Undiagnosed Rare Disease Patients Cost NHS England $4.3 Billion, and then Some

For rare disease patients who have been through a protracted diagnostic odyssey, or those still searching for a name to put to a condition that afflicts them, the emotional and financial burdens of the search of a diagnosis is well known.

Researchers, though, wanted to see if they could put a number to what undiagnosed rare diseases patients cost the National Health Service in England. NHS England collects plenty of data and it is publicly reported in the Hospital Episode Statistics database.

But like many things in the rare disease world, finding the answer was not as straightforward as it might first appear. Imperial College Health Partners, a not-for-profit partnership of NHS providers, universities, and clinical commissioning groups (which are responsible for determining service needs in a local area), conducted the research. Mendelian, a London-based provider of online tools to help accelerate rare disease diagnosis, funded the study.

In terms of the headline, undiagnosed rare disease patients cost the NHS England in excess of $4.3 billion (£3.4 billion) during their diagnosis period over the last 10 years.

The average cost per undiagnosed rare disease patient over a ten-year period reached an average of $16,497 (£13,064), compared to $7,452 (£5,901) for a general hospital population during the same period that also had hospital visits. The undiagnosed rare disease patient had more than twice the cost per patient on average and more costs per treatment on average.

The 38,155 rare disease patients who were at least 10 years old at the time of their diagnosis had an average cost $11,763 (£9,327) compared to (£2,240) for an age-matched comparator group of 3.5 million patients.

“The diagnostic odyssey for rare diseases is often fraught with emotional turmoil and suboptimal care for patients, frustration for clinicians and is also extremely costly for hospitals and associated trusts,” said Peter Fish, head of clinical partnerships at Mendelian. “This new research has given us a rare insight into the extent of this financial burden on the NHS and will hopefully also help us along the path to effective solutions, not just in the U.K. but also globally.”

One problem the researchers faced in conducting their analysis is that the HES database uses the International Classification of Diseases (ICD-10), a coding system that accounts for only about 5 percent of the estimated 7,000 known rare diseases. In addition, multiple diseases (including rare diseases) often share the same coding.

Their solution was to use only discrete codes in the system that identify to ensure that only accurately diagnosed and reported rare disease patients were included in the dataset—some 426 specific rare disease diagnostic codes. The researcher than identified all patients who received their first diagnosis with one of these disease between April 2017 and March 2018, the HES data reporting cycle. Once the cohort was identified, the research gathered ten years of data for these patients. A total of 258,235 patients were in the rare disease group.

A comparator group was created with patients who had any hospital episode between April 2017 and March 2018 and then collected ten-year data for these patients. A total of 27.2 million patients were in the comparator group.

The other issue is that the figures here are based on hospital visits and activity in the lead-up to diagnosis. They exclude the cost of drug treatments and primary care visits, which further understated the cost of being an undiagnosed patient.

“The costs to the NHS highlighted by the research are in fact only the tip of the iceberg, due to the fact that routine healthcare data does not yet fully capture the true extent of the thousands of different rare disease classifications that can pass through the system,” said Julia Wilkins, head of data and analytics at Imperial College Health Partners.

While this may be the tip of the iceberg, we don’t really know how much ice is below the surface. It’s safe to say the report underestimates the actual cost of undiagnosed rare disease to NHS England.

As we move into an age of Big Data, it also illustrates one problem we face with turning all of the data in the health system into actionable information. The coding used in the HES system not only limits its utility for a study like this, but it makes it difficult to determine how big a cost undiagnosed rare disease patients represent within the system. The codes used in the database cover just 5 percent of rare diseases, but it’s unclear what percentage of rare disease patients are covered by those diseases.

Genetic Alliance UK suggests nearly 6 percent of patients in the United Kingdom have a rare disease, but the rare disease patients in the study represents less than 1 percent of the NHS population.

The limits of the data also leave unanswered what is perhaps the most critical question, which is how much money is wasted running the wrong diagnostic tests, treating patients who are misdiagnosed, and the cost of delays in treating progressive conditions that take a bigger toll and require greater interventions than they might otherwise require if caught earlier. That’s not the fault of the study, but the limits of the data.

These are important questions because they will help make a case for the broader use of new diagnostic tools, such as next-generation sequencing, and could help suggest at what price point whole genome sequencing should be broadly deployed.

December 17, 2018
Photo: Julia Wilkins, head of data and analytics at Imperial College Health Partners