ASN's Mission

ASN leads the fight to prevent, treat, and cure kidney diseases throughout the world by educating health professionals and scientists, advancing research and innovation, communicating new knowledge, and advocating for the highest quality care for patients.

learn more

Contact ASN

1401 H St, NW, Ste 900, Washington, DC 20005

email@asn-online.org

202-640-4660

The Latest on Twitter

Kidney Week

Abstract: PO2412

Applying Predictive and Causal Analytics to Design Intelligently Targeted Outreach to Address Underrecognition of CKD

Session Information

Category: CKD (Non-Dialysis)

  • 2102 CKD (Non-Dialysis): Clinical, Outcomes, and Trials

Authors

  • O'Brien, Anthony Terrence, Aetna Inc, Hartford, Connecticut, United States
  • Azari, Ali, CVS Health Corp, Woonsocket, Rhode Island, United States
  • Kipping, Emily, CVS Health Corp, Woonsocket, Rhode Island, United States
  • Culleton, Bruce F., CVS Health Corp, Woonsocket, Rhode Island, United States

Group or Team Name

  • CVS Kidney Care
Background

43% of adults with advanced Chronic Kidney Disease (CKD) are unaware of their diagnosis. 2 priorities for The National Institute of Diabetes and Digestive and Kidney Research are to create diagnostic models of kidney function, and to promote studies with responsive outcomes. We propose a novel approach to identify undiagnosed members and increase CKD testing.

Methods

To target Commercial and Medicare members for CKD testing, we combined 3 techniques: i) machine learning, ii) causal inference, and iii) clinical practice. Our machine learning model predicted members’ risk of stage 3b+ CKD. With causal inference, we calculated the average per member per year costs-difference over 6 years between an early or late CKD test. Using the members’ risk score, their associated costs, and their expected behavior change we created a targeting threshold. When the cost-difference of testing a member was larger than the threshold, we enrolled those members into the study. Our intervention consisted of low-cost member outreach like emails and direct mailers (Dir. M). Members were randomized into a i) control group for Dir. M, ii) a control group for email and Dir. M, iii) Dir. M with interactive voice response (IVR), iv) Dir. M only, v) Dir. M and email, and vi) Dir. M and email with IVR. Logistic regression was used for the primary outcome of testing for CKD.

Results

We enrolled 76,388 members of which 35,933 were allocated to the control group. Baseline features were comparable across test and control (age: 80.4±8.3 vs 80.4±8.3, male: 36% [36.3-36.4%] vs 35.9% [35-36%], diabetes: 40.8% [40.3-41.2%] vs 41.1% [40.5-41.6%], hypertension: 88.8% [88.5-89.1%] vs 88.6% [88.3-88.9%]). Members that received only Dir. M were not ever statistically different from control. In the other test arms, testing increased by 1.6-2.1 percentage points at 90 days (p < 0.05). Post-hoc analysis found an increase in laboratory services, PCP and Nephrologist usage, nephroprotective drug claims, and new or updated CKD diagnosis at 90 and 180 days (all p<0.05). Direct medical costs were unchanged.

Conclusion

Low-cost outreach with individualized targeting led to significant increases in CKD stage diagnosis and care-gap closure. The study was under powered to observe direct medical cost savings.

Funding

  • Commercial Support