7 Data model for integrated patient portals
Abstract: The rate at which patients search for health information online continues to grow, as does the type of content available to them. Patients are currently left to pair their personal health information with other information they find online. This work examines three different methods for disseminating filtered information to patients. This chapter illustrates that different modes have benefits and drawbacks in providing content to patients that is of quality and accessible.
7.1 Introduction
The phrase “patient empowerment” is common in healthcare rhetoric today. There are numerous efforts to support the shift in the healthcare paradigm from patients being passive recipients of care to active participants. There are a range of definitions for patient empowerment. This multitude of ideas demonstrates that patient empowerment is a multi-faceted concept in terms of its definition. However, one common theme across many definitions is that empowered patients are informed.[1] Informed patients is thought to be able to make more competent decisions regarding their health, and to feel more confident doing so. One growing source of information for patients is patient portals. Patient portals are applications that provide patients with direct access to medical record content. However, this content is harvested from medical records, content designed for physicians by physicians, with little change. Thus, portals often provide patients with access to content that is not written specifically for the patient, requiring the patient to translate his or her record.
Patients can also consult publicly available health information, such as MedlinePlus, [2] and the Mayo Clinic website,[3] social networking sites centered around health, such as PatientsLikeMe[4] or other health websites. Patients searching for health information continue to increase,[5] and it currently ranks as the third most popular task completed online.[6] However, the quality of sources is not always readily available.[7] Even when a source is of good quality, its content may not be applicable to the user, meaning it does not meet the user’s information needs. Information needs are defined here as “the recognition that their [the patient’s] knowledge is inadequate to satisfy a goal, within the context or situation that they find themselves in a specific point in time”.[8] Searching for health information requires patients to look through a variety of sources and determine what is relevant to their personal health situation, which has been documented as a difficult task for patients.[2], [9]
Much of the difficulty lies in the differences in language used by professionals versus patients. Often quality content contains medical jargon. In order to select appropriate information and apply it to their care, patients must have a level of health literacy. Health literacy is the ability to search for, consume, understand, reflect on and apply health information.[10] Hence, low levels of health literacy are seen as a significant barrier to accessing health information.[11]
The rate of accessing digital health resources by consumers who have less financial security, have less education, are more culturally diverse and older, will increase. [3] It has been documented that exposing patients to quality health information can lead to improved outcomes. Thus, as the population accessing health resources becomes more diverse, the vocabulary gap between healthcare professionals and consumers becomes more influential.[12]
Some of the best informational health contents on the web, professional medical guidelines, are intended for healthcare practitioners. As they attempt to cover the breadth of a diagnosis and treatment, these guidelines document an expanse of symptoms, test, diagnosis, treatments, clinical studies findings and additional information. This expanse of information requires an expert’s knowledge to determine what portions are applicable to an individual patient, meaning a patient may not be able to determine which professional guideline content applies to him or her. Additionally, both medical content created for consumers and for professionals lack personalized content reflective of the individual’s diagnosis and process of care.
Patient information sources have the potential to provide patients with personalized information. Personalized health information for patients is necessary to give consumers supporting information with which to understand their health. Personalized information contrasts with general guidelines and consumer health content in that it is tailored to the individual and based on their information needs and possibly other content of theirs, such as a medical records or profiles. Thus, personalized content should also try to anticipate the patients’ information needs, or information patients would like to see.
This chapter demonstrates three different methods that refine health information content to provide patients with information relevant to their diagnosis and their personal information needs and preferences. For this illustration, the domain of breast cancer will be used. However, these modes of dissemination have been used in other diagnosis domains as well.
