Organization
CancerCompass / CTCA
Period
2019–2021
Decided by
Attention
Pressure
A family facing a new cancer diagnosis needed the next step and a number to call. The site gave them encyclopedia entries.

Plain-language next steps for 30 million annual visitors

I led digital product for an oncology information and patient-engagement platform with 30 million annual visitors, and turned clinical protocols into plain-language steps patients and families could follow.

a cross-functional R&D team of 14, and the nursing staff who reviewed real patient questions with the product team

The CancerCompass homepage in July 2019: news articles and message-board posts
The homepage in July 2019, as the Wayback Machine archived it: the article-and-message-board site my rebuild replaced.

The platform was overwhelming users with dense clinical prose, and the fix could not be to say less: oncology guidance had to be clinically exact, so every plain-language simplification had to hold up when oncologists and nurses reviewed it. To rebuild the platform, I had to stand up a new technology stack and engineering team while keeping live traffic running.

CancerCompass was a Cancer Treatment Centers of America (CTCA) venture; City of Hope acquired CTCA in 2022, after CancerCompass had wound down. 1

My part

Director of Digital Products & Strategy: the first product and technical hire

  • Led a cross-functional R&D team of 14, onshore and offshore, across three products.
  • Decided which vendors would rebuild the platform, then ran their delivery: three engagements worth more than $2M.
Scope
First product and technical hire; a 14-person R&D team across three products; grew the venture from 2 to 25 FTEs.
The call
Rebuilt the platform around plain-language care pathways and a direct route to a nurse.
What continued
The feedback loops were built so nursing teams could keep tuning the triage rules; the venture wound down before the 2022 acquisition.

Plain-language next steps for 30 million annual visitors · Kanav Jain · kanav.net

I relaunched the platform around plain-language care pathways and direct nursing triage. I put the product team and the nursing staff on a standing review of what patients were asking, so we could retune the triage rules against real questions and still clear oncologist review.

Encyclopedia entries

  1. A family facing a new diagnosis meets dense clinical prose
  2. The site answers with reference material, not a next step
  3. No route to a person who can answer the question

Plain-language pathways and a nurse

  1. Clinical protocols become plain-language steps
  2. A direct route to a nurse
  3. Oncologists and nurses review every simplification before it ships
  4. Nursing teams retune the triage rules against what patients actually ask
The fix could not be to say less: oncology guidance had to stay clinically exact.

Download this comparison

The three products were the health information site, a hospital digital-marketing platform, and a patient-engagement web app. I directed the rebuild of the contact center on Salesforce Health Cloud and set up feedback loops so nursing teams could keep tuning the triage rules and cut avoidable escalations.

Three products, one team
ProductAudienceMy role
Health information site: CancerCompass Patients and families; 30 million annual visitors Relaunched it around plain-language care pathways and direct nursing triage
Hospital digital-marketing platform: CancerCenter, a separate property from CancerCompass Hospital marketing: content, acquisition, and measurement Product direction; defined its place in the rebuild’s stack
Patient-engagement web app: the CancerCompass web application, distinct from the information site above Patients and families Product direction; defined its stack and wrote the Director of Development role to build it
The contact-center rebuild on Salesforce Health Cloud is the layer under both properties in the diagram below.
Web application — CancerCompass the patient-engagement web app Front end React · TypeScript · mobX · Webpack Back end ASP.NET Core · Aurora PostgreSQL Redis · Auth0 Marketing site — CancerCenter the hospital digital-marketing platform Content and acquisition Sitecore Experience Platform Coveo Measurement GA360 · Fullstory · Optimizely Invoca CRM and contact center Salesforce Health Cloud · InRule · QAssign · Cisco IP Communicator Payments: Blackthorn · Stripe · Avalara Video: undecided at time of writing Shared delivery GitLab · Docker · AWS Fargate, ELB, S3, Route 53, CloudWatch · Yuniql QA: BrowserStack · TestRail Web application — CancerCompass the patient-engagement web app Front end React · TypeScript · mobX · Webpack Back end ASP.NET Core · Aurora PostgreSQL Redis · Auth0 Marketing site — CancerCenter the hospital digital-marketing platform Content and acquisition Sitecore Experience Platform Coveo Measurement GA360 · Fullstory · Optimizely Invoca CRM and contact center Salesforce Health Cloud · InRule · QAssign · Cisco IP Communicator Payments: Blackthorn · Stripe · Avalara Video: undecided at time of writing Shared delivery GitLab · Docker · AWS Fargate, ELB, S3, Route 53, CloudWatch · Yuniql QA: BrowserStack · TestRail
The architecture I defined in the July 2020 stack definition, which covers two of the three products: the patient-facing web application and the hospital marketing site, over one shared delivery pipeline and one contact center. In the July 2020 document, the video-chat layer had no vendor chosen yet, so the diagram leaves it unresolved rather than inventing one.

Bounce rate fell 25% and chat conversions rose 267% across 30 million annual visitors after I relaunched the platform and reworked its clinical navigation.

Bounce rate Before 100 After 75 (−25%) Chat conversions Before 100 After 367 (+267%) before = 100 Bounce rate Before 100 After 75 (−25%) Chat conversions Before 100 After 367 (+267%) before = 100

A frightened person is not short of information. The scarce resource is attention and trust, so the product’s job is the next step and a person to call.

Everything I built here assumed it: plain-language pathways in place of encyclopedia entries, and a nurse one step away, with the product team reading what patients asked so the triage rules kept up. The AI products I build now start from the same assumption.

Show sources and citations Hide sources

1 CTCA acquisition

City of Hope agreed to acquire CTCA in December 2021.

Axios’s report of the agreement, on 8 December 2021. City of Hope announced on 2 February 2022 that it had completed the acquisition; the closing itself has no public link in this ledger.

Internal documents

  • Platform tech-stack architecture Defined July 2020 for the rebuild: two properties over one delivery pipeline and one contact center. Redrawn above; the original is an internal CTCA document.
  • Director of Development job description The role I wrote to build the team: a hands-on technology leader for a remote development group (onshore contractors and a contracted team in Eastern Europe), shipping weekly with no separate IT or QA function.

Operating and project records

  • The bounce-rate, chat-conversion, and annual-visitor figures

The CTCA acquisition is sourced to Axios’s report; the venture’s ownership and the traffic, bounce-rate, and chat figures come from my own records.

Advisory and fractional product leadership for healthcare and AI teams — clinical AI deployment, vendor contract review, and zero-to-one healthcare software. I take software from initial demo through hospital IT approval, clinical sign-off, and operational deployment.

Revised

↑ ↓ select ↵ open ⌘K or / toggle