Table 3.
Best single term filters for high sensitivity (keeping specificity ≥50%), high specificity (keeping sensitivity ≥50%), and optimal balance of sensitivity and specificity, and performance of some other single term filters from 24 027 considered. Values are percentages (95% confidence intervals)
| Filters | Sensitivity | Specificity | Precision | Accuracy | ||
|---|---|---|---|---|---|---|
| Ovid† | PubMed translation‡ | Journal set | ||||
| Best sensitivity | ||||||
| kidney$.mp | kidney*[tw] | Development | 78.7 (75.4 to 82.0) | 97.2 (96.5 to 97.9) | 88.9 (86.2 to 91.6) | 93.0 (92.1 to 94.0) |
| Validation | 79.6 (75.25 to 83.9) | 95.6 (94.7 to 96.6) | 78.4 (74.0 to 82.8) | 93.0 (91.9 to 94.1) | ||
| Best specificity | ||||||
| exp *kidney diseases/ | “kidney diseases”[majr] | Development | 57.5 (53.5 to 61.5) | 98.6 (98.1 to 99.1) | 92.4 (89.7 to 95.1) | 89.5 (88.3 to 90.6) |
| Validation | 41.4 (36.1 to 46.7) | 97.6 (96.8 to 98.3) | 77.1 (70.9 to 83.3) | 88.2 (86.8 to 89.7) | ||
| Best optimisation of sensitivity and specificity | ||||||
| kidney$.mp | kidney*[tw] | Development | 78.7 (75.4 to 82.0) | 97.2 (96.5 to 97.9) | 88.9 (86.2 to 91.6) | 93.0 (92.1 to 94.0) |
| Validation | 79.6 (75.25 to 83.9) | 95.6 (94.7 to 96.6) | 78.4 (74.0 to 82.8) | 93.0 (91.9 to 94.1) | ||
| Other single item filters | ||||||
| *kidney transplantation/ | “kidney transplantation” [majr:noexp] | Development | 28.6 (24.9 to 32.2) | 99.7 (99.5 to 99.9) | 96.6 (93.9 to 99.3) | 83.8 (82.4 to 85.2) |
| Validation | 27.0 (22.3 to 31.8) | 99.9 (99.7 to 100) | 97.8 (94.8 to 100) | 87.8 (86.4 to 89.2) | ||
| glomerul$.mp | glomerul*[tw] | Development | 20.8 (17.5 to 24.1) | 99.3 (98.9 to 99.6) | 89.1 (83.9 to 94.3) | 81.8 (80.3 to 83.2) |
| Validation | 24.0 (19.4 to 28.6) | 99.5 (99.1 to 99.8) | 89.9 (83.6 to 96.2) | 86.9 (85.5 to 88.4) | ||
| glomerular filtration rate/ | “glomerular filtration rate”[mh:noexp] | Development | 6.3 (4.3 to 8.2) | 99.8 (99.7 to 100) | 92.5 (84.3 to 100) | 78.9 (77.4 to 80.5) |
| Validation | 7.2 (4.4 to 9.9) | 100 (100 to 100) | 100 (100 to 100) | 84.6 (83.0 to 86.2) | ||
| renal.ti. | renal[ti] | Development | 34.2 (30.4 to 38.0) | 99.5 (99.2 to 99.8) | 94.8 (91.9 to 97.8) | 84.9 (83.5 to 86.3) |
| Validation | 35.1 (30.0 to 40.3) | 98.7 (98.1 to 99.2) | 84.2 (78.1 to 90.2) | 88.1 (86.7 to 89.6) | ||
| *proteinuria/ | “proteinuria”[majr:noexp] | Development | 2.9 (1.5 to 4.2) | 99.9 (99.8 to 100) | 89.5 (75.7 to 100) | 78.3 (76.7 to 79.8) |
| Validation | 1.8 (0.4 to 3.2) | 100 (100 to 100) | 100 (100 to 100) | 83.7 (82.1 to 85.3) | ||
| exp renal replacement therapy/ | “renal replacement therapy”[mh] | Development | 48.1 (44.0 to 52.1) | 99.5 (99.2 to 99.8) | 96.6 (94.5 to 98.7) | 88.0 (86.8 to 89.3) |
| Validation | 50.5 (45.1 to 55.8) | 99.5 (99.1 to 99.8) | 94.9 (91.7 to 98.2) | 91.3 (90.1 to 92.6) | ||
| dialy$.mp. | dialy*[tw] | Development | 26.6 (23.0 to 30.1) | 99.9 (99.7 to 100) | 98.1 (96.0 to 100) | 83.5 (82.1 to 84.9) |
| Validation | 27.0 (22.3 to 31.8) | 99.9 (99.7 to 100) | 97.8 (94.8 to 100) | 87.8 (86.4 to 89.2) | ||
†Ovid fields: exp=exploded search term that automatically includes narrower indexing terms (for example, with exploded term kidney diseases, terms such as glomerulonephritis, renal insufficiency, uremia, and kidney failure are included in search); $=truncation character; mp=multiple posting (term appears in title, abstract, or MeSH); /=MeSH character; *=focused MeSH term; ti=term present in title.
‡PubMed fields: *=truncation character; [majr]=exploded and focused MeSH term; [majr:noexp]=not exploded and focused MeSH term; [tw]=text word present in title, abstract, or MeSH; [ti]=term present in title; [mh]=exploded MeSH term; [mh:noexp]=not exploded MeSH term.