Page 8 Effect of telehealth on quality of life and psychological out comes over 12 months
P. 8
BMJ 2013;346:f653 doi: 10.1136/bmj.f653 (Published 26 February 2013) Page 8 of 20
RESEARCH
Notwithstanding these strengths, some potential caveats should QoL, 84 85 and some evidence indicates that disease specific
86
be acknowledged. All practices in the four primary care trusts measures are more sensitive to clinical change. Forthcoming
were invited to participate in the WSD Evaluation, and 61% analyses will examine the effect of telehealth on these disease
(224/365) agreed and identified at least one patient who met the specific measures. Although patient reported outcomes were
eligibility criteria and participated in either the WSD telehealth the a priori primary endpoint, disease specific clinical markers
trial (179 practices) or WSD telecare trial (217 practices). Data would have afforded a more comprehensive description of the
for pretrial practice characteristics show that participating sample. Such markers include forced expiratory volume ratio
practices differed in practice size, deprivation, ethnic for chronic obstructive pulmonary disease, HbA1c for diabetes,
composition, diabetes prevalence, and WSD site, but had no and the New York Heart Association classification for heart
differences in the prevalence of chronic obstructive pulmonary failure. Unfortunately, there were logistical barriers to the timely
disease or heart failure (web appendix 4). Despite these acquisition of clinical biomarkers. The planned analyses of
differences, recruited practices were not highly selected and disease specific measures of health related QoL (that is the
heterogeneity was preserved (web appendix 4). For example, chronic respiratory questionnaire, diabetes health profile, and
the percentages of non-participating practices categorised as Minnesota Living with Heart Failure questionnaire) will go
having a low, medium, or high proportion of non-white patients some way to describing the clinical severity of the long term
were 32%, 43%, and 25% respectively, while corresponding condition samples.
percentages for participating practices were 25%, 34%, and Despite providing an extensive description of the implemented
41%, respectively. telehealth treatment across sites and conditions (web appendix
Similar concerns could be raised about the representativeness 2; fig 1, web fig 1), there were inevitably some aspects of the
of participants in the questionnaire study. For practical and treatment where detailed data were unavailable. We do not have
ethical reasons, we were unable to collect data on all patients detailed information about changes to treatment or other clinical
who refused to participate in the WSD Evaluation at each stage decisions initiated in response to telehealth data. Detailed
of recruitment. Nevertheless, table 1 shows selection bias in information was not available about the degree to which
those patients who agreed to participate in the nested telehealth participants adhered to their behavioural regimens
questionnaire study relative to the parent trial, and attrition bias (such as monitoring schedules, treatment adherence). We also
in those who were retained at follow-up. Participants allocated do not have detailed information on the degree to which the
to telehealth (in the parent trial) were more likely than those telehealth technology encountered technical problems that
allocated to usual care to agree to participate in the questionnaire interfered with measurement or the exchange of messages
study and to complete one or both assessments. Reasons for the between participants and the monitoring centres. However,
relative advantage of the telehealth arm in recruitment and although about 20% of telehealth participants in the
retention are unclear, though it is consistent with the principle questionnaire study had their equipment removed prematurely
of reciprocity; people receiving a notional benefit (such as during the trial (web fig 2), 80% retained telehealth services for
telehealth) are more likely to comply with subsequent requests. 83 the full 12 months; alternatively, around 10% of telehealth
Potential threats to external validity from self selection or participants had their equipment removed in the first six months
attrition bias underline the need to take care when generalising of the trial while 90% retained it for longer. In web appendix
the results beyond the context of the trial. However, the 2, protocols used by the monitoring centres show that any
relatively high level of practice participation and the large and missing measurement sessions (whether from technical failure
heterogeneous participant sample support our assertion that any or participant non-adherence) were responded to within 72 h.
effect on the external validity of the trial is likely to be minor. Therefore, equipment failure or non-adherence are not plausible
Participants allocated to telehealth did not receive equivalent explanations for the observed null findings.
