Family Practice Vol. 20, No. 3, 289-293
© Oxford University Press 2003
Screening |
Osteoporosis: what are the implications of DEXA scanning ‘high risk’ women in primary care?
a Primary Care Sciences Research Centre, Keele University, Keele, North Staffordshire ST5 5BG
b Staffordshire Rheumatology Centre, The Haywood, Burslem, Stoke-on-Trent, North Staffordshire ST6 7AG, UK.
Correspondence to Dr Elaine Thomas; E-mail: e.thomas{at}keele.ac.uk
Thomas E, Richardson JC, Irvine A, Hassell AB and Hay EM. Osteoporosis: what are the implications of DEXA scanning ‘high risk’ women in primary care? Family Practice 2003; 20: 289–293.
Received 17 September 2002; Accepted 23 December 2003.
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
Background. Current recommended practice for the use of dual X-ray absorptiometry (DEXA) scans in screening for osteoporosis is to concentrate on women at ‘high risk’.
Objective. We have applied such a screening strategy, in a general practice setting, to estimate the number of women requiring scans.
Methods. A two-phase survey was carried out: (i) postal screen of clinical indicators for low bone mineral density (BMD) to define women at ‘high risk’; and (ii) DEXA scanning of the sample at ‘high risk’ set in two general practices in North Staffordshire. Computerized general practice records were used to define a purposive sample of 1001 women, to receive the screening tool, consisting of three equal size groups (i) those with an early hysterectomy; (ii) those receiving oral corticosteroids on repeat prescription; and (iii) those on the practice cervical smear register. A random sample of women defined at ‘high risk’ by the screening tool were invited to have a DEXA scan. The main outcome of interest was the presence of low BMD as measured by a DEXA scanner.
Results. Sixty-five out of 95 women invited (68%) agreed to undergo a DEXA scan: median age = 52 years (interquartile range 44–64 years). Twenty-nine of these 65 women (45%) were classified with low BMD (WHO criteria): 90% had densities below their age-matched mean. Extrapolating from the observed findings to the main study practice (n = 9000 total population), we estimate that 162 women would be defined at ‘high risk’, and, if all were offered a scan, 105 would comply and 56 would be defined with low BMD.
Conclusions. Using this approach, we estimate the unmet need, in women, for DEXA scans to be 180 per 10 000 total practice population. Allowing for scan uptake, this would define 
60 women per 10 000 total practice population with low BMD. The application of this screening strategy has identified a group of women who might benefit from treatment or prophylaxis for osteoporosis.
Keywords. DEXA scan, high risk, osteoporosis, primary care, women.
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Osteoporosis is a progressive disorder characterized by low bone mass, leading to enhanced bone fragility and increased fracture risk.1 The World Health Organization (WHO) published a more quantitative definition as "bone mineral density (BMD) more than 2.5 standard deviations below the norm for healthy young adults".2 The remaining lifetime risk for either a hip or vertebral fracture in a 50-year-old women has been estimated at
15%,3 and these osteoporotic fractures are associated with significant excess morbidity, mortality and economic cost.4 Currently, the best predictor of future osteoporotic fracture is the level of BMD measured by dual X-ray absorptiometry (DEXA) scanning.5 However, debate exists regarding the cost-effectiveness of screening the asymptomatic population.6 In women, recommended practice is to concentrate resources on those at ‘high risk’ of developing osteoporosis, as fracture risk is known to be substantially greater in women with positive clinical indicators and low BMD. Many factors have been shown to accelerate bone loss. including early menopause, vitamin D deficiency, oral steroid treatment and hyperthyroidism.7,8 Other characteristics associated with osteoporosis include age, previous low trauma fracture, lack of weight-bearing exercise, late menarche, smoking and alcohol consumption.7,8 One strategy for general practice might be to identify women with these characteristics and measure their bone density so that preventative strategies or treatment can be targeted towards those at greatest fracture risk.
We report a study that explored the role of measuring BMD, with a DEXA scanner, in women with clinical risk factors to determine the unmet needs associated with osteoporosis in primary care. The specific objectives were: (i) to apply a screening strategy to identify women at ‘high risk’ of low BMD; (ii) to establish the proportion of these ‘high risk’ women meeting the definition of osteopaenia/osteoporosis from DEXA scan results; and (iii) extrapolating from the observed data, to estimate the number of women to be offered a DEXA scan if this screening approach was applied to an average general practice.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The design was a two-phase survey: a postal questionnaire and a DEXA scan. The first phase was a cross-sectional survey of women registered with a local general practice. The age–sex register from the study general practice was used to obtain a sample of women without a current diagnosis of osteoporosis and not receiving bisphosphonates. The sampling process was derived following meetings with local GPs and rheumatologists, and aimed to identify a group of women whom GPs would, in their usual clinical practice, consider investigating for osteoporosis.
