key: cord-0935934-mdbm9p98
authors: Singh, Rani; Pringle, Theresa; Kenneson, Aileen
title: eP534 Telemedicine challenges and strategies for the medical nutrition therapy of patients with inherited metabolic disorders during the COVID-19 pandemic
date: 2021-04-30
journal: Molecular Genetics and Metabolism
DOI: 10.1016/s1096-7192(21)00612-0
sha: c709a747a2485d24ee6b1cd64cf79d1a37d3ab9c
doc_id: 935934
cord_uid: mdbm9p98

nan

$5,300/QALY Incremental Cost-Effectiveness Ratio threshold in Thailand. Conclusion: We strongly support the consideration of germline genetic testing for BRCA1/2 in all women with breast cancer especially those who meet the NCCN criteria. Measure to make the test reimbursable under Universal Health Care coverage should be done. Promotion of family cascade testing and risk-reducing measure should be supported to make the reimbursement becoming more costeffective.

Clinical geneticists' and primary care physicians' responses to hypothetical clinical scenarios Aileen Kenneson, PhD, MS 1 , Tina Truong, MMSc, CGC 1 , Ami Rosen, MS, CGC 1 , Rani Singh, PhD, RD, LD 1 . 1 Emory University

Objective: Primary care providers (PCPs) are considered the gatekeepers of genetic services. Patient benefits from genetic services include genetics evaluation, risk assessment, genetic counseling, genetic testing, interpretation of genetic testing results, medical management, and psychosocially coping with genetic diagnoses and risks. These services help guide precision medicine, provide appropriate treatment, shorten the diagnostic odyssey, and initiate screenings based on a known familial risk. However, many studies show PCPs underutilize or inappropriately utilize such services, possibly due to a lack of genetic knowledge or low comfort levels regarding utilization of genetic services. We sought to characterize PCP knowledge about utilization of genetic services and barriers to genetic services referrals in the current landscape of genomics. Methods: We developed and administered an online survey to collect information about PCPs' practices related to genetic testing and referrals to genetic services. As part of this survey, PCPs were presented with six hypothetical clinical scenarios and asked to assess whether or not they would recommend genetics evaluation, genetic counseling, and/or genetic testing for each patient: (1) unexplained developmental delay, (2) dyslexia, (3) elevated homocysteine, (4) family history suggestive of familial hypercholesterolemia, (5) breast cancer with no family history, and (6) family history of colon and endometrial cancer. We also presented the scenarios to clinical geneticists. Results from the clinical scenarios for which clinical geneticists were in agreement (> = 80%) were analyzed further. For those scenarios, we compared the characteristics of PCPs and their practices based on whether or not they would recommend genetic testing or referral. Statistical significance was defined as P < 0.05, with no correction for multiple comparisons. Trending associations were defined as P < 0.10. Results: Twenty-five clinical geneticists (13 female, 12 male) responded to the hypothetical clinical scenarios. The percent of clinical geneticists who would recommend a genetics referral or genetic testing for each scenario was: (1) unexplained developmental delay: 87%, (2) dyslexia: 24%, (3) elevated homocysteine: 50%, (4) family history suggestive of familial hypercholesterolemia: 58%, (5) breast cancer with no family history: 23%, and (6) family history of colon and endometrial cancer: 95%. Eighty-one PCPs (52 female, 29 male) responded to the hypothetical clinical scenarios. Primary specialty for the physician respondents were: 46 pediatrics, 26 family practice or general practice, 6 internal medicine, and 3 adolescent medicine. In clinical scenario 1 (unexplained developmental delay), 48% of PCPs indicated that they would refer to genetics or order genetic testing. PCPs who would or would not test/refer were distinguished only by a trend in the distance from the nearest clinical genetics center. PCPs whose practice was within 30 miles of the nearest center (n = 38) were more likely to test or refer than PCPs whose practice was further away (n = 11) (68% versus 36%, p = 0.0809). In clinical scenario 6 (family history of colon and endometrial cancer), 71% of PCPs indicated that they would refer to genetics or order genetic testing. PCPs responses differed only by type of practice. PCPs who worked in health maintenance organizations (n = 10) were less likely to report that they would test or refer than were PCPs who worked in other settings (n = 27) (50% versus 93%, p = 0.0091). Conclusion: While there was agreement among clinical geneticists for the unexplained developmental delay and high-risk cancer scenarios, there was lack of consensus for the other four scenarios. There was the least agreement for elevated homocysteine