Skip to main content

Chronic Conditions

Overview​

Code

The Chronic Conditions data mart implements two different chronic condition groupers: one defined by CMS and the other defined by Tuva. We started defining chronic conditions in Tuva after struggling to use the CMS logic, either because certain chronic conditions were missing (e.g. non-alcoholic fatty liver disease, MASH, etc.) or because existing definitions were unsatisfactory (e.g. type 1 and type 2 diabetes are considered the same condition by CMS).

Tuva Chronic Conditions are defined and/or reviewed by medically-trained clinical informaticists.

The Chronic Conditions data mart implements two groupers for classifying chronic condition categories:

CMS: This version uses the logic created by CMS. There are 30 CCW Chronic Condition categories, available for file years 2017 forward, and 40 other chronic health, mental health, substance abuse, and potentially disabling condition categories available for file years 2000 forward. These reference only ICD-10 diagnosis codes and have modified look-back periods, qualifying claims, and codes. You can read more about CMS's logic here.

Tuva: Our own classification. We were unsatisified with some of the CMS definitions and therefore created our own version with the input of medically trained clinical informaticists.

Instructions​

Data Requirements​

The CMS grouper uses the following tables from the Tuva Core Data Model:

  • condition
  • patient
  • procedure
  • medical_claim
  • pharmacy_claim

The Tuva grouper uses the following tables from the Tuva Core Data Model:

  • condition
  • patient

Note: The Tuva Project will generate these Core tables. You just need to map your data to the input layer and run the project.

dbt Examples​

# Runs all marts
dbt build

# Runs only the Chronic Conditions mart
dbt build --select tag:chronic_conditions

# Runs only the CMS grouper
dbt build --cms_chronic_conditions

# Runs only the Tuva grouper
dbt build --tuva_chronic_conditions

Data Dictionary​

cms_chronic_conditions_long​

This table contains one record per patient per chronic condition. For example, if a patient has 3 chronic conditions they will have 3 records in this table. Each record includes the condition category, condition, date of onset, most recent diagnosis, and the total count of diagnosis codes that were recorded that are relevant for the condition.

This table is created by running the CMS chronic conditions data mart on data that's been mapped to the core data model.

Primary Keys:

  • patient_id
  • condition

Foreign Keys:

  • claim_id
ColumnData TypeDescriptionTerminology

cms_chronic_conditions_wide​

This table contains a single record per patient with separate binary (i.e. 0 or 1) columns for every chronic condition. If a patient has a particular chronic condition they will have a 1 in that particular column and 0 otherwise.

Primary Keys:

  • patient_id
ColumnData TypeDescriptionTerminology

tuva_chronic_conditions_long​

This table contains one record per patient per chronic condition. For example, if a patient has 3 chronic conditions they will have 3 records in this table. Each record includes the condition category, condition, date of onset, most recent diagnosis, and the total count of diagnosis codes that were recorded that are relevant for the condition.

Primary Keys:

  • patient_id
  • condition
ColumnData TypeDescriptionTerminology

tuva_chronic_conditions_wide​

This table contains a single record per patient with separate binary (i.e. 0 or 1) columns for every chronic condition. If a patient has a particular chronic condition they will have a 1 in that particular column and 0 otherwise.

Primary Keys:

  • patient_id
ColumnData TypeDescriptionTerminology

Analytics​

Prevalence of Tuva Chronic Conditions

In this query we show how often each chronic condition occurs in the patient population.

select
  condition
, count(distinct patient_id) as total_patients
, cast(count(distinct patient_id) * 100.0 / (select count(distinct patient_id) from core.patient) as numeric(38,2)) as percent_of_patients
from chronic_conditions.tuva_chronic_conditions_long
group by 1
order by 3 desc
Prevalence of CMS Chronic Conditions

In this query we show how often each chronic condition occurs in the patient population.

select
  condition_category
, condition
, count(distinct patient_id) as total_patients
, cast(count(distinct patient_id) * 100.0 / (select count(distinct patient_id) from core.patient) as numeric(38,2)) as percent_of_patients
from chronic_conditions.cms_chronic_conditions_long
group by 1,2
order by 4 desc
Distribution of Chronic Conditions

In this query we show how many patients have 0 chronic conditions, how many patients have 1 chronic condition, how many patients have 2 chronic conditions, etc.

with patients as (
select patient_id
from core.patient
)

, conditions as (
select distinct
  a.patient_id
, b.condition
from patients a
left join chronic_conditions.tuva_chronic_conditions_long b
 on a.patient_id = b.patient_id
)

, condition_count as (
select
  patient_id
, count(distinct condition) as condition_count
from conditions
group by 1
)

select 
  condition_count
, count(1)
, cast(100 * count(distinct patient_id)/sum(count(distinct patient_id)) over() as numeric(38,1)) as percent
from condition_count
group by 1
order by 1