lays do not know what that is, while over-the-counter medications [e.g. NSAIDs] can be found in almost any household. …”
In that case the term NSAID was used when informing a patient with liver cirrhosis and upper gastrointestinal bleeding, that the intake of NSAIDs should be avoided in future. Another participant wrote:
“Abbreviations in the list of diagnoses are o.k., however, when abbreviations are used, they should at least be explained in the text.”
The basic tenor seemed to be that the letters contained much unnecessary and too little necessary information to foster the patient’s comprehension, using too many technical terms and abbreviations, or, according to 2 further comments:
“Therapy recommendations are missing – even if e.g. no drugs were prescribed, this should be mentioned in the discharge letter. …”
“The letter is written as to meet the general practitioner’s needs, but is not sufficiently extrapolated to be understood by lays.”
The only positive feedback stated that the structure was good and the discharge letter very detailed.
PFDL
In the PFDL group, 11 of 35 participants (31%) gave written feedback. Feedback was considered positive in 6 cases, neutral in 2 and negative in 3 cases. Among the positive feedback, three main categories could be identified: structure of the discharge letter, explanation of technical terms and the new list for drug prescriptions. The following statements may serve as examples for each of these categories (some statements referred to 2 or all 3 of them):
Structure/explanation of technical terms/drug prescriptions:
“Good structure, and it is very positive that at there is a simple explanation for lays at the end of each paragraph. Very positive also the additional information on medication.”
Drug prescriptions:
“I believe that the discharge letter is good. Especially the table with the medication is very helpful and I have never seen anything alike before.”
Explanation of technical terms:
“Through information in brackets in ‘non-medical language’ very well understandable for lays.”
Neutral feedback included both positive and negative issues. For example, one participant raised the concern that relevant information could get lost through simplified language, while nevertheless deeming the discharge letter as a whole very patient-friendly:
“To me, not all of the [technical] terms were explained in a comprehensible way. For instance, ‘post-ethylic liver cirrhosis Child B 7 points’ – given the circumstance that this is the main diagnosis, the provided explanation [status post upper gastrointestinal bleeding] might be a little poor. Otherwise, I believe that the discharge letter is very patient-friendly, especially because further necessary treatments are explained in detail.”
Of the three negative feedbacks, one found additional information on the immediate impact of certain diagnoses on the patient’s everyday life (e.g. splenomegaly) lacking, one suggested more background information on prescribed drugs and one raised concern about whether the new discharge letter would be feasible in clinical practice.