Arch Osteoporos (2015) 10: 25 DOI 10.1007/s11657-015-0230-7
ORIGINAL ARTICLE
Validation of a modified FRAX® tool for improving outpatient efficiency—part of the BCatch Before a Fall^ initiative Simon Parker 1 & Maria Ciaccio 2,3 & Erica Cook 4 & Graham Davenport 5,6 & Alun Cooper 7,8 & Simon Grange 9,10 & Peter Smitham 2,3,4
Received: 28 January 2015 / Accepted: 4 August 2015 / Published online: 14 August 2015 # International Osteoporosis Foundation and National Osteoporosis Foundation 2015
Abstract Summary We have validated our touch-screen-modified FRAX® tool against the traditional healthcare professionalled questionnaire, demonstrating strong concordance between doctor- and patient-derived results. We will use this in outpatient clinics and general practice to increase our capture rate of at-risk patients, making valuable use of otherwise wasted patient waiting times. Introduction Outpatient clinics offer an opportunity to collect valuable health information from a captive population. We have previously developed a modified fracture risk assessment (FRAX®) tool, enabling patients to self-assess their osteoporotic fracture risk in a touch-screen computer format and demonstrated its acceptability with patients. We aim to validate the accuracy of our tool against the traditional questionnaire.
Methods Fifty patients over 50 years of age within the fracture clinic independently completed a paper equivalent of our touch-screen-modified FRAX® questionnaire. Responses were analysed against the traditional healthcare professional (HCP)-led questionnaire which was carried out afterwards. Correlation was assessed by sensitivity, specificity, Cohen’s kappa statistic and Fisher’s exact test for each potential FRAX® outcome of Btreat^, Bmeasure BMD^ and Blifestyle advice^. Results Age range was 51–98 years. The FRAX® tool was completed by 88 % of patients; six patients lacked confidence in estimating either their height or weight. Following question adjustment according to patient response and feedback, our tool achieved >95 % sensitivity and specificity for the Btreat^ and Blifestyle advice^ groups, and 79 % sensitivity and 100 % specificity in the Bmeasure BMD^ group. Cohen’s kappa
* Simon Parker
[email protected]
2
Trauma and Orthopaedics, Barnet Hospital, Wellhouse Ln, Barnet, Hertfordshire EN5 3DJ, UK
3
Royal Free London NHS Foundation Trust, London, UK
4
Royal National Orthopaedic Hospital, Brockley Hill, Stanmore, Middlesex HA7 4LP, UK
5
Wrenbury Medical Practice, The Surgery, The Green, Wrenbury, Nantwich, Cheshire CW5 8EW, UK
6
Royal College of General Practitioners, London, UK
7
Bridges Medical Centre, Wassand Close, Three Bridges, Crawley, West Sussex RH10 1LL, UK
Maria Ciaccio
[email protected] Erica Cook
[email protected] Graham Davenport
[email protected] Alun Cooper
[email protected] Simon Grange
[email protected]
1
8
National Osteoporosis Society, Camerton, UK
Peter Smitham
[email protected]
9
Buckinghamshire Healthcare NHS Foundation Trust, Flat 19, 3 St Pancras Way, London NW1 0PB, UK
Trauma and Orthopaedics, Ealing Hospital, Uxbridge Rd, Southall, Middlesex UB1 3HW, UK
10
University of Southampton, Southampton, UK
25 Page 2 of 8
value ranged from 0.823 to 0.995 across all groups, demonstrating Bvery good^ agreement for all. Fisher’s exact test demonstrated significant concordance between doctor and patient decisions. Discussion Our modified tool provides a simple, accurate and reliable method for patients to self-report their own FRAX® score outside the clinical contact period, thus releasing the HCP from the time required to complete the questionnaire and potentially increasing our capture rate of at-risk patients. Keywords Osteoporosis . Osteoporotic fractures . World Health Organisation . Outpatients . Ambulatory care . Efficiency . Self-assessment
Arch Osteoporos (2015) 10: 25
enables patient self-reporting and have previously demonstrated their ability to use such a system effectively, including the elderly population [4, 5]. This follows a change in culture within General Practice centres where patients now regularly use touch-screens to Bcheck-in^ prior to their consultation. In the present study, we aim to validate our tool against the gold standard HCP-led questionnaire. Successful validation will enable us to incorporate it into our touch-screen clinic registration system, so that patients may ascertain their own FRAX® score prior to their consultation, thus improving overall efficiency. We highlight the opportunity for gaining clinically relevant patient data in an effective and efficient manner, which utilises valuable time currently lost in many outpatient clinics.
