1 Overview

1.1 Case description

1.1.1

Short description of the task along with some example images from eCREST. Report period is from the start of eCREST to December 2022.

1.1.2 Task outline

The case that all students completed referred to a female (Mrs Taru Gandhi) presenting to a GP with chest pain. Mrs Gandhi is fifty-eight years old, working as a dinner lady and of Asian background. The students had access to her medical records and tests and were asked to go through the case with thirty-two questions available in total.

An expert panel of GPs identified that:

  • Of the thirty-two questions there were twenty essential questions (3.1) and six relevant questions (3.5).
  • Of the eleven available, the panel identified six essential physical examinations (3.2).
  • Of the nine available, the panel identified three essential bedside tests (3.3).
  • According to the advisory panel there were six relevant diagnoses (4.3 - 4.5), out of the thirteen available.
  • The advisory panel identified that ‘Review in GP’ was the correct follow up management plan (5.1) out of the six possible follow up management plans.
  • The panel also identified three essential follow up investigations out of the six available (5.2).

Lastly, students provided feedback on the usability of eCREST. (6.1)

1.1.3 Example images 1

The Waiting Room and Taru Ghandi case informationThe Waiting Room and Taru Ghandi case information

The Waiting Room and Taru Ghandi case information

Initial perceptions and questionsInitial perceptions and questions

Initial perceptions and questions

Review daignoses and further testsReview daignoses and further tests

Review daignoses and further tests

final diagnoses and management planfinal diagnoses and management plan

final diagnoses and management plan

FeedbackFeedback

Feedback

Additional information and usabilityAdditional information and usability

Additional information and usability

1.2 Summary findings

1.2.1

Short descriptions of key findings by section with full details referenced in the report.

1.2.2 Student Characteristics

Age - in yrs: Overall, students were between 19 and 48 years old, with most around 24. (2.3.2) On average, physician associates were older than UK medical students. (2.3.3)

Gender - male or female: Overall, most students were female. (2.4.2) Physician associates had a higher proportion of female students than UK medical students. (2.4.3)

1.2.3 Patient investigations

Essential questions: Overall, students asked around 70% of essential questions on average. (3.1.2) Both physician associates and UK medical students were similar in the proportion of essential questions asked. (3.1.4) and in how likely they were to ask each of the 20 questions (3.1.5)

Essential Physical exams: Overall, students selected around 55% of essential physical exams on average. (3.2.2) Both physician associates and UK medical students were similar in the proportion of essential physical exams selected. (3.2.4) and in how likely they were to select each of the six exams, though some evidence suggestive that physician associates may have been more likely to select ENT but less likely to select musculoskeletal chest examination. (3.2.5)

Essential Bedside tests: Overall, students selected around 77% of essential bedside tests on average. (3.3.2) Physician associates appeared to have selected a greater proportion of essential bedside tests than UK medical students. (3.3.4)

Inspection of each of the 3 tests showed that oxygen saturation was selected more often than temperature and blood pressure. (3.1.5) In addition, physician associates appeared to have been more likely to select “blood pressure” than UK medical students. (3.3.5)

Relevant questions: Overall, students asked around 50% of relevant questions on average. (3.5.2) Physician associates and UK medical students were similar in the proportion of relevant questions asked. (3.5.4)

Combined measure: Combining the proportion of essential questions asked, physical exams selected, and bedside tests selected in a single measure found physician associates and UK medical students were similar. (3.4.4) This finding remained consistent when relevant questions were added to the combined measure. (3.6.4)

1.2.4 Diagnoses

Lung cancer: Overall, just over 76% of students included lung cancer somewhere in their final diagnosis. (4.1.2)

Both physician associates and UK medical students were similar in the proportion who included lung cancer in their final diagnosis, but there was some evidence that physician associates may have been more likely to have included lung cancer. (4.1.4)

LRTI: Overall, around 83% of students included lower respiratory tract infection somewhere in their final diagnosis. (4.2.2)

Both physician associates and UK medical students were similar in the proportion who included lower respiratory tract infection in their final diagnosis. (4.2.4)

Any relevant diagnosis: Overall, in their initial diagnosis, students included around 45% of the relevant diagnoses on average. (4.3.2) Physician associates included a greater proportion of relevant diagnoses in their initial diagnosis. (4.3.4)

Both physician associates and UK medical students showed significant improvement and had more relevant diagnosis in their final diagnosis than they did initially. (4.5.4) UK medical students had the greatest improvement.

In the final diagnosis, the proportion of relevant diagnoses had increased to about 64%. (4.4.2) At this stage, both UK medical students and physician associates were similar in the proportion of relevant diagnoses included. (4.4.4)

1.2.5 Post-consultation follow up

Management plan: Overall, around 56% of students selected the correct management plan “Review in GP”. (5.1.2)

Both UK medical students and physician associates were similar in how likely they were to select the correct management plan. (5.1.4)

Essential investigations: Overall, the proportion of essential investigations selected was around 36%. (5.2.2) Around 81% of students selected at least one essential investigation in their follow up plan. (5.2.5)

UK medical students included a greater proportion of essential investigations in their follow up plan compared to physician associates. (5.2.4)

Inspection of each of the investigations revealed that UK medical students were more likely to select “Relevant cultures or serological tests” than physician associates. (5.2.5 and 5.2.7)

1.2.6 Feedback

Overall feedback was positive and did not differ between UK medical students and physician associates. (6.1.2) to (6.1.5)

2 Case 3: Student characteristics

2.1 Student institutions

2.1.1

Tables showing the number of students from each institution and then showing the number of students grouped by student type: Medical students (international or UK), Pharmacists, and Physician Associates)

2.1.2 Student institutions

2.1.3 Institutions grouped by student type

df$Type1 n percent
MedStudents_UK 489 74.9%
PhysicianAssociates 164 25.1%
Total 653 100.0%

2.2 Case 3 Completers

2.2.1

Tables showing the number of students who completed each case and the number of students of each student type who completed case 3 (Taru Ghandi).

2.2.2 All Case 3 completers

PatientID n percent
1 249 29.0%
2 163 19.0%
3 213 24.8%
4 233 27.2%
Total 858 100.0%

2.2.3 All Case 3 completers grouped by student type and institutions

label levels MedStudents_UK PhysicianAssociates Total
Total N (%) 159 (74.6) 54 (25.4) 213
Type Barts 3 (1.9) 3 (1.4)
Barts 2020 10 (6.3) 10 (4.7)
Barts2 38 (23.9) 38 (17.8)
UCL 2020 8 (5.0) 8 (3.8)
UCL3 52 (32.7) 52 (24.4)
UEA 2020 19 (11.9) 19 (8.9)
UEA1 22 (13.8) 22 (10.3)
UEA2 7 (4.4) 7 (3.3)
QMUL 7 (13.0) 7 (3.3)
St Georges 40 (74.1) 40 (18.8)
Swansea 4 (7.4) 4 (1.9)
UEA PA 3 (5.6) 3 (1.4)

2.3 Student Age

2.3.1

Descriptive statistics for student age overall and then according to student type and comparison test.

