Evaluating aMultipayerMedical Home Intervention TotheEditorAsphysicians involvedinthePennsylvaniaChronic Care Initiative (PACCI), a medical home pilot, we are concerned that the evaluation of the first phase had many inadequacies that steeredDrFriedbergandcolleagues1 to thewrong conclusions. These conclusions have the potential to impede further payment reform and block continued redesign of primary care practices. This study examined data from the first 3 years of the initiative, a time frame that is too short to detect changes resulting from the intervention. Phase 1 of this program was devotedtobuildingthe infrastructureandculture toearnNational Committee for Quality Assurance (NCQA) Patient-Centered Medical Home (PCMH) recognition and become a true medical home; reduction in costs of carewas not a goal. In phase 2, pilot practices added case management and other quality interventions to the PCMH model. Our practice is just starting to observe reductions in emergency department visits, hospitalizations (includingreductions in30-dayreadmissions),and costs of care attributable to PCMH changes and the PACCI. The use of claims data has significant limitations. For example, one of our laboratories does not report to the payers any claims for bloodwork done for capitated patients; physicians instead complete gap reports via electronicmedical record registries for the payers and submit results to the PACCI. These reports were not included in the analysis. Friedbergetal1 stated that“[p]ointestimatessuggested improved performance among pilot practices relative to comparisonpractices.”However, in thepropensity-weightedanalyses,manymorepatientswere included inpilot vs comparison practices. It ismore difficult to get a larger cohort to goal than a smaller cohort. The smaller number of patients in comparison practices therefore makes it more difficult to find statistically significant differences in outcomes. The authors reported that there was not a strict 1:1 match of pilot vs comparison practices, but they did not adequately highlight potentially important implications of thismismatch. Other factors, such as insurance mix, number of trainees in a practice,andnumberofnewpatientsenrolling inapractice,may have had a substantial effect on results. Our practice enrolls a large number of new patients monthly, many with poor controlofhypertension,diabetes, andcholesterol.Gettingnewpatients to goal is more challenging than keeping a stable cohort at goal. It is unclear if comparison practices faced similar challenges or if the authors accounted for these differences.