In this dissertation, we embody an information-theoretic framework to compress and therefore cluster anatomical connectivity data that avoids many assumptions and drawbacks imposed by previous methods.

Total joint replacement was pioneered by John Charnley in the late 1950's, and has since revolutionised the management of arthritis sufferers. By 1991, an estimated 5 million people had undergone hip replacements. Although relatively successful, the cemented components had some problems, and this led to the development of cementless implants. These implants depend on the ingrowth of bone into a porous coating, to produce a durable method of implant fixation which the normal bone turnover process will maintain. One of the problems with cementless implants is that the type and extent of tissue ingrowth into the porous coating is unpredictable. Movement of the implant relative to the surrounding bone may result in the formation of an interfacial fibrous tissue layer. Hence, numerical modelling has been used to predict tissue ingrowth into such implants. Numerical simulation has the advantage that comprehensive data can be extracted relatively quickly. The finite element method is a powerful tool that has become the preferred method of analysis, and takes into account critical factors such as implant design, bone properties, and loading conditions. However, these models have not been tested extensively. Little attention has been given to comparing numerical models with the actual findings of retrieval studies or radiological imaging studies. This study thus evaluates the potential of one such numerical model. Most numerical models analyse the stress patterns of a particular state of bone ingrowth (i.e. a static case). This model considered the development of the ingrowing material - a dynamic analysis of tissue changes over a period of time. A 2-dimensional, plane stress finite element model was used to predict the ingrowth of bone into the porous coating of the femoral stem of a hip implant. A side plate was incorporated to mimic 3-dimensional characteristics. The evaluation was achieved by comparing the predictions of the numerical model with plane X-ray images of seven patients with Zimmer Anatomic cementless hip implants. The X-rays were scanned at a high resolution, so as to be able to "magnify" the regions to be examined. Several algorithms were developed to analyse the images, and provide a quantitative assessment of the X-ray images. The algorithms were designed to identify regions of bony and fibrous tissue. This involved the identification of the interface between the implant and the surrounding bone, and the extraction of the grayscale values of the X-rays at this interface. Thereafter, various radiographic signs that indicate the presence of fibrous tissue or bony tissue were identified, and these were used to enhance the original grayscale plot. The resulting graph was then modified slightly so as to make its presentation comparable with the numerical model. Plane X-rays proved to be suitable for the task of identifying tissue types. These data were then compared with the predictions of the numerical model. A qualitative correlation was used, as this was deemed to be most appropriate. Several authors in the literature also found a quantitative approach to have limitations. Some agreement between the experimental findings and the numerical simulation was found to exist, although this was limited. The agreement was judged to be less than the "reasonable agreement'' that several studies in the literature concluded. The correlation is better described by "some agreement". Nevertheless, the finite element method was assessed as being a tool with great potential, and modifications to the present model may provide more reliable results. A time study was also undertaken, whereby the tissue density was evaluated at various periods after the operation. The study provided insight into the evolution of the implant-bone interface after surgery, and correlated well with the literature. The phases of repair and remodelling were evident, and it was assessed as being a valuable contribution to this work. The time study may prove to be a more useful method than those used in assessing the "static" images, and could even provide a prognostic tool in assessing implant stability over time.

Includes bibliographical references (leaves 157-168).

