Skip to main content


Table 2 Key question 2 (accuracy of screening tests) study eligibility criteria

From: Screening to prevent fragility fractures among adults 40 years and older in primary care: protocol for a systematic review

PopulationAsymptomatic adults ≥ 40 years in the general population (we will include studies where ≥ 80% of the sample or the sample mean age − 1 standard deviation is ≥ 40 years)
Subgroups for decision-making: age, sex, menopausal status
Methods subgroups: treatment with anti-osteoporosis drugs, baseline predicted fracture risk, length of follow-up
 ■ Adults < 40 years
 ■ > 50% with prior diagnosis of osteoporosis, prior fragility fracture, endocrine or other disorders likely to be related to metabolic bone disease, chronic use of glucocorticoid medications, cancer
InterventionScreening tool to prevent fragility fracture using any of the following approaches:
 ■ Fracture risk assessment alone
 ■ Bone mineral density (BMD) alone by dual-energy x-ray absorptiometry of the femoral neck or lumbar spine (DXA) ± vertebral fracture assessment/spinal radiography
 ■ Fracture risk assessment followed by/incorporating BMD (DXA) ± vertebral fracture assessment/spinal radiography
When a risk assessment tool is used it must have been externally validated to predict fragility fractures in a population within a very high human development index country [119] with a fracture rate similar to Canada (i.e., moderate) [57]
 ■ Risk assessment tools not externally validated (different population than derivation cohort) to predict fragility fractures.
 ■ Validation studies conducted in a population within a country that does not have a very high human development index, and/or has a different fracture rate to Canada
 ■ Other BMD or osteoporosis-related screening tests (e.g., quantitative ultrasound, quantitative computed tomography, peripheral DXA, trabecular bone score, bone turnover markers)
OutcomesCalibration (total/average and by differing estimated risks, e.g., expected vs. observed fractures, goodness-of-fit, calibration slope) for 5- and 10-year fracture risk of:
 ■ Hip fractures
 ■ All clinical fragility fractures
Note: Discrimination (e.g., sensitivity, specificity, area under the receiver operating characteristics curve/c-statistic, positive predictive value, negative predictive value) will not be included, but will be reported as it has been previously in the 2018 USPSTF systematic review as evidence to support contextual and implementation aspects of the guideline. When BMD is used alone, only discrimination outcomes will be reported since calibration is not relevant.
TimingAny length of follow-up; to make predictions for 5- or 10-year fractureNot applicable
Setting ■ Primary health care [117]
 ■ Very high human development index country [119] with a fracture rate similar to Canada (i.e., moderate) [57]
 ■ Long-term care facilities
 ■ Countries that are not very high human development index and/or have a different fracture rate than Canada
Study design and publication status ■ Prospective or retrospective cohort studies with a defined index screen (assessed before fracture measurement); may be randomized comparisons between different index tests, but all patients assessed for fracture and each arm treated separately
 ■ Manuscripts, reports, abstracts, dissertations, and trial registers with data available
 ■ Systematic reviews, meta-analyses, and pooled analyses
 ■ All other primary study designs
 ■ Non-research (e.g., editorials)
 ■ Studies available only as conference proceedings or other grey literature, unless data are available and adequate to assess risk of bias
LanguageEnglish or FrenchAll other languages
Date of publicationAnyNot applicable
  1. *Note that studies of tools (that incorporate mortality in their risk algorithms) that do not consider death hazards in their observed fracture rate will be included but may contribute to downgrading the certainty in the evidence.