Abstract
Front line health care providers (HCPs) play a central role in endemic (e.g., pertussis), epidemic (e.g., influenza) and pandemic (e.g., avian influenza) infectious disease outbreaks. Effective preparedness for this role requires access to and awareness of population-based data (PBD). We investigated the degree to which this is currently achieved among HCPs in Utah by surveying a sample about access, awareness and attitudes concerning PBD in clinical practice. We found variability in the number and nature (national vs. local, pushed vs. pulled) of PBD sources accessed by HCPs, with a subset using multiple sources and using them frequently. We found that HCPs believe PBD improves their clinical performance and that they cannot rely on their own practice to remain informed. These findings suggest that an integrated system, which interprets PBD from multiple sources and optimizes the delivery of PBD, may facilitate preparedness of HCPs through the application of PBD in routine clinical practice.
INTRODUCTION
Frontline health care providers (HCPs), particularly HCPs that provide urgent or acute care, play a critical role in the diagnosis, treatment, and prevention of communicable diseases, including epidemic and endemic respiratory diseases such as influenza and pertussis, which have an important impact on public health. However, current systems are not designed to facilitate the flow of information between the clinical and public health communities and many HCPs may not think about public health issues in their day-today practice. Limited data are available regarding HCPs’ current use of population-based data in clinical practice. A better understanding of the baseline situation is necessary to take the next step to improve HCPs’ use of population-based data in clinical practice and improve communication between health care providers and public health.
BACKGROUND
During the past few years, several events have highlighted the critical role played by health care providers in recognition and control of communicable diseases, including recognition of inhalational anthrax(1), West Nile Virus(2), and SARS.(3). In addition to recognizing unusual cases, health care providers contribute to public health surveillance and disease tracking by reporting notifiable diseases such as pertussis to the health department. HCPs may work with the public health community to identify contacts, conduct diagnostic testing, and prescribe prophylaxis or treatment in an outbreak situation. Finally, HCPs contribute to public health goals by vaccinating their patients and preventing the spread of disease in their individual patients and in the community via herd immunity.
The Importance of Population-based Data to Primary Care Providers’ Clinical Practice:
Ibrahim and colleagues have argued for the value of embracing a population-based perspective in clinical practice. The authors describe 5 key population-based principles that can be applied in clinical practice: “a community perspective, a clinical epidemiology perspective, evidence-based practice, an emphasis on outcomes, and an emphasis on prevention.”(4) For example, a physician with a community perspective and an emphasis on prevention would not only recommend influenza vaccination for his or her individual patients but would participate in community outreach including vaccinating family members or encouraging vaccination in schools. A provider who practices evidence-based medicine would know that the rapid influenza test is not accurate when the level of influenza activity is low and would choose a more accurate test when he or she was suspicious of influenza early in the season.
Examples of population-based data for respiratory pathogens with clinical value for individual patient encounters include patterns of antimicrobial resistance that can guide antibiotic choices and presence of a community outbreak of pertussis that significantly increases the possibility that a chronic cough represents pertussis infection. Similarly, the diagnosis of a patient with influenza-like illness is substantially aided by knowledge of current trends in influenza, parainfluenza, and other illnesses.
Barriers to Use of Population-Based Data in Clinical Practice:
Time pressures in the clinical environment limit HCPs’ attention to public health considerations. Despite the value that knowledge about community infectious disease prevalence and trends can provide, clinical providers’ overwhelming focus is on the individual patient encounter. In addition, public health guidelines are often poorly accessible and lengthy. HCPs desire information that is clinically relevant, locally applicable, and generated by a trusted source.(5) In short, they require decision support tools to translate surveillance data and public health and clinical guidelines into an organized clinical response.
Current Sources of Population-based Data Available to Physicians in Utah:
Web sites and listservs maintained by the Centers for Diseases Control and Prevention (CDC), which includes the Morbidity and Mortality Weekly Report (MMWR), provide state-specific surveillance data about various infectious diseases. In addition to these sources, there are several sources of PBD about respiratory pathogens specific to Utah. Germ Watch, is a system jointly developed by Intermountain Healthcare (IH), the University of Utah Department of Pediatrics and the Utah Department of Health that generates, posts, and delivers epidemic curves to front-line HCPs. Germ Watch gives HCPs access to weekly PBD about respiratory pathogens, including influenza, RSV, parainfluenza, adenovirus, human metapneumovirus, and pertussis based on diagnostic testing performed in all IH facilities.(5) Germ Watch has both a dedicated web site (available from: http://www.intermountainphysician.net/germwatch) and a weekly listserv that includes epidemic curves and very brief synopses of the activity of respiratory pathogens listed above. The Utah Department of Health (UDOH) Division of Epidemiology hosts a web site (available from: http://epi.health.utah.gov/) that provides data about reportable disease activity in Utah: weekly updates about influenza activity surveillance, including influenza-like-illness activity, influenza-associated hospitalizations and school absenteeism. The site also contains links to a variety of resources such as historical surveillance data and guidelines about various reportable diseases. UDOH also maintains an Infectious Disease Update listserv that delivers weekly updates about current disease outbreaks and a range of other information about temporally relevant infectious disease issues ranging from vaccine information to regional activity of drug resistance organisms like MRSA. The listserv delivers this information as a textual summary sent via e-mail to clinicians who have signed up to receive them.
