Soliciting and Responding to Patients' Questions about Diabetes Through Online Sources

Introduction

Type 2 diabetes mellitus (T2DM) is a complex chronic disease that requires effective lifestyle management of diet and exercise as well as medication management with the goal of achieving glycemic control. Effective management is critical to reducing morbidity and mortality associated with microvascular and macrovascular complications. ¹ As patients with T2DM are increasingly encouraged to be actively involved in managing their condition and to participate in medical decisions, access to health information becomes increasingly important. ²

Soliciting and responding to questions about diabetes engages patients in their own care. ¹ It provides pertinent information when there is a specific need at a specific time. Although health professional advice and information is important to supporting diabetes self-management, patients do not always feel comfortable in asking questions during consultations, or they find that information given is not always relevant to their needs concerning day to day issues of living with diabetes. ³

Time constraints during consultations leave little time available for responding to patients if questions are raised or for providing them with information about their health condition. ⁴ Even when questions are answered in a consultation, studies show that patients forget 40%–80% of the medical information they are told, and that nearly half of the information they retain is incorrect or poorly understood. ^{4 –6}

The motives for seeking information among individuals with T2DM vary to include necessity; for reference; to understand body changes; to understand clinical and biomedical aspects of disease and research; and to seek out health service entitlements and service quality. ⁷ While information sources can vary among patients, online health information seeking is increasing. The motivations of online health information seekers include the need for reassurance, obtaining a second opinion to challenge other information, to supplement existing information for better understanding, and perceived barriers to accessing information from conventional sources, including health professionals. ⁸

With specific reference to health professionals or services, patient dissatisfaction with information received or the need to access information without having to wait for scheduled medical appointments has resulted in an increasing use of the internet as a source of health advice and information. ⁹

Diabetes-specific internet sites are reported by patients as a principal source of internet information about diabetes and its management. ¹⁰ Although patients want these sources to be credible, they may not be able to determine the credibility of the information sourced. ¹¹ Furthermore, simple internet searching by patients about their diabetes can be problematic because answers may not be accurate, complete, or relevant to the individual's circumstances. ¹²

Previous research has identified information needs by examining “frequently-asked questions” of patients generated during group meetings within the context of a randomized controlled trial. ³ These questions were considered potentially useful to enhancing peer support in a group context and to enabling patients seek answers from professionals as experts. ³ However, it is not known to what extent patient questions about diabetes are answered adequately through internet sources.

There are notable examples of comprehensive online patient information resources for diabetes. For example, in the United States, there is the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK), in particular the Diabetes Health Topics site (diabetes.niddk.nih.gov/dm/a-z.asp). In the United Kingdom, sites such as the Diabetes.co.uk and Diabetes.org.uk serve people living with or concerned about diabetes. In Ireland, Diabetes Ireland (diabetes.ie) is a national not-for-profit organization providing support and education to those affected by diabetes.

The quality of online information on T2DM has been questioned with accuracy of information seen as a major domain of quality. ¹³ The challenges of distinguishing between diabetes websites in providing accurate information and that are most relevant to the topic prompted the development of a framework for ranking quality of automated information websites by researchers in the United States with reference to content, relevance, and “evidence-based medicine.” ¹⁴

It is proposed that this framework be provided as a service model software for diabetes-related web page searches. A criterion for judging high-quality websites is that they are written in a formal language and in a style of professional writing. ¹⁴ However, health information that is formal and professional in language may be limited in terms of health literacy because plain language is recommended to facilitate understanding of complex health information. ¹⁵

It is not known how readily patients can find answers at such trusted and vetted sources of online information. Healthy People 2020 reports that only 37% of online health-information seekers in the United States easily find what they need, and the goal for 2020 is a modest 10% increase. ¹⁶ Evidence from other health problems such as mental health ¹⁷ has shown that although many answers are undoubtedly available in current online resources, fewer are readily accessible, particularly when the patient has a specific concern as opposed to a general need for information.

To explore this issue for people with T2DM, researchers from the United States, the Republic of Ireland, and Northern Ireland (UK) collected questions at a diabetes clinic and submitted a sample of them to search engines within trusted and vetted websites across three countries to assess the availability and completeness of answers.

Methods

During February to April, 2014, patients attending the diabetes outpatient clinic at the Ulster Hospital in Northern Ireland were asked to respond to an open-ended paper questionnaire. The question was printed on a blank page, which asked the following: What are the one or two most pressing questions about your diabetes that you would like answered?

Following completion of data collection, we identified a sample of questions representative of patient concerns about causes, signs and symptoms, sequelae, psychosocial matters, medication, and symptom management. Ten questions representing these different types of patient concerns were selected (Table 1). We then submitted these questions online (October 24–30, 2014) to the following organizations: the US NIDDK, the Ireland charity Diabetes Ireland, and the UK charity Diabetes.co.uk.

