We want a way to identify (flag or score) the most medically frail rare diseases in Mondo. This came up in a Slack thread with Melissa, Sarah, Tanner and Bradford, and the ask is to start the discussion here so we can track it.
"Medically frail" is not a term most of us use day to day, and it means different things in different places.
In US healthcare policy it has a specific regulatory meaning. Under the ACA / Medicaid, when a state offers an Alternative Benefit Plan that differs from traditional Medicaid, it must exempt people who are "medically frail or otherwise have special medical needs" and let them stay on the traditional plan. The federal floor covers people who are blind or disabled, have a physical/intellectual/developmental disability that limits activities of daily living, have a substance use disorder or a disabling mental disorder, or have "serious or complex" medical conditions. States get wide discretion to define it beyond that floor, and the same status is now being used to exempt people from Medicaid work requirements. So in this world "medically frail" is an eligibility category tied to functional limitation and complexity, not a fixed clinical score.
In clinical medicine "frailty" is a separate, more precise concept: reduced physiologic reserve and increased vulnerability to stressors, measured with validated instruments (Clinical Frailty Scale, Frailty Index, etc.). Historically it is an aging/geriatrics concept, but it is increasingly applied to chronic and severe disease.
The two meanings overlap but are not the same. Tanner made this point in the thread: before we phenotype anything, we need to say who the target audience is and how we are defining a medically frail rare disease. Are we trying to approximate the Medicaid eligibility category (useful for advocacy, coverage, policy)? Or a clinical severity/burden measure (useful for triage, care coordination, research)?
Tanner's suggestion is to use OMIM Phenotypic Series as a starting resource. Phenotypic Series group genetically heterogeneous but clinically similar disorders under one PS accession (e.g. Charcot-Marie-Tooth PS118220, hereditary spastic paraplegia PS303350, DEE PS308350), and OMIM lets you view the clinical synopses of members side by side. That side-by-side phenotype view is a plausible substrate for "phenotyping over" diseases, since some phenotypic features (respiratory failure, feeding dependence, loss of ambulation, severe cognitive impairment, recurrent hospitalization) drive medical frailty far more than others. Mondo could join other sources to this (HPO annotations, Orphanet severity, disability/ADL-relevant phenotypes) to build a first-pass.
Worth noting for the discussion: some state Medicaid programs already publish condition lists that "may qualify" someone as medically frail, but even there the condition alone is not sufficient, severity is still assessed per person. That tension (disease-level flag vs person-level severity) is something we should be honest about up front, since Mondo operates at the disease level.
Related
We want a way to identify (flag or score) the most medically frail rare diseases in Mondo. This came up in a Slack thread with Melissa, Sarah, Tanner and Bradford, and the ask is to start the discussion here so we can track it.
"Medically frail" is not a term most of us use day to day, and it means different things in different places.
In US healthcare policy it has a specific regulatory meaning. Under the ACA / Medicaid, when a state offers an Alternative Benefit Plan that differs from traditional Medicaid, it must exempt people who are "medically frail or otherwise have special medical needs" and let them stay on the traditional plan. The federal floor covers people who are blind or disabled, have a physical/intellectual/developmental disability that limits activities of daily living, have a substance use disorder or a disabling mental disorder, or have "serious or complex" medical conditions. States get wide discretion to define it beyond that floor, and the same status is now being used to exempt people from Medicaid work requirements. So in this world "medically frail" is an eligibility category tied to functional limitation and complexity, not a fixed clinical score.
In clinical medicine "frailty" is a separate, more precise concept: reduced physiologic reserve and increased vulnerability to stressors, measured with validated instruments (Clinical Frailty Scale, Frailty Index, etc.). Historically it is an aging/geriatrics concept, but it is increasingly applied to chronic and severe disease.
The two meanings overlap but are not the same. Tanner made this point in the thread: before we phenotype anything, we need to say who the target audience is and how we are defining a medically frail rare disease. Are we trying to approximate the Medicaid eligibility category (useful for advocacy, coverage, policy)? Or a clinical severity/burden measure (useful for triage, care coordination, research)?
Tanner's suggestion is to use OMIM Phenotypic Series as a starting resource. Phenotypic Series group genetically heterogeneous but clinically similar disorders under one PS accession (e.g. Charcot-Marie-Tooth PS118220, hereditary spastic paraplegia PS303350, DEE PS308350), and OMIM lets you view the clinical synopses of members side by side. That side-by-side phenotype view is a plausible substrate for "phenotyping over" diseases, since some phenotypic features (respiratory failure, feeding dependence, loss of ambulation, severe cognitive impairment, recurrent hospitalization) drive medical frailty far more than others. Mondo could join other sources to this (HPO annotations, Orphanet severity, disability/ADL-relevant phenotypes) to build a first-pass.
Worth noting for the discussion: some state Medicaid programs already publish condition lists that "may qualify" someone as medically frail, but even there the condition alone is not sufficient, severity is still assessed per person. That tension (disease-level flag vs person-level severity) is something we should be honest about up front, since Mondo operates at the disease level.
Related