Outliers: Useless or Critical?

“I don’t see the logic of rejecting data just because they seem incredible.”  –Fred Hoyle

An outlier is an incredible data point, one that is much greater or much smaller than other data points. Outliers can be influential on benchmarks in a small data set, but often they have no effect.  In the case of compensation data, particularly physician contract data, there are two types of outliers:  those that do not compensate for a service (zero value) and those who represent either a very low or very high dollar value.  MD Ranger addresses the “zero” data points by reporting the percent of subscribers who pay for a service.  This statistic helps facilities determine if they should be paying for a service in the first place.

In a large data set like MD Ranger’s, an outlier will have little or no effect on the quantiles.  For example, in a dataset with 50 data points where no provider represents more than 25% of the contracts (to learn more about how we calculate benchmarks, see “Our Approach to Calculating Benchmarks and Market Data”), each data point holds only a 2% weight in the percentile calculation.  If the rest of the data falls fairly close together, then the outlier has a small effect on the quantiles.  The opposite is also true: in a small data set, each data point has a large effect on the percentile calculations and an outlier could greatly affect the percentiles.

julia 110

Print Email

Distribution of Rates in Market Data

Market data can be a fast, reliable, and cost effective way for hospitals and health systems to benchmark their contract rates.  

The distribution of payment rates can be revealing, depending on the attributes of the sample size and specific data set.  When looking at a distribution of market rates, observe the variance in the data.  If the distribution is fairly concentrated and there’s not much difference between rates, this could suggest that rates are very consistent across markets and facility types. However, even if quantile ranges are narrow, it doesn’t mean that there aren’t significant and meaningful outliers.  In a robust data set the outliers don’t impact the range significantly; however, in a large dataset the facilities over the 90th percentile sometimes represent special circumstances that may be valid reasons for higher pay.  Examples might include highly specialized or nationally recognized leaders or physicians with scope of services and hours that are significantly different.  If there is a much larger distribution of rates, it could mean many things.  If the sample size isn’t large and the rates are widely distributed, it could indicate poor quality data or disparate situations.  If the sample size is satisfactory or even large, variation could indicate many things, from a less-competitive market to variations in the types of organizations included in the data.

julia 110

Print Email

Key Considerations for Neonatology Contracts

Neonatology agreements are for professional staffing and physician oversight of licensed neonatal intensive care units.  These agreements provide clinical coverage and administrative oversight of Level 2 or Level 3 neonatal intensive care units, or NICUs, by physicians trained in neonatology.  In some hospitals general pediatricians may provide some staffing, although in such instances a neonatologist is available for consultation at all times.  May contracts in the MD Ranger database pay medical direction only (78%).

Key Factors to Consider While Analyzing Neonatology Contracts:

  • Is the in-house coverage requirement greater than six hours per day or 42 hours per week?
  • Is the proportion of publicly sponsored or uninsured patients extremely high or low?
  • Are physician extenders used and provided by the physician group?
  • Are professional services provider for inter-facility transports to the NICU?
  • Are outreach services provided to other facilities?
  • Are teaching services required and included for academic training programs?

allison

Print Email

Market Data Should Include the Most Helpful Benchmarks

Market data can be a fast, reliable, and cost effective way for hospitals and health systems to benchmark their contract rates.  

If the most important statistics go unreported, market data can be misleading and difficult to apply.  When you know what types of organizations are included within a particular data set, you can find facilities that pay physicians for comparable services.  When looking at surveys, make sure the hospital characteristics that are reported influence payment rates.  MD Ranger has found that the most significant factors influencing rates are trauma status, in-house coverage, and facility size. Urban and rural status can also make a difference.  Additionally, reporting down to the most specific physician specialty is very important.  Within surgical specialties, there is a huge variation in call pay, and some variation in medical director/administrative pay.

julia 110

Print Email

  Product
  MD Ranger Subscription    
  Our Benchmarks    
  Online Platform    
  For Physicians    
  For Post Acute Facilities    
  For Trauma Centers    
  
For Consulting Firms    
      
Company
Team    
Database and Security    
In the News    
Contact Us    
      
      
      
Insights
Contracting    
Compliance    
Webinars    
Briefings and White Papers    
Physician Resources    
The MD Ranger Difference    
Blog    
      
© 2018 MD Ranger, Inc.
1601 Old Bayshore Hwy, Suite 107
Burlingame, CA 94010
650-692-8873
info@mdranger.com
Privacy Policy