Table 3.
Characteristics of the subsamples. Characteristics of the subsamples produced by the subsampling procedures and by simple random sampling applied on the 2000 times resampled bootstrap convenience sample. The Expected Value, Effect Size, Standardized Effect Size and the Bootstrap Intervals [90% BI] are computed as described in the Methods. For the three sampling procedures, the percentage of bootstrap resamples in which the characteristic had a higher value than in the random subsample is shown; values above 50% indicate an overall increase while values below 50% indicate a decrease
| Sample Characteristic | Expected Value 22[90% BI] | Effect Size [90% BI] | Standardized Effect Size | % of subsamples with higher values than random sampling |
|---|---|---|---|---|
| Regional variation coeff. (SD*100/Mean) | ||||
| Reference data | 92% | |||
| Random | 125% [98.5, 152%] | |||
| Distance (G) | 73.2% [55.1, 93.4%] | −53.7% [−83, −22.5%] | −2.88 | 0.23% |
| Distance (S) | 127% [101, 154%] | + 1.94% [−31, 35.3%] | 0.069 | 53.2% |
| Probability | 99.1% [75.2, 127%] | −25% [−58.8, 7.12%] | −1.26 | 10.7% |
| Uniform | 67.2% [56.1, 81.4%] | −58.4% [−86.1, − 31.1%] | −3.52 | 0% |
| Small hospitals (< 200 beds, %) | ||||
| Reference data | 71.1% | |||
| Random | 43.6% [32.7, 54.5%] | |||
| Distance (G) | 43.6% [34.5, 54.5%] | + 0% [− 12.7, 10.9%] | 0.026 | 46.5% |
| Distance (S) | 76.4% [72.7, 80%] | + 32.7% [21.8, 43.6%] | 4.97 | 100% |
| Probability | 30.9% [21.8, 43.6%] | −12.7% [−23.6, 0%] | −1.78 | 3.29% |
| Uniform | 34.5% [29.1, 41.8%] | −9.09% [−18.2, 1.82%] | − 1.32 | 7.19% |
| Medium hospitals (≥ 200, < 500 beds, %) | ||||
| Reference data | 19.5% | |||
| Random | 38.2% [27.3, 49.1%] | |||
| Distance (G) | 36.4% [27.3, 47.3%] | −1.82% [−12.7, 10.9%] | −0.21 | 36.6% |
| Distance (S) | 12.7% [9.09, 18.2%] | −25.5% [−36.4, − 14.5%] | −3.78 | 0% |
| Probability | 41.8% [30.9, 52.7%] | + 3.64% [−9.09, 14.5%] | 0.47 | 62.3% |
| Uniform | 38.2% [30.9, 45.5%] | + 0% [−10.9, 10.9%] | − 0.046 | 41.8% |
| Large hospitals (≥ 500 beds, %) | ||||
| Reference data | 9.35% | |||
| Random | 18.2% [9.09, 27.3%] | |||
| Distance (G) | 18.2% [12.7, 27.3%] | + 1.82% [−7.27, 10.9%] | 0.24 | 54% |
| Distance (S) | 10.9% [10.9, 10.9%] | −7.27% [−16.4, 1.82%] | − 1.36 | 5.35% |
| Probability | 27.3% [18.2, 38.2%] | + 9.09% [0, 18.2%] | 1.61 | 93.2% |
| Uniform | 27.3% [20, 32.7%] | + 9.09% [0, 16.4%] | 1.74 | 94% |
| Quality Score | ||||
| Random | 100 [79, 130] | |||
| Distance (G) | 98 [71, 130] | −6.5 [−37, 24] | −0.34 | 36.5% |
| Distance (S) | 92 [68, 120] | −12 [−42, 18] | − 0.68 | 24.3% |
| Probability | 68 [55, 84] | −35 [− 63, −10] | −2.21 | 0.9% |
| Uniform | 99 [80, 120] | −5.1 [−34, 22] | −0.32 | 38.1% |
| Prevalence (crude, %) | ||||
| Random | 7.44% [6.37, 8.58%] | |||
| Distance (G) | 7.23% [6.07, 8.43%] | −0.21% [−1.43, 1.02%] | − 0.3 | 37.3% |
| Distance (S) | 6.98% [5.75, 8.28%] | −0.47% [−1.91, 1.01%] | −0.51 | 29.7% |
| Probability | 7.87% [6.84, 8.88%] | + 0.45% [−0.63, 1.53%] | 0.69 | 62.7% |
| Uniform | 7.69% [6.79, 8.45%] | + 0.18% [−0.78, 1.29%] | 0.34 | 75.3% |
| Prevalence (hospital’s average, %) | ||||
| Random | 6.35% [5.44, 7.33%] | |||
| Distance (G) | 6.04% [5.07, 7.06%] | −0.31% [−1.34, 0.77%] | − 0.47 | 32.1% |
| Distance (S) | 5.86% [4.96, 6.8%] | −0.47% [−1.49, 0.61%] | −0.72 | 23.8% |
| Probability | 7.11% [6.26, 8.01%] | + 0.8% [−0.33, 1.81%] | 1.23 | 88.5% |
| Uniform | 6.43% [5.61, 7.41%] | + 0.1% [−0.9, 1.09%] | 0.16 | 56.5% |