README

This module produces the demographic part of the HRSA Uniform Data System (UDS) report, which every Health Center Program awardee and look-alike files each year. It covers the Patients by ZIP Code table, Tables 3A, 3B and 4, Table 5 (visits and patients by type of service), Table 6A (selected diagnoses and services), and the eCQM lines of Tables 6B and 7 (quality of care and health outcomes, from OpenEMR's own eCQM engine), following the 2026 UDS Manual (calendar year 2026, due February 15, 2027), as the Universal Report and as the Grant Reports for 330(g), (h) and (i) funding. It also lists the patient records that would make the tables wrong, so staff can fix them before submission, and gives whoever prepares Table 9D (patient service revenue) a worksheet from OpenEMR's accounts receivable.

It doesn't submit anything. Health centers still enter the figures in HRSA's Electronic Handbooks (EHBs), and should check them against their own records first.

Status: 0.17.2, pre-release. The counting rules are tested against hand-worked synthetic cases only. They have not yet been compared with a health center's real UDS submission. If your health center uses OpenEMR and would compare this module's table totals with a report you have already filed (totals only, no patient data shared), please open an issue.

What installing it changes in your database, what it reads and writes, and what it hooks into are listed in FOOTPRINT.md, generated from the source.

Using the module day to day, task by task for each role: see GUIDE.md.

What it gives you

Reports › UDS Report: the tables for a chosen year, in the manual's layout and line numbering. Also:
- a cross-table check (the manual requires the same patient total on every table),
- a CSV download, and
- a "Records to fix" list. Blocking problems keep a patient off a table entirely. Quality problems push a patient into an "unknown" line.
Grant Reports: choose 330(g), 330(h) or 330(i) instead of the Universal Report, on the page and in the CSV. See Grant Reports.
Grant evidence: de-identified counts for grant applications: who the health center serves, and social-needs screening by ZIP, age and insurance. See Grant evidence.
Table 9D worksheet: charges, collections and adjustments by payer line, reclassified as the manual asks, with each adjustment reason and where it suggests it goes. For staff who can open OpenEMR's Collections report. See Table 9D worksheet.
eCQM readiness (Tables 6B and 7): whether this OpenEMR can calculate the year's measures, and which records the calculation can't see, with how to fix each. See eCQM readiness.
Records to fix › patient list: the patients behind each problem, each opening the chart. Viewing the list is written to OpenEMR's audit log.
Settings (administrators only):
- which visit categories count as UDS visits,
- which sites are in the scope of project, school-based, or in or next to public housing,
- the poverty-guideline region,
- the UDS insurance category of each insurance company, and
- the Table 5 line of each provider (suggested from their NUCC taxonomy code) and, optionally, of each visit category.
Three coded demographic fields, added to the Demographics Stats group:
- Housing Status (UDS): housed, shelter, transitional, doubling up, street, permanent supportive housing, other homeless arrangement, homeless (arrangement unknown), or declined.
- Agricultural Worker (UDS): migratory, seasonal, aged or disabled former migratory worker, or declined.
- Military Service (UDS): served (not on active duty now), active duty, has not served, or declined.
OpenEMR's own "Homeless" and "Migrant/Seasonal" fields are free text, which the UDS can't use. They are left untouched and never counted. From 0.16.1, when one of them holds something and the UDS field is empty, Records to fix says so ("the Homeless, etc. field of Demographics has something in it"), in place of "not recorded", so staff record the UDS field from it.