OBJECTIVES
Our overall objective was to assess HCPs’ current perspective on the use of population-based data in clinical practice with a focus on respiratory pathogens. Specifically, we sought to assess HCPs’ 1) current exposure to population-based data related to respiratory pathogens, specifically current data sources, frequency of exposure and number of sources accessed; 2) attitudes about the usefulness of population-based data in their clinical practice (performance expectancy); 3) attitudes about access to and ease of use of population-based data (effort expectancy); and finally, 4) intention to use population-based data for clinical decision-making (use intent).
METHODS
Study population:
The study population comprised all 106 HCP who provided care for patients in any one of the 28 Urgent Care Clinics owned by Intermountain Healthcare (IH) as of 10/01/2007. IH is an integrated health care delivery organization that operates 20 hospitals and over 100 ambulatory care facilities, including 28 urgent care settings, located throughout Utah and southern Idaho.
Survey design and procedure:
We performed a descriptive, cross-sectional survey of the study population and developed our survey based on the Unified Theory of Acceptance and Use of Technology (UTAUT).(6) The UTAUT model employs validated scales to measure three direct determinants of intention to use (performance expectancy, effort expectancy and social influence) and two direct determinants of usage behavior (use intent, facilitating conditions). UTUAT scales were adapted to capture the concept of HCPs’ utilization of electronic information sources of PDB. The survey consisted of 60 individual questions to assess HCPs’ utilization and exposure to PDB and the performance expectancy, effort expectancy, and social influence associated with its use. In addition, the HCPs’ intent to use PBD in the future was also surveyed. Seven demographic questions, including years in practice, practice setting and age distribution were also asked. We pre-tested the survey on public health and physician colleagues, and then pilot tested the survey on 15 physicians.
After obtaining institutional review board approval from the University of Utah and IH, we sent up to three e-mails to HCPs requesting their participation in the survey. The emails contained a link to the internet-based survey. A study coordinator tracked the names of the participants who completed the survey so that respondents would not get additional e-mail messages asking them to complete the survey. The investigators did not have access to the names of respondents or non-respondents, and the database used for analysis did not include participants’ names or any information that could potentially identify an individual respondent. After three attempts to contact participants by email, a paper-based version was sent to the participant’s home address. Surveys were completed between November 1st, 2007 and February 29th, 2008.
RESULTS
Description of respondents:
Among 106 eligible HCPs, 63 (60%) responded to the survey. Half (51%) of the respondents were male. The respondents were an average of 44 years (ranging from 31 to 60 years) and had an average of 11 years in clinical practice (ranging from 0.5 to 30 years). The respondents cared for both adults and children in the urgent care settings. Most (78%) of the respondents reported that during a typical week, more than 25% of their patients were younger than 18 years; 24% of the respondents only cared for pediatric patients. A majority of respondents (90%) reported spending greater than 20 hours per week seeing patients during the 2006–2007 winter ‘respiratory season’. The primary work setting for most of the respondents was reported to be an urgent care setting (71%) or an office-based setting (21%). The remaining (8%) of HCPs primarily worked in emergency medicine (n=2), as a specialist (n=2) or hospitalist (n=1).
Assessment of current exposure to population-based data:
As shown in Table 1, the HCPs use a variety of information sources that either require them to seek out the information (i.e., ‘pulled’) or require that the physician be enrolled in a listserv so the information can be ‘pushed’ when it is available. While most HCPs (63–86%) were accessing information from the national source, Centers for Disease Control and Prevention, only one-third of the physicians were accessing information from state-based resources, primarily the Utah Department of Health. Less than 50% accessed the Germ Watch website and listserv, which is a resource integrated with the clinical information system used by IH physicians. The primary reason reported for a source not being accessed was lack of awareness that it existed. This was particularly the case for local sources.
