Table 1 Characteristics of selected studies

Pain et al. 1996 [35]

Randomised controlled trial with one comparison group

UK; first year medical students.

92 students were allocated to a 15-min video plus the usual lecture; of these 71 saw the video and 85 took the test

93 students were allocated to the usual lecture; 91 took the test

1) Overall performance score out of 68: median (IQR). Intervention group: 42.0 (36.5–47.5); control group: 39.0 (35.0–45.0), p = 0.046;

2) Death certification score out of 44: median (IQR). Intervention group: 26 (22–30); control group: 25 (20–28), p = 0.066

Myers and Farquhar 1998 [34]

Quasi-experimental study with pre and post assessment of death certificates

Canada; Residents assigned to an internal medicine rotation

75-min seminar on proper completion of death certificates. 83 certificates completed after the intervention

146 death certificates completed before the intervention

1) At least one major error (mechanism of death only, improper sequence, competing causes): 48 (32.9%) pre- and 13 (15.7%) post-intervention, p = 0.01;

2) At least one minor error (absence of time intervals, abbreviations, mechanism followed by legitimate underlying cause of death (UCOD)): 123 (84.2%) pre- and 75 (90.4%) post-intervention, p = 0.19;

3) Mechanism of death only: 23 (15.8%) pre- and 4 (4.8%) post-intervention, p = 0.01;

4) Improper sequence: 23 (15.8%) pre- and 5 (6.0%) post-intervention, p = 0.03;

5) Competing causes: 11 (7.5%) pre- and 7 (8.4%) post-intervention, p = 0.81;

6) Absence of time interval: 101 (69.2%) pre- and 63 (75.9%) post-intervention, p = 0.28;

7) Abbreviations: 29 (19.9%) pre- and 15 (18.1%) post-intervention, p = 0.11;

8) Mechanism followed by legitimate UCOD: 67 (45.9%) pre- and 30 (36.1%) post-intervention, p = 0.15

Lakkireddy et al. 2007 [13]

Randomised interventional study with one comparison group

USA; 219 internal medicine residents from five teaching hospitals

Group I (45-min interactive workshop or ‘workshop group’, n = 105), 100 were available for analysis

Group II (printed handout or ‘print group’, n = 114), content was same as in Group I, 100 were available for analysis

1) Mid-America Heart Institute (MAHI) Death Certificate Score > 19: Group 1 20 (20%) pre- and 82 (82%) post-intervention, p < 0.001; Group II 18 (18%) pre- and 58 (58%) post-intervention, p < 0.001; Both groups 38 (19%) pre- and 140 (70%) post-intervention, p < 0.001;

2) MAHI Death Certificate mean Score: Group 1 (n = 100) 13.7 (+/−  5.9) pre- and 24.1 (+/− 4.8) post-intervention, p < 0.001; Group II (n = 100) 14.1 (+/− 4.6) pre- and 19.1 (+/− 5.4) post-intervention, p < 0.001, Both groups (n = 200) 13.9 (+/− 5.3) pre- and 21.6 (+/− 5.7) post-intervention, p < 0.001;

3) Correct identification of cause of death: Group 1 15 (15%) pre- and 91 (91%) post-intervention, p < 0.001; Group II 16 (16%) pre and 55 (55%) post p < 0.001; Both groups 31 (15.5%) pre- and 146 (84.5%) post-intervention, p < 0.001;

4) Erroneously identified cardiac death: Group 1 56 (56%) pre- and 6 (6%) post-intervention, p < 0.001; Group II 64 (64%) pre- and 43 (43%) post-intervention, p = 0.02; Both groups 120 (60%) pre- and 49 (24.5%) post-intervention, p < 0.001

Vilar and Perez-Mendez 2007 [41]

Quasi-experimental study with pre and post assessment

Spain; 166 Medical trainees from various medical specialties (family medicine, internal medicine, anaesthesiology, general surgery, critical care medicine) in seven teaching hospitals

