in progress

Final Research Plan

Osteoporosis to Prevent Fractures: Screening

November 18, 2021

Recommendations made by the USPSTF are independent of the U.S. government. They should not be construed as an official position of the Agency for Healthcare Research and Quality or the U.S. Department of Health and Human Services.

This figure is an analytic framework depicting the key questions (KQs) within the context of the populations, interventions, comparisons, outcomes, time frames, and settings (PICOTS) relative to the effectiveness and harms of screening and treatment for osteoporosis. The figure illustrates the relationship between osteoporosis for adults in primary care settings without known osteoporosis or history of fragility fractures. KQ1 concerns the direct pathway of the relationship between screening and risk assessment and reduced fracture-related morbidity and mortality. On the indirect pathway, the population is shown as being at an increased risk or not at an increased risk based on screening with DXA at a central site or standardized clinical risk assessment tools, or a two-step process using both methods (KQ2). The harms of screening are assessed for the whole population undergoing screening (KQ3). If the population is not at increased risk, the pathway stops. For those at an increased risk, the benefits (KQ 4) and harms (KQ 5) of treatment are assessed relative to fracture-related morbidity and mortality outcomes.

Abbreviation: DXA=dual energy X-ray absorptiometry

1.  Does screening for fracture risk or osteoporosis reduce fractures and fracture-related morbidity and mortality in adults?
2a. What is the predictive accuracy of risk assessment tools for identifying adults who are at increased risk for hip fractures or major osteoporotic fractures?
2b. What is the predictive accuracy of bone mineral density testing with dual X-ray absorptiometry at central skeletal sites for identifying adults who are at increased risk for hip or major osteoporotic fractures?
2c. What is the diagnostic accuracy of risk assessment tools for identifying adults with osteoporosis?
2d. What is the evidence to determine screening intervals, and how do these intervals vary by baseline or current individual fracture risk?"
3. What are the harms of screening for fracture risk or osteoporosis?
4. What is the effectiveness of pharmacotherapy with selected Food and Drug Administration (FDA)–approved medications on fracture incidence and fracture-related morbidity and mortality?
5. What are the harms associated with selected FDA-approved medications?

Contextual questions will not be systematically reviewed and are not shown in the Analytic Framework.

  1. What is the evidence from modeling studies about the effectiveness of risk screening strategies that use different ages at which to start and stop screening and different screening intervals?
  2. How do various fracture risk assessment tools use race and ethnicity in fracture risk calculations?
  3. What is the incidence of fractures among persons of different races and ethnicities in the United States in the last 10 to 15 years, and what factors might explain differences in incidence among different races and ethnicities?
  4. What are the differences in rates of screening or treatment initiation among persons of different races and ethnicities, and what might explain these differences?
  5. What are the implications of using fixed fracture-risk thresholds for decisions regarding stepwise screening or treatment?
  6. What is the evidence for rare harms of bisphosphonate treatment (i.e., osteonecrosis of the jaw, atypical femur fractures) from observational studies that use noneligible control groups or are uncontrolled?
  7. What is the evidence for rebound fractures after discontinuation of denosumab?

The Research Approach identifies the study characteristics and criteria that the Evidence-based Practice Center will use to search for publications and to determine whether identified studies should be included or excluded from the evidence review. Criteria are overarching as well as specific to each of the key questions (KQs).

  Included Excluded
Population KQs 1–3 (Screening benefits, accuracy, harms): Adults age 40 years or older without known osteoporosis or history of fragility fractures

KQs 4, 5 (Treatment benefits and harms): Adults age 40 years or older with osteoporosis, osteopenia, or increased fracture risk (as defined by study authors)

Studies in which less than 50% of the enrolled population includes persons with conditions or medications listed as excluded will be included, and results will be stratified if possible.

Subpopulations of interest include men, women age 65 years or older, and postmenopausal women younger than 65 years.*

All KQs: Studies that exclusively enroll adults younger than age 40 years

KQs 1–3: Studies that exclusively enroll:

  • Adults with known osteoporosis or prior history of fragility fracture.
  • Adults with cancer, metabolic bone diseases, or medical conditions associated with bone loss, including but not limited to hyperparathyroidism, premature ovarian failure, hypogonadism, untreated hyperthyroidism, acromegaly, adrenal insufficiency, Cushing’s syndrome, celiac disease, inflammatory bowel disease, history of gastric bypass surgery, anorexia, chronic liver disease, multiple myeloma, chronic kidney disease, rheumatoid arthritis, lupus, multiple sclerosis, spinal cord injury.
  • Adults taking chronic medications associated with bone loss or strengthening, including glucocorticosteroids, select antiepileptic medications, hypogonadism-inducing agents (e.g., aromatase inhibitors, medroxyprogesterone acetate, gonadotropin-releasing hormone agonists), thiazolidinediones, calcineurin inhibitors, antiretroviral therapy, and testosterone.
KQs 4, 5: Studies that exclusively enroll or in which the majority of the population has:
  • Secondary osteoporosis because of an underlying medical condition or chronic use of a medication associated with bone loss or,
  • Prior fragility fracture.
In addition, studies that exclusively enroll participants who have failed prior medication use for osteoporosis are not eligible.
Screening Interventions KQs 1–3 (Screening benefits, accuracy, and harms):
  • FRA or ORA that has been evaluated in at least two independent cohorts external to the development cohort (unless males are included, then only one independent cohort external to the development cohort is required).
  • DXA measurement of BMD at the femoral neck (T-scores based on NHANES III reference range) or lumbar spine (local reference range).
  • A combination of FRA or ORA and DXA together or in sequence (e.g., two-step approach).