treatments in all sites. Provision of peripheral telehealth devices A further issue concerns the particular version of telehealth
(web figure 1) and response to biometric readings (web appendix evaluated in the parent trial. Telehealth was implemented as
2) varied substantially by site and long term condition, as did monitoring of vital signs done daily (up to five days per week),
the likelihood of having equipment removed prematurely for supplemented by questions assessing health status and symptom
reasons other than death (web figure 2). In a pragmatic trial, severity, plus an educational component. The educational
this heterogeneity reflects the variability of implementation that component consisted of brief textual information delivered
would be observed in a wider rollout of telehealth, thereby through a static telehealth base unit with a small liquid crystal
increasing the generalisability of the findings. display (LCD) screen (Cornwall and Kent) or a dedicated
In line with the original trial protocol, the analysis sought to interactive television channel (Newham). Telehealth participants
draw conclusions about a general class of technology (telehealth) in Newham could also watch short educational videos with
rather than about the effect of specific peripheral devices (pulse disease specific information. Physiological and symptom report
oximeter, glucometer, weight scales, blood pressure monitor) data were transferred to a monitoring centre using store and
for specific long term conditions. Pooling patients with different forward technology. In terms of a recently proposed
40
profiles of long term conditions could mask differential classification of telehealth, the telehealth system evaluated
treatment effects; therefore, planned analyses will examine the here most closely approximates a second generation system.
effect of telehealth on health related QoL, anxiety, and The current findings therefore cannot be generalised to third or
depressive symptoms for three subgroups of participants indexed fourth generation systems that involve both invasive and
to one long term condition (chronic obstructive pulmonary non-invasive physiological monitoring with real time analytical
disease, diabetes, or heart failure). and decision making support by physicians or physician led
We measured health related QoL using three generic scales specialist nurses. Telehealth can only be studied as technology
in use, and research evidence will always lag behind the latest
(SF-12 physical component score, SF-12 mental component technological advances. However, most systems that have
87
score, and EQ-5D) to assess different dimensions of the been tested so far represent first or second generations; third or
construct. It is recommended that assessment of health related fourth generations should be recognised a distinct class of
QoL includes both generic and disease specific measures to
capture the full range of effect of illness on health related
No commercial reuse: See rights and reprints http://www.bmj.com/permissions Subscribe: http://www.bmj.com/subscribe
RESEARCH
Notwithstanding these strengths, some potential caveats should QoL, 84 85 and some evidence indicates that disease specific
86
be acknowledged. All practices in the four primary care trusts measures are more sensitive to clinical change. Forthcoming
were invited to participate in the WSD Evaluation, and 61% analyses will examine the effect of telehealth on these disease
(224/365) agreed and identified at least one patient who met the specific measures. Although patient reported outcomes were
eligibility criteria and participated in either the WSD telehealth the a priori primary endpoint, disease specific clinical markers
trial (179 practices) or WSD telecare trial (217 practices). Data would have afforded a more comprehensive description of the
for pretrial practice characteristics show that participating sample. Such markers include forced expiratory volume ratio
practices differed in practice size, deprivation, ethnic for chronic obstructive pulmonary disease, HbA1c for diabetes,
composition, diabetes prevalence, and WSD site, but had no and the New York Heart Association classification for heart
differences in the prevalence of chronic obstructive pulmonary failure. Unfortunately, there were logistical barriers to the timely
disease or heart failure (web appendix 4). Despite these acquisition of clinical biomarkers. The planned analyses of
differences, recruited practices were not highly selected and disease specific measures of health related QoL (that is the
heterogeneity was preserved (web appendix 4). For example, chronic respiratory questionnaire, diabetes health profile, and
the percentages of non-participating practices categorised as Minnesota Living with Heart Failure questionnaire) will go
having a low, medium, or high proportion of non-white patients some way to describing the clinical severity of the long term
were 32%, 43%, and 25% respectively, while corresponding condition samples.