90-min seminar on the proper completion of death certificates delivered as an interactive workshop, 166 death certificates filled after the intervention

166 death certificates filled before the intervention

1) At least one error: 71.1% pre- and 9.0% post-intervention, p < 0.0001;

2) Mechanism of death only: 71 (42.6%) pre- and 4 (2.4%) post-intervention, p < 0.0001;

3) Improper sequence: 31 (18.7%) pre- and 1 (0.6%) post-intervention, p < 0.0001;

4) Listing two causally unrelated, etiologically specific diseases as the cause of death: 10 (6%) pre- and 5 (3.0%) post-intervention, p = 0.290;

5) Abbreviations: 9 (5.4%) pre- and 5 (3.0%) post-intervention, p = 0.413;

6) Mechanism as UCOD: 22 (13.3%) pre- and 0 (0.0%) post-intervention, p < 0.0001;

7) Listing the cause of death in Part II: 46 (27.7%) pre- and 5 (3.0%) post-intervention, p < 0.0001

Degani et al. 2009 [14]

Quasi-experimental study with pre and post assessment

USA; All third-year medical students from Mercer University School of Medicine rotating at Medical Centre of Central Georgia

129 students were presented with a web-based tutorial lasting approximately 30 min, designed for self-study; 123 death certificates included in analysis

123 death certificates completed before the intervention

1) Modified version of MAHI Death Certificate Scoring system used (n = 123) with score out of 22; mean (SD): 11.75 (3.2) pre- and 18.85 (2.56) post-intervention. Mean difference 7.10 (3.86), p < 0.0001, t = 20.39

Pandya et al. 2009 [36]

Quasi-experimental study with pre and post assessment

India; 43 residents of target postgraduate disciplines at 550-bed teaching hospital

A structured 90-min presentation in one workshop followed by an interactive session. Second and third workshops included group activities. After the intervention 102 death certificates were assessed

96 death certificates from the pre-intervention period

1) Major: Unacceptable UCOD: 38 (39.6%) pre- and 25 (24.5%) post-intervention, p = 0.034;

2) Major: Mechanism only without UCOD: 13 (13.5%) pre- and 1 (1.0%) post-intervention, p = 0.001;

3) Major: Improper sequence: 24 (25.0%) pre- and 6 (5.9%) post-intervention, p = 0.0004;

4) Major: Competing causes: 37 (38.5%) pre- and 26 (25.5%) post-intervention, p = 0.069;

5) Minor: Absence of time interval: 28 (29.2%) pre- and 28 (27.5%) post-intervention, p = 0.91;

6) Minor: Abbreviations: 21 (21.9%) pre- and 34 (33.3%) post-intervention, p = 0.1;

7) Minor: Mechanism followed by legitimate UCOD: 13 (13.5%) pre- and 8 (7.8%) post-intervention, p = 0.28

Pieterse et al. 2009 [37]

Randomised interventional study with one comparison group

South Africa; 24 medical interns who had completed at least 6 months of their internship at an academic tertiary hospital

Death certification educational intervention consisting of a 45-min didactic teaching session and an educational handout (i.e. written guide). 13 were in the group

Written guide only. 11 were in the group

1) Score out of 30 for avoiding minor and major errors; mean (SD): Group 1 11.8 (1.8) pre- and 24.5 (1.0) post-intervention, p < 0.001; Group II 15.5 (1.5) pre- and 25.3 (1.1) post-intervention, p < 0.001;

2) Score out of 30: Group 1 15% pre- and 85% post-intervention, p = 0.004; Group II 9% pre and 91% post-intervention, p = 0.004;

3) Major: Mechanism only: Group 1 69% pre- and 15% post-intervention, p = 0.016; Group II 37% pre- and 27% post-intervention, p = 1.000;

4) Major: Improper sequence: Group 1 54% pre- and 0% post-intervention, p = 0.016; Group II 36% pre- and 36% post-intervention, p = 1.000;

