P6371Network characteristics of a hypertension referral system in western kenya

• Published in: European heart journal Vol. 40; no. Supplement_1
• Main Authors: Thakkar, A, Valente, T, Andesia, J, Njuguna, B, Miheso, J, Mercer, T, Mwangi, E, Pastakia, S D, Pillsbury, M M, Pathak, S, Kamano, J, Naanyu, V, Vedanthan, R, Bloomfield, G S, Akwanalo, C
• Format: Journal Article
• Language: English
• Published: Oxford University Press 01.10.2019
• Subjects: Hypertension - Prevention
• ISSN: 0195-668X; 1522-9645
• DOI: 10.1093/eurheartj/ehz746.0967

Abstract Introduction The Strengthening Referral Networks for Management of Hypertension Across the Health System (STRENGTHS) trial is creating and testing interventions to improve the effectiveness of referral networks for patients with hypertension in Western Kenya. Purpose Network analysis of facility-based healthcare providers was used to understand the existing network of referrals. The ultimate goal was to identify both structural gaps and opportunities for implementation of the planned intervention. Methods A network survey was administered to providers who deliver care to patients with hypertension asking individuals to nominate a) individuals to whom, and b) facilities to which they refer patients, both up and down the health system. We analyzed survey data using centrality measures of in-degree and out-degree (number of links each provider received and sent, respectively), as well as fitting a core-periphery (CP) model. A higher CP indicates a strong referral network, while a lower CP indicates a relatively weaker network. Results Data were collected from 130 providers across 39 sites within 7 geographically separate network clusters. Each cluster consists of a mix of primary, secondary, and/or tertiary facilities. Compared to a perfect CP referral network model (Correlation Score [CP] = 1.00) and a random referral network model (CP = 0.200), the provider referral networks within each cluster showed a weak tendency for CP structure. There was a large range in CP from 0.334 to 0.639. In contrast, cluster-level facility networks showed a strong tendency for CP structure, with a CP range of 0.857 to 0.949. Core Periphery Correlation Scores [CP] Network Cluster 1 Cluster 2 Cluster 3 Cluster 4 Cluster 5 Cluster 6 Cluster 7 Provider Referrals 0.433 0.424 0.334 0.639 0.535 0.448 0.407 Facility Referrals 0.949 0.894 0.871 0.949 0.949 0.904 0.857 Each cluster represents a geographically separate referral network. A random referral network would reveal a CP score of 0.200; while a perfect referral network would give a CP of 1.00. Referral Network Models Conclusions The current health system across Western Kenya does not demonstrate a strong network of referrals between providers for patients with hypertension. While facility-to-facility referrals are more in-line with a perfect referral model, there are gaps in communication between the specific providers. These results highlight the need for STRENGTHS to design and test interventions that strengthen provider referral patterns in order to improve blood pressure control and reduce cardiovascular risk. Acknowledgement/Funding National Institutes of Health: National Heart Lung and Blood Institute, Doris Duke Charitable Foundation:International Clinical Research Fellowship