ARTICLE
Auteur(s) : Timo Rimnera1, Eva Blozika1, Barbra
Fischer Casagrande2, Jan
Von Overbeck1
1Swiss Centre for Telemedicine Medgate,
Gellertstr. 19, CH- 4020 Basel
2Dermatologie-Praxis, Wallisellenstrasse 301a, CH-8050
Zürich-Oerlikon
accepté le 13 Mai 2010
Dermatology is particularly suitable to the use of advanced
electronic communication technologies for delivery of care [1, 2].
In teledermatology, store-and-forward technology is most commonly
used, as opposed to live synchronous communication. It involves
sending digital images to the data storage unit of a specialised
centre, does not demand the presence of both parties at the same
time, and generally does not require expensive equipment [3-5].
The diagnostic accuracy of store-and-forward teledermatology
(SFT) and patient management decisions based on
that technology have been shown to be similar to in-person
clinical encounters [2-9]. Patients undergoing SFT also achieve
similar clinical outcomes [10]. Additionally, SFT technology has a
positive impact on process-of-care outcomes, including a reduction
in time to consultation, shorter waiting times, and shorter
consultations, especially with respect to conservative skin
conditions [3, 11]. Taking into account indirect effects, such as a
reduction in resource use and avoidance of patient-generated costs
such as those associated with travel, lost time from work, or
caregiver reimbursement, some studies have also shown that SFT is
cost-effective [3, 12].
Digital photography for SFT has been shown to produce
high-quality images leading to diagnostic concordance rates that
compare favourably with in-person clinical diagnoses [13]. However,
it has mainly been used for teleconsultations between general
practitioners and dermatologists, or for exchange between
dermatologists, and therefore the photographs were taken by medical
professionals [11, 14]. Digital cameras have become widespread in
the population. Enabling patients to photograph and submit digital
images of their own would be a logical and sensible step towards
patient participation in their dermatological care [15]. However,
the quality of digital skin images submitted by patients and their
value for making a diagnosis have scarcely been investigated, and
the few studies that exist were performed in particular settings or
conditions [15-17].
The aim of this analysis was to investigate the quality of skin
images submitted by patients with no special prerequisites, and to
evaluate the feasibility of making a diagnosis and therapeutic
recommendations.
Materials and methods
Context of the evaluation
The Swiss Centre for Telemedicine Medgate provides
teleconsultations covering the whole spectrum from general
questions about health behaviour to specific acute health problems.
When a person calls, a telemedical assistant records the medical
problem, makes an assessment of its urgency (based on internal
quality standards and under medical supervision), and assigns a
time interval during which a medical doctor shall call the patient
back for a detailed teleconsultation. In 2008, about 30% of the
Swiss population had free access to Medgate teleconsultation
services as part of their obligatory health insurance.
This is a prospective evaluation of patients who received a
teleconsultation from the Swiss Centre for Telemedicine Medgate
because of skin problems in March 2008. Teledermatological
processes of the centre conform to the American Telemedicine
Association's Practice Guidelines for Teledermatology [1]. The
responsible health authorities have authorised Medgate to prescribe
medications in specific pre-defined clinical situations without
physically examining the patient. One of these is that a
dermatologist recommends treatment with a certain medication after
thorough dermatologic assessment of the clinical picture.
Evaluation design and recruitment
Patients (or the guardians if the patients are children) who called
the Swiss Centre for Telemedicine for a skin problem were invited
by the telemedical assistant to submit digital images. Patients
asking for cancer screening of pigmented lesions were directly
transferred to a dermatologist and a priori excluded from
teleconsultations because those patients must undergo a
dermatoscopical examination of the lesion. Eligible patients were
orally instructed on how to take the photographs, and were sent an
email including written instructions. The email did not include the
patient's name or other personal identifiers, except for the case
number for patient re-identification and linkage with clinical
information in the electronic patient record system. Patients were
asked to reply to this email by attaching their images and not to
add personal or clinical information. The oral instructions
included the following information: Patients were asked to use a
digital camera instead of a mobile phone camera if available. They
were told to remove clothing, jewellery and makeup and to ask a
second person for help if needed. Additionally, they were
instructed to apply the macro function for close-up views, to take
pictures in well-lit, preferably ambient light settings, to avoid
flash, and to use a uniform, rather dark background. Images were
not to be compressed before sending to the telemedicine centre. The
written information containing instructions for taking different
views of the skin lesions are found in figure 1. For example,
patients were instructed regarding the minimum distance and angle
of the camera to the skin.
