Optimizing fracture prevention: the fracture liaison service, an observational study
- First Online:
- Cite this article as:
- Eekman, D.A., van Helden, S.H., Huisman, A.M. et al. Osteoporos Int (2014) 25: 701. doi:10.1007/s00198-013-2481-8
- 561 Views
The response rate to the invitation to the fracture liaison service and reasons for non-response were evaluated in 2,207 fragility fracture patients. Fifty-one percent responded; non-responders were most often not interested (38 %) or were hip fracture patients. After 1 year of treatment, 88 % was still persistent and 2 % had a new fracture.
To increase the percentage of elderly fracture patients undergoing a dual energy x-ray absorptiometry (DXA) measurement, and to investigate why some patients did not respond to invitation to our fracture liaison service (FLS).
In four Dutch hospitals, fracture patients ≥50 years were invited through a written or personal invitation at the surgical outpatient department, for a DXA measurement and visit to our FLS. Patients who did not respond were contacted by telephone. In patients diagnosed with osteoporosis, treatment was started. Patients were contacted every 3 months during 1 year to assess drug persistence and the occurrence of subsequent fractures.
Of the 2,207 patients who were invited, 50.6 % responded. Most frequent reasons for not responding included: not interested (38 %), already screened/under treatment for osteoporosis (15.7 %), physically unable to attend the clinic (11.5 %), and death (5.2 %). Hip fracture patients responded less frequently (29 %) while patients with a wrist (60 %) or ankle fracture (65.2 %) were more likely to visit the clinic. In 337 responding patients, osteoporosis was diagnosed and treatment was initiated. After 12 months of follow-up, 88 % of the patients were still persistent with anti-osteoporosis therapy and only 2 % suffered a subsequent clinical fracture.
In elderly fracture patients, the use of a FLS leads to an increased response rate, a high persistence to drug treatment, and a low rate of subsequent clinical fractures. Additional programs for hip fracture patients are required, as these patients have a low response rate.
KeywordFractureFracture liaison serviceOsteoporosisPersistencePreventionResponse
Osteoporotic fractures frequently occur; in the Netherlands, it is estimated that more than 80,000 clinical (i.e., symptomatic) fractures occur each year in patients aged 50 years or higher . The total number is still increasing due to longevity and the increasing number of older individuals . Osteoporotic fractures are associated with increased morbidity and mortality [3–5]. Furthermore, patients with an incident fracture have an increased risk of subsequent fractures [6–8]. In a previous study performed in the Netherlands, 6.5 % of the fracture patients suffered a subsequent fracture in the first year and 10.8 % in the first 2 years after the initial fracture, while in only around 5 %, the presence or absence of underlying osteoporosis was investigated . There is overwhelming evidence that effective drugs, particularly bisphosphonates, can reduce the incidence of new vertebral fractures by 35–70 % and new non-vertebral fractures by around 20–25 % in patients diagnosed with osteoporosis [10, 11]. Screening high-risk patients by measuring bone mineral density (BMD) using dual energy x-ray absorptiometry (DXA) and, if necessary, by treating them with generic bisphosphonates could be a cost-effective strategy and is endorsed by current guidelines [1, 12].
In the last decade, specialized fracture and osteoporosis clinics, the so called fracture liaison services (FLSs) run by nurse practitioners, supervised by a physician, have been started in several countries [13–15]. Typically, patients are screened by performing a DXA and, if necessary, are treated in an outpatient setting or FLSs treatment advice is given to the general practitioner . A considerable percentage of screened patients are diagnosed with osteoporosis: at the first fracture liaison service in the UK, 46 % of elderly fracture patients with a DXA measurement were diagnosed with osteoporosis . Comparable data were found at the Osteoporosis Care Service (OCS) in the USA, 56 % of the patients had a T-score of ≤2.5, and in FLSs in the Netherlands, a prevalence of up to 40.7 % was found [14, 18].