How it counts

Rule (2026 UDS Manual)	Implementation
A patient has at least one countable visit in the year	An encounter in the year with a provider, in a countable visit category (default: every patient category except No Show), at an in-scope site. Record labs, injections, screenings and outreach under categories you leave unticked.
Age is as of December 31	`DOB`
Sex as listed on the birth certificate, with no unknown column (3A)	OpenEMR's `sex` field. Anyone not recorded as Male or Female is listed as a blocking problem.
Race lines 1a–7, ethnicity columns A1–C (3B)	CDC codes from the race and ethnicity lists, rolled up to UDS lines (data/race-ethnicity.json). Several Asian subgroups go to 1g, several Pacific Islander subgroups to 2b, two or more race groups to 6. A race without an ethnicity answer is presumed not Hispanic. The stock ethnicity list has only Hispanic yes/no, so Hispanic patients go to A5 unless the list carries detailed CDC codes.
Limited English proficiency (3B, Line 12)	Preferred language isn't English (or blank or declined), or an interpreter is needed
Income recorded at or within 12 months before the last visit, compared with the HHS poverty guidelines for the health center's location	`monthly_income` × 12 and `family_size`, counted only when the Financial Review Date falls in that window. Monthly income must be one amount (`2000`, `$2,000`, `2,000.50`); a range, a unit or words is counted as Unknown, never guessed. Guidelines are in data/poverty-guidelines.json: 2026 for all regions, 2025 for the 48 contiguous states only.
Primary medical insurance at the last visit, by age 0–17 / 18+	The primary policy in effect on the last visit date. Medicare primary with Medicaid also active counts as dually eligible (Line 9a). The category comes from the clinic's setting, then the payer typology group, then the claim type (data/insurance-categories.json). An insurer that can't be decided is a blocking problem; nothing is guessed.
ZIP codes with 10 or fewer patients combine into "Other ZIP Codes"; people experiencing homelessness without a ZIP use the site's ZIP	As stated
Lines 24 (school-based) and 26 (public housing) depend on the visit site	Sites marked in Settings
Line 25 counts only an affirmative answer, never active duty	"Served" only

Service area (0.17.0)

HRSA's Scope of Project Policy Manual (2026, Chapter 2) says a health center's service area "includes, at a minimum, those ZIP Codes where at least 75 percent of current health center patients reside", from the patient origin it reports in UDS. Under the Patients by ZIP Code table, the report lists the fewest ZIP codes that reach 75% of the year's patients:

How they're chosen: most patients first, and ZIP codes with the same count in ZIP order.
What counts: every ZIP code, including those the table folds into Other ZIP Codes (10 or fewer patients), and patients whose insurance has no UDS column. People experiencing homelessness count at the visit site's ZIP code, as on the table. Unknown residence is left out of both the count and the total, and its number is shown.
Against your Form 5B: enter the service area ZIP codes from your Form 5B in Settings. The report then names any of the 75% ZIP codes that aren't on Form 5B (a change in scope to request, Chapter 10), how many patients live inside Form 5B, and Form 5B ZIP codes where no patient lives (for information: the manual sets a minimum, not a maximum).

The manual changes nothing in what the tables count: its definitions of patient and visit are for scope of project only (Chapter 8, footnote 111).

Table 5: Staffing and Utilization

Visits (Columns B and B2) and patients (Column C) by type of service and personnel. FTEs (Column A) come from HR records, which OpenEMR doesn't hold, so the column is left for the health center to enter in EHB.

Rule (2026 UDS Manual)	Implementation
A visit goes on the line of the provider who saw the patient (page 59)	The line set for the visit's category in Settings, if any; else the line chosen for the provider in Settings; else the one suggested by the provider's NUCC taxonomy code (data/table-5-taxonomy.json, from Appendix A)
Only providers generate visits; medical assistants, LPNs, lab and X-ray staff, pharmacists and others don't (page 64 and Appendix A)	Their lines report no visits, and LPNs and therapy assistants are marked as non-providers on the lines they share with providers. Their encounters are counted as "not Table 5 visits" on the report
Mental health and substance use disorder visits are classified by the primary service, not the provider's title (pages 54–55)	A visit by a mental health or SUD provider whose first diagnosis is an alcohol or other substance use disorder (Table 6A Lines 18–19) goes on Line 21; an SUD specialist's other visits go on Line 20c. A visit with no diagnosis stays on the provider's line, and the report counts them
Medical providers' visits are medical, even for opioid use disorder medication (FAQ 13)	Medical lines aren't reclassified
One visit per patient, service category, site and day; per provider type for other professional and patient support services; two different providers at two different sites make two visits (page 19)	Encounters are grouped that way. A group's visit goes to the line that comes first in the table (a physician before an NP before a nurse), then to the earliest encounter. The report counts the encounters folded into another visit
Virtual visits (Column B2): telehealth CPT and HCPCS codes, modifier 95 or 93, place of service 02 (page 60)	Those, plus place of service 10 and OpenEMR's Virtual Encounter class, listed in data/table-5.json. Also, from 0.16.0, a visit in a category Settings marks as virtual. By default those are the two telehealth categories OpenEMR's Comlink telehealth module creates, whose encounters carry the AMB class and the site's own place of service, so nothing else on them says virtual
Patients are counted once in each service category, so Line 34 normally exceeds Table 3A (page 64, FAQ 16)	As stated. Every UDS patient should still have at least one Table 5 visit; the report lists the number who don't

OpenEMR gives every new user the taxonomy code 207Q00000X (Family Medicine). The report flags each provider still on it, and anyone whose code suggests no line, until a line is chosen for them in Settings.

