Unbiased Analysis of Today's Healthcare Issues

Does Physician income increase from treating more complex patients?

Written By: Jason Shafrin - Jul• 18•11

In Denmark at least, the answer is no.

From the theoretical model we find that higher levels of patient complexity lead GPs [General Practitioners] to choose a lower list size, whereas the effect on income is ambiguous. The effect on total utility (income and leisure) is, however, shown to be negative. Using empirical datafrom 1039 solo practices we find that patient complexity reduces both list size and income and conclude that amixed per capita and fee for service remuneration system does not fully compensate practices with more complexpatients. Differentiated per capita payment may represent a means of ensuring fair and equal income of GPs.”

Differentiated per capita payments may provide a fairer mechanism for compensating physicians for treating more complex patients. This type of reform, however, would also incentivize providers to upcode patient diagnoses in order to increase their per capita payments. Thus, this paper may provide the optimal solution in the case where providers are honest, but this same solution may not be optimal in the case where physicians are potentially dishonest.

The remainder of this post reviews how the authors arrived at the conclusions discussed below.

The Danish Healthcare System

The around 5 million inhabitants in Denmark are served by approximately 3500 GPs organised in 2200 practices. Of those a little more than 1000 GPs are organised in solo practices. The remaining GPs are organised in partnership practices. Denmark operates a public health insurance system, where general practice has the role of gatekeeper for the rest of the healthcare system. The system is tax-funded and covers the whole population. The primary healthcare sector is organised as a list system where all citizens have to be registered with a GP and must contact this GP to receive medical advice and care. Patients are entitled to medical specialist care after referral by their GP only.

All GPs are organised as self-employed private entities and are subjected to the same remuneration scheme, which is based on a mix of fee for service (approximately 70%) and per capita (approximately 30%) payment.

Theoretical Model

The author use a theoretical model in which physicians choose the number of patients to treat. Treating more patients increases income, but decreases leisure. Additionally, the average case mix severity of the GP’s patient roster affects the amount of services–and the amount of time–GPs provide to their patients.

  • Ui = ppcni+pffs (niα(θ)) – C + v(li)

The physician maximizes the utility function above by choosing the number of patients to treat (ni).

The model’s equilibrium predictions are that: 1) list size decreases as patient complexity increases, 2) Overall GP utility decreases as patient complexity increases, and most interestingly 3) differentiated per capita payment can prevent utility loss from serving complex patients. Differentiated capitation payments are larger for sicker patients than for healthier ones.

The data set used contains information on 1039 solo practices from the year 2006. The Health Insurance Register contain data on list size, income and the age and sex of the physician are from the Health Insurance Register. Data on weekly hours of practice personnel and information about whether the practice participates in education of new GPs are from the National Practice Count Survey conducted in December 2005.

Empirical Application

The analytical framework uses two estimation equation.  The first regresses GP characteristics and average patient illness severity on the number of patients on the GP’s rolls.  The second regresses the number of patients on the GP’s rolls, the interaction of the number of patients the GP cares for and the patients’ average illness severity, and GP characteristics on GP income.  The authors measure patient illness severity using the  Danish deprivation index (the DADI index) developed by Vedsted and Sorensen (2009).  To estimate the coefficients of interest, the authors combine the two equation and used a reduced form regression framework, using bootstrapping techniques to derive the confidence intervals.

The key drawback of this paper is that it assumes that patient complexity is exogenous. The authors find that age does have significant explanatory power for predicting patient complexity, but since this variable also has explanatory power in relation to list size, it cannot be used as instrument in the list size equation.

Results

“The GPs’ utility was in the theoretical model assumed to depend on income and leisure. The effect of patient complexity on these two factors in isolation was shown to be ambiguous as both positive and negative effects appear in both cases. However, from the theoretical model we showed that the overall GP utility is always reduced when the level of patient complexity increases. In the empirical section we analysed the impact of patient complexity on income and found that the net impact was negative.  The increase in the fee for service payment, resulting from serving patients with higher complexity, is hence not adequately compensating for the per capita income loss following the reduced list size. This indicates that the remuneration scheme is not sufficiently flexible to ensure equal income across GPs serving different population groups.”

Source:

You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.

One Comment

  1. [...] hospice costs, but if you ask us, a nice cold bath sounds great. Elsewhere, Jason Shafrin asks, “Do physicians make more money when they treat more complex patients?” examining a recent study in Denmark to see whether this has proven true. Denmark sounds [...]

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>