1-s2.0-S0958259214000078-main
3
The Foot 24 (2014) 28–30 Contents lists available at ScienceDirect The Foot journal h om epage: www.elsevier.com/locat e/foot An evaluation of retrospective outcome scores in elective foot and ankle surgery James Widnall ∗ , Peter Ralte, Dave Selvan, Andy Molloy University Hospital Aintree, Lower Lane, Liverpool, Merseyside L9 7AL, UK a r t i c l e i n f o Article history: Received 10 February 2014 Accepted 28 February 2014 Keywords: Outcome Retrospective Research Score a b s t r a c t Background: Patient reported outcome measures are becoming more popular in their use. Retrospective scoring is not yet a validated method of data collection but one that could greatly decrease the complexity of research projects. We aim to compare preoperative and retrospective scores in order to assess their correlation and accuracy. Methods: 36 patients underwent elective foot and ankle surgery. All patients were scored preoperatively using the SF-12 (including both the physical and mental subsets) and FFI. Patients then recorded both PROMs at the 3-month follow up (av. 139 days). Results were then analyzed for statistical significance. Results: 36 patients (av. age 54.6): completed both sets of questionnaires. There were 15 hindfoot and 21 forefoot procedures. No retrospective scores were identical. The mean percentage difference between the preoperative scores was −7.9% (−17.3 to 1.6%, 95% CI) for Physical Component of SF12, −3.2% (−10.3 to 3.9%, 95% CI) for mental component of SF12 and 40.7% (25.3 to 56.1%, 95% CI) for FFI. This retrospective accuracy was statistically significant (p < 0.001). When the scores were plotted against each other, the outcome measurements showed positive correlations (Physical SF 12 p = 0.48, Mental SF 12 p = 0.80 and FFI p = 0.81). With both PROMs mean percentage differences combined, patients undergoing hindfoot procedures (3.5%; −5.0 to 12.1%, 95% CI) were more accurate with retrospective scoring than their forefoot counter- parts (17.5%; 5.0 to 30.0%, 95% CI). This was not statistically significant (p = 0.07). Using regression analysis, we found no significant statistical difference in the retrospective accuracy when compared against both time to retrospective scoring and the outcome measure at 3 months post operatively. Conclusion: Retrospective scoring appears to lack accuracy when compared to prospective methods. How- ever, our data shows the SF12 is recalled more accurately than the FFI (p < 0.001) and both the mental and physical components are recalled to within 10% of the pre-operative score. These results show patients tend to recall their symptoms at a worse level preoperatively than originally described, especially those with forefoot problems. © 2014 Elsevier Ltd. All rights reserved. 1. Introduction Patient reported outcome measures are a popular method of quantifying patient satisfaction following surgical intervention. Routinely patients are scored in the pre-operative period and sub- sequently at certain time periods following surgery. These tools then allow analysis to quantify patient satisfaction but can also lend themselves to research purposes. ∗ Corresponding author at: University Hospital Aintree, Lower Lane, Liverpool, Merseyside, L9 7AL, UK. Tel.: +44 07812755152. E-mail address: [email protected] (J. Widnall). Commonly, in the setting of foot and ankle surgery there are a multitude of scoring systems used as there is yet to be consensus on the most appropriate scoring tool. In our study we have analyzed both the Short Form-12 (SF12) Health Survey – a shorter version of the SF36 with proven correlation [1] – and the Foot Function Index (FFI). They have been quoted, in their various forms, to be used in 13.7% and 5.5% of all foot and ankle research, second only to the American Orthopaedic Foot & Ankle Society (AOFAS) scales and visual analogue scale (VAS) [2]. The SF12 is a generic health scale, not designed purposefully for application to foot and ankle surgery. Its use enables clinicians to assess the perceived impact of the condition in question on the patients overall health status. It is a subset of items from the SF-36 and produces both a physical and mental score [3]. http://dx.doi.org/10.1016/j.foot.2014.02.005 0958-2592/© 2014 Elsevier Ltd. All rights reserved.
