Addressing Vertebral Osteoporosis Incidentally Detected to Prevent Future Fractures

Objective: To improve the quality of care for elderly patients with osteoporosis and vertebral fractures incidentally detected by a chest radiograph in the Emergency Department.

Background: Osteoporosis leads to decreased bone mass, skeletal fragility, and fractures. Osteoporosis affects at least 1.4 million Canadians, 25% of women and 12% of men >60 years of age. The most common osteoporotic fractures are vertebral (spinal), two-thirds of which are asymptomatic. Irrespective of symptoms, these "silent" fractures cause disability, deformity, and death. Compared to those with normal bones and no fracture, the patient with osteoporosis and vertebral fracture has a 20-fold increased risk of future fracture. Guidelines recommend aggressive treatment because rate of re-fracture is 20% within 1 year and treatment reduces risk by 40-50%. Guidelines do not, however, recommend population screening for vertebral fractures. We reported (Arch Intern Med, 2005) that chest radiographs performed in the Emergency Department detected "incidental" fractures in 16% of elderly patients, although only 25% of patients were treated for osteoporosis. The care gap between evidence-based best practice and usual care needs to be urgently addressed.

Hypothesis: The proposed intervention will improve rates of diagnosis and treatment of osteoporosis in high-risk patients. The intervention will be directed at physicians (patient-specific reminders and opinion leader generated 1-page guidelines) with or without patient-activation (leaflets and telephone-based counseling).

Specific Aims: To determine whether exposure to a physician-directed quality improvement intervention can- Aim #1. Increase rates of osteoporosis treatment in patients with newly recognized vertebral fracture(s); Aim #2. Increase patients' self-reported diagnosis of osteoporosis and other related knowledge; Aim #3. Be further augmented by the addition of patient-activation.

Study Design: A single-center prospective nonrandomized controlled trial with blinded (allocation-concealed) ascertainment of outcomes comparing the proposed physician-intervention to usual care controls. We define usual care as notification of the patients' physician of chest radiograph results. Allocation will be on the basis of an alternate-week time series design, with the proposed intervention "on" for all patients for one week, followed by the intervention being "off" (i.e., usual care) the next week. After 3 months, controls will be offered the physician-intervention plus patient-activation, with outcomes recollected 3 months later. Eligible patients will be >60 years of age who present to the Emergency Department: (1) with a chest radiograph reporting the presence of a vertebral fracture, (2) are not taking prescription osteoporosis treatment, and (3) are discharged home. The primary outcome is the proportion of patients starting prescription osteoporosis treatment within 3 months of fracture recognition; secondary outcomes include BMD testing and diagnosis of osteoporosis. The intervention is expected to increase the primary outcome by at least 20% (absolute) over usual care rates of 10%. With alpha=0.05, beta=0.80, 20% losses-to-followup, and additional power to address secondary outcomes and effects of patient-activation, the required sample size is 240 patients.

Pilot Data and Feasibility: Our pilot study (AJR, 2004) found institutional true-positive reporting rates for vertebral fractures were 60%, with 100% specificity, compared with reference standard. A 1-year feasibility study of a random sample of ~5000 radiographs was then undertaken (Arch Intern Med, 2005). Based on these data, we expect to enroll 3-4 patients/week, complete recruitment in ~20 months, ascertain outcomes by 30 months, and finish the study in 3 years. The research team has generated pilot data, systematically reviewed pertinent literature, recruited opinion leaders to create guidelines, trained research staff, refined methods for outcomes ascertainment; and conducted related studies of wrist (Annals Intern Med, 2004) and hip fractures.

Significance: To our knowledge, there are no studies examining strategies to improve care for this vulnerable and high-risk population, and we believe our results will be important regardless of findings. If "negative" it will provide evidence that more costly and laborious interventions will be needed to overcome clinical inertia; if positive, however, our intervention should be widely applicable. In addition, positive results can be easily extended to other related populations (e.g., routine chest radiographs done for admission to nursing homes) and can be adapted for implementation via other modalities (e.g., computerized decision support, telemedicine).