Health information exchange use during dental visits

Abstract

Dental and medical providers require similar patient demographic and clinical information for the management of a mutual patient. Despite an overlap in information needs, medical and dental data are created and stored in multiple records and locations. Electronic health information exchange (HIE) bridge gaps in health data spread across various providers. Enabling exchange via query-based HIE may provide critical information at the point of care during a dental visit. The purpose of this study is to characterize query-based HIE use during dental visits at two Federally Qualified Health Centers (FQHCs) that provided on-site dental services. First, we determine the proportion of dental visits for which providers accessed the HIE. Next, site, patient and visit characteristics associated with query-based HIE use during dental visits are examined. Last, among dental visits with HIE use, the aspects of the HIE that are accessed most frequently are described. HIE use was low (0.17%) during dental visits, however our findings from this study extend the body of research examining HIE use by studying a less explored area of the care continuum.

Introduction

Dental providers and medical providers require similar patient demographic and clinical information for the management of mutual patients¹. For example, when caring for diabetic patients, dental providers need to know the patient’s medications and most recent HbAlc value, a measure that indicates blood sugar control. Uncontrolled diabetes may cause burning sensation and dryness in the mouth, salivary dysfunction, an increased risk for caries and gingivitis, and impaired/delayed wound healing². Additionally, dental providers need to be cognizant of medical conditions that can impact selection of antibiotics, analgesics, and anesthesia for dental procedures and conditions^3-5. Given that poor oral health is associated with several chronic conditions, managing these complex patient comorbidities requires up-to-date health information^6-8.

Despite these information needs in dental practice, medical and dental data are frequently stored in multiple records and locations. As a result, the accuracy of data within dental records is often based on a patient’s ability to recall prescription names, dosages, and details associated with the services rendered from other health providers. This contributes to high observed rates of discrepancies between medical and dental records (72-86%),^1,9-12 which may in turn lead to duplicative and inefficient care, medical errors, and suboptimal patient experiences^1,13,14.

Electronic health information exchange (HIE) bridges gaps in health data spread across various providers, including dentists^8,13-16. The adoption of electronic records by medical and dental providers has increased over the last decade, offering an opportunity to link patients’ dental records with their medical records through a HIE^17,18. Using HIE, a dental provider can search for relevant medical information missing from the electronic dental record (EDR), removing the reliance upon patient recall of this information. One form of HIE – query-based HIE – is particularly well-suited to meet missing information needs, as it does not generally require providers to implement new software within practice settings¹⁹. Providers simply log in to a web-based HIE query portal to search for outside patient health information. Enabling exchange via query-based HIE may reduce information discrepancies across health records – including dental records – and provide critical information at the point of care¹⁹.

Recent federal policies, such as the Health Information Technology for Economic and Clinical (HITECH) act and the 21^st Century Cures Act, provide incentives for HIE development and encourage exchange of electronic health information between all health providers, including dentists^13,20. Taken together, the clinical information needs of dentists combined with this policy emphasis makes dental use of HIE an important area of study; however, dental providers’ use of the HIE is not well understood. Given the relative paucity of evidence, it is important to determine the extent to which dental providers use HIE and what types of outside clinical information are valuable to a dental provider.

The purpose of this study is to characterize query-based HIE use during dental visits at two Federally Qualified Health Centers (FQHCs) that provide on-site dental services. First, we determine the proportion of dental visits for which providers accessed the HIE. Next, we analyze patient and visit characteristics associated with query-based HIE use during dental visits. Last, looking within the sample of dental visits with HIE use, we use HIE audit logs to determine the most frequently accessed HIE information. This is the first study to examine HIE use in the dental context, and our results provide baseline rates of query-based HIE use for dental visits and describe aspects of the nature of that use. This study’s findings are of value to health systems seeking to increase the use of HIE and health information technology broadly across multiple unique providers, including dental practices. HIE system designers, health system administrators, government agencies, and provider organizations working to integrate medical and dental care may gain insights into dental provider information needs. Finally, our study extends the body of research examining HIE use among clinical providers across clinical settings by examining an under-explored area of the care continuum.

Methods

Setting

Two FQHCs – Anthony Jordan Health Center (AJHC) and Oak Orchard Community Health Center (OOCHC) – in the Rochester, New York area had access to the query-based HIE portal of the Rochester Regional Health Organization (RHIO). Each FQHC site provided dental services on-site and utilized a dental module built into the EHR to collect and document data. More specifically, dental professionals would begin in the EHR under the patient’s name and then open the EDR module to document the information specific to the mouth and the patient’s oral health.