-
Upload
androide007 -
Category
Documents
-
view
4 -
download
0
description
Foot Scoring System
Transcript of 1-s2.0-S0958259214000078-main
-
Aa
JU
a
ARA
KORRS
1
qRstt
M
h0
The Foot 24 (2014) 2830
Contents lists available at ScienceDirect
The Foot
journa l h om epage: www.elsev ier .com/ locat e/ foot
n evaluation of retrospective outcome scores in elective foot andnkle surgery
ames Widnall , Peter Ralte, Dave Selvan, Andy Molloyniversity Hospital Aintree, Lower Lane, Liverpool, Merseyside L9 7AL, UK
r t i c l e i n f o
rticle history:eceived 10 February 2014ccepted 28 February 2014
eywords:utcomeetrospectiveesearchcore
a b s t r a c t
Background: Patient reported outcome measures are becoming more popular in their use. Retrospectivescoring is not yet a validated method of data collection but one that could greatly decrease the complexityof research projects. We aim to compare preoperative and retrospective scores in order to assess theircorrelation and accuracy.Methods: 36 patients underwent elective foot and ankle surgery. All patients were scored preoperativelyusing the SF-12 (including both the physical and mental subsets) and FFI. Patients then recorded bothPROMs at the 3-month follow up (av. 139 days). Results were then analyzed for statistical significance.Results: 36 patients (av. age 54.6): completed both sets of questionnaires. There were 15 hindfoot and 21forefoot procedures. No retrospective scores were identical.
The mean percentage difference between the preoperative scores was 7.9% (17.3 to 1.6%, 95% CI)for Physical Component of SF12, 3.2% (10.3 to 3.9%, 95% CI) for mental component of SF12 and 40.7%(25.3 to 56.1%, 95% CI) for FFI. This retrospective accuracy was statistically significant (p < 0.001). Whenthe scores were plotted against each other, the outcome measurements showed positive correlations(Physical SF 12 p = 0.48, Mental SF 12 p = 0.80 and FFI p = 0.81).
With both PROMs mean percentage differences combined, patients undergoing hindfoot procedures(3.5%; 5.0 to 12.1%, 95% CI) were more accurate with retrospective scoring than their forefoot counter-parts (17.5%; 5.0 to 30.0%, 95% CI). This was not statistically significant (p = 0.07).
Using regression analysis, we found no significant statistical difference in the retrospective accuracywhen compared against both time to retrospective scoring and the outcome measure at 3 months postoperatively.
Conclusion: Retrospective scoring appears to lack accuracy when compared to prospective methods. How-ever, our data shows the SF12 is recalled more accurately than the FFI (p < 0.001) and both the mental andphysical components are recalled to within 10% of the pre-operative score. These results show patientstend to recall their symptoms at a worse level preoperatively than originally described, especially thosewith forefoot problems.. Introduction
Patient reported outcome measures are a popular method ofuantifying patient satisfaction following surgical intervention.outinely patients are scored in the pre-operative period and sub-equently at certain time periods following surgery. These tools
hen allow analysis to quantify patient satisfaction but can also lendhemselves to research purposes.
Corresponding author at: University Hospital Aintree, Lower Lane, Liverpool,erseyside, L9 7AL, UK. Tel.: +44 07812755152.E-mail address: [email protected] (J. Widnall).
ttp://dx.doi.org/10.1016/j.foot.2014.02.005958-2592/ 2014 Elsevier Ltd. All rights reserved. 2014 Elsevier Ltd. All rights reserved.
Commonly, in the setting of foot and ankle surgery there are amultitude of scoring systems used as there is yet to be consensus onthe most appropriate scoring tool. In our study we have analyzedboth the Short Form-12 (SF12) Health Survey a shorter versionof the SF36 with proven correlation [1] and the Foot FunctionIndex (FFI). They have been quoted, in their various forms, to beused in 13.7% and 5.5% of all foot and ankle research, second onlyto the American Orthopaedic Foot & Ankle Society (AOFAS) scalesand visual analogue scale (VAS) [2].
The SF12 is a generic health scale, not designed purposefully
for application to foot and ankle surgery. Its use enables cliniciansto assess the perceived impact of the condition in question on thepatients overall health status. It is a subset of items from the SF-36and produces both a physical and mental score [3].
dx.doi.org/10.1016/j.foot.2014.02.005http://www.sciencedirect.com/science/journal/09582592http://www.elsevier.com/locate/foothttp://crossmark.crossref.org/dialog/?doi=10.1016/j.foot.2014.02.005&domain=pdfmailto:[email protected]/10.1016/j.foot.2014.02.005
-
J. Widnall et al. / The Foot 24 (2014) 2830 29
-10
0
10
20
30
40
50
FFI SF12 (physical) SF12 (mental)
Mean % differe nce fr om preopera tive to retrospective scores
Mean % diff erenc e
Fs
fos
trc
2
ebFfahfsah
3
om9
ot1s
aosis
eprc
o
0
10
20
30
40
50
60
0 10 20 30 40 50 60
ero csd ellac eRPre operative
Physical SF12 Regression Plot
0
50
100
150
200
250
0 50 10 0 15 0 20 0 25 0
er ocsd el lace R
Pre Operative
FFI Regression Plot
0
10
20
30
40
50
60
70
0 10 20 30 40 50 60 70
Retr
opec
tive s
core
Pre operative scor e
Mental SF12 Regression Plot
Fig. 2. Regression plots for physical and mental components of the SF12 and theFFI. The lines represent perfect correlation.
