Editor's note: Ann L. Hendrich, RN, PhD, is Vice President of Clinical Excellence Operations and Executive Director of the Ascension Health Patient Safety Organization. Her research focuses on how caregivers' work environment affects patient outcomes. She has become one of the foremost experts on the problem of health care–associated falls, and has developed one of the most widely used risk assessment tools.

Dr. Robert Wachter, Editor, AHRQ WebM&M: What have we learned about assessment of fall risk?

Ann L. Hendrich: The largest predictor of who falls is associated with intrinsic risk, things that we carry around with us that make us much more likely to fall. Certainly an unsafe environment, whether hospital, home, or community, could cause an extrinsic fall. But the majority of injurious patient falls have clear intrinsic risk factors. The risk factors that I have been able to statistically prove and validate are the usual suspects of confusion or dementia, altered elimination needs, impaired gait and mobility, which is very predictive, and dizziness and vertigo. Of all of the hundreds of medications that we tested, there were just two categories that statistically increased the risk of falling, benzodiazepines and antiepileptics as independent risk factors. The way the model was constructed allowed us to detect side effects of medication so that we didn't have a very long list of medications that the clinician had to go through repeatedly. The most common side effects of medications are those very intrinsic risk factors that make us fall, like confusion, impaired gait or mobility, or altered elimination.

RW: What does the best fall assessment program look like?

AH: No matter what tool you use, it should be evidence based. I still find homegrown tools being used that overtarget patients. It doesn't take long to figure out that if you're targeting everyone exactly the same way, there tends to be almost a normalized deviance—if everyone's at risk there can be a perception of then there really nothing we can do. A great example of this is age. Age in and of itself is not a good predictor of falls because there are just as many 75-year-olds playing 18 holes of golf daily as there are falling. Those who are falling, it's because they have intrinsic risk factors that are coupled with age, not age as a single predictor of risk.

RW: In the use of your tool, the Hendrich II, when you apply your tool to an average med–surg unit population, what percentage of patients are deemed to be at high risk?

AH: Depending on the case mix and type of facility, it would not be uncommon to see at least half of the population at risk for falling. What's important about successful programs is taking patients on and off high-risk fall protocols is critical. When surgical patients come back, nearly all of them are going to be high risk for falling. Because they still have drugs on board, they're in pain, and their gait and mobility is altered. But 8 to 12 hours later, that same patient is now clearing, their gait and mobility is returning to normal baselines, and you will see their risk points drop. Their risk points have been statistically calculated into a predictive model that calculates the overall relative risk score of falling based on large sample size. It assists care providers to assess and then score patients as "on and off" the fall protocol.

RW: So given that the risk profile will vary with time in the facility, is a nurse assessing this at the beginning of every shift? How does this get operationalized?

AH: I strongly recommend that every patient be assessed on admission as part of the baseline nursing assessment. But whenever the patient's condition changes, they should be reassessed for fall risk. Many patients have a change in status that causes them to be at high risk for falling, and it goes undetected because that fall risk assessment is only being done once a day. The dynamic of today's hospitalized patients is changing much more frequently than that.

RW: Take us through a nurse doing this every shift, in a paper environment vs. a computerized environment. What does the best practice look like in terms of applying the tool and how it gets embedded in and then disseminated from the medical record?

AH: The ideal would go something like this, as a professional nurse my observation skills are constantly collecting data about my patient. The full risk assessment does not really require me to do that much additional if I'm a good observer and taking an accurate history. It may require me to have the patient get up from a seated position. I'm assessing the patient's mental status as part of the normal baseline. I should be also detecting any signs of chronic or acute onset of depression, altered elimination needs, and dizziness or vertigo. Also one of the risk factors is male. Being a male alone does not make you high risk for falling, but the male gender does add one risk point. In addition to that, the antiepileptic and benzodiazepine medication categories actually do statistically increase the risk of falling above and beyond the other intrinsic risk factors. So if the score is 5 or more, I note on the flow chart what the patient's baseline is, and I constantly monitor for changes in any of those risk factors that would cause me to reassess. If I'm doing it electronically, it's populating as I go through my nursing assessment, and it is mapped over to the fall risk assessment. When I've completed it, it automatically calculates the score and gives me some very basic interventions based on the risk factors.

