To analyze the inquiries sent to an online ask-the-rheumatologist service in order to identify the users' needs and requirements.
To analyze the inquiries sent to an online ask-the-rheumatologist service in order to identify the users' needs and requirements.
The official web site of the German Competence Network Rheumatology (www.rheumanet.org) provided expert information for patients, relatives, and physicians. We analyzed the content of 1,133 inquiries posted over 5 years and the experts' answers were blinded for analyses.
Patients (60.0%), relatives (24.3%), and physicians (15.7%) addressed the experts. Inquiries were predominantly sent by women (62.2%). Distinct rheumatic diseases were mentioned in 40.5% of the inquiries, and 16.3% reported musculoskeletal symptoms without a definite diagnosis. The number of questions ranged from 1–7 per inquiry (mean ± SD 1.58 ± 0.9). Of the inquiries, 33.2% contained personal histories, 24.9% searched for a rheumatologist nearby, and 11.6% asked for a “second opinion.” The questions covered a wide range of interests, including medication (30.8%), diagnosis-related issues (15.7%), laboratory tests (6.9%), (treatment) guidelines (6.2%), sexual and reproductive health issues (4.1%), and clinical trials (3.4%). In more than 50% of the inquiries, the information requested from the experts was already at least partly published on the web site. The experts' answers covered the users' questions completely in 91.8%, partly in 6.1%, and not at all in 2.1%.
A standardized medical web site providing tailored and trustworthy information for all user groups gains from an ask-the-expert service. Only such an interactive online application is able to satisfy users' actual demands: searching for specific individualized information on the internet. Therefore, an ask-the-expert service contributes to optimized patient care.
Online health information and communication services are increasingly used by patients, relatives, and physicians (1–5). Besides general health portals and e-communication with personal health care providers, web-based consultations in expert forums (ask-the-doctor services) have become increasingly important sources of information (1, 2, 5–8).
Various reasons for the use of web-based expert services have been identified (1, 2, 5, 9–11). However, little is known about the users of an ask-the-doctor service, and about the communication between the users and the experts in rheumatology. The official web site of the German Competence Network Rheumatology was launched on www.rheumanet.org in 2000 and had ∼17,000 users per month in 2006 (12). Due to an administrative reconstruction of the network, it was transferred to www.dgrh.de in 2007. The web site offers a large variety of standardized and trustworthy information for rheumatology patients, relatives, and physicians (12). It includes trustworthy information on rheumatic diseases, treatment (options), recommendations for disease monitoring, the ability to search for a local rheumatologist, a tool that provides a risk assessment for inflammatory rheumatic diseases, monthly chats between users and experts, and various other topics and links.
In 2001 a free of charge web-based ask-the-doctor service was implemented. To identify the users' needs and requirements that were addressed besides or in addition to the information already provided on our web site, we decided to analyze the contents of the inquiries sent to our expert service. We report the contents of the inquiries and the quality of the experts' answers for a 5-year period, starting August 2001.
A typical web form offering the possibility for inquiries (placed in a blank limitless cell) was connected to an underlying structured query language database that saved all the inquiries and the experts' replies. Before submission, users categorized themselves as patients, relatives, or physicians and then determined the general topic of their request from a predetermined list (medical, psychological, or technical) on the web form. The inquirers could choose to remain anonymous, but needed to leave an e-mail address and their informed consent to receive an expert reply.
Eight experts (6 rheumatologists, 1 psychologist, and 1 computer scientist) randomly answered the incoming inquiries according to the demanded subjects and were instructed to do so to their best knowledge. The computer scientist provided information only when inquiries contained technical questions (e.g., broken links, questions to the rheumatology documentation software provided on the web site). The psychologist and the medical experts were advised to be cautious about elements that could jeopardize the relationship between the patients and/or their relatives and the physicians caring for the patients. Answers were given according to the Health on the Net conventions (www.hon.ch/HONcode/German/) and suggested guidelines (13, 14). A standardized text informed the inquirers that the e-communication cannot substitute for a personal appointment. No reimbursement was provided to the experts.
