Medtronic Remote Monitoring Study: CONNECT

At the American College of Cardiology 59th annual conference George H. Crossley, MD presented evidence that cardiac patient from remote monitoring (one scheduled in-office visit per year with remote monitoring) verses standard in-office care (four in-office visits per year) cuts the time between the time a cardiac or device related event occurs and when a treatment decision is made.

The title of the study: "The clinical evaluation of the remote notification to reduce time to clinical decision (CONNECT) Trial: The value of remote monitoring."

I present a summary of the method and the results of the study gleaned from the slides presented by Dr. Crossley at the conference.

Hypothesis

Tested hypothesis: Remote monitoring with automatic clinician notifications reduces the time from a cardiac or device event to a clinical decision.

Additionally investigated were rates utilization of the health care system including hospitalization and between treatment groups.

Method

Study participants:  1997 newly implanted CRT-D and DR-ICD patients from 136 US centers were randomly assigned to one of two groups. The first group had 1014 patients assigned to the remotely monitored group and the second had 983 patients assigned to the standard in-office care group. The patients were reasonably well matched for age and gender characteristics.  (A procedure similar to the Biotronik TRUST studies.)

The patients were followed for 12 months.  (On first reading, I found the the time relatively short in that I would not expect enough differentiating events would occur during that time.  However, on further reading, I believe my first impression was incorrect.)

Findings

Time from Event to Clinical Decision

The median time (used nonparametric inferential statistics for the analysis) from the cardiac or device event to clinical decision was 4.6 days in the remote group and 22 days in the in office group. This difference was significant.  The remote group involved 172 patient while the in-office group involved 145 patients.

The cardiac/device events included:

• Atrial Tachycardia/Fibrillation (AT/AF) for 12 hours or more
• Fast Ventricular rate. Of at least 120 beats per minute during at least a 6 hour AT/AFT event
• At least two shocks delivered in an episode
• Lead impedance out of range
• All therapies in a specific zone were exhausted for an episode
• Ventricular Fibrillation detection/therapy off
• Low battery

Total number of events Remote group: 575 and In-office group: 391.  The slides show the breakdowns.

Office Visits

The number of office visits per patient reported are shown below.

                        Scheduled     Unscheduled      All office

Remote group:     1.68              2.24              3.92

In-office group:    4.33              1.94              6.27

The TRUST studies showed a slight increase of more unscheduled visits for the remote group. However, given the nature of the study and that remotely monitored patients would receive only one in-office visit per year, it's remarkable how similar the numbers between the two groups are.

Utilization of the Health Care System

Number of incidents where patients used the health care system show virtually no difference, hospitalization or emergency room. 

However, a remarkable difference was the significant difference in length of stay when there was a hospitalization. The remote group had a mean hospital stay of 3.3 days while the in-office group was 4.0 days with an estimated savings per hospitalization of $1659.

Conclusion

The CONNECT and (Biotronik) TRUST studies show clear benefits from a number of standpoints for remote monitoring.  In addition, the CONNECT study showed clear cost and hospital resource utilization benefits from remote monitoring in that hospitalized patients had shorter stays indicating that they were in better shape than patients in the in-office group when admitted to the hospital.  Quick responses seem to lead to better outcomes as well as cost reductions.

Here is the link to Medtronic's press release of the study: http://wwwp.medtronic.com/Newsroom/NewsReleaseDetails.do?itemId=1268659150410&lang=en_US

Here is another link to an article on the study that includes a link to Dr. Crossley's slides: http://www.cmio.net/index.php?option=com_articles&view=article&id=21235:acc-remote-monitoring-shortens-event-response-time-reduces-hospital-costs

Additional resources

Reimbursement for Remote Patient Monitoring in the United States- Regulatory Antennae Picking Up Demand Signals 

Remote ICD monitoring speeds diagnosis of arrhythmias.(CARDIOVASCULAR MEDICINE)(implantable cardioverter defibrillators)(Report): An article from: Internal Medicine News

Biotronik TRUST Studies: Reprinted Abstracts and Commentary

What follows are published abstracts of the Biotronik studies that provided evidence that Home Monitoring can substitute for quarterly check-ups for ICD patients.  That care of ICD patients can be just as effective with one per year in-clinic check-ups instead of the normal three month in-clinic check-ups.  This was supported primarily by 2008 study.

