SUMMARY: Many of the clinical risk factors used in fracture risk assessment (FRAX) calculator are available in electronic medical record (EMR) databases and are good sources of osteoporosis risk factor information. The EPIC EMR database showed a lower prevalence of FRAX risk factors and, consequently, proportion of patients who would be deemed "high risk." INTRODUCTION: The FRAX tool is underutilized for osteoporosis screening. Many of the clinical risk factors for FRAX may be available in EMR databases and may enable health systems to perform fracture risk assessments. We intended to identify variables in an EMR database for calculating FRAX score in a cohort of postmenopausal women, to estimate absolute fracture risk, and to determine the proportions of women whose absolute fracture risks exceed the National Osteoporosis Foundation (NOF) thresholds. METHODS: Our cohort was selected using an EMR database with demographic, inpatient, outpatient, and clinical information for female patients age≥50 in a family practice, internal medicine, or obstetrics/gynecology clinic in 2007-2008. The latest physician encounter was the index date. Variables, problem and medication lists, diagnosis codes, and histories from the EMR were used to populate the 11 clinical risk factor variables used in the FRAX. These risk factor prevalence and treatment-eligible proportions were compared to those of published epidemiology studies. RESULTS: The study included 345 patients. Mean (SD) 10-year risk for any major fracture was 11.1% (6.8) when bone mineral density (BMD) was used and 11.2% (6.5) when BMI was used. About 10.1% of the cohort exceeded the NOF's 20% major fracture risk threshold and 32.5% exceeded the NOF's 3% hip fracture risk threshold when BMD was used. Overall, the number of treatment-eligible patients was slightly lower when FRAX was calculated using BMD versus BMI (13.6 and 36.8%). CONCLUSION: Our cohort using EMR data most likely underestimated the mean 10-year probability of any major fracture compared to other cohorts in published literature. The difference may be in the nature of EMRs for supporting only passive data collection of risk factor information.
SUMMARY: Many of the clinical risk factors used in fracture risk assessment (FRAX) calculator are available in electronic medical record (EMR) databases and are good sources of osteoporosis risk factor information. The EPIC EMR database showed a lower prevalence of FRAX risk factors and, consequently, proportion of patients who would be deemed "high risk." INTRODUCTION: The FRAX tool is underutilized for osteoporosis screening. Many of the clinical risk factors for FRAX may be available in EMR databases and may enable health systems to perform fracture risk assessments. We intended to identify variables in an EMR database for calculating FRAX score in a cohort of postmenopausal women, to estimate absolute fracture risk, and to determine the proportions of women whose absolute fracture risks exceed the National Osteoporosis Foundation (NOF) thresholds. METHODS: Our cohort was selected using an EMR database with demographic, inpatient, outpatient, and clinical information for female patients age≥50 in a family practice, internal medicine, or obstetrics/gynecology clinic in 2007-2008. The latest physician encounter was the index date. Variables, problem and medication lists, diagnosis codes, and histories from the EMR were used to populate the 11 clinical risk factor variables used in the FRAX. These risk factor prevalence and treatment-eligible proportions were compared to those of published epidemiology studies. RESULTS: The study included 345 patients. Mean (SD) 10-year risk for any major fracture was 11.1% (6.8) when bone mineral density (BMD) was used and 11.2% (6.5) when BMI was used. About 10.1% of the cohort exceeded the NOF's 20% major fracture risk threshold and 32.5% exceeded the NOF's 3% hip fracture risk threshold when BMD was used. Overall, the number of treatment-eligible patients was slightly lower when FRAX was calculated using BMD versus BMI (13.6 and 36.8%). CONCLUSION: Our cohort using EMR data most likely underestimated the mean 10-year probability of any major fracture compared to other cohorts in published literature. The difference may be in the nature of EMRs for supporting only passive data collection of risk factor information.
Authors: John A Kanis; Anders Oden; Helena Johansson; Fredrik Borgström; Oskar Ström; Eugene McCloskey Journal: Bone Date: 2009-02-03 Impact factor: 4.398
Authors: Russel Burge; Bess Dawson-Hughes; Daniel H Solomon; John B Wong; Alison King; Anna Tosteson Journal: J Bone Miner Res Date: 2007-03 Impact factor: 6.741
Authors: C De Laet; J A Kanis; A Odén; H Johanson; O Johnell; P Delmas; J A Eisman; H Kroger; S Fujiwara; P Garnero; E V McCloskey; D Mellstrom; L J Melton; P J Meunier; H A P Pols; J Reeve; A Silman; A Tenenhouse Journal: Osteoporos Int Date: 2005-06-01 Impact factor: 4.507
Authors: Ethel S Siris; Ya-Ting Chen; Thomas A Abbott; Elizabeth Barrett-Connor; Paul D Miller; Lois E Wehren; Marc L Berger Journal: Arch Intern Med Date: 2004-05-24
Authors: Jeffrey R Curtis; Leslie A McClure; Elizabeth Delzell; Virginia J Howard; Eric Orwoll; Kenneth G Saag; Monika Safford; George Howard Journal: J Gen Intern Med Date: 2009-06-24 Impact factor: 5.128
Authors: Christopher C Stahl; Patrick B Schwartz; Glen E Leverson; James R Barrett; Taylor Aiken; Alexandra W Acher; Sean M Ronnekleiv-Kelly; Rebecca M Minter; Sharon M Weber; Daniel E Abbott Journal: Surgery Date: 2020-04-26 Impact factor: 3.982