By Eric Andrew Kristof, RN, CAHIMS | Healthcare IT Professional | HIMSS Arkansas Chapter Member
About This Series: The HIMSS Word of the Day is a running blog series drawing from the HIMSS Dictionary of Health Information and Technology Terms, Acronyms, and Organizations, 6th Edition (2025) — the authoritative reference for healthcare IT professionals and the primary study resource for CAHIMS and CPHIMS certification. Each post paraphrases the term in plain language, grounds it in real-world healthcare IT context, and adds a bedside nursing perspective you won’t find in a textbook.
📖 The Term: ADG (Ambulatory Diagnostic Group)
Source: HIMSS Dictionary of Health Information and Technology Terms, Acronyms, and Organizations, 6th Edition (2025)
🔍 Plain-Language Definition
The HIMSS Dictionary defines an ADG (Ambulatory Diagnostic Group) as a method of categorizing outpatient episodes, and it refers readers to the related terms ACG (Adjusted Clinical Groups) and APG (Ambulatory Payment Group).
The idea behind it is simple. A person seen in clinics has diagnoses on claims and visit records, and the raw list of codes is too large to use directly. A grouping method sorts those codes into a small number of categories that reflect how much illness a person has and what kind of care it may require. Counting the groups, or using them in a model, gives a measure of morbidity, meaning disease burden. That measure is then used to compare populations fairly, set payments, plan resources, and find patients who may need care management.
A note on names helps with reading the research. The literature often calls the same Johns Hopkins method Aggregated Diagnosis Groups, and some older sources call it Ambulatory Diagnostic Groups. In the studies cited here, the system sorts diagnoses into 32 categories, and those categories are combined into larger systems, such as the Adjusted Clinical Groups system, which produces risk scores and utilization bands.
| Element | Role | Example |
|---|---|---|
| Input data | Diagnosis codes from visits and claims | Outpatient and hospital diagnoses |
| Grouping logic | Sorts diagnoses into categories | Groups diagnoses into set categories |
| ADG categories | Summarize morbidity | 32 categories in studies cited here |
| Risk score | Predicts use or cost | Resource utilization band |
| Comparison | Adjusts for patient mix | Fair clinic or plan comparison |
| Related systems | Build on or sit alongside ADGs | ACG, APG |
⚠️ Exam Alert: An ADG is a method of categorizing outpatient episodes used for case-mix and risk adjustment. It is related to ACG (Adjusted Clinical Groups) and APG (Ambulatory Payment Group). It differs from DRG (Diagnosis-Related Group), which classifies inpatient stays for payment. The key idea is that grouping diagnoses lets organizations compare patient populations fairly.
🌐 Why It Matters: Real-World Healthcare IT Context
Risk adjustment matters whenever providers, clinics, or plans are compared. The studies below test how well diagnosis-based groups predict deaths, use, and costs.
A 2011 study in Medical Care by Austin and colleagues examined whether the Johns Hopkins Aggregated Diagnosis Groups predict mortality in a general adult population. The cohort included all 10,498,413 residents of Ontario between ages 20 and 100 who were alive on their birthday in 2007, and the outcome was death within one year. A model with age, sex, and indicators for 28 of the 32 ADG categories had a c-statistic of 0.917 in both derivation and validation samples, which is excellent discrimination, with very good calibration. Models using the Charlson or Elixhauser comorbidity measures performed slightly worse. The data are administrative and from one province.
A 2012 study in Psychiatry Research by Austin and colleagues applied the same approach to 94,466 Ontario adults diagnosed with schizophrenia. A model with age, sex, and 14 of the 32 ADG categories reached c-statistics of 0.845 and 0.836 in the derivation and validation samples, with very good calibration. This shows the groups carry useful information even within a specific diagnostic population.
A 2015 study in Gaceta Sanitaria by Carrera-Lasfuentes and colleagues followed a cohort of people diagnosed with diabetes in Zaragoza in 2006 through 2010. They compared comorbidity measures, including the number of diagnoses, the number of ambulatory diagnostic groups, and the number of major ambulatory diagnostic groups. For predicting mortality, the number of major groups worked best (c = 0.763), and a model with sex, age, major groups, and hospitalizations reached c = 0.818. The authors conclude that the system predicts resource use and mortality in people with diabetes in their setting.
A 2018 study in the European Journal of Internal Medicine by Corti and colleagues applied the ACG system to the population of the Veneto Region in Italy to explain health care costs from routine administrative data. An age and gender model explained only 8 percent of the variance in 2015 total costs. Adding diagnosis-related variables produced a 23 percent increase in explained variance, and pharmacy-based variables an additional 17 percent. The comprehensive model’s adjusted R-squared was six times that of the age and gender model. The authors argue that accumulated chronic conditions, more than aging itself, explain rising costs.
A 2024 study in BMC Health Services Research by Hosar and colleagues tested ACG risk scores in four Norwegian municipalities among 168,285 adults. The scores distinguished well between levels of GP visits (area under the curve up to 0.84), hospitalization (up to 0.8), and specialist outpatient visits (up to 0.72), but not out-of-hours GP visits (up to 0.62). For municipal services such as nursing homes and home services, a simple age and sex model did better than the risk scores. The study shows that a risk method can be valid for some uses and not others.
