Comprehensive genomic profiling (CGP) enables identification of patients eligible for targeted treatments, making it essential in the management of advanced cancer. This retrospective real-world study compared actionable mutations in CGP-tested patients with advanced/metastatic solid tumors to those who received single-gene/small-panel (SP) tests. Patients aged > 18 years with advanced/metastatic solid tumors (including non-small cell lung cancer (NSCLC), colorectal cancer, prostate cancer, breast cancer, or melanoma) with a CGP or SP test reported between 1 January 2018, and 31 December 2022, were included. OncoKB-derived actionability was compared between the two cohorts. Inverse probability of treatment weighting (IPTW) was used to adjust for baseline characteristics. A weighted generalized linear model with log link was used to report actionability ratio (AR) and 95

Precision medicine in oncology using actionable molecular biomarkers to guide treatment selection has been associated with favorable outcomes; however, many potentially eligible patients do not receive it. This Academy of Managed Care Pharmacy Market Insights program sought to characterize unmet needs in biomarker testing among managed care stakeholders, to develop best practice and consensus recommendations to support addressing these needs, and to gain insights on potential quality measures related to biomarker testing. The program used a modified Delphi process and included in-depth interviews with expert panelists, a national survey of managed care professionals, and a consensus survey of experts. Areas of unmet need in biomarker testing identified were education, guidelines and protocols, timeliness, process, and equity. Twenty-two best practices were suggested by managed care experts and other stakeholders; 9 of these best practices achieved consensus. These consensus recommendations addressed biomarker test ordering and test performance, treatment decisions based on biomarker testing, cost-effectiveness of biomarker testing, and health disparities in access to biomarker testing. Opportunities for education and improvements in infrastructure to implement these recommendations were identified. Further investigation is needed to develop quality measures; although, valuable insights were gained.

e20598 Background: Guideline-recommended molecular testing is essential for identifying appropriate targeted therapies (Rx) for treatment of mNSCLC patients (pts). Biomarker testing can be performed by single gene tests, small NGS panels (e.g., < 50 genes), or CGP approaches. There is little evidence on the rate of biomarker testing in the real world setting and the impact of different approaches on Rx utilization and cost of care. This study examined real-world utilization of biomarker testing among mNSCLC pts and outcomes with CGP and non-CGP testing. Methods: De-identified administrative claims data from the Optum Labs Data Warehouse were analyzed to identify newly diagnosed adult mNSCLC pts from 1/2018 to 8/2021; the date of the first claim indicating metastasis was the index date. Continuous enrollment in a commercial (COM) or Medicare Advantage (MA) plan with medical and pharmacy benefits for 12 mo prior to (baseline), and ≥6 mo post index date (follow-up) was required; pts with < 6 mo follow-up due to death were included. Initiation of a line of therapy (LOT1) during follow-up was required. We categorized pts based on receipt and type of biomarker testing prior to LOT1: CGP ( > 50 gene panel), non-CGP (5-50 gene panels or single gene testing), or no testing. Differences in receipt of targeted Rx, overall survival (OS), and total overall per patient per month (PPPM) costs during LOT1 were examined with multivariable regression analyses. Results: 9,945 mNSCLC pts (1,970 COM, 7,975 MA) were identified: 5,484 with no testing, 2,215 with CGP, and 2,246 with non-CGP testing prior to LOT1. Biomarker testing rates prior to LOT1 were low (45%) but increased during the study period from 42% to 48% (p <0.01). Testing rates were higher for the COM vs MA population (49% vs 44%, p <0.01). A higher percent of pts with CGP received targeted Rx compared to the non-CGP and no testing group (17% vs 11% and 5% respectively, p< 0.01); after adjustment, CGP pts were still more likely to receive targeted Rx. Compared to the no testing group, OS was more favorable for the CGP [HR 0.8, 95% CI 0.8-0.9] and non-CGP [HR 0.9, 95% CI 0.8-0.9] groups. There was no significant difference in PPPM costs between tested groups: CGP [CR 1.2, 95%CI 1.1-1.2] vs non-CGP [CR 1.1, 95%CI 1.1-1.2]. Conclusions: Rates of biomarker testing among mNSCLC pts are far from optimal despite well-established guideline recommendations and insurance coverage for testing. There was evidence of improved intermediate outcomes (receipt of targeted Rx) with CGP compared to non-CGP or no testing. In addition, OS was improved for tested pts compared to untested. Interventions to help improve biomarker testing are needed. Given the potential benefits of CGP testing (including assessment of biomarkers that cannot be evaluated using small panels), increasing CGP testing may improve outcomes.

