text
stringlengths 4
179k
| predicted_class
stringclasses 5
values | confidence
float64 0
100
|
|---|---|---|
Our results also illustrate that the errors due to incorrect density correction as a result of motion are distinctly different if the target is on the “liver” side or the “lung” side. For a target region on the “lung” side, depending on the specific phase of the CTAC, an error in the density correction can occur in two distinct ways: either by low density “lung” without the target present (undercorrection), or by higher density “liver” (overcorrection). A target on the “liver” side, however, is surrounded by a material of similar density. Motion due to inspiration may result in an undercorrection if lung is substituted for liver, but motion in the other direction will simply substitute more liver in place of target plus liver, resulting in virtually no error in the correction. As such, a target in this location is generally only subject to undercorrection and not overcorrection, while a target on the “lung” side suffers from both. | other | 32.16 |
To assess the results presented in Figs. 4–7, it is helpful to keep in mind that the primary purpose of 4D imaging is to reduce or eliminate motion artifacts. As such, the effectiveness of a specific 4D imaging and reconstruction method can be evaluated in terms of its ability to effectively stop motion. Within the context of these phantom studies the 3D Static case, in which no motion is present, serves as a useful benchmark for comparison. By contrast, the 3D Dynamic values represent an example of the other extreme — motion is present but not accounted for. | other | 35.62 |
As a consequence of this, when the target is on the higher density “liver” side, a correction based on the CTAC 50% appears to do about the same, or even slightly better, than the 4D CTAC. This does suggest that a strategy of using end expiration 3D CT for correction could be a reasonable alternative to either 4D CINE or 4D CTAC in some situations. However, this approach would require that the target location be reliably known a priori, which will not generally be the case in clinical situations. | other | 32.97 |
Recently, with the change of people's living condition and life style, hyperlipidemia comes to the second major reason of acute pancreatitis. A study involving 2416 cases diagnosed with acute pancreatitis (AP) from 2006 to 2010 in Beijing found that 255 (10.36%) cases were hyperlipidemic acute pancreatitis (HLAP) . An analysis carried by Xu et al. claims that, in the period of 2012 to 2014, HLAP accounted for 19.1% of total AP . Compared with NHLAP, HLAP are characterized by critical condition and high recurrence rate . So, it is important to predict the severity and prognosis of HLAP at early stage, which is beneficial for individualized treatment and prognosis. | other | 33.28 |
There are four frequently used scoring systems of AP, including BISAP (bedside index for severity in acute pancreatitis), Ranson score, MCTSI (modified CT severity index), and APACHE II (acute physiology and chronic health evaluation scoring system). To our knowledge, there is no large-population-based study in assessment of severity and prognosis of HLAP. In this paper, a total of 326 cases diagnosed with HLAP from 2007 to 2015 admitted to a single center were retrospectively analyzed to compare the prediction value of four scoring systems. | other | 37.66 |
While our results show that 4D CT correction methods will reduce the effects of phase mismatch between PET emission and CTAC data, a number of factors complicate the use of 4D CT in a clinical setting. Whereas the addition of a 4D CT series by itself does not add significantly to the acquisition time of a 4D PET study, 4D PET acquisition requires a substantially longer acquisition time. While the acquisition of 4D CT data (either CINE or 4D CTAC) does not add significantly to the total acquisition time, it is important to consider that 4D studies involve substantially increased radiation dose to the patient. The current method used for our study for 4D PET image reconstruction using 4D CTAC also requires significant additional effort on the part of the scanner operator for two specific reasons. The 4D CT data must be phase‐binned using a separate imaging workstation and transferred back to the scanner console and, once transferred, must be manually matched to the appropriate phase of the uncorrected 4D PET. As such, in current form, the 4D CTAC method does not appear to be a practical method for routine clinical use. This is not the case for the 4D CINE method, for which the only drawback is the additional radiation exposure to the patient. | other | 29.16 |
Based on phantom studies using a simulated lung/diaphragm interface, 4D PET images reconstructed using both the 4D CINE and 4D CTAC attenuation correction methods showed fewer motion artifacts when compared to 4D PET images reconstructed using a traditional 3D CTAC. A larger effect was observed when the target was in low‐density material (simulated lung) as opposed to water equivalent (simulated liver) material, and also when motion was according to sin4(x) rather than sin(x). The 4D CTAC correction method was more effective at reducing artifacts than the 4D CINE method for most scenarios, with the exception of a target in high‐density (liver) material and a motion pattern of sin(x) for which the two methods were approximately equivalent. We conclude that when 4D PET images are acquired near a tissue interface such as the lung/diaphragm border, 4D attenuation correction techniques are of value for reducing motion artifacts. | study | 26.34 |
We retrospectively analyzed a series of 326 patients diagnosed with HALP who were admitted to Beijing Chao-Yang Hospital, Capital Medical University, in a period of August 2007 to July 2015 (184 males, 142 females; age ranging from 14 to 85; mean age of 44 years). Of the 326 patients included, 65 (19.9%) had moderately severe acute pancreatitis (MSAP), 27 (8.3%) had severe acute pancreatitis (SAP), 28 (8.59%) had local complications, and 9 died (2.8%). | other | 34.8 |
Local complications include acute peripancreatic fluid collection, pancreatic pseudocyst, walled-off necrosis, infected necrosis, pleural effusion, intestinal fistula, and pancreatic pseudocyst hemorrhage. Of the 28 patients with local complications, 21 had two or more local complications. 72 patients had a relapse, and the recurrence rate is 22.09%. Among them 15 (7.67%) had 3 or more recurrent relapses. | other | 30.69 |
The diagnosis of acute pancreatitis, whether in the presence or absence of underlying chronic pancreatitis, requires two of the following three features: (1) abdominal pain suggestive strongly of acute pancreatitis, (2) serum amylase and/or lipase activity at least 3 times greater than the upper limit of normal, and (3) characteristic findings of acute pancreatitis on transabdominal ultrasonography, contrast-enhanced ECT, or magnetic resonance imaging (MRI). | other | 29.88 |
The grading of severity of AP referred to Chinese guideline for diagnosis and treatment of acute pancreatitis as follows:MAP meets AP diagnostic criteria. MAP requires one of the following three features: no organ failure; no local or systemic complications; and Ranson score < 3 points; APACHE II score < 8; BISAP score < 3 points; and MCTSI score < 4 points.MSAP meets AP diagnostic criteria. At the same time, MSAP should meet one of the following conditions: (1) Ranson score ≥ 3 points; (2) APACHE score ≥ 8 points; (3) BISAP score ≥ 3 points; (4) MCTSI ≥ 4 points; (5) transient organ failure (<48 h); (6) pseudocyst, pancreatic fistula, or pancreatic abscess that needs surgical operation.SAP meets AP diagnostic criteria and have the presence of persistent (>48 h) organ failure (single or multiple organ) and the modified Marshall score ≥ 2 points. | other | 31.98 |
MSAP meets AP diagnostic criteria. At the same time, MSAP should meet one of the following conditions: (1) Ranson score ≥ 3 points; (2) APACHE score ≥ 8 points; (3) BISAP score ≥ 3 points; (4) MCTSI ≥ 4 points; (5) transient organ failure (<48 h); (6) pseudocyst, pancreatic fistula, or pancreatic abscess that needs surgical operation. | other | 29.73 |