In general, low persistence to anti-osteoporotic treatment is an important problem in reducing subsequent fractures and in improving cost-effectiveness [19–23]. There is evidence, however, that initiation of treatment in a fracture liaison service has a positive effect on persistence. In a French study, 279 patients who were seen at the FLS after a fragility fracture were contacted by telephone to answer a questionnaire. If they could not be reached themselves, their general practitioner was contacted. The answers of 155 patients were analyzed, which revealed that 80 % was still persistent with medication 1 year after initiation of treatment .
Reasons for not responding to invitation are rarely documented [18, 25]. To our knowledge, no study has been performed prospectively investigating the reasons for not attending a FLS. Therefore, we investigated what reasons were given by non-responders for not attending the clinic, and whether this proportion could be increased by contacting non-responders by telephone.
Furthermore, patients diagnosed with and treated for osteoporosis were followed during 1 year to monitor drug persistence and subsequent fracture rates.
Patients and methods
Situation before start of the study—the outpatient clinic
Four Dutch hospitals with an established fracture and osteoporosis outpatient clinic participated in this study: Maastricht University Medical Center (MUMC), Sint Franciscus Gasthuis in Rotterdam (SFG), University Medical Center in Utrecht (UMCU), and VU University Medical Center in Amsterdam (VUmc). These clinics were run by different departments; in Maastricht by the trauma department in collaboration with orthopedic surgeons and rheumatologists (a “surgical” FLS), in Rotterdam and Utrecht by the rheumatology department and the geriatric department, respectively, and in Amsterdam by a collaboration between the rheumatology and internal medicine/endocrinology department (“non-surgical” FLSs). All men and women, aged 50 years or older, presenting at the emergency department of these centers with a low trauma (fall from standing height or less) clinical fracture (other than a hand, foot or skull fracture) were invited for a DXA measurement (total hip and spine or wrist). Patients residing outside the region allocated to the hospital, and/or patients older than 90 years were not included.
After a DXA measurement was performed, a fracture nurse evaluated risk factors for osteoporotic fractures and discussed the DXA results with the patient. Arbitrarily, osteoporosis was defined as the occurrence of a fracture in combination with a T-score of −2 or lower at either the spine (L1–L4) and/or total hip or as height loss of 25 % or more in at least one of vertebrae T5–T10 or 20 % or more height loss in vertebrae T11-L4, according to the method of Genant et al. , regardless of the T-score. Since a previous fracture independently increases the risk of subsequent fractures (beyond that explained by a low BMD) , we chose to start anti-osteoporotic treatment in fracture patients with a T-score of −2 or lower, instead of the standard −2.5 cut-off point, in accordance with McLellan et al. and Oyen et al. [15, 28]. If other reasons for anti-osteoporosis treatment existed, e.g., chronic glucocorticoid use or multiple fractures, initiation of therapy was left to the discretion of the physician.
Further analyses of possible secondary causes of osteoporosis were performed on indication, and anti-osteoporotic treatment was prescribed under supervision of a specialized physician. Except for patients with gastrointestinal complaints or contraindications for treatment with oral bisphosphonates, all patients received an oral bisphosphonate and, if necessary, supplementation with calcium and vitamin D. The intake of calcium was estimated by counting the number of units of dairy products a patient consumed daily; for instance, a glass of milk or bowl of yogurt represents one unit. In case of gastrointestinal complaints or contraindications for oral bisphosphonates, zoledronic acid or strontium ranelate was prescribed. Every 3 months during 1 year of therapy, an appointment (at the clinic or by telephone) was scheduled, to evaluate therapy (checking for adherence and possible side effects), to answer additional questions of patients, and to motivate them to continue therapy.
Situation before start of the study—invitation
Due to the fact that the four outpatient clinics were already established when the study was initiated and they are managed by different departments, there is a slight difference in the way in which patients are invited for osteoporosis screening. In Maastricht (MUMC), all fracture patients are invited by the fracture nurse during the control visit for their fracture at the surgical outpatient clinic or during the admission for the treatment of their fracture. In the other three centers, selected patients (only low trauma fractures, no hand, foot, or skull fractures) are sent a written invitation within 3 months after the fracture, including an explanatory letter, practical information about the fracture and osteoporosis clinic, and a form and contact information for the department of radiology.