| Source | Number (%) of respondents that access each resource | Frequency of access during the 2006–2007 respiratory season | ||||
|---|---|---|---|---|---|---|
| Never | Once a quarter | Once a month | Twice a month | Once a week | ||
| Data must be ‘pulled’ by the user | ||||||
| Centers for Disease Control and Prevention (CDC) website | 54 (86%) | 15% | 54% | 15% | 13% | 4% |
| Morbidity and Mortality Weekly Report (MMWR) | 40 (63%) | 35% | 38% | 18% | 3% | 8% |
| Utah Department of Health Communicable Disease Epidemiology website | 19 (30%) | 21% | 47% | 5% | 5% | 21% |
| Germ Watch website | 27 (43%) | 11% | 30% | 26% | 19% | 15% |
| Data is ‘pushed’ to the user | ||||||
| UDOH “Weekly Infectious Disease Update” listserv | 20 (32%) | 10% | 15% | 10% | 15% | 50% |
| Germ Watch listserv | 41 (65%) | 0% | 15% | 24% | 22% | 39% |
Thirty percent reported accessing sources of PBD other than those specifically asked about in the survey. ‘Other’ sources included searches using Google, PubMed, Up-to-Date, MD Consult, Flu Watch, various journals (e.g., Pediatric Infectious Disease Journal), information posted on web sites (e.g., the National Association of Pediatric Nurse Practitioners, American Academy of Nurse Practitioners, and the American Academy of Pediatrics), newspapers, textbooks, and colleagues.
Frequency of exposure:
Seventy-eight percent of HCPs accessed at least 1 source monthly and 40% accessed at least 1 source weekly. While the national sources of PBD--CDC web site and the MMWR--were the most frequently accessed sources, less than one third of respondents accessed these sources at least once per month. In contrast, HCPs accessed the local sources of PBD, which are delivered by weekly e-mail (UDOH and Germ Watch listservs) more frequently (50% and 39%, respectively).
Number of sources accessed:
During the 2006–2007 respiratory season, 100% of respondents reported accessing at least 1 source of PBD. The most frequently reported number of sources accessed was 4 (30% of respondents). Ninety-four percent used 2 or more sources with a maximum use of 7 sources.
Performance Expectancy:
A majority, >90% of respondents agreed that they would find PBD about respiratory pathogens 1) useful in clinical practice, 2) useful for describing “what’s going around” to patients, 3) describing treatment decisions and 4) improving diagnostic accuracy. A lesser majority agreed that they would find PBD about respiratory pathogens useful for deciding whether or not to 1) order viral tests or testing for pertussis (79%), 2) prescribe antiviral medication (71%) and 3) prescribe antibiotics (67%). A minority of respondents agreed that they would find PBD about respiratory pathogens useful for deciding whether or not to admit a patient to the hospital (38%) and order tests that are useful when concern exists for bacterial infection (blood culture 41%, complete blood count 40%). Nearly half of the respondents agreed that they would find PBD about respiratory pathogens useful for enabling them to accomplish tasks more quickly (49%).
Two thirds (68%) agreed that current sources of PBD reflect what they are seeing in their clinical practice. Only 15% of HCPs agreed that their clinical practice was sufficient to understand “what was going around” in their communities.
Effort expectancy:
Nearly a third (62%) of the respondents believed that it was easy to access PBD about respiratory pathogens circulating in their area. However, over half either agreed with (19%) or were neutral about (33%) the statement, “it takes too long to interpret PBD to make it worth the effort,” and just under half agreed (14%) were neutral (30%) about the statement, “using PBD about respiratory pathogens takes too much time away from caring for individual patients.” A majority (73%) did not agree that it was difficult to understand how PBD can be applied to their clinical practice.
Social Influence:
Two thirds (67%) agreed that their patients and colleagues expected them to know about PBD for respiratory pathogens. Over half agreed that their colleagues use PBD in clinical practice (57%).
Use Intent:
A high proportion of respondents (79%) intended to use PBD during the 2007–2008 winter respiratory season.
DISCUSSION
To our knowledge, this is the first study to address the question of what internet-based sources of PBD about respiratory pathogens front-line HCPs are currently accessing. With respect to HCPs’ utilization of population-based data in clinical practice, it is also the first study to evaluate the three direct determinants of intention to use (performance expectancy, effort expectancy, and social influence) and one direct determinant of usage behavior (use intent) posited by the UTUAT.
