Analysis of medication consultation networks and reporting medication errors: a mixed methods study

BACKGROUND: To examine characteristics of verbal consultation about medication within social networks of hospital inpatient medication system, and their associations with medication error reporting. METHOD: The setting was a 90-bed provincial district hospital with 4 wards, 7 physicians, 5 pharmacists, 44 nurses, 5 pharmacist assistants, and 4 unskilled ancillary workers. A mixed method comprising (i) a prospective observational study for investigating incidences and the nature of reporting medication errors, and (ii) a social network analysis for patterns of interaction. RESULTS: Out of 5296 prescriptions, 132 medication errors were reported during the one month study period: an incidence rate of 2.5%. Every incident of medication errors was formally documented through pharmacists. The most frequent medication errors were in pre-transcribing (n = 54; 40.9%). The pharmacists were central in the whole network of consultation on medication with the mean in-degree centrality of 35 (SD 14.9) and mean out-degree centrality of 15.4 (SD 11.1). Two bridging participants were identified who were influential communicators connecting the network (betweenness centrality). Medication error reporting were influenced by (i) participants whose advice is sought and viewed as trustworthy (in-degree centrality; p < 0.001), (ii) sex (p = 0.01), and (iii) level of education (p = 0.04). CONCLUSION: In-degree centrality was the most important network characteristic. A culture of medication safety can be fostered by encouraging consultation about the medication of in-patients within the hospital network where reporting of medication errors is essential. ELECTRONIC SUPPLEMENTARY MATERIAL: The online version of this article (10.1186/s12913-018-3049-2) contains supplementary material, which is available to authorized users.