and a suggested family history of familial hypercholesterolemia. There was less agreement among PCPs when compared to geneticists as to whether or not to pursue genetic testing or referrals. Distance of primary care practice to a clinical genetics center and type of primary practice may influence PCPs' decision making. In conclusion, there is a need for clear recommendations from clinical geneticists about when PCPs should conduct genetic testing or refer to a clinical geneticist. The lack of consensus within the geneticist population that seemed to increase as the cases became less classic indicates that the genetics community may need to spend some time discussing and developing recommendations. Additionally, PCP education about appropriate genetics practices may be helpful, but since previous studies have indicated that increased genetic knowledge does not necessarily result in increased referrals, other potential barriers and facilitators to genetics evaluations should be explored. Objective: The clinical management of patients with inherited metabolic disorders (IMDs) includes medical nutrition therapy (MNT) by registered dietitians (RDs). Disruptions to care of these patients can lead to cognitive impairment, coma or death, depending on the IMD. Furthermore, some IMDs can lead to metabolic decompensation in response to illness. Thus, it is critical that patients with IMDs continue to receive MNT while limiting risk of exposure to the SARS-CoV-2 virus. We sought to characterize the practices of RDs treating patients with IMDs during the COVID-19 pandemic and identify challenges and unmet needs. Methods: We conducted an online survey of RDs who provide MNT for patients with IMDs, including both quantitative and qualitative questions. The survey consisted of questions about the respondent, their practice, their use of telemedicine for direct patient care, barriers to telemedicine, billing and reimbursement related to telemedicine, and institutional emergency preparedness. Qualitative data were analyzed using thematic analysis. Results: We received responses from 117 RDs, who reported using alternate methods to provide MNT for this vulnerable population, including telemedicine. One hundred RDs reported that their institution provided telemedicine to patients, and 96 of these reported that they provided telemedicine visits for MNT for patients with IMDs. Seventy-six RDs reported that there were barriers to the full implementation of telemedicine for patients with IMDs. Barriers included the limitations of virtual visits (inability to conduct physical exams and collect blood samples), time, lack of patient knowledge of technology, audio problems, insurance coverage and billing issues, and limited patient access to internet, computers, or smartphones. RDs reported a variety of strategies to address these barriers, such as limiting non-urgent visits, extending prescriptions without a medical exam, relying on local facilities for blood draws, increasing the number of patients that use at-home filter papers for blood monitoring, and increasing the use of phone calls and emails. Approximately two-thirds of RDs (68.4%) reported that there are patient populations that cannot be seen via telemedicine, including newly diagnosed patients (45.5%), patients in hospitals (33.3%), out-ofstate patients (22.7%), and other (54.5%). The other types of patients included patients without access to technology (n = 19), patients with language barriers (n = 7), patients who lack knowledge or comfort with technology (n = 4), patients using Palynziq (n = 2), patients who need labs or exams that cannot be otherwise coordinated (n = 2), and children in foster care (n = 1). Patients seen via telemedicine who required a blood draw for monitoring received phlebotomy services at the following locations: the originating site (46.9%), a local laboratory (87.5%), their primary care provider's office (37.5%), and other (36.5%). Other locations included home-based bloodspots on filter papers (n = 22), at clinic or hospital (n = 4), and home-based blood draws (n = 1). Seventy-eight of the RD respondents indicated a need related to improving implementation of telemedicine, including training, patient education materials, and provider education materials. Reported benefits of telemedicine included the ability to see patients that previously would not come to clinic due to travel burdens, and patients benefit from decreased travel time. Other benefits to the RDs included increased insight into how patients/families manage their disorders at home. Conclusions: Using telemedicine can be time-consuming, including training of staff, scheduling of patients, and preparing patients for the visit. This increase in workload comes at a time when clinic staff may be redeployed to assist with COVID-19 patients. As a result, RDs report that metabolic clinics are limiting non-urgent visits