5) Major: Competing causes: Group 1 69% pre- and 8% post-intervention, p = 0.008; Group II 73% pre- and 9% post-intervention, p = 0.039;

6) Minor: Absence of time interval: Group 1 77% pre- and 23% post-intervention, p = 0.016; Group II 64% pre- and 18% post-intervention, p = 0.063;

7) Minor: Abbreviations: Group 1 62% pre- and 8% post-intervention, p = 0.016; Group II 73% pre- and 9% post-intervention, p = 0.039

Hemans-Henry, Greene and Koppaka 2012 [31]

Non-randomised experimental study with one comparison group

USA; postgraduate year 1 (PGY1) internal medicine and general surgery residents (n = 114) and postgraduate year 2 (PGY2) internal medicine, emergency medicine, and general surgery residents (n = 113)

PGY1 residents completed a pre-test, e-learning course, post-test, and course evaluation. 59 completed all evaluations

74 PGY2 residents completed the same pre-test

The test consisted of 10 multiple-choice questions. The PGY1 and PGY2 average pre-test scores were comparable (59% and 61%, respectively). The average PGY1 post-test score was higher than both the average PGY1 pre-test score (72% vs 59%, respectively; p = 0.01); and the average PGY2 pre-test score (72% vs 61%, respectively; p = 0.001)

Walker et al. 2012 [15]

Quasi-experimental study with pre and post assessment

Fiji; Medical students in their final year who were undertaking their final week of education at the university

WHO training tool plus access to the online certification module. Participants completed the death certification module in the Fiji School of Medicine computer laboratory. 13 case vignettes were used in the post-test assessment. Responses of 23 participants were included

13 case vignettes were used in the pre-test assessment. Responses of 23 participants were included.

1). Quality index score and % were used (total score 15 per certificate ×  13 certificates = 195; lower is better). Pre-test (n = 23) mean: 57.22 (22.91); post-test (n = 23) mean: 30.30 (11.66); mean change in quality index 26.91 (11.25); individual scores available and SD can be calculated;

2) Mean error rate: 33.14% pre- and 20.27% post-test;

3) Abbreviations: 19.40% improvement between pre- and post-test;

4) Reporting a legitimate sequence of events in Part I: 19.06% improvement;

5) Reporting only one cause per line: 18.06% improvement;

6) Reporting a disease and not a mode of death: 17.3% improvement;

7) Legibility: 1.67% improvement

Ali and Hamadeh 2013 [26]

Quasi-experimental study with pre and post assessment

Bahrain; 27 secondary healthcare physicians

Interactive workshop. Post-workshop death certificates were used, with each participant (n = 27) completing one certificate

Pre-workshop death certificates, with each participant (n = 27) completing one certificate

1) Listing mechanism without underlying disease: 2 (7.4%) pre- and 0 (0.0%) post-intervention, p = 0.491;

2) Improper sequence 1 (3.7%) pre- and 2 (7.4%) post-intervention, p = 1.0;

3) Listing two causally unrelated, etiologically specific diseases as the cause of death: 3 (11.1%) pre- and 0 (0.0%) post-intervention, p = 0.236;

4) Listing mechanism of death followed by proper UCOD: 18 (66.7%) pre- and 9 (33.3%) post-intervention, p = 0.009;

5) Listing the cause of death as one of the other significant conditions contributing to the death but not causally related to the immediate cause of death: 1 (3.7%) pre- and 0 (0.0%) post-intervention, p = 1.0;

6) Abbreviations 0 (0.0%) pre- and 0 (0.0%) post-intervention;

7) No error: 2 (7.4%) pre- and 16 (59.3%) post-intervention, p < 0.001

Azim et al. 2014 [28]

Quasi-experimental study with pre- and post-assessment (described as an observational study: audit- intervention and a re-audit)

India; 12 resident doctors undergoing their subspecialty training in critical care medicine

