Ace Health Insurance Inc. (AHI) offers health insurance to millions of customers in the US. AHI has created a self-service web portal for its member. Members can find their relevant health insurance information on the portal, such as see the status of their health insurance claims, explore member benefits, and look for treatment options. Members can also conduct transactions with AHI on the web portal, such as ordering a new health insurance card. The portal would allow members to find the desired information and make fewer calls to AHI’s call center.

The AHI members have started using the web portal. The AHI management wants to understand the true effect of web portal adoption and usage on member calls.

The AHI management designed a quasi-natural experiment on the sample of 2475 members, as described below in Figure 1

Figure 1: Quasi-experimental settings

 

Out of the total 2475 members. 1289 members (treated members) started using the web portal in the pre-period, but the remaining 1186 members (control members) did not. The AHI collected data the following data for these members in the pre-and post-periods.

·        Ntel : number of telephone calls made by the member

·        Nweb: number of web portal visits by the member

·        Nclm: number of insurance claims filed for the member

·        Clmamt: Amount of claims in USD filed for the member

The AHI management collected claims information because a member with a higher number claims (higher claim amount) is more likely to call.

The AHI also collected some other socio-demographic information about the members, such as their age, gender, type of health insurance plan (PPO or nonPPO), and their total past claims.

Use this data and the experimental setup to estimate the average treatment effect (ATE) of web portal adoption on member calls.

1.      Estimate the difference-in-means ATE of web portal adoption on member calls based on data for (1) only treated customers and (2) treated and control customers in the post period.

2.     Conduct the balance check to see if the treated and control customers are similar.

3.     Choose appropriate variables for members to compute the propensity scores for their web adoption, and then estimate the matching ATE of the web portal adoption based on the post-period data.

4.     Estimate the difference-in-difference (DID) ATE of the web portal adoption by selecting appropriate variables that may affect member calls.  

5.     Estimate the DID matching ATE of the web portal adoption.

6.     Estimate the DID matching ATE of the web portal usage on member calls. [Hint: use number of web portal visits as the treatment variable instead of the Treat indicator variable.]  

Treated Members (1289) (Started using the web portal in pre-period) Control Members (1186) (Did not start using the web portal in pre-period) One year Pre-period Ntel/Nweb Ntel /Nweb One year Post-period Ntel/Nweb Ntel/Nweb