But the real key is that these risk factors help us detect underlying root causes of falls like overmedicated patients, or a patient who is dehydrated and that's causing confusion. Once the risk factors are identified, the evidence-based practice piece of this should prompt me as a clinician to match an intervention against the risk factors present. If the patient has confusion, you could imagine that a PharmD is consulting with the nurse and care team about all the medications that the patient is on and doing medication reconciliation, looking for dose levels and also polypharmacy.

RW: Any idea why males have a higher risk?

AH: This research has been going on for about 20 years, but in the last study I did, which had more than 800 patients in it, we detected this and thought it's just one of those statistical findings that's really not grounded. But I personally started interviewing male patients and qualitatively I can tell you that there were two common themes expressed by every male patient who was not confused. Number one, I don't like to ask for assistance—I want to do this for myself and doing anything less than that makes me less of a man. Number two, I really did not want a female to accompany me to the toilet.

RW: Sounds like that had face validity. Early on you talked about how some of the tools are too broad and your rates are too high and lead to normalization of deviance. But if it turns out that 50% to 70% of people are in an at-risk category, how do you approach this?

AH: The risk factors are pointing to something that is potentially reversible or can be improved with this patient that will not only reduce their fall risk and harm injury, but also may prevent other complications and improve their quality of life. An example, the confusion and the disorientation, is it acute? There could be something metabolic going on with electrolytes or dehydration, especially in the older adult, which can be very subtle. Is there a smoldering urinary tract infection that's gone undetected that's going to cause a big episode of illness and hospitalization? There have been so many stories about these risk factors and how they actually identified something that's reversible, that's the real secret. The hospitals and health care systems who really get it are using the risk factors in just that way.

RW: Talk a little bit about one of the tensions in the safety field. Sometimes we do something that is a good idea but creates another risk. In this area it seems that the pressure for early mobilization seems like a very good idea, but probably does increase the risk of falls. How do you balance that tension?

AH: One of the most common interventions that a physician will write is bed rest for a patient that's at high risk for falling. Everyone is so fearful that the patient may fall but this can actually increase instability. The second most common is to get the patient up with assistance and walk them. The availability of physical therapists or occupational therapists to work with these patients who've been immobile is limited, unfortunately. And it doesn't take weeks or a lot of days for a negative effect to occur with immobility—2 or 3 days of bed rest for any adult is going to cause changes in muscle strength, gait, and mobility. The key is, unless there is some contraindication medically, the human body is intended to be upright, moving, and going. That's why the gait and mobility assessment is crucial. Helping folks to understand the importance of simple strength training that they can do even while they're at bed rest and most importantly getting an exercise program for when they go home.

RW: We've spoken a lot so far about assessment of risk and some of the risk factors that can be modified, particularly drugs that might be causing confusion. Let's shift a little bit to other proposed prevention strategies, particularly in environmental ones, everything from bed alarms to beds on the floor to changes in flooring. What's the state of the evidence in all of those? Do you use them, and what do you think about them?