Educational intervention programme consisting of a lecture followed by an interactive session. 75 death certificates post-intervention were audited

75 pre-intervention death certificates

1) Major error: Unacceptable UCOD: 74 (98.6%) pre- and 31 (41.3%) post-intervention, p = 0.001;

2) Major: Mechanism only: 45 (60%) pre- and 11 (14.6%) post-intervention, p = 0.001;

3) Major: Improper sequence: 67 (89.3%) pre- and 27 (36.0%) post-intervention, p = 0.001;

4) Major: Competing causes: 66 (88.0%) pre- and 10 (13.3%) post-intervention, p = 0.001;

5) Minor: Absence of time interval: 75 (100.0%) pre- and 17 (22.6%) post-intervention, p = 0.001;

6) Minor: Abbreviations: 67 (86.3%) pre- and 22 (29.3%) post-intervention, p = 0.001;

7) Minor: Mechanism followed by legitimate UCOD: 12 (16.0%) pre- and 7 (6.6%) post-intervention, p = 0.55

Alonso-Sardon et al. 2015 [27]

Quasi-experimental study with pre- and post-assessment

Spain; 308 sixth year medical students

A formative intervention that included a five-hour on-site seminar-workshop, consisting of both theoretical and practical parts. Five completed death certificates were selected for comparison

Five death certificates filled before the intervention were selected for comparison

1) Major indexes consisted of assessment of underlying, intermediate and immediate causes;

2) Minor index: Mechanisms of death instead of causes;

3) Minor index: Inappropriate and vague terms;

4) Minor index: Abbreviations;

5) Minor index: Existence of multiple UCODs;

6) Minor index: Capital letters

Spain; 62 practising family doctors and interns

A formative intervention including a five-hour on-site seminar-workshop with two parts; theoretical and practical. Five completed death certificates were selected for comparison

Five death certificates completed before the intervention were selected for comparison

1) Major indexes consisted of assessment of underlying, intermediate and immediate causes;

2) Minor index: Mechanisms of death instead of causes;

3) Minor index: Inappropriate and vague terms;

4). Minor index: Abbreviations;

5) Minor index: Existence of multiple UCODs;

6) Minor index: Capital letters

Miki et al. 2018 [32]

Quasi-experimental study with pre and post assessment

Peru; Doctors received either

1. an online intervention; or

2. an online intervention and a training intervention

1.’Online intervention’ - one hour on the online system (SINADEF) (900 death certificates)

2. ‘Online and training intervention’ - one hour on SINADEF and one-hour training on certification of cause of death (900 death certificates)

Pre intervention’ 300 pre-intervention death certificates

1): Major: Multiple causes per line: 1. Pre: 2.0%; 2. Online: 1.3%; 3. Online and training: 0.6%, p > 0.05;

2) Major: Absence of time interval: 1. Pre: 96.0%; 2. Online: 47.1%; 3. Online and training: 30.0%, p < 0.01;

3) Major: Incorrect sequence of events leading to death: 1. Pre: 40.3%; 2. Post: 25.9%, p < 0.05; 3. Online and training: 17.9%, p < 0.01;

4) Major: Ill-defined condition entered as UCOD: 1. Pre: 52.0%; 2. Post: 45.4%; 3. Online and training: 38.9%, p < 0.01;

5) Minor: Presence of blank lines within the sequence of events: 1. Pre: 11.3%; 2. Post: 0.2%; 3. Online and training: 0.3%, p < 0.01;

6) Minor: Abbreviations 1. Pre: 11.7%; 2. Post: 4.6%; 3. Online and training: 4.1%, p < 0.01);

7) Minor: Additional errors on the certificate: 1. Pre: 32.3%; 2. Post: 26.6%, p > 0.05; 3. Online and training: 21.0%, p < 0.01

Sudharson et al. 2019 [40]

Quasi-experimental study (described as a cross sectional study)

