Skip to main content
ClinicalSim
Back to Insights
6 min read

HCAHPS can't see a single conversation. Here's what can.

Across 3,909 US hospitals, HCAHPS doctor communication rose 0.8 points through 2019, while the overall hospital rating rose 8.5 (Beckett, 2024).

ClinicalSim Team

ClinicalSim

Hospitals that report the HCAHPS survey publicly see a doctor communication score on Care Compare, and physician leaders get asked to raise it. Nationally, that score rose 0.8 points in more than a decade, while the overall hospital rating rose 8.5 [1]. The physicians aren't the problem here. The score answers a different question from the one a training program needs answered.

What does the HCAHPS doctor communication score measure?

HCAHPS doctor communication is a composite of three survey questions asking how often doctors treated the patient with courtesy and respect, listened carefully, and explained things in a way the patient could understand [2]. Courtesy and a clear explanation sit close to the acknowledge and explain steps of AIDET. The questions ask about "doctors" as a group, and the patient answers after going home. In the July 2026 public report, the national top-box score (the share of patients answering "always") was 80% across 3,949 hospitals, level with the nurse communication score [3]. A separate post covers what the overall rating tracks most closely.

The questions are fair ones. The trouble is that one answer covers the hospitalist, the overnight resident, the consultants and the surgeon together, so a patient who saw five physicians gives one rating to all of them.

Why the score barely moves

Across 3,909 US hospitals, the HCAHPS doctor communication top-box score rose 0.8 percentage points from the start of public reporting to 2019, the smallest change of any HCAHPS measure. Over the same period, the overall hospital rating rose 8.5 points and nurse communication rose 6.5 [1]. Doctor communication was publicly reported, counted in Medicare's value-based purchasing program and fed the star ratings like the other measures, and it still barely moved.

Attribution is the best-tested explanation. At Cleveland Clinic, researchers observed 28 hospitalists on rounds and coded their bedside communication with the Four Habits Coding Scheme, a published observation instrument. The correlation between those observed skills and the same hospitalists' HCAHPS doctor communication scores was 0.098, which was not significant, across 1,003 survey responses [4]. When the analysis kept only patients cared for by a single hospitalist, the empathy habit correlated 0.442 with the doctor communication score [4]. The skill was there to be measured, but once the patient's experience was spread across several physicians, the survey could no longer see it.

What happened when hospitals trained physicians

Several hospitals have trained physicians in communication and then looked at HCAHPS. At Baystate Medical Center, 128 physicians took part in simulation-based communication coaching. HCAHPS doctor items showed no significant difference against surgical controls, but a separate survey that named the individual physician rose significantly, including an adjusted 14.8% shift on "using words the patient could understand" [5]. At Cleveland Clinic, physicians who took an eight-hour communication course scored higher on an outpatient physician survey, but the HCAHPS doctor communication composite didn't differ significantly (P = .22) [6].

The study that did move HCAHPS gave each physician feedback on their own patients. At UCLA, internal medicine residents attended a one-hour conference and then received real-time satisfaction feedback from their own patients, monthly recognition and small rewards. Compared with other departments, their patients' HCAHPS doctor communication score rose 6.6 points more and "definitely recommend" rose 5.6 points more [7]. The authors judged that the workshop, which about half the residents missed, had minimal impact, and they pointed to individualized feedback instead [7]. The study was not randomized and covered six months at one hospital.

Across these studies, the change showed up where measurement reached the individual clinician: a survey that named the physician, or feedback from each resident's own patients. None of this shows that communication practice doesn't matter. It shows that a hospital-wide score, with nothing feeding back to the clinician, is a poor way to tell whether practice changed how anyone talks with patients.

What measuring the conversation looks like

Measuring the conversation means scoring what a clinician said, against a named framework, close enough to the moment that the clinician can use the feedback. One encounter from our example library shows what that produces.

