SurvdigitizeR: an algorithm for automated survival curve digitization.

BACKGROUND: Decision analytic models and meta-analyses often rely on survival probabilities that are digitized from published Kaplan-Meier (KM) curves. However, manually extracting these probabilities from KM curves is time-consuming, expensive, and error-prone. We developed an efficient and accurate algorithm that automates extraction of survival probabilities from KM curves. METHODS: The automated digitization algorithm processes images from a JPG or PNG format, converts them in their hue, saturation, and lightness scale and uses optical character recognition to detect axis location and labels. It also uses a k-medoids clustering algorithm to separate multiple overlapping curves on the same figure. To validate performance, we generated survival plots form random time-to-event data from a sample size of 25, 50, 150, and 250, 1000 individuals split into 1,2, or 3 treatment arms. We assumed an exponential distribution and applied random censoring. We compared automated digitization and manual digitization performed by well-trained researchers. We calculated the root mean squared error (RMSE) at 100-time points for both methods. The algorithm's performance was also evaluated by Bland-Altman analysis for the agreement between automated and manual digitization on a real-world set of published KM curves. RESULTS: The automated digitizer accurately identified survival probabilities over time in the simulated KM curves. The average RMSE for automated digitization was 0.012, while manual digitization had an average RMSE of 0.014. Its performance was negatively correlated with the number of curves in a figure and the presence of censoring markers. In real-world scenarios, automated digitization and manual digitization showed very close agreement. CONCLUSIONS: The algorithm streamlines the digitization process and requires minimal user input. It effectively digitized KM curves in simulated and real-world scenarios, demonstrating accuracy comparable to conventional manual digitization. The algorithm has been developed as an open-source R package and as a Shiny application and is available on GitHub: https://github.com/Pechli-Lab/SurvdigitizeR and https://pechlilab.shinyapps.io/SurvdigitizeR/ .

Health-care practitioners' preferences for the return of secondary findings from next-generation sequencing: a discrete choice experiment.

PURPOSE: Health-care practitioners' (HCPs) preferences for returning secondary findings (SFs) will influence guideline compliance, shared decision-making, and patient health outcomes. This study aimed to estimate HCPs' preferences and willingness to support the return (WTSR) of SFs in Canada. METHODS: A discrete choice experiment estimated HCPs' preferences for the following attributes: disease risk, clinical utility, health consequences, prior experience, and patient preference. We analyzed responses with an error component mixed logit model and predicted WTSR using scenario analyses. RESULTS: Two hundred fifty participants of 583 completed the questionnaire (completion rate: 42.9%). WTSR was significantly influenced by patient preference and SF outcome characteristics. HCPs' WTSR was 78% (95% confidence interval: 74-81%) when returning SFs with available medical treatment, high penetrance, severe health consequences, and patient's preference for return. Genetics professionals had a higher WTSR than HCPs of other types when returning SFs with clinical utility and patient preference to know. HCPs >55 years of age were more likely to return SFs compared with younger HCPs. CONCLUSION: This study identified factors that influence WTSR of SFs and indicates that HCPs make tradeoffs between patient preference and other outcome characteristics. The results can inform clinical scenarios and models aiming to understand shared decision-making, patient and family opportunity to benefit, and cost-effectiveness.

Cost-Effectiveness Analysis of Using Loss of Heterozygosity to Manage Premalignant Oral Dysplasia in British Columbia, Canada.

BACKGROUND: Management of low-grade oral dysplasias (LGDs) is complicated, as only a small percentage of lesions will progress to invasive disease. The current standard of care requires patients to undergo regular monitoring of their lesions, with intervention occurring as a response to meaningful clinical changes. Recent improvements in molecular technologies and understanding of the biology of LGDs may allow clinicians to manage lesions based on their genome-guided risk. METHODS: We used a decision-analytic Markov model to estimate the cost-effectiveness of risk-stratified care using a genomic assay. In the experimental arm, patients with LGDs were managed according to their risk profile using the assay, with low- and intermediate-risk patients given longer screening intervals and high-risk patients immediately treated with surgery. Patients in the comparator arm had standard care (biannual follow-up appointments at an oral cancer clinic). Incremental costs and outcomes in life-years gained (LYG) and quality-adjusted life-years (QALY) were calculated based on the results in each arm. RESULTS: The mean cost of assay-guided management was $8,123 (95% confidence interval [CI] $2,973 to $23,062 in 2013 Canadian dollars) less than the cost of standard care. This difference was driven largely by reductions in resource use among people who did not develop cancer. Mean incremental effectiveness was 0.18 LYG (95% CI 0.08 to 0.39) or 0.64 QALY (95% CI 0.46 to 0.89). Sensitivity analysis suggests that these findings are robust to both expected and extreme variation in all parameter values. CONCLUSION: Use of the assay-guided management strategy costs less and is more effective than standard management of LGDs. IMPLICATIONS FOR PRACTICE: The findings of this study strongly suggest that the use of a risk-stratification method such as a genomic assay can result in improved quality-adjusted survival outcomes for patients with low-grade oral dysplasia (LGD). The use of such an assay in this study provides "precision medicine," allowing for a change in follow-up frequency or early intervention as compared with current standard care. As genomic technologies become more common in cancer care, it is hoped that such an assay, once validated, will become part of a new model for the standard management of LGDs in similar health systems.