India; Teaching faculty post-graduates, junior residents and interns (who have completed medicine and surgery postings) (n = 80)

Lecture. Death certificates completed post-intervention based on a case scenario (n = 80)

Death certificates completed pre-intervention based on a case scenario (n = 80)

1) Major: Incorrect sequence of events: 48 (60.0%) pre- and 3 (3.75%) post-intervention;

2) Major: Unrelated causal events in sequence 6 (7.5%) pre- and 0 (0.0%) post-intervention;

3) Major: At least 1 major error: 51 (63.75%) pre- and 3 (3.75%) post-intervention;

4) Minor: Missing time interval: 68 (85.0%) pre- and 0 (0.0%) post-intervention;

5) Minor: Mechanism followed by legitimate UCOD: 66 (82.5%) pre- and 1 (1.25%) post-intervention;

6) Minor: Abbreviations: 18 (22.5%) pre- and 0 (0.0%) post-intervention;

7) Minor: At least 1 minor error: 78 (97.5%) pre- and 1 (1.25%) post-intervention

Hart et al. 2020 [30]

Comparison paper comparing multiple countries; ll quasi experimental studies with pre- and post-assessment

PNG; Physicians

Direct training of physicians on completion of death certificates. 378 post-training MCCODs

948 baseline MCCODs

1) Major: Multiple causes per line: 16.3% pre- and 7.9% post-intervention;

2) Major: Incorrect sequence: 41.7% pre- and 20.3% post-intervention;

3) Major: Illegible handwriting: 4.3% pre- and 1.6% post-intervention;

4) Major: Ill-defined cause as UCOD: 39.1% pre- and 18.7% post-intervention;

5) Major: Additional information on neoplasm not available: 4.5% pre; and 2.3% post-intervention;

6) At least one major error: 55.6% pre- and 30.7% post-intervention;

7) Minor: Abbreviations: 19.8% pre- and 5.4% post-intervention;

8) Minor: Absence of time interval: 74.7% pre- and 42.3% post-intervention;

9) Minor: Additional errors on the certificate: 5.3% pre- and 5.1% post-intervention;

10) At least one error: 86.4% pre- and 60.6% post-intervention

Philippines; Physicians

Training of trainers and then direct training. 959 post-training MCCODs

975 baseline MCCODs

1) Major: Multiple causes per line: 21.2% pre- and 6.0% post-intervention;

2) Major: Incorrect sequence: 27.1% pre- and 12.4% post-intervention;

3) Major: Illegible handwriting: 0.3% pre- and 1.1% post-intervention;

4) Major: Ill-defined cause as UCOD: 28.6% pre- and 15.5% post-intervention;

5) Major: Additional information on external causes not available: 4.8% pre- and 1.2% post-intervention;

6) Major: Additional information on neoplasm not available: 2.3% pre- and 1.9% post-intervention;

7) At least one major error: 41.6% pre- and 22.6% post-intervention;

8) Minor: Abbreviations: 7.1% pre- and 0.8% post-intervention;

9) Minor: Absence of time interval: 37.4% pre- and 23.7% post-intervention;

10) Minor: Additional errors on the certificate: 5.3% pre- and 1.1% post-intervention;

11) At least one error: 72.9% pre- and 43.6% post-intervention

Myanmar; Physicians

Training of trainers and then direct training. 600 post-training MCCODs assessed

595 baseline MCCODs assessed

1) Major: Multiple causes per line: 24.4% pre- and 10.8% post-intervention;

2) Major: Incorrect sequence: 7.9% pre- and 5.8% post-intervention;

3) Major: Illegible handwriting: 4.2% pre- and 2.8% post-intervention;

4) Major: Ill-defined cause as UCOD: 44.5% pre- and 32.7% post-intervention;

5) Major: Additional information on neoplasm not available: 1.4% pre- and 0.3% post-intervention;

6) At least one major error: 63.2% pre- and 44.8% post-intervention;

