Abstract
Background: A number of reforms have been implemented in Swedish health care to support integrated care for frail older people and to reduce utilization of hospital care by this group. Outcomes and process indicators have been used in pay-for-performance (P4P) schemes by both national and local governments to support developments.
Objective: To analyse limitations in the use of outcome and process indicators to incentivize integrated care for elderly patients with significant health care needs in the context of primary care.
Method: Data were collected from the Region Skåne county council. Eight primary care providers and associated community services were compared in a ranking exercise based on information from interviews and registered data. Registered data from 150 primary care providers were analysed in regression models.
Results and conclusion: Both the ranking exercise and regression models revealed important problems related to risk-adjustment, attribution, randomness and measurement fixation when using indicators in P4P schemes and for external accountability purposes. Instead of using indicators in incentive schemes targeting individual providers, indicators may be used for diagnostic purposes and to support development of new knowledge, targeting local systems that move beyond organizational boundaries.