Is it time to consider population screening for fracture risk in postmenopausal women? A position paper from the International Osteoporosis Foundation Epidemiology/Quality of Life Working Group

Osteoporotic fractures are undoubtedly a common public health problem, particularly in ageing societies in terms of patient’s health, quality of life and social care costs [9‐12]. At the age of 50 years, the remaining lifetime probability of at least one of the major osteoporotic fractures is 22% in men and 46% in women. In 2000, it was estimated that there were approximately 9 million fractures annually worldwide, with over one-third of all osteoporotic fractures occurring in Europe [13]. The latter accounted for 2 million disability-adjusted life years (DALYs) annually in Europe, a burden that exceeded that of hypertensive heart disease or rheumatoid arthritis. The number and burden of osteoporotic fractures is rising in many countries, partly related to the increased longevity of the populations. The age- and sex-specific incidence of fracture has also increased in some but not all countries [14, 15].

In 2019, 4.3 million new fragility fractures were estimated to have occurred in the EU, comprising approximately 827,000 hip fractures, 663,000 clinical vertebral (spine) fractures, 637,000 forearm fractures and 2,150,000 fractures at other sites (i.e. pelvis, rib, humerus, tibia, fibula, clavicle, scapula, sternum and other femoral fractures) [11]. About two-thirds of all incident fractures occurred in women. In 2019, the number of deaths causally related to fractures was estimated to be 248,487, with about half of these attributable to hip fractures. The number of fracture-related deaths was comparable or exceeded those from some of the most common causes of death such as lung cancer, diabetes and chronic lower respiratory diseases. In 2000, it was estimated that osteoporotic fractures in Europe accounted for more DALYs (2,006,000) than rheumatoid arthritis (1,048,000), but less than for osteoarthritis (3,088,000) [13]. In 2016, the estimate placed the total DALY loss related to fragility fractures at more than 2.6 million, which was higher than that estimated for chronic obstructive pulmonary disease and ischaemic stroke, but lower than that for lung cancer or dementia [12].

In 2019, the cost of osteoporosis, including pharmacological intervention in the EU was estimated at €56.9 billion (1000 million), with two-thirds derived from the treatment of fractures and only 3% representing the costs of pharmacological intervention. Excluding the latter cost, hip fractures represented 54% of the costs. The total cost including values of quality adjusted life years (QALYs) lost was estimated at €113 billion, a figure that is expected to rise to €121 billion in 2025. The cost of osteoporotic fractures accounts for approximately 3.5% of healthcare spending (i.e. €55.3 billion in 2019) indicating a very substantial impact of fragility fractures on the present healthcare budgets of European countries.

The epidemiology of osteoporosis is well characterized and described in detail elsewhere [14, 16‐19]. The incidence of fragility fractures increases markedly with age, and the vast majority of osteoporotic fractures occur in older women. Compared with other fractures, a great deal of information is available on the epidemiology of hip fracture as nearly all such patients are admitted to hospital with data available through surgical and discharge records. In patients with fractures at other skeletal sites, only a minority are admitted, though they may attend hospital on an outpatient basis.

The occurrence of hip fracture, widely regarded as the most serious osteoporotic fracture, reflects decreasing bone strength with increasing age, with a concomitant increase in falls risk [20]. The risk of falling is somewhat higher in elderly women than in elderly men with about one-third of elderly individuals falling annually [21]. Hip fractures are relatively rare at the age of 50 years but become the predominant fracture from the age of 75 years [22]. Patients with hip fracture often have significant co-morbidities, and up to 20% of patients die in the first year following hip fracture, mostly as a result of these underlying medical conditions [23, 24]; nonetheless, it is estimated that approximately 30% of deaths are causally related to the fracture event [25].

In summary, there is ample evidence that osteoporosis and, in particular, osteoporosis-related fractures are a significant and growing health problem for individuals and healthcare providers. Fractures are highly prevalent and are associated with significant morbidity and mortality.