7.2 The current situation: needs, gaps and challenges
7.2.1 Patient information needs and preferences
In general, cancer patients are more positive about the idea of electronic health records and online health information than those patients without cancer, and tend to cope better with a diagnosis when they have accurate information regarding their disease and treatment.[13] Patients use health information based on their records to review findings, process information and communicate with their healthcare practitioner. [14] found that cancer patients wanted to be more involved in their care, desiring to have an active role. However, [15] found that as a patient’s condition deteriorated, they tended to want less information, although still wanted to know information that was positive in nature. Yet, [16] documented that patients wanted as much information as possible regarding their diagnosis, regardless of whether the nature of it was positive or negative. Patients specifically wanted information on both diagnosis and treatment. Numerous citations show that patients are interested in diagnosis, treatment, symptoms, prognosis, survival information and side effects of treatments.[14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26]
Health information preferences can also vary by demographics. Younger patients tend to search for more information on their own than older patients.[5] Those with higher levels of education tend to search more than those with less. Women tend to search more than men.[5] In terms of ethnicities, Hispanics have the lowest rates of searching for information at 44%. In comparison, 51% of African Americans search for health information and 65% of whites access digital health information.
While these rates of access are likely to stay stable in the short term, just as access to health information online has increased across demographics, so has it increased within individual groups and is likely to continue to do so.[5] With the benefits in health outcomes and prevention in an educated population now documented,[27] more public and private institutions are attempting to decrease the digital divide and provide information online geared towards a variety of users. This means that content must strive to be beneficial to users across demographic lines and yet anticipate the nuanced variation in needs between groups. As the older, less educated and those of lower socio-economic status are anticipated to increase their rate of using digital health resources, so must those resources make information not simply available but accessible, meaning that they give the user the opportunity to use the content to make health decisions. While no consumer health tool will benefit all patients without complications, digital resources must strive to be as inclusive as possible.
7.2.2 Portrait of the breast cancer patient
The domain of diagnosis and the demographics associated with it can also influence information preferences. For this domain of breast cancer, it is worth noting that the disease primarily affects women, although men can also develop breast cancer. The average age for developing breast cancer is 61. White, non-Hispanic women tend to have the highest incident rate of breast cancer. However, women of various ages and all ethnicities can develop breast cancer.[28] Increases in mammogram screenings have led to a higher rate of women being diagnosed while in situ stage, meaning the cancer was less advanced. Being diagnosed at an earlier stage increases chances of survival.[28]
7.2.3 Information sources
Breast cancer screening patients have numerous sources of traditional supporting information including support groups, informational sessions and brochures. General supporting health information in electronic format is now also widely available.[2] [3] [4], [24] In addition, patients’ personal medical information is becoming available electronically via personal health records (PHRs) and patient portals. While both types of content are readily available, there remains a deficit in information integration in that there are few sources that merge the two types. The Health Insurance Portability and Accountability Act (HIPAA) mandates that patients have access to their medical record. While the content of their medical record is available to patients, the next step is to make it comprehensible to them. To ensure that patients understand the information contained within their medical record, patient portals cannot display the data formatted as it was captured. The content within a portal was originally documented by physicians with the intent that it be used by other medical practitioners. Although patients may want to see some of the same information as a physician, they do not necessarily want it displayed in the same way, nor will they apply it in the same context. In order to make the content understandable to consumers, changes are required in the visualization, vocabulary used and information abstraction level.
Usefulness of quality of a source varies by the user’s expectation.[29], [30] While no tool will be able to anticipate all user needs, a needs assessment of a population is necessary to help estimate the types of information required by a specific population. Types of needs assessments include: literature reviews, surveys, focus groups and interviews.[30]
7.3 Proposed solutions
7.3.1 Tailored patient guidelines
The linkage of supporting content based on medical record data is standard practice for practitioners within numerous institutions. Similar features are now also available for patients via portals and web services (i.e. MedlinePlus Connect). Although these options often use unique concept identifiers provided by the Unified Medical Language System or another ontology, incorrect and inexact matches still occur. Definitions are often context dependent: consider the difference in the definition of diaphragm in a gynecological versus thoracic domain. While a consumer health source may provide correct information on a topic, it does not demonstrate how the topic relates to the patient’s screening process or their health record. Here, an existing information model generating process that was used to link guideline content and patient information needs and preferences for non-small cell lung cancer (NSCLC) screening,[31] was applied to the domain of breast cancer screening. This model was created based on the content that falls within the overlap of two domains: professional guidelines and consumer health information needs. The model content must be contained in both spheres of information in order to be simultaneously reflective of patient information needs and guideline content, with the intent that this model will be used to harvest content from professional guidelines that are reflective of patient needs and can be linked to actual concepts found within the patient’s record.