The patterns of population-based data access reported by this sample of urgent care providers indicates that front-line HCPs in Utah have significant, ongoing interest in timely and regional PBD about respiratory pathogens. Our data suggest substantial variability in the number and nature (national vs. local, pushed vs. pulled) of PBD sources accessed by HCPs, but a majority of the sources being electronic and a subset of HCPs using multiple sources and using them frequently. This finding is important because it represents an opportunity to capitalize on existing information-seeking support systems such as information dashboards and information buttons to provide context-specific PBD at the point of care.(7)
The fact that the local PBD sources that are ‘pushed’ via e-mail--Health Department and Germ Watch listservs--were accessed more frequently then ‘pulled’ resources, is not surprising given the time pressures on clinicians providing acute care. HCPs need local data in a quickly accessible and easily interpreted format. The providers’ reports of lack of awareness that PBD sources exist, suggest that better approaches to advertising and disseminating these resources are needed. These results have already led to the expansion of the UDOH and GermWatch listservs and improvements in the placement of links to the web pages.
Our results suggest that a significant majority (75%) of HCPs understand how PBD can be applied in their clinical practice and access it because they believe the data to be useful in clinical practice for improving diagnostic accuracy, improving communication with patients about pathogens circulating in their regions, and describing their treatment decisions. However, providers’ responses that using PBD takes too much time away from caring for individual patients and that PBD takes too long to interpret suggest that techniques for aiding interpretation are important. Such techniques could include presenting a graph accompanied by a few bullet points with interpretation and recommendations about how to apply the data in clinical practice.
While most providers reported that they intend to use PBD in their practice, future research will include measures of their actual use and changes in clinical decision-making.
Some of the conclusions drawn from this study may not be generalizable across care settings or to other areas of the country. The local PBD sources, specifically Germ Watch, accessed by IH providers may be unique in that the data come from clinical testing being done throughout a large, integrated care delivery network that provides a majority of the health care delivered in the state of Utah. This level of penetration provides a covered population that approximates the entire population of Utah and thus the information content of this data source may be viewed as more useful by this sample of physicians. Furthermore, a response of 60% introduces uncertainty about the representativeness of our sample and likely introduces a bias towards HCPs that are enthusiastic about PBD. We think this bias is likely to overestimate providers’ performance expectancy, underestimate effort expectancy, and overestimate intent to use PBD in clinical practice.
CONCLUSIONS
The results of our survey suggest that HCPs are interested in PBD from both national and local sources and believe it improves their clinical performance because they cannot rely on their own practice to remain informed. An integrated system that interprets PBD from multiple sources, including healthcare delivery systems, may facilitate preparedness of HCPs through the application of PBD in routine clinical practice.
Acknowledgments
This research was funded by the Centers for Disease Control and Prevention through the Utah Center of Excellence in Public Health Informatics (Grant # 1P01CD000284-01). This project was also supported in part by the Children’s Health Research Center at the University of Utah and Primary Children’s Medical Center Foundation. We thank Marjorie Carter and Amyanne Wuthrich for their support with conducting the survey, and the providers and administration at Intermountain Healthcare for supporting this research.
REFERENCES
- 1.Bush LM, Abrams BH, Beall A, Johnson CC. Index case of fatal inhalational anthrax due to bioterrorism in the United States. N Engl J Med. 2001 Nov 29;345(22):1607–10. doi: 10.1056/NEJMoa012948. [DOI] [PubMed] [Google Scholar]
- 2.Asnis DS, Conetta R, Teixeira AA, Waldman G, Sampson BA. The West Nile Virus outbreak of 1999 in New York: the Flushing Hospital experience. Clin Infect Dis. 2000 Mar;30(3):413–8. doi: 10.1086/313737. [DOI] [PubMed] [Google Scholar]
- 3.Srikantiah P, Charles MD, Reagan S, Clark TA, Pletz MW, Patel PR, et al. SARS clinical features, United States, 2003. Emerg Infect Dis. 2005 Jan;11(1):135–8. doi: 10.3201/eid1101.040585. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Ibrahim MA, Savitz LA, Carey TS, Wagner EH. Population-based health principles in medical and public health practice. J Public Health Manag Pract. 2001 May;7(3):75–81. doi: 10.1097/00124784-200107030-00012. [DOI] [PubMed] [Google Scholar]
- 5.Gesteland PG, Samore MH, Pavia ATA, S R, Korgenski K, Gerber K, et al. Informing the Front Line about Common Respiratory Viral Epidemics. Proceedings of the Annual American Medical Informatics Association Symposium. 2007:274–8. [PMC free article] [PubMed] [Google Scholar]
- 6.Venkatesh V, Morris MG, Davis FD. User Acceptance of Information Technology: Toward a Unified View. MIS Quarterly. 2003 Sep;27(3):425–78. [Google Scholar]
- 7.Del Fiol G, Haug PJ. Infobuttons and classification models: A method for the automatic selection of on-line information resources to fulfill clinicians’ information needs. J Biomed Inform. 2007 Dec 8; doi: 10.1016/j.jbi.2007.11.007. [DOI] [PMC free article] [PubMed] [Google Scholar]