and extending prescriptions without visits. Limiting visits may result in delayed care with unknown long-term implications at this time. Despite limitations, comments revealed significant support for telemedicine effectiveness and its potential to improve access to care, including people living in rural areas, those without transportation, and those unable to take time off work. By allowing patients to receive services while remaining at home, telemedicine decreases the risk of exposure to the SARS-CoV-2 virus. However, creative strategies are needed to address remaining barriers and unmet needs so that MNT telemedicine services are optimized for patient care, improved outcomes, and sustainability post pandemic. Whole genome sequencing (WGS) requires patient phenotype information for the purpose of variant prioritization and interpretation. Recipients of outpatient genetic testing develop most of their relevant phenotype prior to being seen by a geneticist. This is in contrast to patients undergoing rapid WGS (rWGS) (Sanford et al. 2019) or other inpatient testing, whose full phenotype may not be apparent at the time of admission. A critically ill or medically complex patient's phenotype is likely to evolve well beyond their initial presentation either because of disease evolution or because of diagnostic studies (e.g., MRI). Deeper Phenotyping can enable clinicians to provide genetic testing laboratories with better input data or help themselves assess whether genetic testing is expected to be a productive avenue of inquiry. However, manually curating phenotype terms from medical records on a daily basis is an exhaustive use of human time, and fundamentally at odds with the time-sensitive nature of rWGS. To strike a balance between premature phenotyping and exhaustive manual phenotyping, we should use Natural Language Processing (NLP) tools to aid in the analysis of clinical notes. The Human Phenotype Ontology (HPO) is a standard knowledge base used by practitioners and researchers to discuss phenotypes in a uniform and consistent way (Kohler et al. 2021) . Its hierarchical structure admits many measures of the specificity and concreteness of a given term. One such measurement which has come into common use is the notion of Information Content (IC) (Sanchez et al. 2011) . IC aims to express the level of surprise of a particular event-in this case, the occurrence of a given patient phenotype. The particular formulation of IC that we use here is measured in terms of the number of genes associated with each HPO term (Jagadeesh et al. 2019 ). We examined a cohort consisting of patients under one year of age who were enrolled in a rWGS study at Nationwide Children's Hospital (NCH), N = 22. For each patient, we requested from the Enterprise Data Warehouse all notes from the visit during which the patient enrolled in the rWGS study. These spanned many categories, including care plans, consults, procedures, progress notes, and the text from radiology test results. These notes were processed using Clinphen, an NLP-driven tool designed to automatically extract HPO terms from bodies of text (Deisseroth et al. 2019) . Organizing all terms by the date of their first appearance, we then calculated the information content of the set of terms for each date. This yields a time series that can be used to glean both quantitative and qualitative insights into a patient's phenotypic trajectory. Specifically, we sought to identify inflection points in the IC-time relationship. Inflection points on the Phenotype IC Accumulation curve signal a slowdown in the accumulation of new phenotype information. While multiple inflection points can appear, the first often appears within 10 days of hospital admission (21/22, 95.5%). Moreover, patients accumulated over 50% of the total information content of their hospital admission by the time of their first inflection point in 14 of 22 cases (63.3%). We also saw patients accrue 50% of their total information content within 10 days in 15 of 22 cases (68.2%). Thus, detecting inflection points or specifying a threshold number of days can serve as a meaningful starting point for determining when to put human effort toward careful manual phenotyping in advance of or in parallel to genetic testing. Conversely, this work may help to identify a population of critically ill neonates and children who are unlikely to develop phenotypes informative for purposes of genetic testing in the near term and for whom testing decisions should be made under the assumption that little or no further phenotypic information will be forthcoming.

ClinPhen extracts and prioritizes patient phenotypes directly from medical records to expedite genetic disease diagnosis

Phrank measures phenotype sets similarity to greatly improve Mendelian diagnostic disease prioritization

The Human Phenotype Ontology in 2021

Ontology-based information content computation