Table 6A: Selected Diagnoses and Services Rendered

Counted from the billing codes on each encounter (the fee sheet's codes). The manual names billing systems and health records as the sources, and its FAQ asks health centers to code a service even when there's no charge, rather than leave it out.

Rule (2026 UDS Manual)	Implementation
Diagnoses (Lines 1–20g) count at countable visits	ICD-10 codes on countable visits (same test as above)
Tests, screenings, dental and support services (Lines 21–42) count when provided at any point in the year to a health center patient	Codes on any encounter in the year at an in-scope site, for patients with at least one countable visit. A flu shot at a nurse-only visit counts; one given to someone with no countable visit doesn't
Column A: one visit per line, however many of its codes match; a visit counts on every line it matches	As stated
Column B: a patient once per line	As stated
Age limits use the age at the visit (Lines 15–17, 26, 26a, 26e)	`DOB`; an unknown birth date can't satisfy an age limit
The codes for each line	data/table-6a.json, copied in the manual's own notation and checked against the manual code by code
Dental services (Lines 27–32) only by licensed dentists, dental hygienists and dental therapists, or staff working under their direct supervision; fluoride also at a medical visit (page 73)	A visit counts on these lines only when its provider is dental staff, placed as Table 5 places them: the line set for the visit category, else the line chosen in Settings, else the one their NUCC taxonomy code suggests. A dental assistant counts as dental staff, though not a Table 5 provider. Fluoride (Line 31) counts at a medical visit too. The patients with a dental service that isn't counted are listed in Records to fix
Immunizations (Lines 24–24b), listed by CPT code	The CPT codes on the fee sheet, and vaccines in the Immunizations screen by their CVX code. Each line's CVX codes are those CDC's CPT-to-CVX crosswalk maps from its CPT codes. A vaccine counts when it was given here: completed or with no status, recorded as a new immunization (or with no source), not refused, and not marked as added in error. A historical record of a vaccine given elsewhere doesn't count. One recorded without an encounter counts at the patient's visit that day
Social needs screening (Line 39)	A coded screening: a code from the line's value set on a visit, a coded questionnaire, or the SDOH history with its screening tool. OpenEMR's older Social Screening Tool encounter form (2022) records no instrument code, so it can't count; from 0.16.0 Records to fix lists the patients screened only with it, once Line 39 is counted
PrEP (Line 21e): Z29.81, only for patients with an active prescription for FTC/TDF, FTC/TAF, cabotegravir or lenacapavir (page 72)	Z29.81 at a visit of a patient prescribed one of those drugs in the year (started in the year, or earlier with an end date in the year or later) or given one from OpenEMR's inventory. Drugs are found by name, generic or brand (Truvada, Descovy, Apretude, Yeztugo). The patients with a Z29.81 visit and no such prescription are listed in Records to fix

Value sets

The manual defines 7 lines only by value sets from the National Library of Medicine's Value Set Authority Center (VSAC): 26c2, 26c3 and 38–42 (tobacco cessation drugs, medications for opioid use disorder, language assistance, and the upstream drivers of health lines). It adds value sets to 10 more lines, 21 value sets in all.

The module doesn't ship their codes. The UMLS licence doesn't allow passing CPT codes on (category 3: internal use only), and its SNOMED CT terms require sub-licences and access controls a public repository can't give. So each site fetches the codes under its own UMLS licence:

an administrator does it in Settings › Table 6A value sets, or from the command line with bin/uds-value-sets.php;
it takes the site's UMLS API key (the account is free; apply at uts.nlm.nih.gov);
the key is used for the fetch and not saved, and the codes are kept in two module tables;
fetch again each year, when HRSA publishes a new manual.

Until then, those lines are left blank, or count only the codes the manual lists.

