Article

Does your EHR need a clinical decision support layer, or is it enough on its own?

Published

Content pillar P7 Grow Your Practice

Most cash-pay clinics that see complex patients need both, because they do different jobs. The EHR is your legal record and your ordering, prescribing, and billing system. A clinical decision support layer reads across the EHR, outside records, labs, and wearable data, then drafts reasoning for you to review. It sits beside your chart, so you keep your EHR.

On October 26, 2026, HHS finalized the HTI-5 rule. Among other changes, it removes the requirement that certified EHRs let users see "source attributes" for the decision support tools built into them, Health System CIO reported. Those attributes, added in late 2023, covered who built a tool, what data it used, and how it was tested.

For a small clinic, the point is practical. Your EHR's certification will no longer tell you how the decision support inside it reasons. You have to ask, whether the tool is part of the EHR or sits beside it.

This page separates the two jobs, explains how a decision layer connects to an EHR, sets out what FDA and HHS rules say about decision support software, and gives you the questions to ask before you buy.

What is the difference between an EHR and a clinical intelligence layer?

The EHR records what happened. A clinical intelligence layer helps you decide what should happen next. You need the first in any practice. You need the second when the patient's data is spread across more sources than you can review by hand before a visit.

What is the EHR for?

The EHR is where the legal medical record lives: notes, orders, prescriptions, results, consent, and billing. Nearly every practice has one. In 2024, 95.0% of office-based physicians used an EHR and 83.6% used a certified EHR, according to the CDC's National Center for Health Statistics.

EHRs are built to store and retrieve. They are good at showing you what is in your own chart. They are less good at combining your chart with records, labs, and device data that came from somewhere else.

What is a clinical intelligence layer?

A clinical intelligence layer is decision support software. It pulls together the patient's information from several sources, organizes it, and prepares a recommendation or summary that a licensed clinician reviews.

FDA's description of decision support that is not a medical device fits this job. In its Clinical Decision Support Software guidance, FDA describes software that displays or analyzes medical information about a patient, supports recommendations to a health care professional, and lets that professional review the basis for those recommendations.

"Clinical intelligence layer" is a product category name, not a regulatory term. What matters is what the tool does and whether you can check its reasoning.

Why can't the EHR answer the clinical question on its own?

Because much of the data it needs is not in a form the EHR can use. Three published findings show why:

  • Outside records often arrive but don't get used. In 2021, 38% of office-based physicians integrated patient health information from outside sources into their EHR, and 16% engaged in all four parts of interoperability (sending, receiving, finding, and integrating), according to ONC's quick stats.
  • The EHR already produces more than clinicians can process. In a survey of VA primary care practitioners published in 2013, Singh and colleagues reported a median of 63 notifications a day. Of those surveyed, 69.6% said they received more than they could manage, and 29.8% said they had missed abnormal results that delayed care.
  • Chart review takes real time. In a study of about 100 million visits, Overhage and McCallie found physicians spent 16 minutes and 14 seconds in the EHR per encounter, a third of it on chart review.

None of this means the EHR is failing. It means the EHR was designed to keep the record, and finding a pattern across records is a separate task.

Why is this harder in cash-pay care?

Cash-pay longevity, hormone, and peptide patients usually bring more outside data than a typical primary care patient. In our experience at Scriptura, a single new patient may arrive with notes from a PCP on one EHR, a consumer lab panel as a PDF, and months of wearable data in an app.

The AMA's August 2026 report on patients' wearable data found a gap between the clinical value physicians see in that data and how much of it reaches practice. Dr. Steven Steinhubl called for "a set way to provide a summary of that data in the patient's health record."

A PDF of outside labs can be filed in the EHR, but it usually can't be trended there. That trend is the question a decision layer is built to answer.

Do you have to replace your EHR to get clinical decision support?

No. A decision layer is meant to run beside the EHR, and the EHR stays the record. There are two common ways to connect them, and a third option when neither is available.

How do SMART on FHIR and CDS Hooks connect a tool to an EHR?

They are open HL7 standards that let outside software work with an EHR:

  • SMART App Launch "authorizes a user-facing client application ("App") to connect to a FHIR Server." An app can be launched from a user's session inside the EHR, so it opens with the right patient already selected.
  • CDS Hooks is a pattern "for invoking decision support from within a clinician's workflow." At set moments, such as opening a chart or signing an order, the EHR calls an outside service. The service sends back "cards" with information, suggestions, or a link to launch an app.