Situation during the study—non-response
From October 2007 until December 2008, reasons for non-response to invitations for the FLS were recorded. In MUMC, patients were invited personally by the fracture nurse for screening during their outpatient clinic control visit for their fracture at the surgery department or during their admission for the fracture. If they declined, the reason for their refusal was registered. In the three other centers, patients who did not respond within 8 weeks after written invitation were contacted by telephone to ask for the reason of non-responding. If patients could not be reached after three attempts at different times of the day, they were recorded as “unable to reach.” Reasons for not attending the clinic were registered and divided into six categories: death, not interested, already screened/under treatment for osteoporosis, physically unable to attend the clinic, insufficient mastery of the Dutch language and residing in retirement home/admitted to hospital.
In VUmc, SFG, and MUMC, the type of fracture for all invited patients was recorded and in VUmc, the results of bone density measurements (normal bone mineral density, osteopenia or osteoporosis) of all patients who visited the clinic were recorded as well.
Situation during the study—follow-up
Compliance was evaluated and falls and new fractures were registered in all fracture patients diagnosed with and treated for osteoporosis after 3 months at their control visit at the outpatient clinic. They were sent short questionnaires concerning these topics 6 and 9 months after start of treatment. After 12 months, a final visit took place at the outpatient clinic or by telephone. At VUmc and SFG, a second DXA measurement was performed prior to this visit.
Response to invitation to the outpatient clinic
Baseline characteristics and reasons for non-response of 2,207 patients aged >50 years presenting with a fracture at the emergency department
Number of patients invited
Female gender, N (%)
Mean age, years (SD)
Fracture type N (%)a
(total N = 2,034)
Visit to outpatient clinic N (% of total nr of fractures)
No visit to outpatient clinic
Reasons for not attending (N, % of no visit to clinic)
Could not be reached
Already screened and/or under treatment for osteoporosis
Physically unable to visit clinic
Does not read or speak Dutch
Admitted to hospital or residing elsewhere
The majority of the 57 patients who had died roughly within 3–6 months after the fracture, had suffered a hip fracture (N = 27, 47 %). The remaining patients had various fractures, e.g., humerus, wrist, and multiple fractures. Of the four patients who were deceased, no fracture data was available. The mean (SD) age of the patients who were deceased was 80 years (9), which was significantly higher than the age of the patients that were alive at the time of the invitation for additional diagnostic measurements (mean 70 years, SD 11, p < 0.0001).
Although in all centers the mean age and gender of the patients were comparable, and the most frequent fracture localization was the wrist (roughly one third of all fractures), there were some striking differences between the centers as well. For instance, hip fractures seemed to be the second most frequent fracture in VUmc and SFG (17.4 and 24.4 %, respectively), while in MUMC, humerus and ankle fractures were more frequent (12.5 and 12 %, respectively) than hip fractures (10 %).
The response rate (the percentage of patients in which a DXA was performed) was by far the highest in MUMC (71 %), versus 37 % (Rotterdam) and 42 % (Amsterdam) at the other non-surgical FLS services (and 16 % in Utrecht, in which only a very small number of patients were enrolled). The percentage of patients that started with anti-osteoporotic treatment was much lower at the MUMC; only 108 of 637 patients that visited the surgical FLS started anti-osteoporotic drugs (17 %), while in the other three centers, 229 patients of 526 patients that visited started treatment (44 %).
The proportion of patients in whom osteoporosis was diagnosed in MUMC (after 12 months since inviting fracture patients according to study protocol) was comparable to VUmc (38.4 vs 34.9 %).
In UMCU, the rate of patients who were already screened and/or under treatment for osteoporosis was higher (31.7 %) compared to the other centers (11.7–15.3 %), as well as the percentage of patients who were physically unable to visit the clinic (26 vs 2.6–17.6 %).