Once all of a line's value sets are fetched, the line also counts:

Codes on the fee sheet, and in care plans: the value sets' ICD-10-CM, CPT, HCPCS and SNOMED CT codes. Care plan entries count when marked Completed (a service provided, not planned), and not goals or health concerns; this applies to the listed SNOMED CT codes of Lines 35 and 36 too, fetched or not. A code the line's own entry excludes still doesn't count: Line 9's value set includes gestational diabetes (O24.4-), which the manual excludes from Line 9.
Medications (RxNorm), Lines 26c2 and 26c3: prescriptions, by their start date, and medications given from OpenEMR's inventory.
Screenings and tests (LOINC), Lines 21, 26f and 39:
- the screening instrument of an SDOH assessment in History, unless marked declined;
- questionnaires answered, unless their encounter form was deleted;
- Observations, unless deleted;
- lab results, unless cancelled or entered in error.

A record tied to an encounter counts at that visit. One with no encounter counts at the patient's visit that day, if there is one (one at an in-scope site first), so it isn't a second visit; otherwise it counts as a visit of its own on its date.

What it can't count yet (each line is marked on the report, with the reason):

Medications coded by ingredient (Lines 26c2, 26c3). A medication counts only when its RxNorm code is in the value set, which lists products. A prescription coded only by ingredient (OpenEMR's RxNorm lookup offers those), or with no code, isn't counted. For MOUD the manual counts buprenorphine, methadone and naltrexone in any form (FAQ 15). The record problems list names the patients with such a medication, so staff can choose the product instead.
Interventions not marked Completed (Lines 35–42). A care plan entry still Active or Draft isn't counted, since the manual counts services provided; one written only in a note isn't seen.
Codes outside the fee sheet on other lines. For example, a test recorded only in a note or a scanned document isn't seen. The manual's preferred fix is to put the code on the encounter with a zero charge.
Dental codes are recognised by their shape (D plus four digits) under any code type except ICD, since OpenEMR has no CDT code type.

Grant Reports

A health center funded under more than one Health Center Program type files a Grant Report for each special population it is funded for, alongside the Universal Report (pages 11–12 and 15–16):

Grant Report	Patients	Matches the Universal Report's
Migratory and Seasonal Agricultural Workers, 330(g)	Agricultural Worker (UDS) is migratory, seasonal, or an aged or disabled former migratory worker	Table 4, Line 16
Homeless Population, 330(h)	Housing Status (UDS) is any homeless arrangement	Table 4, Line 23
Residents of Public Housing, 330(i)	Seen at a site marked in Settings as in or next to public housing; it is a site-based count, so where the patient lives doesn't matter	Table 4, Line 26

Each covers those patients and all their in-scope visits and services, whatever paid for them (FAQ 2), on Tables 3A, 3B, 4, 5 and 6A. Table 5 has no FTEs, and the Patients by ZIP Code table isn't part of a Grant Report. The membership test is the one Table 4 uses, so a Grant Report's total patients equals the Universal Report's line for that population, and no Grant Report cell can be larger than the Universal Report's, as the manual requires.

Grant evidence

Reports › UDS Report › Grant evidence gives the counts a grant application or report asks for, from the same patients as the UDS Report, for the Universal Report or any Grant Report population, with a CSV download:

Who we serve: patients by age band, sex at birth, insurance (the Patients by ZIP Code table's columns), income as a share of the poverty guideline (Table 4), and ZIP code; homeless patients, migratory or seasonal agricultural workers, veterans, and patients best served in a language other than English.
Social needs screening and services: patients screened for social needs (Table 6A Line 39), and patients given services for food insecurity (40), housing instability (41) and financial insecurity (42), overall and by ZIP, age band and insurance, each as a share of all patients. These are Table 6A's own counts, so they agree with it, and they show only once the site has fetched the value sets for those lines. A screening counts when it is coded, including a questionnaire answered with its LOINC code. Lines 40 to 42 count services given (coded on the fee sheet or in a completed care plan entry), not positive screens or Z59 diagnoses.
Quality measures: links to Tables 6B and 7 and eCQM readiness. The page doesn't recalculate the measures.

The figures are de-identified for use outside the health center:

Counts only: no names, dates or record numbers.
Small cells hidden: a count from 1 to 10 shows as "<11", following CMS's cell-size rule. Within a breakdown, when exactly one count is hidden, the next smallest is hidden too, so the total can't give it away. A share is hidden with its count.
ZIP codes with fewer than 11 patients are folded into "Other ZIP codes".
The limit: suppression is per breakdown. Figures for different populations or years can still be subtracted from each other (the Universal Report's children less the homeless population's, say), so check what you publish together.

Opening it takes the same permission as the UDS Report.