Certified EHR developers must publish the FHIR service base URLs that patients can use to access their electronic health information, under ONC's API conditions of certification. Whether your EHR lets a third-party decision tool connect for clinicians is a separate question. Ask your EHR vendor which connections it supports and what they cost.

What if your EHR has no open connection?

Then the two systems run side by side. Records come in through patient uploads, fax or PDF, lab portals, and device connections. The clinician reviews the output in the decision tool and documents the final plan in the EHR.

Patients can help close the gap. Under HIPAA, individuals have a right to access the protected health information in providers' designated record sets, including clinical notes and lab reports, as HHS explains. A patient can request records from each prior practice and bring them to you.

Scriptura's suggestion: whichever setup you use, the signed plan belongs in the EHR. The decision tool supports your reasoning. It is not a second chart.

What does FDA consider non-device clinical decision support software?

Software that meets all four criteria in section 520(o)(1)(E) of the FD&C Act. FDA's current guidance was issued in January 2026. As Covington summarizes it, the software must:

  • Not acquire, process, or analyze a medical image or a signal from an in vitro diagnostic device or a signal acquisition system.
  • Display, analyze, or print medical information about a patient, or other medical information such as studies and guidelines.
  • Support or provide recommendations to a health care professional about prevention, diagnosis, or treatment.
  • Let the professional independently review the basis for those recommendations, so they don't rely primarily on the software.

The January 2026 revision added an enforcement discretion policy for software whose output is a single recommendation when only one option is clinically appropriate. It also moved the discussion of time-critical decisions under the fourth criterion. FDA's position is that software used for time-critical decisions may not let the clinician independently review the basis.

The fourth criterion is the one to test when you evaluate a vendor. If you can't see why a tool recommended something, you can't review it, and you can't defend relying on it.

What changed for decision support inside certified EHRs?

Federal transparency requirements for decision support built into certified EHRs have narrowed. Here is the sequence:

  • HTI-1, finalized in December 2023, replaced the old clinical decision support criterion with a "decision support interventions" criterion. ONC's HTI-1 fact sheet lists 13 source attributes for evidence-based tools and 31 for predictive ones, such as intended use, training data, validation, and performance measures.
  • HTI-5, finalized October 26, 2026, removes the old criterion at § 170.315(a)(9) and revises the decision support interventions criterion at § 170.315(b)(11), according to ONC's removals and revisions chart. Health System CIO reports that the revision removes the source-attribute requirements HTI-1 added.
  • In total, HTI-5 removes 14 certification criteria and revises seven, the American Hospital Association reported. Some changes apply from the rule's effective date, and others on January 1, 2027.

Reed Smith noted that ONC finalized this change although a "large majority of commenters" opposed it. Without the requirement, information about how an embedded tool was built and tested would have to come through contracts.

For a clinic, the result is the same whether the tool is inside your EHR or beside it. Transparency now depends on what the vendor will show you.

When do you need both, and when is the EHR enough?

It depends on how much outside data you have to reason across before each visit. This table is Scriptura's suggestion, not a standard:

Your situationEHR aloneAdd a decision layer
Most data originates in your own chartUsually enoughOptional
Patients bring outside labs from several labsStores the PDFsTrends and compares them
Patients share wearable dataRarely structuredSummarizes the trend for review
Patients see several clinicians on different EHRsHolds what is sent to youPuts the history in one view
You run hormone, peptide, or metabolic protocols that need monitoringRecords the planChecks results against the protocol
You need a dated regulatory check in the chartManualCan bring the source into the plan

If most of your rows land in the right-hand column, the time you spend gathering and reconciling data before visits is the cost to weigh against the tool's price.

What should you ask a decision support vendor?

Ask in writing, and keep the answers. Scriptura suggests these questions:

  • Can I see the source behind every recommendation, down to the specific lab, note, or guideline?
  • How does the tool show missing or conflicting data?
  • Which EHRs do you connect to today, and by what method (SMART on FHIR, CDS Hooks, or another interface)?
  • If you don't connect to my EHR, how do records get in, and how does the final plan get into my chart?
  • Will you sign a HIPAA business associate agreement?
  • How do you meet FDA's four non-device criteria, especially independent review of the basis?
  • Does the tool ever act without a clinician's review and signature?
  • How do you update clinical and regulatory content, and how is each update dated?

