Cost and Workflow Impact of a Primary Care Based Multi-Level Pediatric Oral Health Intervention

Abstract

Objectives:

A recent trial demonstrated the effectiveness of a primary care-based multilevel intervention to increase dental attendance in 3- to 6-year-old Medicaid-insured children. We estimate the cost and workflow impact for real-world practices wishing to implement this intervention.

Methods:

Intervention practices from the trial integrated oral health (OH) questions into their electronic medical records (EMR). Providers received theory-based training on delivering OH education and provided “prescriptions” for dental visits and a list of Medicaid-accepting dentists. EMR enhancement and training costs were estimated by applying nationally-representative, role-specific hourly labor costs to reported time spent by study participants performing each activity. Study staff timed the OH portion of 2 to 3 randomly selected encounters per provider.

Results:

Twenty-eight providers from 9 intervention pediatric practices participated. The percentage of Medicaid patients in these practices ranged from 22% to 86%. Practices corresponding in size to the smallest, median, and largest in the intervention group can expect first-year implementation costs of $579.79, $863.86, and $1482.15, respectively, with subsequent annual maintenance costs of $167.11, $451.18, and $1069.47. Encounter time for the intervention averaged 38 s longer than for standard care (control group).

Conclusions:

Implementation of this effective pediatric OH intervention appears to entail modest costs and lengthen encounters minimally.

Introduction

By school age, 21.8% of U.S. children will have dental caries; more than one-third of these cases will be undiagnosed. ¹ Among 2- to 5 year-old children living in households whose income is below 200% of the Federal Poverty Level (FPL), prevalence rises to 29.6%, and children in this category are more than twice as likely to have untreated caries relative to their counterparts in wealthier households (13.9% vs 6.0%) ² Beyond income, significant racial and ethnic disparities in prevalence and treatment rates are observed in this age group. In fact, Black/African American and Mexican American children are more than twice as likely as White children to suffer untreated dental caries (14.8%, 15.1%, and 6.1 %, respectively) in this age group. ²

Among 3- to 5 year-old children enrolled in Medicaid, only an estimated 48% receive recommended preventive dental services. ³ This gap has spurred intervention studies in community^4,5 and school^6-8 settings and has motivated the National Institute for Dental and Craniofacial Research (NIDCR) to launch the Multidisciplinary and Collaborative Research Consortium to Reduce Oral Health Disparities in Children. ⁹ This initiative has funded 4 community trials of multilevel interventions to promote pediatric oral health (OH) and improve provision of preventive and restorative care in vulnerable pediatric populations. One of these studies, Pediatric Providers Against Cavities in Children’s Teeth (PACT), was a cluster randomized trial among 18 Northeast Ohio pediatric practices that investigated the effectiveness of a practice- and provider-level intervention to improve delivery of education and OH screening services for 3- to 6-year-old children enrolled in Medicaid. Practices in the intervention arm of the trial integrated OH assessment questions into their electronic medical records (EMR), and their providers were trained to deliver Common-Sense Model of Self-Regulation (CSM) ¹⁰ theory-based OH facts during well-child visits (WCV), a prescription to go to the dentist (informal referral) and a list of Medicaid-accepting local dentists. Children in intervention practices had 34% (OR = 1.34, 95% CI = 1.07, 1.69) higher odds of dental attendance resulting in sealants and/or fillings compared to controls. Additionally, a reduction in the mean number of teeth with untreated decay (OR = −0.27, 95% CI = −0.56, 0.02) was found in the intervention group compared to controls. ¹¹

While the intervention appears effective, it is important to consider the practical implications of implementing such an intervention in a real-world practice setting where resources (financial and human) are often tightly constrained. Here, we use PACT trial data to estimate the expected real-world implementation cost and workflow impact for practices wishing to implement the PACT intervention.

Methods

The present study stems from the recently completed PACT cluster randomized clinical trial ¹¹ (ClinicalTrials.gov Identifier: NCT03385629) examining the effectiveness of a multi-level intervention (practice and provider levels) to increase dental utilization among 3- to 6-year-old children enrolled in Medicaid. From November 2017 to August 2019, 18 participating Ohio primary care pediatric practices in Cuyahoga County, Ohio (home to Cleveland) and the adjoining rural and suburban counties of Lorain, Medina, and Ashtabula were recruited and randomly assigned to either the intervention or a control group. Within each practice, all pediatricians and nurse practitioners (NPs) with 2 or more days per week of clinical practice participated. Dyads consisting of English-speaking adult caregivers and their 3- to 6-year-old, Medicaid-enrolled children attending WCVs at these practices were also recruited to participate. Participating children had to be free from serious medical/behavioral conditions that would preclude participation in the dental screening performed as part of study data collection. Participating parent/child dyads were followed to completion of 3 WCVs.