  • FRAs or ORAs that are not publicly available.
  • Studies of FRAs or ORAs using split sample validation.
  • Fall risk assessments (i.e., instruments validated to predict falls, not fractures).
  • FRAs or ORAs using risk factors not readily available or feasible within primary care settings.
  • Quantitative ultrasound.
  • Quantitative CT.
  • Magnetic resonance imaging.
  • Trabecular bone score.
  • Vertebral fracture assessment.
  • DXA measured at peripheral skeletal sites (e.g., radius, wrist, heel).
  • DXA measured at central skeletal sites, but hip T-scores based on local reference ranges.
  • Bone turnover biomarkers.
  • Finite element analysis.
  • Hip structural analysis.
  • Opportunistic screening for osteoporosis on images taken for other indications (e.g., dental X-rays, abdominal CT).
Screening Comparators KQs 1, 3 (Screening benefits and harms):
  • No screening.
  • FRA/ORA or BMD or both, but no results shared with patient or their primary care provider.

KQ 2 (Accuracy):

  • For predictive accuracy: Observed fracture incidence from nationally representative and verified sources.
  • For diagnostic accuracy: DXA-measured BMD at the femoral neck (T-scores based on NHANES III reference range) or lumbar spine.
KQs 1, 3:
  • No control group.
  • Another screening strategy (active comparator).

KQ 2: Any comparator not specifically identified as included.

Treatment Interventions KQs 4, 5 (Treatment benefits and harms):

Bisphosphonates with FDA-approved indications for the treatment of osteoporosis (i.e., alendronate, ibandronate, risedronate, zoledronic acid), denosumab

Males only: Teriparatide, abaloparatide, and romosozumab are also eligible
KQs 4, 5:
  • Bisphosphonates that do not have FDA-approved indications for the treatment of osteoporosis (e.g., etidronate, pamidronate).
  • Estrogen (with or without progesterone), raloxifene, or bazedoxifene.
  • Females only: Teriparatide, abaloparatide, or romosozumab.
  • Medications that are sometimes used off-label to treat osteoporosis (e.g., testosterone, tamoxifen).
  • Treatments that are no longer used in practice or that have been recalled, specifically calcitonin and parathyroid hormone 1-84.
  • Vitamin D or calcium supplements alone (these are considered adjuncts to treatment).
  • Dietary supplements.
  • Nonpharmacologic treatments (e.g., exercise, fall prevention interventions).
Treatment Comparators KQs 4, 5 (Treatment benefits and harms): Placebo, vitamin D or calcium or both, no treatment
  • Active drug comparators (e.g., head-to-head comparisons of active drugs or comparisons of multiple drugs in combination or in sequence with monotherapy).
  • Nonpharmacologic interventions (e.g., exercise).
Outcomes KQs 1, 4 (Screening and treatment benefits):
  • All-cause mortality.
  • Fracture-related mortality.
  • Fractures (all-cause, hip, major osteoporotic fractures, clinical vertebral fractures, any clinical fragility fractures).
  • Fracture-related morbidity (e.g., disability).

KQ 2 (Accuracy):

  • Predictive accuracy: Calibration outcomes (e.g., observed vs. expected ratio, calibration slope, calibration plot, Hosmer-Lemeshow goodness-of-fit, gradient of risk [risk ratio per standard deviation change in risk score]), overall prediction model performance (e.g., Brier score, explained variation [R2]).
  • Discrimination outcomes (e.g., c‑statistic, discrimination slope, sensitivity, specificity, area under the receiver operating characteristic curve).

KQ 3 (Screening harms):

  • Overdiagnosis.
  • Unnecessary treatment from inaccurate risk prediction.
  • Radiation exposure.
  • Anxiety from labeling.

KQ 5 (Treatment harms):

  • Total adverse events.
  • Total serious adverse events.
  • Specific serious adverse events: Major cardiovascular events (i.e., myocardial infarction, stroke, cardiovascular death), atrial fibrillation, osteonecrosis of the jaw, atypical femur fractures, incident gastrointestinal cancer, serious gastrointestinal events, rebound fractures after discontinuing denosumab treatment.
  • Discontinuations because of adverse events.
KQs 1, 4:
  • Radiographic (i.e., morphometric) vertebral fractures.
  • Fractures based on patient self-report without verification/confirmation.
  • BMD.
  • Other outcomes not specifically identified as included.