2.3.2 Descriptive stats

  • The descriptive statistics for student age overall.
    vars n mean sd median trimmed mad min max range skew kurtosis se
    X1 1 209 24.57 4.97 24 23.76 1.48 19 48 29 2.72 8.83 0.34
  • Then showing the descriptive statistics for student age according to student type.
    group1 n mean sd median trimmed mad min max range skew kurtosis se
    MedStudents_UK 156 23.69 3.57 23 23.25 1.48 19 42 23 2.55 10.05 0.29
    PhysicianAssociates 53 27.17 7.18 25 25.77 2.97 19 48 29 1.80 2.35 0.99

2.3.3 Follow up tests (MDEs)

  • The table showing all the pairwise comparisons with mean difference estimates (95%CIs). Raw and holm corrected p values are provided.
Group1: Mean Group2: Mean Mean difference estimate CI lower CI higher raw p value Holm p value
MedStudents_UK 23.69 PhysicianAssociates 27.17 -3.48 -5.53 -1.42 0.001 0.001
  • The figure below plots student age by student type. Points represent individual data points, the point estimate represents the mean estimate with error bars, bar chart overlaid.

2.4 Student Gender

2.4.1

Tables for students’ self-reported gender showing the number and (%) overall and then according to student type, the omnibus tests for association between gender and student type, and all pairwise follow-up tests of associations.

2.4.2 Descriptive tables

Overall and according to student type
label levels Female Male Total
Total N (%) 127 (59.6) 86 (40.4) 213
Type1 MedStudents_UK 84 (52.8) 75 (47.2) 159 (100)
PhysicianAssociates 43 (79.6) 11 (20.4) 54 (100)

2.4.3 Omnibus tests (Chi2 & Fisher’s Exact Test)

  • We conducted chi-squared test with Yates continuity correction and Fishers exact test to assess whether student self-reported gender was associated with student type. Both tests indicate a significant association between self-reported gender and student type.

(Odds Ratios (95%CIs) for the Fisher’s Exact Test and rounded Raw p values are provided.)

Odds ratio CI lower CI upper raw p value
UK medical students vs. Physician associates 0.29 0.12 0.62 0.001
  • The figure below plots student gender by student type. Columns represent %s with total ns provided.

  • Alternative figure with truncated axis to emphasize differences.

3 Case 3: Patient investigations

3.1 Essential Questions only [EQs]

3.1.1

Twenty of the thirty-two available questions were identified by experts during development as essential and are highlighted in bold font below.

Category Question Question
Essential Does anything make your pain worse? Have you ever coughed up any blood?
Essential Does anything make the pain better? Have you had a fever?
Essential Does the pain come and go? Are you more breathless than normal?
Essential How long has the pain been there? Any recent immobility?
Essential Where do you feel the pain? Do you have pets?
Essential Does the pain travel anywhere else? Do you smoke?
Essential I can see you also have a cough. Can you tell me more about it? How many cigarettes do you smoke per day?
Essential Have you noticed any other symptoms? How long have you been a smoker?
Essential When did the cough start? Do you have any ideas about what is causing this?
Essential Do you produce any phlegm? How is this affecting you?
Relevant Does your position alter the pain? Have you started any new medications recently?
Relevant Have you had any injury to your chest? Do you drink alcohol?
Relevant What do you do for a living? Have you lost or gained any weight?
Other Can you identify anything that makes the cough worse? What do you do with your spare time?
Other Does anything make the cough better? Do you take medication?
Other Have you noticed a change in your mood at all? Has your appetite changed?

Tables 3.1.2 to 3.1.4: Descriptive statistics for the number and % of EQs asked overall and then according to student type, and tests for differences in the % of EQs asked by student type.

Tables 3.1.5 to 3.1.8: Tables for showing the number and (%) of students who asked each EQ overall and then according to student type, the omnibus tests for association between % of students who asked each EQ and student type, all pairwise follow-up tests of associations, and all figures.

3.1.2 Descriptive stats

  • Descriptive statistics for the total number of essential questions asked overall
vars n mean sd median trimmed mad min max range skew kurtosis se
X1 1 213 14.18 4.43 15 14.57 4.45 0 20 20 -0.72 0.1 0.3
  • Descriptive statistics for the proportion of essential questions asked overall
vars n mean sd median trimmed mad min max range skew kurtosis se
X1 1 213 70.92 22.16 75 72.87 22.24 0 100 100 -0.72 0.1 1.52
  • Descriptive statistics for the proportion of essential questions asked by student group
group1 n mean sd median trimmed mad min max range skew kurtosis se
MedStudents_UK 159 70.13 23.17 75.0 72.02 22.24 0 100 100 -0.67 -0.20 1.84
PhysicianAssociates 54 73.24 18.89 72.5 74.55 18.53 10 100 90 -0.78 1.14 2.57

3.1.3 Follow up tests (MDEs)

  • The table shows all pairwise comparisons of the proportion of essential questions asked by student type.
Group1: Mean Group2: Mean Mean difference estimate CI lower CI higher raw p value Holm p value
MedStudents_UK 70.13 PhysicianAssociates 73.24 -3.11 -9.38 3.15 0.326 0.326

3.1.4 Descriptive tables

  • Table showing the proportion of students who asked each essential question, with comparisons described in the final column

3.1.5 Omnibus tests (Chi2 & Fisher’s Exact Test)

  • First we ran chi-square and fisher omnibus tests for all questions. Rows for which at least one test result was statistically significant are highlighted in bold font.
X-squared df p value Fisher p value
Does anything make your pain worse? 0.00 1 1.000 1.000
Does anything make the pain better? 1.33 1 0.250 0.221
Does the pain come and go? 1.56 1 0.211 0.184
How long has the pain been there? 0.00 1 1.000 0.824
Where do you feel the pain? 0.08 1 0.780 0.816
Does the pain travel anywhere else? 2.56 1 0.110 0.104
I can see you also have a cough. Can you tell me more about it? 0.62 1 0.430 0.455
Have you noticed any other symptoms? 0.65 1 0.419 0.389
When did the cough start? 0.27 1 0.603 0.590
Do you produce any phlegm? 1.47 1 0.226 0.170
Have you ever coughed up any blood? 1.37 1 0.243 0.255
Have you had a fever? 1.02 1 0.313 0.289
Are you more breathless than normal? 0.61 1 0.434 0.408
Any recent immobility? 0.81 1 0.368 0.335
Do you have pets? 0.03 1 0.872 0.861
Do you smoke? 0.06 1 0.799 0.746
How many cigarettes do you smoke per day? 0.00 1 0.973 0.876
How long have you been a smoker? 0.06 1 0.813 0.744
Do you have any ideas about what is causing this? 0.18 1 0.670 0.599
How is this affecting you? 2.59 1 0.107 0.085

3.1.6 All figures

  • Figures showing the proportion of students who asked each essential question

3.2 Physical exams only [PEs]

3.2.1

Six of the eleven available physical examinations were identified by experts during development as essential and are higlighted in bold font below.