The PACT Intervention

The 9 practices assigned to the intervention arm of the PACT trial implemented practice-level changes by integrating 4 OH assessment questions with yes/no responses (does the child have white or brown spots, did the child see a dentist in the past 12 months, did the provider communicate oral health facts to caregiver, did the provider give caregiver a prescription for and list of dentists in the area?) into the EMR. Providers in these practices received CSM theory-based training to provide core OH facts ¹¹ about the chronicity of dental caries to parents/children during well-child visits as well as a “prescription” (informal referral) to visit the dentist and a list of local dentists accepting Medicaid. Providers attended in-person sessions consisting of a 1-h didactic training and a 1-h skills training with standardized patients prior to enrollment of parent-child dyads into the RCT. A subsequent 1-h in-person booster didactic session was conducted in year 2. For purposes of estimating real-world implementation costs, we assume that, in practice, all trainings (initial and booster) will be delivered through an online toolkit (that has been developed by the study team) featuring the educational curriculum and a video simulation of a provider delivering the OH messaging (total duration of 1 h), plus downloadable educational materials and a dental visit prescription template. Because we assume that future trainings will be delivered asynchronously online at no charge, no costs for trainer time are included in the analysis.

PACT trial control practice providers attended a 1-h didactic training session covering basic oral health education based on American Academy of Pediatrics (AAP) recommendations, ¹² followed by a 1-h booster didactic training in the second year.

Measures

For each practice in the intervention group, we report the number of providers, practice setting (academic medical center [AMC], AMC-affiliated private practice, or non-AMC-affiliated private practice), proportion of practice patients enrolled in Medicaid, and Area Deprivation index (ADI) ¹³ of the census block group where the practice is located. The ADI is a commonly-used index of social deprivation incorporating 17 separate factors available from U.S. Census data which cover domains of education, employment, income, housing (costs, crowding), and transportation access. ¹³ Higher ADI values indicate greater degrees of deprivation. ADI was normalized to a mean score of 100 based on a reference population of all Ohioans from the 2015 to 2019 U.S. Census American Community Survey. ¹⁴

For estimating real-world implementation costs to practices, we divided costs into EMR enhancement, training, and supply cost categories. It was assumed that time spent on EMR enhancement and training would result in a corresponding amount of lost patient care time and revenue. Therefore, 1-time EMR enhancement and annual training costs were estimated by multiplying job category-specific hourly labor costs (wages plus a private sector fringe benefit rate of 42.0% ¹⁵ ) by the reported number of hours spent performing each activity by intervention group participants. EMR enhancement costs are based on the reported time spent by a Medical Director in each practice to create the necessary EMR template to embed the 4 oral health questions. These enhancements are assumed to be a 1-time expense accrued in the first year of implementation. Wage data was drawn from the most recent available (May 2023) U.S. Occupational Employment and Wage Statistics (OEWS) Survey from the U.S. Bureau of Labor Statistics (BLS), ¹⁶ which provides national hourly wage data for hundreds of job categories. We used median hourly wage figures for general pediatricians (Standard Occupational Code 29-1221) and nurse practitioners (Standard Occupational Code 29-1171)—the 2 types of providers participating in the PACT trial—as point estimates; we used corresponding 25th and 75th percentile hourly cost figures to establish low and high bounds, respectively. The resulting range of labor costs is included to account for regional heterogeneity in the costs of labor in different parts of the U.S.; such ranging of inputs is known as sensitivity analysis. ¹⁷ Labor costs were inflated to April 2024 U.S. dollars based on the Physician Services component of the Medical Consumer Price index (CPI) ¹⁸ Appendix 1 provides details on hourly labor cost estimation.

