Assessing primary care based on "guideline adherence" is wrong

I just reviewed the slides of a presentation Dr. Barbara Starfield gave at a conference this past week in Vancouver Canada.  She wrote:

Patients’ health problems are not synonymous 
with their diagnoses. The health problems of 
people and populations  are not the same as 
the sum of their individual “diseases”. 
Diseases:

•  are professionally defined entities, not 
pathophysiologic “truths” 
•  have variable manifestations, pathways of 
genesis, and prognosis 
•  do not exist in isolation from each other 

Resource use is determined by 

the extent of co-morbidity, not 

by the presence or absence of 

particular diagnoses. 

Management focused primarily on diseases 

does not make sense for primary care. 

The benefits of primary care (person-focused, 

comprehensive, and coordinated) are greatest 

for people with high morbidity burdens. 

This is at least part of the reason why disease 

management has not proven useful in 

improving health. Even the chronic care model 

will not be useful unless it is carried out in the 

context of good primary care. 

Primary care performance measurement using 

disease-oriented guidelines will 

•  create inappropriate incentives in caring for 

people with multiple conditions 

•  create incentives based on flawed evidence 

(inadequate outcomes; unrepresentative 

populations) 

•  create perverse incentives based on easily 

measured rather than important 

characteristics 

•  create incentives for avoiding the care of 

disadvantaged populations 

Gordon Moore