Study raises questions about home health monitoring

By: Neil Versel | Apr 26, 2012        

Tags: | | | | | | |  |

Neil_Versel_LargeMaybe home-based patient monitoring is not quite ready to take off after all.

Some of you may be aware that I was looking for the right piece of monitoring technology not too long ago to help my family keep tabs on my 93-year-old grandmother, who had been living alone far from the rest of us. That’s not an issue for the moment because she’s been staying with my parents for the last month or so, and she may not be able to go back into her home anyway because a wireless monitor can’t cook dinner or do laundry.

Meantime, I had heard anecdotally that another elderly relative had opted for – and then stopped using – a wireless accelerometer because the device had returned several false positives. Maybe it’s better to be safe than sorry, but the system automatically called for help, thinking the wearer had fallen, when this relative had sat down a little too quickly.

I really have no idea if false alarms are a chronic problem with personal accelerometers, but a recently published study provides some scientific evidence that home monitors need work and are not appropriate in many cases.

Researchers at the Mayo Clinic and the Purdue University School of Industrial Engineering observed greater mortality in elderly patients with multiple health issues who were provided Intel Health Guide base stations and several pieces of monitoring equipment, compared to those receiving typical, “patient-driven” care. In fact, the mortality rate was three times higher.

According to the paper, published last week in the Archives of Internal Medicine, 14.7 percent of those remotely monitored died within 12 months of being enrolled in the study, while just 3.9 percent in the control group died during the same timeframe.

There wasn’t much difference in the number of hospitalizations or visits to the emergency room, whether the patients had home monitoring or not, and the researchers did not have an explanation for the higher mortality. Also, they noted that patients without the monitoring systems had regular telephone contact with nurse case managers, so these patients may have been more aware of serious symptoms than others who did not talk to nurses often.

The researchers further suggested that there might not be enough infrastructure in place yet to “fully optimize” case management in this vulnerable population. (The mean age of study participants was 80.3 years.)

Still, the findings raise a lot of questions, and not just from reporters. In fact, two physicians from the University of Iowa Carver College of Medicine provided commentary in the same Annals of Internal Medicine issue.

“The results of this study are important and sobering and warrant careful consideration. It might be tempting to discount the lack of benefit of the telehealth intervention for any number of reasons. For example, it might be possible that the lack of effect in the telehealth intervention compared with usual care was a consequence of the fact that usual care was already outstanding because the comparator group was already receiving care from a top-tier integrated delivery system (Mayo Clinic, Rochester, Minnesota),” write Dr. Scott R. Wilson and Dr. Peter Cram.

“We would caution against such discounting of this study and its negative findings. In contrast, we would argue that this study joins a growing body of literature suggesting that home telehealth does not reduce readmissions or ED visits,” they add. However, they caution against reading this study as a “blanket indictment of a potentially useful technology.”

What was the study trying to accomplish, they wonder? It is unclear whether the researchers were trying to reduce hospitalizations and ED visits or save lives. Nor do they know if Mayo Clinic staff received proper training on the technology. “It seems almost certain that these subtle differences in strategy and training of telehealth providers would have major influences on study results, but these sorts of details are often difficult to describe adequately within the constraint of published articles,” Wilson and Cram.

Clearly, telemonitoring is not appropriate in all cases, but I don’t have to be a scientist to know that it might be an effective tool for a lot of people.

Wilson and Cram offer some very good advice: “While awaiting the answers to these questions, we would advise payers and physicians to move slowly in implementing telehealth programs on a wide scale.” But I see no good reason not to try if clinicians or insurers have specific goals in mind and they figure out what types of patients stand the best chance of benefiting from home monitoring technology.

  • Lawrence PhD

    Wow!  Go to Medapps they have been awarded for their remote patient monitoring devices. With FCC and FDA involved in terms of regulations its unclear what will be their potential interventions impact will be. Remote patient monitoring has potential but validation of these devices is needed especially in terms of false positives or negatives.

  • Laurie Orlov

    Do you really think this study proves anything one way or the other? See comments on this blog post, too small a study to draw any conclusions — except context of care provided makes all the difference :

  • Brian Dolan

    +1 Laurie, your post and the resulting comments are a must read — esp. comment that VA figured this out years ago.

