Medicare has a low administrative rate. Medicare pays roughly 4% to 5% of total expenses for administrative costs. The best run health insurers have slightly higher administrative expense ratios (8% to 10%). Is this a slam dunk that Medicare is inherently more efficient than private insurers? Not neccessarily. I want us to consider the following vignette that I am prepping for a new hire class.
Mayhew Insurance has an extremely good idea how much each hospital in the network should be paid each month for pediatric emergency room visits. Big City Academic Hospital gets 17% of total spend, St. Anthony’s the Forgetful gets 2.3%, East Nowhere Community Medical Center tends to get about 1.5% and North Suburban General usually sees 30 admissions and 1.1% of spend. These numbers will bounce around a little bit, especially at the smaller community hospitals as a run of food poisoning or extremely slippery monkey bars will create local clusters. We expect that. We look for the things that are odd and inexplicable.
A data analyst has been reviewing pediatric ER claims and is noticing a spike for the past three months at North Suburban General for pediatric ER. Instead of 30 plus or minus a few, we are now paying for 45 pediatric ER visits a month. The excess incremental ER utilization is costing $10,000 more per month. The analyst notifies her boss who agrees that this looks odd, and they call the medical management team to figure out what is going on as this is expensive and unexpected.
After a week of data diving and talking with the hospital the following things are noted.
- The incremental ER diagnosis codes are low level, unspecified complaint codes.
- Starting about four months ago, North Suburban General has instituted a policy of given a $10 meal voucher to any adult who stays at the ER for more than 4 hours and $8 meal vouchers for kids.
- Most of the increased utilization is on Tuesday night.
- The cafeteria food is not that bad and if you are selective, you can easily find 150 calories per dollar. You can also find 20 calories per dollar. Tuesday night is burrito night in the cafeteria.
- Most of the increased utilization is coming from a cluster of families where the mothers all once had the same last name.
What is going on here?
Four months ago, a family brought their son in for a broken leg. They received X-rays and had it casted up after about five hours in the ER. The mom, kid and his little sister also got meal vouchers. They asked and were told that any ER visit that lasted more than four hours came with meal vouchers.
The family is food and income unstable. Some times the bank account lasts the month, some times it doesn’t. Some times, the parents are working 40 hour in the week, some times they are working 15 hour per week . An effectively free meal that costs four hours of time at the ER is an effective hack. They started to visit the ER when money was getting tight as the waiting room had a TV and some games for the other siblings, and someone could complain about something. The hack worked, the family was fed for the night. The mother told her sisters, and they embraced the hack.
Mayhew Insurance’s plan to beat the hack was to send a social worker with a large stack of burritos and fajitas to the grandmother’s house for the entire family to have a burrito night on us. The social worker has helped a couple of the families hook up with other sources of aid. The other families were given Mayhew Insurance expense account to order pizza or burritos once a week. The total cost to us for these families is now under $125 per month. In the past six months, there was one ER visit from this family and it was legitimate.
The fundamental problem being presented here is not a medical problem. It is a problem of resources. These families lacked sufficient money to buy food, so they found a way to take care of their families given the constraints that they faced. Our solution is a non-medical solution to avoid massive medical costs…..
This is a vigentte for data analytics, but it is also illustrative of the incompleteness of the administrative cost calculation. Traditional Medicare is a fee for service system. They pay a claim as it comes in. If the sudden utilizers were covered by Medicare, they would be generating multiple $750 to $1,000 claims that Medicare would quickly turn around and pay out for an administrative cost of less than a dollar per claim. This elevated ER utilization would have continued for years. Extremely efficient but this is also extremely wasteful.
Using the simple formula of Non-medical Expenses/(non-medical expenses + medical claims expenses) to calculate the administrative cost ratio, Medicare looks awesome. They are paying a buck per claim and they have a high number of mid-dollar claims. Mayhew Insurance looks extremely inefficient as we’re paying far fewer claims, and solving this problem cost about $4,000 in labor plus another $1,500 per year in pizza/burritos. None of that is considered medical expense. Yet our total medical spend is lower and the people who were going to the ER to get a meal are better off as they are not wasting their Tuesday evenings as well as getting a meal.
Fee for service auto-adjudication is inherently cheaper than any system that involves population management. There is some significant value to both the patient (avoiding needless hospital visits) and the population manager (not paying for needless hospital visits) but to get that value costs money.