As a clinical endpoint, osteoporotic fracture has multiple factors contributing to the risk of fracture that can enable identification of high risk prior to the event occurring. In many ways, this is a different latency to that identified under established screening programmes in diseases such as breast and cervical cancer or diabetic eye disease, where the focus is on early detection of the disease itself. At a simplistic level, it might be assumed that low or decreasing bone mass is the only latent stage of note for osteoporotic fracture but, as described below, this excludes a wealth of information on other factors that contribute to fracture risk. This awareness has led to the development of risk calculators that take account of several risk indicators as used in the management of other diseases such as cardiovascular disease [32‐34].

Examples of ‘non-BMD’ factors that contribute to fracture risk include ethnicity, age, sex, body mass index (BMI) [35], a prior fracture [36], a family history of fracture [37], long-term use of systemic glucocorticoids [38], a history of rheumatoid arthritis [39], diabetes [40‐42], other causes of secondary osteoporosis [43] and lifestyle risk factors such as physical inactivity [44], falls [45, 46], smoking [47] and alcohol consumption[48]. It is important to note that some of these risk factors are partially or wholly independent of BMD, so that their use with BMD can enhance the information provided by BMD alone. Conversely, some strong BMD-dependent risk factors, for example BMI, can be used for fracture risk assessment in the absence of access to BMD tests. This is discussed in more detail in the section on the test.

In summary, the epidemiology and natural history of osteoporosis, including the risk factors that contribute to the clinical endpoint of fracture, are well documented and understood. The recognition of identifiable risk factors and the quantitation of that risk are now possible allowing preventative interventions to be targeted to those at highest fracture risk.

Thus, whereas patients will continue to receive general advice in the absence of cost-effective primary prevention interventions, there remains a need to implement other proven, cost-effective strategies and interventions.

In brief, the FRAX® tool (https://​www.​sheffield.​ac.​uk/​FRAX/​index.​aspx) is a computer-based algorithm developed by researchers at the then WHO Collaborating Centre for Metabolic Bone Diseases at the University of Sheffield [43, 70, 71]. The calculation incorporates information gathered from easily obtained clinical risk factors (Fig. 1) to estimate the probability of sustaining a fracture in the next 10 years. Femoral neck BMD can also be added to the calculation to enhance fracture risk prediction. It has two outputs, namely the probability of sustaining a hip fracture in the next 10 years, or a major osteoporotic fracture. The outputs are calibrated to age and sex-specific rates of fracture and mortality in a substantial number of countries/territories (currently 85 models covering 77 countries).

FRAX has been widely used for the assessment of fracture risk since the launch of the FRAX website in 2008. Following regulatory review by the US Food and Drug Administration, FRAX was incorporated into DXA scanners to provide FRAX probabilities at the time of DXA scanning. For those without internet access, hand-held calculators and an application for Apple smartphones have been developed. A paper-based FRAX pad in several languages allows patients to document risk variables prior to a visit with healthcare professionals.

It is also worthy of note that the availability and access to densitometry in many countries is low [72‐75], so that a major advantage of FRAX is the ability to assess fracture risk where BMD is unavailable. In the UK, National Institute for Health and Care Excellence (NICE) has recommended that FRAX (or QFracture) should be used to estimate 10-year predicted absolute fracture risk prior to deciding on the need for DXA measured BMD [76]. As a completely non-invasive test, the tool is safe for users.

The characteristic of major importance, for the purpose of risk assessment, is the ability of a tool to predict the occurrence of new fractures, more usefully expressed as the increase in relative risk per standard deviation (SD) unit increase in risk score, termed the gradient of risk. The gradient of risk for the use of the FRAX clinical risk factors alone, femoral neck BMD alone, and the combination is shown in Table 2 [77]. It is relevant that the performance of FRAX is enhanced by the use of BMD tests; FRAX without BMD has a predictive value for fractures that is comparable to the use of BMD alone. Overall, the predictive value compares very favourably with other risk engines such as the Gail score for breast cancer [78]. The choice of screening modality to identify high hip fracture risk can be adapted to local resources; for example, where BMD measurements are readily available, it can play a predominant role in screening, while screening with FRAX risk factors alone can readily be applied in the absence of access to BMD.

