Gábor Benedek

The Importance of Social Embeddedness: Churn Models at Mobile Providers

This article argues the importance of social embeddedness at mobile providers by examining the effects of customers’ network topological properties on churn probability—the probability of a customer switching from one telecommunication provider to another. This article uses data from regional snowball sampling—the only practically feasible network sampling method—to identify groups with significantly different churn ratios for customers with different network topological properties. Clear evidence indicates that individual network characteristics (node-level metrics) have considerable impact on churn probabilities. The inclusion of network-related measures in the churn model allows a longer-term projection of churners and improves the predictive power of the model. With no possibility to carry out repeated sampling, sample stability was checked through simulation results. On the one hand, this article highlights the importance and effectiveness of the providers tailored marketing campaigns by showing that customers targeted by direct marketing campaigns are less threatened by churn than nontargeted customers. On the other, this article shows that social embeddedness blocks the impact of the very same marketing efforts. This article forwards the idea that social embeddedness, also prevalent in vendor switching, can be extended to understanding the development of professional societies threatened by membership churn.

The importance of social embeddedness

This article argues the importance of social embeddedness at mobile providers by examining the effects of customers’ network topological properties on churn probability—the probability of a customer switching from one telecommunication provider to another. This article uses data from regional snowball sampling—the only practically feasible network sampling method—to identify groups with significantly different churn ratios for customers with different network topological properties. Clear evidence indicates that individual network characteristics (node-level metrics) have considerable impact on churn probabilities. The inclusion of network-related measures in the churn model allows a longer-term projection of churners and improves the predictive power of the model. With no possibility to carry out repeated sampling, sample stability was checked through simulation results. On the one hand, this article highlights the importance and effectiveness of the providers tailored marketing campaigns by showing that customers targeted by direct marketing campaigns are less threatened by churn than nontargeted customers. On the other, this article shows that social embeddedness blocks the impact of the very same marketing efforts. This article forwards the idea that social embeddedness, also prevalent in vendor switching, can be extended to understanding the development of professional societies threatened by membership churn.

Pareto improvement and joint cash management optimisation for banks and cash-in-transit firms

Improving the ATM cash management techniques of banks has already received significant attention in the literature as a separate optimisation problem for banks and the independent firms that supply cash to automated teller machines. This article concentrates instead on a further possibility of cost reduction: optimising the cash management problem as one single problem. Doing so, contractual prices between banks and the cash in transit firms can be in general modified allowing for further cost reduction relative to individual optimisations. In order to show the pertinence of this procedure, we have determined possible Pareto-improvement re-contracting schemes based on a Baumol-type cash demand forecast for a Hungarian commercial bank resulting in substantial cost reduction.

Formal Professional Relationships Between General Practitioners and Specialists in Shared Care: Possible Associations with Patient Health and Pharmacy Costs

BackgroundShared care in chronic disease management aims at improving service delivery and patient outcomes, and reducing healthcare costs. The introduction of shared-care models is coupled with mixed evidence in relation to both patient health status and cost of care. Professional interactions among health providers are critical to a successful and efficient shared-care model.ObjectiveThis article investigates whether the strength of formal professional relationships between general practitioners (GPs) and specialists (SPs) in shared care affects either the health status of patients or their pharmacy costs. In strong GP–SP relationships, the patient health status is expected to be high, due to efficient care coordination, and the pharmacy costs low, due to effective use of resources.MethodsThis article measures the strength of formal professional relationships between GPs and SPs through the number of shared patients and proxies the patient health status by the number of comorbidities diagnosed and treated. To test the hypotheses and compare the characteristics of the strongest GP–SP connections with those of the weakest, this article concentrates on diabetes—a chronic condition where patient care coordination is likely important. Diabetes generates the largest shared patient cohort in Hungary, with the highest frequency of specialist medication prescriptions.ResultsThis article finds that stronger ties result in lower pharmacy costs, but not in higher patient health status.ConclusionOverall drug expenditure may be reduced by lowering patient care fragmentation through channelling a GP’s patients to a small number of SPs.

Social network influence on new drug diffusion: Can the data-driven approach provide practical benefits?1

This article studies the determinants of pharmaceutical innovation diffusion among specialists. To this end, it investigates the influences of six categories of factors—social embeddedness, socio-demography, scientific orientation, prescribing patterns, practice characteristics, and patient panel composition—on the use of 11 new drugs for the treatment of type 2 diabetes mellitus in Hungary. The Cox proportional hazards model identifies three determinants—social contagion (in the social embeddedness category) and prescribing portfolio and insulin prescribing ratio (in the prescribing pattern category). First, social contagion has a positive effect among geographically close colleagues—the higher the adoption ratio, the higher the likelihood of early adoption—but no influence among former classmates and scientific collaborators. Second, the wider the prescribing portfolio, the earlier the new drug uptake. Third, the lower the insulin prescribing ratio, the earlier the new drug uptake—physicians’ therapeutic convictions and patients’ socioeconomic statuses act as underlying influencers. However, this finding does not extend to opinion-leading physicians such as scientific leaders and hospital department and outpatient center managers. This article concludes by arguing that healthcare policy strategists and pharmaceutical companies may rely exclusively on practice location and prescription data to perfect interventions and optimize budgets.

Lower fragmentation of coordination in primary care is associated with lower prescribing drug costs-lessons from chronic illness care in Hungary

Improving patient care coordination is critical for achieving better health outcome measures at reduced cost. However, assessing the results of patient care coordination at system level is lacking. In this report, based on administrative healthcare data, a provider-level care coordination measure is developed to assess the function of primary care at system level. In a sample of 31 070 patients with diabetes we find that the type of collaborative relationship general practitioners build up with specialists is associated with prescription drug costs. Regulating access to secondary care might result in cost savings through improved care coordination.

A gyógyszerkiadás és a betegek egészségi állapota a háziorvosi és szakorvosi kapcsolatok függvényében = Formal professional relations between general practitioners and specialists. Possible links with patient health and pharmacy cost

Arra a kerdesre keressuk a valaszt, hogy a szoros haziorvosi-szakorvosi szakmai kapcsolatoknak van-e hatasuk a betegek gyogyszerkiadasara, illetve egeszsegi allapotara. Az orvosok kozotti szakmai kapcsolatok szorossagat a kozosen gondozott betegek szama alapjan hataroztuk meg, mig a betegek egeszsegugyi allapotat a diagnosztizalt es kezelt tarsbetegsegek szamaval mertuk. Hipotezisunk egyreszt az volt, hogy a hatekonyabb koordinacionak koszonhetően a szoros kapcsolatban kezelt betegek jobb egeszsegi allapotuak, masreszt kezelesuk az erőforrasok hatekonyabb felhasznalasa miatt kisebb gyogyszerkoltseggel jar. E ket hipotezist a cukorbetegekre teszteltuk. Azert esett erre a kronikus betegsegre a valasztasunk, mert itt a haziorvosok es a szakorvosok egyuttműkodese elsődleges fontossagu. Magyarorszagon a cukorbetegek eseteben a legnagyobb a kozosen kezelt betegek populacioja, valamint itt a legmagasabb a szakorvosi javaslatra felirt haziorvosi receptek szama. Azt az eredmenyt kaptuk, hogy a szoros kapcsolatban kezelt betegek nem rendelkeznek sem jobb, sem rosszabb egeszsegi allapottal, mikozben a kapcsolodo gyogyszerkiadasuk szignifikansan alacsonyabb. Journal of Economics Literature (JEL) kod: C12, H51, I19.