percentages for participating practices were 25%, 34%, and Despite providing an extensive description of the implemented
41%, respectively. telehealth treatment across sites and conditions (web appendix
Similar concerns could be raised about the representativeness 2; fig 1, web fig 1), there were inevitably some aspects of the
of participants in the questionnaire study. For practical and treatment where detailed data were unavailable. We do not have
ethical reasons, we were unable to collect data on all patients detailed information about changes to treatment or other clinical
who refused to participate in the WSD Evaluation at each stage decisions initiated in response to telehealth data. Detailed
of recruitment. Nevertheless, table 1 shows selection bias in information was not available about the degree to which
those patients who agreed to participate in the nested telehealth participants adhered to their behavioural regimens
questionnaire study relative to the parent trial, and attrition bias (such as monitoring schedules, treatment adherence). We also
in those who were retained at follow-up. Participants allocated do not have detailed information on the degree to which the
to telehealth (in the parent trial) were more likely than those telehealth technology encountered technical problems that
allocated to usual care to agree to participate in the questionnaire interfered with measurement or the exchange of messages
study and to complete one or both assessments. Reasons for the between participants and the monitoring centres. However,
relative advantage of the telehealth arm in recruitment and although about 20% of telehealth participants in the
retention are unclear, though it is consistent with the principle questionnaire study had their equipment removed prematurely
of reciprocity; people receiving a notional benefit (such as during the trial (web fig 2), 80% retained telehealth services for
telehealth) are more likely to comply with subsequent requests. 83 the full 12 months; alternatively, around 10% of telehealth
Potential threats to external validity from self selection or participants had their equipment removed in the first six months
attrition bias underline the need to take care when generalising of the trial while 90% retained it for longer. In web appendix
the results beyond the context of the trial. However, the 2, protocols used by the monitoring centres show that any
relatively high level of practice participation and the large and missing measurement sessions (whether from technical failure
heterogeneous participant sample support our assertion that any or participant non-adherence) were responded to within 72 h.
effect on the external validity of the trial is likely to be minor. Therefore, equipment failure or non-adherence are not plausible
Participants allocated to telehealth did not receive equivalent explanations for the observed null findings.
treatments in all sites. Provision of peripheral telehealth devices A further issue concerns the particular version of telehealth
(web figure 1) and response to biometric readings (web appendix evaluated in the parent trial. Telehealth was implemented as
2) varied substantially by site and long term condition, as did monitoring of vital signs done daily (up to five days per week),
the likelihood of having equipment removed prematurely for supplemented by questions assessing health status and symptom
reasons other than death (web figure 2). In a pragmatic trial, severity, plus an educational component. The educational
this heterogeneity reflects the variability of implementation that component consisted of brief textual information delivered
would be observed in a wider rollout of telehealth, thereby through a static telehealth base unit with a small liquid crystal
increasing the generalisability of the findings. display (LCD) screen (Cornwall and Kent) or a dedicated
In line with the original trial protocol, the analysis sought to interactive television channel (Newham). Telehealth participants
draw conclusions about a general class of technology (telehealth) in Newham could also watch short educational videos with
rather than about the effect of specific peripheral devices (pulse disease specific information. Physiological and symptom report
oximeter, glucometer, weight scales, blood pressure monitor) data were transferred to a monitoring centre using store and
for specific long term conditions. Pooling patients with different forward technology. In terms of a recently proposed
40
profiles of long term conditions could mask differential classification of telehealth, the telehealth system evaluated
treatment effects; therefore, planned analyses will examine the here most closely approximates a second generation system.
effect of telehealth on health related QoL, anxiety, and The current findings therefore cannot be generalised to third or
depressive symptoms for three subgroups of participants indexed fourth generation systems that involve both invasive and
to one long term condition (chronic obstructive pulmonary non-invasive physiological monitoring with real time analytical
disease, diabetes, or heart failure). and decision making support by physicians or physician led
We measured health related QoL using three generic scales specialist nurses. Telehealth can only be studied as technology
in use, and research evidence will always lag behind the latest
(SF-12 physical component score, SF-12 mental component technological advances. However, most systems that have
87
score, and EQ-5D) to assess different dimensions of the been tested so far represent first or second generations; third or
construct. It is recommended that assessment of health related fourth generations should be recognised a distinct class of
QoL includes both generic and disease specific measures to
capture the full range of effect of illness on health related
No commercial reuse: See rights and reprints http://www.bmj.com/permissions Subscribe: http://www.bmj.com/subscribe