The performance characteristics of the FRAX clinical risk factors, with and without BMD, have been validated in eleven independent population-based cohorts [77]. Notably, the gradients of risk and age-adjusted area under the curves (AUC) from receiver-operated characteristic (ROC) analysis were comparable in the validation cohorts compared with the original cohorts. In contrast to separate analyses of sensitivity and specificity, the AUC provides an index of overall test accuracy if sensitivity and specificity have equal weights. A test with no predictive (or discriminative) ability would have an AUC of 0.5 (or 50%), while a perfect predictive test would have an AUC of 1.0 (or 100%). For hip fracture prediction without BMD, the mean AUC at age 70 years was 0.66 in the validation cohorts compared with 0.67 in the original cohorts. With the addition of BMD, the mean AUC was 0.74 and 0.78, respectively [77]. Since that original validation, a number of independent reviews, including four systematic reviews, have included detailed descriptions of the performance of the FRAX tool [76, 79‐82]. Bearing in mind the flaws that can arise from cross-study and within-study comparisons of the AUC [83, 84], the reviews have confirmed a significant predictive ability of FRAX for future fractures, especially of the hip. For example, in the most recent of the systematic reviews, conducted for the US Preventive Services Task Force, the AUCs for hip fracture prediction with FRAX ranged from 0.76, in 12 studies comprising just under 200,000 women where FRAX was calculated without BMD, to 0.79 in 10 studies comprising approximately 162,000 women when FRAX was calculated with BMD [79]. In a recent independent comparative study in a single population sample, the predictive value for hip fractures was better for the FRAX tool (AUC: 0.841, 95% CI 0.795–0.887) than for the Garvan fracture risk tool (AUC: 0.769, 95% CI 0.702–0.836, p = 0.01) [85].

In summary, the FRAX tool fulfils the criteria of being a simple, safe, precise and well-validated test. While the AUCs for future hip fracture prediction are perhaps lower than existing screening tests that detect existing or early disease, they suggest excellent performance suitable for application in clinical practice.

The distribution of test values in the target population should be known and a suitable cut-off level defined and agreed.

Screening programmes are usually designed and adopted at regional or national levels. The same is true for the development and setting of intervention thresholds; this necessarily remains a more localised remit, given that each health care system will consider local/national factors such as reimbursement issues, health economic assessment, willingness to pay for health care in osteoporosis and access to DXA. The use of health economic analyses to derive intervention thresholds is fraught with problems, the most common of which is that they are time limited due to changes, usually reductions, in the costs of treatment. In the presence of very inexpensive medications, absurd situations can arise with a recent example from the UK where, following a multiple technology appraisal (MTA) on bisphosphonate use in osteoporosis, NICE recommended that treatment with oral bisphosphonates may be instituted in those with a 10-year probability of major osteoporotic fracture of 1% or more [86]. If interpreted as an intervention threshold, virtually all women aged ≥ 65 and men ≥ 75 years would be recommended treatment [87]. Shortly thereafter, NICE endorsed the assessment and intervention thresholds proposed within the NOGG guidance [50, 88].

A systematic review in 2016 identified assessment guidelines for osteoporosis that incorporated FRAX, utilising either age-independent (i.e. fixed) or age-dependent thresholds for further assessment and/or treatment [67]. In most guidelines, treatment for osteoporosis is recommended in individuals with prior fragility fractures, especially fractures at the spine or hip, with the FRAX thresholds reserved specifically for use in those without such fractures. The majority of guidelines had utilised a fixed intervention threshold, frequently as a component of more complex guidance (e.g. BMD thresholds). A substantial proportion (about 25%) had adopted an age-dependent threshold, with a small minority proposing a combination of age-dependent and fixed thresholds (so called hybrid thresholds).