e23109 Background: Immune checkpoint inhibitors (IO) have become a powerful precision therapy option to treat advanced stage cancer with biomarkers such as PD-L1, tumor mutational burden (TMB), and microsatellite instability (MSI) associated with improved patient response rates. Despite this, many patients do not receive genomic testing for all IO biomarkers. Providence, a large US community health system, developed a pathologist-directed testing protocol where comprehensive genomic profiling (CGP) was routinely used at time of diagnosis for advanced cancer patients. Methods: Advanced cancer patients who received CGP (ProvSeq 523) and IHC testing for PD-L1 between 2019-2023 were included in the study. Patients were required to be treated at Providence and were assessed for presence of IO biomarkers and subsequent therapy selection. Real world data were curated from patient charts and genomic laboratory data by employing a novel natural-language processing (NLP) approach to accelerate abstraction. Results: The study included 2,502 patients (53% female, median age 68y, 83% white). Top 3 tumor types tested were lung (40%), colorectal (9%), and breast (8%). Overall, 58% (N = 1,455) of patients had presence of ≥1 IO biomarker, with 46% (N = 1,155) being PD-L1 positive, and 27% (N = 682) being TMB-H. In PD-L1 negative patients (N = 1,347; 54%), 300 (22%) were TMB-H and 18 (1%) were MSI-H. 53% (N = 767) of patients who harbored an actionable IO biomarker received IO-based therapy. Fewer patients with an IO biomarker received chemotherapy compared to patients without an IO biomarker (23% vs 35%, p < 0.001). Patients who possessed ≥2 IO biomarkers received an IO precision therapy 68% of the time vs 47% in patients with 1 IO biomarker (p < 0.001). In PD-L1 negative patients, 26% (N = 78) of TMB-H patients received IO monotherapy compared to 5% (N = 49) of TMB-L patients. Conclusions: IO biomarker presence is associated with increased precision therapy use. CGP identified many TMB-H patients for IO monotherapy that would have been missed with PD-L1 testing alone. More than half of patients eligible for IO therapy don’t receive it; ongoing analyses will evaluate the impact of pathology-directed reflex testing on IO as well as targeted therapy approaches. [Table: see text]

e13508 Background: ASCO and NCCN guidelines strongly encourage participation in clinical trials though participation rates remain low. Although up to 60% of trials require biomarker information as a prerequisite to participation, guidelines are silent on biomarker testing to assess eligibility. To help inform policy development, we analyzed commercially available NGS panels to assess the capture rate of the specific gene variants required for targeted therapies in clinical trials. Methods: Publicly available gene variant information on 12 high volume NGS tissue and plasma commercial assays representative of panel size was accessed. The JAX-CKB database was queried to identify genomic prerequisites in active phase I-III clinical trials; mRNA and protein expression biomarkers were excluded. We used unpaired t-testing to compare the ability of large versus small panels to detect variants required for targeted treatments. (Somatic gene variant annotations and related content have been powered by The Jackson Laboratory Clinical Knowledgebase (JAX-CKBTM)). Results: Overall, 185 genes were linked to eligibility for targeted therapies in clinical trials (Table). Large panel assays (>80 genes) had a higher average capture rate compared to smaller panels, (84% vs. 46%, p value < 0.000002). Similarly, capture rates in select cancers were higher in larger panels: lung (85% vs 59%), breast (92% vs 54%), prostate (86% vs 34%), pancreatic (84% vs 54%), colorectal (88% vs 52%). All results were statistically significant (p < 0.05). Smaller panels were more likely to exclude large genes (e.g. BRCA), genome-wide signatures (e.g. TMB), fusions, and copy number variants. Conclusions: NGS-based panels with more than 80 genes had higher capture rates for biomarkers required for clinical trials of targeted therapies. Clinicians searching for all relevant treatment options for patients should include large genomic sequencing panels to maximize identification of potential clinical trial eligibility. Professional societies and policy makers interested in enhancing clinical trial enrollment rates should consider the panel size-related variability in capture rates of genetic variants when developing recommendations. [Table: see text]

