Clinician Survey to Determine Knowledge of Dengue and Clinical Management Practices, Texas, 2014

Introduction

Dengue is a mosquito-borne acute febrile illness that is caused by one of four dengue viruses (DENV-1, 2, 3, and 4).1 Dengue is endemic throughout the tropics and subtropics worldwide, including the Americas where the incidence of dengue has increased greater than 4-fold during the last three decades.2 During the same period, the region has experienced an increase in both frequency of dengue outbreaks and disease severity.

In the United States, dengue is endemic in the territories of Puerto Rico and the U.S. Virgin Islands.3 The majority of dengue cases identified in the 50 states are travel associated, and dengue is the leading cause of febrile illness among U.S. travelers returning from southeast Asia, Latin America, and the Caribbean.4 Travel-associated dengue cases have led to outbreaks with local DENV transmission in Hawaii,5 Florida,6 and Texas.7,8 Since 1980, Texas border communities have experienced multiple dengue outbreaks concurrent with epidemics in Mexico.8–10 The possibility of dengue outbreaks in Texas is a continual threat because of its proximity to Mexico, substantial cross-border traffic, presence of Aedes aegypti mosquitoes,11 and a predominantly nonimmune U.S. population.

Although the majority of DENV infections result in asymptomatic infection or mild acute febrile illness, a limited number (∼5%) of symptomatic infected persons will develop severe dengue, a potentially life-threatening illness with a vascular leakage syndrome resulting in hypovolemic shock and hemorrhage.12–14 Fatality rates for hospitalized patients with severe dengue can be as high as 10%; however, studies have reported that mortality rates can be reduced to < 0.5% with early case identification and timely initiation of correct clinical management.1,15,16

During May–October 2013, the Mexican state of Tamaulipas, which borders Texas, experienced a dengue epidemic and ∼5,500 suspected dengue cases were reported.17 During the same period in Texas, 53 laboratory-confirmed dengue cases (median patient age = 28 years, range = 1–85 years) were detected in south Texas border counties; 26 (49%) were locally acquired.18 Although sporadic cases of dengue have been detected in south Texas during past decades, this outbreak was the largest since the 2005 outbreak in Brownsville, TX.9,19 Because of the substantial increase in dengue cases reported along the United States–Mexico border, the Texas state and local public health officials partnered with Centers for Disease Control and Prevention (CDC) to develop strategies to reduce morbidity and mortality associated with dengue. One objective was to survey local Texas clinicians' knowledge of dengue and its clinical management to guide intervention strategies.

Materials and Methods

Physicians, physician extenders (nurse practitioners [NPs], and physician assistants [PAs]) practicing medicine in the Texas counties of Cameron, Hidalgo, Starr, Willacy, and Zapata (collectively known as south Texas) were invited by e-mail to participate in the dengue clinical survey. We also invited clinicians from the Houston metropolitan area (HMA) to participate because of identification of both locally acquired and travel-associated dengue cases in the HMA.20 We specifically sought participation from clinicians in general practice, family medicine, internal medicine, pediatrics, emergency medicine, infectious disease, hematology, obstetrics, and intensive care, because these medical specialties might be more likely to encounter a patient with dengue.

We developed our survey instrument by using multiple data sources to formulate the survey questions. We conducted key informant interviews with treating clinicians and a medical record review of confirmed and suspected dengue patients at south Texas hospitals to identify clinical management practices and whether these practices varied from the World Health Organization 2009 dengue guidelines21. Findings of deficiencies in recommended clinical management practices (e.g., failure to recognize dengue and warning signs for developing severe dengue, inaccurate understanding of dengue diagnostics, and use of nonsteroidal anti-inflammatory drugs [NSAIDs]) were used to develop survey questions.