7) Minor: Presence of blank lines within the sequence of events: 0.2% pre- and 0.3% post-intervention;

8)Minor: Abbreviations: 50.8% pre- and 31.0% post-intervention;

9) Minor: Absence of time interval: 93.4% pre- and 65.3% post-intervention;

10) Minor: Additional errors on the certificate: 1.6% pre- and 0.7% post-intervention;

11) At least one error: 99.8% pre- and 74.8% post-intervention

Sri Lanka; Physicians

Training of trainers and then direct training. 558 post-training MCCODs assessed

517 baseline MCCODs assessed

1) Major: Multiple causes per line: 38.9% pre- and 20.8% post-intervention;

2) Major: Incorrect sequence: 37.1% pre- and 17.0% post-intervention;

3) Major: Illegible handwriting: 0.6% pre- and 0.0% post-intervention;

4) Major: Ill-defined cause as UCOD: 4.4% pre- and 10.6% post-intervention;

5) Major: Additional information on neoplasm not available: 4.3% pre- and 0.5% post-intervention;

6) At least one major error: 58.8% pre- and 37.5% post-intervention;

7) Minor: Presence of blank lines within the sequence of events: 2.1% pre- and 2.7% post-intervention;

8) Minor: Abbreviations: 36.0% pre- and 20.3% post-intervention;

9) Minor: Absence of time interval: 87.0% pre- and 53.2% post-intervention;

10) Minor: Additional errors on the certificate: 2.9% pre- and 0.2% post-intervention;

11) At least one error: 95.4% pre- and 68.5% post-intervention

Wood, Weinberg and Weinberg 2020 [42]

Quasi-experimental study with pre, immediate-post and 2-month-post assessment

Canada; 63 residents and nine staff physicians participated in the pre-survey; 67 residents and eight staff in the immediate-post survey; 18 residents and six staff in the 2-month-post survey

60-min didactic session with case scenarios at grand rounds. 372 mock death certificates completed at immediate-post survey and 103 at 2-month-post survey

351 mock death certificates completed pre-intervention

1) Mechanism of death used as underlying cause of death: Error Occurrence (EO) Rate (%): 17 pre-; 1 immediate-post; 3 at 2 months, p < 0.05;

2) Absence of UCOD: EO Rate: 15 pre-; 2 immediate-post; 10 at 2 months, p < 0.05;

3) Incorrect manner of death recorded: EO Rate: 23 pre-; 2 immediate-post; 2 at 2 months, p < 0.05;

4) Abbreviations: EO Rate: 29 pre-; 26 immediate-post, p > 0.05; 5 at 2 months, p < 0.05;

5) Signs and symptoms listed: 1 pre-; 2 immediate-post; 0 at 2 months, p > 0.05;

6) Illogical sequence: 4 pre-; 2 immediate-post; 4 at 2 months, p > 0.05;

7) UCOD not in last line: 23 pre-; 5 immediate-post; 7 at 2 months, p < 0.05;

8) Part 2 items listed in part 1 (all errors preceding this in the row): 13 pre-; 3 immediate-post; 11 at 2 months, p < 0.05;

9) Listing medical conditions: 8 pre-; 0 immediate-post; 0 at 2 months, p < 0.05;

10) Part 1 items listed in Part 2: 14 pre-; 4 immediate-post; 2 at 2 months, p < 0.05;

11) Incorrect manner of death recorded: 23 pre-; 2 immediate-post; 2 at 2 months, p < 0.05;

12) More than once condition per line in Part 1: 1 pre-; 1 immediate-post; 1 at 2 months, p > 0.05

Abos et al. 2006 [25]

Quasi-experimental study with pre- and post-assessment

Spain; Group of 135 physicians assigned to practice in the reformed network of primary care

90-min seminar (BEDTAR programme); post-intervention assessment of 3 cases

Pre-intervention performance

1) Error item ‘immediate cause’ in relation to each case;

2) Error item ‘cardiopulmonary arrest’ in relation to each case;