The Rochester RHIO is a not-for-profit health information organization in Western New York State that provides HIE services to 13 counties in the region. The organization has developed and maintained a secure web-based portal that affords authorized providers and clinical staff members from participating health care organizations access to patient health information provided by its more than 100 data-contributing organizations. Contributor organizations include hospitals, reference laboratories, physician practices, public health organizations, payers. During the time of our study, more than two-thirds of hospitals and physician practices in the region participated in the HIE, which contained information on more than 1 million patients. These data consist of discharge summary documents, diagnoses, medication history, laboratory results, radiology reports and images, and information from payers. After users are provisioned with an account by the RHIO, users log in to the portal to search for a patient and confirm that the patient has consented to data sharing. New York is an “opt-in” HIE state, meaning that patients must explicitly give permission for their data to be shared between providers²¹. Almost all (97%) of patients in the Rochester RHIO consent to HIE²². Once in the HIE, the user is presented with a summary tab including the patient’s recent laboratory values, radiology results and reports, admission-discharge-transfer (ADT) notifications, and medication history. From this landing page, users can navigate to section tabs specific to laboratory results, radiology results, reports, or ADT notifications, with additional detail available on those tabs. If users require a more detailed or higher-intensity look at patient data, they may also access documents specific to a single laboratory test, medication, radiology report, or ADT notification.

Data & Sample

Clinical EHR data was retrieved from the two FQHCs during January 2012 through December 2015. We identified dental visits via the visit type and description fields of the EHR data. This data also included an anonymized patient identifier for matching with HIE use log data, the patient’s gender, age, visit date, and the unique employee identifier for the provider seen during the visit.

The Rochester RHIO provided detailed access log data documenting use of the query portal by all users at the two FQHCs in our study during the same four-year period. An HIE event observation includes click-level data capturing the unique portal user, the site with which that user was affiliated, the portal page(s) viewed by the user, timestamps for each action (i.e. click) and a unique anonymized patient identifier allowing for linkage to the patient’s clinical EHR data. These observations are divided into two types: records of section tab viewing and records of document viewing. Within the query portal, there are tabbed sections for a patient summary, laboratory results, radiology results, radiology reports, ADT alerts, and medications. Document viewing is logged whenever a user clicks to view a specific result, for example a comprehensive metabolic panel result from a laboratory or an ADT report from a recent emergency department visit. (See Figure 1)

Figure 1.

Data generation process

We linked the HIE query portal data and dental visit data by matching the patient identifier and dental visit dates to all HIE use taking place in the two days prior to the visit or on the visit date. We limited qualifying HIE use to use by a user from the site at which the visit occurred. We limited the sample of patient dental visits to those that did not have medical visits in the two days prior to the dental visit or on the same day as the dental visit, to avoid erroneously attributing medical-visit HIE use to the dental visit. This allowed us to examine HIE query portal use for patients with a dental visit.

Measures

Our outcome measure was a binary indicator of HIE use for the dental visit, defined as HIE use that occurred on the same day as the dental visit or in the two days preceding the dental visit by a provider at the same site as the dental visit. Visits meeting these criteria served as our numerator for calculating prevalence of HIE use. The denominator included all dental visits during the study period that did not have a co-occurrent medical visit or a medical visit in the two days prior. We describe these visits henceforth as “isolated dental visits.” For our analysis of patient and visit factors associated with HIE use, our independent variables included patient age at visit, patient sex (female, male), year (2012, 2013, 2014, 2015), FQHC site (OOCHC; AJHC), whether the patient had a chronic disease (yes, no), and days since the patient’s last visit to the FQHC (no previous visit, within last month, within last 3 months, within last 6 months, within last year, over a year ago). We also created binary indicators for all ICD-9/10-CM diagnoses codes specifically grouped under “Diseases of the oral cavity, salivary glands, and jaws”. To measure the nature of HIE use, we created binary variables for each of the HIE portal section tabs (e.g. summary or lab) and categories of documents (e.g. ADT alert, lab result, radiology report) that HIE users accessed, to identify the information viewed during dental visits.