BISAP and APACHE II score were calculated in 24 hours after admission. Ranson score was calculated in 48 hours. MCTSI score was calculated in patients with contrast-enhanced CT scan within 3 days of onset. The score is calculated based on the worst value of each criteria [8–10]. | other | 34.75 |
Local complications were evaluated by contrast-enhanced computed tomography (CECT). CECT scans were retrospectively and independently reviewed by two experienced abdominal radiologists who were unaware of presenting signs and symptoms or of patient outcomes. Kappa statistic was calculated for measuring agreement between two radiologists and the result indicates good agreement. We obtained approval from Institutional Review Board of Beijing Chao-Yang Hospital for this study. | other | 35.3 |
The data was statistically analyzed using SPSS software 19.0. BISAP, Ranson, MCTSI, and APACHE II were compared in predicting severity, location complications, and mortality of HLAP, using chi-square testing, Fisher's exact probability test, and receiver operating characteristic curve. Odds ratio (OR), sensitivity, specificity, positive predictive value (PPV), positive likelihood ratio (PLR), Youden index, and area under ROC curve (AUC) were calculated. A p value < 0.05 was considered as statistically significant. | other | 34.12 |
Tables 1, 2, and 3 show that the incidence of MSAP and SAP, local complications, and mortality in patients with BISAP score ≥ 3, Ranson score ≥ 3, APACHE II score ≥ 8, and MCTSI score ≥ 4 were significantly higher than those in BISAP score < 3, Ranson score < 3, APACHE II score < 8, and MCTSI < 4 (p < 0.05). | other | 31.73 |
Table 4 and Figure 1(a) show the sensitivity, specificity, PPV, PLR, Youden index, and AUC of BISAP, Ranson, MCTSI, and APACHE II in predicting severity of HLAP. In assessment of severity, the sensitivity and AUC of APACHE II were 57% and 0.814, which were the highest. BISAP was second with sensitivity of 54%. And the AUC of BISAP was 0.795. | other | 32.12 |
Table 4 and Figure 1(b) show the sensitivity, specificity, PPV, PLR, Youden index, and AUC of BISAP, Ranson, MCTSI, and APACHE II in predicting location complications of HLAP. In assessment of local complications, the sensitivity and AUC of MCTSI were 68% and 0.791, which were the highest. BISAP was second with AUC of 0.731. | other | 31.06 |
Table 4 and Figure 1(c) show the sensitivity, specificity, PPV, PLR, Youden index, and AUC of BISAP, Ranson, MCTSI, and APACHE II in assessment of mortality of HLAP. In assessment of mortality, the sensitivity and AUC of BISAP were 89% and 0.867, which were the highest. The second was APACHE II with AUC of 0.854. Both APACHE II and BISAP had the highest sensitivity of 89%. | other | 31.25 |
Compared with NHLAP, HLAP has the following features: (1) high recurrence rate: the higher the blood lipid level is, the greater the possibility of recurrence; serum TG lower than 5.56 mmol/L can prevent episodes of pancreatitis; (2) serum TG above 11.3 mmol/L; (3) xanthomata in the limbs, buttocks and back, retinal lipemia, hepatosplenomegaly, and fatty liver which can be found in patients with severe hypertriglyceridemia (HTG) because of lipid deposition; (4) patients with HLAP having younger age of onset. Uncontrolled diabetes, obesity, alcoholism, pregnancy, family history of hyperlipidemia are thought to be the risk factors for HLAP [11, 12]. | other | 33.03 |
It has been reported that the incidence of sever acute pancreatitis (SAP) and organ dysfunction, recurrence rate, and mortality of HLAP were significantly higher than those of acute biliary pancreatitis . So, it is extremely important to predict and evaluate the severity of HLAP at early stage. The clinical scoring system is a practical tool for the doctor to find potential patients who need intensive care. However, there is little research on this aspect. A total of 129 cases of HLAP show that it is fairly accurate to predict SAP, complications, organ failure, and prognostic of HLAP, using BISAP, Ranson, SIRS, and MCTSI (area under the curve ranges from 0.938 to 0.668) [13, 14]. | other | 36.97 |
BISAP, Ranson, APACHE II, and MCTSI are the most commonly used scoring systems to evaluate acute pancreatitis, but they still have limitations. Firstly, APACHE II scoring system is not convenient to operate, because it has too many parameters to collect. Secondly, different scoring systems contain common parameters; for example, SIRS is a composite parameter used in both BISAP and APACHE II. However, different scoring systems have different way to evaluate one parameter. Taking blood urea nitrogen (BUN) as an example, Ranson takes the increasing level of BUN as criteria, but BISAP takes the absolute value as the criteria. So it takes a lot of time using variety of scoring systems to predict the prognosis. | other | 32.12 |
Our study showed that the incidence of MSAP and SAP, local complications, and mortality were significantly higher in BISAP score ≥ 3, Ranson score ≥ 3, APACHE II score ≥ 8, and MCTSI score ≥ 4 than in BISAP score < 3, Ranson score < 3, APACHE II score < 8, and MCTSI < 4 (p < 0.05). Therefore, all four scoring systems can be used to predict the severity, local complications, and mortality of HLAP. | study | 29.05 |
In 1974, Ranson et al. selected 11 indicators associated with the severity of AP by screening 43 clinical and biochemical indicators . Our study showed that Ranson score was poor in predicting severity and prognosis of HLAP. The AUC of it ranked the third in every aspect we considered. In assessment of MSAP and SAP and mortality, the AUC of Ranson was higher than MCTSI. And Ranson had the lowest positive predictive value, positive likelihood ratio, and Youden index compared with other scoring systems. In assessment of local complications, Ranson was only better than APACHE II. | study | 30.61 |
APACHE II is a frequently used scoring system to assess severity of AP. It consists of three parts, namely, acute physiology score, age, and chronic health score. Our study showed that the APACHE II had highest accuracy in predicting MSAP and SAP and did a good job in predicting mortality. But APACHE II was poor in assessment of local complications. In assessment of MSAP and SAP, the AUC and Youden index of APACHE II were 0.814 and 0.46, respectively, which were the highest among these four scoring systems. In assessment of mortality, it had the second highest AUC (0.854) and Youden index (0.67), lower than BISAP. In terms of assessing local complications, APACHE II had the lowest AUC (0.580), PPV (15), PLR (2.0), and Youden index (0.21). | other | 30.39 |
MCTSI is a clinical radiological imaging scoring system for evaluating the mortality and local complications of AP. Contrast-enhanced CT is the gold standard for diagnosis of necrotizing pancreatitis and acute peripancreatic fluid collection . In our study, MCTSI had outstanding performance in predicting location complications, with AUC of 0.791. However, it was poor in predicting severity and mortality. The sensitivity and AUC of MCTSI in assessment of MSAP and SAP and mortality were 36%, 0.654 and 78%, 0.839, respectively, which were lower than the other three scoring systems. According to the retrospective study by Bollen et al., there were no significant differences in prediction accuracies between 7 CT scoring systems and two clinical scoring systems (APACHE II, BISAP), so contrast-enhanced CT scan was not recommended for severity assessment on admission . That conforms to our study. | other | 31.3 |