AH: There are some very basic things that we must be aware of as it relates to the older adult in a hospitalized or institutional environment. Lighting is crucial to the aging eye. It's really important to have adequate lighting, especially after the sun goes down. There have been some carefully controlled studies around the floor surface and the tension underneath the feet and whether or not it causes additional slippage. I think some of the more practical things are around having flat soles, firm shoes, not letting patients get up without something on the foot because especially the older adult when they try to get up and they slip on the floors. Then there's the ongoing controversy around bed height. It's so important to have a bed that matches the height of the individual. I have seen so many 4'8" women on these high beds with a step stool beside it and that actually becomes like an accident waiting to happen because they trip off of that or they trip trying to get up on it. Low bed height has been one of the simple, better interventions that we've done, and it costs very little. Having some hip padding for patients who are unpredictable about getting up or at increased risk for fractures. I know some of the longer-term care facilities have had good luck with that. I think the hip protectors can work, although I've read studies that go both ways. I think the basics of lighting, not having floors that are highly waxed and polished, making sure that there's only half of a side rail up, particularly for the patient who may not alert the nurses when he or she has to get up. Making sure that there's a bed exit alarm feature and that staff respects that and responds to it when the alarm goes off.

RW: The bed exit alarm is built into the bed, and will audibly signal that the patient is getting up. Practically, how does that work? Does the patient need to be right next to the nurses' station?

AH: I think it can work. In today's complex hospital environment, there are so many alarms going off and nurses are pulled in so many directions. There has to be a real respect for the alarm, when it goes off, if there's a high-risk patient at the end of that, the intervention needs to be swift. Otherwise we simply get an alarm and find that patient has already had a fall by the time we get there. One of the things that we're working on at Ascension Health right now that I'm really excited about is connecting the electronic record with the bed with the nurse communication. The way this would work is when a high risk fall patient is identified in the electronic record, the bed is talking to the electronic record and saying my wheels are locked, my side rail is half way up, and—oops—the patient is getting up, and they're high risk! Then it immediately alarms to the nurse who has that patient. So creating a closed loop system around communication.

RW: What about identifying the patient with strategies around different colored shoes or signs on the wall or the door. What do some of those look like and how helpful are they?

AH: It's really all over the page. I've seen everything from falling stars to falling leaves to big red signs. I think the idea is right, but at the end of the day what I see that makes the most difference when the patient is on the unit is that the nurses and other members of the care team understand what fall risk means and the value of reassessment. They understand the consequences of an injurious patient fall, that it can be life altering and life ending. You don't get a second chance with that. So it's really how those clinicians value that and monitor the patients versus having more signs and things on the wall. Where I think the identification is very helpful is when patients are moving around quite a lot within the hospital to treatment areas or diagnostic areas—then having a universal symbol that denotes the patient's high risk for something can be very helpful in those ancillary areas. Because many patients are often parked in the hallways and they fall off carts or they're not being observed when they leave the unit.

RW: Reflect on the Medicare policy now that's also being embraced by other insurers around not paying for adverse events. Hospital falls is on that list, which implies that they are identifiable and largely preventable. As a policy strategy, do you think falls should have been on that list?

AH: Should public policy be driving quality through economic models? The fact that we now have CMS, The Joint Commission, the National Quality Forum, and others aligning around universal quality measures and metrics that encourage health care assistants to become very focused on the preventable aspects of care is a good thing. Are all falls preventable because the policy assumes that they are? The answer is no. Because even on the best day with the best risk factor predictor and with the best intentions, there will still be an occasional patient who does fall. But they should not receive an injury as a result of a fall.

RW: Do falls meet the characteristics of an adverse event that should be subsumed under that umbrella, which at least to my mind is that it's maybe not 100% preventable but largely preventable by achievable evidence-based practices?

AH: I do agree. The same as with pressure ulcers. We have nearly eliminated pressure ulcers here at Ascension, and it was all built around a strategy of using the evidence, sustaining that practice, and measuring it.

RW: What innovations do you see coming down the pike that may transform the way we approach falls?

AH: We are going to see more devices built into the hospital bed. I've seen some prototypes that are really innovative; for example, a bed that actually goes upright and a bicycle arm pops out at the bottom so you can actually start strength training before leaving the hospital, which prevents a lot of those falls that cause readmissions. I think the future really is about how to sustain and strengthen mobility, reduce risk factors, and bring the rehab right to the patient when in an acute care environment, in the patient room versus trying to get the patient transported to a rehab area.