The classification of long bone osteomyelitis
<p>This thesis investigates the classification of long bone osteomyelitis in adults and aims to develop a new classification system for the guidance of management and prognosis.</p> <p>Osteomyelitis can be complex and carry a significant burden for patients and healthcare services...
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Thesis |
| Language: | English |
| Published: |
2018
|
| _version_ | 1797105378770550784 |
|---|---|
| author | Hotchen, AJ |
| author2 | McNally, MA |
| author_facet | McNally, MA Hotchen, AJ |
| author_sort | Hotchen, AJ |
| collection | OXFORD |
| description | <p>This thesis investigates the classification of long bone osteomyelitis in adults
and aims to develop a new classification system for the guidance of management
and prognosis.</p>
<p>Osteomyelitis can be complex and carry a significant burden for patients and
healthcare services. Recent developments in treatment options and the multidisciplinary approach to management have resulted in improved outcomes over
the last 40 years. However, currently there is no evidence based or validated
method of stratifying these patients into those that require specialist management
or defining disease prognosis.</p>
<p>Systematic review of the literature highlighted 13 classification systems that have
been presented in the medical literature ranging from 1970 to 2015. Detailed analysis
highlighted the important variables that may be included in a new classification
system. Combining these with the multi-disciplinary approach to management, the
four most important variables were presented. These were: the bone involvement,
the microbiology (anti-microbial options), the coverage of the soft tissues and the
host status. This was incorporated into the acronym of BACH.</p>
<p>Version 1 of the BACH classification system was applied retrospectively to 96
cases of long bone osteomyelitis. The classification successfully correlated with
clinical and patient reported outcome measures, demonstrating that a higher BACH
score increases the chance of a poor prognosis. However, there were limitations
demonstrated in the application of the classification system. These issues related
to the use of certain terminology, the application of the ESCMID criteria for antimicrobial resistance patterns and the definition of host status. Further analysis of
the microbiology demonstrated that isolates with ≥4 resistant or <80% sensitive
susceptibility tests had a 96.6%-100% likelihood of being multi-drug resistant.
Version 2 of the BACH classification system was assessed using an inter-user
assessment which included 30 clinicians from around the world who were asked to
classify 20 cases of long bone osteomyelitis. Users consisted of orthopaedic surgeons,
plastic surgeons, infectious disease physicians, microbiologists and anaesthetists.</p>
<p>The overall accuracy of BACH classification was 86.2% (SD6.2% [95% CI 83.9-
88.6%]). The anti-microbial options variable scored the highest Fleiss’ κ (F κ) (0.815
[95% CI 0.811 - 0.819], almost perfect agreement) but as the complexity of the isolate
increased, the classification became less accurate. It was hypothesised that this
was due to the application of the ESCMID criteria. The bone involvement variable
had the lowest accuracy (77.0% [95% CI 71.2-82.8%]) and agreement amongst users
(0.479 [95% CI 0.475-0.483], fair agreement). This was thought to be secondary
to the presentation of the CT and MRI series as single slices, thus not allowing
a 3-dimensional visualisation to the user. Further assessment of this variable was
conducted by asking 9 users to classify the bone involvement variable using a
webPACS as in real-world clinical practice. This demonstrated a significantly higher
accuracy (p<0.01, <em>ANOVA with Dunnett’s post hoc test</em>) in the group who used
the webPACS compared to those who did not.</p>
<p>On the basis of these assessments, version 3 of the classification was presented.