KQs 2, 3, 5:  Outcomes not specifically identified as included

Timing KQs 1, 4 (Screening and treatment benefits): Followup for at least 1 year

KQ 2 (Accuracy): 

  • For predictive accuracy of FRAs or ORAs, observed fracture incidence over at least a median or mean of 80% of the time specified by the FRA (e.g., at least 8 years for a tool designed to predict 10-year risk). For FRAs or ORAs that do not specify a prediction interval, a minimum of 3 years of observed incidence is required.
  • For predictive accuracy of DXA, observed fracture incidence over at least 1 year.
  • For diagnostic accuracy of risk assessments, no longer than 8 weeks between FRA or ORA and BMD measurement.
KQs 3, 5 (Screening and treatment harms): Any length of followup
Timing that does not meet inclusion criteria
Study Design KQs 1, 3 (Screening benefits and harms): RCTs, clinical controlled trials, or systematic reviews of RCTs or controlled trials. Cohort studies and systematic reviews of cohort studies are also eligible for KQ 3 only.

KQ 2 (Accuracy): Recent (published in the last 5 years) systematic reviews of cohort or test accuracy studies, cohort studies designed for evaluating predictive accuracy (i.e., prognosis for fracture risk) or diagnostic accuracy (for identification of osteoporosis), comparative studies in which a single group is treated as a cohort for purposes of evaluating predictive or diagnostic accuracy are also eligible.

KQs 4, 5 (Treatment benefits and harms): RCTs and controlled trials (including those in which participants serve as their own controls); controlled cohort studies are also eligible for KQ 5 only
All KQs: Case series; case reports; case-control studies; conference abstracts, posters, or proceedings without data or information available to assess risk of bias; unpublished data; editorials; commentaries; narrative reviews

KQs 4, and 5: Systematic reviews are not eligible but will be hand searched to identify studies potentially missed by our search

Settings KQs 1, 3, 4, 5 (Screening and treatment benefits and harms): Primary care settings in countries designated as “very high” on the 2020 Human Development Index (as defined by the United Nations Development Programme)

KQ 2 (Accuracy): Predictive accuracy: United States or countries with similar hip fracture incidence as the United States§ for synthesis of any primary research studies

KQs 1, 3, 4, 5: Long-term care settings such as nursing homes, inpatient settings

KQs 1, 3: Specialty medical settings (e.g., endocrinology, rheumatology)

KQ 2: Predictive accuracy: studies in single countries with high or low fracture incidence
Study Quality All KQs: Good or fair quality as determined by standard risk of bias instruments and existing USPSTF criteria tailored to study design All KQs: Poor quality

* For the purposes of this review, we will use the terms men and women consistent with how they are typically used in the underlying evidence base for this topic. Men refers to persons assigned male sex at birth. Women refers to persons assigned female sex at birth. Studies that include gender-diverse individuals, including those who have undergone gender-affirming therapy (e.g., transmen, transwomen), are not excluded from the scope of this review. However, studies that exclusively enrolled populations who take hormone therapy that affects bone density are excluded from this review, consistent with our criteria that exclude studies that focus on populations with secondary osteoporosis or who take chronic medications that have known effects on bone metabolism. For such populations, individualized clinical decisions about bone density testing in the context of condition and medication management are required.
† This review is not intended as a comprehensive review of all available pharmacologic therapies. Second-line therapies (abaloparatide, teriparatide, romosozumab) are excluded for women because the USPSTF is likely to have sufficient evidence to determine the net benefit of treatment based on the evidence for FDA-approved bisphosphonates and denosumab, as determined by the most recent review before this update. We will only consider these drugs for men given the paucity of treatment studies generally available for men. Although romosozumab is not currently FDA-approved for men, it is currently in Phase 3 studies for men, so it will be included in this update. Hormone therapy and selective estrogen receptor modulators are reviewed in a separate USPSTF review on hormone therapy, so they will not be included in this update.
‡ Major osteoporotic fracture is typically defined as fractures of the hip, wrist, and humerus and clinical vertebral fractures.
§ Countries with “moderate” hip fracture incidence in addition to the United States include Finland, France, Canada, New Zealand, Lithuania, Malaysia, South Korea, Portugal, Japan, Israel, Australia, Russia, The Netherlands, Kuwait, Spain, Mexico, Estonia, Poland, Chile, and Thailand.1

Abbreviations: BMD=bone mineral density; CT=computerized tomography; DXA=dual-energy X-ray absorptiometry; FDA=Food and Drug Administration; FRA=fracture risk assessment; KQ=key question; NHANES=National Health and Nutrition Examination Survey; ORA=osteoporosis risk assessment; RCT=randomized, controlled trial; USPSTF=U.S. Preventive Services Task Force.

The draft Research Plan was posted for public comment on the USPSTF website from August 12, 2021, to September 8, 2021. In response to comments, the USPSTF included additional outcomes, added two contextual questions regarding rare but serious harms, and specified subpopulations of interest. The USPSTF also made several minor additions and wording changes to improve the clarity and specificity of the research approach.

  1. Kanis JA, Oden A, McCloskey EV, et al. A systematic review of hip fracture incidence and probability of fracture worldwide. Osteoporos Int. 2012;23(9):2239-2256.