In 21 % of the patients visiting the clinic at VUmc, a normal BMD was found, while 36 and 43 % were diagnosed with osteopenia and osteoporosis, respectively.
Clinical, BMD, and treatment characteristics at baseline
Baseline characteristics of screened fracture patients diagnosed with and under treatment for osteoporosis (n = 337)
N = 108
N = 100
N = 13
N = 116
Female gender (N, %)
Mean age, years (SD)
Fracture type (%)
Osteoporosis in family, N (%)a
Calcium intake, units daily, mean (SD)
Smoking, N (%)
Alcohol intake, units weekly, mean (SD)
2.51 (n = 95)
5.11 (n = 93)
8.77 (n = 93)
7.65 (n = 111)
Prednisone use, N (%)
2 (1.9) (n = 105)
3 (3) (n = 98)
>2 falls in the last 12 months, N (%)
Height (mean, cm)
Weight (mean, kg)
BMI (mean, cm/kg2) (SD)
BMD spine (mean, g/cm2) (SD)
T-score spine mean (SD)
BMD left hip mean (g/cm2) (SD)
T-score left hip mean (SD)
In the majority of patients anti-osteoporotic treatment was started with an oral bisphosphonate (90 %, n = 304; 142 risedronate, 137 alendronate, 25 ibandronate). In nine patients, only supplementation with calcium and vitamin D was started because of renal failure (n = 2) or patient refusal to take bisphosphonates (n = 7). Thirteen patients started treatment with strontium ranelate, and in nine patients, the treatment initiated was not documented.
Adherence to therapy, subsequent fractures, and BMD after 1 year
Follow-up data are available from 280 patients: 2 patients died, 19 stopped study participation (for various reasons but mostly because they experienced the study as too burdensome) and from 36 patients, the follow-up data after 12 months was incomplete due to changes in personnel at the various centers. There were no differences in age or gender between the patients with a complete follow-up and those with missing data. Of the 280 patients, 88 % (246) were still persistent with medication after 1 year. Reasons for discontinuing anti-osteoporotic medication were: gastrointestinal side effects (n = 11), musculoskeletal side effects (n = 4), other side effects (n = 5), problems with following the instructions concerning intake (n = 3), stopped without a reason (n = 5), and unknown reason (n = 6). Six patients (2.1 %) had a subsequent fracture: 1 wrist, 1 ankle, 1 hip, 1 rib, and 2 hand fractures. Data concerning falls during the past 12 months were available from 215 patients: 13 patients had fallen 2–5 times and 3 had fallen >5 times.
In VUmc and SFG, a second DXA measurement was performed. In a paired samples t test, a significant increase in BMD was found in both hip (N = 166, increase from a mean (SD) BMD of 0.763 (0.108) to 0.781 (0.102) g/cm2; 2.3 %, p < 0.001) and spine (N = 178, increase mean (SD) BMD from 0.818 (0.114) to 0.850 (0.120) g/cm2; 4 %, p < 0.001).
The major findings of the present study are, that the use of an FLS in elderly fracture patients leads to a substantially increased response rate, a high persistence to anti-osteoporotic drug treatment, and a decreased rate of subsequent fractures. We also gained more insight in the reasons for non-response, the most important reasons including: not interested (38 %), already screened/under treatment for osteoporosis (15.7 %), physically unable to attend the clinic (11.5 %), and death (5.2 %).
The response rate in the present study was almost 51 % of the 2,207 invited fracture patients, which percentage is much higher than the around 5 % of fracture patients who were screened before the implementation of FLSs . The highest response rate was found in the MUMC (71 %), were patients were invited personally. One of the explanations for this difference could be that the personal contact with the surgeon directly after the fracture might persuade patients to visit the surgical FLS at the phase they suffer from fracture pain and limitations. However, the percentage of patients that started with anti-osteoporotic treatment was much lower at the MUMC (17 %) than in the other centers (44 %), The reason for this striking difference could be that that those patients at the non-surgical FLS who reacted positively to an invitation letter are more intrinsically motivated.