  • CMS Telehelath, LLC

    First let me comment that we agree there are problems with false positives with the accelerometer and that is why we will not provide this to our clients. Unfortunately, many of the comapnies who offer this are well aware and I would like to think that they tell this to their clients before they purchase it. Second, Laurie’s analysis was excellent on the study that you cite in your article. Seems that just because a “name” institution puts out a study doesn’t make it gospel-in fact many use that their name that carries weight rather the quality of their reseracher,data and accurate findings. It wasn’t long ago that Yale did a study on Telehealth which was one of the poorest studies ever done.In fact,Telehelath without proper follow-up with companies and individuals who are proactive in monitoring and follow-up are often the reasons for poor outcomes.Additionally, offer a one-size fits all hub and peripherals is not often the answer-as we have seen in many cases.
    Third,there are times as well as that patients  should not be provided a Telehealth solution and there are times that is an excellent tool to supplement what a caregiver does.But,like all things one needs to deal with reputable companies and individuals who understand patient care and technology and want to do what’s right. We often see good sales people with little knowledge of patient care and the ability to offer the best solution because they have only one product and other times we see good technology but with company reps who have little or no understanding of patient care. It’s about time we bring integrity back to healthcare and do what’s right for patients’

  • Geoffrey Clapp

    Laurie’s article was excellent and I agree with 99% of her postion. My position is obviously (or at least, previously)  biased, however, there are two points we need to make sure are not thrown out (so to speak) with the bathwater in our agreement (to be clear, I’m not saying Laurie said these things):

    1. The point about the pedometer that Niel makes is important. There are a lot of *very* bad home medical devices in the market. We had to integrate with many of them, and they did not work well. One thing we worked on with the VA on was limiting the set of devices to higher quality devices that fit a work flow – and even then, many of the devices/sensors were not up to par. This is still a market that needs quality improvement, metrics, and rating system.  For example, we’re missing a British Hypertension style device rating (there, for BP cuffs) to keep bad devices out of remote patient monitoring in the home for high-risk patients.

    2. The hypothesis about lacking infrastructure to support home patient monitoring is spot on. As the market struggles to get EMR integration into workflows, telehealth and RPM will (and DO) struggle to get into clinical workflows and systems. This is still, very much, an uphill battle. It does not mean Telehealth is bad – far from it – but this is a real issue. Despite success at the VA (and the work they have done is Herculean) , nearly 10 years after the first national contract (not including the 4 years of pilots before that) and MOST of the telehealth data is not integrated into VistA, including subjective data. This is still a hard problem and the EHR market is not currently focused on – nor should it be, given their current incentives and priorities.

    So, while of course I believe fully and completely in telehealth and remote patient monitoring – rather than simply separate into two camps of “right” and “wrong” about the study, we’ll be better served to pick up on a few of the points that were made that have not been solved – and solve them.

  • Lawrence PhD

    Your comments are very well taken. I work in evaluationn research/health systems research in the US Government/World Health Organization and rest assured when I indicate that if evaluation tools are used reports studies etc will fail and results tainted. At numerous events I raise the question as to evaluation that your product works etc. and golden silence prevails. From my perspective patient and health education is the key and am working in this area.  

  • Jofjuu22

    It seems ridiculous to me that anybody believes that having an “added” technology system would cause more deaths…. that’s like implying a baby monitor causes more infants to die. I understand that telehealth technologies are scary (Accelerometers are not really considered telehealth, by the way), but remember they are just a tool in providing care. If people think they are no longer accountable or responsible, that’s ignorant. This study does nothing, but convince people that we should not innovate and go backwards as a society.

  • Neil Versel

    Laurie, I saw your piece after I finished writing mine. I don’t think it proves anything other than this field still needs a lot more study.

  • Neil Versel

    Yes, I remember that Yale study very well. I actually wrote a commentary in Columbia Journalism Review taking the LA Times to task for calling that study a good idea that just didn’t work out.

  • Neil Versel

     P.S., Your commentary was great.

  • georgemargelis

    I agree, what needs to be evaluated is not the technology, but the associated clinical model. Capturing patient data without having a clinical model on the response leads to exactly what is shown in this study. If you know the status of a patient continuously you are more likely to find times when they are sub optimal. If our response is, go to the ED, then you will see more ED visits. If on the other hand you have the ability to respond in a way that brings the patient back to a better state, whilst keeping them at home, and maybe even using educating them on how to avoid getting to that state again, then we have a good clinical model.

  • Guest

    well this is showing the other side of using technology. The main target must be save lives along with reducing the visits to ED.

  • Bob Pyke Jr

    Not every pt is a canidate for telehealth, if you don’t have the care coordinated or have a back up plan, then all you have is a piece of equipmet sitting there and a pt at risk…

  • Davesandb

    In the UK we have just had a randomised control trial on a national bases – three different areas of the UK. Participants  3000 or so. The headline results are very much worth you having a look at especialy the mortality figures which wee in y opinion exceptional – in a good way may I say. Please see this link.