In "48 Hours in the Hallway," the learner plays an emergency department nurse starting a shift. The AI patient is Tiffany, a daughter who has spent the night beside her 78-year-old father, who has been boarding in a hallway bed for 48 hours while he waits for a procedure and an inpatient bed. Partway through the conversation, which ran 4 minutes 42 seconds, she says: "I asked for the doctor in charge overnight, and a young doctor came, and then another nurse, and they both said the same thing. I just need somebody to actually know." That sentence is the attribution problem in a family member's words. A patient survey after discharge would fold the doctor and both nurses into answers about "doctors" and "nurses" in general.

The conversation was scored against two frameworks. The first, NURSE (Naming, Understanding, Respecting, Supporting, Exploring), describes empathic responses and isn't scored. The learner used four of the five. The clearest Understanding statement was "I'm sorry you've been told 'soon' so much and it's been nothing but later and later." Exploring didn't appear, and the feedback names the moment to try it: when Tiffany said her father "looks small in that stretcher" and that she didn't know whether to go to work.

The second framework, Ask-Tell-Ask with teach-back, is scored from 1 (poor) to 5 (excellent) on four dimensions. The learner scored 4 on responsiveness for committing to a concrete time and being candid about not having seen the labs yet, 3 on plain language, and 2 on both asking what Tiffany already understood and confirming the plan.

Each score comes with the learner's own words and suggested wording for next time. For the teach-back, the suggestion is: "Just so I know I explained this clearly, can you tell me in your own words what's going to happen over the next twenty minutes?" The learner gets what went well, one or two things to practice next, and the reason for each.

The example comes from nursing practice, but the method is the same for physicians. The person whose conversation is scored is the person who gets the feedback, and the next practice session shows whether the habit took hold.

How to connect practice data to HCAHPS honestly

Conversation-level measures and HCAHPS answer different questions, and a program needs both. The conversation score tells a clinician and an educator whether a specific habit is changing. HCAHPS tells leadership how patients experienced the whole stay, across every clinician they met.

A program that wants to look for a link between the two can set it up before practice starts:

  1. Name the HCAHPS measures in advance, such as doctor communication or the overall rating, and the units or services where participating clinicians work.
  2. Pick a comparison group, such as a similar unit or service that hasn't started practice yet, so a national upward trend isn't mistaken for a program effect.
  3. Agree on a time window long enough to collect a meaningful number of surveys, and on what result would count as no change.
  4. Keep the conversation scores formative, so that clinicians practice without worrying that a hard first attempt follows them into an evaluation.

ClinicalSim doesn't predict HCAHPS scores, and no current evidence shows that practice scores change them. What a program can show directly is whether clinicians are using the skills they practiced. For the wider case, see how communication connects to hospital economics and how to build a training ROI case without inventing a number.

References

  1. Beckett MK, Quigley DD, Cohea CW, et al. Trends in HCAHPS survey scores, 2008-2019: a quality improvement perspective. Med Care. 2024;62(6):416-422. 2024. doi:10.1097/MLR.0000000000002001
  2. Centers for Medicare & Medicaid Services. Crosswalk of updated HCAHPS survey questions to sub-measures. HCAHPS Online. 2025.
  3. Centers for Medicare & Medicaid Services. Summary of HCAHPS survey results: October 2024 to September 2025 discharges. HCAHPS Online. July 2026 public report. 2026.
  4. Velez VJ, Kaw R, Hu B, et al. Do HCAHPS doctor communication scores reflect the communication skills of the attending on record? A cautionary tale from a tertiary-care medical service. J Hosp Med. 2017;12(6):421-427. 2017. doi:10.12788/jhm.2743
  5. Seiler A, Knee A, Shaaban R, et al. Physician communication coaching effects on patient experience. PLoS One. 2017;12(7):e0180294. 2017. doi:10.1371/journal.pone.0180294
  6. Boissy A, Windover AK, Bokar D, et al. Communication skills training for physicians improves patient satisfaction. J Gen Intern Med. 2016;31(7):755-761. 2016. doi:10.1007/s11606-016-3597-2
  7. Banka G, Edgington S, Kyulo N, et al. Improving patient satisfaction through physician education, feedback, and incentives. J Hosp Med. 2015;10(8):497-502. 2015. doi:10.1002/jhm.2373