In computer-aided medical diagnosis, obtaining labeled medical image data is expensive, while there is a high demand for model interpretability. However, most deep learning models currently require a large amount of data and lack interpretability. To address these challenges, this paper proposes a novel data augmentation method for medical image segmentation. The uniqueness and advantages of this method lie in the utilization of gradient-weighted class activation mapping to extract data efficient features, which are then fused with the original image. Subsequently, a new channel weight feature extractor is constructed to learn the weights between different channels. This approach achieves non-destructive data augmentation effects, enhancing the model's performance, data efficiency, and interpretability. Applying the method of this paper to the Hyper-Kvasir dataset, the intersection over union (IoU) and Dice of the U-net were improved, respectively; and on the ISIC-Archive dataset, the IoU and Dice of the DeepLabV3+ were also improved respectively. Furthermore, even when the training data is reduced to 70 %, the proposed method can still achieve performance that is 95 % of that achieved with the entire dataset, indicating its good data efficiency. Moreover, the data-efficient features used in the method have interpretable information built-in, which enhances the interpretability of the model. The method has excellent universality, is plug-and-play, applicable to various segmentation methods, and does not require modification of the network structure, thus it is easy to integrate into existing medical image segmentation method, enhancing the convenience of future research and applications.
Objective To systematically review the health utility scores for patients with breast cancer in China. Methods PubMed, EMbase, The Cochrane Library, Web of Science, CNKI, WanFang Data, CBM and VIP databases were searched from inception to November, 2016 to collect studies for health utility scores for breast cancer in China. Two reviewers independently screened literature, extracted data, and assessed the risk of bias of included studies. Meta-analysis was performed using Stata 12.0 software. Results A total of 6 studies were included. Compared with modeling studies, the range of utility scores from cross-sectional studies was narrower. The results of meta-analysis showed the utility scores for breast cancer were 0.77 (95%CI 0.67 to 0.87), 0.77 (95%CI 0.68 to 0.86) at stage 0 to Ⅰ, 0.76 (95%CI 0.66 to 0.85) at stage Ⅱ, 0.74 (95%CI 0.65 to 0.83) at stage Ⅲ and 0.73 (95%CI 0.65 to 0.81) at stage Ⅳ, respectively. Meanwhile, we descripted the median of utility scores of all studies, and the corresponding values were 0.81 (range 0.65 to 0.83), 0.82 (range 0.65 to 1.00) at stage 0, 0.82 (range 0.65 to 0.90) at stage Ⅰ, 0.77 (range 0.67 to 0.86) at stage Ⅱ, 0.70 (range 0.64 to 0.82) at stage Ⅲ and 0.70 (range 0.30 to 0.80) at stage Ⅳ, respectively. Conclusion Research on the health utility scores of breast cancer is still limited in China. Utility scores are decreasing among different clinical stages, with lowest score at late stage cancer. These findings can be used in further cost-utility evaluation on various breast cancer interventions.
ObjectivesThis study aimed to study the economic effect of five kinds of detection systems for nucleic acid, which were based on five kinds of working electrodes: gold electrode, glassy carbon electrode, carbon paste electrode, screen printing electrode, and indium-tin-oxide (ITO) glass electrode.MethodsThe cost of completing a single test was taken as the cost of economic analysis. The Youden index was used to represent the effect of cost-effectiveness analysis (CEA). Meanwhile, the cost-utility analysis (CUA) and incremental cost-effectiveness ratio (ICER) were used for the economic analysis of the corresponding system.ResultsThe cost of five detection systems based on gold electrode, glass carbon electrode, carbon paste electrode, screen printing electrode, and ITO glass electrode was 3.70 yuan/unit, 4.20 yuan/unit, 5.25 yuan/unit, 33.98 yuan/unit and 5.01 yuan/unit, respectively. The Youden indexes of all five systems were 1. The cost effectiveness (C/E) were 3.70, 4.20, 5.25, 33.98, and 5.01, respectively. The cost utility (C/U) were 6.61, 6.89, 9.91, 62.93, and 9.45, respectively. The C'/E and C'/U of the gold electrode detection system were the minimum (2.96 and 5.29). Compared with the system applying the gold electrode, the system using the glassy carbon electrode had ΔC >0 and ?E0 >0; When carbon paste electrode, screen printing electrode, and ITO glass electrode system were used, ?C was >0 and ?E0 was <0.ConclusionsFrom the perspective of CEA and CUA, the system using the gold electrode has the best economic effect. The sensitivity analysis proved the reliability of CEA and CUA results. According to the ICER, gold electrode or glassy carbon electrode can be used in clinical practice with the choice depending on the user.
