Data flow for sites from directory sources and providers

Our 2021Q1 prototype is sites-centered. So our most important production table is **sites_all**. It will get these kinds of information from datasource audit tables [fields in brackets], fields for 3. onwards = series of boolean columns:

1. **Location** [address, zip | lat, long]
2. **Direct info** [website | phone]
3. **Types of MAT** [mat_bupe, mat_mtd, mat_ntrex, mat_other, mat_any]
4. **Setting** [hospital, outpatient, residential, otp, etc.]
5. **Clients/groups served** [pregnant women, veterans, lgbtq, adolescents, seniors, active duty military, clients referred from court/judicial system, etc.]
6. **Insurance/payment options** [medicaid, private insurance, medicare, military insurance, low cost sliding scale, private pay, etc.]
7. **Services** [telehealth, primary care, childcare, housing, mental health counseling, individual counseling, group counseling, Spanish, etc.]

Provider data CONTRIBUTES this information to sites focus:
1. Additional **sites** with buprenorphine (including practices not listed in any site directory + some listed sites that directories don't recognize as offering bupe)
2. **Medical specialties**
3. A rough idea of **prescribing capacity** [waiver limit, reached patient limit]
4. Possibly helpful info re: **insurance** (I'm not sure we'll be able to process this information enough to do anything with it in our February delivery, but it's worth noting as a component of NPI entries that public health stakeholders may be able to use)

Providers get LINKED to sites through two factors:
- **Location** [address, zip | lat, long | sometimes phone, fax]
- **Buprenorphine** waivers
Noting the second factor matters, because sites that offer only **methadone** and/or **naltrexone** for MAT have no provider links. In 2020, we gave those less attention, because HFP had asked us to focus on buprenorphine. But for an MAT (treatment) finder, the other two options are just as important.

...

issue に記載されている sites_all の production table、datasource の audit fields、provider-to-site linking factors の説明から始めます。実装のエントリポイントを特定する前に、想定されるデータフローと 2 月の提供範囲を明確にします。完了条件には、記載された MAT の区別を維持しながら directory データと provider データを組み合わせるための合意済みのアプローチが含まれている必要があります。

索引モデルが issue の本文から書いたものです。