Supply cost estimates were based on study records of the cost of materials related to the intervention itself, ignoring research costs related to subject recruitment, data collection, or study administration. These costs were varied ±50% in sensitivity analysis to account for regional cost variation, and all supply costs were inflated to April 2024 dollars based on the Other Goods and Services component of the CPI. ¹⁴

In order to provide a spectrum of estimated costs for practices of different sizes, total implementation costs for first and subsequent years were calculated for the “Median” practice in the intervention arm (employing the median number of general pediatricians and the median number of nurse practitioners), for the smallest enrolled practice, and for the largest enrolled practice.

Workflow Impact and Implementation Considerations

Research staff recorded the time required to deliver oral health facts during observation of 2 to 3 randomly selected well-child patient encounters conducted by each participating provider (for a total of 123 WCVs). We did not estimate costs related to additional time spent by providers delivering the intervention relative to standard practice under the assumption that substituting the intervention for standard-of-care OH counseling would not reduce the number of visits completed in a day.

Results

Participating Practices

Twenty-eight providers from the 9 intervention practices participated in the original trial. ¹¹ Most practices were private clinics with an academic affiliation, most were located in census block groups with ADI below the Ohio mean, and the percentage of patients receiving Medicaid varied from 22% to 86% (Table 1). A hypothetical “median” practice was defined based on the median numbers of general pediatricians and of nurse practitioners employed in participating intervention arm practices: 2 and 1 of each, respectively.

Table 1.

Intervention Arm Practice Descriptions.

Intervention practice number	Number of pediatricians in practice	Number of nurse practitioners in practice	Census Block Group Area Deprivation Index a	Percentage of patients enrolled in Medicaid (%)	Practice setting
1	1	1	90.7	80	Non-AMC affiliated private
2	2	0	124.5	24	AMC-affiliated private
3	3	0	85.1	22	AMC-affiliated private
4	1	0	72.6	68	AMC-affiliated private
5	5	2	79.1	23	AMC-affiliated private
6	3	0	77.6	45	AMC-affiliated private
7	2	1	74.8	34	AMC-affiliated private
8	1	2	105.7	63	AMC-affiliated private
9	3	1	167.9	86	AMC-based

Abbreviations: ADI, Area Deprivation Index; AMC, Academic Medical Center.

a

Normalized to a mean score of 100 based on a reference population of all Ohioans.

Implementation and Maintenance Costs

Table 2 summarizes estimates of labor costs for 1-time EMR modifications and annual provider training. EMR modifications required a median 3.0 h across the 9 intervention practices. For each practice, modifications were performed by a physician in the role of Medical Director. For purposes of projecting future implementation costs, we assume that providers will participate in a 1-h training session during the first year and in a 1-h booster training in subsequent years as was done in the main trial. Median hourly employment costs, inclusive of fringe benefit costs and adjusted for inflation, were $137.56 and $87.41 for general pediatricians and NPs, respectively (see Appendix 1 for calculation details). Given the estimated time requirements above, the estimated 1-time cost for EMR modification was $412.68 per practice (with low and high bound estimates of $292.92 and $456.60, respectively).

Table 2.

Labor Costs.

	Staff involved	Hours	Total inflation-adjusted hourly labor cost—median (25th, 75th percentile)	Total estimated cost—median (25th, 75th percentile)
Task				Per practice (one time)	Per provider (annually)
EMR modifications	General pediatrician	3.0	$137.56 ($97.64, $153.20)	$412.68 ($292.92, $459.60)	N/A
Training	General pediatrician	1.0	$137.56 ($97.64, $153.20)	N/A	$137.56 ($97.64, $153.20)
	Nurse practitioner	1.0	$87.41 ($74.04, $97.34)	N/A	$87.41 ($74.04, $97.34)

Inflation-adjusted supply costs covering prescription pads (“prescription” to see a local dentist), flip chart and oral health fact card educational materials, and sheets listing local dentists accepting pediatric patients insured by Medicaid were estimated at $29.55 per provider per year. The low and high values used in sensitivity analysis (±50%) were $14.78 and 44.33, respectively.

Table 3 shows point estimates, and high and low estimates from sensitivity analysis, for first-year total implementation cost and subsequent annual maintenance cost for 3 different practice sizes. a “median” practice with 2 pediatricians and 1 NP, the smallest practice in the intervention arm (single pediatrician), and the largest practice in the intervention arm (5 pediatricians and 2 NPs). For the median practice, first-year implementation would cost an estimated $863.86 (ranging in sensitivity analysis from $606.57 to $996.32), with subsequent annual maintenance costs of $451.18 (ranging from $313.65 to $536.72 in sensitivity analysis).

