AHRQ Surveys on Patient Safety Culture (SOPS)

The Agency for Healthcare Research and Quality’s Surveys on Patient Safety Culture® (SOPS®) program has run since 2001, and its survey instruments are now the most widely used tool in the United States for measuring how staff actually perceive patient safety in their own unit — not what a policy manual says should happen, but what respondents report happening. For an infection preventionist, patient-safety officer, or quality director deciding whether and how to run one, the practical questions are narrower than the program’s history: which survey version applies to which setting, what a composite measure actually is, how administration mechanics interact with CMS reporting requirements, and what benchmarking against AHRQ’s database can and cannot tell you.

Choosing the right SOPS survey for your setting

AHRQ maintains separate survey instruments for different care settings, and each has its own version history. Using the wrong one, or an outdated version, undermines both internal trending and any benchmarking comparison.

Hospital Survey: Version 1.0 (2004) vs. Version 2.0 (2019)

AHRQ released the original SOPS Hospital Survey in 2004. In 2019 it released Hospital Survey 2.0, which AHRQ now encourages over 1.0, though the original version remains available. The two are not directly comparable item-for-item — a published transition guide documents what changed — and as of this writing the AHRQ SOPS Hospital Database only accepts submissions built on Version 2.0, which matters if benchmarking is part of the reason you’re running the survey at all.

Hospital 2.0 administration also intersects with a federal reporting requirement. The CMS Patient Safety Structural Measure (PSSM), Domain 3, Statement A, requires hospitals to conduct a hospital-wide culture-of-safety survey either annually or every two years. Hospitals on the two-year cycle must administer a shorter “pulse” survey on target units during the off year. AHRQ’s guidance for building a pulse version is specific: pull a subset of composite measures and single-item measures from the full instrument, but for any composite measure you select, include every item within it — partial composites aren’t valid. Composites and items should stay in the order they appear in the core survey, and AHRQ notes that a pulse survey built this way remains comparable to prior full-survey administrations and to the SOPS Hospital Database, whereas administering the full survey to only selected units in the off year is offered as an alternative approach.

Nursing Home Survey: Version 1.0 (2008) vs. Version 2.0 (2025)

AHRQ released the original Nursing Home Survey in 2008 and a substantially revised Version 2.0 in 2025. The revision was not cosmetic: it removed every negatively worded item, cut the instrument from 44 items to 25, and reduced the composite structure from 12 composite measures down to 8 while adding one new composite, “Speaking Up.” AHRQ publishes a crosswalk document mapping the two versions for organizations that need to reconcile historical data against the new instrument.

Setting fit matters here specifically: AHRQ designs this survey for facilities with long-term care beds, short-stay beds, or that offer rehabilitation services — nursing homes and skilled nursing facilities. It is explicitly not designed for independent living, assisted living, or community care facilities, which sit outside the population the item wording and composite structure were validated against.

Community Pharmacy Survey

Unlike the hospital and nursing home instruments, the Community Pharmacy Survey has a single current version, developed specifically for community pharmacy staff and management in response to requests for a setting-appropriate instrument. It is not a trimmed-down hospital survey; the item set and composites reflect community pharmacy workflow directly.

AHRQ also maintains Medical Office and Ambulatory Surgery Center versions of SOPS. If your organization spans more than one of these settings — a health system with an owned pharmacy or ambulatory surgery center, for example — each site type needs its own matching survey rather than the hospital instrument stretched to cover it.

The composite-measure structure

A composite measure, in SOPS terminology, is a grouping of two or more survey items that assess the same underlying area of safety culture; scores are typically reported at the composite level rather than the individual item level, both because it’s more stable statistically and because it’s more actionable for a unit that needs to know which area to work on.

The Hospital Survey 2.0 assesses ten composite measures across 32 survey items:

• Teamwork (3 items)
• Staffing and Work Pace (4 items)
• Organizational Learning – Continuous Improvement (3 items)
• Response to Error (4 items)
• Supervisor, Manager, or Clinical Leader Support for Patient Safety (3 items)
• Communication About Error (3 items)
• Communication Openness (4 items)
• Reporting Patient Safety Events (2 items)
• Hospital Management Support for Patient Safety (3 items)
• Handoffs and Information Exchange (3 items)

Alongside the composites, the Hospital Survey includes single-item measures that are reported individually rather than combined: Number of Events Reported, Patient Safety Rating, plus six background/demographic items used to break results out by respondent role, unit, and tenure rather than to score culture directly.

The Nursing Home Survey 2.0’s eight composites follow the same logic — grouped items measuring one construct each, including the new “Speaking Up” composite added in the 2025 revision — though the exact item groupings differ from the hospital instrument because nursing home workflow and staffing structure differ from a hospital unit’s. Don’t assume composite names or item counts carry across surveys; check the setting-specific items-and-composites document before scoring.

Administration mechanics that affect data quality

A few mechanics determine whether the resulting composite scores are trustworthy enough to act on or benchmark:

• Sampling. AHRQ’s user guides walk through whether to survey the full eligible staff population or a sample, and how unit-level reporting requirements should shape that choice — a unit needs enough respondents to report a composite score without identifying individuals.
• Mode. Both web-based and paper administration are supported and documented in the survey-specific user guides; mixed-mode administration is common in hospitals with a large frontline non-desk workforce.
• Response rate. Low response rates, or response rates that are uneven across units, distort composite scores and undermine unit-to-unit comparison even before benchmarking against the external database enters the picture. AHRQ’s improvement and action-planning resources treat response-rate management as a first-order administration task, not an afterthought.
• Version consistency. Comparing this year’s Version 2.0 results to a prior Version 1.0 administration requires using the published crosswalk rather than treating composite names as equivalent across versions — the item wording, and in the nursing home survey’s case the composite count itself, changed.