Age-dependent intervention thresholds, first developed by the National Osteoporosis Guideline Group (NOGG), are based on the rationale that if a woman with a prior fragility fracture is eligible for treatment, then, at any given age, a man or woman with the same fracture probability (i.e. at the ‘fracture threshold’) should also be eligible, even in the absence of fracture [50, 89, 90]. By design, this fracture threshold increases with age with a plateau or decline at older ages due to the competing hazard of death. Age-dependent thresholds have since been adopted into European guidelines [51, 91] and elsewhere [67, 92, 93]. The same intervention threshold is applied to men, since the effectiveness and cost-effectiveness of interventions in men are broadly similar to that in women for equivalent risk [94]. In the UK, it was noted that the age-dependent probability thresholds introduced inequalities in access to therapy at older ages (≥ 70 years) for women without a prior fracture, leading to the development of a hybrid model which reduced the disparity [69]. For the proposed screening strategy, the approach would be to adopt the current NOGG thresholds in the UK. The thresholds at age 70 years for major osteoporotic fracture and hip fracture in the UK and some other high-risk countries, if using a similar approach, are shown in Table 3. Other countries that have an average 10-year hip fracture probability threshold of 5% or more include Austria, Denmark, Germany, Greece, Iceland, Iran, Ireland, Israel, Italy, Kazakhstan, Kyrgystan, Malta, Moldova, Norway, Singapore, Slovakia, South Korea, Sweden, Switzerland, Taiwan and Uzbekistan.

In summary, there are a number of approaches to defining cut-off levels for the FRAX tool as intervention thresholds within a screening program. Of these, the probability thresholds at age 70 in women with a prior fracture seem intuitive, as all current guidelines would recommend treatment in such individuals.

The data to date suggest that assessment of fracture risk by the FRAX tool, with subsequent BMD measurement where indicated, is acceptable to the populations in which it has been used.

In clinical settings, this assessment of BMD, usually by DXA measurements at the lumbar spine and hip, is used not just for risk prediction, but also for diagnosis, selection of patients for treatment and monitoring of patients on treatment. A requirement has been estimated of approximately 11 DXA units per million of the general population to permit implementation of practice guidelines [72], a level which is only achieved in about 60% of countries in the EU [11, 100]. Strategies that target the use of DXA to those with fracture risks at or near an intervention threshold reduce the need for DXA while maintaining the identification of those at high risk. For example, a comparison of the NOGG strategy, comprising risk assessment followed by targeted DXA, with the previous guidance issued by the Royal College of Physicians (RCP) where the presence of a risk factor simply mandated the use of DXA, NOGG identified a similar proportion of women at high risk (average 34.6% vs. 35.7% across all ages), but with lower numbers of scans required at each age [68]. Thus, the NOGG approach required only 3.5 scans at the age of 50 years to identify one case of hip fracture, whereas the previous RCP approach required 13.9. At 75 years, the corresponding numbers were 0.9 and 1.5. Compared to the older strategy, the FRAX-based NOGG strategy used BMD resources more efficiently with lower acquisition costs and lower costs per hip fracture averted [68].

Most clinical guidelines agreement the need for additional radiological, haematological, biochemical and immunological assessments in patients at high risk of fracture or with recently diagnosed osteoporosis [50, 51, 53]. Broadly, these seek to identify or exclude any undetected underlying disease that might contribute further to fracture risk and its management (e.g. identify a reversible or treatable cause such as primary hyperparathyroidism or coeliac disease), and to identify any contraindications to a particular treatment (e.g. renal impairment in the case of bisphosphonates or hypocalcaemia in the use of potent bisphosphonates or denosumab).

In summary, there is much agreement across clinical guidelines on the further diagnostic investigation of individuals at increased risk of fracture and on the treatment choices available to such individuals (see next section).