'Real-world' health care priority setting using explicit decision criteria: a systematic review of the literature.

BACKGROUND: Health care decision making requires making resource allocation decisions among programs, services, and technologies that all compete for a finite resource pool. Methods of priority setting that use explicitly defined criteria can aid health care decision makers in arriving at funding decisions in a transparent and systematic way. The purpose of this paper is to review the published literature and examine the use of criteria-based methods in 'real-world' health care allocation decisions. METHODS: A systematic review of the published literature was conducted to find examples of 'real-world' priority setting exercises that used explicit criteria to guide decision-making. RESULTS: We found thirty-three examples in the peer-reviewed and grey literature, using a variety of methods and criteria. Program effectiveness, equity, affordability, cost-effectiveness, and the number of beneficiaries emerged as the most frequently-used decision criteria. The relative importance of criteria in the 'real-world' trials differed from the frequency in preference elicitation exercises. Neither the decision-making method used, nor the relative economic strength of the country in which the exercise took place, appeared to have a strong effect on the type of criteria chosen. CONCLUSIONS: Health care decisions are made based on criteria related both to the health need of the population and the organizational context of the decision. Following issues related to effectiveness and affordability, ethical issues such as equity and accessibility are commonly identified as important criteria in health care resource allocation decisions.

Cost-Effectiveness of Molecularly Guided Treatment in Diffuse Large B-Cell Lymphoma (DLBCL) in Patients under 60.

BACKGROUND: Classifying diffuse large B-cell lymphoma (DLBCL) into cell-of-origin (COO) subtypes could allow for personalized cancer control. Evidence suggests that subtype-guided treatment may be beneficial in the activated B-cell (ABC) subtype of DLBCL, among patients under the age of 60. METHODS: We estimated the cost-effectiveness of age- and subtype-specific treatment guided by gene expression profiling (GEP). A probabilistic Markov model examined costs and quality-adjusted life-years gained (QALY) accrued to patients under GEP-classified COO treatment over a 10-year time horizon. The model was calibrated to evaluate the adoption of ibrutinib as a first line treatment among patients under 60 years with ABC subtype DLBCL. The primary data source for efficacy was derived from published estimates of the PHOENIX trial. These inputs were supplemented with patient-level, real-world data from BC Cancer, which provides comprehensive cancer services to the population of British Columbia. RESULTS: We found the cost-effectiveness of GEP-guided treatment vs. standard care was $77,806 per QALY (24.3% probability of cost-effectiveness at a willingness-to-pay (WTP) of $50,000/QALY; 53.7% probability at a WTP of $100,000/QALY) for first-line treatment. Cost-effectiveness was dependent on assumptions around decision-makers' WTP and the cost of the assay. CONCLUSIONS: We encourage further clinical trials to reduce uncertainty around the implementation of GEP-classified COO personalized treatment in this patient population.

Cost-effectiveness analysis of primary human papillomavirus testing in cervical cancer screening: Results from the HPV FOCAL Trial.

The Human Papillomavirus FOr CervicAL cancer (HPV FOCAL) trial is a large randomized controlled trial comparing the efficacy of primary HPV testing to cytology among women in the population-based Cervix Screening Program in British Columbia, Canada. We conducted a cost-effectiveness analysis based on the HPV FOCAL trial to estimate the incremental cost per detected high-grade cervical intraepithelial neoplasia of grade 2 or worse lesions (CIN2+). A total of 19,009 women aged 25 to 65 were randomized to one of two study groups. Women in the intervention group received primary HPV testing with reflex liquid-based cytology (LBC) upon a positive finding with a screening interval of 48 months. Women in the control group received primary LBC testing, and those negative returned at 24 months for LBC and again at 48 months for exit screening. Both groups received HPV and LBC co-testing at the 48-month exit. Incremental costs during the course of the trial were comparable between the intervention and control groups. The intervention group had lower overall costs and detected a larger number of CIN2+ lesions, resulting in a lower mean cost per CIN2+ detected ($7551) than the control group ($8325), a difference of -$773 [all costs in 2018 USD]. Cost per detected lesion was sensitive to the costs of sample collection, HPV testing, and LBC testing. The HPV FOCAL Trial results suggest that primary HPV testing every 4 years produces similar outcomes to LBC-based testing every 2 years for cervical cancer screening at a lower cost.