After a personalized login, the site with the inquiries and the responses was open only to all experts of the service. Inquirers received the individualized expert's answer through an e-mail that included the original inquiry. Followup questions were possible by either sending an e-mail that could be directed to the expert who had answered, or by submitting a new inquiry. Followup questions sent directly to the experts were not available for analyses. An open frequently asked questions (FAQ) site was built on the questions of the incoming inquiries and the experts' answers, but it never displayed the original inquiries or the inquirers' names.
A coding system for the inquiries and the replies was developed using a content analysis approach. Categories were not fully exclusive of one another. However, each inquiry was coded for its dominant theme and then placed in the category that seemed to best capture the essence of the inquiry. The length of the inquiries and the replies were recorded, as well as the time lag of the experts' responses. Inquiries were coded across domains as message type, number of requests per inquiry, and demographic data.
Reported rheumatic diseases were categorized according to the standard documentation set of the national database of the German Rheumatism Research Center in Berlin if reported from physicians (15). If the diagnoses were reported by lay persons and were not clearly allocatable to the given standard documentation set, the authors reassessed these diagnoses and classified them in a consensus decision. Questions regarding medical treatment were subcategorized with respect to the active components (e.g., disease-modifying antirheumatic drugs [DMARDs], biologic agents, nonsteroidal antirheumatic drugs [NSAIDs]), their indications, effects, and side effects. Further categories included orthopedic surgery, medical diagnostics (e.g., laboratory tests, computed tomography, and magnetic resonance imaging scans), sexual and reproductive health issues, clinical trials, physiotherapy, rehabilitation programs, nutrition and diet, and miscellaneous others. We also verified whether the inquirer might have found the information on the web site without using the ask-the-expert service.
The experts' answers were blinded for the analyses and were classified as “completely” answering the users' inquiry if all items of the question(s) were covered. Answers were classified as “partly” answering the inquiry if one or more item(s) were missing, and as “not at all” if misconceptions of the question or missing answers occurred. Independently from each other, 2 physicians of the ask-the-expert service team retrospectively analyzed the inquiries stored in the database over 5 years.
We also intended to evaluate the users' perceptions of our ask-the-expert service. Therefore, in 2006 we sent out an e-mail enclosing a link to an online questionnaire and a text file containing the same questions. The number of responses (n = 47) did not seem representative to us and therefore they were not considered for further analyses. Ethical approval had been received from the local ethic committee.
Statistical evaluation was made using the Statistical Package for the Social Sciences, version 15.0 (SPSS). Values are expressed as percentages for discrete variables, or as mean ± SD, range, or median for continuous variables. Nonparametric tests (chi-square test or Mann-Whitney U test, as appropriate) were performed as 2-tailed tests. P values less than 0.05 were considered significant. Interrater reliability was assessed using Cohen's kappa.
Between August 2001 and August 2006 we received 1,298 inquiries. Of those, 165 inquiries were excluded from further analyses because they included no or fragmentary content (n = 144), were sent repeatedly (n = 4), or included essay-type personal reports without any question (n = 17). Therefore, a total of 1,133 e-mails entered the final analyses. Fifty-eight (5.1%) individuals sent more than one inquiry (range 2–8 inquiries). Cohen's kappa coefficients reflecting the interrater reliability of the inquiries' classification ranged from 0.88–0.97.
The inquiries were sent from Germany (93.9%), Austria (2.9%), Switzerland (1.5%), and other countries by patients (60.0%), relatives (24.3%), and physicians (15.7%). Age was reported by 113 patients; mean ± SD age was 37.8 ± 12.6 years (range 17–72 years). When inquiries were sent on behalf of someone else, the patient's age was reported in 85 inquiries (mean ± SD age 40.8 ± 26.0 years, range 2–96 years). Questions regarding children were asked by 22 (1.9%) inquirers. Information on sex was available in 1,019 inquiries; users were predominantly (62.2%) women.