BISAP was proposed to construct a simple and accurate clinical scoring system to estimate the mortality risk of AP at early stage. BISAP analyzed the in-hospital mortality risk of AP using classification and regression tree analysis, and finally five variables were selected. This scoring system was established from a study involving 17992 cases diagnosed with AP from 212 hospitals. Compared with the APACHE II, the validity of the BISAP score was confirmed using a data of 18256 cases diagnosed with AP from 177 hospitals . Singh et al. applied BISAP to evaluate the severity and mortality in a prospective study enrolling 397 cases diagnosed with AP and came to the same conclusion that BISAP was effective to evaluate the severity of AP . A study by Papachristou et al. indicated that the sensitivity and specificity of BISAP were not worse than “traditional” scoring system (Ranson score, APACHE II, and MCTSI) . In our study, BISAP had high accuracy in predicting MSAP and SAP, local complications, and mortality. In predicting mortality of HLAP, the sensitivity (89%), Youden index (0.69), and AUC (0.867) were the highest in all scoring systems. In predicting MSAP and SAP, the AUC of BISAP was slightly lower than APACHE II. In predicting local complications, the AUC of BISAP was slightly lower than MCTSI. So, BISAP performed better than other three scoring systems in every aspect we compared in this study. | other | 32.1 |
In conclusion, Ranson did not have significant advantage in predicting severity and prognosis of HLAP. APACHE II was the best in predicting severity, but it had shortcoming in predicting local complications. MCTSI had outstanding performance in predicting local complications. But it is poor in predicting severity and mortality. BISAP score had high accuracy in predicting severity, local complications, and mortality of HLAP. And BISAP score has the advantages of having less parameters and being easy to operate. Further studies in comparison to these four scoring systems appear to be needed, since our study was a single-center study. Nevertheless, we recommended using BISAP to predict the severity and prognosis of HLAP. | study | 27.7 |
However, the accuracy of BISAP should be promoted in further study. In assessment of severity, for example, the AUC value of APACHE II was the highest (0.814), while BISAP was slightly lower (0.795). Mounzer et al. had shown that the current scoring system for acute pancreatitis had reached the maximum efficacy in prediction of organ failure. It was more accurate to predict the severity of combined use of several scoring systems, but it was inconvenient for clinical practice. Unless new scoring system was proposed, it was difficult to improve the prediction accuracy of AP . | study | 32.28 |
Another study showed that, compared with NHLAP, the C-reaction protein (CRP) in patients with HLAP was significantly higher on days 1, 2, 3, 4, and 6. That means CRP has predictive value in patients with HLAP [13, 14]. We believed that the diagnostic accuracy will be improved by combining the scoring systems with biochemical parameters correlating with the severity of HLAP, such as CRP and high density lipoprotein. And that is the direction of our further studies [13, 19]. | other | 35.38 |
Previous works based on machine learning and computer vision [1–4] have shown the commercial potential and the practical value of symptoms detection and classification using computing devices. A generalized algorithm is useful as an independent step before higher-level algorithms like recognition and prediction; the existing recognition algorithms are usually based on assumptions and trained for specific symptoms, therefore the performance and utility are constrained by lacking training data of unusual symptoms. | other | 32.88 |
This study makes several contributions, includingAnalyzing and quantifying common facial features which are generally shared among human beings regardless of race, gender and age. The data and results are produced upon applying computer vision algorithms and statistical analysis on faces databases . The actual data in use include more than 8200 frontal face images following gender, age, and race distributions of the adult U.S. population .Detecting and categorizing suspected illness features on the testing data by adopting the semi-supervised outliers based on the statistical facts obtained from normal faces dataset. The illness featuring data are collected from UCSD School of Medicine and VA Medical Center , The Primary Care Dermatology Society , and other multiple online resources . The testing dataset is consisted of 237 pictures of more than 20 diseases (Central CN 7 Palsy, Cervical Adenopathy, Hematoma of the Scalp with Cellulitis, Parotitis, Peripheral CN7 Palsy, Submandibular Abscess, Zoster and Cellulitis, Corneal Ulcer, Cyanosis, Extraocular Muscle Entrapment (Inf Rectus), Horner's Syndrome, Icterus, Muddy Brown Sclera, Periorbital Cellulitis, Periorbital Echymosis, Scleritis, Subconjunctival Hemorrhage and different types of Acnes) which can be reflected as abnormal facial features under a variety of different conditions, and 237 pictures of normal randomly picked from databases [9–13].Unifying multiple symptom-detecting processes for different diseases into one automatic procedure by a relatively simple implementation, such that the recognition of specific diseases can be isolated as an independent module with less assumptions. | other | 34.75 |
Analyzing and quantifying common facial features which are generally shared among human beings regardless of race, gender and age. The data and results are produced upon applying computer vision algorithms and statistical analysis on faces databases . The actual data in use include more than 8200 frontal face images following gender, age, and race distributions of the adult U.S. population . | other | 36.56 |
Detecting and categorizing suspected illness features on the testing data by adopting the semi-supervised outliers based on the statistical facts obtained from normal faces dataset. The illness featuring data are collected from UCSD School of Medicine and VA Medical Center , The Primary Care Dermatology Society , and other multiple online resources . The testing dataset is consisted of 237 pictures of more than 20 diseases (Central CN 7 Palsy, Cervical Adenopathy, Hematoma of the Scalp with Cellulitis, Parotitis, Peripheral CN7 Palsy, Submandibular Abscess, Zoster and Cellulitis, Corneal Ulcer, Cyanosis, Extraocular Muscle Entrapment (Inf Rectus), Horner's Syndrome, Icterus, Muddy Brown Sclera, Periorbital Cellulitis, Periorbital Echymosis, Scleritis, Subconjunctival Hemorrhage and different types of Acnes) which can be reflected as abnormal facial features under a variety of different conditions, and 237 pictures of normal randomly picked from databases [9–13]. | other | 36.7 |
Figure 1 displays the workflow of the proposed methods.Fig. 1The proposed framework adopting our methods: training data processing and feature extracting are introduced in section 3.1 and 3.2; detecting process running on testing dataset is introduced in section 3.3 | study | 30.94 |
The results of this research are expected to be a practical tool for preliminary diagnosis. It could be used as a component of health systems and increase the efficiency of treatment process and makes use of previously unused data. It is important to note that the algorithms introduced in this paper are intended to be a supplementary tool for existing medical assessment and treatment mechanisms, not a replacement. | other | 36.62 |
Early works investigated the utility of systems based on supervised learning, which provide gratifying performance but also require significant feature engineering and high quality training data. Quentin Ferry et al. introduced SVM classifier and PCA to extract phenotypic information from ordinary non-clinical photographs to model human facial dysmorphisms in a multidimensional 'Clinical Face Phenotype Space' ; Jane Reilly Delannoy and Tomás E. Ward proposed a computer vision based system for automatically measuring patients’ ability to perform a smile , where the degree of facial paralysis can be identified with the aid of Active Appearance Models; Mingjia Liu and Zhenhua Guo introduced an approach to detecting jaundice by investigating skin color with reasonable accuracy ; Lilian de Greef et al. introduced a system on mobile phone to monitoring newborn jaundice by analyzing the skin conditions of infants along with color calibration cards . Compared to previous works, our methods focus on detecting and classifying ill faces without assuming specific targeting symptoms by adopting semi-supervised anomaly detection. | other | 33.16 |