During the two week evaluation period, 51 patients (or in
case of children: their guardians) called the telemedicine centre
because of a skin problem. Of those, 5 patients (10%) were not
willing or not equipped to send in skin images. All patients (N =
46) who submitted digital images of their skin problems were
included in the evaluation. Images were transferred to a
dermatologist from an independent office together with the
patient's history. The medical history included information on the
incidence, localisation, spreading, and progression of the skin
problem, other symptoms such as pain or itching, the cause or
exacerbating factor suspected by the patient, therapeutic
interventions already applied and their effects, exposition to
cosmetics, therapeutics, animals, sun, etc., visits abroad, sexual
anamnesis, prior dermatologic diseases, medication, and social
anamnesis. This information was obtained by the doctor of the
telemedicine centre during the initial teleconsultation. The
dermatologist (co-author BFC) assessed the image quality and
feasibility of making a dermatological diagnosis. The
dermatological diagnosis, together with the dermatologist's
therapeutic recommendations, were sent back to the telemedicine
centre by email. Then, a doctor from the telemedicine centre
discussed these results with the patient in a second telephone
consultation and initiated teleprescription if it had been
recommended by the dermatologist. If it had not been possible to
make a diagnosis on the basis of the images, the patient was
referred to a dermatologist for face-to-face consultation.
Technical specifications
For image acquisition, the minimal resolution was 800 × 600 pixels.
The emails sent to the telemedicine centre containing patients’
images were kept in a separate database at the original acquired
resolution. The case number provided in the email was used to link
the individual images with personal and clinical information from
the electronic patient record. Access to the electronic patient
record system as well as to the image mailbox was possible only
after personal authentication by smart cards and individual
passwords. Transmission of images and clinical information between
the dermatologist and the telemedicine centre were done via email
without inclusion of personal identifiers except for the case
number.
For image display, monitors with a 0.29-dot pitch were used. The
brightness of the monitors was 300 cd/m2, and the
contrast ratio was 1000:1. When viewing images, brightness,
contrast, zoom, pan, rotate and flip functions were available.
Monitors were positioned in ambient room light fortified by
artificial room light. In case of critical connectivity, the Swiss
Centre for Telemedicine has redundant and recovery systems as well
as infection control procedures in place. Telemedicine procedures
are under continuous quality assessment and control and comply with
the relevant safety laws and regulations in Switzerland.
Assessment of images
Image quality was assessed by rating image focus, display detail,
and availability and helpfulness of an oblique image. As patients
were advised to send in several images, image focus was rated for
the whole package of images. Based on the image package, the
dermatologist attempted to make a diagnosis, in a first step
unaware of the history, in a second step in combination with the
history of the patient. If making a diagnosis was possible, the
dermatologist recommended one or more therapeutic interventions.
The overall quality of the image package was evaluated by a
composite measure including all three dimensions of image quality
assessed (image focused, display details adequate, and oblique
image available).
Statistical analyses
Statistical analyses included simple counting and frequency
distributions of the demographic characteristics of the patients,
and of image quality and diagnosis characteristics. Analyses were
done using Microsoft Office Excel 2003.
Results
Characteristics of the patient population
Ages ranged from 1 month to 65 years. The mean age was
30 years. Slightly more than half of the patients (26 of
46 patients, 56%) were female.
Image quality
Of the 46 image packages, 26 (57%) were rated to be focused,
and 12 (26%) were not focused. For another 8 (17%) some of the
images in the package were focused, the others were not. Display
details of the majority of image packages (33/46 or 72%) were rated
to be adequate, whereas in 13 (28%) the display of the skin problem
was not sufficient. Fifteen patients sent an oblique image (33%),
but 31 (67%) did not (table 1). The
dermatologist felt an oblique image would have been helpful in
33 cases (72%), in 14 of those (42%) there was an oblique
image available. For case examples of patient-provided image
packages see figures 2, 3, and 4.