Response rates above 50 % have also been reported from studies in The Netherlands (71 %), UK, (88–96 %), Spain (55.6 %), and USA (61 %) confirming that an FLS is a big step forward in the secondary prevention of subsequent fractures in elderly patients with an initial fracture. The differences in response rates between centers might probably be related to differences in patient characteristics and differences in design of the FLS [15, 18, 25, 30]. However, our study was not designed to study these differences.
The importance of utilizing an FLS has also been underlined in recently published guidelines, among which, the American Society for Bone and Mineral Research taskforce report on secondary fracture prevention. This report calls for worldwide implementation of FLSs as these services have been proven the most effective tool for screening elderly fracture patients, preventing subsequent fractures, and improving adherence and compliance [1, 12].
However, there is still room for improvement since 49 % of the patients in our study did not respond. Reasons for non-response have rarely been documented in previous studies. In our study, we prospectively collected these data in an effort to discover opportunities for improvement. Most non-responders indicated to be not interested (38 %), others were already under treatment (15.7 %), or physically unable to visit the clinic (11.5 %). In two previous studies performed in the USA and Spain, with a substantial smaller number of invited patients (n = 370 and 683, respectively, vs 2,207 in our study), reasons for non-response were recorded as well. However, these studies were not performed to primarily evaluate invitation response rates. Our data generally confirm their results; a similar percentage of patients declined the invitation (40–50.2 %). The percentage of patients that was deceased/in poor health was comparable; in our study 16.7 %, in the USA study 14 %, and was not documented in the Spanish study. A similar percentage of the patients invited for the secondary fracture prevention program in Spain were not reached by telephone (21.1 vs 23.8 % in our study). Strikingly, in our study, 15.7 % of the patients were already under treatment for osteoporosis, while this percentage was much lower in Spain (4.9 %) and substantially higher in the USA (36 %) [18, 25].
To further improve the percentage of responders, it seems worthwhile to invest more extensively in better informing the “not interested” group. A frequently mentioned reason for not attending the clinic was the patients' opinion that they were too old to visit the outpatient clinic and that treatment would not be useful anymore. Furthermore, the mean age of not responding patients was significantly higher than the mean age of responding patients. Both these facts suggest that the elderly fracture population is less likely to be screened and treated for osteoporosis, while it is known that they will benefit greatly from treatment and fracture prevention.
Possibly, providing more extensive written information in a brochure, for instance concerning prevalence of osteoporosis and the usefulness of treatment to prevent fractures in elderly people, could stimulate fracture patients to consider screening. However, oral information to non-responders resulted in only 6 % more patients at the outpatient clinic. Alternative ways to improve response rates might be by informing through the general practitioner and/or by improving public awareness about the consequences of fractures and osteoporosis, although the latter goal may be difficult to achieve.
In patients who were already under treatment (15.7 %), screening obviously is no longer relevant. Patients who are physically unable to visit the clinic are a potentially high-risk group for subsequent fractures because reduced mobility is associated with both an increased fall risk and a reduction in bone density. In our study, this subgroup consisted of patients who were older and of whom the majority had suffered a hip fracture. Only 29.6 % of all hip fracture patients visited the clinic. Some of these patients already had a reduced mobility before their fracture, but a large proportion was immobilized as a consequence of their fracture; especially in patients with a hip fracture. These findings illustrate that our current outpatient clinic strategy is not effective for patients with hip fractures and more suitable for patients with wrist and ankle fractures, of whom, respectively, 60 and 65.2 % visited the clinic. The highest response rate for hip fracture patients was seen in the MUMC (56.7 % response, vs 18.1 and 21.1 % in the VUmc and SFG, respectively), where the clinic is primarily organized by the trauma surgery department in collaboration with the orthopedic surgery and rheumatology department and patients are approached during admission or an outpatient clinic visit for their fracture. For immobilized patients, including hip fracture patients, this is a more effective strategy, as has previously been suggested in the Blue Book, in which a multidisciplinary approach and routine access to acute orthogeriatric medical support during the time of admission is recommended . In Europe, as well as in the USA, Australia, and New Zealand, several efforts are undertaken to improve post-hip fracture care in a similar way [32–36]. After discharge from the hospital, follow up of treatment can take place at an FLS as described in this article.