For the purpose of detecting multiple symptoms and the future extensibility of our algorithm, we avoided using techniques which are sensitive to specific symptoms only, like the House-Brackman scoring system ; instead, we relied on studying the statistical models of general facial features, e.g. color and proportion, as those are likely to be distorted by infections and disorders. Machines perform more sensitive to the eccentricity of statistical data than human beings do, therefore the dependency on special calibrations, like House-Brackman scoring system mentioned above, can be reduced and replaced by those general calibrations with a relatively low cost. | other | 38.03 |
The training dataset is composed of 8278 pictures of normal frontal face images following gender, age, and race distributions of the adult US population ; we further collected 237 pictures of faces with symptoms [6–8] paired with 237 pictures randomly picked from normal face datasets [9–13] as our testing dataset. | other | 34.94 |
The training dataset is composed of 8278 pictures of normal frontal faces . We used active shape models (ASMs) to label this dataset. The algorithm adopted in this study is a reimplementation of Face Alignment by Explicit Shape Regression , licensed by MIT. The version of Face Alignment algorithm used in this experiment is trained by the Helen Database with 194 landmarks. | other | 31.9 |
The testing dataset is composed of 237 pictures of ill face expanded from from UCSD School of Medicine and VA Medical Center and The Primary Care Dermatology Society and 237 pictures randomly picked from normal faces datasets [9–13]; 474 pictures in total. 20 diseases are featured in this dataset (Central CN 7 Palsy, Cervical Adenopathy, Hematoma of the Scalp with Cellulitis, Parotitis, Peripheral CN7 Palsy, Submandibular Abscess, Zoster and Cellulitis, CN3 Palsy, Corneal Ulcer, Cyanosis, Extraoccular Muscle Entrapment (Inf Rectus), Horner's Syndrome, Icterus, Muddy Brown Sclera, Periorbital Cellulitis, Periorbital Echymosis, Scleritis, Subconjunctival Hemorrhage and different types of Acnes). | other | 36.28 |
Source URLs for our collected testing dataset were converted to shortened versions for the purposes of publication using TinyUrl (http://tinyurl.com/). The links provided are expected to decay with time and should only be considered exemplars of database composition. | other | 28.2 |
We paired 237 pictures of face with symptoms with equal amount of normal faces data because we had no information about the prior probabilities of various diseases. On the other hand, it is common to evaluate the performance of a system by assuming an equal prior when the actual prior is highly skewed because a trivial classifier that always predicts the popular class will seemingly do extremely well. | other | 34.1 |
The 237 pictures of face with symptoms in the testing dataset were hand labeled. Most of those images were collected along with mosaics or clipped to protect personal privacy; therefore, ASMs were not applicable to them. Applying ASM algorithm on images with those unpredictable conditions is another different challenging problem. Since it is not directly related to the challenge addressed in this paper, we decided to hand label this testing dataset for simplifying purpose. We plan on addressing this problem in future work. | other | 34.3 |
The labels we made in the pictures of the training dataset and the testing dataset suggested the polygons that bounded all related pixels for certain face components, for example, left eye and lips. For each set of labels, we obtained its binary imaging to represent its corresponding facial component. | other | 36.7 |
Because of the limits of the ASMs with 194 landmarks, some labels overlapped with each other, therefore one pixel could be incorrectly presented in more than one binary feature; for example, the upper lip might share regions with the lower lip. The overlapping pixels usually represent neither lips, but the teeth and tongue on a smiling face, which are not the region of interest in our experiments. To avoid including errors, we further sanitized the features by removing those overlapping pixels. | other | 34.25 |
We transformed the original picture from the RGB color space into the CIELAB color space. The A channel and B channel of CIELAB color space allows an approximately linear scale describing the redness and yellowness of the features to flag the potential symptoms on faces. Combining with extracted binary features, we could have a better understanding of the size, color, proportion and even relative position of those face components. | other | 39.62 |
For one picture of a face, we extracted six possible binary features: face/skin, upper lip, lower lip, nose, left eye and right eye (Fig. 4). The extracted binary features obtained from labeled data were used in future steps to generate variants for anomaly detection algorithm.Fig. 4A set of binary features on a face corresponding to left eye, right eye, face contour, upper lip, lower lip and nose | other | 31.11 |
Because the prior probabilities of diseases were unknown, we instead assumed Guassian distribution on the features of our normal face data in this preliminary study. We defined an outlier as one observation containing at least one variant that appearing to deviate markedly from the obtained mean value of the samples in the training dataset. | other | 38.78 |
Table 1 illustrates the variants we used in the outlier detection and their statistical summarization obtained. For abbreviation, α represents the aggregate value of the CIELAB alpha channel (red-green channel) of the feature; β represents the aggregate value of the CIELAB beta channel (yellow-blue channel) of the interested feature; Σ represents the total count of all the pixels belonging to the feature; H is the process of applying the well-known Hough Transform on the CIELAB feature of the skin area, and then further applying a counting function to count how many circular structures we found; the mechanism is based on Size Invariant Circle Detection .Table 1Variants for the outlier detection algorithm, with their mean values and corresponding standard deviationVariantμδ1α(Eye)/Σ(Eye)138.42612.4122β(Eye)/Σ(Eye)138.21413.3453α(Lip)/Σ(Lip)150.7259.7524Σ(LFace)/ Σ(RFace)0.9620.1865Σ(LEye)/ Σ(REye)0.9580.0716H(Face)2.2333.141 | study | 28.03 |
For Variant 6, the K-Means Clustering algorithm was applied on the CIELAB feature before applying size invariant circle detection in our experiments. The clustered features of symptom featuring faces are usually rigid; we further applied Hough Transform on the clusters to find potential circular structures. | other | 34.53 |
Figure 5 displays a comparison of exploring Variant 6 on a normal example and an abnormal example.Fig. 5Two sets of features for Variant 6. 1st row illustrates the CIELAB features and their clusters of a normal face; 2nd row illustrates the CIELAB features and their clusters of a face with acne. Yellow circles are the results of applying Size Invariant circle detection. For above examples, 0 circle was found for the normal face; 119 circles were found for the acne featuring face | other | 30.69 |
Two sets of features for Variant 6. 1st row illustrates the CIELAB features and their clusters of a normal face; 2nd row illustrates the CIELAB features and their clusters of a face with acne. Yellow circles are the results of applying Size Invariant circle detection. For above examples, 0 circle was found for the normal face; 119 circles were found for the acne featuring face | other | 29.5 |
The values of those variants listed in Table 1 can be easily computed by investigating binary features and the corresponding CIELAB feature. We further summarized the mean values (μ) and standard deviations (δ) of the data in training dataset. An outlier is hence defined as a variant whose value is not in μ ± t × δ, where t is the multipler we used to tighten the degree of normality. We applied the threshold μ ± t × δ on our observations with assumed distribution function and eventually divided the testing dataset into flagged group and unflagged group with respect to different t values. | other | 37.25 |