3) Error item ‘intermediate cause’ in relation to each case;

4) Error item ‘root cause’ in relation to each case available;

5) Error item ‘double fundamental cause’ in relation to each case;

6) Error item ‘Other processes’ in relation to each case;

7) Error item ‘Abbreviations’ in relation to each case;

8) Error item ‘Legible letter’ in relation to each case;

9) Error item ‘logical sequence’ in relation to each case;

10) Error item ‘use all information’ in relation to each case;

11) Error item ‘Invention’ in relation to each case;

12) Error item ‘Poor defined entity’ in relation to each case;

13) Error item ‘Use of lowercase’ in relation to each case

Canelo and Gonzalez 1995 [29]

Quasi-experimental study with pre- and post-assessment

Spain; 173 sixth year medical students

Seminar; six post-intervention death certificates completed by each participant

Six pre-intervention death certificates completed by each participant

1) Basic or fundamental cause is correct: 937 (90.26%) pre- and 1012 (97.49%) post-intervention;

2) Logical sequence is correct: 683 (65.79%) pre and 906 (87.28%) post-intervention;

3) Various basic causes of death are correct: 981 (94.50%) pre- and 1027 (98.94%) post-intervention;

4) Mechanisms/cause of death is correct: 879 (84.68%) in pre and 1004 (96.72%) in post; 5) No imprecise terms: 1026 (98.84%) pre- and 1033 (99.51%) post-intervention;

6) No Abbreviations or acronyms: 882 (84.97%) pre- and 1031 (99.32%) post-intervention;

7) Legible and lowercase: 491 (47.30%) pre- and 999 (96.24%) post-intervention

Selinger, Ellis and Harrigton 2007 [38]

Quasi-experimental study with pre- and post-assessment (described as a clinical audit)

England; Senior house officers (SHOs), staff grades, specialist registrars and consultants

Education was in three forms: (1) Presentation of the findings of the pre-assessment during a clinical governance meeting; (2) Each doctor was given individualised performance data and (3) the topic was highlighted during the induction of new doctors. Post-intervention, 85 case notes were assessed

140 case notes

1) Consultants’ name not given: 48.6% pre- and 18.0% post-intervention;

2) At least one mistake or omission: 58.6% pre- and 20.0% post-intervention;

3) Completed by doctors who did not meet the requirements of being involved in the patient’s care: 13.6% pre- and 2.4% post-intervention, p = 0.01

Myers and Eden 2007 [33]

Quasi-experimental study with pre- and post-assessment

Canada; 25 family physicians.

Half-day workshop with case scenarios; 16 completed the post-test

21 completed the pre-test

1) Decline in use of mechanisms of death as the UCOD;

2) Increased use of more specific diseases as the UCOD;

3) More knowledgeable about not using old age as a cause of death

Suarez et al. 1998 [39]

Quasi-experimental study with pre- and post-assessment

Spain; Medical students, interns and trainees in family and community medicine

120-min teaching programme. 472 post-intervention exercises

472 pre-intervention exercises

1) Correct immediate cause: 89.8% pre- and 98.5% post-intervention;

2) Correct intermediate cause: 78.2% pre- and 97.7% post-intervention;

3) Correct initial or fundamental cause: 83.5% pre- and 97.9% post-intervention;

4) Correct other processes: 91.3% pre- and 94.1% post-intervention;

5) Correct basic cause of death: 78.4% pre- and 97.2% post-intervention;

6) Legible: 98.1% pre- and 98.5% post-intervention;

7) Logical sequence: 97.9% pre- and 99.8% post-intervention;

8) No abbreviations or acronyms: 80.9% pre- and 82.6% post-intervention;

9) No omission of diseases: 82.6% pre- and 91.3% post-intervention;

10) Absence of causes not described: 88.1% pre- and 97.0% post-intervention;

11) Correct causal sequence: 82.0% pre- and 98.7% post-intervention