Analysis

Descriptive statistics and the prevalence of HIE use among dental visits were estimated, using all isolated dental visits during the study time frame as the denominator for calculations. Bivariate and multivariable regression analyses were conducted to examine patient and visit characteristics associated with HIE use for dental visits. We used Chi-square tests for bivariate analyses and multivariate logistic regression for our adjusted analyses. We used robust standard errors clustered at the FQHC level and included site and year fixed effects to adjust for time-invariant differences across the study sites and secular trends. We report average marginal effects for this model and 95% confidence intervals. Finally, we calculated the most frequently viewed section tabs and documents within the sample of dental visits with HIE use. We calculated the proportion of isolated dental visits in which the HIE user viewed each available section tab and document type, to identify the most common clinical data categories viewed in the HIE for dental visits.

All data preparation and management, construction of HIE use measures, and analysis were done in the RStudio development environment using the R statistical programming language.²³ The tidyverse suite of packages and data.table package were the primary software libraries used to construct the analytical data file. The mfx package was used in the regression analysis.

Results

Our final study sample included 97,460 dental visits across 21,444 unique patients. A total of 164 visits (0.17%) included use of the HIE query portal. Table 1 contains the results of our bivariate analyses examining the unadjusted relationship between HIE use and site, patient, and visit characteristics. Table 2 presents bivariate analyses. The rate of HIE use differed across FQHC, years, days since last visit, and certain dental disease diagnoses.

Table 1.

Site, patient, and visit characteristics associated with health information exchange use during dental visits (2012-2015).

		HIE Use	No HIE Use	p-value
Number of Dental Visits		164	97,296
FQHC Site*** (%)				<0.001
	Anthony Jordan Community Health Center	39 (23.1)	60,853 (62.5)
	Oak Orchard Community Health Center	125 (76.2)	36,443 (37.5)
Patient Sex (%)				0.788
	Female	94 (52.8)	57,070 (58.7)
	Male	70 (47.2)	40,226 (41.3)
Age at Visit (%)				0.37
	18 to 25	8 (7.2)	15,077 (14.2)
	26 to 64	90 (81.1)	80,580 (75.7)
	65 & Older	13 (11.7)	10,720 (10.1)
Chronic Disease (%)		2 (1.2)	495 (0.5)	0.466
Diseases of the oral cavity, salivary glands, and jaws
	Disorders of tooth development and eruption*	41 (25.0)	32,396 (33.3)	0.03
	Diseases of hard tissues of teeth or dental caries	0 (0.0)	109 (0.1)	1.00
	Diseases of pulp & periapical tissues	16 (9.8)	12,701 (13.1)	0.256
	Gingival & periodontal diseases	1 (0.6)	953 (1.0)	0.933
	Dentofacial anomalies*	9 (5.5)	11,271 (11.6)	0.021
	Diseases of the jaws	0 (0.0)	55 (0.1)	1.00
	Diseases of the salivary glands	15 (9.1)	7,399 (7.6)	0.551
	Diseases of the oral soft tissue	0 (0.0)	55 (0.1)	1.00
	Diseases of the tongue	0 (0.0)	1 (0.0)	1.00
Days Since Last Visit**				0.004
	No Past Visit	63 (38.4)	23,970 (24.6)
	Greater than 1 Year	5 (3.0)	4,158 (4.3)
	Previous year	16 (9.8)	11,678 (12.0)
	Previous 6 months	11 (6.7)	7,710 (7.9)
	Previous 90 days	24 (14.6)	14,863 (15.3)
	Previous month	45 (27.4)	34,917 (35.9)
Year***				<0.001
	2012	14 ( 8.5)	23516 (24.2)
	2013	38 (23.2)	24488 (25.2)
	2014	65 (39.6)	24381 (25.1)
	2015	47 (28.7)	24911 (25.6)

Appendix.

Logistic Regression Results (Marginal Effects) of select site, patient, and visit characteristics associated with HIE use during dental visits.