A 2017 study in Medical Care by Chang and colleagues looked at whether prescription fill rates, derived by comparing EHR prescriptions with pharmacy claims, improve risk models built on the Johns Hopkins ACG system. In 43,097 primary care patients, adding fill rates modestly improved the explanation of concurrent and total medical costs, and the authors conclude that population-level gains in predicting use were limited. It illustrates how EHR data can supplement claims-based grouping.
Together, the studies show that diagnosis grouping adds a lot beyond age and sex for predicting deaths and costs, and that results depend on the data source, the outcome, and the population.
Common risk adjustment problems:
- Incomplete or inconsistent diagnosis coding
- Models applied to outcomes they were never validated for
- Comparing providers without adjusting for patient mix
- Using claims only, without pharmacy, lab, or EHR data
- Treating a risk score as a clinical judgment about one patient
- Data that lag too far behind to guide current care
What supports sound use of ADG-based tools:
- Complete, accurate diagnosis documentation and coding
- Validation in the local population before use
- Combining claims with pharmacy and EHR data where possible
- Using scores to flag patients for review, not to decide care alone
- Transparency about how groups and scores are calculated
- Regular recalibration as populations and coding change
🏥 The Nurse’s Perspective: Clinical, Bedside, & Workflow
Nurses meet ADG-based tools when care management lists, panel reports, and risk scores appear in population health dashboards.
At the Bedside
Diagnoses recorded at visits become the raw material for these groups. Complete problem lists and accurate documentation of chronic conditions help ensure that a patient’s real illness burden is counted.
In Clinical Workflow
Nurse care managers use risk scores to decide whom to call first, and nurse informaticists help teams review the lists with clinical judgment. They check that high-risk flags lead to a defined action, such as outreach or a visit, and not only to a report.
In EHR Implementation (Go-Live Perspective)
Risk scores and registries rely on data captured during workflows tested before an Epic Go-Live, especially problem lists and diagnosis entry. At-the-elbow support analysts note where staff skip structured diagnosis fields, and that feedback helps leaders protect the quality of the data behind the scores.
🎓 CAHIMS / CPHIMS & HIMSS Perspective
CAHIMS Exam Domain Mapping
ADGs appear in data, analytics, and payment questions:
| CAHIMS Domain | How ADGs Appear |
|---|---|
| Healthcare & Technology Environments | Payment systems, population health |
| Clinical Informatics | Problem lists, diagnosis coding, care management |
| Healthcare Information & Systems Management | Administrative data, risk models, reporting |
| Management & Leadership | Resource planning, performance comparison |
Key Terms to Know Alongside ADG
- ACG (Adjusted Clinical Groups) — case-mix system built on diagnosis groups to predict use and cost
- APG (Ambulatory Payment Group) — outpatient classification used for payment
- Case-Mix — the mix of patients, by illness and need, that a provider treats
- Risk Adjustment — statistical correction for differences in patient risk
- DRG (Diagnosis-Related Group) — classification of inpatient stays for payment
- Multimorbidity — presence of two or more chronic conditions
HIMSS Organizational Position
HIMSS offers education and research on health information and technology topics through its Resource Center. Members can find related information on the HIMSS Resource Center page.
🔗 Explore Further: External References
- Austin et al. (2011), Medical Care — Using the Johns Hopkins Aggregated Diagnosis Groups (ADGs) to predict mortality in a general adult population cohort in Ontario, Canada https://pubmed.ncbi.nlm.nih.gov/21478773/
- Austin et al. (2012), Psychiatry Research — Using the Johns Hopkins Aggregated Diagnosis Groups (ADGs) to predict mortality in a population-based cohort of adults with schizophrenia in Ontario, Canada https://pubmed.ncbi.nlm.nih.gov/22364931/
- Carrera-Lasfuentes et al. (2015), Gaceta Sanitaria — Comorbidity as a predictor of health services utilization and mortality in patients with diabetes https://pubmed.ncbi.nlm.nih.gov/25200482/
- Corti et al. (2018), European Journal of Internal Medicine — A case-mix classification system for explaining healthcare costs using administrative data in Italy https://pubmed.ncbi.nlm.nih.gov/29514743/
- Hosar et al. (2024), BMC Health Services Research — Validity of the Johns Hopkins Adjusted Clinical Groups system on the utilisation of healthcare services in Norway: a retrospective cross-sectional study https://pubmed.ncbi.nlm.nih.gov/39448990/
- Chang et al. (2017), Medical Care — Evaluating the Impact of Prescription Fill Rates on Risk Stratification Model Performance https://pubmed.ncbi.nlm.nih.gov/29036011/
- HIMSS — Resource Center https://www.himss.org/resource-center
Term source: HIMSS Dictionary of Health Information and Technology Terms, Acronyms, and Organizations, 6th Edition (2025). All definitions in this series are paraphrased for editorial purposes. Readers are encouraged to consult the primary source for exact language.
Eric Andrew Kristof, RN, CAHIMS is a Healthcare IT professional and HIMSS Arkansas Chapter member based in Hot Springs Village, AR, with hands-on Epic Go-Live experience and a diverse healthcare and IT background. He writes at kristof.org at the intersection of clinical care and healthcare technology.
Leave a Reply