General practice records
The practice computerized records were searched to establish three, equally sized groups of women: (i) those documented as having a hysterectomy (with or without a oopherectomy) prior to the age of 45 years; (ii) those aged over 18 years receiving oral corticosteroids on repeat prescription; and (iii) those aged over 50 years on the practice cervical smear register. The first two groups contained women with at least one established risk factor for osteoporosis and the third contained women with a greater cross-section of risk factor profiles. This enriched sample was chosen on purpose so that a large group of women at ‘high risk’ of low BMD, with various risk factor profiles, could be determined.
As the main study practice had only 231 women who fulfilled criterion (i), an additional 103 women were sampled randomly from 232 women who fulfilled this criterion at a second general practice. In the main study practice, groups (i) and (iii) contained more than the needed 333 women (356 and 1280, respectively) and so a simple random sampling procedure was used within these two groups to select the required sample size.
Recruitment survey
A questionnaire was posted to all women in the three general practice groups, with a reminder questionnaire sent after 2 weeks. The questionnaire gathered information on demographics and 16 clinical indicators of low BMD (see Table 1
).
|
DEXA scan sample
In the second phase of the survey, the clinical indicator data from the postal questionnaire were used to define a group of women at ‘high risk’ of low BMD, i.e. those women who reported the presence of: (i)
5 clinical indicators; or (ii) 2–4 clinical indicators, one of which was either early hysterectomy or oral steroid use
All women with at least five clinical indicators and an 50% random sample of those with 2–4 clinical indicators were sent a letter inviting them to undergo a DEXA scan, together with a patient information leaflet outlining the procedure. Those agreeing to a scan were telephoned, and a convenient appointment time was arranged. Women who declined were replaced, randomly, until the required sample size was achieved.
DEXA scanning
DEXA scans were taken of the lumbar spine (L1–L4) and the proximal right femur at the Staffordshire Rheumatology Centre, and were reported on by a single consultant rheumatologist (AH) according to a standard protocol. The scanner used was an Hologic QDR 4500A fan-beam X-ray bone densitometer. The BMD measurements in the study population were compared with the Hologic normal population database for the spine and the NHANES database for the hip. Women were classified as having low BMD if they had osteopaenia or osteoporosis at either site.2
Ethical approval for all stages of the study was granted by the North Staffordshire Research Ethics Committee.
Statistical analysis
Independent funding was secured for between 60 and 70 DEXA scans. From a pilot study using general practice data, it was determined that 15% of women would have at least one documented risk factor for osteoporosis. Assuming a 60% response rate to the survey, a sample of 1000 women would then yield detailed risk factor information on 600 women, of which 90 (15%) would have a minimum of one risk factor for osteoporosis. This sample size of 1000 would then provide sufficient numbers of subjects for the DEXA scanning phase.
Simple percentages of the presence of the self-reported risk factors and the number of women defined at ‘high-risk’ are presented for the study responders. In women undergoing a DEXA scan, the frequency (%) of low BMD was calculated at both the femur and the spine. Extrapolating from the observed data, estimates of the number of women to be offered a DEXA scan, when applying this ‘high risk’ strategy to the total practice population, are presented. These calculations were performed within the three groups, i.e. early hysterectomy, oral steroid use and smear register, and included a factor to allow for estimated scan uptake. All analysis was carried out using Stata 6.0 statistical software package.9
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
A total of 1001 women (334 early hysterectomy, 334 steroid use and 333 cervical smear register) were mailed a questionnaire; 847 (85%) were returned completed and responders had a median age of 59 years (range 18–92 years). Response rates were similar across the three groups (85, 83 and 86%). Table 1
presents frequencies (%) of the 16 clinical indicators of low BMD recorded on the postal questionnaire. The most common clinical indicators (other than those that were part of the sampling process) were ever smoking, low calcium intake and a low oestrogen state (i.e. late menarche/early menopause). There were low frequencies of renal failure, anorexia, liver disease and partial gasteroctomy/colectomy.
Of the responders, 138 fulfilled our criteria for ‘high risk’ (44 had 5 risk factors, and 94 satisfied the conditions for those with 2–4 risk factors). Of these, 16 women did not give permission to be re-contacted, and so 122 were available for DEXA scan invitation. Initially, 87 women were offered a DEXA scan, of which 59 completed a scan, 19 refused the invitation (one on her GP’s recommendation) and nine could not be contacted. In order to achieve our sample of 60–70 scans, eight more subjects were contacted, of which six completed a scan and two refused. Therefore, a total of 65 women were scanned (scan uptake rate = 68%); median age of 52 years (range 31–91 years).