Background: Comparisons in morphological shape/form across population groups could provide population differences that might assist in making decisions on diagnosis and prognosis by the clinician. Geometric morphometrics (GM) is one of the fields that help to provide such population comparisons. In medical imaging and related disciplines, GM is commonly done using annotated landmarks or distances measured from 3D surfaces (consisting of triangular meshes). However, these landmarks may not be sufficient to describe the complete shape. This project aimed to develop GM for analysis that consider all vertices in the triangular mesh as landmarks. The developed methods were applied to South African and Swiss shoulder bones (scapula and humerus) to analyse morphological differences. Methods: The developed pipeline required first establishing correspondence across the datasets through a registration process. Gaussian process fitting was chosen to perform the registration since it is considered state-of-the-art. Secondly, a novel method for automatic identification of vertices or areas encoding the most shape/form variation was developed. Thirdly, a principal component analysis (PCA) that addressed the high dimensionality and lower sample size (HDLSS) phenomenon was adopted and applied to the dense correspondence data. This approach allowed for the stabilisation of the distribution of the data in low-dimensional form/shape space. Lastly, appropriate statistical tests were developed for population comparisons of the shoulder bones when dealing with HDLSS data in both form and shape space. Results: When the mesh-based GM analysis approach was applied to the training datasets (South African and Swiss shoulder bones), it was found that the anterior glenoid which is often the site of the shoulder dislocation is the most varied area of the glenoid. This has implications for diagnosis and provides knowledge for prosthesis design. The distribution of the data in the modified PCA space was shown to converge to a stable distribution when more vertices/landmarks are used for the analysis. South African and Swiss datasets were shown to be more distinguishable in a low-dimensional space when considering form rather than shape. It was found that left and right South African scapula bones are significantly different in terms of shape. Discussion: In general, it was observed that the two populations means can be significantly different in shape but not in form. An improved understanding of these observed shape and form differences has utility for shoulder arthroplasty prosthesis design and may also be useful for orthopaedic surgeons during surgical preoperative planning.

Includes bibliographical references. / TB is an airborne infectious disease which is spread by droplet nuclei, carrying Mycobacterium tuberculosis, in the air. The droplet nuclei small enough to enter human respiratory pathways are 1-5 μm in size and are able to travel long distances (Hodgson, et al., 2009) (WHO, 1999), and can be distributed widely throughout (hospital) buildings (Beggs, Noakes, Sleigh, Fletcher, & Siddiqi, 2003). These droplet nuclei may remain suspended in the air until they are removed by dilution ventilation or other disinfection methods (Parsons, Hussey, Abbott, & de Jager, 2008) (National Department of Health, 2007). Dilution ventilation refers to the dilution of contaminated air with “clean” air (ACGIH, 2005), thereby reducing the concentration of contaminants in the room. One of the recognised approaches for minimising the risk of transmission of TB is to adequately ventilate the contaminated room/space. A higher ventilation rate can provide higher dilution capability, in turn reducing the risk of airborne infections (WHO, 2009). The parameters of concern in ventilation design are ventilation flow rate and airflow pattern in the room (and building). The former reduces contaminant concentration while the latter aims to move uncontaminated air to high risk areas, and contaminated air away from occupied areas, usually to the outside. The shortcomings of conventional natural ventilation strategies are well documented. The aim of this research project is to review and study the effectiveness of natural ventilation design supplemented by a turbine ventilator. The project was divided into two components: a field study and laboratory experiments. In the field study, a turbine ventilator was installed into a bedroom of a low-income house in Pretoria. Tracer gas (concentration decay) tests were performed to determine the ventilation flow rates, mean age of air and air change efficiency of four natural ventilation configurations. These included infiltration/leakage (IL), two cases of single-sided ventilation (SS1 and SS2), and crossventilation (CV). Three baseline (without the turbine ventilator) and three turbine ventilator tests were performed, one each in the morning, noon and afternoon. The tests were performed between February and April 2011 on typical summer days. The turbine ventilator was then tested in a laboratory environment under wind, buoyancy and a combination of wind and buoyancy forces. The wind speeds were low, ranging from 0.0 to 0.5 m/s (0.0 to 1.8 km/h), and the temperature differential tested was in the range of 5.5 to 9.3˚C. The in-duct velocities and centreline velocities were investigated to establish if, under the subjected force(s), a capture envelope described by Dalla Valle’s equation could be measured. This envelope would be used to determine if the turbine ventilator could potentially reduce the concentration of airborne contaminants in the test volume. In the field study baseline tests, IL, SS1, CV and SS2 mean – and range of - ventilation flow rates of 0.6 [0.5 – 0.6], 8.1 [6.8 – 9.3], 16.9 [14.7 – 19.0] and 7.4 [7.0 – 7.9] ACH, respectively, were reported. The baseline tests highlight the potential of cross-ventilation where, by simply opening windows and doors, a ventilation rate exceeding IPC recommendations was obtained. All configurations, save An investigation into turbine venti lators as a potential environmental co ntrol measur e to minimise the risk of transmission on TB Page IV SS1, appear to have approached the fully-mixed case.SS1 also showed the greatest variability in ventilation flow rates. This finding is not unexpected, as air exchange in single-sided ventilation is due to wind pressure fluctuations, which varied across each test. In addition, in all tests it was found that the ventilation flow rate was dependant on the natural ventilation configuration and openable area, and not necessarily environmental conditions. In the turbine ventilator tests, the mean ventilation flow rates for IL, SS1, CV and SS2 were 1.8 [1.6 – 2.1], 5.4 [5.2 – 5.7], 17.7 [16.0 – 18.6] and 9.5 [8.5 – 10.1] ACH, respectively. The mean ventilation flow rate increased in IL and SS2 with the installation of the turbine ventilator, while in SS1 a decrease was reported. The increase in ventilation flow rate in IL was found to be due to natural convection, where the turbine ventilator merely facilitated the exhaustion of warm air. The results of the field study are specific to the environmental conditions at the time of the test, and are not generalizable. In the laboratory experiments, the in-duct velocity increased with an increase in wind speed and temperature differential. For a given temperature differential, an increase in wind speed resulted in a decrease in in-duct velocity. Across all tests, no centreline velocity profile, described by the Dalla Valle equation, could be measured. In the wind speed tests, no capture envelope could be established. This was due to the low wind speed test range, where the resulting centreline velocity was beyond the limit of detection of the thin-film sensors. In the buoyancy forces test, a turbulent region near the base of the turbine ventilator was realised, where the magnitude and direction of the air flowing at 1.5D continuously changed. This turbulent region was again observed in the combined wind and buoyancy forces tests, though the magnitude was smaller and occurrence less frequent. The results of the laboratory experiments are specific to the parameters tested, and are not generalizable. By correlating the field study, laboratory experiments, and previous (similar) studies, it was concluded, that, under the tested conditions, adding a turbine ventilator as a supplement to natural ventilation system will not reduce the concentration of contaminants in the occupied zone in a room.

