Table 2.
Experts, method of elicitation, and method of aggregation
| Study |
Experts and recruitment |
Approach and method of elicitation |
Aggregation |
|||||
|---|---|---|---|---|---|---|---|---|
| Type of experts | Recruitment | No. of experts | VIM or FIM | Method (summaries elicited) | Aggregation approach | Weights used in main exercise? | Nature of weights | |
| Garthwaite et al. [14] | NR | NR | 4 | VIM | Median and quartiles | Mathematical | No | – |
| Leal et al. [10] | Clinicians | Purposive | 6 | FIM | Four complementary intervals | Mathematical | No | – |
| Girling et al. [15] | Clinicians | Purposive | 5 | VIM* | NR | Consensus | – | – |
| Stevenson et al. [16] | NR | NR | NR | NR | NR | NR | NR | NR |
| Meads et al. [12] | Clinicians | Purposive | 21 | FIM | Chips and bins† | Mathematical | No | – |
| McKenna et al. [19] | NR | NR | 5 | FIM | Chips and bins | Mathematical | No | – |
| Haakma et al. [13] | Clinicians | Purposive | 14 | VIM | Mode and 95% CI | Mathematical | Yes | Objective weights |
| Stevenson et al. [17] | Clinicians | Purposive | 3 | VIM | Median and quartiles | Consensus | – | – |
| Speight et al. [25] | Clinicians | NR | 9 | FIM | Chips and bins† | NR | NR | |
| Sperber et al. [22] | Clinicians | NR | NR | VIM | Median and quartiles | Mathematical | Yes, but no details provided | Performance-based |
| Colbourn et al. [28] | NR | NR | 4 | VIM | Mean and 95% CI | Mathematical | NR | NR |
| Soares et al. [9] | Clinicians | NR | 23 | FIM | Chips and bins | Mathematical | No, but explored in a pilot | Performance-based weights explored |
| Bojke et al. [18] | Clinicians | Purposive | 5 | FIM | Chips and bins | Mathematical | Yes | Performance-based weights |
| Cao et al. [11] | Clinicians | NR | 2 | FIM | Mode and one percentile | Mathematical | No | – |
| Fischer et al. [23] | Clinicians | Purposive | 19 | VIM | Median and 80% CI | Mathematical | No, but explored | Performance-based weights explored |
| Poncet et al. [27] | Clinicians | NR | 13 | FIM | Chips and bins | Mathematical | No | – |
| Grigore et al. [24] | Clinicians | Purposive | 7 | FIM | Chips and bins + four complementary intervals | Mathematical | Yes, alongside equal weighting | Performance-based weights explored |
| Wilson et al. [20] | Clinicians + policy strategist | NR | 6 | NR | NR | NR | NR | NR |
| Meeyai et al. [21] | Clinicians + epidemiologists | NR | 10 | VIM | Mode and quartiles | Consensus | – | – |
| Grimm et al. [35] | NR | NR | 3 | NR | NR | Mathematical | NR | NR |
CI, confidence interval; FIM, fixed interval method; NR, not reported; VIM, variable interval method.
*Unclear, but description of results suggests that VIM has been used.
†
Includes studies that list the following methods: chips and bins, frequency chart, and histogram method.