Table 5 Differences over time in proportions answering knowledge questions correctly (n, %), Friedman's Test
Questions (abbreviated) | Time 1 Pre (n = 114) | Time 2 Post (n = 106) | % change | Time 3 Follow-up (n = 51) | % change | DF    | χ 2    | p    |
|---|---|---|---|---|---|---|---|---|
Q1. A good clinical question consists of an intervention, target population, outcome, and comparison intervention (Agree) | Â | Â | Â | Â | Â | Â | Â | Â |
   Correct | 53 (46.9)* | 105 (99.1) | +55 | 48 (94.1) | +5 | 2 | 44.9 | 0.0001** |
   Incorrect | 60 (53.1)* | 1 (0.9) |  | 3 (5.9) |  |  |  |  |
Q2. Databases of primary medical literature, such as Medline/Pubmed generally contain a compilation of only high quality evidence (Disagree) | Â | Â | Â | Â | Â | Â | Â | Â |
   Correct | 8 (7.0) | 42 (39.6) | +7 | 35 (70.0) 2 | +30 | 2 | 21.8 | 0.0001** |
   Incorrect | 106 (93.0) | 64 (60.4) |  | 15 (30.0)2 |  |  |  |  |
Q3. An electronic database such as PubMed can only be accessed from hospital and university libraries (Disagree) | Â | Â | Â | Â | Â | Â | Â | Â |
   Correct | 26 (22.8) | 102 (96.2) | +73 | 47 (92.2) | -4 | 2 | 68.7 | 0.0001** |
   Incorrect | 88 (77.2) | 4 (3.8) |  | 4 (7.8) |  |  |  |  |
Q4. The Cochrane Database is a good place to look for reviews of high quality research based on clearly stated criteria, in many areas of clinical practice (Agree) | Â | Â | Â | Â | Â | Â | Â | Â |
   Correct | 46 (40.4) | 98 (92.5) | +53 | 45 (88.2) | -5 | 2 | 40.5 | 0.0001** |
   Incorrect | 68 (59.6) | 8 (7.5) |  | 6 (11.8) |  |  |  |  |
Q5. Confidence intervals are a measure of clinical significance, and provide a way of estimating where the 'true' result for any population lies (Agree) | Â | Â | Â | Â | Â | Â | Â | Â |
   Correct | 20 (17.5) | 92 (86.8) | +69 | 41 (80.4) | -7 | 2 | 63.5 | 0.0001** |
   Incorrect | 94 (82.5) | 14 (13.2) |  | 10 (19.6) |  |  |  |  |
Q6. The p-value is a measure of reliability (Disagree) | Â | Â | Â | Â | Â | Â | Â | Â |
   Correct | 10 (8.8) 1 | 61 (57.5) | +49 | 20 (40.0) | -18 | 2 | 33.3 | 0.0001** |
   Incorrect | 103 (91.2)1 | 45 (42.5) |  | 30 (60.0) |  |  |  |  |
Q7. Single case designs are regarded as a similar level of evidence to randomised controlled trials (Disagree) | Â | Â | Â | Â | Â | Â | Â | Â |
   Correct | 72 (63.7) 1 | 101 (95.3) | +31 | 50 (98.0) | +3 | 2 | 22.5 | 0.0001** |
   Incorrect | 41 (36.3)1 | 5 (4.7) |  | 1 (2.0) |  |  |  |  |