		Outcome: Any HIE Use During Dental Visit
		AME	95% CI Low	95% CI High	p-value
Sex (ref: Female)
	Male	0.000020	-0.000419	0.000459	0.92803
Age (years)		0.000005	-0.000012	0.000023	0.55851
Chronic Disease***		0.000271	0.00014212	0.000400	p<0.001
Previous Visit Timing (ref: more than 1 year)
	No Past Visit***	0.001340	0.000728	0.001953	p<0.001
	Previous Year	-0.000256	-0.001076	0.000564	0.54012
	Previous 6mo	-0.000116	-0.000580	0.000348	0.62381
	Previous 3mo	0.000317	-0.000505	0.001139	0.45021
	Previous 1mo	0.000243	-0.000086	0.000572	0.14733
Periodontal Diseases
Disorders of tooth development and eruption***		-0.000942	-0.001090	-0.000794	p<0.001
Diseases of hard tissues of teeth or dental caries		0.000120	-0.001485	0.001726	0.88307
Diseases of pulp & periapical tissues		-0.000238	-0.001569	0.001094	0.72645
Gingival & periodontal diseases***		0.000098	0.000059	0.000138	p<0.001
Dentofacial anomalies***		-0.000936	-0.001083	-0.000788	p<0.001
Diseases of the jaws***		-0.000936	-0.001083	-0.000788	p<0.001
Diseases of the salivary glands***		-0.000930	-0.001076	-0.000783	p<0.001
Diseases of the oral soft tissues***		-0.000934	-0.001080	-0.000787	p<0.001
Diseases of the tongue***		-0.000938	-0.001088	-0.000789	p<0.001
Site (ref: AJHCa)
	OOCHCb***	0.002317	0.002165	0.002469	p<0.001
Year (ref: 2012)
	2013***	0.001593	0.001444	0.001742	p<0.001
	2014***	0.002617	0.002038	0.003196	p<0.001
	2015*	0.001919	0.000365	0.003473	p<0.05

^aAnthony Jordan Health Center, New York

^bOak Orchard Community Health Center, New York

In adjusted analyses, HIE use was not associated with patient sex or age. However, the probability of HIE was 0.02 percentage points higher for patients with a chronic medical condition (p<0.001), holding all other characteristics constant. Among visit characteristics, the probability of HIE use was 0.13 percentage points higher if the visit was also the patient’s first visit to the FQHC, compared to patients with previous visits more than a year prior (p<0.001). HIE use was more likely for dental visits that included diagnoses of gingival and periodontal diseases (0.01 percentage points, p<0.001). The probability of HIE use was 0.09 percentage points less if the dental visit included any diagnoses related to the following 6 categories of ICD-9/10-CM diagnoses codes: “Disorders of the tooth development and eruption”, “Diseases of the tongue”, “Diseases of the salivary glands”, “Diseases of the oral soft tissues”, “Diseases of the jaws”, and “Dentofacial anomalies” (all p-values <0.001). Figure 2 presents the results of our multivariable logistic regression. (Appendix includes full regression results).

Figure 2.

Marginal effects of site, patient and visit characteristics on the rate of HIE use during dental visits.

In our analysis of HIE information viewed for dental visits, the most frequent sections of the HIE viewed were the summary tab page (n=85; 52.4%), the laboratory tab (n=47; 28.7%), and the radiology tab page (n=21; 12.8%) (Figure 3). The most frequent HIE documents viewed included patient index documents (n=36; 22%), laboratory document (n=31; 18.9%), %), and radiology documents (8.5%). All other document types accessible in the HIE were viewed in less than 10% of the visits (Figure 4).

Figure 3.

Sections tabs viewed in the HIE portal of the Rochester RHIO during dental visits (n=164).

Figure 4.

Documents viewed in the HIE portal of the Rochester RHIO during dental visits (n=164).

Discussion

We measured the prevalence of query-based HIE use related to dental visits and examined site, patient, and visit characteristics associated with that HIE use. Finally, we analyzed what HIE information was most frequently assessed. Overall, the rate of HIE use during dental visits is 0.17%. While this rate is low, low rates of query-based HIE use are consistent with other studies of safety-net providers and settings, which find HIE use is between 2.3 and 3.1% of visits.^24,25 Query-based HIE is not the only form of potential interoperability between providers. Query-based HIE is complementary to directed HIE, which allows providers to “push” or send data (such as structured documents, images, and laboratory results) to another provider.²⁴ Given regulatory incentives for directed HIE, it is likely that dentists engage in this form of HIE more often than query-based HIE, as is the case in primary care practices.²⁴ Thus low HIE use in this study may be related to providers selecting to use directed HIE to share data or seek information. Low query-based HIE use may also be partially explained by FQHC dental providers’ existing access to each patient’s EHR data, which may meet most of providers’ clinical health information needs. With access to the FQHC’s health record of the patient, the utility for query-based HIE among these dental providers would be reduced. Therefore, our estimate of dental HIE use is likely to be conservative, compared to what might be observed in less integrated clinical contexts. Convenient access to medical data contained in an EHR is a rare asset for many dentists, who generally operate in solo or small group practices that are not co-located or integrated with medical care. Very few private practicing dentists share data with larger medical networks.²⁶ Therefore, the HIE use we observe in these two FQHC settings goes beyond information needs contained within an EHR and provides important insights into the information needs related to dental visits that are unlikely to be met via clinical information integration alone. Additionally, our findings of overall low prevalence of HIE use underscore the importance of integrating outside information into EDR workflows, via directed or query-based methods.