To assess its application, use in management and ability to offer prognosis, a
prospective analysis has been planned. In this assessment, three end points will
be assessed. These are (i) the ability of BACH to guide need for referral, (ii) the
application of individual variables and (iii) the ability of BACH to offer prognosis
in terms of clinical and health related outcome measures. A pilot of 40 cases is
included to assess the feasibility of implementation.</p>
<p>The BACH classification has been developed using a number of analyses that
are presented as part of this thesis. The development of the BACH classification
system has led to a simple, non-invasive intervention that can be used in clinical
practice and ultimately improve the care that is offered to patients who have
long bone osteomyelitis.</p> |
| first_indexed | 2024-03-07T06:46:41Z |
| format | Thesis |
| id | oxford-uuid:fb15bf49-4636-471b-9ccd-cab8f9f32dd2 |
| institution | University of Oxford |
| language | English |
| last_indexed | 2024-03-07T06:46:41Z |
| publishDate | 2018 |
| record_format | dspace |
| spelling | oxford-uuid:fb15bf49-4636-471b-9ccd-cab8f9f32dd22022-03-27T13:11:13ZThe classification of long bone osteomyelitisThesishttp://purl.org/coar/resource_type/c_db06uuid:fb15bf49-4636-471b-9ccd-cab8f9f32dd2EnglishORA Deposit2018Hotchen, AJMcNally, MASendi, P<p>This thesis investigates the classification of long bone osteomyelitis in adults and aims to develop a new classification system for the guidance of management and prognosis.</p> <p>Osteomyelitis can be complex and carry a significant burden for patients and healthcare services. Recent developments in treatment options and the multidisciplinary approach to management have resulted in improved outcomes over the last 40 years. However, currently there is no evidence based or validated method of stratifying these patients into those that require specialist management or defining disease prognosis.</p> <p>Systematic review of the literature highlighted 13 classification systems that have been presented in the medical literature ranging from 1970 to 2015. Detailed analysis highlighted the important variables that may be included in a new classification system. Combining these with the multi-disciplinary approach to management, the four most important variables were presented. These were: the bone involvement, the microbiology (anti-microbial options), the coverage of the soft tissues and the host status. This was incorporated into the acronym of BACH.</p> <p>Version 1 of the BACH classification system was applied retrospectively to 96 cases of long bone osteomyelitis. The classification successfully correlated with clinical and patient reported outcome measures, demonstrating that a higher BACH score increases the chance of a poor prognosis. However, there were limitations demonstrated in the application of the classification system. These issues related to the use of certain terminology, the application of the ESCMID criteria for antimicrobial resistance patterns and the definition of host status. Further analysis of the microbiology demonstrated that isolates with ≥4 resistant or <80% sensitive susceptibility tests had a 96.6%-100% likelihood of being multi-drug resistant. Version 2 of the BACH classification system was assessed using an inter-user assessment which included 30 clinicians from around the world who were asked to classify 20 cases of long bone osteomyelitis. Users consisted of orthopaedic surgeons, plastic surgeons, infectious disease physicians, microbiologists and anaesthetists.</p> <p>The overall accuracy of BACH classification was 86.2% (SD6.2% [95% CI 83.9- 88.6%]). The anti-microbial options variable scored the highest Fleiss’ κ (F κ) (0.815 [95% CI 0.811 - 0.819], almost perfect agreement) but as the complexity of the isolate increased, the classification became less accurate. It was hypothesised that this was due to the application of the ESCMID criteria. The bone involvement variable had the lowest accuracy (77.0% [95% CI 71.2-82.8%]) and agreement amongst users (0.479 [95% CI 0.475-0.483], fair agreement). This was thought to be secondary to the presentation of the CT and MRI series as single slices, thus not allowing a 3-dimensional visualisation to the user. Further assessment of this variable was conducted by asking 9 users to classify the bone involvement variable using a webPACS as in real-world clinical practice. This demonstrated a significantly higher accuracy (p<0.01, <em>ANOVA with Dunnett’s post hoc test</em>) in the group who used the webPACS compared to those who did not.</p> <p>On the basis of these assessments, version 3 of the classification was presented. To assess its application, use in management and ability to offer prognosis, a prospective analysis has been planned. In this assessment, three end points will be assessed. These are (i) the ability of BACH to guide need for referral, (ii) the application of individual variables and (iii) the ability of BACH to offer prognosis in terms of clinical and health related outcome measures. A pilot of 40 cases is included to assess the feasibility of implementation.</p> <p>The BACH classification has been developed using a number of analyses that are presented as part of this thesis. The development of the BACH classification system has led to a simple, non-invasive intervention that can be used in clinical practice and ultimately improve the care that is offered to patients who have long bone osteomyelitis.</p> |
| spellingShingle | Hotchen, AJ The classification of long bone osteomyelitis |
| title | The classification of long bone osteomyelitis |
| title_full | The classification of long bone osteomyelitis |
| title_fullStr | The classification of long bone osteomyelitis |
| title_full_unstemmed | The classification of long bone osteomyelitis |
| title_short | The classification of long bone osteomyelitis |
| title_sort | classification of long bone osteomyelitis |
| work_keys_str_mv | AT hotchenaj theclassificationoflongboneosteomyelitis AT hotchenaj classificationoflongboneosteomyelitis |