Drs Ash and Ellis present a new method for producing risk-adjusted reimbursement for primary care services. Their analysis is timely. Several key health reform initiatives currently in play seek to significantly change how primary care services are reimbursed. The new payment systems that underlie Patient Centered Medical Homes, Accountable Care Organizations, and other current health care reform initiatives require robust risk adjustment methods. Fair reimbursement for bundled services requires methods that accurately distinguish health care service needs among patients in primary care provider groups with disparate patient populations.
In this study, expenditures are risk adjusted for patient level differences in age and sex and for hundreds of hierarchical condition category scores. The results demonstrate that the statistical models developed in the analysis perform well and are not overfit to the data. The model explains a very large proportion of the observed variation in patient level expenditures (depending on the subgroup considered) and also of the observed variation among primary care practices in mean per capita expenditures. Predicted expenditures aggregated by primary care practices are well calibrated with the observed expenditures for these practices over the full range of predicted values. From the perspective of a statistical model, the developed risk adjustment method accounts for large differences in the amount of expenditures for primary care attributable to differences in patient case mix. This is an important result, because the fairness of reimbursement for bundled services depends on the extent to which differences in patient needs are accounted for by these adjustments.
The methodology they present for the risk-adjusted reimbursement of bundled primary care services is unique in that it embraces an abstract concept as an outcome measure: the cost of the services that primary care practices should provide to their patients. This abstract concept incorporates both the amount of primary care services that should be provided and the total expenditure in terms of dollars that should be spent for these services. Both components of this concept are difficult to measure.
Expenditures for the components of primary care considered in this study reflect those reported for a large population of commercially insured persons. These are the expenditures for the care that was provided. Expenditures are a poor resource for determining the amount of primary care services that should be provided. Expenditures directly reflect only prices for health care services, and prices provide a highly distorted measure of health care service utilization.1 Even if expenditures accurately measured utilization, it would still be difficult to use this information to measure the amount of services that should be provided. Prior research that has compared actual patient treatment histories with standards of care for the prevention and treatment of acute and chronic conditions has demonstrated that adults in the United States receive about one half of the total care processes recommended for them.2
In their study, expenditures were weighted by the estimated proportion of time that should be required for each component, overall, across all patients. These weights were obtained by asking 5 primary care physicians about the time that they spend on components of primary care, including core services, management of prescription drugs, and coordination of specialty, hospital, and emergency care. The mean experience of these 5 physicians was used to weight the relative value of expenditures in the composite measure. The sum of these weighted expenditures is used as a proxy measure of the dollars that should be available for the delivery of comprehensive primary care in the study population.
Both the expenditure data and the specific weights used in the proxy measure provide a very limited resource for estimating the cost of the services that primary care practices should provide to their patients. The authors acknowledge the limitations of their proxy measure. They indicate that the value of the proxy measure is primarily as an innovation for “conceptualizing, implementing, and testing a credible and flexible approach to predicting primary care need from age and sex and the diagnoses and costs recorded in claims data” and that future research should explore other methods of constructing the primary care activity level proxy outcome.
Standards of care for specific types of acute and chronic illnesses are known, and the amount of resources required to deliver the standard of care can be accurately measured by observing the care delivery process. Direct observation of care delivery in representative primary care practices would provide more valid measurements of the amount and cost of the care that should be provided to specific types of patients. Direct observation has been used to successfully measure the categories and durations of the direct and indirect patient care activities that make up the components of a primary care physician’s work day.3 This approach has also been used to successfully measure the amount of time spent by primary care physicians on direct patient care and other categories of clinical activities in studies assessing the effects of electronic health record systems.4
Methods for the risk-adjusted reimbursement of primary care services are needed. More valid measurements of the amount and cost of the primary care services that should be provided to patients are also needed. Both are required for the fair reimbursement of bundled primary care services.
George J. Stukenborg, PhD, MA
Department of Public Health Sciences, Division of Patient Outcomes, Policy, and Epidemiologic Research, University of Virginia School of Medicine, Charlottesville, VA
1. Reinhardt UE. The many different prices paid to providers and the flawed theory of cost shifting: is it time for a more rational all-payer system? Health Aff (Millwood). 2011;11:2125–2133
2. McGlynn EA, Asch SM, Adams J, et al. The quality of health care delivered to adults in the United States. N Engl J Med. 2003;348:2635–2645
3. Gottschalk A, Flocke SA. Time spent in face-to-face patient care and work outside the examination room. Ann Fam Med. 2005;3:488–493
4. Pizziferri L, Kittler AF, Volk LA, et al. Primary care physician time utilization before and after implementation of an electronic health record: a time-motion study. J Biomed Inform. 2005;38:176–188