The purpose of this study was to investigate the ergonomics and biomechanics of rifle shooting from the standing position. At present, the scientific literature on shooting contains primarily qualitative descriptions of the various aspects of the sport. Quantitative data on the kinetic and kinematic aspects of standing rifle shooting was collected in the present thesis. Transducers were developed to measure foot forces in the vertical and horizontal plane, recoil force on the shoulder and the grip force of the trigger hand. Kinematic data was collected with a video camera. The study revealed that recoil energy was dependent on the attributes of the shooter. Handgrip forces were found to be well below the maximum handgrip strength. An exploratory investigation of the interrelationships among foot forces, rifle recoil and angular and linear displacements was carried out. No horizontal foot forces to counteract the recoil force were observed. A possible explanation for the finding is proposed. The findings are discussed with reference to the ergonomic implications for rifle stock design. Further investigations of the relationships between shooting performance and the man-rifle interface are indicated.

The use of technology in health-care today is increasing dramatically with a corresponding increase in cost and complexity to provide and support it. The degree to which a hospital manages this technology affects its ability to treat patients, to perform research, to teach and to attract competent staff. This thesis project has identified the role that clinical engineering could play in health-care technology provision and support in South Africa. A system synthesis technique was employed to develop an idealized clinical engineering model (ICE) that would satisfy South African technological requirements. An extensive literature survey of the current status of clinical engineering in both developed and developing countries was undertaken to provide input to the synthesis process. Surveys were then conducted to determine the actual current status of clinical engineering and its environment in the RSA. To enable such an idealised department to function as defined, it must be supported by appropriate and timeous information. The information needs of the idealised clinical engineering model were analysed and a corresponding decision support system (DSS) defined. Further surveys were conducted to test the applicability and acceptability of the idealised clinical engineering model. The feasibility of implementing the idealised clinical engineering model in South Africa was investigated and recommendations were made based on the research results of this thesis to bring the actual status of clinical engineering closer to the idealised model. ii