11154 Background: Despite concerted efforts to improve clinical trial participation (CTP), enrollment remains low. Perceived additional costs to the patient or health plan may be barriers to CTP; low biomarker testing rates may be another. This study assessed differences in characteristics of patients with CTP versus patients without CTP in an insured population with advanced cancer. Methods: A retrospective analysis was conducted using de-identified administrative claims data from commercially insured and Medicare Advantage (MA) enrollees in the Optum Labs Data Warehouse. Patients ≥18y with claims evidence of advanced cancer and systemic therapy between 01/01/2018 and 02/28/2022 were stratified into 2 groups: 1) with CTP: ≥1 claim with a CTP diagnosis (ICD10 Z00.6) and ≥1 claim with a Q modifier on or within 90 days after index date; and 2) without CTP: no claims with a CTP diagnosis or Q modifier. Patients with >1 primary cancer, T-cell therapy, or <360 days baseline or follow-up (unless death) enrollment were excluded. Biomarker testing, targeted therapy use, and healthcare costs per patient per month (PPPM) were assessed in the baseline period. Results: Of 61,490 patients identified, 1753 (3%) had CTP and 59,737 (97%) did not have CTP; 4% of commercial and 3% of MA patients had CTP. Most common tumor types were breast (27%; 2% had CTP), lung (16%; 3% had CTP), digestive tract (15%; 2% had CTP), and prostate (15%; 3% had CTP). Patients with CTP were younger and more frequently had documented biomarker test use in the baseline period than patients without CTP (Table). In unadjusted analyses, the CTP group had higher healthcare costs in the baseline period prior to CTP regardless of insurance type, driven by 51% higher systemic cancer therapy costs and 35% higher ambulatory visit costs. An ongoing propensity score matched analysis will evaluate the impact of CTP on cost of care in the follow-up period. Conclusions: Consistent with prior studies, the overall CTP rate among patients with advanced cancers was low (3%). Baseline costs (prior to CTP) were higher for patients that later enrolled in clinical trials than for patients without CTP. Further research is ongoing to assess differences in follow-up healthcare costs for patients with vs without CTP adjusting for baseline characteristics. [Table: see text]

3124 Background: Next generation sequencing (NGS)-based panels are routinely used in the diagnosis of mNSCLC, guidelines recommending assessing 12 genes for small variants (SV), (EGFR, BRAF, KRAS, MET, ERBB2); fusions (F), (ALK, ROS1, RET, NTRK1, NTRK2, NTRK3); and copy number variants (CNV), (MET) to assess eligibility for targeted drug therapies. Reimbursement is often limited to targeted panels; yet, a comprehensive review of targeted panels’ ability to capture all recommended biomarkers has not been reported. Methods: Targeted panels were classified as ≤50 and 51-100 genes. We searched the DEX Diagnostics Exchange Registry for NGS-based lung cancer and solid tumor targeted panels. Each panel was compared against the 12 recommended genetic variants. If the type of genetic analysis (SV/F/CNV) was unclear, labs were contacted for clarification. Fusion assays by PCR or FISH with reflex testing to NGS were included. Results: 30 targeted NGS panels from 21 commercial labs were identified; 13 (43%) were specific to lung cancer. Six of 13 panels assessing fusions used FISH or PCR not NGS alone. 12 panels reported only SV, 8 small SV/F, 5 SV/CNS, and 5 all three variant types. Only 1 of the 30 panels (51-100 genes) captured all 12 biomarkers with correct variant types. Of 20 panels reporting ALK, RET, and ROS1, only 11(55%) reported fusions. Of the 5 reporting all three variant types, NTRK fusions were most commonly absent. On average, solid tumor panels sequenced twice as many genes (51 vs. 24) but were no more likely to capture all recommended biomarkers ((1/17 (6%) vs. 0/13 (0%), (p>0.05)). The average biomarkers captured in lung vs. solid tumor panels were 7.9 and 6.6, respectively (p>0.05), while in ≤50 gene panels vs. 51-100 gene panels, 7.0 and 7.6, respectively (p>0.05). Conclusions: Of 30 commercially available NGS panels with <100 genes, only one captured all recommended gene variant types for mNSCLC. Failure to capture all biomarkers was independent of panel size or specificity to lung cancer. NGS panels including small variants, fusions, and copy number variants assessed more biomarkers. Medical oncologists and clinical pathologists ordering targeted NGS panels for the management of patients with mNSCLC must ensure they assess all biomarkers relevant to targeted therapies. Insurers reimbursing targeted panels should acknowledge the inadequacy of <100 gene panels alone to provide guideline-concordant molecular diagnostic testing. [Table: see text]

