MedicalResearch.com Interview with:
Rashid Bashshur, PhD
Senior Advisor for eHealth
Professor Emeritus, Department of Health Management and Policy
Executive Director, UMHS eHealth Center
University of Michigan Health System
Medical Research: What are the Main findings of the study?
Dr. Bashshur: The main findings can be summarized as follows:
- The prevalence of chronic disease is increasing due to the changing demographic composition of the population and unhealthy life styles. Chronic diseases are expensive, accounting for about 70% of health care expenditures; but they are amenable to telemedicine interventions. These interventions consist of ongoing monitoring of patients living in their own home environments, engaging them in managing their health, providing them with educational materials and the necessary tools to manage their life style mostly by avoiding risky behaviors and adopting healthy ones.
- Telemedicine interventions consist of various configurations of technology (telephone, video, wired or wireless, automated or manual) , human resources (physicians, nurses, combinations, and patient populations at various levels risk levels or severity of illness, The interventions investigated to date also vary in terms of study design, duration of application (during the study period), and frequency of administration of the intervention. Hence, the findings must be viewed from the perspective of the specific methodological attributes of the studies that were performed.
- Using only robust studies from 2000 to the near present, and limited to congestive heart failure, stroke and COPD. With notable exceptions, overall there is substantial and fairly consistent evidence of:
- A decrease in use of service, including hospitalization, rehospitalization, and emergency depart visits as a result of the telemedicine intervention.
- Improved health outcomes.
- Improved event timing for stroke patients.
- Reduced exacerbations of symptoms for COPD patients
- Some interesting trends to be considered:
- Frail and elderly patients, those with co-morbid conditions, and those in advanced stages of illness are not likely to benefit from telemedicine interventions, as compared to their counterparts.
- Visual information as in videoconferencing proved superior to telephone only connection for the diagnosis and treatment of stroke patients.
- There could be a tradeoff between extra nursing time versus reduced physician time in some settings, but the net effect is cost savings.
- There seems to be a paradoxical telemedicine effect in terms of increasing longevity but no reduction in hospitalization, reported in few studies.
Medical Research: Are there unexpected findings?
Dr. Bashshur: Yes, we did not expect to find this level of consistency and strength in findings among studies conducted in some 14 countries over the span of about 14 years. The evidence supporting the use of telemedicine in chronic disease management is strong, consistent and based on valid research (randomized clinical trials and large samples). There were exceptions. In a minority of studies, there was no evidence of reductions in hospitalization/hospitalization but a definite improvement in health outcomes. But here again, one study that recruited frail and elderly patients with multiple comorbidities observed a negative health effect in the telemedicine intervention.
Medical Research: What should clinician and patients take away from this report?
Dr. Bashshur: I think clinicians need to know that:
(1) the acute care model that has been in practice for the past century has outlived its utility or usefulness, and the revolving door arrangement in treating the chronically ill is clinically ineffective and costly. A new model for on-going management of chronic illness is needed, and telemedicine offers a viable and effective alternative.
(2) it would be folly to hold physician’s and other clinicians accountable for the health of the population, when their domain (or their service) can account for less than 30% of a population’s health.
Hence, (3) patients have to be fully engaged in managing their own health, including the adoption of a healthy life style, medication management, and informed decision making in health matters. Patients must assume more responsibility for managing their health, especially when they are afflicted with chronic diseases, singly or in some combination. When given good information on their health status, symptoms and appropriate courses of action, they must embrace this new modality of care and participate in managing their care as a partner with their provider. Patients can be heartened to learn that the symptoms and severity of most chronic conditions can be alleviated, even prevented in few instances. The merit of “shared decision making’ between patient and provider is well validated by research evidence.
Medical Research: What recommendations for future research do you have as a result of this work?
Dr. Bashshur: Recommendations for future research include:
Research is a scientific enterprise that has many tools for uncovering the truth, or what we call evidence. As everyone knows, the rules of scientific discovery are explicit and rigid. But they are there to assure both internal and external validity, namely that what we observe cannot be explained by a rival hypothesis, and also not limited to the setting where the research is conducted. Therefore, researchers have the obligation of considering all validi threats in designing their studies, both internal and external. The telemedicine literature is inundated with inferior or flawed studies. Their sheer volume, by itself, has not advanced the science behind telemedicine nor accounted for advancing the adoption of telemedicine by mainstream medicine. We don’t need more feasibility , pilot or acceptance studies, except when exploring brand new areas. Sample size must provide adequate statistical power to test hypotheses. Typically, samples of less than 150 may not yield adequate statistical power. Randomization of subjects is far superior to “cluster randomization.” And importantly, we must be very cautious and frugal in utilizing the “intent to treat’ escape valve. As in sample surveys, missing data can introduce bias into estimates to be made. Data imputation is used sparingly in sample surveys’, and good researchers tend to do their best in assessing its effect on bias. If used improperly, data imputation is a palatable term for data fabrication. In other words, when we incur the potential for bias as a result of loss of cases or non-participation, we should do our best to exercise extreme caution in data imputation, and we must attempt to estimate its effect on introducing bias into the findings. In sample surveys, calculating statistical or sampling error while ignoring bias is hazardous at best. In experimental studies, ignoring the loss of cases as a result of non-participation is similarly egregious. The percentage of “intent to treat” must be quite small, and some effort must be made to ascertain the nature of the bias it might introduce.
Finally, the sheer large volume of studies that we ignored to report on in this analysis suggests perhaps philosophical questions that must be addressed:
Should we go ahead and continue to conduct small and inconclusive studies from which we can get a publication or two but whose results we cannot trust?
Is flawed research better than no research ?
It seems to be that if we are trying to build the science behind telemedicine, the answer is clear.
Bashshur RL1, Shannon GW, Smith BR, Alverson DC, Antoniotti N, Barsan WG, Bashshur N, Brown EM, Coye MJ, Doarn CR, Ferguson S, Grigsby J, Krupinski EA, Kvedar JC, Linkous J, Merrell RC, Nesbitt T, Poropatich R, Rheuban KS, Sanders JH, Watson AR, Weinstein RS, Yellowlees P.
Telemed J E Health. 2014 Sep;20(9):769-800. doi: 10.1089/tmj.2014.9981.
Epub 2014 Jun 26.