Category Physical Examination Physical Examination
Essential Cardiovascular Lower limb and foot
Essential Respiratory Lymphatic System
Essential Musculoskeletal chest examination ENT
Other Abdomen Neurological (central and cranial nerves)
Other Eyes Neurological (peripheral)
Other Genital examination

Tables 3.2.2 to 3.1.4: Descriptive statistics for the number and % of PEs selected overall and then according to student type, and tests for differences in the % of PEs asked by student type.

Tables 3.2.5 to 3.2.8: Tables for showing the number and (%) of students who selected each PEs overall and then according to student type, the omnibus tests for association between % of students who selected each PEs and student type, all pairwise follow-up tests of associations, and all figures.

3.2.2 Descriptive stats

  • Descriptive statistics for the total number of physical exams selected overall
    vars n mean sd median trimmed mad min max range skew kurtosis se
    X1 1 213 3.31 1.15 3 3.35 1.48 0 6 6 -0.44 0.9 0.08
  • Descriptive statistics for the proportion of physical exams selected overall
    vars n mean sd median trimmed mad min max range skew kurtosis se
    X1 1 213 55.16 19.21 50 55.75 24.71 0 100 100 -0.44 0.9 1.32
  • Descriptive statistics for the proportion of physical exams selected by student group
    group1 n mean sd median trimmed mad min max range skew kurtosis se
    MedStudents_UK 159 54.72 18.58 50 55.43 24.71 0 100 100 -0.60 1.35 1.47
    PhysicianAssociates 54 56.48 21.09 50 56.82 24.71 0 100 100 -0.14 -0.21 2.87
group1 n mean sd median trimmed mad min max range skew kurtosis se
MedStudents_UK 159 3.28 1.11 3 3.33 1.48 0 6 6 -0.60 1.35 0.09
PhysicianAssociates 54 3.39 1.27 3 3.41 1.48 0 6 6 -0.14 -0.21 0.17

3.2.3 Follow up tests (MDEs)

  • The table shows all pairwise comparisons of the proportion of physical exams selected by student type.
group1 estimate1 group2 estimate2 estimate conf.low conf.high p p.adj
MedStudents_UK 54.72 PhysicianAssociates 56.48 -1.76 -8.18 4.65 0.586 0.586

3.2.4 Descriptive tables

  • Table showing the proportion of students who selected each essential physical exam

3.2.5 Omnibus tests (Chi2 & Fisher’s Exact Test)

  • First we ran chi-square and fisher omnibus tests for all physical exams.
X-squared df p value Fisher p value
Cardiovascular system 0.14 1 0.711 0.735
ENT 3.43 1 0.064 0.054
Lower Limb and foot 0.28 1 0.598 0.492
Lymphatic System 0.39 1 0.530 0.520
Musculoskeletal chest examination 3.66 1 0.056 0.053
Respiratory 0.00 1 1.000 1.000

3.2.6 All figures

  • Figures showing the proportion of students who asked each essential question

3.3 Bedside tests only [BTs]

## # A tibble: 1 × 12
##     estimate .y.     group1   group2    n1    n2 stati…¹     p conf.low conf.h…²
## *      <dbl> <chr>   <chr>    <chr>  <int> <int>   <dbl> <dbl>    <dbl>    <dbl>
## 1 -0.0000309 BT_prop MedStud… Physi…   159    54    3410 0.013 -5.46e-5 -5.94e-6
## # … with 2 more variables: method <chr>, alternative <chr>, and abbreviated
## #   variable names ¹​statistic, ²​conf.high
## # A tibble: 1 × 6
##   .y.         n statistic    df     p method        
## * <chr>   <int>     <dbl> <int> <dbl> <chr>         
## 1 BT_prop   213      6.17     1 0.013 Kruskal-Wallis
##    
##     MedStudents_UK PhysicianAssociates
##   0       6.289308            7.407407
##   1      10.062893            1.851852
##   2      37.735849           24.074074
##   3      45.911950           66.666667
## Warning in stats::chisq.test(x, y, ...): Chi-squared approximation may be
## incorrect
## 
##  Pearson's Chi-squared test
## 
## data:  table(df_C3$BT_ttl, df_C3$Type1)
## X-squared = 9.0702, df = 3, p-value = 0.02837
## 
##  Fisher's Exact Test for Count Data
## 
## data:  table(df_C3$BT_ttl, df_C3$Type1)
## p-value = 0.02413
## alternative hypothesis: two.sided

3.3.1

Three of the nine available bedside tests were identified by experts during development as essential and are listed below.

Category Test Test Test
Essential Blood pressure Temperature Oxygen saturation
Other Blood Glucose Peak Flow Weight
Other Height Urinalysis Patient Health Questionnaire (PHQ-9)

Tables 3.3.2 to 3.3.4: Descriptive statistics for the number and % of BTs selected overall and then according to student type, and tests for differences in the % of BTs asked by student type.

Tables 3.3.5 to 3.3.8: Tables for showing the number and (%) of students who selected each BTs overall and then according to student type, the omnibus tests for association between % of students who selected each BTs and student type, all pairwise follow-up tests of associations, and all figures.

3.3.2 Descriptive stats

  • Descriptive statistics for total number of bedside tests selected overall
    vars n mean sd median trimmed mad min max range skew kurtosis se
    X1 1 213 2.3 0.88 3 2.46 0 0 3 3 -1.2 0.72 0.06
  • Descriptive statistics for proportion of bedside tests selected overall
    vars n mean sd median trimmed mad min max range skew kurtosis se
    X1 1 213 76.68 29.21 100 81.87 0 0 100 100 -1.2 0.72 2
  • Descriptive statistics for proportion of bedside tests selected by student group
    group1 n mean sd median trimmed mad min max range skew kurtosis se
    MedStudents_UK 159 74.42 29.10 66.67 78.81 49.42 0 100 100 -1.03 0.37 2.31
    PhysicianAssociates 54 83.33 28.78 100.00 90.15 0.00 0 100 100 -1.81 2.48 3.92

3.3.3 Follow up tests (MDE)

  • The table shows all pairwise comparisons of the proportion of essential bedside tests selected by student type.
Group1: Mean Group2: Mean Mean difference estimate CI lower CI higher raw p value Holm p value
MedStudents_UK 74.42 PhysicianAssociates 83.33 -8.91 -17.94 0.12 0.053 0.053

3.3.4 Descriptive tables

  • Table showing the proportion of students who selected each bedside test

3.3.5 Omnibus tests (Chi2 & Fisher’s Exact Test)

  • First we ran chi-square and fisher omnibus tests for all bedside tests.
X-squared df p value Fisher p value
Blood pressure 3.60 1 0.058 0.048
Temperature 2.41 1 0.121 0.096
Oxygen saturation 0.00 1 1.000 1.000

3.3.6 Follow up tests (Fisher’s Exact Test)

We conducted follow up tests using Fishers exact test for all pairwise comparisons by student type, which showed:

  • Physician associates were more likely to select “Temperature test” than International medical students and pharmacists. UK medical students were more likely to select this test than pharmacists.
Odds ratio CI lower CI upper raw p value
UK medical students vs. Physician associates 2.06 0.87 5.47 0.096

  • All student types were more likely to select “Oxygen saturation test” than pharmacists
Odds ratio CI lower CI upper raw p value
UK medical students vs. Physician associates 0.96 0.34 3.14 1

  • Some evidence that physician associates may have been more likely to select this test.
    Odds ratio CI lower CI upper raw p value
    UK medical students vs. Physician associates 2.06 0.99 4.54 0.048

3.4 EQs, PEs, and BTs combined

3.4.1

Descriptive statistics for the number and % of combined EQs, PEs, and BTs asked/selected overall and then according to student type, and tests for differences in the % of combined EQs, PEs, and BTs asked/selected student type.