In contrast, many randomised, placebo-controlled trials have shown that treatments directed at preventing bone loss and/or improving bone mass can significantly reduce the incidence of fracture [108] at vertebral (relative risk, RR 0.4–0.6) and non-vertebral sites (RR 0.6–0.8), including reducing the incidence of hip fracture by 30–45% [109‐111]. Classes of drugs demonstrating efficacy in the prevention of fragility fracture include bisphosphonates (e.g. alendronate, ibandronate, risedronate and zoledronate), selective oestrogen receptor modulators [SERM] (e.g. raloxifene, bazedoxifene), parathyroid hormone peptides and derivatives (e.g. teriparatide and abaloparatide), menopausal hormone therapy (MHT) and humanized monoclonal antibodies (e.g. denosumab and romosozumab). Recent comparative clinical trials have provided evidence of enhanced anti-fracture efficacy of anabolic compared with antiresorptive therapies [112‐114], prompting considerations of starting treatment with an anabolic agent in patients at very high risk of fracture as a more appropriate means of rapidly reducing fracture risk [115‐117].

The availability of effective therapies has long given rise to considerations of approaches for fracture prevention; for example, almost 20 years ago, a simulation model investigated the cost-effectiveness of a 5-year treatment to prevent hip fracture in women at average risk [118]. Major determinants of cost-effectiveness included the age of the individual and the cost of treatment. Assuming an efficacy of a 35% reduction in relative risk, it was cost-effective to treat all women above the age of 81 years with an intervention cost of $650/year. For cheaper treatments, e.g. $250, it was cost-effective to give treatments at the age of 70 years, and for the cheapest treatment to treat nearly all postmenopausal women. The authors suggested that controlled prospective studies in the apparently healthy population were worthy of consideration, particularly in the elderly. Since then, several randomised, controlled studies have examined the use of osteoporosis interventions in older populations, unselected for osteoporosis. In the Women’s Health Initiative study, the use of MHT was examined in two studies, one in women with an intact uterus and one in women with a history of hysterectomy [119, 120]. The effects on fracture risk were similar between the two studies. In women age 50–79 years with an intact uterus, the use of conjugated equine oestrogen and medroxyprogesterone acetate over a mean of 5.2 years follow-up reduced the incidence of hip fracture by 34% (hazard ratio, HR: 0.66; 95% CI: 0.45–0.98], with a 23% reduction in other osteoporotic fractures (HR: 0.77; 95% CI: 0.69–0.86) [119]. Importantly, this effect was independent of baseline BMD [121] suggesting that treatment efficacy was not dependent on the presence of osteoporosis. A similar BMD-independent effect on fracture risk reduction was observed in a 3-year randomised, placebo-controlled study of the oral bisphosphonate, clodronate, in women age 75 years and older, again unselected for osteoporosis [122]. This study was underpowered for the outcome of hip fractures, but treatment was associated with a 23% reduction in osteoporotic fractures. A post hoc analysis demonstrated that, while the effect was independent of BMD, treatment was more effective in women at higher baseline fracture risk assessed by the FRAX tool [123]. Thus, at a probability of 15% (25th percentile), the relative risk for fracture was reduced by 8% (RR 0.92, 0.69–1.24), whereas at probabilities of 24% and 30% (the 75th and 90th percentiles), the reductions were 27% (RR 0.73, 0.58–0.92) and 38% (HR 0.62, 0.46–0.84) respectively. Several other studies of a variety of osteoporosis therapies have also shown no interaction between treatment efficacy and underlying BMD [124] supporting the use of treatment in those with high fracture risk even before a BMD in the osteoporosis range has been reached.

The success of such population-based approaches does not necessarily imply that risk assessments to target therapies are not worthwhile, since it is likely to be even more cost-effective to target individuals at higher-than-average risk. Certainly, the more expensive the treatment and the cheaper the risk assessment, the stronger the case for selecting high-risk segments of the population. The age at which intervention might be offered depends critically upon the costs of treatment, efficacy and the average risk of hip fracture, as well as additional skeletal and extraskeletal effects. The aforementioned cost-effectiveness analysis in 2001 suggested that strategies aimed at intervention over the age of 70 years required testing [118].

In summary, there is an extensive clinical and economic evidence base for effective bone-targeted treatments that can reduce fracture risk. Unselected population-based studies suggest that fracture risk can be also significantly reduced with such treatments; targeted intervention of relatively inexpensive, safe treatments to those within these populations that have identifiable increased risk will intuitively lead to better outcomes than treatment delayed until after the subsequent occurrence of fracture.