For instance, we found HIE use was more likely for patients with no previous visit at the FQHC, and therefore no available medical EHR data. We also found that HIE use was more likely for patients who had a chronic disease. Given the relationship between oral and systemic health⁷, this finding fits with dental-medical integration models which emphasize information sharing in particular for complex patients, such as those managing chronic diseases. Among certain diagnoses associated with diseases of the oral cavity, salivary glands, and jaws, we found only one category of diagnoses codes under the heading “Diseases related to the oral cavity, salivary glands, and jaws” related to a higher likelihood of HIE use – “Gingival and periodontal diseases”. While additional qualitative research is needed to explore the motivations underlying this finding, greater HIE use in this context may be rooted in known associations between periodontal disease (which begins as gingivitis) and numerous chronic diseases.^6,7,27 The exact periodontal mechanisms behind these disease relationships remains unknown. However, increased HIE use when gingival and periodontal diseases are present may indicate information seeking for other associated chronic diseases.

Future participation in HIEs by dental providers will require well thought-out workflow design and information need considerations. These findings provide useful information for policy makers, healthcare administrators, and HIE designers and indicate summary information, laboratory results, and radiology to be the most frequently accessed elements of a HIE during dental visits. These particular features of an HIE may be of high value to dentists and should be considered when designing HIE portals for use related to dental care. Customizing the HIE portal interface based on the information most pertinent to certain types of providers in particular settings may increase use. Finally, recent federal efforts to encourage HIE have resulted in state-level efforts to further integrate data across clinical settings, including dentistry. As an example, North Carolina has implemented a state designated HIE (NC HealthConnex) that requires Medicaid provider participation and sharing of electronic health data in order to receive Medicaid reimbursements for services.²⁸ These requirements apply to all health providers seeking Medicaid payments, including dentists, who must meet NC Health Connex’s expectations for participation by 2021²⁹.

This study is novel as the first study to explore HIE use related to dental visits. Nevertheless, our findings should be interpreted in the context of several limitations. First, our setting includes two FQHCs with on-site dental services, which serve a low-income population more likely to have multiple medical needs or chronic diseases than the general population. Second, our analysis focuses on one community HIE system, so the generalizability of our findings to other query-based HIE portals is limited. The unique display and interface of the Rochester RHIO may impact usage behavior and patterns. In addition, the characteristics of the users accessing the HIE during dental visits are unknown and therefore we were unable to determine the impact of a user’s level of computer skills and their perception of the value of HIE use for dental related services.

Despite these limitations, our study makes important contributions to the literature. We study sought to close a gap in the HIE and interoperability research, which rarely explores the value of medical information to other health services and providers such as dentists. Future research should explore HIE use related to dental visits among additional patient populations and settings, such as private or group dental practices commonly separated from medical care or within hospital emergency department rooms. Qualitative analyses of any perceived value of HIE to dental providers is also needed to further inform HIE design. Further, given the Centers for Medicare and Medicaid (CMS) Merit-based Incentive Payment System (MIPS) 2019 Promoting Interoperability (PI) performance measures, which require bi-directional provider-to-provider exchange, research is needed to explore direct health information exchange among dental providers and associated meaningful use measures tied to this form of exchange^30,31.

Conclusion

Dental visit related HIE use in FQHCs is low but is more common for patients with chronic diseases and periodontal disease, as well as for patients unfamiliar to the dental site. Laboratory and radiology information were the most common information types explored in the HIE for dental visits. Future work should explore the motivations underlying HIE use among dental providers to understand the most important aspects of HIE for dental providers and in turn inform system implementation and design.