3.4.2 Descriptive stats

  • Descriptive statistics for the proportion of Essential questions, physical exams, and bedside tests selected overall
    vars n mean sd median trimmed mad min max range skew kurtosis se
    X1 1 213 68.25 17.59 72.41 69.87 15.34 0 97 97 -0.97 1.27 1.21
  • Descriptive statistics for the proportion of Essential questions, physical exams, and bedside tests selected by student group
    group1 n mean sd median trimmed mad min max range skew kurtosis se
    MedStudents_UK 159 67.38 18.37 68.97 69.15 20.45 0.00 96.55 96.55 -1.0 1.15 1.46
    PhysicianAssociates 54 70.82 14.91 72.41 71.55 15.34 27.59 96.55 68.97 -0.5 0.27 2.03

3.4.3 Follow up tests (MDEs)

  • The table shows all pairwise comparisons of the proportion of essential questions, physical exams, and bedside tests by student type.
Group1: Mean Group2: Mean Mean difference estimate CI lower CI higher raw p value Holm p value
MedStudents_UK 67.38 PhysicianAssociates 70.82 -3.44 -8.39 1.52 0.172 0.172

3.5 Relevant Questions only [RQs]

3.5.1

Six of the thirty two available questions were identified by experts during development as relevant and are highlighted in bold font below.

Category Question Question
Essential Does anything make your pain worse? Have you ever coughed up any blood?
Essential Does anything make the pain better? Have you had a fever?
Essential Does the pain come and go? Are you more breathless than normal?
Essential How long has the pain been there? Any recent immobility?
Essential Where do you feel the pain? Do you have pets?
Essential Does the pain travel anywhere else? Do you smoke?
Essential I can see you also have a cough. Can you tell me more about it? How many cigarettes do you smoke per day?
Essential Have you noticed any other symptoms? How long have you been a smoker?
Essential When did the cough start? Do you have any ideas about what is causing this?
Essential Do you produce any phlegm? How is this affecting you?
Relevant Does your position alter the pain? Have you started any new medications recently?
Relevant Have you had any injury to your chest? Do you drink alcohol?
Relevant What do you do for a living? Have you lost or gained any weight?
Other Can you identify anything that makes the cough worse? What do you do with your spare time?
Other Does anything make the cough better? Do you take medication?
Other Have you noticed a change in your mood at all? Has your appetite changed?

Descriptive statistics for the number and % of RQs asked overall and then according to student type, and tests for differences in the % of RQs asked by student type.

3.5.2 Descriptive stats

  • Descriptive statistics for the total number of relevant questions asked overall
vars n mean sd median trimmed mad min max range skew kurtosis se
X1 1 213 3.03 1.79 3 3.03 1.48 0 6 6 0.02 -0.96 0.12
  • Descriptive statistics for the proportion of relevant questions asked overall
vars n mean sd median trimmed mad min max range skew kurtosis se
X1 1 213 50.55 29.82 50 50.49 24.71 0 100 100 0.02 -0.96 2.04
  • Descriptive statistics for the proportion of relevant questions asked by student group
group1 n mean sd median trimmed mad min max range skew kurtosis se
MedStudents_UK 159 49.48 30.26 50 49.35 24.71 0 100 100 0.07 -1.01 2.40
PhysicianAssociates 54 53.70 28.53 50 54.17 24.71 0 100 100 -0.07 -0.86 3.88

3.5.3 Follow up tests (MDEs)

  • The table shows all pairwise comparisons of the proportion of relevant questions asked by student type.
Group1: Mean Group2: Mean Mean difference estimate CI lower CI higher raw p value Holm p value
MedStudents_UK 49.48 PhysicianAssociates 53.7 -4.23 -13.29 4.83 0.357 0.357

#Essential questions, physical exams, bedside tests, and relevant questions combined

3.6 EQs, PEs, BTs, and RQs combined

3.6.1

Descriptive statistics for the number and % of combined EQs, PEs, BTs, and RQs asked/selected overall and then according to student type, and tests for differences in the % of combined EQs, PEs, BTs, and RQs asked/selected student type.

3.6.2 Descriptive stats

  • Descriptive statistics for the proportion of Essential questions, physical exams, bedside tests, and relevant questions selected overall
    vars n mean sd median trimmed mad min max range skew kurtosis se
    X1 1 213 65.22 18.6 65.71 66.43 16.94 0 97 97 -0.63 0.42 1.27
  • Descriptive statistics for the proportion of Essential questions, physical exams, bedside tests, and relevant questions selected by student group
    group1 n mean sd median trimmed mad min max range skew kurtosis se
    MedStudents_UK 159 64.31 19.32 65.71 65.58 21.18 0.00 97.14 97.14 -0.65 0.30 1.53
    PhysicianAssociates 54 67.88 16.19 68.57 68.51 12.71 25.71 97.14 71.43 -0.34 0.07 2.20

3.6.3 Follow up tests (MDEs)

  • The table shows all pairwise comparisons of the proportion of Essential questions, physical exams, bedside tests, and relevant questions selected by student type.
Group1: Mean Group2: Mean Mean difference estimate CI lower CI higher raw p value Holm p value
MedStudents_UK 64.31 PhysicianAssociates 67.88 -3.57 -8.89 1.75 0.186 0.186

4 Case 3: Diagnoses

Six of the thirteen available diagnoses were identified by experts during development as relevant and are highlighted in bold font below. **Lower respiratory tract infection was considered the top diagnosis for students to consider.

Category Diagnosis Diagnosis
Relevant Lower respiratory tract infection** COPD
Relevant Pulmonary Embolism Lung cancer
Relevant Pulmonary Tuberculosis Costochondritis
Other Acute Coronary Syndrome Pneumothorax
Other Interstitial lung disease Psychogenic cough
Other Musculoskeletal injury Stable Angina
Other Pericarditis

4.1 Lung cancer

4.1.1

Tables for inclusion of lung cancer as a final diagnosis showing the number and (%) overall and then according to student type, the omnibus tests for association between lung cancer diagnosis and student type, and all pairwise follow-up tests of associations.