Osteopaenia or osteoporosis (hereafter called low BMD) was recorded in 28 (43%) subjects at the lumbar spine and 12 (18%) at the femur (Table 2). All but one subject with low BMD at the femur also had low BMD at the spine, and 11 were concordant for both sites. In total, 29 subjects were classified with low BMD (prevalence of 45% in this ‘high risk’ group) and their median Z-score was –0.71 (range –1.9 to 0.9), with 90% having densities below the age-matched mean (Fig. 1).
|
|
Applying the screening strategy, described above, to the main study general practice (total population = 9000) resulted in an estimated 162 women who would fulfil the criteria for DEXA scanning. Of these, we estimated that 105 women (64%) would agree to a scan and 56 (53%) would be classified with low BMD (Table 3
).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The principal question addressed here relates to the application and clinical usefulness of DEXA scanning in primary care. We quantified the degree of ‘unmet need’ by determining the number of women with low BMD identified by our screening strategy. Thus the results presented apply to women without a current diagnosis of osteoporosis who were defined at ‘high risk’ of low BMD on the basis of the clinical indicators and strategy used in this study.
The strategy applied has shown that it would be successful in identifying women at ‘high risk’, as almost all of the women identified had BMDs below their age-matched mean (Fig. 1). Additionally, applying age-specific prevalence rates of osteoporosis, derived from a random sample in a UK population,10 to the women scanned in this study (n = 65) led to a lower expected prevalence of 2.5/65 (3.8%) women when compared with the observed prevalence of 6/65 (9.2%).
We estimate that the application of our screening strategy would define 180 women per 10 000 total practice population at ‘high risk’ of low BMD. If DEXA scans were to be offered to all these, the observed uptake rate would suggest that 115 would complete a scan, and 60 women would be identified in whom intervention may be indicated. Following a record review of the main study practice, 71 women were not included in the sampling procedure because they had a GP diagnosis of, or were on bisphosphonates for, osteoporosis at the time of the study. Thus, combining these two sources of information, a minimum estimate of the total number of women at the study practice with low BMD is 130 per 10 000 total practice population. In addition to this ‘cross-sectional’ information, the number of scans needed per year will also need to include a factor to reflect the number of repeat scans needed for women with an established diagnosis of osteoporosis in order to monitor treatment response.
Methodological issues
The generalizability of our results might be questioned as the study was based in only two local general practices. Although subjects at the practices are almost exclusively Caucasian, they do have a varied socio-economic mix. The choice of clinical indicators and definition of ‘high risk’ was pragmatic and derived by consultation with local GPs and rheumatologists. Essentially, it was chosen to identify a group of women whom GPs, in the course of a routine consultation, might consider for DEXA scanning. The starting point was women with well-recognized, easily identifiable risk factors for osteoporosis, i.e. those with early hysterectomy or those on long-term (repeat prescription) oral steroids, that can be identified by searching the computerized practice database. In addition, we wanted to supplement these groups with women who had other recognized risk factors for osteoporosis such as previous fracture or inactivity. A pilot study in three local GP practices had demonstrated that such information could not be reliably obtainable from computerized records, hence, a postal survey was used.
The WHO criteria2 used to define low BMD are relative and depend on the mean and standard deviation of the values in the reference population. Minor differences in the reference data can have important effects on patient classification; Ahmed11 applied two different normative data sets to their study data and found a 2-fold difference in the prevalence of osteoporosis (6% compared with 15%). Nevertheless, the normative data sets utilized in this study are widely used in clinical practice in the UK.
The estimate of the number of scans needed based on our data is considerably higher than previous estimates from the UK.12,13 Our estimate was based on the results from systematically surveying women with potential ‘unmet need,’ whereas previous estimates appear to have been based on numbers scanned in active units using standard criteria for scanning.
Implications for primary care
Fracture is the clinically important consequence of osteoporosis. Survival rate in the 5 years following hip or vertebral osteoporotic fracture is 80% of that expected for the same age subject with no fracture.4 Symptomatic fractures are the only way that patients with osteoporosis are identified clinically. However, this is too late to help the patient or to reduce the public health burden. There is a need, therefore, to find a feasible and acceptable way of identifying individuals at future risk of fracture so that appropriate prophylaxis can be recommended.
In our study, low BMD, defined by DEXA scanning, was found to be common amongst the women chosen on the basis of predefined risk factors for osteoporosis. In terms of yield, the group with an early hysterectomy yielded the greatest number of women at ‘high risk’ for low BMD and, in practice, this seems to be a good starting point for a practice-based policy. Additionally, in this hysterectomized group, the percentage of women with low BMD was similar in those who had and had not, at some time, taken hormone replacement therapy (HRT). This suggests that it is necessary to scan all women with an early hysterectomy and not restrict it to those not taking HRT. The lowest yield, perhaps as expected, was from the practices’ cervical smear registers. This may suggest a simpler method for defining ‘high risk’ that could be implemented within all general practices through the use of routinely recorded data.