Table 1Analyzing the impact of both time to retrospective scoring and the 3-month outcomescore on retrospective accuracy. No statistically significant relationship was found.
PROM % difference Variable Correlation (r)
SF 12 Time to retrospective score 0.303 month post op score 0.10ig. 1. Mean percentage difference between preoperative and recalled outcomecores for physical and mental components of the SF12 and the FFI.
The FFI was originally validated on a patient population sufferingrom rheumatoid arthritis and it reveals the impact of foot pathol-gy on function. This is calculated by the use of 23 visual analoguecales relating to pain, difficulties and limitations [4,5].
While outcome scores are commonly used prospectively we aimo assess their potential for retrospective use, comparing the cor-elation and accuracy of both the SF12 and FFI. This method of dataollection could greatly simplify future research.
. Methods
From November 2012 to March 2013, 59 patients underwentlective foot and ankle procedures at our centre. All patients hadeen scored pre-operatively at the time of listing by the SF12 andFI. 23 patients had either declined to continue or were lost toollow up by the 3-month stage. 36 patients (26 female, 10 male,verage age 54.6 years) had therefore been scored pre-operatively,ad their procedure and completed both PROMs at the 3-monthollow up. At the time of follow up, the patients were asked both tocore themselves as they remembered they were pre-operativelynd also as at that stage postoperatively. There were 15 (41.7%)indfoot and 21 (58.3%) forefoot procedures.
. Results
Average follow up was 137 (range 73262) days. The mean pre-perative FFI score was 103.1 points. Recalled at 3 months, theean retrospective score was 135.1 points, a 40.7% (25.3 to 56.1%,5% CI) increase in symptom level.With respect to the SF12, the average physical component pre-
perative score dropped by 3.8 points from 37.4 pre-operativelyo 33.6 retrospectively. This is the equivalent of a 7.9% (17.3 to.6%, 95% CI) drop in score, again akin to an increase in reporting ofymptoms.
The mental component was recalled more accurately with anverage 3.2% (10.3 to 3.9%, 95% CI) change in value from pre-perative to retrospective score (49.447.0 points). This is againynonymous with an increase in symptom level. This differencen accuracy of recall between the SF12 and FFI was statisticallyignificant (p < 0.001, Fig. 1).
With the preoperative and retrospective scores plotted againstach other to show correlation the FFI returned a coefficient
= 0.81. The physical component of SF12 showed moderate cor-
elation, p = 0.48 and the mental component similar to the strongorrelation of the FFI, p = 0.80 (Fig. 2).
Performing regression analysis allowed evaluation of the impactf both time to retrospective scoring and the actual 3-month postFFI Time to retrospective score 0.053 month post op score 0.28
operative score on the retrospective accuracy. The SF12 scores werecombined into one to enable ease of statistical testing. There was nostatistical relationship of note between the retrospective accuracyof the SF12 or FFI and either of the confounders in question (Table 1).
With the PROMS combined, analysis of the patients could besplit into forefoot and hindfoot cohorts. The forefoot cohort hadretrospective scores within 8.8% and 4.2% error of their originalphysical and mental SF12 scores respectively. The FFI was recalled
less accurately, to within 55% of the preoperative value on average.The hindfoot patients showed greater accuracy with the physical SF12 component (average retrospective error 6.6%) but less accuracy
-
30 J. Widnall et al. / The Foo
-8.8
4.2
53
18
-6.6
-13.7
25.2
6.2
-20
-10
0
10
20
30
40
50
60
Physica l SF12 Mental SF1 2 FFI Comb ined
%
Forefoot vs Hindfoot accuracy
Forefoo tHindfoo t
Fa
wth
f6w
4
adSmataotua
[
[
[
[
[
[
ig. 3. Bar chart comparing retrospective scoring accuracy between the hindfootnd forefoot patient cohorts.
ith the mental component (average error 13.7%). With respecto FFI the hindfoot patients recalled their scores on average 25.2%igher than their actual preoperative score.With all PROMS combined the retrospective accuracy of the
orefoot group was less then that of the hindfoot group (18.0 and.2%). This approached but did not reach statistical significanceith p = 0.08 (Fig. 3).
