RARE DISEASE DIAGNOSTICS

Among the rare diseases representing about 90% of presently diagnosed rare disease victims, those rare diseases are likely less rare than they are  rarely diagnosed. This leaves their undiagnosed victims to suffer needlessly while their disease festers, gathering comorbidities and demanding more and more expensive care.

There is a solution now available that healthcare systems, insurers, legislators and regulators lag behind in recognising. They must work together to solve a problem costing society an estimated two trillion dollars in each of Europe and the United States.

To date, training well over a million clinicians to identify even hundreds out of thousands of rare medical conditions has been simply impossible, which is why, despite the best efforts of so many nonprofit champions of those diseases, has failed.

I know. As the chair of the Advocacy Committee of one of the most notorious, Complex Regional Pain Syndrome, or CRPS, also known as RSD, the subject of the hit Netflix documentary “Take Care of Maya,” even this one disease with a solution at hand has been an uphill battle. And not for lack of trying. I am uniquely motivated because I am a patient.

CRPS is recognised as the most painful chronic disease known to medicine, scoring a 42 out of 50 on the McGill University Pain Scale. It is disabling with a suicidal ideation of 75% and high comorbidity of medication resistant depression and PTSD, all according to peer reviewed studies. Its most effective drug, ketamine, again highlighted in numerous peer reviewed studies, is controversial and not  approved for use by regulatory authorities. It is, paradoxically, viewed as a drug of abuse despite its record as an opiate sparing drug.

CRPS is just one example of over 10,000 recognised rare diseases in 2024. The vast majority are so rare they have populations of one or two. In the world. Five years ago, there were seven thousand.

The economics are shocking, just shy of one trillion US Dollars in 2019 according to a major US study. More on that in a moment because the technology to diagnose the bulk of them, the most destructive to people and society, a silent mass disabler and killer, is available, but healthcare, from payers to providers, are surprisingly disinterested.

The total number of distinct diseases is less relevant than their collective impact.

A 2019 study funded by the EveryLife Foundation for Rare Diseases and conducted by the Lewin Group found that the economic burden in the United States of just 379 of the 10,000 rare diseases with the most patients (a rare disease is defined as occurring in less than one in about 1,700 people) was just shy of one trillion US dollars annually. Europe being larger than the US in population is likely to meet or exceed the US 2019 number. Today, considering the impact of the pandemic, inflation and other economic changes, both Europe and the US are likely to have easily topped the one trillion US dollar and Euro mark in 2023 and continue to be rising. (Insert Fig. 1 here)

Furthermore, just these rare diseases were estimated to consume about fourteen percent of all US healthcare spending. These outsized costs should make the most common and costly rare diseases a top priority for early detection, diagnosis, and proper treatment when they are most effectively managed, yet there are no economic incentives to do so. In the US, government grants are met with little to no interest by legislators and relevant agencies. Public payers actively resist. Regulators are inflexible, introducing chaos into rare disease healthcare communities with off-label drugs that impoverish patients and invite profiteering. Money for research, effective medications, and cures is non-existent because the large investment required for approval cannot be recovered with so few customers at the back end. Rare, or orphan diseases as they are also known, draw little attention and fewer financial support from private sources.

Despite the public/private disinterest, there is a relatively inexpensive and readily available solution: Employing AI based sub-routines in existing analytic programs to identify rare disease candidates using symptoms and test results common to those diseases. Identifying rare disease candidates is relatively simple as the following five step program shows using healthcare data lakes, matched to a patient’s physician through their electronic medical record (EMR) system using unique patient numbers:

1. Establishing clinical identifiers in test, imaging, and natural language to identify potential candidates,
2. Matching the candidate patient with their physician and uploading the proper diagnostic protocol to their physician’s electronic medical record system,
3. Alerting a central database to download the latest diagnostic protocols to walk the clinician through the proper diagnostic steps, and,
4. If the patient is positive, downloading treatment protocols from the same database, then, finally,
5. Connecting the physician to the relevant nonprofit support group to assist in finding proper specialist support.

Sound complicated? Not really. At scale, a small data center dedicated to the task can “fish” the multiple existing data lakes yielding extraordinary results.

What is a healthcare data lake?

A healthcare data lake is a centralized storage center for vast amounts of structured and unstructured data from disparate sources within the healthcare industry. These include electronic health records from physician’s offices (EHRs), medical imaging reports and images, laboratory and other test results, patientgenerated data, clinical trials data, wearable device data, genomic data, billing records, and other historical data on thousands or millions of patients.

Here's an explanation of key aspects and benefits of a data lake in healthcare:

1. Centralized Storage: A data lake provides a centralized location for storing diverse datasets without needing to pre-structure or format them. This allows for easy access to a wide variety of data types without the need for extensive data transformation.

2. Scalability: Data lakes are designed to scale horizontally, meaning they can handle increasing volumes of data seamlessly, an important factor since health data pours into them at a ferocious rate. This scalability is essential in healthcare, where data volumes are growing rapidly due to advancements in medical technology, increased adoption of electronic health records, and the proliferation of healthcare-related wearable devices feeding data into EHRs.