For those data with no certain binary features because of the data quality issue, some variants were not applicable, e.g., Variant 1 (the redness of eye) could not be applicable because no corresponding binary feature of eyes was available for this picture. In addition, color related variants require colored images; proportion related variants require frontal face images. | other | 34.56 |
In this study, we picked the threshold t from t = 0.0 to t = 3.0, with the interval of 0.1, 31 sets of experiments in total. The statistical results are shown in Table 2.Table 2Statistical results collected by choosing thresholds from t = 0.0 to 3.0 t =TPFPPrecisionRecallAccuracyF-10.02372370.5001.0000.5000.6670.12372370.5001.0000.5000.6670.22362370.4990.9960.4980.6650.32322350.4970.9790.4940.6590.42312290.5020.9750.5040.6630.52292210.5090.9660.5170.6670.62221860.5440.9370.5760.6880.72181750.5550.9200.5910.6920.82161540.5840.9110.6310.7120.92121250.6290.8950.6840.7391.02091100.6550.8820.7090.7521.1207970.6810.8730.7320.7651.2200770.7220.8440.7590.7781.3196710.7340.8270.7640.7781.4191650.7460.8060.7660.7751.5186580.7620.7850.7700.7731.6183400.8210.7720.8020.7961.7176350.8340.7430.7970.7861.8171330.8380.7220.7910.7761.9163250.8670.6880.7910.7672.0160220.8790.6750.7910.7642.1157190.8920.6620.7910.7602.2155180.8960.6540.7890.7562.3151170.8990.6370.7830.7462.4145140.9120.6120.7760.7322.5144130.9170.6080.7760.7312.6141100.9340.5950.7760.7272.713990.9390.5860.7740.7222.813650.9650.5740.7760.7202.913450.9640.5650.7720.7133.013340.9710.5610.7720.711 | other | 29.44 |
We collected 60 pictures of 20 different diseases from UCSD School of Medicine and VA Medical Center and The Primary Care Dermatology Society as our starting point, and then expanded this dataset by collecting the images with the same descriptions from other online resources. We eventually obtained 237 pictures of faces with symptoms . In this way the professional suggestions and symptom descriptions [6, 7] are also applicable to this expanded dataset. | other | 36.3 |
Figure 6 displays the ROC curve computed with the 31 sets of experiments displayed in Table 2, using the maximum likelihood fit of a binormal model . The fitted ROC Area (AUC) is 0.846; the Area under curve (AUC) evaluates the overall performance of the algorithm.Fig. 6The ROC curve computed by statistical data in Table 2, and its statistical summary | other | 27.52 |
Figures 7 and 8 display some examples of the detection of the true positive data.Fig. 7True positive examples flagged by the outlier detection. Left: Periorbital Cellulitis; right: Cyanosis Fig. 8True positive examples flagged by the outlier detection. Left: Periorbital Cellulitis; right: Cervical Adenopathy | study | 30.33 |
Figure 7 displays two outliers at t = 1.0 captured because of color information; the left picture was flagged as an outlier by Variant 1 (i.e., redness of eyes, value = 168); the right picture was flagged as an outlier by Variant 3 (i.e., lips color, value = 139). | other | 36.16 |
Figure 8 displays two outliers at t = 1.0 captured because of proportion information; the left picture was flagged as an outlier by Variant 5 (i.e., proportion of eyes, value = 3.03); the right picture was flagged as an outlier by Variant 4 (i.e., proportion of face, value = 1.71). | other | 35.2 |
We also recorded the variant flagged each outlier and the its value; we compared these factors with the ground truth; the flagged cases were counted as true positive reports reflected in Table 2 only if the variants matched the ground truth; we further classified these true positive reports into multiple categories by the reporting variants. The results were displayed in Table 3.Table 3Six categories corresponding to their flagging reasonsFlagging ReasonSuspected SymptomsCategory 1Variant 1 > μ + t × δScleritis, Subconjunctival Hemorrhage, Corneal Ulcer, Extraocular Muscle Entrapment (Inf Rectus), Muddy Brown Sclera, Periorbital Cellulitis, Periorbital EchymosisCategory 2Variant 2 > μ + t × δIcterusCategory 3Variant 3 < μ + t × δCyanosisCategory 4Variant 4 > μ + t × δ or Variant 4 > μ - t × δCentral CN 7 Palsy, Cervical Adenopathy, Parotitis, Peripheral CN7 Palsy, Submandibular AbscessCategory 5Variant 5 > μ + t × δ or Variant 5 > μ - t × δCentral CN 7 Palsy, Peripheral CN7 Palsy, Extraocular Muscle Entrapment (Inf Rectus), Horner’s Syndrome, Periorbital Cellulitis, Periorbital EchymosisCategory 6Variant 6 > μ + t × δAcnes, Hematoma of the Scalp with Cellulitis, Zoster and CellulitisEach flagged outlier was classified into one of the six categories according to its reporting variant. Although most of the categories contain more than one suspected symptoms, the classified category helps to narrow down the possible medical reasons of the anomaly detection | other | 31.95 |
Each flagged outlier was classified into one of the six categories according to its reporting variant. Although most of the categories contain more than one suspected symptoms, the classified category helps to narrow down the possible medical reasons of the anomaly detection | other | 33.75 |
For the purpose of our study, a dataset containing a wide range of diversities of symptoms with roughly equal amount of each is required for testing; similar data scarcity challenge is also faced by many other studies on image recognition-based diagnosis [1–4]. We address this problem by using semi-supervised anomaly detection which produced promising results. Given the diversity, imbalance, and noise in the dataset, as well as a simple methodology, the statistical results we achieved in this study confirm the promise of our approach and future possibilities. | other | 36.66 |
However, semi-supervised learning also restrained the performance. Algoritms for medical usages often require high recall with relatively high precision, which is still beyond the overall summary statistics of our current methods. There are other semi-supervised anomaly detecting mechanisms could be used . We investigated Gaussian Model-Based detecting mechanism in our preliminary study; applying other semi-supervised anomaly detecting models on our variants should result in similar performance. We plan to improve the performance of our algorithm in future work. | other | 31.9 |
The better results could be obtained by combining multiple variants; as implied in Table 3, some diseases have symptoms reflected by multiple variants. However, it would be nontrivial to learn such correlation for the number of variants without supervision. Given that our proposed system is motivated by avoiding using expensive supervised learning, exploiting the correlation between multiple variants is out of the scope of this study. | other | 34.34 |
Our algorithm can be integrated into a multi-cue diagnosis system, i.e. Visual Clinial Decision Support System (CDSS), to help a clinician make a final, reliable diagnosis decision combining with temperature, lab test and other observations. We have initiated some collaborations on automated skin lesion characterization in the context of CDSS; we plan to deploy our methods to industrial pipelines to validate and improve our methods. The anomality detecting mechanism introduced in this study can also be extended to assist other health related research like detecting and recognizing psycho-behavioral signals . In addition, while our study focuses on the faces, the algorithm itself is readily extended to body and limbs. | other | 33.47 |
Globally, there are many crucial gaps in research and knowledge of the health of transgender people . In South Africa, despite gender affirming care (GAC) being, in accordance with international standards and the South African National Health Act , “legal, ethical medical practice” , and such services being—theoretically—available in both the private and public sectors, access remains severely limited and unequal. As there are no national policies or guidelines, little is known about how individual health care professionals (HCPs) provide GAC. In an effort to fill this gap, this article presents original research with HCPs in South Africa to explore their knowledge, practices and constraints in the provision of GAC to transgender individuals in the absence of national policies or guidelines. | other | 27.9 |