A vendor that answers "trust the output" to the first question fails the fourth FDA criterion.

What is already on record, and what should you watch?

On record as of October 30, 2026:

  • FDA's current Clinical Decision Support Software guidance dates from January 2026.
  • HTI-5 was finalized on October 26, 2026. It removes the § 170.315(a)(9) criterion and revises § 170.315(b)(11) to drop the HTI-1 source-attribute requirements.
  • In 2024, 95.0% of office-based physicians used an EHR (CDC).
  • In 2021, 38% of office-based physicians integrated outside health information into their EHR (ONC).
  • SMART App Launch (version 2.2.0) and CDS Hooks (version 2.0.1) are the main open standards for connecting outside tools to EHRs.

Watch:

  • The HTI-5 effective dates and any follow-up guidance from ONC.
  • ONC's statement that it may restore some source-attribute requirements in future rulemaking, as Health System CIO reported.
  • Information blocking changes. Reed Smith reports that ONC did not finalize the proposed changes in HTI-5 and may address them later.
  • Any FDA update to its decision support guidance or its examples.
  • Your EHR vendor's list of supported SMART apps and CDS Hooks services.

Frequently asked questions

Do I have to replace my EHR to get clinical decision support?

No. Decision support can run beside the EHR or connect to it through SMART on FHIR or CDS Hooks. The EHR stays the legal record.

What is the difference between an EHR and a clinical intelligence platform?

The EHR stores the record and runs ordering, prescribing, and billing. A clinical intelligence platform reasons across the record and outside data and drafts recommendations for you to review.

Is a clinical intelligence layer a second chart?

It shouldn't be. Document the final, signed plan in your EHR.

What is SMART on FHIR?

An HL7 standard that lets an outside app connect to an EHR's FHIR server, often launched from inside the EHR with the patient already selected.

What are CDS Hooks?

An HL7 standard that lets the EHR call an outside decision support service at set moments, such as opening a chart, and show the service's response as cards.

What does FDA consider non-device clinical decision support software?

Software that meets all four statutory criteria, including that the clinician can independently review the basis for each recommendation.

Did HTI-5 remove decision support rules for EHRs?

It removed the old clinical decision support criterion and dropped the source-attribute transparency requirements from the decision support interventions criterion. The decision support interventions criterion itself remains, in revised form.

How do I get one picture of a patient who sees clinicians on three different EHRs?

Request outside records, or have the patient request them under HIPAA's right of access, then use a tool that combines them into one view. The EHR alone usually stores them as separate documents.

Is a decision support tool worth the cost for a two-provider cash-pay practice?

Compare its price with the clinician time you spend gathering and reconciling outside data before each visit. If that time is small, the EHR may be enough.

What should I ask a decision support vendor first?

Whether you can see the source behind every recommendation. If you can't, you can't review it.

About the author

Sabrina Runbeck

Co-Founder and Chief Strategy Officer, Scriptura Health

Sabrina Runbeck is an active clinician licensed in eight states. She has 16 years of clinical practice in cardiothoracic surgery, emergency medicine, and hospitalist medicine, and a public health background. Her work covers healthcare compliance, governance, and regulatory risk. She translates FDA and state board guidance into defensible clinical protocols for longevity and functional medicine practitioners working across states with different rules.

Last reviewed: October 30, 2026

How can Scriptura help clinics that already have an EHR?

If your patients' records, labs, and device data are spread across more places than your EHR can bring together, book a demo to see how Scriptura's clinical decision support prepares a plan with each source shown, for you to review and sign. Clinicians verify their license before platform access.

Book a Scriptura demo.

Educational and regulatory information only

This article is educational content only. It is not medical or legal advice and does not create an attorney-client or clinician-patient relationship. Privacy rules depend on your clinic's status and state. Confirm your obligations with a healthcare attorney licensed in your state before acting.

Other articles you might be interested in

All articles
  1. New Research, New Evidence

    Does more patient data improve care, or add to information overload in healthcare?

  2. Grow Your Practice

    Clinical decision support systems vs AI chatbots

Does Your EHR Need a Clinical Decision Support Layer?