Ruiz 2016.

Study characteristics
Patient Sampling	Retrospective case‐control study; unclear how the data were collected.
Patient characteristics and setting	Participants were grouped as follows. Normal: eyes that had no systemic or ocular diseases, prior ocular surgery, or regular astigmatism ≤ 1.5 D Astigmatism: symmetric and straight bowtie pattern on the anterior keratometry with AST < 1.5 D After refractive surgery: people who had undergone laser‐assisted subepithelial keratectomy for myopia, hyperopia, or astigmatism correction. The spherical corrections ranged from –1 D to –6 D, and cylindrical corrections ranged from –1 D to–3 D. Forme fruste: non‐symptomatic Pentacam tomography and elevation, but fellow eye has manifest keratoconus Keratoconus: diagnosis of mild or moderate keratoconus taking into account the presence of a zone of increased power with abnormal thinning and elevation in the anterior and posterior surfaces combined with patient and family history, changes in refraction, and slit‐lamp examination Exclusion criteria: any systemic disease, a history of ocular surgery (except for laser refractive surgery in the refractive surgery group), cross‐linking, ametropia > 610 D (for the normal group), and very advanced keratoconus with corneal scarring. Note that keratoconus suspect cases (i.e. eyes that the investigators could not reliably categorize as either normal or keratoconus based on Pentacam or slit‐lamp examination) were removed from the analysis, as they could not be properly categorized for training purposes.
Index tests	Support vector machine First, the data were preprocessed to ensure there were no impossible values or missing data. The second step was a correlation‐based hierarchical clustering. The output of this analysis is dendrogram, in which variables are organized along branches according to their degree of correlation with each other. Next, the dendrogram was revised to select only 1 variable from each branch, which was typically the variable with the highest clinical relevance. This procedure led to a data set of 22 variables. All classifications were performed using a linear kernel support vector machine; this method provides a linear decision boundary for binary classification problems.
Target condition and reference standard(s)	An experienced keratoconus specialist and an experienced optometrist classified the eyes into 5 groups using criteria based on patient history and corneal tomography maps from the Pentacam.
Flow and timing	All participants received the same reference standard and were included in a 2 × 2 table.
Comparative	Not applicable
Notes	Study supported by an Agency for Innovation by Science and Technology grant.
Methodological quality
Item	Authors' judgement	Risk of bias	Applicability concerns
DOMAIN 1: Patient selection
Was a consecutive or random sample of patients enrolled?	Unclear
Was a case‐control design avoided?	No
Did the study avoid inappropriate exclusions?	Yes
Could the selection of patients have introduced bias?		Unclear risk
Are there concerns that the included patients and setting do not match the review question?			High
DOMAIN 2: Index test (All tests)
Were the index test results interpreted without knowledge of the results of the reference standard?	Yes
If a threshold was used, was it pre‐specified?	Unclear
Was the model designed in an appropriate manner?	Yes
Could the conduct or interpretation of the index test have introduced bias?		Low risk
Are there concerns that the index test, its conduct, or interpretation differ from the review question?			Low concern
DOMAIN 3: Reference standard
Is the reference standard likely to correctly classify the target condition?	Yes
Were the reference standard results interpreted without knowledge of the results of the index tests?	Yes
Could the reference standard, its conduct, or its interpretation have introduced bias?		Low risk
Are there concerns that the target condition as defined by the reference standard does not match the question?			Low concern
DOMAIN 4: Flow and timing
Did all patients receive the same reference standard?	Yes
Were all patients included in the analysis?	Yes
Could the patient flow have introduced bias?		Low risk
DOMAIN 5: Comparative