The 2009 study is a logical follow-up to the 2008 study. This study provided evidence that the Biotronik remote monitoring (Home Monitoring) system can provide early-warning notifications of significant cardiac events faster and more effectively than quarterly, in-clinic visits.  This study has wider implications.   It provides evidence that remote monitoring can provide the kind of care that at one time could only be provided in hospitals.  Furthermore, it demonstrates the kind of capability necessary to provide the kind of early warning that can keep specific, targeted populations out of hospital, thus providing more economical and more desirable health care.

These studies are reprinted with permission from Biotronik.  (I have no affiliation with Biotronik.) 


2008 Study


Evaluation of Efficacy and Safety of Remote Monitoring for ICD Follow-Up:

The TRUST Trial

Authors: Niraj Varma, Cleveland Clinic, Cleveland, OH; Andrew Epstein UAB Medical Center, Birmingham, AL; , Univ of Alabama Birmingham Medical Center, Birmingham, AL; Robert Schweikert Cleveland Clinic, Cleveland, OH; , Akron Medical Center, Akron, OH; Charles Love, Davis Heart and Lung Research Institute, Columbus, OH; Jay Shah, Carolina Cardiology Associates, Rock Hill, SC; Anand Irimpen; Tulane University Medical Center, New Orleans, LA

Background: Remote monitoring (RM) of ICDs may provide daily, automatic device and patient status data and cardiac event notifications. TRUST tested the hypothesis that RM was safe and effective for ICD follow-up for 1 year in a prospective, randomized controlled clinical trial.

Methods: 1282 patients were randomized 2:1 to RM or to conventional (RM disabled) groups.Follow up checks occurred at 3, 6, 9, 12 and 15 months post-implant. In the RM arm, RM was used before office visits (OVs) at 3 and 15 months. At 6, 9 and 12 months, RM only was used but followed by OVs if necessary. Conventional patients were evaluated with OVs only. Follow up was “actionable” if system reprogramming/revision or change in anti-arrhythmic therapy occurred. Scheduled and unscheduled OVs (including responses to event notifications in RM) were quantified for each individual patient per year (pt yr) of follow up. Incidence of death, strokes and surgical interventions (morbidity) was tracked in both groups. 

Results: RM and conventional patients were similar in age (63.3 ± 12.9 vs 64.1 ± 12.0 yrs, p = 0.30), gender (71.9% vs 72.4% male, p =; 0.89), pathology (LVEF 29.1 ± 10.8% vs 28.6 ± 9.8%, p = 0.47;coronary artery disease 64.5% vs 71.4%, p = 0.02), medications (Beta blockers 79.5% vs 75.9%, ACE inhibitors 42.4% vs 46.8%, ARBs 7.8% vs 9.9%, p = NS), indication (primary prevention 72.3% vs 74.2%, p = 0.50), and dual chamber implants (57.9% vs 57.0%, p = 0.76). RM reduced scheduled OVs by 54% and total OVs by 42% without affecting morbidity. Event notifications were managed using RM alone in 92% of cases. Of the remainder resulting in unscheduled OVs, 52.2% were actionable. RM improved adherence to follow-up. 

Conclusions: TRUST demonstrated that remote monitoring is safe, decreases the need for in-office visits, provides early detection of significant problems, and improves ICD surveillance without increasing unscheduled office visits. In conclusion, remote monitoring is a safe alternative to conventional care.


2009 Study


EARLY DETECTION OF ICD EVENTS USING REMOTE
MONITORING: THE TRUST TRIAL
 
Authors: Niraj Varma, MD, FRCP, Andrew Epstein, MD, Anand Irimpen, MD, Robert Schweikert, MD, Jay Shah, MD, Lori Gibson, DVM and Charles Love, MD. Cleveland Clinic, Cleveland, OH, University of Alabama Birmingham Medical Center, Birmingham, AL, Tulane University Medical Center, New Orleans, LA, Akron Medical Center, Akron, OH, Carolina Cardiology, Rock Hill, SC, Biotronik, Inc., Lake Oswego, OR, Davis Heart & Lung Research Institute, Columbus, OH

Introduction: ICDs have extensive self-monitoring capability with diagnostic data available at interrogation. Remote Monitoring (RM) may facilitate data access but this has not been tested. The secondary endpoint of the TRUST trial tested the hypothesis that RM with automatic daily surveillance can provide rapid notification thereby facilitating prompt physician evaluation.