Cost-effectiveness is accepted in many countries as a primary driver of decisions about who should be offered treatment with particular medications. Such approaches arose from the perceived need to limit access to expensive treatments across a range of diseases. Within osteoporosis, for example, health technology assessments of available treatments in the UK by the NICE have traditionally stratified the risk levels at which treatments could be initiated [125‐129]. This stratification was usually based on age, BMD T score and the presence of one or more individual clinical risk factors; the cost-effectiveness thresholds were subsequently assumed by the clinical community to be intervention thresholds. More recently, NICE appraisals have embraced the use of absolute fracture risk, but the continued application of cost-effectiveness to developing thresholds for now inexpensive drugs led to counterintuitive and potentially harmful guidance [130].

In summary, many countries have access to evidence-based policies and guidelines covering which individuals should be offered treatment and the appropriate treatment to be offered.

A standardized approach to post hip fracture care was first published in a collaboration between the British Orthopaedic Association and British Geriatric Society in 2007 [131], followed soon after by the establishment of the National Hip Fracture Database to audit hospital performance. The database is centrally funded via government but is run independently. Importantly, the programme was subsequently supported by the creation in 2010 of a Best Practice Tariff (BPT) to financially incentivize improved care. Subsequently, NICE incorporated much of this acute hip fracture management initiative into their own guideline in 2011 [132]. The guidelines state that patients should be managed with combined orthopaedic and orthogeriatric care in a hip fracture programme with all patients preferably being seen preoperatively; due to difficulties in accessing orthogeriatric services at weekends, the target is that all patients are reviewed within 72 h from admission. Notably, 30-day mortality fell by 7.6% per year in the 4 years after the introduction of the National Hip Fracture Database compared to a 1.8% per year decrease in the 4 years preceding its introduction [133]. Similar systems are being established in other European countries [134].

In the next tier of the pyramid in Fig. 2, it was recognized that patients presenting with a fragility fracture, related or unrelated to a fall, should be assessed for osteoporosis and receive effective management to improve their bone health and reduce future fracture risk. The preferred solution is the establishment of Fracture Liaison Services, which are usually hospital-based and provide a co-ordinator to identify patients aged 50 and over with a first fracture [135, 136]. The identification is followed by risk assessments and initiation of evidence-based interventions for bone health and falls prevention. Ideally, the service would also monitor adherence and any recurrent events, but this is frequently passed back to primary care services. Treatment includes prescribing bone strengthening medicines, as well as referral to falls risk assessment and prevention services where appropriate. To date, no randomised control trials have demonstrated the effectiveness of FLS in reducing fracture risk, but such trials have shown increased DXA testing, treatment initiation and early adherence in FLS-like settings [137‐139]. Observational studies have shown reductions in fracture risk, though somewhat inconsistently, with difficulty in interpreting the results due to a number of biases in such study designs [140]. Many of these biases were overcome in a recent Swedish study, comparing fracture rates before and after FLS implementation, with non-FLS hospitals as an added comparator [141]. The risk of recurrent fracture was 18% lower in the period post FLS implementation compared with the control period (hazard ratio = 0.82, 95% confidence interval [CI] 0.73–0.92, p = 0.001), with no change in recurrent fracture rate in the non-FLS hospitals [141]. The number and standard of FLS worldwide continues to increase under the auspices of the International Osteoporosis Foundation’s Capture The Fracture ® programme (https://​www.​capturethefractu​re.​org/​).