4.1.2 Descriptive tables

  • Proportion of students including a lung cancer diagnosis overall and according to student type.
label levels Lung cancer included Lung cancer not included Total
Total N (%) 163 (76.5) 50 (23.5) 213
Type1 MedStudents_UK 119 (74.8) 40 (25.2) 159 (100)
PhysicianAssociates 44 (81.5) 10 (18.5) 54 (100)

4.1.3 Omnibus tests (Chi2 & Fisher’s Exact Test)

  • We conducted chi-squared test with yates continuity correction and Fishers exact test to assess whether proportion of students who considered lung cancer as a possible diagnosis was associated with student type. Both tests indicate no significant association between proportion of students who considered lung cancer as a possible diagnosis and student type.

(Odds Ratios (95%CIs) for the Fisher’s Exact Test and rounded Raw p values are provided.)

Odds ratio CI lower CI upper raw p value
UK medical students vs. Physician associates 0.68 0.28 1.53 0.358

4.2 Lower respiratory tract infection

4.2.1

Tables for inclusion of LRTI as a final diagnosis showing the number and (%) overall and then according to student type, the omnibus tests for association between LRTI diagnosis and student type, and all pairwise follow-up tests of associations.

4.2.2 Descriptive tables

  • Proportion of students including a LRTI diagnosis overall and according to student type.
label levels LRTI included LRTI not included Total
Total N (%) 176 (82.6) 37 (17.4) 213
Type1 MedStudents_UK 133 (83.6) 26 (16.4) 159 (100)
PhysicianAssociates 43 (79.6) 11 (20.4) 54 (100)

4.2.3 Omnibus tests (Chi2 & Fisher’s Exact Test)

  • We conducted chi-squared test with yates continuity correction and Fishers exact test to assess whether proportion of students who considered lung cancer as a possible diagnosis was associated with student type. Both tests indicate a significant association between proportion of students who considered LRTI as a possible diagnosis and student type.
X-squared df p value Fisher p value
Test of LRTI diagnosis by Student type 0.22 1 0.6416 0.5348

(Odds Ratios (95%CIs) for the Fisher’s Exact Test and rounded Raw p values are provided.)

Odds ratio CI lower CI upper raw p value
UK medical students vs. Physician associates 1.31 0.54 3.02 0.535

4.3 Times changed diagnosis

myFile6 <- paste0(myDir, "df_Diagchange.csv") #Load the dataset where we just took each diagnoses and put them next to each other (removing the spaces between so we can compare each diagnosis to its preceeding one)
df_Diagchange <- readr::read_csv(myFile6, na = "0")
## Rows: 213 Columns: 8
## ── Column specification ────────────────────────────────────────────────────────
## Delimiter: ","
## chr (8): InitialDiag, Diag1, Diag2, Diag3, Diag4, Diag5, Diag6, Diag7
## 
## ℹ Use `spec()` to retrieve the full column specification for this data.
## ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.
#This code is to create new variables where no diagnosis is 0 and a new diagnosis is 1. This does not include initial diagnosis as these are just changes
df_Diagchange <- df_Diagchange %>% mutate(Diag1_num = ifelse(Diag1 == "", 0,1))
df_Diagchange <- df_Diagchange %>% mutate(Diag2_num = ifelse(Diag2 == "", 0,1))
df_Diagchange <- df_Diagchange %>% mutate(Diag3_num = ifelse(Diag3 == "", 0,1))
df_Diagchange <- df_Diagchange %>% mutate(Diag4_num = ifelse(Diag4 == "", 0,1))
df_Diagchange <- df_Diagchange %>% mutate(Diag5_num = ifelse(Diag5 == "", 0,1))
df_Diagchange <- df_Diagchange %>% mutate(Diag6_num = ifelse(Diag6 == "", 0,1))
df_Diagchange <- df_Diagchange %>% mutate(Diag7_num = ifelse(Diag7 == "", 0,1))

#This code sums them all.
df_Diagchange$DiagChangesRaw <- df_Diagchange$Diag1_num + df_Diagchange$Diag2_num + df_Diagchange$Diag3_num + df_Diagchange$Diag4_num + df_Diagchange$Diag5_num + df_Diagchange$Diag6_num + df_Diagchange$Diag7_num

#This is to make blank spaces NAs so they are not matched as blanks
df_Diagchange$Diag1[df_Diagchange$Diag1 == ""] <- NA
df_Diagchange$Diag2[df_Diagchange$Diag2 == ""] <- NA
df_Diagchange$Diag3[df_Diagchange$Diag3 == ""] <- NA
df_Diagchange$Diag4[df_Diagchange$Diag4 == ""] <- NA
df_Diagchange$Diag5[df_Diagchange$Diag5 == ""] <- NA
df_Diagchange$Diag6[df_Diagchange$Diag6 == ""] <- NA
df_Diagchange$Diag7[df_Diagchange$Diag7 == ""] <- NA

#This is to check if the diagnosis matches whatever was the preceding diagnoses
df_Diagchange <- df_Diagchange %>% mutate(InitStep1Diag_Match =  ifelse(Diag1 == InitialDiag, 1,0))
df_Diagchange <- df_Diagchange %>% mutate(Step1Step2Diag_Match = ifelse(Diag2 == Diag1, 1,0))
df_Diagchange <- df_Diagchange %>% mutate(Step2Step3Diag_Match = ifelse(Diag3 == Diag2, 1,0))
df_Diagchange <- df_Diagchange %>% mutate(Step3Step4Diag_Match = ifelse(Diag4 == Diag3, 1,0))
df_Diagchange <- df_Diagchange %>% mutate(Step4Step5Diag_Match = ifelse(Diag5 == Diag4, 1,0))
df_Diagchange <- df_Diagchange %>% mutate(Step5Step6Diag_Match = ifelse(Diag6 == Diag5, 1,0))
df_Diagchange <- df_Diagchange %>% mutate(Step6Step7Diag_Match = ifelse(Diag7 == Diag6, 1,0))

df_Diagchange <- df_Diagchange %>% mutate(DiagSameRaw = select(., InitStep1Diag_Match:Step6Step7Diag_Match) %>% rowSums(na.rm = TRUE))
 
df_Diagchange$DiagChangesFinal <- df_Diagchange$DiagChangesRaw - df_Diagchange$DiagSameRaw

tabyl(df_Diagchange$DiagChangesFinal) %>% adorn_totals()
##  df_Diagchange$DiagChangesFinal   n     percent
##                               0  10 0.046948357
##                               1  33 0.154929577
##                               2  45 0.211267606
##                               3  60 0.281690141
##                               4  41 0.192488263
##                               5  20 0.093896714
##                               6   3 0.014084507
##                               7   1 0.004694836
##                           Total 213 1.000000000
df_C3 <- cbind(df_C3, df_Diagchange[25])

tabyl(df_C3$DiagChangesFinal) %>% adorn_totals() %>% adorn_pct_formatting()
##  df_C3$DiagChangesFinal   n percent
##                       0  10    4.7%
##                       1  33   15.5%
##                       2  45   21.1%
##                       3  60   28.2%
##                       4  41   19.2%
##                       5  20    9.4%
##                       6   3    1.4%
##                       7   1    0.5%
##                   Total 213  100.0%

4.3.1

Tables X to X: Descriptive statistics for the number and % of times changed diagnoses overall and then according to student type, and tests for differences in the number of times changed diagnoses by student type.