We assessed the information provided in the three top-ranked documents returned at each website. The documents emerged by conducting an electronic search of each website using the search engine available at the website. This search engine is accessed through the “search box” that is typically presented in the top right hand corner of the site's Home Page. Many sites (e.g., Diabetes.co.uk) use Google search within their website. The order of the documents (or “pages”) returned by a Google search is determined by a proprietary algorithm that ranks the search-results pages in terms of their priority, which is commonly interpreted to mean their importance or relevance to what was entered at the search box. Other search engines used at other sites (e.g., Diabetes.ie and Niddk.nih.gov) operate on the same principles.

Each patient question was assessed independently by two researchers as follows:

(1) Is an answer there? Yes/No

The process of question submission was repeated during June 1–6, 2016 to the same three online websites. Again, each patient question was assessed independently by two researchers using the above questions (1) to (3) on the information provided in the two top-ranked documents returned. On this occasion, we limited our assessment to the two top-ranked documents because our earlier assessment showed no additional benefits in identifying adequate answers in the third-ranked document returned. Furthermore, the number of pages searched by information seekers has been found to rapidly decline from the second and following pages returned in website searches. ¹⁸

Results

One hundred and sixty-four questions were collected from 120 patients at 12 outpatient clinics. Typical attendance at each clinic was 23 patients, with an average of 16 questionnaires returned at each clinic. Most of the questions were about diabetes (n = 155), whereas a smaller number related to clinic operation (n = 9). Most of the patients attending the clinic had T2DM (95%). We eliminated questions about clinic operations and type 1 diabetes, following which 152 questions about T2DM remained. Ten questions selected as representative of patients' concerns (Table 1) were submitted to the trusted and vetted websites across the three countries.

In the first assessment in 2014, we found that there was no instance, in which an adequate answer was found in the three top-ranked documents returned, as judged by two independent researchers. A synopsis of the search results is given in Table 1. The amount of information was overpowering in terms of volume for some questions, notably absent for other questions, and more often than not completely out of context. In five instances, no document was returned. For 66 of the 75 documents that were returned and examined, relevant information was either too general to provide an answer or too specific, restricted to some limited aspect of the question.

For question 4 (How is my blood sugar controlled? Do I have to do anything more?), the Ireland and UK sites provided an answer to some extent, but the information was very general. For question 6 (Why is it getting harder for me to control my blood sugars as I get older?), there was information on blood sugar control at the US site, but it was difficult to find and did not refer to age. The UK site contained a reference to age and control, but it was difficult to find among all the information provided, much of it irrelevant to the specific question submitted. In some instances, spurious matches to specific words produced irrelevant results. For question 8 (Is there a lifespan to having diabetes?), “lifespan” produced NIDDK committee reports.

For question 7 (How on earth I ever got diabetes in the first place. Never overweight, blood pressure always fine, never eat sweet food.), the words “blood pressure” (Ireland site) and “weight” (UK site) produced documents about those topics, but the question was about “never” having weight, blood pressure, or sweet-food problems. For question 9 (How to remain more positive about my condition. I get fed up constantly finger pricking etc.—then I do realize how lucky I am to have all the help I need!), the Ireland site referred to somewhat hopeful future technological progress, but the relevant information was buried in much irrelevant material.

In the second submission of the questions in 2016, out of 60 documents returned, there was one that provided an adequate answer to one of the questions. For question 3 (How do I stop the need for sugar/chocolate?), the top-ranked UK document was an article specifically on sugar cravings titled “Which dietary techniques can (and cannot) reduce your sugar cravings?” A relevant but very limited response was provided by the top-ranked US document. A long article on food portions contained the following single item of relevant advice: “Try drinking water or herbal tea without sugar or eating a low-fat snack such as an apple if a craving hits you.”

Conclusion

Ninety documents (10 questions, 3 sites, 3 top results) failed to provide answers to 10 patient questions in our 2014 assessment. In our 2016 assessment, 60 documents (10 questions, 3 sites, 2 top results) provided one document that answered one of the questions and one additional document that answered the question, in part. We conclude that for patients who have questions on T2DM, the search engine capabilities of available trusted and vetted online patient information resources very seldom provide useful answers.

Despite the increasing interest in the use of technology to empower patients in understanding their own health, our findings that are available online resources are unable to provide appropriate answers to questions from people with diabetes show that the current situation is even more limited than that previously shown for questions on mental health. ¹⁷ An adequate answer could be readily found for only 26% of the mental health questions submitted to a trusted and vetted online source. Our results on diabetes warrant further investigation of the questions patients have about T2DM and ways to provide answers rather than general or out-of-context information online.