Table 9D worksheet

Table 9D follows the health center's audited financial statements, and part of it comes from the general ledger (FAQ 8, page 169), so the module gives a worksheet for whoever prepares it, not the table. It reads OpenEMR's accounts receivable: fee sheet charges and the payments and adjustments posted to them. It opens from the Universal Report; Grant Reports have no Table 9D.

Rule (2026 UDS Manual)	Implementation
Charges (Column A) by date of service in the year, from the fee schedule (pages 164–165, FAQ 6)	Active fee sheet lines of visits dated in the year at in-scope sites. Sales tax lines are left out, and so are old COPAY lines (payments recorded as negative charges), which are reported
Charges are reclassified: each payer is charged what it accepted, and the rest moves to the next payer and finally to Self-Pay (page 165, FAQ 4)	Each payer is charged what it settled on the visit (the payments and adjustments posted for it, up to the day the report runs). The balance still open goes to whoever owes it now: the insurer at the level after the last one closed, or the patient. The page also shows each line's charges before reclassification
Collections (Column B) and adjustments (Column D) by posting date in the year, for services of any year (FAQ 6)	Postings dated in the year: the date entered with the posting, else its batch's Post To Date, else the day it was saved. Takebacks and refunds, posted as negative payments, reduce collections (page 168)
A payer line follows the payer, including managed care plans run by private insurers (pages 160–161)	The insurer that sent the payment, else the patient's insurer at that level on the date of service, by the UDS insurance categories set in Settings. Patient postings are Self-Pay
Contractual adjustments, sliding fee discounts and hardship waivers are Column D; bad debt is Line 15, not Column D (pages 162–167)	By the reason posted with each adjustment (data/table-9d-adjustments.json). Reasons with no clear treatment, and reasons the file doesn't know, are listed for review and kept out of Column D
Bad debt by the allowance method where possible; direct write-off otherwise (page 163)	Direct write-offs posted in the year. An allowance comes from the general ledger
Pharmacy revenue is Line 17, not charges (pages 164–166)	Product sales from OpenEMR's inventory and their payments are reported apart, because OpenEMR doesn't tell pharmaceuticals from supplies such as eyeglasses, which are charges

Lines 17 and 18, capitation, settlements and wrap payments are left to the general ledger. Insurers not mapped to a UDS category, and insurers mapped to None/Uninsured (a program such as Title X is uninsured on Table 4 but Other Public on Table 9D), get their own rows so the total stays whole. The worksheet also reports payments not yet applied to any visit, and visits paid more than was charged.

eCQM readiness

Most of Tables 6B and 7 are eCQMs, which OpenEMR calculates itself, and the module reports them. Their results are only as good as what the engine can see, so Reports › UDS Report › eCQM readiness shows what to fix first.

It lists the 16 measures UDS 2026 uses (data/ecqm-measures.json), the version the manual names, and whether that version is installed. It also counts, for the year's countable visits and the patients seen at them, the records OpenEMR's eCQM calculation can't see:

Records	Why (measured on OpenEMR master c11dd1b15c with CMS165; the codes are from the 2026 measures' value sets)
Visits with no coded Encounter Type	Visits are found only by the code of their Encounter Type, not by billing codes. OpenEMR's stock Encounter Types have no codes; the page lists those that don't
Problems not classified as Diagnosis, or with no code	Problems are read only when their Classification Type is Diagnosis, which a new problem doesn't have
Blood pressures only in the Vitals form	Blood pressure is read only from Observations of type Physical Exam Performed, a type the stock list has turned off. OpenEMR master reads the Vitals form too since 2026-10-18 (openemr/openemr#14656); there this row and the BMI one go away
BMI only in the Vitals form	BMI (CMS69) and children's BMI percentile, height and weight (CMS155) are read the same way, only from Physical Exam Performed Observations
Tobacco use only in History	Tobacco screening (CMS138) is read only from Observations of type Assessment coded 72166-2 (or 39240-7, 68535-4, 68536-2). The Tobacco field in History isn't read
PHQ-9 forms with no depression screening the measure reads	Depression screening (CMS2) is read only from Assessment Observations coded 73832-8 (adults) or 73831-0 (adolescents), not from the PHQ-9 form. Shown when the PHQ-9 form is installed
A measure's screening or exam code recorded where the measure doesn't read it	The codes above, and falls screening's (CMS139), count only as an Observation of the type named. The page names each kind found only as a lab result or as an Observation of another type
Ongoing problems, and visits with no end time	OpenEMR treats a missing end date as unknown, so these never count (openemr/openemr#14526). The page asks OpenEMR's own builder, so the row goes away once OpenEMR is fixed. Adding end dates to ongoing problems isn't a fix: an end date marks a problem resolved