Benchmarking against the AHRQ SOPS database

AHRQ hosts a separate voluntary database for each survey setting — the SOPS Hospital Database, Nursing Home Database, Community Pharmacy Database, and so on — as a central, free repository of submitted results. A few mechanics of how it works are easy to get wrong if you’re relying on secondhand summaries rather than AHRQ’s own submission guidance:

• Participation is voluntary and free, open to any organization that has administered the corresponding survey. Organizations that also administer AHRQ’s approved supplemental item sets (health IT, workplace safety, value and efficiency, for the hospital survey) can submit that data too.
• The comparison you get back is a feedback report showing your own results against the database’s aggregate results — not a percentile rank against every hospital in the country and not a regulatory benchmark.
• The database is not representative of all U.S. facilities in that setting. AHRQ states this explicitly: it reflects whichever organizations chose to submit, which skews toward organizations already engaged enough in safety-culture measurement to participate. Treat a favorable or unfavorable comparison against the database as directional, not as proof of where your organization stands nationally.
• Submission runs on a defined schedule through AHRQ’s SOPS Databases Submission System, with its own forms and specifications per survey — check the current submission window before assuming your data will be included in the next published comparative report.
• Version matters for eligibility. For the hospital setting, only Version 2.0 data is currently accepted into the database submission process; a Version 1.0 administration won’t benchmark against current aggregate results even though the survey itself is still published and usable internally.

Turning composite scores into unit-level action

A completed survey and a benchmarking report are inputs, not outcomes. Three practices separate organizations that use SOPS data to move culture from those that file the report:

• Report at the level people can act on. A hospital-wide composite score for, say, Handoffs and Information Exchange hides which units are actually driving the average down. Breaking composites out by unit (respecting the minimum-n threshold needed to protect respondent anonymity) is what turns a survey result into a work item for a specific team.
• Trend, don’t snapshot. A single administration tells you where you stand; repeated administrations on a consistent version tell you whether an intervention worked. This is the practical reason the pulse-survey option exists for hospitals on a two-year full-survey cycle — it keeps a comparable data point flowing in the off year instead of leaving an 18-month gap in the trend line.
• Close the loop with a documented action plan. AHRQ’s own Action Planning Tool and improvement-resource library exist specifically because a low composite score with no follow-up plan trains staff to see the survey as symbolic. Pairing each low-scoring composite with an owner, an intervention, and a re-measurement point on the next administration is what connects the survey to the culture it’s meant to describe.

How this fits with other safety-culture and quality-reporting work

SOPS results sit alongside, rather than replace, several other measurement and improvement structures already covered in this cluster. A daily safety huddle surfaces near-term issues that a periodic culture survey won’t catch between administrations; a root cause analysis investigates a specific event rather than measuring the underlying culture that shaped how staff responded to it; and the improvement cycle used to act on a low composite score is typically the same PDSA cycle used elsewhere in a quality program. Organizations building toward high-reliability organization maturity generally treat SOPS composite trends as one of several leading indicators, not a standalone scorecard. SOPS is also a distinct instrument from patient-experience surveys like HCAHPS — HCAHPS asks patients about their own care experience, while SOPS asks staff about the culture they work inside, and the two shouldn’t be conflated when a board or executive team is reviewing “survey results” without specifying which one.

Frequently asked questions

How often should a hospital administer the SOPS Hospital Survey?

CMS’s Patient Safety Structural Measure, Domain 3 Statement A, sets the floor: annually, or every two years with a shorter pulse survey administered on target units during the off year. Many hospitals that aren’t specifically tracking PSSM compliance still use an annual or biennial cadence because it’s the interval AHRQ’s own guidance and the database submission schedule are built around.

Is submitting data to the AHRQ SOPS database required?

No. Participation in every SOPS database is voluntary. You can administer any SOPS survey purely for internal use and never submit to AHRQ; the tradeoff is that you lose the external comparison in the feedback report, and any comparative claim about how your results stack up will need a different data source.

Can we compare Version 1.0 results to a new Version 2.0 administration?

Only with care. AHRQ publishes crosswalk documents (for both the hospital and nursing home transitions) that map the two instruments’ items and composites, but the surveys are not identical, and for the nursing home survey the composite count itself changed from 12 to 8. Treat a version transition as a break in the trend line that needs the crosswalk methodology applied, not a straightforward year-over-year comparison.

Which SOPS survey should a skilled nursing facility use?

The Nursing Home Survey, not the Hospital Survey. AHRQ designed the Nursing Home Survey specifically for facilities with long-term care beds, short-stay beds, or rehabilitation services, and explicitly excludes independent living, assisted living, and community care facilities from its intended use — those settings sit outside the population the instrument was built and validated for.

What’s the difference between a composite measure and a single-item measure?

A composite measure combines two or more related survey items into one scored construct (for example, three items combine into the Teamwork composite on the Hospital Survey). A single-item measure, like Patient Safety Rating or Number of Events Reported, is reported on its own because it doesn’t group naturally with other items into a broader construct.