Friday, November 19, 2010

Remote Cardiac Monitoring: The Debate Begins

An article in this month’s New England Journal of Medicine (NEJM) (http://www.nejm.org/doi/full/10.1056/NEJMoa1010029) reported the results of a six month clinical trial which evaluated the effectiveness of a telephone monitoring service for patients that suffered heart failure. The authors of the study, based at Yale University, reported that this service had no impact on the health of the patients when compared with a control group.

In the weeks to come, this article is certain to generate a lot of comment and debate. It provides some important results and we need to assess the issues fairly. But we may see a considerable amount of unwarranted conclusions and unfair criticism of remote cardiac monitoring, of telemedicine and even of the study itself. As we have seen before, once the facts and scientific analyses end, the generalizations and misstatements begin.

Beware of the headlines. The study was narrowly focused with biases in its design that do not allow broad generalizations. I have included below some facts about the study, and a few of the lessons learned and finally comments on the validity of the conclusions.

THE FACTS

The study used a random sample of 1,653 heart failure patients with a median age of 61 who were divided evenly into a test and control group. Over six months, the test group was asked to telephone a call center daily and follow automated prompts to a series of questions about their health status. At the end of the six months a comparative study was done on the rates of hospital readmission or death of the two groups. Using statistical analyses, it was determined that there were no appreciable differences between the groups. Based on these findings, the authors concluded: “Among patients recently hospitalized for heart failure, telemonitoring did not improve outcomes.”

Despite pre-study training, approximately 15 percent of the patients in the test group never started calling in for monitoring. By the end of the study 45 percent of the remaining test patients stopped making their monitoring calls. The comparative statistical analysis was based on the entire test group, including those that failed to participate in making the phone calls.

The authors of the study decided not to use automated data collection devices in the test patients’ homes nor were medication minder devices included. Although attempts were made to remind patients who failed to make their telephone calls, the study was based on the patients initiating the calls and self-reporting their information including weight gain or loss.

LESSONS

One of the most glaring outcomes (as well as one of the most glaring errors in the study design and data analysis) was the lack of patient participation in the test group. Of the 826 patients in the original test group, only 707 ever started making the monitoring calls and by the end of the study only 390 kept making the calls. Perhaps some types of monitoring devices in the home (or maybe just an annoying reminder beep) could have improved participation. Certainly greater efforts need to be made to engage patients in their healthcare, whether that is reducing weight, taking their medicine or using technology to monitor and report on their conditions.

An accompanying editorial in the NEJM focused on several concerns. For example, telephone monitoring services may not be asking the right questions of patients, suggesting that we need to reevaluate the appropriate ongoing physiological data that should be measured for heart failure. Also more timely staff follow-up with the test patients for corrective action may make a difference. Alternatively, smart software automatically generating diagnoses and treatment plans based on the data and supplementing health provider support, may lead to significantly improved outcomes.

COMMENT

In the Discussion portion of the article, the authors chose to take a swipe at vendors and use an incredibly broad generalization about all patient monitoring: “In an environment in which vendors promote their products to health systems that are under increasing pressure to reduce readmission rates, the knowledge that telemonitoring is ineffective suggests the need to consider alternative approaches to improving care.” Even worse, one of the authors chose to write an article on the study for Forbes.com with the headline: “Why Telemedicine is Overhyped.” This raises serious doubts about any intent to be fair and balanced.

Unfortunately, we don’t know whether the test group patients that participated in telemonitoring to the end of the six months showed any differences in their rate of re-hospitalization. It would have been helpful and relatively easy to show this data. It is troubling why this was not included.

The authors made a point of criticizing previous studies as using too small of a sample, but it does not appear that they ever conducted a literature search on the subject. Despite quoting several other studies, they failed to reference a landmark study of telemonitoring by the Veterans Administration, concluding positive results, that was based on a broad grouping of patients over a longer time span.

I am concerned that in the days to come we may witness misinformed discussions and articles about remote patient monitoring as a result of some unfortunate printed malware that slanders all of telemedicine. Let’s not throw the baby out with the bathwater, especially when the bathwater may just need adjusting.