First a literature review on cancer patient information needs, particularly breast cancer patient information needs, had to be conducted. Then the guidelines from breast cancer screening had to be reviewed. The literature review of patient information needs was used to define the class set, as the primary objective was to create a model relevant to patient information needs. Themes found within the literature review included the desire to hear information on diagnosis, treatment options, symptoms, diagnostic tests and common side effects of treatment.[15] [16] [17], [20] [21] [22] [23] [24] [25] [26], [32] [33] [34] [35]
Professional guidelines were used as the inspiration for candidate concepts that could be used to populate the classes (themes from the literature review). In this information model, the smallest units are concepts; concepts are the normalized instances of terms (e.g. “lump”). To create a list of concepts that would have appropriate contextual explanations for the domain of breast cancer screening, the breast cancer screening guidelines from the National Collaborating Center for Cancer and UpToDate [36], [37] were reviewed. All concepts belong to one or more classes, the class being reflective of the common clinical feature or set of features the concepts share (e.g. “Symptoms”). With this candidate list, the information model was organized, with the constraint that concepts be included only if they are reflective of classes made evident by the literature review on patient information needs. In turn, classes could only be included if they were reflective of the steps in the screening process.
Figure 2: Simplified version of practitioner guidelines for breast cancer screening
The information model for breast cancer screening patient guidelines consists of three classes, as seen in Table 1. The Diagnosis class consists of all the possible diagnoses that can be given to a breast cancer screening patient (Tx–T4). The symptoms class does not contain every possible symptom experienced by a patient undergoing breast cancer screening. Rather, it contains just the most common, as documented in [38]. Analysis of the guidelines documented that not all patients experience symptoms, however the symptoms topic is included here, in the event that a patient who is experiencing symptoms should have access to diagnosis relevant information. The diagnostic test contains all tests that a patient may undergo during screening; this was done in order to assist patients to better understand the screening process. The concepts within this model can now be linked to content from practitioner guidelines and used to annotate patient records. Concepts found within a patient’s record can be used to pull the guideline appropriate to that concept.
Table 1: Model of concepts and classes.
| Class | Number of sources citing information need | Guideline concepts mapped to class |
|---|---|---|
| Diagnosis | 7 | Tx, T0, Tis, T1, T2, T3, T4 |
| Symptoms | 3 | Lump in breast, lump in underarm, thickening or swelling of breast, irritation of breast skin, dimpling of breast skin, redness in nipple area, flaky skin in nipple area, nipple discharge (other than blood), any change in shape of breast, any change in size of breast, pain in any area of breast |
| Diagnostic tests | 4 | Mammogram, ultrasound, MRI, biopsy |
7.3.2 Tailoring content based on user profiles
The BCKOnline portal uses metadata tags to create both metadata for information sources and for patient profiles in the domain of breast cancer,[29, 39] in order to match patients with information. To create these user-centric tags, breast cancer patients and family members were surveyed on their information needs and preferences. Demographic information on participants was also collected. The survey found that information should be tailored by precision based on a user’s circumstance (e.g. stage), the user’s choice in format (e.g. article, short-length) and provision of resource quality (e.g. government institution).
Information on breast cancer patients’ information preferences, their demographics, resource content, resource format and resource quality were then used to create specific tags that were then used to index documents. Patients were able to create a user profile based on their information preferences, demographics and specifics regarding resource content in order to retrieve documents that best fit their search, based on the metadata tags. By prioritizing what types of information they prefer, the BCKOnline portal is able to limit information overload, excluding articles that do not fit the user profile.