The term ‘transgender’ has become an umbrella term for people who experience incongruence between the sex assigned to them at birth and their gender identity. Those who are gender diverse (that is, individuals who experience a difference between their gender identity, expression, or role and social norms conferred upon that individual’s sex), who experience gender incongruence, and either temporarily or permanently live in their identified gender with or without pursuing GAC can all identify as ‘transgender’ or ‘trans persons’ [4, 5]. The diversity of lived experiences of trans persons and the corresponding variety of medical needs has traditionally posed challenges for health authorities in setting diagnostic criteria and guidelines for providing GAC, although agreement on treatment standards has been achieved internationally . While not all transgender individuals will pursue GAC, for those who do, it can be integral to realizing their gender identity, both personally and socially. | other | 29.66 |
GAC works towards alleviating the potential distress or dysphoria of trans people through exploring and locating a gender identity that is suited to each individual . GAC usually takes the form of medical interventions (which are discussed in the following section). Whilst relying on medical skills and technologies, and access to healthcare systems, the relationship between gender diversity (including trans identities) and biomedicine is a fraught one. Within medical discourses and health systems, gender diversity has been regarded as a mental health condition since the 1980s, including being listed in the International Classification of Diseases (ICD) and the Diagnostic and Statistical Manual of Mental Disorders (DSM) [8, 9]. Trans activists have consistently asserted that gender outside of the strict male/female binary is simply a part of human gender diversity, is not an illness, and ought to be understood outside of biomedical frameworks . To this end, over the last decade, trans activists have advocated for the removal of gender diversity-related diagnoses from diagnostic manuals, culminating in changes to the 11th revision of the ICD. In the suggested forthcoming ICD-11, a gender diversity-related diagnosis will be maintained, however, it will be moved out of the mental illness chapter, and will be rephrased to ‘gender incongruence’ rather than ‘transsexualism’ . It is hoped that this configuration will facilitate access to the medical interventions necessary for GAC within existing health systems, but also reduce pathologising language and stigma. | other | 27.31 |
Currently, however, individuals who experience significant distress and wish to access GAC are frequently assigned the ICD-10 diagnosis of ‘transsexualism’, listed under the category of ‘gender identity disorders’ (GID)—“a desire to live and be accepted as a member of the opposite sex, usually accompanied by a sense of discomfort with, or inappropriateness of, one’s anatomic sex and a wish to have hormonal treatment and surgery to make one’s body as congruent as possible with the preferred sex” —or the DSM-V diagnosis of ‘gender dysphoria’—“a marked incongruence between one’s experienced/expressed gender and assigned gender, of at least 6 month’s duration” where the “condition is associated with clinically significant distress or impairment in social, occupational, or other important areas of functioning” . Both the World Health Organization (WHO) and the American Psychiatric Association’s (APA) diagnostic guidelines indicate that relief of these symptoms is positively associated with changes to the primary and/or secondary sex characteristics as a means of diminishing the level of incongruence with the individual’s gender identity . | other | 27.47 |
There are various psychological and physiological medical approaches that trans persons can pursue as GAC. The World Professional Association for Transgender Health (WPATH), in the seventh edition of its Standards of Care (SOC-7), outlines its recognized treatment options for those seeking care for gender dysphoria. These are inclusive of: | study | 29.72 |
changes in the gender expression and role (which may involve living part time or full time in another gender role, consistent with one’s gender identity); hormone therapy to feminize or masculinize the body; surgery to change primary and/or secondary sex characteristics (e.g. breasts/chest, external and/or internal genitalia, facial features, body contouring); psychotherapy (individual, couple, family, or group) for purposes such as exploring gender identity, role and expression; addressing the negative impact of gender dysphoria and stigma on mental health; alleviating internalized transphobia enhancing social and peer support; improving body image; or promoting resilience . | other | 28.12 |
Despite the individualized nature of GAC, however, it is a clinically recognized necessity for those who wish to align their body to their identified gender. This necessity has been acknowledged by the WHO, APA, WPATH, and various health districts around the world. For example, the Royal College of Psychiatrists uses the SOC-7 to inform its United Kingdom standards of care , and the American Medical Association (AMA) shares WPATH’s position that GAC is neither cosmetic nor experimental and advocates for coverage of associated care by insurers, since “GID, if left untreated, can result in clinically significant distress, dysfunction, debilitating depression, and for some people without access to appropriate medical care and treatment, suicidality and death” . Likewise, the APA’s official position on access to GAC is that it “opposes categorical exclusions of coverage for such medically necessary treatment when prescribed by a physician” . | study | 28.94 |
While a diagnosis of GID/GD would suggest that poor health amongst transgender populations is a result of internal distress and discomfort with their gender incongruence, studies show that it is the social environment that individuals navigate in their daily lives which most significantly affects their health and wellbeing . Accordingly, GAC–surgery in particular–is recognised as improving social recognition and reducing experiences with distress and dysphoria . | other | 30.86 |
The South African post-apartheid state enacted a multitude of laws and policies to prevent discrimination and ensure the health of its people. The Constitution guarantees every citizen: the enjoyment of rights regardless of one’s “race, gender, sex, […] sexual orientation, age, […]”; the respect and protection of their inherent dignity; the right to life; freedom and security of their person including protection from inhumane and degrading treatment and the preservation of “bodily and psychological integrity,” including sovereignty in decisions regarding reproduction and control over the body; and the right to access healthcare services . Case law has interpreted this as including protection based on gender identity [18, 19]. | other | 29.64 |
Access to GAC, however, is hindered by the realities of the South African health system. South Africa has both a public health system–of about 400 hospitals and 4000 primary care facilities , and, in line with state policies, a private system consisting of general practitioners (GPs) and about 200 hospitals, which are largely funded by private medical schemes . | other | 31.33 |
There is significant inequity between the two sectors: just thirty per cent of the country’s doctors work in the public health system responsible for providing care to the overwhelming majority of the population who do not have private health insurance (over 40 million individuals) and, despite the already high concentration of doctors in the private sector, public sector doctors have lately acquired the ability to also work a portion of their time privately . | other | 29.6 |
