My Letter to Novitas About My Own Targeted Probe & Educate (TPE) Audit

Dear Drs Patterson and Capeheart,

The facility where I practice, CHI St. Luke’s in The Woodlands, Texas, has been selected for a targeted probe and educate audit (TPE) on Hyperbaric oxygen therapy (HBOT). I attached the letter I received. I understand the reasons why Novitas (and other MACs) have launched TPEs for HBOT, and I do not object to an audit.

Because no one is more interested than I am at stopping improper use of HBOT, I am sending this letter to express my concern that, based on the methods revealed in the attached letter, Novitas will be unable to identify the facilities most at risk of improper use of HBOT.

Here are the reasons I fear the Novitas TPE process is fatally flawed when it comes to identifying facilities for audit:

1. The Novitas TPE is NOT designed to identify excess use of HBOT services:
   • I assumed that the TPE process would target high volume providers of HBOT which would be defined as a large NUMBER OF PATIENTS TREATED WITH HBOT.
   • Clearly the Novitas TPE process is not designed to identify excess use of HBOT if CHI SLEH was targeted for treating a total of FOUR Medicare fee for service patients in a year.
2. The Novitas method of selecting centers for TPE is not designed to identify excessive numbers of treatments per patient:
   • Improper use of HBOT happens if patients receive HBOT who do not need it (generally reflected in a high number of patients treated). We have established that Novitas is not identifying high volume centers since it targeted CHI SLEH The Woodlands.
   • Improper use may happen if patients undergo more treatments than needed on a PER PATIENT basis or if the facility provides TOO FEW treatments for a given condition. If, for example, the average number of treatments needed in appropriately selected diabetic foot ulcers is 34, then providing, for example, only 10 treatments (absent a clinical reason to stop) is not likely to benefit the patient and therefore wastes Medicare resources.
   • The Novitas method to identify clinics for TPE is to target the average number of an incremental portion of a hyperbaric treatment.
   • The average hyperbaric treatment is billed with 4 units of G0277 but in some cases, 5 are needed because patients have trouble ear clearing or a longer protocol is clinically indicated.
   • Some conditions are properly treated with relatively few TOTAL treatments, such as failing flaps.
   • A facility that treats a typical number of HBOT patients annually (which is NOT CHI SLEH The Woodlands with only 4 in a year) finds that about 10%- 20% of patients undergo one or two treatments and then do not return (problems with their ears, claustrophobia, other issues).
   • A center with a high percentage of patients who do not complete a course of therapy is providing poor quality of care
   • Facilities that provide longer hyperbaric treatments will have a higher number of G2077 treatments while providing fewer HBOT treatments per patient. These WILL be targeted for TPE, simply because their treatments are LONGER.
   • If a facility has a high volume of services but has many patients fail to complete HBOT, they WOULD HAVE A LOW AVERAGE TREATMENT NUMBER and thus NOT undergo a TPE, even if they have a high volume of treatments and if other patients are being treated with more treatments per patient than indicated.
3. The Novitas TPE process is not statistically valid:
   • The 4 patients seen at CHI SLEH will be compared to regional norms.
   • In all the other CMS programs of which I am aware (including analysis involved in MIPS), CMS requires a minimum of 20 patients in a group for statistical comparison to be valid.
   • It is not possible to draw a statistically valid conclusion when 4 patients are compared to a region since a very slight anomaly in only one patient will skew the entire dataset.

The simple solution for the above issues is to use the mean number of G services if it is not possible to count specific treatment numbers, and set a threshold of volume below which a TPE is not indicated.

The average number of hyperbaric TREATMENTS provided per patient patient at St. Luke’s The Woodlands (among the FOUR patients treated) is 31. This is the national average, based on the National Hyperbaric Registry which I direct. As you see from the table below, of the FOUR patients being audited, one patient received 49 treatments which skewed the dataset to a higher average. The fact that nearly all patients underwent the planned course of treatment resulted in a “high” per patient average of the “G code”.

The CHI SLEH The Woodlands facility may have the lowest utilization rate for HBOT in the entire Novitas jurisdiction. It has followed the Novitas guidelines religiously and has provided such high quality care that it ensured all patients completed their course of treatment. The Novitas TPE methodology has targeted this facility for TPE. A busy facility that has a high percentage of patients who do not finish treatment will NOT be targeted, based on the methods as described.

I think I can confidently say that the Novitas TPE process will not identify the facilities most at risk of improper use. This is unfortunate for all concerned.

Caroline E. Fife, MD
Medical Director
CHI SLEH The Woodlands, TX