The number of distinct questions asked of the expert panel ranged from 1–7 per inquiry with a mean ± SD question of 1.58 ± 0.9 per inquiry. A total of 62.4% of the users sent 1 question per inquiry, 23.5% sent 2, and 9.9% sent 3 questions per inquiry. Inquiries had a mean length of 525.1 characters (range 19–5,625 characters, median 366.0 characters); the corresponding experts' replies had a mean length of 600.6 characters (range 11–4,671 characters, median 504.5 characters). A total of 13.7% of the inquiries were answered within the same day, 56.7% within 2 days, and 73.9% within 4 days. The median time lag between inquiry and response was calculated to 2 days.
Very detailed and personal histories with only a few questions embedded requiring very specific and individual answers were given in 33.2% of the inquiries, 131 (11.6%) users explicitly asked for a “second opinion,” and 282 (24.9%) were looking for a rheumatologist nearby.
Distinct rheumatic diseases (predominantly rheumatoid arthritis [RA; 33.1%], connective tissue diseases [17.9%], vasculitides [13.1%], spondylarthritides [10.0%], degenerative diseases [5.9%], psoriatic arthritis [PsA; 5.9%], and fibromyalgia [3.1%]) were mentioned in 459 (40.5%) inquiries. Median disease duration of established diseases was 35 months (range 1–528 months). Diagnosis-related questions covered 15.7% of the inquiries. Musculoskeletal symptoms without a definite diagnosis were reported by 185 (16.3%) users.
Inquiries about drugs or drug monitoring (indications, effects, and side effects) totaled 30.8% (n = 349). Two hundred eighteen (19.2%) requests referred to DMARDs and biologic agents, and 131 (11.6%) referred to NSAIDs, coxibs, corticosteroids, antibiotics, or other medications (Tables 1, 2, and 3).
|Substances requested||Complete group (n = 218)||Patients (n = 138)||Relatives (n = 51)||Physicians (n = 29)|
|Immunosuppressive substances||147 (67.4)||98 (71.0)||34 (66.7)||15 (51.7)|
|Azathioprine||12 (5.5)||8 (5.8)||2 (3.9)||2 (6.9)|
|Chloroquine||17 (7.8)||13 (9.4)||2 (3.9)||2 (6.9)|
|Cyclosporin||1 (0.5)||1 (0.7)||0||0|
|Cyclophosphamide||3 (1.4)||3 (2.2)||0||0|
|Gold||5 (2.3)||1 (0.7)||2 (3.9)||2 (6.9)|
|Leflunomide||11 (5.0)||9 (6.5)||1 (2.0)||1 (3.5)|
|Methotrexate||60 (27.5)||41 (29.7)||14 (27.5)||5 (17.2)|
|Mycophenolate mofetil||4 (1.8)||3 (2.2)||0||1 (3.5)|
|Penicillamine||1 (0.5)||0||0||1 (3.5)|
|Sulfasalazine||11 (5.0)||10 (7.2)||1 (2.0)||0|
|No specific substance mentioned||22 (10.1)||9 (6.5)||12 (23.5)||1 (3.5)|
|Biologic agents||71 (32.6)||40 (29.0)||17 (33.3)||14 (48.3)|
|Anakinra||6 (2.8)||2 (1.5)||3 (5.9)||1 (3.5)|
|Adalimumab||6 (2.8)||2 (1.5)||2 (3.9)||2 (6.9)|
|Etanercept||16 (7.3)||9 (6.5)||2 (3.9)||5 (17.2)|
|Infliximab||20 (9.2)||14 (10.1)||3 (5.9)||3 (10.3)|
|General TNFα-blocking agents, no specific substance mentioned||23 (10.6)||13 (9.4)||7 (13.7)||3 (10.3)|
|Concomitant medication inquiries||Complete group (n = 131)||Patients (n = 91)||Relatives (n = 29)||Physicians (n = 11)|