There is some overlap between the next two tiers of the strategy shown in Fig. 2, namely individuals at highest risk of fragility fracture and older people, as age itself is a prominent risk factor for fracture. The individuals encompassed by these tiers are largely those where the burden of osteoporosis assessment and management falls on primary care health professionals, and it is this constituency to which the proposed screening approach would be largely directed. There is good evidence that, in this setting, treatment rates are very low. For example, in a recent study, treatment rates in women ≥ 70 years, deemed to be at high risk of fracture, were examined in primary care in 8 European countries [142]. The women were enrolled when visiting their general practitioner, regardless of the reason for attendance, and data captured on fracture risk and osteoporosis treatments. Increased risk of fracture was characterized as one or more of the following, a history of fracture, FRAX 10-year probability of fracture above country-specific thresholds, or a T score ≤  − 2.5 at the spine or hip. In the 3798 enrolled patients, median FRAX probability (calculated without BMD) was 7.2% for hip fracture and 16.6% for major osteoporotic fracture. Overall, 2077 women (55%) met one or more of the criteria for increased risk of fracture (median 10-year probabilities of hip and major osteoporotic fracture 11.2% and 22.8% respectively). An osteoporosis diagnosis was recorded in 804 patients (21%); most (80%) of these were at increased fracture risk. The treatment gap (the proportion at high risk but untreated) was 75%, varying from 53% in Ireland to 91% in Germany [142]. The treatment gap was somewhat lower in women with an osteoporosis diagnosis compared to those without (31% vs. 94% respectively). These data are very similar to osteoporosis treatment gaps assessed in other recent studies [11, 143], and reflect the need for integrated approaches to fracture prevention [144].

As described above, strides are being made in the clinical management of fracture risk, but the impact on patient and healthcare outcomes remains sub-optimal. The proposed screening program, if well implemented, would actually address the current deficiencies in care of a high-risk subgroup within the population, i.e. it should be regarded as a means to optimize healthcare in a greater proportion of those at risk.

There should be evidence from high quality randomised controlled trials that the screening programme is effective in reducing mortality or morbidity. Where screening is aimed solely at providing information to allow the person being screened to make an “informed choice”, there must be evidence from high quality trials that the test accurately measures risk. The information that is provided about the test and its outcome must be of value and readily understood by the individual being screened.

As stated earlier, the proposed screening programme is based on the randomised, controlled SCOOP study which is described in more detail below [3]. Two additional randomised studies, namely the Risk-stratified Osteoporosis Strategy Evaluation study (ROSE) from Denmark [145] and the SALT Osteoporosis Study (SOS) from the Netherlands [146], have used FRAX-based approaches for population screening and are also discussed below.

As part of a systematic review of randomised studies of screening for high fracture risk, the investigators of the SOS study subsequently undertook a meta-analysis of the three studies described in detail above [149]. This analysis comprised a total of 42,009 participants, reflecting all the women participating in the SCOOP and SOS studies, plus all the women providing a FRAX-based assessment at entry to the ROSE study. The latter group was deemed to be comparable to the participants in the other two studies. The proportions receiving treatment recommendations in the studies were 13% in ROSE, 14% in SCOOP and 25% in SOS, with subsequently 11%, 15%, and 18% respectively commencing therapy. Despite these relatively low treatment rates, the meta-analysis showed a statistically significant decrease of 5% in the incidence of osteoporotic fractures (HR 0.95, 95% confidence interval (CI) 0.89–1.00) when comparing screening to usual care. Most notably, there was a 20% decrease in hip fractures (HR 0.80; 95%CI 0.71–0.91) (Fig. 4A). There was also a 9% decrease in major osteoporotic fractures (HR = 0.91; 0.84–0.98), but this analysis didn’t include the outcome of major osteoporotic fractures in 9 (HR 0.88, 0.79–0.98, p = 0.018) [150]. Addition of the SCOOP result to the meta-analysis resulted in a similar overall reduction in major osteoporotic fractures (Fig. 4B). There was no difference in all-cause mortality (HR 1.04; 0.95–1.14). It is worth noting that both ROSE and SOS used the FRAX MOF output to target assessment; if the FRAX Hip output had been used as in 9, it is possible that the reduction in hip fractures would have been somewhat greater.

In the pooled cohort, the NNS and NNT for hip fractures were 272 and 28 respectively. The meta-analysis clearly showed population screening to be effective, with the biggest reduction observed in the outcome of hip fracture, leading the authors to conclude that implementation of screening in older women should be considered a serious option.

In summary, three randomised controlled studies and a subsequent meta-analysis have established efficacy for the use of FRAX screening to reduce the risk of hip fracture, as well as other fracture outcomes.