This thesis investigates a mechanism for air flow limitation in children with croup. Croup is a common condition affecting many young children. Infection (usually viral) causes swelling of the mucosa in the subglottic region of the airway with consequent narrowing of the airway. Although researchers have investigated croup for the past sixty years, there is still very little information available on how croup affects air flow dynamics. The current theory assumes that the stenosis formed by croup in the subglottis of infants leads to a dynamic collapse of the extrathoracic trachea (Chernick, 1990). According to this literature, the dynamic collapse of the extrathoracic trachea will limit the inspiratory flow. It was believed that in severe cases of croup, the dynamic collapse may even temporarily block the airways. In order to investigate the mechanism for air flow limitation in croup the author used the intrathoracic pressure - flow traces from twenty patients with croup, four patients who had been intubated for croup and five normal subjects. Laryngeal X-rays from another twenty patients with croup were analysed as well as five videos, made during laryngoscopy, of the subglottic cross-sectional area of an additional five patients with croup requiring intubation. All data used in this project was collected by an experienced paediatrician from the Red Cross War Memorial Children's Hospital who is also the supervisor of this thesis. Both the video and the X-ray data showed that the dynamic collapse of the trachea contributes much less to airflow obstruction than the subglottic swelling itself. The hypothesis investigated in this thesis is that air flow becomes restricted due to wave speed limitation. According to the theory of wave speed limitation, an increase in driving pressure (the intrathoracic pressure) does not increase the flow if the speed of the air particles exceeds the wave speed. In our case the wave speed is the speed of sound within the lumen of the compliant, narrowed airway. In order to test that theory, it was necessary to obtain the flow, the driving pressure in the subglottis and the cross-sectional area of the subglottis of patients with croup. Unfortunately, the measurement of subglottal cross-sectional areas from videos made during laryngoscopies, proved to be impossible due to both ethical and practical constraints. The measurement of the subglottal cross-sectional areas from X-rays was also difficult in practice. Therefore, the cross-sectional area is calculated. The general orifice equation is modified m order to calculate the subglottal cross-sectional areas in patients with croup. Two methods are used to test the hypothesis of wave speed limitation: i) The wave speed limitation formula. The wave speed limitation formula directly calculates the maximum flow from the pressure - flow data. Hereafter the calculated maximum flow is compared with the measured flow. ii) A lumped component model. A nonlinear, lumped component model has been used to calculate the flow from the driving pressure (intrathoracic pressure). Flow is not limited in this model and an increase in driving pressure will result in a corresponding increase in flow. The flow which is calculated using this model has also been compared to the measured flow. It was found that, in children with croup, there is a good correlation (r=0.82) between calculated and measured values of maximum flow using the wave speed limitation model. The slope of the linear fit using a least square's approximation is 0.98 and this linear relationship is valid for a 0.05 level of significance for Conover's nonparametric test (Daniel and Terrell, 1989). The lumped component model was able to fit the inspiratory flow data with a small sum of square error in the case of both normal ((7.56 ± 0.86) · 10⁻⁹ (ml/s)²) and intubated patients ((3.2 ± 0.75)·10⁻⁹ (ml/s)²). However, the error rose dramatically in patients with croup ((2.04 ± 0.5) -10⁻⁸ (ml/s)²) thus indicating that the lumped component model is no longer valid in these patients. It is concluded that the measured flow velocities in patients with croup approach the calculated velocity of sound in the region of the subglottic swelling, and that the wave speed theory accurately describes the flow limitation. Further support of this is the fact that the lumped component model, which does not incorporate a flow limiting mechanism, breaks down in patients with croup.

Includes bibliographical references (leaves 155-160).

Includes bibliographical references / Heparin (Hep) and heparan sulfate (HS) have been shown to possess anticoagulative properties, inhibit smooth muscle cell proliferation, moderate inflammation and control angiogenesis by stabilization and potentiation of growth factors (GF). These properties are potentially very useful for the treatment of cardiovascular diseases, especially when delivered as injectable hydrogels that can form in situ. This project focused on developing Hep and HS hydrogels for localized GF delivery. Hep and HS were acrylated, characterized and crosslinked with PEG tetra-thiols, either directly (10m% Hep/HS-Ac, Type 1) or by copolymerization with 20PEG8Ac or 20PEG8VS (4m% copolymer; 1.5% Hep/HS-Ac) to form degradable (Type 2D) or non-degradable (Type 2N) gels, respectively. Gelation times, viscoelasticity, swelling, mesh size, Hep/HS elution and activity, as well as GF incorporation and release were studied in vitro. Type 2D gels with covalently incorporated (CI) Hep and GFs were evaluated in vivo as ingrowth matrices in porous polyurethane (PU) scaffolds for healing response in a rat subcutaneous model.