AnalysIs of Social Networks Among PhysICIans Employed at a MedIcal - PowerPoint PPT Presentation

AnalysIs of Social Networks Among PhysICIans Employed at a MedIcal School MIE 2014 – Istanbul, Turkey YK Yüce, N Zayim, B Oguz, S Bozkurt, F Isleyen, KH Gulkesen Tue, 2nd Sep 2014

Outline ¤ Introduction n Problem: Technology acceptance n Opinion leaders? n Social Network Analysis n Objective of the study ¤ Material and methods n Social network data collection n Setting and Method Identification n Tools and Measures ¤ Results ¤ Discussion

Introduction

The Problem: Technology Acceptance ¨ Acceptance of Information and Communication Technologies (ICT) in healthcare settings ¨ More specifically; physician resistance to ICT ¤ Serious concerns behind it n Workflow change n Spending additional time to patient interactions

Solution Approach n What do we know? n Physicians listen to their colleagues n And no one but their colleagues! n Who are those being referred/listened to? n Opinion leaders n Why not to develop an OL-based technological change? n How to find them? n Social network (analysis ) of physicians n Let’s have a look at these coins: Opinion Leaders and Social Network

Opinion Leaders? n Those whom are asked for their opinions/guidance regarding a subject, course or action n Influentials n Their behavior pattern n Wait for the first responses from the public n Experience the product, idea, course or action n Conceive the climate of it n Realize the advantages and disadvantages n Decide to favor/support or disfavor the idea n Exercise the influence within their surroundings

Opinion Leaders? n They are informally defined – not formally defined n They are reported to n Help overcome issues regarding diffusion of innovations n Increase the rate of diffusion of innovations n Be the change agents

Social Network? n Social structure consisting of entities and relations (ties) among them n Relation can be based on occupation, religion, club (e.g. sports club), etc. n Social Network Analysis n Patterns and dynamics within the social structure

Objective ¤ Identify and analyze the elements of the social network among physicians that can be significant for the development of an Opinion Leaders (OLs) based diffusion strategy to overcome physician resistance towards ICT n Some social network characteristics and opinion leaders among physicians n Medical practice n Technology ¤ Explore the impact of the “mentor system” on the formation of social networks among physicians

Research Questions n Who are monomorphic OLs in the physicians’ social network? n Physician OLs in medical practice n Physician OLs in technology n Who are polymorphic OLs in the physicians’ social network? n Is there a relationship between being an OL in medical practice and being an OL in technology? n How much do physicians interact regarding a technology related subject compared to a medical subject? n How cohesive do the physicians act in their profession and technology?

Material and Methods

Social Network Data Collection n Already available data source? n Existing mass communication/relationship data n E.g. Scientific publications database n Nope? Then collect social network data

Method & Instrument Identification - I n Which social network data collection method? And why? n A total of 10 methods reported in the literature n Five basic methods + five derivative methods n Guideline by Kim D. n Offering a methodological comparison model of available methods based on n Nature and characteristics of methods n Research conditions n Setting

Method & Instrument Identification - II n Best suited method for this study n Sociometric method n Allows unfolding the social network n Reliable and valid n Basic instrument: a questionnaire

The Setting n School of medicine of a state university and its hospital n Three divisions, 39 departments and other units attached to them n Six building blocks within 300 m. Radius n All employed physicians n Regardless of the academic title and amount of time spent working for n N = 757

The Data Collection Instrument - I n A self-completed questionnaire with two questions n A series of steps to optimize for and adapt the questions to the setting n Format n Expressions

The Data Collection Instrument - II n Format n Preferred using “free recall” type of questionnaires n “Roster/recognition” type of questionnaires was not favorable due to organizational structure n In the literature, both types were found equally reliable and valid n Expressions (questions) n Focusing on usual transactions and routine relationships n Found more reliable than questions asking about “specific events in specific time frames” n Addressing informal relations while excluding formal relations among physicians n Formal relations (consultation) among physicians must be ignored

The Data Collection Instrument - III n Name those colleagues to/from whom you would n “ask her/his opinion” n “seek advice or guidance” n “go for advice” regarding; n Technological or computer-related issues/matters n Medical practice/professional issues/matters n Number of potential responses was not set to a priori

The Data Collection Instrument - IV n For each name that the participant provides n Department of the colleague n A communication frequency n daily or almost daily n once or twice a week n once or twice a month n a few times a year n Demographic data - participant’s gender, graduated university, exact date of employment

Measures and Tools for Analyses - I n Social Network Analysis n Individual measures (Opinion Leaders) n In-degree centrality of an individual à the number of individuals who say that they are connected to her/him n Cattel’s method n Whole network measures n Structural cohesion à Network density, average degree n Pajek: a free Social Network Analysis Software n Statistical Analysis n SPSS

Results

Participation n 394 physicians out of 757 (response rate of 52%) n 151 Females (38.3%) n 243 Males (61.7%) n 238 Residents (60.4%) n 156 Faculties (39.6%) n Social networks n Medical practice 522 nodes n Technology 407 nodes

Relations Based on Gender à 73% of ties in medical practice and 86% of ties in technology are to males à Both networks appear to be male-driven Medical Practice Technology Female Male Female Male Female 196 (41.8%) 273 (58.2%) 60 (26%) 171 (74%) Male 131 (18.4%) 581(81.6%) 20 (5.7%) 328(94.3%)

Opinion Leaders - I Medical Practice Technology

Opinion Leaders - II n Cattel’s point of inflection rule on in-degree n Point of inflection for medical practice network is 5 n Point of inflection for technology network is 4 n For being an OL, 6 and 5 are set as minimum in-degrees, respectively n 66 OLs in medical practice, 31 OLs in technology n No female OLs in technology n 16 polymorphic OLs (OL in both medical practice and technology)

Opinion Leaders - III n Weak association between being an OL in technology and being OL in medical practice (φ=0.3, ¡ p<0.05) n OLs are mostly n Faculties (Prof.s and Assoc. Prof.) for medical practice n Residents and Asst.Prof.s for technology

Structural Cohesion à Physicians act more cohesive in their profession than they do in technology Average Input Average All Density Degree Degree Medical Practice 0.0044 2.2625 4.5249 Technology 0.0035 1.4246 2.8452

How About the Mentor System? à Statistically, in both networks, there is a relationship between being a faculty or resident and addressing a peer (p<0.001) Medical Practice Technology Resident Faculty Resident Faculty Resident 294 (42%) 406 (58%) 259 (72.6%) 98 (27.4%) Faculty 4 (0.8%) 477(99.2%) 12 (5%) 224(95%)

Discussion

Discussion n Relatively low density in both networks n Might be due to scattered work places in relatively big campus n In both networks, faculty members prefer to go to their peers for advice in both subjects (95% and 99.2% of ties) n Is “faculties not asking residents for technology related advice” a result of the mentor system? n In medical practice network, residents go for advice to their peers as much as they go to their mentors (faculties) n Some sub-network among residents?

Discussion & Future Work n Other key players - those bridging one sub-network with another, i.e. Brokerage n Betweenness n Components (sub-networks)

Discussion & Conclusion n A diffusion strategy should n Aim to catalyze and increase the interaction of females n Female touch n Increase the interaction on technology between faculties and residents n Can lead to an increase on the interaction on medical practice between faculties and residents

Thank you J