6633 Background: Guideline-recommended molecular testing has become essential for biomarker-guided clinical decision making, particularly for patients with advanced (adv) disease. Several biomarkers have indications that are tumor type-agnostic (starting with MSI in 2017, NTRK in 2018, TMB in 2020, and RET and BRAF in 2022). Biomarker testing is covered by insurers, both Medicare (covers comprehensive genomic profiling [CGP] and non-CGP panels) and commercial (at least non-CGP). This study aimed to understand utilization of biomarker testing across tumor types. Methods: This retrospective analysis used de-identified administrative claims from Optum Labs Data Warehouse. Adult Commercial (COM) and Medicare Advantage (MA) enrollees diagnosed with any 1 of 6 adv cancer types from 1/2018 to 8/2021 were identified; the date of the first claim indicating adv cancer was the index date. Continuous enrollment for 12 months prior to (baseline), and ≥6 months post-index date, unless they died (follow-up) was required. Biomarker testing was captured using Current Procedural Terminology (CPT) codes indicating CGP (> 50 gene panels), non-CGP (at most 5-50 gene panels), or CPT code 81479 (unlisted molecular pathology procedure) during the study period. The primary analysis evaluates testing in the follow-up only, with secondary analyses evaluating testing including the baseline (to account for testing prior to the index date). Results: We identified 16,931 breast (BC), 16,838 non-small cell lung (NSCLC), 8,755 colorectal (CRC), 4,244 pancreatic (PC), 2,610 ovarian (OC), and 1,231 gastric (GC) adv cancer patients meeting study criteria. Overall biomarker testing rates in the follow-up period were: 37% NSCLC, 19% BC, 41% CRC, 35% PC, 51% OC, 35% GC. The Table shows testing rates by cancer, insurance type, and panel size during follow-up. Testing rates were lower among MA patients compared to COM patients. Even considering baseline and follow-up periods, overall biomarker testing rates were low, and lower among MA compared to COM : 48% NSCLC (53% COM, 47% MA), 27% BC (34% COM, 22% BC), 56% OC (61% COM, 53% MA),42% PC (53% COM, 38% MA), 47% CRC (56% COM, 42% MA), 41% GC (54% COM, 35% MA). Conclusions: Considering guideline recommendations, rates of biomarker testing across tumor types are far from optimal despite insurance coverage for testing. Identification of barriers to biomarker testing and interventions to overcome these are needed to improve adherence to biomarker testing guidelines. [Table: see text]