All three dimensions of image quality (image focused, display
details adequate, and oblique image available) were present in only
12 patients (26%) (table 1).
However, if only those images for which the presence of an oblique
image was rated helpful were taken into account, 23 patients
(50%) submitted image packages of global good quality.
Table 1 Image quality, feasibility of making
a dermatological diagnosis and recommendation
of therapeutic interventions (N = 46)
|
Outcome characteristic
|
N
|
(%)
|
|
Image quality
|
|
All images focused
|
26
|
(57%)
|
|
Partly focused
|
8
|
(17%)
|
|
Display details adequate
|
33
|
(72%)
|
|
Oblique image available
|
15
|
(33%)
|
|
Dermatological diagnosis
|
|
Feasible
|
31
|
(67%)
|
|
Differential diagnosis feasible
|
8
|
(17%)
|
|
Recommendation of therapeutic interventions
|
|
|
|
Feasible
|
31
|
(67%)
|
Dermatological diagnosis
In none of the image packages, was making a reliable diagnosis
feasible without knowing the anamnesis of the patient. However,
being aware of this information, making a diagnosis was feasible in
31 cases (67%). The diagnoses comprised allergic (urticaria,
contact dermatitis, atopic disease), infectious (tinea corporis,
erythema chronicum migrans, furuncle, viral exanthema, scarlet
fever, secondary bacterial infection), tumorous (granuloma
pyogenicum, nevus, histiocytoma), or other (psoriasis, hematoma,
pityriasis rosea Gilbert, acne neonatorum, pulpitis sicca)
diseases. In another 8 cases (17%) the dermatologist was able
to narrow down the differential diagnoses, but not to make a
specific diagnosis. In 7 patients (15%), it was not possible
to specify potential diagnoses at all (table 1).
Recommendation of therapeutic interventions
For all of the 31 patients (67%) for whom it was possible to
make a diagnosis based on the medical history and patient-provided
image packages, a recommendation concerning further treatment or
management was made (table 1).
Recommendations included application of topical steroids,
fungicides, antibiotics, vitamin D derivates, or disinfectants,
oral antihistaminic or antibiotics drugs, application of lipid
regulating cream, stop of home care measures and medicines,
watchful waiting, counselling, and referral to a dermatologist for
a face-to-face consultation.
In 20 patients (65% of 31), according teleprescriptions
were recommended. These were initiated by the physician at the
telemedicine centre in the second telephone consultation, and these
patients were not referred for a face-to-face consultation. In
2 patients (6% of 31) counselling, reassurance and watchful
waiting were sufficient and no further diagnostics, therapies or
referral were needed according to the dermatologist's assessment.
Nine patients (29% of 31) had to be referred for a face-to-face
consultation because the condition was not appropriate for
teletherapy, or because further diagnostics were needed (e.g.
secondary bacterial infection, scarlet fever, allergic reaction,
histiocytoma, evaluation of a nevus, atopic disease, psoriasis,
contact dermatitis). These patients were given recommendations for
the time interval between the second telephone consultation and the
recommended face-to-face consultation if appropriate (e.g.
teleprescription of a vitamin D derivate for psoriasis treatment,
recommendation to continue with penicillin treatment for scarlet
fever). Some were counselled and reassured (e.g. need for excision
of a histiocytoma, need for allergologic evaluation after allergic
reaction).
In the 8 patients for whom it was only possible to narrow
down the diagnoses (17%) and the 7 patients for whom it was
not possible to specify potential diagnoses at all (15%), the
dermatologist did not make a treatment recommendation. These
patients were therefore referred to a dermatologist for a
face-to-face consultation.
Discussion
These results show that digital skin images submitted by patients
are generally of adequate quality and enable specialist doctors to
make a diagnosis and to recommend therapeutic interventions in the
majority of cases. Still, when considering all dimensions of image
quality there is considerable potential for improvement. However,
all patients (or their proxies) were able to photograph their skin
problems and to transmit image packages by PC and internet,
indicating that there are no significant technical obstacles even
in a diverse sample of callers, if persons are willing and
adequately equipped.
