The survey mysteries.

Here's a simple question: Are OBRA surveys good or bad for nursing homes?

The answer, of course, is "yes."

By now you have no doubt heard about, or even read, a brace of studies published in a recent issue of the Journal of the American Geriatrics Society indicating that the survey process has worked in improving nursing home care. Data showed, for example, that from 1990 to 1993, after implementation of the Resident Assessment Instrument, hospitalization of frail residents declined by 28%, residents experienced slower declines in ADLs and cognitive performance, and prevalence of dehydration declined by 50%. The word "indicating" is key, however - in research lingo, this means that the study, in this case because of its design, did not prove that OBRA surveys caused the improvements. Rather, the investigators found the coexistence of the RAI and the improvements to be "suggestive."

Two accompanying editorials by Joseph G. Ouslander, MD, and John F. Schnelle, PhD - long-term care research luminaries in their own right - note the discrepancy and propose that future studies of nursing home quality improvement target actual nursing home practices and their effects, rather than focus only (or primarily) on their documentation. A third editorial, by Gwen C. Umen, RN, PhD, sums up the current state of knowledge quite neatly with the line, "What is really going on here?"

The crux of the matter is that the OBRA "black box" of documentation vs. reality has yet to be penetrated. If nursing homes really are improving - and they certainly seem to be - no one is quite sure how; possibilities include improved care techniques, a focusing of resources on the most tractable problems, or simply creative documentation. It seems reasonable to suppose that, at the very least, the existence of OBRA guidelines has focused nursing homes' attention on needed improvements. OBRA has not yet succeeded, though, in developing solid information on how to achieve them.

In a sense, though, none of this matters, because continuous quality improvement (CQI) has not been OBRA's principal focus. Though there are, of course, instances in which state surveyors have helpfully collaborated with facilities in upgrading performance, policing for and punishing violations has been OBRA's major thrust. Nursing home staffs know that, as evidenced most recently by the spate of letters generated by Beth Klitch's July/August "Survey Survival" column ("What's Wrong With the Survey Process?" p. 8), comments that Beth addresses in her current installment (p. 12). What emerges from these communications is a picture of OBRA surveys as a "gotcha" process, subject in its fair execution to the varied experience levels, training, people skills and even political whims of the surveyors. This approach produces striking data (e.g., two-thirds of nursing homes are "out of compliance") - but is this any way to encourage quality improvement?

That doesn't really matter, either, though, because one of departing HCFA Administrator Bruce C. Vladeck's parting shots was that OBRA surveys, if anything, were going to get tougher. He also rejected a South Dakota Medicaid waiver application that would have introduced CQI methods into the process. For nursing homes, the only glimmer of hope for relief these days is HCFA's ongoing study of the "deemed status" issue, due for report this December.

"Relief" is another well-chosen word. Even while nursing homes are apparently improving, nursing home staffs are feeling dismayed, frustrated and fed up. Meanwhile, the resources available to effect these improvements continue to decline, particularly with the recent demise of the Boren amendment. What sorts of trade-offs might lie ahead?

It's not that nursing homes don't want to improve, it's that the system in place for accomplishing this is so needlessly onerous. How can government reconcile its policing responsibilities with the tenets of quality improvement? That, of all questions, is the key. It will take far more creative and courageous policymakers than have recently been on display to resolve it.