The 28-question survey included 10 optional demographic questions and 18 technical questions. The technical questions were divided into three knowledge areas, including dengue prevention messages and anticipatory guidance, clinical presentation and course, and dengue clinical management. We also asked questions to assess clinical practice habits. We piloted the survey to assess for clarity and level of difficulty by using a convenience sample of nine physicians and incorporated feedback. The survey was administered online through SurveyMonkey^® (SurveyMonkey, Palo Alto, CA).22 Answering all technical questions was required to advance and complete the survey. After survey completion, participants were redirected to a webpage that provided the correct answers with explanations. Participants had the option to redirect to the CDC dengue website (http://www.cdc.gov/dengue) for additional information. Data were analyzed using IBM SPSS Statistics for Windows (Version 20.0; IBM Corporation, Armonk, NY) and difference in responses among groups was analyzed using Pearson's χ² test. Time in practice was analyzed using clinicians in practice > 30 years as the comparator. Some of the 18 technical questions had multiple correct answers, so each answer was treated like a true or false question. We decided a priori to exclude surveys from clinicians who were no longer practicing clinical medicine, practicing outside the catchment area, and did not return complete survey instruments. We chose to exclude the first and second categories of clinicians since our goal was to evaluate clinical decision-making in dengue-prone areas.

Potential participants included all members of local medical societies in south Texas and HMA. In addition, we partnered with the two major health systems in HMA to distribute our survey via e-mail to their clinician employees. All participants were invited via e-mail either through their membership with a local medical society or in the case of the HMA clinicians; many received an invitation from their employer. We were unable to access a list of all licensed clinicians in the catchment area. A letter of endorsement from the Regional Medical Director of south Texas was included with the e-mail containing the survey website link. As an incentive, participants completing the survey were entered in a drawing to win a tablet device provided by the CDC U.S. Mexico Unit. Participants were asked to complete the survey within 2 weeks; a reminder e-mail was sent after the first week had passed, if the survey had not been completed.

This survey was reviewed by CDC for human subjects protection, and was deemed to be nonresearch.

Results

The survey was sent to 2,375 clinicians, including 1,410 (59%) physicians, 400 (17%) NPs, and 565 (24%) PAs (Figure 1). A total of 339 (14%) clinicians responded, although 71 (21%) did not complete all the technical questions and were excluded from analysis. Of 268 completed surveys, 51 (15%) were excluded because the participant was either not active in direct patient care or practiced outside of the defined catchment area. Data from the remaining 217 (9%) surveys were included in our analysis.

Clinician enrollment for dengue clinical survey in Texas during 2014.

Citation: The American Society of Tropical Medicine and Hygiene 96, 3; 10.4269/ajtmh.16-0367

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Clinician enrollment for dengue clinical survey in Texas during 2014.

Citation: The American Society of Tropical Medicine and Hygiene 96, 3; 10.4269/ajtmh.16-0367

• Download Figure
• Download figure as PowerPoint slide

Clinician enrollment for dengue clinical survey in Texas during 2014.

Citation: The American Society of Tropical Medicine and Hygiene 96, 3; 10.4269/ajtmh.16-0367

• Download Figure
• Download figure as PowerPoint slide

Response rates differed by the following geography and clinician type: HMA physicians had the highest response rate (16%), followed by HMA physician extenders (15%), south Texas physician extenders (8%), and south Texas physicians (4%). Participants from all regions were approximately equally split between physicians (54%) and physician extenders (46%) (Table 1). The majority (73%) of participants had been in practice for ≤ 20 years and reported working in an outpatient setting (86%). The majority (90%) of participants practiced in a primary care specialty, and a substantial number (42%) were pediatricians. The majority (81%) of participants had never diagnosed a case of dengue; 25 (12%) participants had diagnosed 1–4 cases. Seven (3%) participants (all south Texas physicians) reported having diagnosed > 20 dengue cases during their career; the majority of these participants were trained in infectious disease. Deficiencies in knowledge of dengue were present among all groups, and no differences were noted among specialties.