Previous studies investigating images submitted by patients
reported very low proportions of photographs for which image
quality was too poor to allow a diagnosis (2-6%) [16, 17]. In our
evaluation this proportion was markedly higher. However, in one
study only 14% of participants were able to handle the mobile phone
camera and to take images of their skin problems so that in fact
the majority of images were taken by professionals [16]. The other
study did not predefine technical requirements and admitted only a
very small file size [17]. As a result, diagnostic concordance was
quite low which, in turn, challenges the accurateness of the
diagnostic procedure. In contrast, our evaluation was performed
during routine operations of the telemedicine centre. In order to
prevent unexpected outcomes, the certainty level applied for
teledermatological diagnosis and subsequent telecare
recommendations was rigorous and conservative, i.e. when in doubt
the image package was judged as insufficient for a diagnosis, and a
physical consultation with a dermatologist was recommended.
Qureshi and colleagues investigated the effects of practical
training or an online tutorial on the quality of skin images
submitted by patients [15]. This study, performed in acne and
rosacea patients from a dermatology clinic, found good quality of
the submitted skin images irrespective of the training method, with
almost all images being in focus and well framed. The fact that our
evaluation, in contrast, exhibits a considerable number of
image packages of suboptimal quality may indicate the positive
impact of training on image quality. We chose a simple and
easy-to-administer approach as opposed to elaborate training, to
minimise obstacles for participation. However, training efforts
should probably be raised in order to obtain higher numbers of
usable images. One possible approach would be to specifically
tailor the instructions to the needs of inexperienced
photographers. Training the telemedical assistants who instruct
callers on the telephone could also result in improved image
quality. This would probably reduce the proportion of image
packages for which it is not possible to make a teledermatological
diagnosis, with potentially favourable effects on telecare rates
and referral rates for face-to-face consultations.
There are several limitations that have to be considered when
interpreting the results of this evaluation. Firstly, the image
quality was assessed by a single dermatologist. The results may
have differed if several outcome assessors had rated the images.
However, the assessment was done by a competent specialist
practicing in an independent dermatological office. Secondly, as
the evaluation focussed on intermediate outcomes such as the image
quality and the feasibility of making a diagnosis and recommending
therapy, we did not validate the diagnoses. Therefore, we can not
make conclusions regarding the diagnostic accuracy and the clinical
effectiveness of therapeutic interventions initialised by such
teleconsultations. However, previous research has proven that SFT
as a general method of health care delivery leads to similar
clinical outcomes as face-to-face dermatology [2-12]. Thirdly, due
to feasibility constraints, this analysis included a relatively
small sample size. Data was collected during a limited time
interval in a single centre in Switzerland. This may limit the
generalisability of the results to other geographic regions or
health care systems. However, a large proportion of persons from
all over the four culturally different regions in Switzerland had
access to services of the Swiss Centre for Telemedicine. In
contrast to previous studies performed in trial settings with
carefully selected patients, this evaluation investigated running
operations without selection of study participants based on
predefined clinical criteria. Therefore, the patients participating
in this study represent a broad sample of callers from newborns to
elderly persons, with a broad range of clinical questions from
general practice to specialist care problems. This may increase the
generalisability of our results to routine settings in
non-specialised telemedical centres.
There are several starting points for future research.
Longitudinal large scale studies involving several dermatologists
are needed to evaluate safety and efficacy of teledermatological
procedures. Additionally, the effects on overall health service
use, hospitalisation rates, and overall health status have to be
investigated further.
Our results have implications for teledermatological practice.
Although there is potential for improvement in the rate of overall
good quality image packages, SFT using digital skin images
submitted by patients provides a promising tool to reduce waiting
times for dermatological face-to-face consultations. Because of its
independence from real-time interactions between patients and
health professionals and the low cost of the required
infrastructure, SFT may help to optimise the use of the available
health resources and foster reasonable access to health
professionals and services [3]. The underlying rationale is that
patients presenting uncomplicated problems may be managed by
telecare, and it is only necessary for patients to use face-to-face
consultations or other treatment modalities in more complex
situations, or when additional examinations are needed. More
specifically, SFT may play an important role in managed care
models. The time is ripe to develop and evaluate dermatological
teletriage and telecare models using SFT and patient-provided
images.