|Nonsteroidal antiinflammatory drugs||22 (16.8)||17 (18.7)||3 (10.3)||2 (18.2)|
|Coxibs||8 (6.1)||7 (7.7)||1 (3.5)||0|
|Corticosteroids||61 (46.6)||38 (41.8)||18 (62.1)||5 (45.5)|
|Bisphosphonates, calcium, and vitamin D||0||0||0||0|
|Antibiotics||11 (8.4)||7 (7.7)||2 (6.9)||2 (18.2)|
|Other additional medications||29 (22.1)||22 (24.2)||5 (17.2)||2 (18.2)|
|Topic||Complete group (n = 862)||Patients (n = 478)||Relatives (n = 224)||Physicians (n = 160)|
|Competence centers||79 (9.1)||11 (2.3)||7 (3.1)||61 (38.1)|
|Contact to rheumatologists||282 (32.7)||166 (34.7)||95 (42.4)||21 (13.1)|
|Radiographs/CT/MRI scans||12 (1.4)||8 (1.7)||0||4 (2.5)|
|Nutrition/diet suggestions||23 (2.7)||18 (3.8)||5 (2.3)||0|
|Orthopedic surgery||22 (2.6)||19 (4.0)||2 (0.9)||1 (0.6)|
|Osteoporosis||6 (0.7)||4 (0.8)||1 (0.4)||1 (0.6)|
|Prevention||14 (1.6)||10 (2.1)||4 (1.8)||0|
|Psychosomatic/logical problems||25 (2.9)||20 (4.2)||4 (1.8)||1 (0.6)|
|Rehabilitation||15 (1.7)||9 (1.9)||6 (2.7)||0|
|Self-help groups||25 (2.9)||17 (3.6)||7 (3.1)||1 (0.6)|
|Physiotherapy||43 (5.0)||22 (4.6)||17 (7.6)||4 (2.5)|
|Medication recommendations 1†||93 (10.8)||60 (12.5)||31 (13.8)||2 (1.3)|
|Medication recommendations 2‡||46 (5.3)||33 (6.9)||11 (4.9)||2 (1.3)|
|Treatment guidelines||70 (8.1)||37 (7.7)||15 (6.7)||18 (11.3)|
|Web site||79 (9.2)||23 (4.8)||14 (6.3)||42 (26.3)|
|Coverage of medication costs by health insurances (DMARDs n = 5, biologics n = 3)||19 (2.2)||14 (2.9)||4 (1.8)||1 (0.6)|
|Vaccinations (tetanus, hepatitis, influenza)||9 (1.0)||7 (1.5)||1 (0.4)||1 (0.6)|
Ongoing clinical trials offering new treatment options were addressed by 38 users (18 patients, 7 relatives, and 13 physicians) referring to the following diseases: n = 5 (13.2%) for RA, n = 2 (5.3%) for PsA, n = 2 (5.3%) for spondylarthritides, n = 3 (7.9%) for connective tissue diseases, n = 2 (5.3%) for vasculitides, and 1 patient with osteoarthritis. Nineteen (50.0%) patients did not mention a specific diagnosis or experienced unspecific musculoskeletal symptoms (n = 4).
Forty-seven (4.1%) inquiries referred to fertility and pregnancy issues, whereas worries about the intake of oral contraceptives and hormone replacement therapies each were verbalized by 4 (0.4%) users. The underlying diseases concerning sexual and reproductive health issues distributed as follows: RA = 14, connective tissue disorders = 10, PsA = 3, vasculitides = 2, and others = 7. Nine inquiries did not mention a diagnosis, and 10 had not yet been diagnosed as having a rheumatic disease.
Questions concerning laboratory tests (n = 78, 6.9%) were mainly asked by patients and relatives (92.3%). Most of them addressed the association or diagnostic value of a specific laboratory test to a certain rheumatic disease when detected recently (n = 48) or during followup (n = 16). Inquiries about recommended and necessary blood tests under immunosuppressive medications were brought up by 14 (1.2%) users. Table 3 lists the additionally addressed topics and matching frequencies in further detail.