Discussion

This survey among HMA and south Texas clinicians revealed critical knowledge gaps of dengue and its clinical management. Overall knowledge of anticipatory guidance and prevention messages for dengue and its clinical management was low among all participants regardless of clinician type, years of practice, and experience managing dengue patients. However, numerous clinicians did report correctly avoiding NSAIDS and ordering dengue diagnostic testing for patients suspected of having dengue. Participants understood dengue clinical presentation and course. Specific differences were noted between physicians and physician extenders in the identification of early indicators of shock and indications for blood transfusion. Physician extenders performed better concerning questions related to dengue identification than questions concerning inpatient clinical management, which might be related to training and practice sites (physician extenders may be more commonly involved in primary care versus specialty training). In addition, those in clinical practice > 10 years were more likely to identify the indications for blood transfusion in dengue patients than those practicing ≤ 10 years. Participants in clinical practice > 30 years were less likely to recognize scenarios of plasma leakage and more likely to prescribe corticosteroids to dengue patients than those practicing ≤ 30 years.

This survey was subject to multiple limitations, including participants' response and survey completion rates were low and varied with regard to specialty and geographic region. Specifically, HMA clinicians were more likely to participate, which might be attributable to differing recruitment methods in HMA and south Texas (i.e., the former received survey invitations from their employer and the latter from local medical societies). Second, the number of participating physicians from south Texas was low and prevented direct comparisons between groups. Third, less knowledge of inpatient clinical management might be attributable to an overrepresentation of outpatient clinicians. Fourth, the survey did not allow participants to leave any question blank, forcing a response to maximize the number of completed surveys for analysis but may have skewed the sample toward those with more knowledge on dengue. Fifth, the sampling method may have selected for clinicians with more interest in dengue and potentially biasing toward those with more dengue knowledge. Finally, we were unable to describe clinicians who did not complete the survey, which limited our ability to fully account for bias.

Response rates in clinician surveys have traditionally been poor with the best response rates being roughly 40–50%.23 Despite the low response rate, clear deficiencies in knowledge about dengue were evident among all participants. Given the abovementioned limitations and biases, it is possible that knowledge was skewed toward those with greater dengue interest and knowledge, magnifying knowledge deficiencies. Clinicians who reported experience in diagnosing and managing dengue, albeit with a limited number of patients, still scored poorly. We identified knowledge gaps reported from similar clinical surveys administered in dengue-endemic areas; clinicians lacked knowledge about dengue clinical characteristics,24 severe disease warning signs, discharge criteria,25 and prevention methods.26 Limited recognition of the critical phase of disease by clinicians was also demonstrated in a survey conducted in Puerto Rico.16 These surveys were conducted in endemic areas and reported higher response rates than our survey. However, it is likely that physicians working in areas which see only sporadic dengue (such as south Texas) have less knowledge about dengue, especially since U.S. medical education often lacks tropical medicine training.

These findings highlight a need for clinical management training for U.S. clinicians across all points of care in regions likely to treat dengue patients. Training is especially important in south Texas, because studies have demonstrated a continued risk for dengue outbreaks attributable to a highly mobile and susceptible population, regular DENV introduction, and presence of competent mosquito vectors in south Texas.11 Further, all four DENV have been reported in the region, thereby increasing risk for secondary infections among the susceptible resident U.S. population.9 Because dengue patients experiencing secondary DENV infection are associated with increased risk for developing severe dengue, compared with patients with primary infection,1 future outbreaks in south Texas might be associated with increased incidence of severe dengue.8

Focus on clinician training and education in areas at risk for DENV transmission or among clinicians who serve at-risk travelers is an important public health strategy to decrease morbidity and mortality associated with dengue. Previous physician trainings on dengue clinical management in Puerto Rico demonstrated improved management of hospitalized dengue patients after the implementation of a classroom-style physician training course.27 Medical curriculum should emphasize identified knowledge gaps, including warning signs for severe dengue and the clinical management of severe dengue. Online curriculum such as the CDC's Dengue Clinical Management Course (http://www.cdc.gov/dengue/training/cme.html) is an attractive option for education with free continuing medical education. Clinicians in Texas might also benefit from cross-border collaboration with clinicians in Mexico who have extensive dengue experience; possible projects might include clinical exchange programs or clinical research collaboration.