Users create their profile by answering the multiple choice questions presented to them when they first log onto the site. The age group question determines what age range the patient falls into, as information preferences have been shown to vary by age. For instance, younger women have been shown to be more likely to want information on fertility issues.[39] The disease stage question steers users towards information appropriate for a specific stage, while the information preference question allows the user to choose content based on a matrix of types of information (plain and brief, plain and detailed, scientific and brief, scientific and detailed). The user type question filters information by the type of user (patient, family member or friend). User testing demonstrated that the portal was well received. Patients rated the portal as easy to use, indicating that it filtered irrelevant information out and provided information patients perceived as relevant.[39]
7.3.3 Information content via social networking
Breast cancer patients who are socially isolated have higher rates of mortality.[40] Alongside providing human contact, social networking sites have been shown to have similar outcomes to other more formal health interventions, discouraging unhealthy behavior and encouraging healthy behavior.[41] Social networking sites, such as PatientsLikeMe, allow for patients to connect with other patients that have the same diagnoses and symptoms.[42] These sites have been shown to have been of particular value to patients with breast cancer.[42] In addition to providing individuals with other individual’s information, these sites also can permit crowd sourcing, so that a user can get a summary of many others’ experiences.
[43] found that breast cancer patients tended to join a social networking site to find information, but stayed as a member even after they had progressed into remission in order to support others. Breast cancer patients who are member of a social networking site for breast cancer tend to be active participants, meaning they both read and publish content.[1] This cycle of information supply among patients makes participants both the source and the recipient of the information. [43] reported that users of the social networking site were able to communicate more freely then they would in face-to-face conversation, having less fear to do so. As members of these sites, patients are allowed to remain as anonymous as they desire, but conversely they can share as much as they want to.
7.4 Discussion
This chapter has shown three modes to produce personalized informational content for patients, with all three modes filtering out information based on an anticipation of what types of information will interest the patient. The method to produce tailored content taken from guidelines is based on [31], and applied here to a second domain of breast cancer screening. This method filters out content, based on the overlap that occurs between documented patient information needs and professional guidelines.
[29], [39] demonstrate how a user profile can be used to filter information within a portal for breast cancer. To better anticipate the user’s need, the content can be filtered by filling out a short survey after logging on. Users are then linked to content based on the tags they create for themselves via the profile being matched to tags assigned to informational content.
Information can also be filtered based on diagnosis in social networking sites such as in [42], [43]. Within these sites information is also filtered by patients both as users and contributors. Using the site to find information, patients can search for other profiles that resemble their own or that contain content they are interested in. As suppliers of information, users filter the type of content available based on what they want to share.
7.5 Conclusion
This chapter has examined three different modes of supplying supporting information to patients who are participating in breast cancer screening or treatment. Social networking sites have been rated highly by participants and allow users a significant amount of individual control over what information they access. Content from these cites is written by other patients, making the language accessible to other users. However, concerns regarding the quality and population relevance (e.g., if it’s true for one, is it true for all) have been raised. Variations in the content’s ability to apply to an individual patient make these sites unsuitable as the sole source of supporting content.
A user profile of criteria to match patients to content permits some preferences from individual users to be incorporated into source selection. However, information preference patterns demonstrated over a population may not hold true for an individual patient. E.g., while many women in their twenties being screened for breast cancer may want information on fertility, an individual user may not. Similarly, providing content found in medical guidelines ensures that certain concepts relevant to a diagnosis are further explained, but does nothing to consider individual patient information preferences. Further work must be done to design tools that simultaneously ensure the quality and relevance of information for an individual while accounting for that patient’s individual information needs and preferences.
References
Van Uden-Kraan CF, Drossaert CHC, Taal E, Shaw BR, Seydel ER, Van de Laar MAFJ. Empowering processes and outcomes of participation in online support groups for patients with breast cancer, arthritis, or fibromyalgia. Qualitative Health Research. 2008;18(3).
Keselman A, Slaughter L, Smith C, Kim H, Divita G, Browne A, et al., editors. Towards consumer friendly PHRs: patients’ experience with reviewing their health records. American Medical Informatics Association Symposium 2007. 399–403.
Zielstorff RD. Controlled vocabularies for consumer health. Journal of Biomedical Informatics. 2003;36(4–5):326–33.
PatientsLikeMe. PatientsLikeMe Homepage. 2013 [cited 2013 Jan 9]; Available from: http://www.patientslikeme.com
Fox S, Jones S. The social life of health information; Pew Research Center; 2009.
