How can a home-care agency tell whether a referral record is trustworthy before it reaches intake, scheduling, or a manager's queue? This study treats referral data provenance as an operations question: can a reviewer trace each important field to a source, a received time, and an accountable next action? The objective is not to assign a universal quality score. It is to show a repeatable way to separate a complete record from a record that merely looks complete.
Research question and evidence scope
The question is narrow: which evidence fields most often distinguish a usable home-care referral from an item that needs clarification? The study uses public guidance from the Centers for Medicare and Medicaid Services, the Agency for Healthcare Research and Quality, and the Department of Health and Human Services. CMS describes quality information for home health providers. AHRQ frames home care as a setting with patient-safety risks that depend on the home environment and handoffs. HHS explains why access and disclosure should be limited to the minimum necessary for a task.
These sources establish context, not a referral benchmark. No source reviewed here publishes a national pass rate for referral records. The findings below are therefore a measurement design and an interpretation framework, not a claim about the performance of any agency.
Methodology
Select one referral stream and one review period. Do not combine hospital referrals, family inquiries, payer requests, and physician office referrals in the first sample because each stream may carry different source fields and urgency rules. Pull the full population if it is small. Otherwise use a documented sample and record the denominator before reviewing individual records.
For each record, preserve the original arrival point, the received date and time, the referral source, the requested service, the stated contact preference, and the current disposition. Add a link or identifier for the approved system of record. The reviewer should not copy an entire clinical narrative merely to prove that a referral arrived. HHS guidance supports limiting access and disclosure to information needed for the purpose.
Score five evidence dimensions separately:
| Dimension | Evidence question | Interpretation |
|---|---|---|
| Source identity | Can the sender or origin be verified? | Unknown source requires clarification |
| Arrival time | Is the received date and time preserved? | Date-only records limit response analysis |
| Request | Is the requested service stated in the source? | Do not infer a service from a diagnosis |
| Ownership | Is the next responsible role named? | A queue without an owner is not a handoff |
| Disposition | Is the current status supported by a note or source? | Status labels need evidence |
Calculate a separate proportion for each dimension. A single composite result hides the difference between weak source capture and weak follow-up. Reviewers should record reasons for missing evidence, such as an unreadable fax, an incomplete portal submission, a duplicate entry, or a request that falls outside the agency's service scope.
What the public evidence supports
The public sources support three cautious conclusions. First, home-care operations depend on information moving across settings, so the receiving agency needs a clear record of what was received and what remains unknown. Second, safety and quality work is not only a clinical concern. An administrative handoff can create risk when the wrong person, task, or timing is carried forward. Third, privacy controls should shape the sample itself. A reviewer needs enough information to test provenance, not every detail available in the originating record.
Those are evidence-led principles. They do not prove that a particular field will improve an agency's conversion rate, staffing level, or visit outcome. They identify what can be observed and reviewed.
Implications for a home-care staffing operation
A back-office assistant can maintain the provenance queue when the agency defines the permitted fields, source systems, and escalation rules. The assistant may label an item as waiting for source, identify a duplicate, send an approved request, and prepare a daily aging view. It should not decide clinical eligibility, interpret a payer rule, promise service availability, or select a caregiver for a client.
The useful handoff contains the original source, the unresolved question, the last action, and the next review date. If a source document conflicts with a phone note, preserve both and route the conflict to the authorized owner. Replacing one value with the other would make the record look cleaner while reducing its evidentiary value.
The route-local review should retain the source field exactly as received before normalizing names or categories. Keep a change note when a referral is merged, corrected, or reassigned, and record the reviewer and date in the approved system. A second pass on a small sample can distinguish an ambiguous source from a transcription error. Report missing provenance separately from missing follow-up so that an agency does not mistake a busy referral queue for evidence that its source data is reliable.
Limitations
This method does not compare agencies, predict referral acceptance, or measure quality of care. It is sensitive to the design of the source systems and to the definition of a referral. A record can pass an administrative provenance check and still be clinically incomplete. Conversely, a record may be safe for a narrow administrative purpose while lacking information needed by a clinician. Any local target should be set after a baseline review and discussed with the agency's privacy, compliance, and clinical owners.
Evidence-led conclusion
Referral data is more trustworthy when a reviewer can follow its chain from source to receipt, request, owner, and disposition without guessing. The strongest first study is not a broad scorecard. It is a small, source-specific sample with separate evidence dimensions, explicit limitations, and preserved exceptions. For home-care agencies, that design gives managers a defensible view of where intake work is incomplete while keeping clinical and legal decisions with the people authorized to make them.