Of the 280 patients with complete follow-up data until 12 months after start of treatment, only 2 % had endured a subsequent clinical fracture, which percentage is considerably lower than the 6.5 % with subsequent fractures reported from fracture patients who were not screened and treated for osteoporosis after their initial fracture . The percentage found in the present study could be an underestimation, however, since it is possible that particularly, patients who were healthy enough to visit the outpatient clinic for diagnostic tests and participate in a study were included. This finding suggests a selection bias towards fracture patients with a better prognosis. In Spain and in the USA, similar percentages were found; only 4 out of 187 fracture patients (2 %) who were treated with a bisphosphonate during 4 years endured a subsequent fracture , and in the OCS, 2 % of the patients experienced a new fracture within 2 years of follow-up after initiation of treatment .
We found a persistence rate of 88 % after 1 year of treatment, while in a large systematic review and meta-analysis adherence rates of 40–50 % were found in real-world settings [21, 37]. The high adherence rate in fracture patients might be explained by the fact that these patients have experienced the main complication of osteoporosis, namely a fracture. Osteoporotic patients without a recent fracture are “asymptomatic” and may therefore lack motivation to continue treatment that involves complex intake instructions and frequent occurring side effects. Indeed, in an earlier study in Canada, a persistence of 82 % after 1 year of treatment with bisphosphonates was found in 102 patients who were screened and diagnosed with osteoporosis after a wrist fracture . In addition, in a group of 187 fragility fracture patients who were treated with bisphosphonates in Spain, persistence after a mean of 4 years was 71 % . These findings suggest that both persistence as well as subsequent fracture rate is influenced beneficially by screening and follow-up by a specialized nurse.
Our study has some limitations. A relatively large number of non-responding patients could not be reached (23.8 %), even after several attempts. The proportion of patients diagnosed with fractures with high mortality rates and prolonged recovery, such as hip fractures, was greater in the subgroup that could not be contacted than in the subgroup of patients who were contacted (32.4 vs 19.6 %), respectively. Possibly, some patients had deceased, or were still recovering from their fracture at a rehabilitation center. In addition, in other studies, comparable percentages of non-responders have been reported [18, 25].
Furthermore, data concerning DXA outcomes of responding patients visiting the clinic were only available from VUmc. In VUmc, 43 % of the patients were diagnosed with osteoporosis. Possibly, this percentage is not representative for the total population; however, it is in the same range as percentages reported in previous studies performed in The Netherlands (40–67 %) [13, 16, 39] and comparable to the percentages found in other countries (46–56 %) [15, 18].
Unfortunately, of the 316 patients who were followed during 1 year after their initial fracture, 36 were lost to follow-up due to changes in personnel. Since these patients did not differ in age, sex, or type of fracture from the patients of whom complete data were available, we assume that the missing data does not have a substantial influence on persistence and the number of subsequent fractures.
In conclusion, a fracture liaison service improves the rate of screening of fracture patients aged 50 years or older substantially to 51 %. The most important reasons for not responding included not interested (38 %), already screened/under treatment for osteoporosis (15.7 %), and physically unable to attend the clinic (11.5 %). The clinic seems more suitable for patients who remained in relative good mobility after their fracture, such as patients with wrist fractures, than for hip fracture patients. One of the challenges for the future will be to implement a strategy to improve the response rate for hip fracture patients. Persistence with anti-osteoporotic treatment in patient with an initial fracture is remarkably high, and the incidence of subsequent clinical fractures after 1 year is much lower in the setting of an FLS than reported previously.
Conflicts of interest