Abstract: Objective To evaluate the clinical effects and health economics of lung volume reduction surgery（LVRS）, single lung transplantation（SLTx） and bilateral lung transplantation（BLTx） for patients with end-stage emphysema. Methods A total of 61 patients with end-stage emphysema, including 39 patients who underwent LVRS（LVRS group）, 14 patients who underwent SLTx（SLTx group）, and 8 patients who underwent BLTx（BLTx group） from September 2002 to August 2008 in Wuxi People’s Hospital, were analyzed retrospectively. Lung function, arterial blood gas analysis and 6-minute walk distance（6-MWD）were assessed before their surgery and 6 months, 1-year and 3-year after their surgery respectively. Their 1-year and 3-year survival rates were observed. Cost-effectiveness analyses were made from a health economics perspective. Results Compared with their preoperative results, their mean forced expiratory volume in 1 second（FEV1.0）in LVRS group increased by 75%, 83% and 49% at 6 months, 1-year and 3-year postoperatively, by 176%, 162% and 100% in SLTx group, and by 260%, 280% and 198% in BLTx group respectively. Their mean forced vital capacity（FVC）in LVRS group increased by 21%, 41% and 40% at 6 months, 1-year and 3-year postoperatively, by 68% , 73% and 55% in SLTx group, and by 82%, 79% and 89% in BLTx group respectively. Their exercise endurance as measured by 6-MWD increased by 75%, 136% and 111% in LVRS group at 6 months, 1-year and 3-year postoperatively, by 513%, 677% and 608% in SLTx group, and by 762%, 880% and 741% in BLTx group respectively. The 1-year and 3-year survival rates after operation were 74.40% and 58.90% in LVRS group, 85.80% and 64.30% in SLTxgroup, and 62.50% and 50.00% in BLTx group respectively. The three years’ cost utility of SLTx group was significantly higher than that of BLTx group（1 668.00 vs.1 168.55, P＜ 0.05）and LVRS group （1 668.00 vs. 549.46, P＜ 0.05）. Conclusion SLTx and BLTx are better than LVRS in improving patients’ lung function and exercise endurance for end-stage emphysema patients. LVRS is more cost-effective than SLTx and BLTx in the early postoperative period. With the development of medical technology and decreased expenses of lung transplantation and immunosuppressive agents, lung transplantation will become the first surgical choice for end-stage emphysema patients.
ObjectivesTo systematically review approaches to derive disability weights (DWs) based on EQ-5D instrument.MethodsPubMed, EMbase, Web of Science, CNKI and WanFang Data databases were electronically searched to collect studies on the approaches to derive disability weights based on EQ-5D from inception to June 1st, 2019. Two reviewers independently screened literature, extracted the basic information and evaluated risk of bias of included studies. Then, systematic review on approaches to derive DWs based on EQ-5D instrument was performed.ResultsA total of 18 studies were included, which were published between 2003 and 2018. The included studies involved a variety of diseases, mostly focusing on quality of life and the burden of disease. The approaches to derive DWs based on EQ-5D health instrument were as follows: DWs=health utility scoreNormal or Control?health utility scoreDisease (7 studies), DWs=1?VAS score/100 (6 studies), DWs=1?health utility scoreDisease (3 studies), linear regression model (1 study), and mapping (1 study).ConclusionsAmong all the included studies using EQ-5D-based disability weight measurement methods involves a variety of diseases, with relatively low comparability. More methodological studies are from abroad. Among all the applied approaches, DWs equally to health utility scoreNormal or Control minus health utility scoreDisease is the most commonly used.
Objectives To determine the health benefit of elbasvir/grazoprevir versus peginterferon combing with ribavirin (PR regimen) for Chinese chronic hepatitis C patients with genotype 1b infection. Methods Markov cohort state-transition models were constructed to conduct cost utility analysis. Sensitivity analyses were performed based on base-case analysis. Results Elbasvir/grazoprevir was dominant versus PR, resulting in higher QALYs and lower costs for both noncirrhotic patients (13.867 5 QALYs, 82 090.82 RMB vs. 12.696 2 QALYs, 122 791.55 RMB) and cirrhotic patients (12.841 6 QALYs, 225 807.70 RMB vs. 8.892 4 QALYs, 326 545.01 RMB). Elbasvir/grazoprevir was economically dominant in nearly 100% among all patients within the range of threshold from 0 to 161 805 RMB/QALY. Conclusions Elbasvir/grazoprevir was dominant in treatment of genotype 1b chronic hepatitis C infection in China.