Smith M, R. Saunders, L. Stuckhardt, J.M. McGinnis. Best Care at Lower Cost: the Path to Continuously Learning Health Care in America. Washington D.C.: Institute of Medicine; 2012.
Meric F, Bernstam EV, Mirza NQ, Hunt KK, Ames FC, Ross MI, Kuerer HM, Pollock RE, Musen MA, Singletary SE. Breast cancer on the world wide web: determinants of web site popularity. Proceeds from the American Society of Clinical Oncology. 2001;20(39b).
Case DO. Looking for Information: A Survey of Research on Information Seeking, Needs and Behavior. San Diego: Academic Press Elsevier Science; 2002.
Tse T, Soergel, D ed. Exploring medical expressions used by consumers and the media: an emerging view of consumer health vocabularies. Proceeds from the American Medical Informatics Association: Symposium; 2003. 674–678.
Services USDoHaH. HHS Releases National Plan to Improve Health Literacy. 2010 [cited 2012 Sept 9]; Available from: http://www.hhs.gov/ash/news/20100527.html
Zeng QT, T.Tse, J.Crowell, G.Divita, L.Roth, A.C. Browne, editor. Identifying consumer-friendly display names for health concepts. American Medical Informatics Association Symposium; 2005;854–63.
Zeng QT, T. Tse. Exploring and developing consumer health vocabularies. Journal of Medical Informatics Association. 2006;13(1):24.
Beckjord E, Rechis R, Nutt S, Shulman L, Hesse B. What do people affected by cancer think about electronic health information exchange? Results From the 2010 LIVESTRONG Electronic Health Information Exchange Survey and the 2008 Health Information National Trends Survey. Journal of Oncology Practice. 2011;7(4):237–41.
Davidson J, Brundage M, Feldman-Stewart D. Lung cancer treatment decisions: patients’ desire for participation and information. Psycho-Oncology. 1999;8:511–20.
Butow P, Maclean M, Dunn S, Tattersall M, Boyer M. The dynamics of change: cancer patients preferences for information, involvement, and support. Annals of Oncology. 1997;8:857–63.
Jenkins V, Fallowfield L, Saul J. Information needs of patients with cancer: results from a large study in UK cancer centers. British Journal of Cancer. 2001;84(1):48–51.
Leydon G, Boulton M, Moynihan C, Jones A, Mossman J. Cancer patients information needs and information seeking behavior: in-depth interview study. British Medical Journal. 2000;320:909–13.
Gore J, Brophy C, Greenstone M. How well do we care for patients with end stage chronic obstructive pulmonary disease (COPD)? A comparison of palliative care and quality of life in COPD and lung cancer. Thorax. 2000;55:1000–6.
Murray S, Boyd K, Kendall M, Worth A, Benton T, Clausen H. Dying of lung cancer or cardiac failure: prospective qualitative interview study of patients and their carers in the community. British Medical Journal. 2002;325.
Slaughter L, Ruland C, Rotegard A. Mapping cancer patients’ symptoms to UMLS concepts. American Medical Informatics Association Symposium; 2005;699–703.
Clauser S, Wagner E, Aiello Bowles E, Tuzzio L, Greene S. Improving modern cancer care through information technology. American Journal of Preventive Medicine. 2011;40(5s2):s198–s207.
Grant R, Wald J, Poon E, Schnipper J, Gandhi T, Volk L, et al. Design and implementation of a web-based patient portal linked to an ambulatory care electronic health record: Patient Gateway for diabetes collaborative care. Diabetes Technology and Therapeutics. 2006;8(5):576–86.
Hess R, Bryce C, McTigue K, Fitzgerakd K, Zickmund S, Olshansky E, et al. The diabetes patient portal: patient perspectives on structure and delivery. Diabetes Spectrum. 2006;19:106–9.
Koch-Weser S, Bradshaw Y, Gualtieri L, Gallagher S. The internet as a health information source: findings from the 2007 health information national trends survey and implications for health communication. Journal of Health Communications. 2010;15(S3):279–93.
Bass S, Ruzek S, Gordon T, Fleisher L, McKeown N, Moore D. The relationship of internet health information use with patient behavior and self efficacy: experiences of newly diagnosed cancer patients who contact the National Cancer Institute’s Cancer Information Service. Journal of Health Communications. 2006;11:219–36.
