Table 4 Number of studies meeting individual STARD modified criteria for validating health administrative data

TITLE, KEYWORDS, ABSTRACT

Identify article as study of assessing diagnostic accuracy

7/7

Identify article as study of administrative data

7/7

INTRODUCTION:

State disease identification & validation one of goals of study

7/7

METHODS:

Participants in validation cohort:

Describe validation cohort (Cohort of patients to which reference standard was applied)

7/7

Age

6/7

Disease

6/7

Severity

1/7

Location/Jurisdiction

7/7

Describe recruitment procedure of validation cohort

6/7

Inclusion criteria

6/7

Exclusion criteria

6/7

Describe patient sampling (random, consecutive, all, etc.)

7/7

Describe data collection

7/7

Who identified patients and did selection adhere to patient recruitment criteria

5/7

Who collected data

6/7

A priori data collection form

6/7

Disease classification

7/7

Split sample (i.e. re-validation using a separate cohort)

0/7

Test Methods:

Describe number, training and expertise of persons reading reference standard

6/7

If >1 person reading reference standard, quote measure of consistency (e.g. kappa)

6/7

Blinding of interpreters of reference standard to results of classification by administrative data e.g. Chart abstractor blinded to how that chart was coded

6/7

Statistical Methods:

Describe methods of calculating/comparing diagnostic accuracy

7/7

RESULTS:

Participants:

Report when study done, start/end dates of enrollment

4/7

Describe number of people who satisfied inclusion/exclusion criteria

6/7

Study flow diagram

2/7

Test results:

Report distribution of disease severity

1/7

Report cross-tabulation of index tests by results of reference standard

7/7

Estimates:

Report at least 4 estimates of diagnostic accuracy

5/7

Diagnostic Accuracy Measures Resported:

Sensitivity

5/7

Spec

5/7

PPV

6/7

NPV

4/7

Likelihood ratios

1/7

kappa

4/7

Area under the ROC curve/C-statistic

0/7

Accuracy/agreement

1/7

Report accuracy for subgroups (e.g. age, geography, differen sex, etc.)

2/7

If PPV/NPV reported, ratio of cases/controls of validation cohort approximate prevalence of condition in the population

2/7

Report 95 % confidence intervals for each diagnostic measure

5/7

DISCUSSION:

Discuss the applicability of the validation findings

7/7