In addition to the structural inequality, the attitudes of HCPs toward transgender and gender diverse people present a considerable barrier to access to care, including GAC. Whilst there is little documentation of transphobia in the healthcare system specifically, trans advocacy organisations provide important first person accounts of gender-based discrimination [23, 24], and a recent study on heteronormativity in healthcare facilities demonstrates that gender diverse people experience severe and specifically gender-bias motivated discrimination by HCPs in South Africa . | other | 29.33 |
The aim of this study was to determine how HCPs in South Africa provide GAC in the absence of national guidelines and, from this acquired understanding, to evaluate the potential role for national guidelines to assist in their work. Specifically, the study examines what GAC services are available in South Africa and where; which tools HCPs employ to inform their provision of GAC and how this impacts upon clients; clients’ ability to access services as well as HCPs access to other GAC-providing professionals; and the diverse challenges of providing GAC in South Africa. | other | 32.3 |
Based on an initial policy review and service mapping, the study employed semi-structured interviews with a snowball sample of twelve HCPs between December 2014 and February 2016. The authors are based in a South African research unit that works specifically on policy development and evidence-based advocacy for gender, sexual health and reproductive rights, and violence prevention in the South African and Southern African context. AM in particular has worked in the field of transgender health in South Africa for several years, and directed the team toward the relevant health policies. In addition, we examined all healthcare policy, including National Strategic Plans and treatment guidelines, on the website of the South African National Department of Health, and consulted other health policy experts as well as organisations working in transgender health and rights advocacy. South Africa experienced a radical policy change from 1991 due to the repeal of many apartheid laws and, subsequently, the implementation of laws that would govern the country under the new democratic dispensation. We reviewed all Department of Health policies since 1991 for content on GAC. For the service mapping, AM provided the initial database of HCPs, which we cross-checked against a list compiled by Gender Dynamix, the largest transgender advocacy organisation in South Africa. Additionally, we asked all participating HCPs if they were aware of GAC services provided elsewhere in the country, and searched newspaper articles about transgender health-related topics published within the last 10 years for information about HCPs that we may have missed. | other | 28.73 |
Through the policy review for this study, prior work with transgender support organizations and other HCPs, and media reports, we compiled a database of clinicians providing various aspects of gender affirming care. We identified eight professionals across the country, which covered the spectrum of GAC including a surgeon, clinical social worker, clinical psychologist, psychiatrist, sexologist, endocrinologist and GP, and worked in either the private or public health system, or both. Another five HCPs were referred to us by participating HCPs (Table 1). Because GAC is seen as highly specialized and is provided by a small group of professionals as identified through our database, this sample is not only indicative of the scope of expertise available but also captures the vast majority of GAC-providing HCPs in the country—although there may be GPs providing ad hoc services. Our recruitment goal was to interview all HCPs currently providing GAC in South Africa. To our knowledge, we succeeded in interviewing most HCPs working in the tertiary public sector, and in private. | other | 30.72 |
All interviewees were approached by email and provided with information about the background, methodology and objectives of the study, allowing the researchers to introduce the study and arrange an in-person or telephonic interview, depending on the interviewee’s availability and location. Thirteen professionals were contacted, and indicated interest, but only twelve proceeded to participate. Interviews took place at locations of the participant’s choosing, either at their place of work (for participants located in the wider Cape Town area) or by phone (for participants living in other parts of South Africa). | other | 35.47 |
Two authors (AM and TM) conducted interviews. Both interviewers are employed as academic staff at their institution and have more than seven years qualitative research experience each. Both interviewers are female and cisgender. Rapport was easily established by the fact that the interviewers are researchers in a health science faculty and have intimate knowledge of the healthcare system and the field of specialisation of the providers interviewed. One of the interviewers (AM), a physician herself, had an existing professional relationship with most participants. The interviews were semi-structured, based upon a standardised interview guide, and conducted in a private conversation with the participant. Briefly, they ascertained the interviewee’s provision of services for transgender people; their use of ICD codes in service provision; the barriers and/or facilitators they experience in the provision of GAC; the guidelines that inform their provision of such care; and their understanding of the necessity of GAC. The interviews ranged between fifteen and sixty minutes in duration, were conducted in English, and were audio recorded for data accuracy. In two instances, follow-up interviews were conducted to clarify and/or expand upon an initial interview; otherwise, interviews were not repeated. Given that interviewees were practising healthcare providers with a busy schedule, we did not ask participants to read their transcripts for further comments. | other | 36.66 |
The interview recordings were transcribed and proofread. Additionally, the interviewing researchers (AM and TM) took notes during the interview. All data were analysed for thematic fields related to the research question by each researcher, making use of a content analysis framework. The principal researcher (AM) read all data in a comparative analysis to identify key themes around which further analysis was structured. Following this, the principal researcher and one co-investigator (SS) determined codes for all data and harmonized the key themes. Analysis was then done by one researcher (SS) using MSOffice (Word and Excel, Microsoft Corp.) and NVivo (QSR International) software for qualitative data analysis. | other | 36.56 |
Approval for this research was obtained from the Human Research Ethics Committee of the Faculty of Health Sciences, University of Cape Town (HREC 857/2014). Participants were interviewed after giving written informed consent. In order to protect participants’ anonymity, we attribute information given by them through the interview codes listed in Table 1. | other | 37.22 |
The South African Department of Health has not issued any guidelines or policies on GAC, nor is transgender health recognized as a medical specialty. Six public hospitals in urban centres provide various components of GAC (Table 2); however, the waitlist, particularly for surgery, is long and growing in the public sector—up to 25 years at Groote Schuur Hospital in Cape Town, as, due to limited resources, the provincial Department of Health only allocates four theatre days a year for gender affirming surgeries (HC004) [26–28]. | other | 28.39 |
Alternatively, psychosocial support, hormone therapy, and, to a much more limited extent, gender affirming surgeries are available through the private sector, although generally not covered by health insurance (HC004), as none of these procedures are recognised under the prescribed minimum benefits that regulate essential coverage . Anecdotal evidence suggests that health insurances oppose claims for GAC on the grounds that they deem it ‘cosmetic’ [26, 28]. | study | 28.86 |