4.3.2 Descriptive stats

  • Descriptive statistics for the total number of times changed diagnoses (after initial diagnosis)
vars n mean sd median trimmed mad min max range skew kurtosis se
X1 1 213 2.78 1.41 3 2.75 1.48 0 7 7 0.08 -0.42 0.1
  • Descriptive statistics for the total number of times changed diagnoses (after initial diagnosis)
group1 n mean sd median trimmed mad min max range skew kurtosis se
MedStudents_UK 159 2.78 1.40 3 2.77 1.48 0 7 7 0.06 -0.34 0.11
PhysicianAssociates 54 2.78 1.45 3 2.73 1.48 0 6 6 0.13 -0.72 0.20

4.3.3 Follow up tests (MDEs)

  • The table shows all pairwise comparisons of the total number of times changed diagnoses (after initial diagnosis) by student type.
Group1: Mean Group2: Mean Mean difference estimate CI lower CI higher raw p value Holm p value
MedStudents_UK 2.78 PhysicianAssociates 2.78 0 -0.45 0.45 0.993 0.993

4.4 Changed diagnoses at least once

4.4.1

Tables for whether they changed diagnoses at least once or not showing the number and (%) overall and then according to student type, the omnibus tests for association between student type, and all pairwise follow-up tests of associations.

4.4.2 Descriptive tables

  • Proportion of students including a LRTI diagnosis overall and according to student type.
label levels Did not change once Changed at least once Total p
Total N (%) 10 (4.7) 203 (95.3) 213
Type1 MedStudents_UK 8 (5.0) 151 (95.0) 159 (100) 0.979
PhysicianAssociates 2 (3.7) 52 (96.3) 54 (100)

4.4.3 Omnibus tests (Chi2 & Fisher’s Exact Test)

  • We conducted chi-squared test with yates continuity correction and Fishers exact test to assess whether proportion of students who considered lung cancer as a possible diagnosis was associated with student type. Both tests indicate a significant association between proportion of students who considered LRTI as a possible diagnosis and student type.
X-squared df p value Fisher p value
Test of whether changed diagnosis at least once or not by Student type 0 1 0.9791 1

(Odds Ratios (95%CIs) for the Fisher’s Exact Test and rounded Raw p values are provided.)

Odds ratio CI lower CI upper raw p value
UK medical students vs. Physician associates 1.38 0.26 13.71 1 

## [1] "Did not change once"   "Changed at least once"
##  df_C3$DiagChangesFinal   n percent
##                       0  10    4.7%
##                       1  33   15.5%
##                       2  45   21.1%
##                       3  60   28.2%
##                       4  41   19.2%
##                       5  20    9.4%
##                       6   3    1.4%
##                       7   1    0.5%
##                   Total 213  100.0%
##  df_C3$DiagChangedOnce   n percent valid_percent
##    Did not change once   0    0.0%             -
##  Changed at least once   0    0.0%             -
##                   <NA> 213  100.0%             -
##                  Total 213  100.0%          0.0%

4.5 Relevant initial diagnosis

4.5.1

Tables 4.3.2 to 4.3.4: Descriptive statistics for the number and % of relevant initial diagnoses selected overall and then according to student type, and tests for differences in the % of relevant initial diagnoses selected by student type.

4.5.2 Descriptive stats

  • Descriptive statistics for the total number of relevant initial diagnoses
vars n mean sd median trimmed mad min max range skew kurtosis se
X1 1 213 2.73 0.89 3 2.74 1.48 1 5 4 0.03 -0.41 0.06
  • Descriptive statistics for the proportion of relevant initial diagnoses overall
vars n mean sd median trimmed mad min max range skew kurtosis se
X1 1 213 45.46 14.76 50 45.61 24.71 17 83 67 0.03 -0.41 1.01
  • Descriptive statistics for the proportion of relevant initial diagnoses by student group
group1 n mean sd median trimmed mad min max range skew kurtosis se
MedStudents_UK 159 43.40 14.41 50 43.80 24.71 16.67 66.67 50.00 -0.03 -0.71 1.14
PhysicianAssociates 54 51.54 14.21 50 51.14 12.35 16.67 83.33 66.67 0.19 -0.10 1.93

4.5.3 Follow up tests (MDEs)

  • The table shows all pairwise comparisons of the proportion of relevant initial diagnoses by student type.
Group1: Mean Group2: Mean Mean difference estimate CI lower CI higher raw p value Holm p value
MedStudents_UK 43.4 PhysicianAssociates 51.54 -8.15 -12.61 -3.69 0 0

4.6 Relevant final diagnosis

4.6.1

Tables 4.4.2 to 4.4.4: Descriptive statistics for the number and % of relevant final diagnoses selected overall and then according to student type, and tests for differences in the % of relevant final diagnoses selected by student type.

4.6.2 Descriptive stats

  • Descriptive statistics for the total number of relevant final diagnoses
vars n mean sd median trimmed mad min max range skew kurtosis se
X1 1 213 3.82 0.92 4 3.89 1.48 1 5 4 -0.59 0.09 0.06
  • Descriptive statistics for the proportion of relevant final diagnoses overall
vars n mean sd median trimmed mad min max range skew kurtosis se
X1 1 213 63.62 15.27 66.67 64.81 24.71 17 83 67 -0.59 0.09 1.05
  • Descriptive statistics for the proportion of relevant final diagnoses by student group
group1 n mean sd median trimmed mad min max range skew kurtosis se
MedStudents_UK 159 62.89 15.45 66.67 64.21 24.71 16.67 83.33 66.67 -0.63 0.21 1.23
PhysicianAssociates 54 65.74 14.63 66.67 66.67 24.71 33.33 83.33 50.00 -0.39 -0.72 1.99

4.6.3 Follow up tests (MDEs)

  • The table shows all pairwise comparisons of the proportion of relevant final diagnoses by student type.
Group1: Mean Group2: Mean Mean difference estimate CI lower CI higher raw p value Holm p value
MedStudents_UK 62.89 PhysicianAssociates 65.74 -2.85 -7.49 1.79 0.226 0.226

4.7 Change in relevant diagnosis

4.7.1

Tables 4.5.2 to 4.5.4: Descriptive statistics for the number and % for the change in relevant diagnoses selected by the students (from initial to final) overall and then according to student type, and tests for differences in the % change in relevant diagnosis by student type.

Tables 3.3.5 to 3.3.8: Tables for showing the number and (%) of students with changes in relevant diagnosis overall and then according to student type, the omnibus tests for association between % of students who had more relevant diagnoses and student type, all pairwise follow-up tests of associations, and all figures.