11 comments:

  1. Interesting discussion. Could you please post a link to the VA study?

    ReplyDelete
  2. I completely agree that the study is somewhat flawed both technically and methodologically, and moreover, might have even been contemplated with an "a priori" negative bias, which is reflected in overly simplified, exaggerated and unduly generalized concluding remarks. The data on comparatively new modality, obtained from a single study, however large and of RCT design, conflicting with results of previous studies (including such time-honored format as Cochrane analysis), merely warrant further investigation - which in fact would have been much more balanced conclusion. Thank Mr. Linkous for his comments; we believe there is an urgent need to voice concerns regarding premature conclusions undermining importance of an entire direction in modern health care, a direction that has not once shown promise and value, which is b.t.w. absolutely intuitive to any health specialist.

    ReplyDelete
  3. The days of reimbursement for devices are numbered. Health plan coverage for services that may incorporate devices are the future and within these services utilization, technique and outcomes focused programs form the basis of how we may all benefit from RPM technological advancements.

    ReplyDelete
  4. We have to take this study's results constructively: it's all about good ol' patient compliance. It's up to the industry and health care structure to solve this problem. The study points out to a real 44% improvement in the earlier study when proper controls and motivations were in place. So, the principle is sound, but the implementation is a hurdle. What else is new?

    ReplyDelete
  5. From the time Luis Katchis invented and patented the loop cardiac trans-telephonic monitor, an entire industry emerged, and countless of lives were saved. You can Google remote cardiac monitoring and see how effective this technology is. It is just common sense and does not take a rocket scientist to understand that remote monitoring and telemedicine can improve patient outcomes. For that matter, the problem with all of medicine is patient compliance. If existing technology using automated data gathering and reporting was used, the study would have been a lot more viable.

    ReplyDelete
  6. these monitors are a great idea, however they sshould all automatically report an arrythimia, i am having to use one of my now 4 year old daughter and having to press a button and transmit each time i think she is having a symptom is ridiculous, she rarely tells me when anything is wrong and her rate has been as high at 256 at times. if i do not know when it is happening then how can they be accurately recorded. any i cannot expect a 4 year old to understand the importance of pressing this herseld or remembering to tell me she has pressed it if needed for that matter.

    ReplyDelete
  7. i agree complaince is an issue, however i feel these should automatically detect and record arrythmias, i would not mind transmitting it myseld each evening but sometimes people do not know they are having the problems that are going. how can we fault the patient for that.

    ReplyDelete
  8. You cannot extrapolate imaginary results from what could have been tested (imaginary intervention) in this trial.

    The study tested an intervention (simple tele-monitoring device with reporting of weights and patient self-reported symptoms) using the gold standard of an RCT by a group of investigators well trained and respected in this area.

    Unfortunately, the results are what they are. Not what the critics would like them to be. Interestingly, patient compliance appeared to be a major driver of device use and ultimately any benefit that may have been derived from therefrom!

    The naysayers would be wise to take a deep breath and remain objective...where do we go from here?

    ReplyDelete
  9. It's a stretch to consider this study "remote cardiac monitoring" and almost an embarrassment that it made it to the NEJM. There are many aspects to the potential patient outcomes using technology, the least of which is early detection of A-Fib, for example. A few post-op phone calls aren't going to cut it.

    ReplyDelete
  10. On first pass I don't know how you can even label this study as "monitoring". Self reporting is not monitoring. Flawed uses of terminology in this way cast incorrect biases into the industry particularly for payers who will use any opportunity to find cause not to pay. Correct the basis for the definition of the study from monitoring to "self assessed, and self reported patient compliance tallying".

    ReplyDelete
  11. In response to your comment: "Unfortunately, we don’t know whether the test group patients that participated in telemonitoring to the end of the six months showed any differences in their rate of re-hospitalization", I suggest you read-up on intention-to-treat analysis.

    Regarding data from the VA, the most recent study I see from them ("Long-term effect of home telehealth services on preventable
    hospitalization use", Huanguang et al, 2009) shows an increase in rate of hospitalization among patients with CHF.

    Within the medical field, there is no shortage of great ideas - and no matter how wonderful they sound in theory, they usually fail in practice.

    ReplyDelete