Shorter hospital stay may worsen quality of treatment

Figure 1

Risk of death and live discharge by calendar year, relative to 2004 (dashed line). Year-specific odds ratios and 95% confidence intervals were adjusted for age (85+ vs < 85), sex, preadmission residence (residence vs home), comorbidity (at least 1 acute admission for heart failure, chronic obstructive pulmonary disease, ischemic heart disease, hypertension, or diabetes in the year prior to admission for hip fracture), fracture type (transcervical vs other), procedure type (fixation vs implant), weekday and time of admission, hospital type and size (teaching, large, medium, or small community), and hospital volume of hip fracture surgery (higher vs lower), demand at index admission (number of admissions in the week of initial hospitalization, larger vs smaller than hospital weekly capacity), time to surgery (2 days or more days vs less), and province of admission. Trend test for odds ratios of death z = −3.9, P < .001. Trend test for odds ratios of discharge z = 5.5, P < .001.