Table 3.
The Association between Hospice Concentration and the Occurrence of Disagreement in Pain Rating on Nonhospice Residents between NH and Study Nurses: Severe Pain versus Nonsevere Pain (Cutoff=4)
| False Positive (FP) | False Negative (FN) | |
|---|---|---|
| Null models with random intercepts | ||
| Unexplained variance at facility level | ||
| Variance | 0.87 (0.17)* | 0.88 (0.18) |
| Covariance | −0.12 (0.12) | |
| Intraclass correlation | 0.19 | 0.19 |
| Full models with random intercepts | ||
| Hospice concentration (%) | ||
| 0−3 | —† | — |
| 3–5 | 0.39 (0.21–0.73) | 0.54 (0.29–1.00) |
| 5+ | 0.56 (0.33–0.95) | 1.00 (0.61–1.65) |
| Age (centered) | ||
| Linear term | 0.94 (0.78–1.13) | 1.03 (0.85–1.25) |
| Quadratic term | 0.99 (0.92–1.07) | 1.00 (0.91–1.10) |
| Cognitive impairment | ||
| No or mild (CPS=0 or 1) | — | — |
| Moderate (CPS=2 or 3) | 0.57 (0.42–0.79) | 0.80 (0.57–1.13) |
| Severe (CPS=4, 5, or 6) | 0.34 (0.21–0.55) | 0.72 (0.47–1.11) |
| Having cancer | 1.04 (0.69–1.57) | 1.55 (1.02–2.35) |
| Having dementia | 1.02 (0.68–1.52) | 1.06 (0.73–1.54) |
| Male | 0.75 (0.57–1.00) | 0.85 (0.62–1.15) |
| Nurse hours/resident | 1.04 (0.86–1.26) | 1.00 (0.81–1.23) |
| Occupancy rate <80% | 1.50 (1.01–2.25) | 0.75 (0.48–1.18) |
| Part of a chain | 1.00 (0.71–1.41) | 1.29 (0.90–1.86) |
| Having 100+ beds | 0.81 (0.53–1.25) | 0.66 (0.43–1.02) |
| Number of health deficiencies >50th percentile in the state | 0.96 (0.65–1.43) | 0.99 (0.66–1.49) |
| Location of the study nursing home | ||
| California | 2.07 (1.11–3.86) | 0.52 (0.24–1.11) |
| Illinois | 1.25 (0.66–2.36) | 1.40 (0.74–2.65) |
| Missouri | 1.76 (0.79–3.93) | 4.07 (1.99–8.30) |
| Ohio | 2.37 (1.25–4.49) | 1.21 (0.62–2.39) |
| Pennsylvania | 1.39 (0.75–2.58) | 1.41 (0.77–2.58) |
| Tennessee | — | — |
| Unexplained variance at facility level | ||
| Variance | 0.41 (0.12) | 0.44 (0.14) |
| Covariance | −0.07 (0.10) | |
*
Standard error of variance at the level of nursing home.
†
Referent group.