4.7.2 Descriptive stats

  • Descriptive statistics for the total change in relevant diagnoses (positive numbers indicate more relevant diagnoses were included in the final diagnosis than in the initial diagnosis)
vars n mean sd median trimmed mad min max range skew kurtosis se
X1 1 213 1.09 1 1 1.04 1.48 -1 4 5 0.27 -0.36 0.07
  • Descriptive statistics for the percentage change in relevant diagnoses
vars n mean sd median trimmed mad min max range skew kurtosis se
X1 1 213 18.15 16.72 16.67 17.35 24.71 -17 67 83 0.27 -0.36 1.15
  • Descriptive statistics for the percentage change in relevant diagnoses by student group
group1 n mean sd median trimmed mad min max range skew kurtosis se
MedStudents_UK 159 19.5 16.58 16.67 18.73 24.71 -16.67 66.67 83.33 0.23 -0.22 1.32
PhysicianAssociates 54 14.2 16.64 16.67 13.26 24.71 -16.67 50.00 66.67 0.40 -0.77 2.26

4.7.3 Follow up tests (MDEs)

  • The table below shows the percentage change in relevant diagnoses from initial to final within each student group.
Mean change SD CI lower CI higher df t raw p value
UK Medical students 19.5 16.58 16.90 22.09 158 14.83 0
Physician Associates 14.2 16.64 9.66 18.74 53 6.27 0

  • The table shows all pairwise comparisons of the percentage change in relevant diagnoses by student type.
Group1: Mean Group2: Mean Mean difference estimate CI lower CI higher raw p value Holm p value
MedStudents_UK 19.5 PhysicianAssociates 14.2 5.3 0.1 10.5 0.046 0.046
  • Alternative figure with the percentage of relevant initial and final diagnoses

*For interpretation, note that, on the x axis, “100%” corresponds to including 6 out of 6 relevant diagnoses. Consequently, “50%” corresponds to including 3 out of 6.

4.7.4 Descriptive tables

Table showing the change scores in relevant diagnoses from initial to final and the number of students with each.

  • For instance, one student had two fewer relevant diagnoses in their final diagnoses that they had included in their initial diagnoses and five students included four more relevant diagnoses in their final diagnoses.
label levels 4 3 2 1 0 -1 Total
Total N (%) 2 (0.9) 14 (6.6) 57 (26.8) 74 (34.7) 60 (28.2) 6 (2.8) 213
Type1 MedStudents_UK 2 (1.3) 11 (6.9) 45 (28.3) 59 (37.1) 38 (23.9) 4 (2.5) 159 (100)
PhysicianAssociates 3 (5.6) 12 (22.2) 15 (27.8) 22 (40.7) 2 (3.7) 54 (100)
  • Table showing the number of students whose change scores in relevant diagnoses from initial to final were better, worse, or the same.
label levels More Same Fewer Total
Total N (%) 147 (69.0) 60 (28.2) 6 (2.8) 213
Type1 MedStudents_UK 117 (73.6) 38 (23.9) 4 (2.5) 159 (100)
PhysicianAssociates 30 (55.6) 22 (40.7) 2 (3.7) 54 (100)
  • Showing the number of students whose change scores in relevant diagnoses from initial to final were better, worse, or the same.
label levels Had more relevant diagnoses Had the same or fewer relevant diagnoses Total
Total N (%) 147 (69.0) 66 (31.0) 213
Type1 MedStudents_UK 117 (73.6) 42 (26.4) 159 (100)
PhysicianAssociates 30 (55.6) 24 (44.4) 54 (100)

4.7.5 Omnibus tests (Chi2 & Fisher’s Exact Test)

  • We conducted chi-squared test with yates continuity correction and Fishers exact test to assess whether the number of students who had more relevant diagnoses was associated with student type. Both tests indicate a significant association between the number of students who had more relevant diagnoses and student type.
X-squared df p value Fisher p value
Test of number of students having more relevant diagnoses at final than initial by Student type 5.31 1 0.0212 0.0171

4.7.6 Follow up tests (Fisher’s Exact Test)

  • The table showing all the pairwise comparisons with Odds Ratios (95%CIs). Raw p values are provided.
Odds ratio CI lower CI upper raw p value
UK medical students vs. Physician associates 0.45 0.23 0.9 0.017

5 Case 3: Post-consultation follow up

5.1 Management plan

5.1.1

Review in GP was identified by experts during development as the correct management plan out of the 6 available and is highlighted in bold font below.

Category Management plan
Other Immediate referral to acute secondary care
Other Non-urgent referral to secondary care
Other Refer as a 2 week wait to secondary care
Correct Review in General Practice
Other She does not require follow-up
Other Urgent referral to secondary care
Other Refer for psychological therapy or counselling

Tables for students’ management plans showing the number and (%) overall and then according to student type, the omnibus tests for association between management plan and student type, and all pairwise follow-up tests of associations.

5.1.2 Descriptive tables

  • Showing the number of students who chose each option overall and by student group. (Review in GP was identified as the correct management plan by the expert GP panel)
label levels Immediate referral to acute secondary care Non-urgent referral to secondary care Refer as a 2 week wait to secondary care Review in General Practice She does not require follow-up Urgent referral to secondary care Total
Total N (%) 14 (6.6) 11 (5.2) 41 (19.2) 119 (55.9) 2 (0.9) 26 (12.2) 213
Type1 MedStudents_UK 13 (8.2) 9 (5.7) 27 (17.0) 86 (54.1) 1 (0.6) 23 (14.5) 159 (100)
PhysicianAssociates 1 (1.9) 2 (3.7) 14 (25.9) 33 (61.1) 1 (1.9) 3 (5.6) 54 (100)
  • Showing the number and % of students who either did or did not selecte review in GP overall and by student group.
label levels Did not select review in GP Selected review in GP Total
Total N (%) 94 (44.1) 119 (55.9) 213
Type1 MedStudents_UK 73 (45.9) 86 (54.1) 159 (100)
PhysicianAssociates 21 (38.9) 33 (61.1) 54 (100)

5.1.3 Omnibus tests (Chi2 & Fisher’s Exact Test)

  • We conducted chi-squared test with Yates continuity correction and Fisher’s exact test to assess whether student self-reported gender was associated with student type. Both tests indicate a significant association between correct management plan selection and student type.
X-squared df p value Fisher p value
Test of Correct management plan by Student type 0.55 1 0.4597 0.4288

(Odds Ratios (95%CIs) for the Fisher’s Exact Test and rounded Raw p values are provided.)

Odds ratio CI lower CI upper raw p value
UK medical students vs. Physician associates 1.33 0.68 2.65 0.429

5.2 Essential investigations [EIs]

5.2.1

Three of the six available investigations were identified by experts during development as essential and are listed below.

Category Follow up investigation Follow up investigation Follow up investigation
Essential Electrocardiogram Relevant cultures or serological tests Prescription of appropriate medication
Other CXR Bloods PEFR, Spirometry and Reversibility testing

Tables 5.2.2 to 5.2.4: Descriptive statistics for the number and % of EIs selected overall and then according to student type, and tests for differences in the % of EIs selected by student type.