A second list covers records the calculation still counts that the rest of OpenEMR treats as deleted or wrong. Each row shows only while the installed OpenEMR's eCQM builder reads them, judged from its own query, and says how to fix the records where there's a way:

Records	What the measures do with them (OpenEMR master 44cbd6af98, 2026-10-18)
Problems marked Entered in Error or Refuted	Still read as diagnoses when classified Diagnosis. Clearing the Classification Type keeps them out
Immunizations marked Entered in error	Still read, so Childhood Immunization Status (CMS117) can count a dose that wasn't given
Prescriptions marked not active, with no end date	Read as still taken. Giving a stopped prescription its end date fixes it
Observations, care plan entries and orders in forms deleted from the visit	Still read (openemr/openemr#14587 covers the observations)
Lab results reported as error or cancelled	Read whatever their status, for example an HbA1c in the diabetes measure (CMS122)

Tables 6B and 7

Reports › UDS Report › Tables 6B and 7 shows the eCQM lines of both tables for the year, calculated by OpenEMR's own eCQM engine with the measure versions the manual names, and a CSV download.

The 2026 UDS Manual (pages 84–85)	What the module does
Count patients seen for a countable UDS visit who meet the eCQM's initial population and qualifying encounter	The engine is run for the year's UDS patients (the same countable-visit rule as the other tables)
Column A: the denominator, less exclusions, less exceptions who don't meet the numerator	As stated, per patient
Column B: records reviewed	Equal to Column A: the whole EHR is reviewed
Column C: the numerator; each patient counts once per measure	As stated
Line 12 (CMS155): BMI percentile and counseling for nutrition and for physical activity	A child in all three of the measure's numerators
Line 14a (CMS138): screened and not a user, or a user who received cessation intervention	The measure's third population set
Lines 23a and 23b (CMS137): initiation and engagement, over one denominator	The measure's first and second population sets
Table 7 by race and Hispanic, Latino/a, or Spanish ethnicity	Each patient's Table 3B cell, by Appendix B's crosswalk to the Table 7 line, with the subtotals

Calculating. The engine is run once per year and the results are stored, because it is slow: about a tenth of a second per light synthetic patient, and about a second per patient with realistic records (Synthea patients with years of visits and lab results). 20,000 such patients would take about five and a half hours. An administrator starts it from the page, which works through the patients 50 at a time and can be resumed if closed; for anything but a small health center, run it on the server instead:

HTTP_HOST=localhost php interface/modules/custom_modules/oe-module-uds-report/bin/uds-quality.php --year=2026

Patients the engine refuses. OpenEMR's engine accepts only UCUM units, and one lab result in any other unit (mEq/L, x10E3/uL, IU/L, or the UNK OpenEMR's C-CDA import stores) makes it refuse that patient for every measure. Those patients are left out of every line and named on the page, with the engine's reason, so their records can be corrected and the year calculated again; the CSV gives their number and reasons. The readiness page counts them beforehand, with the UCUM unit for each common one. It also warns when OpenEMR reads lab results as whole numbers, which keeps an HbA1c of 9.4 out of CMS122's poor-control count (it reaches the measure as 9).

What it needs.

OpenEMR's PHP eCQM engine (openemr/openemr#14469), on OpenEMR master after 8.4.
The 2026 measure files, in OpenEMR 8.5.0 (openemr/openemr#14536, merged 2026-10-13; not in 8.4.1). On an OpenEMR without them, the page names each measure that isn't installed.
The readiness page cleared, above all openemr/openemr#14526: until it's fixed, ongoing problems and visits with no end time don't count, and the page says so.

Not calculated. Line 18 (CMS164v7, which the manual keeps at an old version no measure bundle has); the lines with no eCQM (prenatal care, Lines 1–9; HIV linkage to care, Line 20; dental sealants, Lines 22a and 22b); and Table 7 Section A (deliveries and birth weight). The page lists them, to be reported from the health center's records.