Conclusion
The results of this evaluation increase trust in the use of skin
images submitted by patients. This approach extends the potential
of teledermatology by facilitating dermatological telecare
procedures.
Acknowledgements
This study was funded by the Swiss Centre for Telemedicine Medgate.
We thank Samantha Foulger for her support in data collection.
Conflicts of interests: TR is working as an employee of
Medgate. EB is working as a employee of the Swiss Centre for
Telemedicine Medgate and main investigator of this study. BFC is
working as consultant for dermatology for Swiss Centre of
Telemedicine Medgate. JVO is working as an employee of the Swiss
centre for Telemedicine Medgate.
References
1 Krupinski E, Burdick A, Pak H, et al.
American Telemedicine Association's Practice Guidelines for
Teledermatology. Telemed J E Health 2008; 14: 289-302.
2 Burg G, Soyer HP, Chimenti S. Teledermatology. Eur J Dermatol
2009; 19: 656-8
3 Deshpande A, Khoja S, Lorca J, et al. Telehealth:
systematic review of analytic studies and environmental scan of
relevant initiatives. Ottawa: Canadian Agency for Drugs and
Technologies in Health (CADTH), 2008.
4 Agency for Healthcare Research and Quality. Telemedicine for
the Medicare population. Rockville, MD: Agency for Healthcare
Research and Quality (AHRQ), 2001.
5 Andalusian Agency for Health Technology Assessment.
Teledermatology. Systematic review and economic assessment.
Seville: Andalusian Agency for Health Technology Assessment. AETSA,
2008.
6 Kvedar JC, Edwards RA, Menn ER, et al. The
substitution of digital images for dermatologic physical
examination. Arch Dermatol 1997; 133: 161-7.
7 Lyon CC, Harrison PV. A portable digital imaging system
in dermatology: diagnostic and educational applications. J Telemed
Telecare 1997; 3 Suppl 81-3.
8 High WA, Houston MS, Calobrisi SD,
Drage LA, McEvoy MT. Assessment of the accuracy of
low-cost store-and-forward teledermatology consultation. J Am Acad
Dermatol 2000; 42: 776-83.
9 Du Moulin MF, Bullens-Goessens YI, Henquet CJ,
et al. The reliability of diagnosis using store-and-forward
teledermatology. J Telemed Telecare 2003; 9: 249-52.
10 Pak H, Triplett CA, Lindquist JH,
Grambow SC, Whited JD. Store-and-forward teledermatology
results in similar clinical outcomes to conventional clinic-based
care. J Telemed Telecare 2007; 13: 26-30.
11 Klaz I, Wohl Y, Nathansohn N, et al.
Teledermatology: quality assessment by user satisfaction and
clinical efficiency. Isr Med Assoc J 2005; 7: 487-90.
12 Pak HS, Datta SK, Triplett CA,
Lindquist JH, Grambow SC, Whited JD. Cost
minimization analysis of a store-and-forward teledermatology
consult system. Telemed J E Health 2009; 15: 160-5.
13 Krupinski EA, LeSueur B, Ellsworth L,
et al. Diagnostic accuracy and image quality using a digital
camera for teledermatology. Telemed J 1999; 5: 257-63.
14 Moreno-Ramirez D, Ferrandiz L, Nieto-Garcia A,
et al. Store-and-forward teledermatology in skin cancer
triage: experience and evaluation of 2009 teleconsultations. Arch
Dermatol 2007; 143: 479-84.
15 Qureshi AA, Brandling-Bennett HA, Giberti S,
McClure D, Halpern EF, Kvedar JC. Evaluation of
digital skin images submitted by patients who received practical
training or an online tutorial. J Telemed Telecare 2006; 12:
79-82.
16 Ebner C, Wurm EM, Binder B, et al. Mobile
teledermatology: a feasibility study of 58 subjects using
mobile phones. J Telemed Telecare 2008; 14: 2-7.
17 Eminovic N, Witkamp L, Ravelli AC, et al.
Potential effect of patient-assisted teledermatology on outpatient
referral rates. J Telemed Telecare 2003; 9: 321-7.
|
Version imprimable
Version PDF