Methods: 1312 patients were randomized 2:1 to RM or to conventional (C) groups. Follow up checks occurred at 3, 6, 9, 12 and 15 months post-implant. RM was used before office visits (OVs) at 3 and 15 months in RM group. At 6, 9 and 12 months, RM only was used but followed by OVs if necessary. C patients were evaluated with OVs only. Unscheduled checks between these time points were tracked. The hypothesis was tested by determining time elapsed from first event occurrence in each patient to physician evaluation.

Results: RM and C patients were similar (age 63 ±13 vs 64 ±12 yrs; gender 72 vs 73% male, NYHA class II 56 vs 61%, pathology LVEF 29 ±11 vs 28 ± 10%; CAD 65 vs 72%, amiodarone 14 vs 14%, primary prevention indication 72 vs 74%, and DDD implants 58 vs 57%). Median time to evaluation was < 3 days in RM compared to < 30 days in C (p < 0.001) for all arrhythmic events (figure) including silent episodes eg AF. System (lead/ generator) problems were infrequent (20 events in RM +C).

Conclusions: Remote monitoring with automatic daily surveillance provides rapid detection and notification of both symptomatic and asymptomatic arrhythmic events, enabling early physician evaluation.

Commentary

2008 Study


Of significant interest would be morbidity rate.  The remote monitoring group showed a .9% higher death rate than the conventional group.  This result was also nonsignificant.  The nonsignificant difference appears to be expected outcome.  Demonstrating a negative - or no difference - is always a concern in research because of the logical problem in demonstrating that something did not happen or that there are no difference between the groups.


I have an additional concern with respect to the unbalanced design.  Unbalanced designs have lower  power of your statistical power - that is, the ability to reject the null hypothesis - than balanced designs. And that would be of concern in study where the expected outcome is no difference.  However, the numbers are extremely large that should off-set the reduction in statistical power created by the unbalanced design.  Since I do not have the raw data, I cannot be sure.  Nevertheless, this seems reasonable. 


The remote monitoring group did have a slightly higher rate of unscheduled appointments - .6 per year in the RM group and .5 in the conventional group; and the actionable percentage was .7% higher in the RM group. The differences could be considered marginally significant with a p = .104. If I understand the circumstances correctly, it seems reasonable that remotely monitored patients would have a higher rate of unscheduled appointments.  Remote monitoring should have the capability of earlier detection of arrhythmic events.  Thus a detected cardiac event would trigger the patient's to request that the patient come to the clinic as soon as possible thus an unscheduled appointment would be registered.  One might expect remotely monitored patients would have appointments that are more demand or situation based than regular, scheduled appointments.




In spite of this difficulties of this design, the conclusions of this study seem reasonable in that the remotely monitored patients who received in-clinic check-ups once per year had similar outcomes to those who receives conventional care with four in-clinic visits per year.


A point of interest.  No comparisons were made between Biotronik and remote monitoring systems provided by other companies such as Medtronic, St. Jude Medical or Boston Scientific. I understand the difficulties and roadblocks in the attempt to assess whether the other systems would be just as effective.  However, Biotronik effectively side-stepped the issue by comparing their home monitoring system against conventional care thus avoiding comparisons with other remote monitoring systems.  Biotronik focused on effectiveness against conventional care and in this case they were successful.


2009 Study

As a study to show the effectiveness of remote monitoring, I believe this study is more effective.  First, it's a better design in that the expected outcome is to reject the null hypothesis - that is, to find a significant difference.  Second, there is a clear case made by the findings that remote monitoring leads to earlier discovery of an adverse event.  One truism in medicine, particularly when it comes to cardiac events, is the earlier the discovery, the better the outcome.  Another thing, ICD patients have been identified as a vulnerable population and rapid reports of adverse events within this population are particularly welcome.


In theory, over time patients remotely monitored should show better outcomes than those who are not.  The data in these two studies does not show that.  However, data from other studies are starting to demonstrate that remotely monitored patients are less likely to be admitted to the hospital.  This is a new area of technology and more research is required.  However, the trends are favorable for remote monitoring.


In this study, the Biotronik remote monitoring system reported arrhythmic events.  The data reported was not early warning or predictive.  The capability to collect predictive data would increase the value of remote monitoring.  Predictive data would allow the clinic (or computer system) following the patient to intervene before the adverse event occurs.  In this study, this was not the case.