. Discussion
Our study shows that retrospective symptom scoring is nots accurate as prospective methods. However, there is significantifference in accuracy between the outcome measurement tools.ymptoms recalled via the SF12 are replicated with significantlyore accuracy than those using the FFI tool (p < 0.001). This is prob-bly due to the fewer elements within the SF12 when comparedo the FFI creating less opportunity for error. Both the physicalnd mental components are recalled, on average, to within 10%
f the prospective score. It is up to the clinician to determine ifhis magnitude of error is acceptable. Interestingly, retrospectivese of both tools produced results consistent with worse preoper-tive symptoms than previously recorded. This phenomenon has
[
t 24 (2014) 2830
also been seen with the AOFAS scores in two previous papers [6,7].This exaggeration of symptoms would need to be taken account ofif retrospective scoring was ever used for research purposes.
Neither the time elapsed between obtaining the two scores orthe patients current state had any effect on the retrospective accu-racy. This was the case for both the FFI and SF12.
To the best of our knowledge, accuracy between forefoot andhind foot patients has not been previously analyzed. We observedthat those patients with forefoot procedures were more likely toexacerbate the symptoms when recalling them than their hindfootcounterparts. This was not statistically significant, however, this isunsurprising given the small numbers involved in this study.
This study shows retrospective scoring will lack accuracy com-pared to prospective data collection. A margin of error of roughly10% can be expected with retrospective methods. This means thoseprojects using retrospective scores are likely to exaggerate theimpact of the intervention in question. Further studies on other out-come reporting tools are needed to evaluate retrospective scoringand its suitability for research.
Conflict of interest
The authors have no conflict of interest to declare.
References
1] Jenkinson C, Layte R. Development and testing of the UK SF-12 (short form healthsurvey). J Health Serv Res Policy 1997;2(1):148.
2] Hunt KJ, Hurwit D. Use of patient reported outcome measures in foot and ankleresearch. J Bone Joint Surg Am 2013;95(16):e118(1-9).
3] Ware JE, Kosinski M, Keller SD. A 12-Item Short-Form Health Survey: con-struction of scales and preliminary tests of reliability and validity. Med Care1996;34(3):22033.
4] Budiman-Mak E, Conrad K, Roach K. The foot function index; a measure of footpain and disability. J Clin Epidemiol 1991;144(6):56170.
5] Budiman-Mak E, Conrad K, Stuck R, Matter M. Theoretical model and Rasch anal-ysis to develop a revised Foot Function Index. Foot Ankle Int 2006;27(7):51927.
6] Toolan B, et al. An evaluation of the use of retrospectively acquired preoperative
AOFAS clinical rating scores to assess surgical outcome after elective foot andankle surgery. Foot Ankle Int 2001;22(10):7758.
7] Schnieder W, Knahr K. Poor agreement between prospective and retrospec-tive assessment of hallux surgery using the AOFAS hallux scale. Foot Ankle Int2005;26(12):10626.
http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0005http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0005http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0005http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0005http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0005http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0005http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0005http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0005http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0005http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0005http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0005http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0005http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0005http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0005http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0005http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0005http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0005http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0005http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0005http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0005http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0005http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0005http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0005http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0010http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0010http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0010http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0010http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0010http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0010http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0010http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0010http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0010http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0010http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0010http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0010http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0010http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0010http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0010http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0010http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0010http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0010http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0010http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0010http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0010http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0015http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0015http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0015http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0015http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0015http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0015http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0015http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0015http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0015http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0015http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0015http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0015http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0015http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0015http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0015http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0015http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0015http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0015http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0015http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0015http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0015http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0015http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0015http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0015http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0015http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0015http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0015http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0020http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0020http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0020http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0020http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0020http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0020http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0020http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0020http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0020http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0020http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0020http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0020http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0020http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0020http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0020http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0020http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0020http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0020http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0020http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0020http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0020http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0020http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0020http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0025http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0025http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0025http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0025http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0025http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0025http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0025http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0025http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0025http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0025http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0025http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0025http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0025http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0025http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0025http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0025http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0025http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0025http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0025http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0025http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0025http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0025http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0025http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0025http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0025http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0025http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0025http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0030http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0030http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0030http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0030http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0030http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0030http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0030http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0030http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0030http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0030http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0030http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0030http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0030http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0030http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0030http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0030http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0030http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0030http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0030http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0030http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0030http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0030http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0030http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0030http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0030http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0030http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0030http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0030http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0030http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0030http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0030http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0030http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0030http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0035http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0035http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0035http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0035http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0035http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0035http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0035http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0035http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0035http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0035http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0035http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0035http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0035http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0035http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0035http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0035http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0035http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0035http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0035http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0035http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0035http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0035http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0035http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0035http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0035http://refhub.elsevier.com/S0958-2592(14)00007-8/sbref0035
An evaluation of retrospective outcome scores in elective foot and ankle surgery1 Introduction2 Methods3 Results4 DiscussionConflict of interestReferences