3. Flexibility: Unlike traditional data warehouses, which require data to be formatted and structured before storage, data lakes accept raw, unprocessed data in its original form, parsed by unique patient identifying numbers. This flexibility allows healthcare organizations to store data from disparate sources without worrying about data format or schema changes while allowing it to remain tied to individual patients.

4. Analytics and Insights: Data lakes enable healthcare organizations, insurers, and analysts to perform advanced analytics and obtain insights from their data limited only by their needs and imagination. By integrating data from various sources, including clinical, operational, and financial data, healthcare providers can gain a comprehensive understanding of patient populations, treatment outcomes, operational efficiency, and, in the case of classifying individual disease candidates and grouping disease clusters to quickly identify outbreaks like COVID quickly and geocentrically.

5. Data Governance and Security: Despite the massive mixing of tens or hundreds of millions of individual data points, robust data governance and security measures are built into data lakes ensuring patient privacy. Data lakes allow organizations to implement granular access controls, encryption, and auditing capabilities to ensure data privacy and regulatory compliance such as HIPAA laws in the United States.

6. Machine Learning and AI: Data lakes serve as a foundation for implementing machine learning and artificial intelligence algorithms. By leveraging the vast amounts of data stored in data lakes, which can be connected to create data oceans, analysts, researchers and scientists can develop predictive models for disease diagnosis, treatment optimization, patient monitoring, and personalized medicine. These capabilities are critical to develop effective treatments and even cures for rare diseases, most which center around personalized medicine strategies.

7. Interoperability: Interoperability is a significant challenge in healthcare, as data is often siloed across different systems and institutions. Data lakes facilitate interoperability by serving as a central data repository where data from various sources can be integrated and accessed seamlessly.

More realistically, even a small nonprofit support group can yield big results on a modest budget with an analytics partner.

Let us pick an excellent candidate, Complex Regional Pain Syndrome, or CRPS, which I have personally worked on a program for from beginning to implementation. CRPS impacts fewer than one of every 2,000 people. A condition that a physician may recognize once or twice in their entire career.

CRPS has an internationally established protocol for diagnosis. While incurable and worsening over time, if diagnosed early, it has a history of being put into remission for years and potentially a lifetime.

Rare diseases like CRPS are rarely diagnosed because physicians are trained to “look for horses when they hear hoofbeats,” otherwise, like Occam’s Razor, the simplest solution is the most likely.

Using indicators and a sub-routine in a powerful analytics program like those available through Optum, Innovaccer, EPIC and many others sifting through data lakes, CRPS candidates can be identified, matched with their primary care physicians, and partnered with downloaded diagnosis and treatment protocols with relative ease.

For example, matching four or five of the following CRPS/RSD Indicators can reliably indicate a candidate:

1. A major or minor trauma or surgery that heals but the pain does not go away and gets worse over time.
2. Pain from the injury spreads to another area or limb.
3. Abnormal hair, nail or skin growth including discoloration and abnormal digit growth such as a toe or finger or multiple toes or fingers.
4. Unexplained edema and/or venial insufficiency.
5. Body temperature instability and temperature differential between one limb to another in excess of 1 degree C.
6. Parkinsonian type tremors or myoclonic spasms.
7. Weakness and/or dystonia in the affected limb.
8. Extreme fatigue.
9. Algesia or hyperalgesia or loss of sensation.

Since CRPS frequently strikes children aged 9 to 14 years old, when remission can be achieved through the use of steroid therapy protocols sparing a lifelong disabling disease, they can be separately be identified using common symptoms including:

1. Severe, Prolonged Limb Pain: This is often described as burning, shooting, or stabbing in nature, or may feel like a "pins and needles" sensation.
2. Allodynia: Pain caused by stimuli that are usually not painful, such as light touch.
3. Hyperalgesia: An increased sensitivity to painful stimuli.
4. Swelling and Changes in Skin Color: Affected limbs may exhibit swelling and changes in skin color, including dry, mottled skin.
5. Functional Impairment: The pain can induce functional impairment, making it difficult for the child to use the affected limb normally.
6. Deep Limb Pain: The pain is often felt deep inside the limbs with a burning, stinging, or tearing sensation.

This is true for many of the most common rare diseases (ironic as it may seem, less than 400 rare diseases represent the vast majority of rare disease patients), virtually all of which have common indicators that can be used to identify candidates using the same method.

Given the massive cost to healthcare systems and society that can be mitigated with a relatively de minimis investment, public/private partnerships with nonprofit rare disease support organizations and academic institutions that have the necessary information to begin these programs and public institutions that have the infrastructure and resources to carry out their implementation must be a pressing priority.

Furthermore, with these mechanisms in place, identifying candidates to participate in studies leading to safe and effective treatments and even cures solve the most difficult element in carrying out those studies with the greatest efficiency and timing.

What is learned here can be shared and implemented throughout the world.

Technology has finally caught up with a great and compelling need. It is time for legislators, government leaders and health officials to act because when they do, a silent epidemic of needless human suffering and impoverishment can be addressed and vanquished.

--Issue 03--