  • Evan Freedman

    I have some line-by-line criticism/questioning of the study’s methodology and possible uncontrollable variables, but I can’t post a full PDF with my annotations without violating something, so here’s a selection of critiques (I’ll try to put things in context for those who don’t have access to the full paper):
    Does each clinic provide equal quality of care?See “Study Design,” page E2 of PDF The study participants were patients at four different Mayo Clinic Employee and Community Health program sites. It is hard to determine from the paper if the random assignment of participants is done on the total scale, or at each individual site so that one site does not have an overwhelming percentage of the Telehealth patients. If the distribution at each site is not near 50/50, then potential differences of the quality of care delivered at each site may influence results.
    Telehealth and lack of social interaction?The device worked asynchronously… A registered nurse oversaw approximately 100 patients and communicated with the individuals via phone or videoconference if alerts arose.“Telemonitoring,” page E2. I know the asynchronous nature of telemedicine is one of its strong points, but perhaps the delay in social interaction between the time of recording symptoms and hearing from a nurse had an effect on the telehealth patients. We know social interaction is especially important when it comes to geriatric health. Particularly face-to-face. I don’t mean to suggest that telemedicine is incapable of delivering social interaction, but as mentioned later in the paper, In total, 3942 phone calls were made to participants using a telemonitor.“Results,” page E3.3942 calls to 102 patients over the study period? That’s about 38.5 calls per patient over 12 months, or a call to an individual patient every 9.5 days. I did not catch any mention of using videoconferencing features.
    Telehealth influencing physical activity?Usual care patients had access to primary and specialty office visits.“Usual Care” on page E2. Telehealth patients were brought in for visits at the discretion of their physician. Perhaps the fact that telemedicine patients had less obligation to leave their home affected their level of physical activity? Decreased physical activity will typically result in earlier mortality.
    No change in telemedicine technology over period of study. As a control, the researchers used the same 2009 Intel Health Guide for all Telehealth patients, even for those who had enrolled as late as July 2011. This is obviously necessary when conducting a study. But we must acknowledge the improvement in telemedicine technology since November 2009 (when the study began), such as the transition from the Intel-GE project to what is now Care Innovations (as Laurie mentions in her blog post). This study uses 2009 tech. Just keep that in mind.
    What’s with the big drop in hospitalizations for “Usual Care” patients post-enrollment? The number of hospitalizations for the “Usual Care” patients dropped from 109 pre-study to 85 post-enrollment. What kind of care were these patients receiving before enrolling in the study? From the paper’s description of recruitment methods, it sounds as if they were receiving care from the same Mayo Clinic location prior to the study. What’s the explanation then for this drop in hospitalizations? Were doctors and nurses aware of a patient’s involvement in the study (or made aware by patients who like to talk) and therefore were subject to the “Hawthorne Effect” (which the authors do acknowledge in the conclusion of the paper), wherein the act of being monitored changes behavior, i.e. doctors and nurses provide greater care to the “Usual Care” patients?
    I have so many more questions and critiques, but this comment is already egregiously long.

  • georgemargelis

    Hi Evan
    I like your systematic review of the article. There are lots of unanswered questions, in particular for me whether a model of care that focussed on keeping the patient at home was at the heart of the project. If the standard of care was referral to an ED or clinic if parameters were abnormal, then telehealth is a great way of building up those referrals. If however a clinical model is in place that looks for trends, responds with a relevant interaction that is designed to keep the patient at home, and is supplemented by an effective home health service, then I suspect you would see a different outcome.

    The question of the use of video conferencing is also very relevant. Phone based interventions have been shown to be ineffective, a recent Australian report showed no effect on ED admissions. Video conferencing can provide a much more effective communication method for evaluating a patient’s status,

  • Garry Welch

    It may be that this study is helping us understand that we early in the journey and have a lot to learn about the delivery of telehealth and how to configure chronic care nurses, patients, caregivers, and providers into a clinicallly valuable system that works for everyone. We are rolling out a diabetes telehealth program at a community health center with bluetooth/wireless devices to measure bg, bp, and track meds. The most interesting thing is not the great technology that send the data up to the cloud and across to our clinical application without dialups and modems but the process we have just started of figuring out how clinical care is currently delivered in the community health center and how we can configure the alerts, reminders, reports, and protocols to work and keep everyone one happy and feeling like this helps them solve their problems.

  • Pingback: Elderly still not using apps for health, but tomorrow’s seniors might | mobihealthnews()