The information demanded from the experts was already published on the web site in 30.5% (n = 346) of the inquiries, and in a further 27.9% (n = 316) inquiries the information could have been retrieved at least partly from the web site. In contrast, 8.4% (n = 95) of the information asked for was not available on the web site and was predominantly requested by patients (n = 51), followed by physicians (n = 27), and relatives (n = 17) (Table 4). In addition, 376 (33.2%) inquiries contained personal histories that necessitated individual experts' answers.
|Topics not provided||No. (%)|
|Known rheumatic diseases||10 (10.5)|
|Laboratory tests||6 (6.3)|
|Radiographs/CT/MRI scans||2 (2.1)|
|Orthopedic surgery||6 (6.3)|
|Synoviorthesis, yttrium||2 (2.1)|
|Biologic agents||2 (2.1)|
|Other medication†||8 (8.4)|
|Competence Network Rheumatology||9 (9.5)|
|Coverage of medication costs by health insurance||5 (5.3)|
|Psychosomatic/-logical problems||6 (6.3)|
|Self-help groups||2 (2.1)|
|Technical support web site||6 (6.3)|
The experts' replies covered the users' questions completely in 91.8% (n = 975) of the inquiries, partly in 6.1% (n = 65), and not at all in 2.1% (n = 22). This relation was consistent for all participating experts (P = 0.862 by chi-square test). Neither the number of distinct questions nor the character length of the completely and partly answered inquiries differed significantly from those that had been categorized as answered not at all (P = 0.578 and P = 0.834, both by Mann-Whitney U test).
The experts left 71 inquiries unanswered. Twenty-four concerned DMARDs (n = 17) and biologic agents (n = 7). Others focused on corticosteroids (n = 4), NSAIDs (n = 2), coxibs (n = 2), antibiotics (n = 2), and various other drugs (n = 4). The study investigators were unable to assess the underlying reason(s) for leaving the inquiries unanswered. All of them contained “easy to answer” contents and had comparable character length to those that were answered.
We implemented a web site for the noncommercial Competence Network Rheumatology including an ask-the-expert service (12). This special service is used by ∼20 of 17,000 web site users per month. To our knowledge, this is the first study analyzing the impartial and unsolicited inquiries sent to such an ask-the-expert service in rheumatology. The analysis revealed important lessons for us that have already improved the quality of the web site, but the implications seem to be useful for the construction and design of specialized medical web sites in general.
Along with patients, relatives and physicians are also important clients. While differences in the topics addressed between patients and physicians are not surprising, it is noteworthy that relatives address different issues than patients (Tables 1, 2, and 3). Therefore, group-specific interests have to be taken into account, and the provided information content must be consistent throughout groups (6, 9). The relatively high proportion of questions asked of the online expert panel by proxies also raise ethical, privacy, patient security, and legal concerns not occurring to the same extent in face-to-face health care, which have to be considered and solved like those discussed for e-communication in general (9, 11, 16, 19, 20).
Similar to published data from other medical expert services and e-mail consultations, many inquirers asked common questions on (rheumatic) symptoms/diseases, diagnostic procedures, and medication/treatment (1, 2, 5, 17, 21, 22). These standard issues are either not completely met in regular health care or they arise after face-to-face appointments (2, 9, 17, 19, 22). Since patients and/or relatives obviously use more than one source of information, physicians should explicitly address and critically comment on the retrieved (online) information during face- to-face visits in order to avoid confusion and/or misconception.
Interestingly, more than 50% of the requested information was, in principle, available on our web site, but was either not identifiable or did not meet the users' expectations. Suitable prerequisites for optimization are a clearly structured medical web site combined with an intelligent user guidance that allows targeted navigation through well-tailored online information (5, 8, 23–25). Additionally, the web site should include an easy-to-use search function that accounts for terms searched by laymen, as well as professionals, and a user-friendly FAQ section. As part of the quality assurance process, some of these recommendations had been implemented when the Competence Network web site was transferred to www.dgrh.de.