Resource utilization and costs during the initial years of lung cancer screening with computed tomography in Canada.

BACKGROUND: It is estimated that millions of North Americans would qualify for lung cancer screening and that billions of dollars of national health expenditures would be required to support population-based computed tomography lung cancer screening programs. The decision to implement such programs should be informed by data on resource utilization and costs. METHODS: Resource utilization data were collected prospectively from 2059 participants in the Pan-Canadian Early Detection of Lung Cancer Study using low-dose computed tomography (LDCT). Participants who had 2% or greater lung cancer risk over 3 years using a risk prediction tool were recruited from seven major cities across Canada. A cost analysis was conducted from the Canadian public payer's perspective for resources that were used for the screening and treatment of lung cancer in the initial years of the study. RESULTS: The average per-person cost for screening individuals with LDCT was $453 (95% confidence interval [CI], $400-$505) for the initial 18-months of screening following a baseline scan. The screening costs were highly dependent on the detected lung nodule size, presence of cancer, screening intervention, and the screening center. The mean per-person cost of treating lung cancer with curative surgery was $33,344 (95% CI, $31,553-$34,935) over 2 years. This was lower than the cost of treating advanced-stage lung cancer with chemotherapy, radiotherapy, or supportive care alone, ($47,792; 95% CI, $43,254-$52,200; p = 0.061). CONCLUSION: In the Pan-Canadian study, the average cost to screen individuals with a high risk for developing lung cancer using LDCT and the average initial cost of curative intent treatment were lower than the average per-person cost of treating advanced stage lung cancer which infrequently results in a cure.

Erlotinib or best supportive care for third-line treatment of advanced non-small-cell lung cancer: a real-world cost-effectiveness analysis.

UNLABELLED: Erlotinib has been approved as a third-line treatment for advanced non-small-cell lung cancer (NSCLC) in British Columbia (BC). A cost-effectiveness analysis was conducted to compare costs and effectiveness in patients who received third-line erlotinib to those in a historical patient cohort that would have been eligible had erlotinib been available. METHODS: In a population of patients who have been treated with drugs for advanced NSCLC, overall survival (OS), progression-to-death survival (PTD) and probability of survival one year after end of second-line (1YS) were determined using a Kaplan-Meier survival analysis. Costs were collected retrospectively from the perspective of the BC health care system. RESULTS: Incremental mean OS was 90 days (0.25 LYG), and incremental mean cost was $11,102 (CDN 2009), resulting in a mean ICER of $36,838/LYG. Univariate sensitivity analysis yielded ICERs ranging from $21,300 to $51,700/LYG. CONCLUSION: Our analysis suggests that erlotinib may be an effective and cost-effective third-line treatment for advanced NSCLC compared to best supportive care.

Erlotinib or docetaxel for second-line treatment of non-small cell lung cancer: a real-world cost-effectiveness analysis.

INTRODUCTION: Erlotinib was recently approved in British Columbia (BC) as a second-line treatment for advanced NSCLC. A cost-effectiveness analysis was conducted which compares costs and effectiveness in patients who received second-line erlotinib with those in patients who received docetaxel. METHODS: In a population of patients who have been treated with drugs (either erlotinib or docetaxel) for advanced NSCLC, overall survival (OS), progression-free survival (PFS), and probability of survival 1 year after beginning of second-line treatment (1YS) were determined using Kaplan-Meier and Cox proportional hazard analysis, as well as χ test. Costs were collected retrospectively from the perspective of the BC health care system. RESULTS: Incremental mean OS was 1 day, and incremental mean cost was $2891. Neither costs nor effectiveness were statistically significantly different between groups. PFS and 1YS were also nonsignificantly different. Cox proportional hazard models were used to evaluate multivariate confounding. CONCLUSIONS: Erlotinib and docetaxel are statistically equivalent in terms of treatment cost and overall survival. As treatment practice patterns change, docetaxel may become more frequently prescribed. Therefore, the choice of whether to use erlotinib or docetaxel should be based on factors relating to patient preference rather than costs or effectiveness.