Interviewees reinforced that GAC is available in the public and private sector, but that access is limited to the urban centres of South African provinces with better infrastructure. The majority of interviewees worked in public hospitals. While a number of these public sector specialists also provide care in the private sector, those interviewees providing GAC exclusively in private care do so through their own private practices as GPs. | other | 28.83 |
In Gauteng, GAC services are being provided at Steve Biko Academic Hospital in Pretoria, Helen Joseph Hospital in Johannesburg and Chris Hani Baragwaneth Hospital in Soweto. Professionals at the latter—all mental health professionals—formed a transgender health team over the last few years in response to a lack of guidance for trans patients or HCPs: | other | 27.17 |
It started off with one psychologist [who] was the first one to come and try and help with transgender patients. And because endocrine was being so difficult and access to hormones or even surgery or any other gender affirming kind of treatment […] we decided to start a little bit of a panel (HC008). | other | 33.44 |
The panel conducts preliminary assessments to determine the level of psychosocial support necessary for an individual, provides necessary mental health services, and facilitates access, often ad hoc, to the endocrinology and surgery departments (HC008, HC010, HC012). | other | 30.83 |
In the Western Cape, GAC is available at the Red Cross War Memorial Children’s Hospital and Groote Schuur Hospital, both in Cape Town. At Red Cross, a private sector psychiatrist provides pro-bono services once a week and, through their advocacy, adolescents can also access paediatric endocrinology services or be referred for further care at Groote Schuur Hospital (HC001, HC002). Groote Schuur Hospital is the home of the Multidisciplinary Transgender Unit, which provides GAC through a team of specialists . The Unit comprises psychologists and psychiatrists, endocrinologists, a surgeon, and a clinical social worker (HC003, HC004, HC007). Occasionally, it has also included a laser therapist and gynaecologist (HC007). One distinct feature of the Unit is its strong ties to local transgender community organizations, which help to inform its provision of services and clients’ access to the Unit (HC003, HC007). This successful model is owed to the initiative of a few dedicated HCPs: | other | 27.3 |
[The] clinic had been run since the 1980s but it was run in a very disjointed manner [by two doctors]. And then both retired and then there was a gap from about 2001 through to about 2009. [Patients] came through but they were seen by a variety [of professionals] and it just was a bit of a mess and we didn’t have a surgeon. […] then 2009, we got together, told ourselves we have to do something about this. And then we got everyone […] we got a group, which hadn’t happened in other [places] so we were quite pleased […]. We got people from outside and social workers, […] and it’s been working and we called that a transgender clinic and then the hospital accepted that as such, that we would do [provide GAC] and so that’s how we started (HC007). | other | 32.06 |
While access to GAC is largely concentrated in a few urban centres, there are also some GPs providing certain aspects of GAC on a case-by-case basis in a variety of healthcare settings across the country. Aware of the potential negative consequences of the lack of guidelines for access to GAC, one of the university-based professionals we interviewed had developed a pamphlet of basic guidance for such GPs (HC006). Such care is limited to psychosocial support and gender affirming hormone therapy, given that only tertiary institutions can conduct surgeries (HC009). In some cases, clients will travel to a tertiary institution for initial services and can then be monitored by a local GP or by the specialist from a distance (HC004, HC012). As a surgeon pointed out: | study | 29.1 |
In addition to traditional healthcare services, the key informants all highlighted the role of community organizations—specific to either transgender or lesbian, gay, bisexual, transgender, intersex, and queer (LGBTIQ)—as integral to the provision of GAC. The work being done by Triangle Project and Gender DynamiX in Cape Town (Western Cape), Transgender Intersex Africa (TIA) and OUT in Pretoria (Gauteng), and Social, Health and Empowerment Feminist Collective of Transgender and Intersex Women of Africa (S.H.E.) in East London (Eastern Cape) has become integral to locating and accessing services (HC001-007, HC009-010, HC012). Additionally, these organisations are actively engaged in advocacy for the expansion of services for transgender clients by health professionals and institutions, and have contributed to improved services by directly providing training and information to HCPs about current best-practices in the provision of GAC (HC007, HC009). For example, Gender Dynamix has published three clinical guidelines on gender-affirming care adapted for primary care providers in the South African context [3, 30, 31]. These organisations, like the healthcare institutions providing GAC, are mostly located in urban centres (with the exception of S.H.E.) and whilst they may provide outreach to a wider area, their influence is largely evident in the urban healthcare facilities where they can constantly lobby and educate HCPs. For example, the Multidisciplinary Transgender Unit at Groote Schuur Hospital has had a long and successful partnership with Gender DynamiX, up to the inclusion of Gender DynamiX representatives at its review meetings (HC007), which has seen it take a truly multidisciplinary coordinated patient-centred approach . | study | 29.86 |
The interviewees referenced a variety of nationally and internationally produced guidelines and standards of care which guided their provision of GAC. All but two of the professionals explicitly acknowledged the WPATH Standards of Care as an important resource to their provision of care. Though the majority of these references were to the most recent SOC-7, some noted that they or colleagues they worked with are consulting prior editions (SOC-6). | other | 27.2 |
In addition, service professionals also referenced guidelines put forth by the International Endocrine Society (HC006, HC011); the Centre for Excellence for Transgender Health in San Francisco (HC006); Callen-Lorde Community Health Centre in New York (HC006, HC011); the Transgender Health Information Programme in British Columbia (HC011); the Psychological Society of South Africa (HC009); WHO (HC001-005, HC007-009, HC011-012); APA (HC001, HC003, HC004, HC007, HC009); Gender DynamiX [3, 30, 31] (HC006, HC009); and various academic journal articles (HC009, HC001). | study | 26.95 |
With the exception of a position statement by the Psychological Society of South Africa and guidelines by the community organization Gender DynamiX [3, 30, 31], there are no national guidelines for the provision of GAC and no state-recognized standards of care (HC001, HC002, HC004, HC012). Further, use of the international guidelines that the HCPs referenced as informing their practices are not mandated (HC001, HC002, HC008, HC009). Thus, HCPs are left to seek out guidance and best-practice guidelines in their own time and of their own volition. | other | 28.64 |
A number of HCPs noted that the current system, whilst extremely uneven, can facilitate the pursuit of GAC for clients (HC002-004, HC007-009). Accessing services does cover the spectrum of services; however, access to mental health professionals, which form the basis of GAC in South Africa, is much more available than access to gender affirming surgeries. In part, this is a product of the knowledge and will of specific HCPs and resources available, including surgery theatre availability and plastic surgeons (HC004, HC007), but it is also reflective of the affordability and health insurance coverage of such procedures (HC004, HC005, HC007, HC008). | other | 28.52 |
Subsets and Splits