Moreover, the high proportion of users with undiagnosed symptoms (comparable to data published by others) who try to sort out their condition prior or in addition to their face-to-face appointments underlined the need for an online instrument that provides a risk assessment for inflammatory rheumatic diseases (6, 10, 17, 26). Our “RheumaCheck,” a frequently used online patient questionnaire, now provides this kind of information on our web site (27). In combination with a search function for a local rheumatologist and an ask-the-expert service, it might reduce inappropriate expert consultations and the congestion of the health system, as well as enhance early detection and referral to rheumatologists, and thereby contribute to improved quality of medical care (7, 11, 27–29).
Our findings reconfirm that an ask-the-expert service is inevitably necessary in order to meet rare but relevant users' requirements: it is used as an auditorium for complex case histories and, as reported by others, as a “new” opportunity to gain a qualified second expert opinion (2, 6, 9, 16). Therefore, the expert service as an example of an interactive online application might support quality of care, although internet-based consultations usually act as a complement and cannot replace a personal visit (6, 7, 17, 22). Nevertheless, the web site provider has to decide whether he is interested in giving this kind of information when adding an ask-the-expert service to the array of e-communication options (30). In addition to the published advantages, disadvantages, and barriers of e-communication, Figure 1 summarizes the pros and cons of an ask-the-expert service (6, 8, 19, 31–33).
Since inquiries covered a wide spectrum of topics and often contained very specific and sometimes (possibly due to the relative online anonymity) even sensitive issues, an expert panel should be adequately staffed and integrate appropriate professions and disciplines (6, 8, 16, 21, 34).
From an economic point of view, having inquiries answered by experts consumes valuable resources: reading and understanding a “typical” inquiry and providing the answer demands ∼15 minutes from an expert and thereby costs ∼25€ per inquiry (technical costs not included in this estimate) (21, 22). Therefore, limitation for inquiries needing sophisticated answers seems necessary. The prescreening and filtering of inquiries by well-trained medical staff and/or sorting via detailed web forms (the user fills in the form to pre-describe the content in detail when the inquiry is submitted) might help to prevent experts' overload and save money (1, 2, 8, 22). One might conclude that the benefits may not outweigh the costs since only a minority of the web site visitors used the service. In our scenario, its realization was possible due to the funding by the German Competence Network Rheumatology and the experts' assignment without reimbursement for their work. Brockes et al reported that the introduction of a service fee resulted in a decrease in requests, while at the same time a greater number of older people used the service and asked more complex questions (18). The provision of tax-free services seems more reasonable since this might ensure inquiries from all ages and socioeconomic groups. In addition, financing options (e.g., by health insurances, grants) and the cost-effectiveness of the services should be further evaluated, and provision might be reserved to community-based, well-funded, and/or commercial web sites (6, 22).
Our study has several limitations. It is hampered by the fact that we do not know whether the identified topics reflect the concerns of all users or only a minority of users since only a small subgroup used the expert service. Numerous reasons for not sending inquiries to a web-based expert forum have been published (16). Furthermore, our data may not fully reflect the information needs of patients experiencing musculoskeletal conditions seen in rheumatology practices in different countries across the world. Further studies that address the users' satisfaction with the experts' answers and the service are necessary. Our coding criteria developed in this study are preliminary, and objectivity and generality need further validation.
In conclusion, our results offer valuable insights for the construction and maintenance of a medical web site; main aspects are summarized in Figure 2. The integration of an ask-the-expert service seems warranted given the number of inquiries containing personal histories that cannot be covered by standardized information on a medical web site, and given its function as a valuable feedback tool for the web site provider, thereby supporting the quality assurance process. Implementation of such interactive online applications might contribute to optimized patient care in the future.
All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be submitted for publication. Dr. Richter had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study conception and design. Richter, Becker, Schalis, Koch, Willers, Specker, Monser, Schneider.
Acquisition of data. Richter, Becker, Schalis, Koch, Willers, Specker, Monser, Schneider.
Analysis and interpretation of data. Richter, Becker, Schalis, Koch, Willers, Specker, Monser, Schneider.
We thank all participating ask-the-expert experts for providing answers to the patients' inquiries over the years without being reimbursed for their work.