From a health economics perspective, the results of this study may help practices when making decisions regarding treatment strategies, i.e. should they treat all those at ‘risk’ or concentrate treatment on those who have low BMD in addition to clinical indicators. In this study, of the 65 women scanned, 36 had normal T-scores, indicating that the additional step of including a DEXA scan in the identification process has ‘saved’ these women from undergoing unnecessary treatment. The DEXA scan appears to be a simple, feasible and acceptable method for assessing low BMD as part of a screening process; 70% of women invited completed a scan, an uptake rate that may be improved upon in a non-research setting, and no complaints were received regarding the scanning procedure.
One practical way of implementing our screening strategy would be to use a step-wise approach in primary care to define women at risk of low BMD: first, targeting women with recorded evidence of an early hysterectomy or repeat prescription for oral steroids. These women are easily identifiable from GP computerized databases and produce a high yield of women with low BMD. Secondly, a more opportunistic strategy can be used where the nurse or GP can gain information on clinical indicators on women aged over 50 years. Women then identified at ‘high risk’ should be offered a DEXA scan to determine those in need of prophylaxis or treatment for osteoporosis.
|
Acknowledgments |
|---|
We would like to thank Lesley Radford for performing the DEXA scans, Tracy Whitehurst and Rhian Hughes for preparing the general practice records, Martyn Lewis for initial analysis, Professor Peter Croft for his constructive criticism during the manuscript development, and the patients and GPs who participated in this research. Funding of the DEXA scans was by a £2500 grant from Proctor and Gamble. Funding for Dr Alison Irvine’s protected time as a GP research fellow at the Primary Care Sciences Research Centre was provided as part of a grant from the North Staffordshire District Health Authority for the North Staffordshire GP Research Network.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
1 Riis BJ. Biochemical markers of bone turnover. II: diagnosis, prophylaxis, and treatment of osteoporosis. Am J Med 1993; 9: 17S–21S.
2 WHO Study Group. Assessment of Fracture Risk and its Application to Screening for Postmenopausal Osteoporosis. WHO Technical Report Series no. 843. Geneva: WHO, 1994: 1–129.
3 Melton LJ III, Chrischilles EA, Cooper C, Lane AW, Riggs BL. How many women have osteoporosis? J Bone Miner Res 1992; 7: 1005–1010.[Web of Science][Medline]
4 Cooper C, Atkinson EJ, Jacobsen SJ, O’Fallon WM, Melton LJ. Population-based study of survival after osteoporotic fractures. Am J Epidemiol 1993; 137: 1001–1005.
5 Marshall D, Hailey D, Jonsson E. Health policy on bone mineral density measurement technology in Sweden and Australia. Health Policy 1996; 35: 217–228.[CrossRef][Web of Science][Medline]
6 Sheldon T, Raffle A, Watt I. Why the report of the Advisory Group on Osteoporosis undermines evidence based purchasing. Br Med J 1996; 312: 296–298.
7 Melton LJ III. Epidemiology of osteoporosis: predicting who is at risk.Ann NY Acad Sci 1990; 94: 646–650.[CrossRef]
8 Cummings SR, Nevitt MC, Browner WS et al. Risk factors for hip fracture in white women. Study of Osteoporotic Fractures Research Group. N Engl J Med 1995; 332: 767–773.
9 StataCorp. Stata Statistical Software: Release 6.0. College Station (TX): Stata Corporation, 1999.
10 Petley GW, Cotton AM, Murrills AJ et al. Reference ranges of bone mineral density for women in southern England: the impact of local data on the diagnosis of osteoporosis. Br J Radiol 1996; 69: 655–660.
11 Ahmed AI, Blake GM, Rymer JM, Fogelman I. Screening for osteopaenia and osteoporosis: do the accepted normal ranges lead to overdiagnosis? Osteoporos Int 1997; 7: 432–438.[CrossRef][Web of Science][Medline]
12 Compston JE, Cooper C, Kanis JA. Bone densitometry in clinical practice. Br Med J 1995: 310: 1507–1510.
13 Royal College of Physicians. Osteoporosis: Clinical Guidelines for Prevention and Treatment. Suffolk (UK): The Lavenham Press Ltd, 1999.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  M. Gattellari, N. Donnelly, N. Taylor, M. Meerkin, G. Hirst, and J. E Ward Does 'peer coaching' increase GP capacity to promote informed decision making about PSA screening? A cluster randomised trial Fam. Pract., June 1, 2005; 22(3): 253 - 265. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. C Richardson, A. B Hassell, E. Thomas, and E. M Hay GPs' perceptions of the role of DEXA scanning: an exploratory study Fam. Pract., February 1, 2004; 21(1): 51 - 53. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||