Sarkar U, Karter A, Liu J, Alder N, Nguyen R, Lopez A, et al. The literacy divide: health literacy and the use of an internet-based patient portal in an integrated health system – results from the diabetes study of northern California (DISTANCE). Journal of Health Communication. 2010;15(S2):183–96.
DeWalt D, Berkman N, Lohr K, Pignone M. Literacy and health outcomes. Journal of General Internal Medicine. 2004;19(12):1228–39.
American Cancer Society. Breast Cancer Facts and Figures 2011–2012. Atlanta: American Cancer Society; 2012.
McKemmish S, Manaszewicz R, Burstein F, Fisher J. Consumer empowerment through metadata-based information quality reporting: The Breast Cancer Knowledge Online Portal. Journal of the American Society for Information Science and Technology. 2009;60(9):1792–807.
Eysenbach G. Design and evaluation of consumer health information web sites. In: Lewis D, Eysenbach G, Kukafka R, Stravi Z, Jimison HB, ed. Consumer Health Information: Informing Consumers and Improving Health Care. NY: Springer; 2005.
McNamara M, Arnold C, Sarma K, Aberle D, Bui A. Method to produce a data model for personalized guidleines: an exploratory study. American Medical Informatics Association: Washington D.C.; 2014. In press.
Hack T, Degner L, Dyck D. Relationship between preferences for decisional control and illness information among women with breast cancer: a quantitative and qualitative analysis. Social Science Medicine. 1994;39(2):279–89.
Degner L, Kristjanson L, Bowman D, Sloan J, Carriere K, O’Niel J, et al. Information needs and decisional preferences in women with breast cancer. Journal of American Medical Association. 1997;277(18):1485–92.
Rees C. The information needs and source preferences of women with breast cancer and their family members: a review of the literature published between 1988 and 1998. Journal of Advanced Nursing. 1999;31(4):833–41.
Wallberg B, Michelson H, Nystedt M, C B, Degner L, Wilking N. Information needs and preferences for participation in treatment decisions among Swedish breast cancer patients. Acta Oncologica. 2000;39(4):467–76.
Fletcher SW. Screening for breast cancer: strategies and recommendations. 2014 [cited 2014 Feb 1]; Available from: http://www.uptodate.com/contents/screening-for-breast-cancer-strategies-and-recommendations?source=search_result&search=breast+cancer+screening& selectedTitle=1~49
Network NCC. NCCN clinical pratice guidelines in oncology (NCCN Guidelines) breast cancer. 2013 [cited 2014]; Available from: http://www.nccn.org/professionals/physician_gls/pdf/breast.pdf
Prevention CfDCa. Symptoms of breast cancer. Atlanta: Center for Disease Control and Prevention; 2013 [cited 2014]; Available from: http://www.cdc.gov/cancer/breast/basic_info/symptoms.htm
Burstein F, McKemmish S, Fisher J, Manaszewicz R, Malhotra P. A role for information portals as intelligent decision support systems: breast cancer knowledge online experience. In: N.D. J, Gupta GAF, Manuel Mora T, editors. Intelligent Decision-making Support Systems: Foundations, Applications and Challenges: Springer-Verlag; 2006.
Kroenke CH, Kubzansky LD, Schernhammer ES, Holmes MD, Kawachi I. Social networks, social support, and survival after breast cancer diagnosis. Journal of Clinical Oncology. 2006;24(7):1105–11.
Johnson K, Jimison HB, Mandl KD. Consumer health informatics and personal health records. In: Shortliffe EH, Cimino JJ, editors. Biomedical Informatics. London: Springer; 2014.
Frost J, Massagli M. Social uses of personal health information within PatientsLikeMe, an online patient community: what can happen when patients have access to one another’s data. Journal of Medical Internet Research. 2008;10(3):e15.
Shaw BR, McTavish F, Hawkins R, Gustafson DH, Pingree S. Experiences of women with breast cancer: exchanging social supprt over the CHESS computer network. Journal of Health Communication: International Perspectives. 2000;5(2):135–59.