Table 3.

First Year Implementation and Subsequent Maintenance Costs.

Practice size	First year implementation cost (EMR changes, training, supplies)	Subsequent year maintenance cost (training, supplies)
“Median” practice (2 pediatricians; 1 nurse practitioner)
Point estimate	$863.86	$451.18
Low-cost	$606.57	$313.65
High-cost	$996.32	$536.72
Smallest Practice (1 pediatrician; 0 nurse practitioners)
Point estimate	$579.79	$167.11
Low-cost	$405.34	$112.42
High-cost	$657.13	$197.53
Largest practice (5 pediatricians; 1 nurse practitioner)
Point estimate	$1,482.15	$1,069.47
Low-cost	$1,032.63	$739.71
High-cost	$1,730.56	$1,270.96

Impact of Intervention on Provider Workflow

In the intervention group, the mean time required to deliver oral health counseling was 3.51 min (SD = 2.90), 38 s longer than the 2.89 min (SD = 3.72) spent by control group providers delivering standard oral health education recommended by AAP.

Discussion

Using data collected from pediatric providers in the intervention arm of a cluster randomized trial, we have estimated the real-world implementation and maintenance costs for an effective multi-level (practice and provider levels) pediatric oral health intervention. The intervention involves a brief, theory-based training (with annual booster trainings) for providers around delivering oral health education at well-child visits for 3- to 6-year-old Medicaid-enrolled patients using flip chart pictorial visual aids. Providers were given lists of local dentists accepting Medicaid to distribute to patients as well as prescription pads for writing a “prescription” to visit a dentist. An RCT has demonstrated that this intervention likely increases dental visits for sealants and fillings and likely decreases the amount of untreated caries. ¹¹ We estimate that the first-year implementation cost and the annual maintenance cost for the intervention will be minimal. Total costs are estimated to be under $900 in the first year and under $500 in subsequent years for a median practice.

Participants in PACT intervention arm practices spent approximately 38 additional seconds (3 min 30 s total) providing OH education compared to their control arm counterparts delivering standard guideline based OH education. This compares favorably to the nearly 5 min required for a similar intervention in an Indian Health Service clinic. ¹⁹ Our results are consistent with the results of a separately published qualitative study among 21 intervention arm providers, in which participants broadly reported that delivering the intervention did not interfere with their productivity, and that incorporating it was not cost-prohibitive. ²⁰

Understanding the financial and workflow impacts of implementing a new intervention during well-child visits is critical given the numerous competing demands for precious time with patients and parents during encounters. This is particularly true when the intervention will not generate additional revenue. Previous studies describing implementation of OH educational interventions for preschool and school-aged children have not provided implementation cost information,^21-24 and few have provided information on implementation time requirements. ¹⁹

A key strength of this work is that participating practices varied greatly in their proportion of patients receiving Medicaid, and we therefore believe the results of this analysis can be applied somewhat broadly. Another strength is that we accounted for cost variation stemming from heterogeneity in provider salaries and supply costs.

Limitations

Practices implementing the intervention were mostly academically-affiliated private practices—a model made increasingly common by the intensive consolidation in the healthcare systems of many regions. ²⁵ The above-noted wide range of Medicaid penetration does offer reassurance that practices involved represent a broad socioeconomic range. Another potential limitation is that we did not estimate costs of delivering the intervention itself during clinical encounters. The intervention required 38 additional seconds beyond the current standard of care. We assume that this 38 s will be “made up” elsewhere in the encounter or during the workday. Were we to assign a cost to this time, based on the median hourly employment costs described in our results ($137.56 and $87.41 for general pediatricians and NPs, respectively), these 38 s would cost $1.45 (pediatricians) or $0.92 (NPs) per Medicaid well-child visit among 3- to 6-year-olds. Finally, we did not account for cost associated with periodic updating of lists of Medicaid-accepting dentists by office staff. This cost would be expected to be minor.

Conclusion

Here, we have described the implementation costs and workflow impact of a simple practice- and provider-level intervention to improve oral health education for 3- to 6-year-old Medicaid enrollees during well-child primary care visits. Both cost and workflow impacts appear minimal.