The Financial Review Date pitfall

OpenEMR's Financial Review Date is a plain text box on a date column, and OpenEMR saves with the database's strict checking off. A date typed as 2026-09-15 is kept; typed any other way (9/15/2026, Sept 15) it is silently stored as no date at all, with no message on screen. The patient's income then can't be used and counts as Unknown on Table 4. The Records to fix list names this case on its own ("Financial Review Date was not saved"), so staff can re-enter those dates.

Reference data

The published figures and code mappings are in data/, in plain files that people who prepare UDS submissions can check without reading PHP: the poverty guidelines, the race and ethnicity roll-up, the insurance categories, the Table 5 lines and taxonomy mapping, and the Table 6A codes. The counting rules stay in the code. A broken data file stops the report with a message naming the file and entry; it never runs on a bad edit.

The expected results are data too (data/cases/): line and column placements, insurance categories, poverty thresholds, and whole tables built from synthetic patients worked out by hand from the manual. Any program implementing the same rules can be held to them. The tests run every case (tests/ConformanceTest.php). The full report has also been checked end to end against a running OpenEMR with synthetic patients whose every cell was worked out by hand first.

Not done yet, or approximated

Other tables: 8A and 9E and the two forms aren't produced yet. Tables 6B and 7 cover their eCQM lines only (above), and Grant Reports don't include them. Table 9D is a worksheet, for the reasons above.
Table 5 FTEs (Column A) aren't computed; they come from HR records. A person who holds two roles (a nurse who is also a case manager, say) has one line per person, plus any line set by visit category.
Countable visits: whether a visit is truly countable (independent judgment, a credentialed provider, not a nurse-only or ancillary service) is approximated by visit category and provider. Each health center has to set the categories.
Homelessness and farmworker look-back: the manual counts someone as homeless for 12 months after homelessness was last documented, and as a farmworker if that was their work within 24 months. The module uses the current value of each field; it has no history of past values.
Medicaid versus CHIP: the two can't be told apart automatically. They're reported as Medicaid unless a plan is set to CHIP in Settings, which is what the manual says to do when they can't be distinguished.
Validation: tested only with synthetic data. It has not been compared with a health center's real submission.

Install

From the OpenEMR directory:

composer require anun333/oe-module-uds-report

That places it in interface/modules/custom_modules/oe-module-uds-report/. Then go to Modules › Manage Modules and Register, Install and Enable UDS Report. Installing adds the settings tables and the three demographic fields. Uninstalling leaves them in place, because they hold data staff entered.

Viewing needs patients/demo; the Table 9D worksheet needs acct/rep_a, as OpenEMR's Collections report does; Settings and the Table 6A value sets need admin/super. Fetching the value sets needs the site's UMLS API key, and HTTPS from the server to cts.nlm.nih.gov.

Compatibility

PHP security. PHP's php2shell advisory (2026-10-21; CVE-2026-109618 and CVE-2026-109619) is remote code execution without logging in, and it is being exploited. Run OpenEMR on PHP 8.2.35, 8.3.36, 8.4.27, 8.5.12 or later, whatever the module needs. The official OpenEMR 7.0.2 and 7.0.3 Docker images can't get that fix: they carry PHP 8.2.16 and 8.3.15 on Alpine releases that no longer get updates. Run 7.0 only on a patched PHP, or upgrade OpenEMR. OpenEMR rebuilt its 8.4.1, 8.3.0, 8.2.0 and 8.0.0.3 images with the fix on 2026-10-22 (forum t/27157); a container started before then needs pulling again.

Older OpenEMR: on 7.0.3 and older (no OEGlobalsBag), the module stays switched off rather than stop OpenEMR's pages; checked on fresh 7.0.2 and 7.0.3 installs with every module enabled. On fresh official 8.1.0, 8.3.0 and 8.5.0 images (2026-10-23) it installs through Module Manager and its pages that need no patient load, with every module enabled and with it alone (dev/module_compat_check.sh). 7.0.4 to 8.0 haven't been tried.