Tables 5.2.5 to 5.2.8: Tables for showing the number and (%) of students who selected each EI overall and then according to student type, the omnibus tests for association between % of students who selected each EIs and student type, all pairwise follow-up tests of associations, and all figures.

5.2.2 Descriptive stats

  • Descriptive statistics for the total number of essential investigations selected
vars n mean sd median trimmed mad min max range skew kurtosis se
X1 1 213 1.1 0.7 1 1.11 0 0 3 3 0.1 -0.44 0.05
  • Descriptive statistics for the proportion of essential investigations selected overall
vars n mean sd median trimmed mad min max range skew kurtosis se
X1 1 213 36.62 23.45 33.33 36.84 0 0 100 100 0.1 -0.44 1.61
  • Descriptive statistics for the proportion of essential investigations selected by student group
group1 n mean sd median trimmed mad min max range skew kurtosis se
MedStudents_UK 159 38.36 24.07 33.33 39.02 49.42 0 100 100 -0.03 -0.69 1.91
PhysicianAssociates 54 31.48 20.90 33.33 30.30 0.00 0 100 100 0.49 1.08 2.84

5.2.3 Follow up tests (MDEs)

  • The table shows all pairwise comparisons of the proportion of essential investigations selected by student type.
Group1: Mean Group2: Mean Mean difference estimate CI lower CI higher raw p value Holm p value
MedStudents_UK 38.36 PhysicianAssociates 31.48 6.88 0.09 13.68 0.047 0.047

5.2.4 Descriptive tables

  • Showing the number of students who selected 0, 1, 2, or 3 of the essential investigations.
label levels 0 1 2 3 Total
Total N (%) 40 (18.8) 115 (54.0) 55 (25.8) 3 (1.4) 213
Type1 MedStudents_UK 29 (18.2) 79 (49.7) 49 (30.8) 2 (1.3) 159 (100)
PhysicianAssociates 11 (20.4) 36 (66.7) 6 (11.1) 1 (1.9) 54 (100)
  • Showing the number of students who selected at least one of the essential investigations.
label levels Did not select any essential investigations Selected at least one essential investigation Total
Total N (%) 40 (18.8) 173 (81.2) 213
Type1 MedStudents_UK 29 (18.2) 130 (81.8) 159 (100)
PhysicianAssociates 11 (20.4) 43 (79.6) 54 (100)
  • Table showing the proportion of students who selected each of the essential investigations.

5.2.5 Omnibus tests (Chi2 & Fisher’s Exact Test)

  • We conducted chi-squared test with yates continuity correction and Fishers exact test to assess whether selecting at least one essential investigation was associated with student type. Both tests indicate no significant association between selecting at least one essential investigation and student type.
X-squared df p value Fisher p value
Test of selecting at least one essential investigation by Student type 0.02 1 0.8848 0.6926
  • We then ran chi-square and fisher omnibus tests for all essential investigations.
X-squared df p value Fisher p value
Electrocardiogram 0.11 1 0.738 0.743
Relevant cultures or serological tests 4.74 1 0.029 0.025
Prescription of appropriate medication 0.00 1 0.998 0.868

5.2.6 Follow up tests (Fisher’s Exact Test)

  • The table showing all the pairwise comparisons for having selected at least one essential investigation with Odds Ratios (95%CIs). Raw p values are provided.
Odds ratio CI lower CI upper raw p value
UK medical students vs. Physician associates 0.87 0.38 2.11 0.693

We conducted follow up tests using Fishers exact test for all pairwise comparisons by student type:

Odds ratio CI lower CI upper raw p value
UK medical students vs. Physician associates 0.85 0.41 1.69 0.743

  • Physician associates were the less likely to select “Relevant cultures or serological tests” compared to UK medical students.
Odds ratio CI lower CI upper raw p value
UK medical students vs. Physician associates 0.45 0.21 0.92 0.025

Odds ratio CI lower CI upper raw p value
UK medical students vs. Physician associates 1.06 0.51 2.13 0.868

6 Case 3: Student feedback

Students were asked to indicate their level of agreement with the following statements:

  • “It was easy to navigate through eCREST”,
  • “The level of difficulty of the material was appropriate”,
  • “eCREST should be used to supplement traditional teaching”,
  • “eCREST helped me to learn clinical reasoning skills that I could apply to my clinical work”,
  • “Overall, using eCREST enhanced my learning”,
  • “I would use eCREST for my future professional development”,

1=Strongly disagree, 2=Disagree, 3=Neither agree or disagree, 4=Agree, 5=Strongly agree

For analyses these were re-coded as (1=Agree or strongly agree vs. 0=Stongly disagree, disagree, or neither)

6.1 eCREST Usability

6.1.1

Table 6.2 shows the number and proportion of students who agreed or strongly agreed with each statement.

Tables 6.3 to 6.6 show figures for each statement.

6.1.2

  • Table showing the number and percentage of students who agreed or strongly agreed with each statement (shown in bolded font).
label levels MedStudents_UK PhysicianAssociates Total p
Total N (%) 119 (74.8) 40 (25.2) 159
It was easy to navigate through eCREST Agree or strongly agree 116 (97.5) 38 (95.0) 154 (96.9) 0.224
Neither 3 (2.5) 1 (2.5) 4 (2.5)
Strongly disagree or disagree 1 (2.5) 1 (0.6)
The level of difficulty of the material was appropriate Agree or strongly agree 114 (95.8) 37 (92.5) 151 (95.0) 0.217
Neither 5 (4.2) 2 (5.0) 7 (4.4)
Strongly disagree or disagree 1 (2.5) 1 (0.6)
eCREST should be used to supplement traditional teaching Agree or strongly agree 110 (92.4) 35 (87.5) 145 (91.2) 0.631
Neither 7 (5.9) 4 (10.0) 11 (6.9)
Strongly disagree or disagree 2 (1.7) 1 (2.5) 3 (1.9)
eCREST helped me to learn clinical reasoning skills that I could apply to my clinical work Agree or strongly agree 99 (83.2) 35 (87.5) 134 (84.3) 0.811
Neither 16 (13.4) 4 (10.0) 20 (12.6)
Strongly disagree or disagree 4 (3.4) 1 (2.5) 5 (3.1)
Overall, using eCREST enhanced my learning Agree or strongly agree 103 (86.6) 34 (85.0) 137 (86.2) 0.938
Neither 14 (11.8) 5 (12.5) 19 (11.9)
Strongly disagree or disagree 2 (1.7) 1 (2.5) 3 (1.9)
I would use eCREST for my future professional development Agree or strongly agree 4 (100.0) 36 (90.0) 40 (90.9) 0.803
Neither 3 (7.5) 3 (6.8)
Strongly disagree or disagree 1 (2.5) 1 (2.3)

6.1.3 Easy

6.1.4 Difficulty

6.1.5 Teaching

6.1.6 Reasoning

6.1.7 Overall

6.1.8 Use