First Phase of Integrated Care Initiative PAIK in Estonia: Lessons Learned

Abstract

Background: Patients with chronic conditions often experience fragmented care and suboptimal health outcomes. Therefore, initiatives to improve the care-continuum of high-risk patients have emerged in Estonia. PAIK project was the first which tried to integrate social-services sector, specialist care and primary healthcare to co-ordinate the care for high-risk elderly patients with social health determinants. Although we measured improved perception of quality of care and level of integration, pragmatic analysis with matched controls failed to improve health outcomes and reduce costs.
 Aim: To analyse and describe lessons learned from PAIK initiative and if intervention or implementation failure explains the negative results of the project.
 Method: A qualitative analysis using focused interviews and expert opinion was used.
 Results: Firstly, the PAIK project was more a pragmatic service design and feasibility exercise than a full-fledged study designed to demonstrate better health outcomes. It was started without power analysis and analytical components were added later. No control arm was used and having social health determinants as inclusion criteria prevented retrospectively finding an unbiased propensity matched control group. Analysis showed that majority of the patients included, belonged to a very high-risk stratum (top 1%), where care-management is less effective. Also we found that not all components of the intervention were aligned with reducing hospitalisations and ER-visits.
 Another factor was that the healthcare sector lacked any prior experience with integrated care practice implementation and scaling, so the lack of existing workforce resulted in training activities continuing well into the duration of the project. Mid-study there was a change in clinical leadership, which also resulted in target population adjustment and process redesign.
 Thirdly we failed to identify and recruit champions from the primary care and social services sides, therefore engagement of these sectors was modest and the service was dominantly specialist care focused. The Covid pandemia also coincided causing resource conflicts in the sector both on hospital and primary care side. 
 Despite these above mentioned shortcomings, we consider the study generally a success, as we feel that focusing on social health determinants is important in such services and was very welcomed by both the social service providers and GP nurses. Also due to our project the role of care managers was defined in Estonia and educational programs implemented. Also, a functional care-management software module was developed and tested during the service. Finally our experience has been seen as valuable and has attracted specialist in several other counties to invest in integrated care efforts and there is greater readiness to change.
 Conclusions: Despite the PAIK intervention in Viljandi being well received by the patients and showed better patient reported outcomes, it failed to convincingly reduce health resource utilisation and hospitalisations. Both components of intervention failure and implementation failure are probably responsible. Experience from the project however might improve the success of future initiatives and be part of the foundation for care-delivery model innovation in Estonia.