OpenEMR	Tables, Grant Reports, 9D worksheet, eCQM readiness	Tables 6B and 7	How it was checked
8.5.0 and master (8.6.0-dev)	works	works: the 2026 measures are in 8.5.0 (#14536, merged to master 2026-10-13; 8.5.0 was cut from master on 2026-10-25). Ongoing problems and visits with no end time don't count until #14526 is fixed	end to end, synthetic patients; rechecked on master 76fa2f1e88 (2026-10-13): CMS165v14 for nine synthetic patients matched each one's prediction
8.4.1	works	no PHP eCQM engine; the page says so	end to end, synthetic patients
8.4.0, 8.3.0, 8.2.0, 8.1.0	installs, every page loads	no PHP eCQM engine	official Docker images: installed through Module Manager, every page loaded as an administrator with no PHP error (no patients)
8.0.0 and older	stays switched off (before this release, enabling it on 7.0.4 to 8.0.x stopped every page): OpenEMR classes it uses don't exist there (`AccessDeniedHelper`, `BC\ServiceContainer`, and others)		which classes exist at each release tag

PHP 8.3 or later is needed. On 8.1.0 to 8.3.0 the module reads the request itself, because OpenEMR's CurrentRequest only arrived in 8.4.0.

The Grant evidence page (0.15.0) passes dev/uds_grants_check.py in the openhealthcare repository on 8.4.1 and on master 1faefbbeb9 (2026-10-18). That covers the total matching the UDS Report, no count from 1 to 10 in the CSV, and the social-needs lines from synthetic value sets.

The Table 6A value sets were checked on master (c9b422d608):

Upgrade: Module Manager's upgrade path (SQLUpgradeService) from 0.8.0 created the two tables.
Fetch: the page and the command line each fetched all 21 value sets from VSAC (1,548 codes). A wrong key stopped at the first one and left the stored sets as they were, and a form without a valid CSRF token was refused.
Counting: synthetic records of every kind the module reads were counted on the lines they belong to, at their visit or on their own date. Records left out on purpose weren't counted: a declined screening, an Observation on a deleted form, an ingredient-only prescription, gestational diabetes on Line 9, and a patient with no countable visit. With the value sets removed, the same records counted only Line 9's listed code.

The 8.4.1 checks: Tables 5 and 6A from codes saved through OpenEMR's own billing functions, a Table 5 line chosen through the Settings page, and the Table 9D worksheet from payments and adjustments posted through OpenEMR's EOB posting screen. The eCQM readiness page was also checked against OpenEMR master (c11dd1b15c), where eCQMs are calculated in PHP, with and without a local fix for #14526. Tables 6B and 7 were checked on master (7001d5008b) with #14536's 2026 measures: the module's results for each patient are identical to OpenEMR's own openemr:cqm-calculate, and for 2,009 synthetic patients the CMS165 denominator (1,100) and numerator (443) match what each patient's records predict, by race and ethnicity too, calculated both from the page and from the command line.

The package declares no Composer constraint on openemr/openemr: OpenEMR reports itself to Composer as version 1.0.0+no-version-set, so any minimum version written there would block every install.

Development

composer install
vendor/bin/phpunit                 # the counting rules and data/cases/; no database needed
php tools/footprint.php --check    # FOOTPRINT.md must match the source

Code systems

data/table-6a.json lists codes and code ranges as the 2026 UDS Manual prints them for Table 6A, without their descriptions. The module's own licence covers its code, not these code systems, which belong to their publishers:

SNOMED CT® is a registered trademark of SNOMED International.
ICD-10-CM (National Center for Health Statistics) and HCPCS Level II (CMS) are published by the U.S. government.
CVX codes (vaccines administered) and the CPT-to-CVX crosswalk the module's CVX lists come from are published by the CDC.

The value sets that Table 6A names are not in the module: each site fetches them under its own UMLS licence (Value sets). The value-set page shows the notices that licence asks for, for the code systems fetched.

Other modules for OpenEMR

From the same project. Each installs with composer require anun333/<module> and is then enabled in Modules › Manage Modules.

Module	What it does
Clinic Forms	Printable clinic paperwork: consents, privacy notices, referral letters and validated screening questionnaires, with the clinic's own details, in English and Spanish
Staff Worklist	One queue of outstanding staff work: unsigned results, orders with no result, referrals with no reply, screening and claims follow-up
Sliding Fee Discount	HRSA's sliding fee discount program: the schedule checked against Chapter 9, income assessments, and the discount applied
Care Management	CCM, PCM and BHI enrollment with CMS's consent elements, monthly time logging, and the codes each month meets
Migration	Records from another EHR, read from the FHIR export certified EHRs offer, checked before import and safe to rerun
Layout Export	Layouts and lists exported to files you can keep in version control, compared with the site or another file; reads the old Backup page's SQL files without running them

License

GPL-3.0-or-later.

anun333 / oe-module-uds-report

Package info

Security

Statistics