Introduction Methods Across all specialties, a significant amount of a doctor’s working week is dedicated to outpatient clinics. Yet, despite the importance of this service in overall patient care, it is often viewed as an area of inefficiency, where patients can expect to wait significantly longer than their 10-min allocated appointment in order to be seen, and longer still for any investigation and management. A contributing factor is time spent by doctors addressing administrative tasks, rather than consulting patients [1]. It is a common cause of patient frustration [2]. Waiting time could be used to gain valuable patient health information or determine outcome measures on a condition through patient-reported outcome measurement tools. One area that should always be targeted within a fracture clinic is that of future fracture risk assessment and prevention. The World Health Organisation (WHO) Fracture Risk Assessment Tool (FRAX®) is a validated questionnaire based on individual patient models that integrates clinical risk factors to produce a 10-year probability of hip fracture (HF) and major osteoporotic fracture (MOF) [3]. This is traditionally a questionnaire that would be completed by a Healthcare Professional (HCP) with the patient present, and the answers transferred to a computer programme to perform a risk calculation. In our clinic, this may be carried out by a specialist nurse prior to doctor consultation or indeed within the doctor consultation itself. Alternatively, at-risk patients may be highlighted in a letter to the General Practitioner, to be assessed and followed up in the community. This is a timeand resource-consuming process. New formats such as the FRAX® Pad have been developed (available at www. iofbonehealth.org) to enable the capture of risk factor variables prior to consultation, but this still requires input onto computer software. Furthermore, no studies have been carried out assessing the reliability or validity of patientreported data using such a tool for self-assessment. In order to address this issue, we have developed a modified FRAX® tool formatted for touch-screen computers which
Following registration with Barnet R&D department, 50 patients over 50 years of age within the fracture clinic were randomly selected in order of arrival and consented to involvement in the study. These patients had all been referred to the fracture clinic for having a suspected fracture whether this was a high energy or a low energy fracture. Population number was chosen following a power calculation which determined that to obtain 80 % power to detect a kappa of 0.4, a total sample size of 39 patients would be needed, two-tailed. The proportion of positive ratings was set at 0.9. This cohort could generally be considered a high-risk population, as patients are taken from a standard district general hospital fracture clinic, where the majority will be present for fracture management. Each patient independently completed a paper equivalent of our modified FRAX® questionnaire (Fig. 1). They were then asked to complete the traditional questionnaire with the assistance of an HCP (doctor or specialist nurse). Answers from our modified FRAX® tool were collated by the HCP and used to produce a FRAX® score, as would be carried out automatically on our touch-screen computer system. Results were then analysed against those produced by the traditional tool. The questionnaires were always carried out in this order so that the traditional HCP-led questionnaire had no influence on patient response to our modified tool. Data was collated using Microsoft Excel© software. Statistical methods The FRAX® tool provides one of three management advice outcomes based on the calculated 10-year probability of major osteoporotic fracture: Blifestyle advice only^ for low risk Fig. 1 The paper equivalent of our modified FRAX® tool. These questions are the same as those that appear in our user-friendly touchscreen computer software
Arch Osteoporos (2015) 10: 25
Page 3 of 8 25
Osteoporosis is a disease of the bones that leads to an increased risk of fracture due to the thinning of the bone. Osteoporosis can affect both men and women and is very common in people over 50. Answering these questions may help to assess your risk of breaking a bone due to osteoporosis. You will have the opportunity to have a copy of your answers at the end of the questionnaire. Personal Information Gender Male Female How old are you? Under 54yrs 55- 59yrs 60-64yrs 65-69yrs 70-74yrs 75-79yrs 80–84yrs 85-89yrs Over 90 yrs What is your weight (in Kg)? _________ What is your height (in cm)? _________ Do you drink alcohol? No 1.
Yes How many units per day? ______
Glass of wine or pint of lager is approximately 2 units
2.
Large glass of wine or pint of strong lager is approximately 3 units
3.
Pint of strong cider is approximately 5 units
Do you smoke? No
Yes
How many years? _____ How many per day? _____
Medical History Have you broken a bone in the last 20 years? Yes No don’t know Did either of your parents break a hip? Yes No don’t know Have you been diagnosed with Rheumatoid Arthritis? Yes No don’t know Have you been diagnosed with Osteoporosis? Yes No don’t know Have you been diagnosed with Secondary Osteoporosis or Insulin Dependent diabetes? Yes No don’t know Do you feel you are getting shorter or have lost more than 5 centimetres (2 inches) in height? Yes No don’t know Medications Are you currently taking, or has your doctor ever prescribed you prednisolone or other steroids? Yes No don’t know Do you currently take calcium supplements? Yes No don’t know Do you currently take a bisphosphonate? E.g. Fosamax or Actonel. Yes No don’t know
25 Page 4 of 8
patients, Bmeasure bone mineral density (BMD)^ for those of moderate risk, and Btreat^ for those of high risk. For each outcome, the sensitivity, specificity, positive predictive value and negative predictive value of our modified FRAX® tool was assessed, as compared to the gold standard HCP-led questionnaire. Cohen’s kappa statistics were then used to measure the degree of agreement between the two tools for each outcome, whereby agreement due to chance is factored out [6]. The strength of agreement is defined as Bpoor^ (κ=90 %. The only measure to be below 80 % was sensitivity of Bmeasure BMD^, at 79 %. Cohen’s kappa was then used to assess the degree of agreement between our modified FRAX® tool and the HCP-led questionnaire (Table 1). With corrected responses our modified tool has achieved the highest level of agreement with the traditional tool, with a kappa value in the Bvery good^ strength of agreement range for each potential outcome. The distributions of patient and doctor decisions were compared using SPSS Crosstabs procedure. As there were many cells with low expected values, a Fisher’s exact test was computed (based on the hypergeometric distribution). This revealed that the distributions of patient and doctor decisions were very significantly associated (p