550 Background: Biomarker testing to direct individualized therapy options can optimize cancer patient outcomes, particularly for patients with advanced disease. Guideline-recommended molecular testing, comprehensive genomic profiling [CGP] and non-CGP testing, is covered by both Medicare and commercial private insurers (at least non-CGP).. This study examined real-world utilization of biomarker testing among cancer patients with metastatic disease who received anti-cancer systemic therapy. Methods: A retrospective analysis was conducted using de-identified administrative claims data from the Optum Labs Data Warehouse. Adults identified with 1 of 6 advanced cancer types from 1/2018 to 8/2021 were identified; index date was the first claim date for advanced disease. Patients with diagnosis codes indicating only lymph node involvement around the primary cancer were excluded. Continuous enrollment in a commercial (COM) or Medicare Advantage (MA) health plan with both medical and pharmacy benefits was required for 12 months prior to the index date (baseline), and ≥6 months after the index date (follow-up); patients with <6 months follow-up due to death were included. Receipt of anti-cancer systemic therapy during the follow-up was required. Biomarker testing was captured using Current Procedural Terminology (CPT®) codes indicating CGP (> 50 gene panels) or non-CGP (≤50 gene panels or single gene tests) during the study period. Rates of biomarker testing by cancer and insurance type, and receipt of targeted therapy and/ were assessed. Results: Result: There were 27,434 metastatic cancer patients meeting study criteria: 10,320 non-small cell lung (NSCLC), 5,525 breast (BC), 5,429 colorectal (CRC), 2,314 ovarian (OC), 872 gastric (GC) and 2,974 pancreatic (PC) cancer patients. 66% of patients were MA enrollees. The median age was 70 years and ranged from 67 years for BC patients to 71 years for NSCLC and PC patients. The median follow-up was 383 days, ranging from 246 days (for PC patients) to 564 days (for BC patients). Overall biomarker testing rates in the follow-up period were: 47% for NSCLC, 34% BC, 58% CRC, 61% OC, 48% GC, and 47% PC. Testing rates were lower among MA patients compared to COM patients (46% vs 53%). Receipt of targeted/monoclonal antibody therapy was higher (47% vs 33%, p<0.01) among patients with biomarker testing compared to those without. Conclusions: Rates of biomarker testing across metastatic tumor types are far from optimal despite guideline recommendations and insurance coverage for testing, and may affect quality of care. To improve adherence to biomarker testing guidelines, interventions to help overcome obstacles to biomarker testing are needed. Future analysis with this cohort will examine patient management and outcomes by receipt of testing, timing of testing and type of therapy received.

Several external hardware upgrades have been developed for the APOGEE Spectrographs as part of the Sloan Digital Sky Survey-V (SDSS-V) to improve their radial velocity (RV) precision from a floor of 100-200 m/sec to approx. 30 m/sec. The upgrades include: (1) Back Pressure Regulator (BPR) systems to stabilize the internal instrument LN2 tank boil-off pressure, lessening induced movement of the APOGEE optical bench; (2) Fabry-Perot Interferometer (FPI) calibration sources to improve wavelength calibration; and (3), the use of octagonal core fiber segments in the fiber train to improve radial scrambling. We discuss the fabrication, commissioning, and early performance of these upgrades.

We discuss the field retermination of high-fiber count MTP fiber connectors used with the APOGEE spectrograph at Apache Point Observatory (APO) in 2021. We address lessons-learned, wear-analysis of removed MTPs, and throughput of the fiber train with the newly terminated fibers in SDSS-V. For the past decade the spectrograph at APO, as part of multiple incarnations of the Sloan Digital Sky Survey (SDSS), has relied upon rapid changes of ten MTP connectors, each containing 30 terminated fibers, and all contained within a custom gang connector system. These rapid changes enable the iterative plugging of the gang connector into multiple cartridges with different plug plates to observe various survey fields throughout the night. While robotic Focal Plane Systems have been developed for SDSS-V to replace plug plates, which will minimize the fiber connector cycles, we nonetheless reterminated the most heavily used MTP connectors. The connector cycles had far exceeded manufacturer lifetimes and the overall system throughput was degrading.

Group decisions are crucial in solving complex problems successfully across many socio-technical contexts. The development of distributed group decision systems are problem-specific, literally with no systematic support and so a high level of variability in terms of their effects. There is a lack of generic and comprehensive manner that can support the description of the actual decision-making processes precisely. In this paper, we propose a group decision description language. The language consists of the constructs of agent, protocol, decision rule, and constraint for representing group decisions generically, flexibly, while maintaining local autonomy. A protocol of Triple Assessment for breast cancer has been used as an example to demonstrate its application. The language has been further applied to a Consensus Protocol and a Coordinator Election Protocol to demonstrate its expressiveness.

