Hospital/vendor A (n=50) | Hospital/vendor B (n=50) | |||||
---|---|---|---|---|---|---|
Content criteria | Reviewer 1 | Reviewer 2 | IRR* | Reviewer 1 | Reviewer 2 | IRR* |
Meaningful use view-download-transmit | ||||||
Patient name | 50 (100%) | 50 (100%) | κ=1.00 | 50 (100%) | 50 (100%) | κ=1.00 |
Admission/discharge: date and location† | 50 (100%) | 50 (100%) | κ=1.00 | 25 (50%) | 25 (50%) | κ=1.00 |
Reason for hospitalisation | 46 (92%) | 39 (78%) | κ=0.57‡ | 49 (98%) | 46 (92%) | κ=0.88‡ |
Inpatient care team (primary and consultants) | 0 | 0 | κ=1.00 | 0 | 0 | κ=1.00 |
Procedures performed during admission | 1 (2%) | 0 | κ=0.96‡ | 0 | 0 | κ=1.00 |
Problem list (updated) | 50 (100%) | 50 (100%) | κ=1.00 | 50 (100%) | 50 (100%) | κ=1.00 |
Medication list (reconciled) | 50 (100%) | 50 (100%) | κ=1.00 | 50 (100%) | 50 (100%) | κ=1.00 |
Discharge instructions | 39 (78%) | 38 (76%) | κ=0.94 | 49 (98%) | 44 (88%) | κ=0.72‡ |
Manually entered by provider | 27 (54%) | 27 (54%) | κ=1.00 | 19 (38%) | 26 (52%) | κ=0.48 |
Generic/third-party only (‘boiler-plate’) | 33 (66%) | 32 (64%) | κ=0.87 | 38 (76%) | 42 (84%) | κ=0.75 |
Overall IRR | κ=0.92 (95% CI=0.88–0.96) | κ=0.95 (95% CI=0.93–0.98) | ||||
---|---|---|---|---|---|---|
Reviewer 1 | Reviewer 2 | IRR* | Reviewer 1 | Reviewer 2 | IRR* | |
Society of hospital medicine | ||||||
Follow-up appointments scheduled before discharge‡§ (%) | 18/115 (16%) | 23/131 (18%) | ρ=0.89 | 46/110 (42%) | 49/112 (44%) | ρ=0.98 |
Advise about anticipated problems (‘red flags’) | 27 (54%) | 25 (50%) | κ=0.60 | 23 (46%) | 0 | κ=0.24‡ |
Provide specific 24/7 call-back number | 0 | 0 | κ=1.00 | 0 | 0 | κ=1.00 |
*IRR=inter-rater reliability, calculated as Cohen's κ for categorical data and Spearman's rank correlation coefficient (ρ) for discrete data.
†Hospital/vendor B provided clinical summaries to patients without a discharge date, information that was included in the first two pages, but retained by the hospital (see online supplementary appendix 2).
‡Based on prevalence-adjusted κ.
§Differences in the percentage of follow-up appointments scheduled before hospital discharge may reflect variations in discharge planning practices at each site. Denominator reflects total number of follow-up appointments, which was >1 for many patients.