The discrete choice experiment (DCE) is a stated preference analysis method used to evaluate the impact of multiple factors on individual choice, which has been explored by scholars around the world for health utility measurement. This method is considered to reduce the cognitive burden of traditional utility measurement methods and has high development potential. By examining empirical studies conducted domestically and internationally that employ DCE for measuring health utility, and drawing on methodological guidelines for constructing DCE models, this paper provides an overview of the methodological background of DCE, the practical process used for measuring health utility, and discusses relevant challenges in its application.
Objective To assess the cost-utility study of renal transplantation compared with nemodialysis (HD) and peritoneal dialysis (PD). Methods A prospective study of end-stage renal disease patients was followed up for 3 months after renal replacement therapy. The study population included 196 patients (renal transplant [RT] n=63, hemodialysis [HD] n=82 and continious ambulatory peritoneal dialysis [CAPD] n=51) from 6 hospitals of Sichuan province. Health-related quality of life was assessed by using the WHOQOL-BRIEF questionnaire. Utility scores were obtained so as to conduct CUA (cost-utility analysis). Costs were collected from financial department and by patient interview. Results The utility values were 0.539 9± 0.013 for RT, 0.450 8± 0.014 for HD, 0.512 2±0.099 for CAPD, respectively. The mean direct cost of the first three months of renal transplant was significantly higher than dialysis (RT and CAPD). Over 3 months, the average cost per quality-adjusted life year (QALY) for patients after CAPD was lower than HD and RT. Compared to HD, incremental cost analysis showed that CAPD was more ecnomical than RT. Sensitive analysis showed that CAPD was more effective than RT when ΔQALY varied in the limit of 95% confidence interval. However, the cost-utility of RT vs HD and CAPD vs HD was varied with ΔQALY level. Conclusions Cost-utility analysis showed that CAPD was a more favorable cost-utility ratio when compared to RT at early stage RT vs HD and CAPD vs HD, but which cost-utility ratio is better, we can not draw a certain conclusion.
ObjectiveTo evaluate the efficacy and cost-effectiveness of thymalfasin (Tα1) as an adjuvant therapy for hepatitis B virus (HBV)-related hepatocellular carcinoma (HCC) after surgery.MethodsPatients with HBV-related HCC who underwent hepatectomy from February 2007 to December 2015 in West China Hospital of Sichuan University, Chengdu Military General Hospital, or the Third Affiliated Hospital of Sun Yat-sen University were retrospectively collected and divided into the Tα1 group and the observation group. Log-rank test and Kaplan-Meier curve were used to assess the overall survival, recurrence-free survival, and safety of patients. A Markov model was used to calculate the quality-adjusted life years (QALYs) and incremental cost-effectiveness ratio (ICER) of the Tα1 group compared with the observation group. Cost data was from the hospital information system of the three hospitals. Utility scores mainly came from published data. Sensitivity analyses were applied to explore the impact of essential variables.ResultsA total of 208 patients with HCC after liver resection were enrolled, among them 48 received Tα1 treatment (the Tα1 group) and 160 were only followed up (the observation group). There was no significant difference between the two groups in the baseline characteristics. The median overall survival of the Tα1 group and the observation group was 102.0 months [95% confidence interval (CI) (74.8, 129.2) months] and 81.6 months [95%CI (65.7, 97.6) months], respectively, and the difference was statisitically significant (P=0.047); the median recurrence-free survival was 66.7 months [95%CI (17.3, 116.1) months] and 37.4 months [95%CI (28.7, 46.2) months], respectively, and the difference was statistically significant (P=0.044). There were no grade Ⅲ-Ⅳ adverse events and no treatment-related death occurred. The ICER of Tα1 group was ￥108 050.02/QALY, which was less than the willingness to pay (￥177 785.25/QALY).ConclusionsTα1, as an adjuvant therapy for HBV-related HCC patients, can improve the prognosis of the patients, and the cost is within the acceptable level in our country, so this strategy is likely to be a cost-effective option compared with the observation group.