Abstract Background Multidisciplinary teams (MDT/tumour boards) were first introduced in the 1990s and have experienced little change to their methodology since. Universally used for the treatment of prostate cancer (CaP) in the UK new interventions are proposed to improve MDT efficiency and patient outcomes. Clinical practice guidelines (CPGs), designed to increase the uptake of evidence-based practice, suffer from lack of proper implementation. There is increasing evidence to suggest a gap between CPGs and actual treatment and the use of artificial intelligence an (AI) systems can help to increase the efficiency of the MDT. Methods We evaluated differences in MDT concordance with guidelines in the primary treatment of localised or locally advanced prostate cancer using the Deontics AI based custom clinical decision support software (CDSS) software. 59 paper cases were created by an expert clinician, 9 of which were excluded as they did not meet eligibility criteria. The remaining 50 cases were provided to the CDSS for evaluation. Simultaneously, two physicians assessed each patient case and provided their treatment recommendations. The results were assessed for concordance with UK, European and American guidelines, inter-rater reliability and trends in concordance based on patient variables.Results Overall clinician concordance with guidelines was 76%, while total concordance with all three guidelines was 28%. Overall concordance was highest with NICE guidelines, while total concordance was highest for NCCN guidelines. Inter-rater reliability was highest for the NCCN guidelines. Age < 75 (p <0.001; odds ratio [OR], 35.000), prostate volume < 46.5ml (p =0.047; OR, 4.909), and a Gleason score ≠ 8 (p =0.013; OR, 12.333), were all significantly associated with increased guideline concordance in this study. Conclusions Concordance with CPGs needs to be improved in specific patient groups. This may reflect cognitive bias, cognitive overload, or conflicting guideline recommendations and evidence base. One potential solution may be the integration of CDSS technology into the MDT setting. Citation Format: Vishal Santis, Deborah Enting, Vivek Patkar, Anastasia Chalkidou, John Fox, Danny Ruta, Jonathan K. Makanjuola. The PROState AI Cancer–Decision Support (PROSAIC-DS) pilot study: Clinical decision support technology and its role in prostate cancer MDT meetings [abstract]. In: Proceedings of the AACR Virtual Special Conference on Artificial Intelligence, Diagnosis, and Imaging; 2021 Jan 13-14. Philadelphia (PA): AACR; Clin Cancer Res 2021;27(5_Suppl):Abstract nr PO-095.

Introduction:We report a pathfinder study of AI/knowledge engineering methods to rapidly formalise COVID-19 guidelines into an executable model of decision making and care pathways. The knowledge source for the study was material published by BMJ Best Practice in March 2020.Methods:The PROforma guideline modelling language and OpenClinical.net authoring and publishing platform were used to create a data model for care of COVID-19 patients together with executable models of rules, decisions and plans that interpret patient data and give personalised care advice.Results:PROforma and OpenClinical.net proved to be an effective combination for rapidly creating the COVID-19 model; the Pathfinder 1 demonstrator is available for assessment at https://www.openclinical.net/index.php?id=746.Conclusions:This is believed to be the first use of AI/knowledge engineering methods for disseminating best-practice in COVID-19 care. It demonstrates a novel and promising approach to the rapid translation of clinical guidelines into point of care services, and a foundation for rapid learning systems in many areas of healthcare.

OBJECTIVE:OpenClinical.net is a way of disseminating clinical guidelines to improve quality of care whose distinctive feature is to combine the benefits of clinical guidelines and other human-readable material with the power of artificial intelligence to give patient-specific recommendations. A key objective is to empower healthcare professionals to author, share, critique, trial and revise these 'executable' models of best practice.DESIGN:OpenClinical.net Alpha (www.openclinical.net) is an operational publishing platform that uses a class of artificial intelligence techniques called knowledge engineering to capture human expertise in decision-making, care planning and other cognitive skills in an intuitive but formal language called PROforma.3 PROforma models can be executed by a computer to yield patient-specific recommendations, explain the reasons and provide supporting evidence on demand.RESULTS:PROforma has been validated in a wide range of applications in diverse clinical settings and specialties, with trials published in high impact peer-reviewed journals. Trials have included patient workup and risk assessment; decision support (eg, diagnosis, test and treatment selection, prescribing); adaptive care pathways and care planning. The OpenClinical software platform presently supports authoring, testing, sharing and maintenance. OpenClinical's open-access, open-source repository Repertoire currently carries approximately 50+ diverse examples (https://openclinical.net/index.php?id=69).CONCLUSION:OpenClinical.net is a showcase for a PROforma-based approach to improving care quality, safety, efficiency and better patient experience in many kinds of routine clinical practice. This human-centred approach to artificial intelligence will help to ensure that it is developed and used responsibly and in ways that are consistent with professional priorities and public expectations.

Artificial Intelligence in Medicine is looking for novelty in the methodological and/or theoretical content of submitted papers. Such kind of novelty has to be mainly acknowledged in the area of AI and Computer Science. Methodological papers deal with the proposal of some strategy and related methods to solve some scientific issues in specific domains. They must show, usually through an experimental evaluation, how the proposed methodology can be applied to medicine, medicallyoriented human biology, and health care, respectively. They have also to provide a comparison with other proposals, and explicitly discuss elements of novelty. Theoretical papers focus on more fundamental, general and formal topics of AI and must show the novel expected effects of the proposed solution in some medical or healthcare field.

Economic theory predicts a decrease in valuation as the availability of substitutes increases. This paper describes a contingent valuation (CV) survey that investigates the effect of substitutes on valuation of private market goods. Using an approach that compares willingness to pay (WTP) values elicited from a CV question that accounts for substitutes with WTP values elicited from a similar question without substitutes, we find that allowing for substitutes can moderate WTP values. For the item valued in this study, a hamburger sandwich, allowing for substitutes was associated with a reduction of from 10% to 16% in stated values.

We examine whether allowing field substitutes to be simultaneously valued with a good of interest in a hypothetical valuation survey moderates behavioral bias between hypothetical statements and actual behaviors. We use a within-sample experiment and compare values in a hypothetical survey with values in a non-hypothetical experimental auction. Especially, we test whether a common calibration function exists for the good of interest with and without the presence of substitutes. We then estimate each calibration function and compare calibration factors. We find that hypothetical bias differs with and without the presence of substitutes and allowing the field substitutes to be simultaneously valued with the product of interest in a hypothetical valuation survey reduces hypothetical bias. Our results suggest that considering substitutes in the application of stated preference survey can help identify the true underlying demand conditions.

Background Live donor nephrectomy is an operation that places the donor at risk of complications without the possibility of medical benefit. Rigorous donor selection and assessment is therefore essential to ensure minimization of risk and for this reason robust national guidelines exist. Previous studies have demonstrated poor adherence to donor guidelines. Methods We developed a clinical decision support system (CDSS), based on national living donor guidelines, to facilitate the identification of contraindications, additional investigations, special considerations, and the decision as to nephrectomy side in potential living donors. The CDSS was then tested with patient data from 45 potential kidney donors. Results The CDSS comprises 17 core tasks completed by either patient or nurse, and 17 optional tasks that are triggered by certain patient demographics or conditions. Decision rules were able to identify contraindications, additional investigations, special considerations, and predicted operation side in our patient cohort. Seventeen of 45 patients went on to donate a kidney, of whom 7 had major contraindications defined in the national guidelines, many of which were not identified by the clinical team. Only 43% of additional investigations recommended by national guidelines were completed, with the most frequently missed investigations being oral glucose tolerance testing and routine cancer screening. Conclusions We have demonstrated the feasibility of turning a complex set of national guidelines into an easy-to-use machine-readable CDSS. Comparison with real-world decisions suggests that use of this CDSS may improve compliance with guidelines and informed consent tailored to individual patient risks.

Summary In recent years, there has been massive progress in artificial intelligence (AI) with the development of deep neural networks, natural language processing, computer vision and robotics. These techniques are now actively being applied in healthcare with many of the health service activities currently being delivered by clinicians and administrators predicted to be taken over by AI in the coming years. However, there has also been exceptional hype about the abilities of AI with a mistaken notion that AI will replace human clinicians altogether. These perspectives are inaccurate, and if a balanced perspective of the limitations and promise of AI is taken, one can gauge which parts of the health system AI can be integrated to make a meaningful impact. The four main areas where AI would have the most influence would be: patient administration, clinical decision support, patient monitoring and healthcare interventions. This health system where AI plays a central role could be termed an AI-enabled or AI-augmented health system. In this article, we discuss how this system can be developed based on a realistic assessment of current AI technologies and predicted developments.