Considerations for US coverage and reimbursement: Which digital health products to evaluate? Findings from a modified Delphi study

Abstract

BACKGROUND:

The rapid growth of digital health tools, including digital applications, wearables, sensors, diagnostics, digital therapeutics (DTx), and prescription DTx, offers new ways to treat patients and close gaps in care. Payers need transparent, credible, and efficient processes to differentiate products for potential reimbursement from the larger universe of digital health products.

OBJECTIVE:

To identify areas of agreement, disagreement, and rationale for payers to determine which digital health products should be evaluated for formulary consideration and to develop generalizable criteria for health care decision-makers developing policies and approaches for digital health products.

METHODS:

Experts from the Academy of Managed Care Pharmacy DTx Advisory Group Payer Evaluation subcommittee rated whether a pharmacy and therapeutics committee, health technology assessment group, or an innovation center within a health plan or pharmacy benefit manager should consider 14 hypothetical products for potential formulary coverage. Using a 4-step modified Delphi approach, experts rated whether it was appropriate for a payer to evaluate each product on a scale of 1 (strongly disagree) to 9 (strongly agree). Quantitative agreement was assessed using terciles of responses, medians, and the distribution of appropriateness scores. The corresponding discussions are summarized to identify generalizable criteria for payers to consider as they develop approaches to determine which digital health products to evaluate.

RESULTS:

Among the 14 hypothetical products, 4 achieved quantitative agreement that payers should evaluate the product. 5 products had quantitative disagreement, and the remaining were indeterminant. Payers were most likely to review a product if it (1) was reviewed by the US Food and Drug Administration, (2) required a prescription, (3) was intended to be paid for using premium dollars, (4) treated rather than diagnosed or monitored a clinical condition, (5) had a low clinical opportunity cost, and (6) could address population health metrics.

CONCLUSIONS:

The rapid availability of digital health and DTx options can be daunting for health care decision-makers when determining which products to evaluate. These generalizable criteria can help payers develop a more efficient process.

Plain language summary

An increasing number of digital health products, including applications, wearables, sensors, digital therapeutics, and prescription digital therapeutics, are in development. Using 14 hypothetical product profiles and a consensus process, we identified areas of agreement (and disagreement) and corresponding rationale to identify 6 generalizable criteria for payer organizations to consider when developing approaches for selecting digital products for evaluation for coverage and reimbursement.

Implications for managed care pharmacy

Payers need to differentiate digital therapeutics with documented clinical benefit from the universe of available digital health applications. These criteria help create a more transparent, credible, and efficient formulary and policy development review process. Greater clarity on which products are likely to be considered by payers can help digital therapeutics manufacturers develop market access strategies and more effectively plan for evidence needs and coverage and reimbursement consideration.

The emergence of digital health technologies (including digital diagnostics, sensors, and therapeutics) can alter health care delivery, improve patient outcomes, lower costs, and close gaps in care. However, the rapid market growth and new categories of digital health products present unique challenges for payers as they conduct horizon scanning for emerging products to accurately forecast and set rates for benefits, premiums, and formulary coverage.^1,2

Digital products can vary in purpose, evidence standards, and commercialization approaches. For example, more than 350,000 digital health tools are available as wearables and for download as mobile applications for nutrition, fitness, or medication wellness. These products are not regulated by the US Food and Drug Administration (FDA) and may have limited clinical evidence.³ In contrast, digital therapeutics (DTx) are evidence-based therapeutic interventions driven by high-quality software programs to prevent, manage, or treat a medical disorder or disease.^2,4 Within the DTx category, a subset of products are reviewed by the FDA for safety and efficacy claims and are available as prescription DTx (PDTs).⁵ Although fewer than a dozen PDTs are marketed in the United States, more than 100 products are in development.³

Payers seek ways to evaluate and optimally use digital products to improve patient engagement and value-based care.⁶ Some payers and pharmacy benefit managers (PBMs) have developed digital formularies, which include applications for diabetes, smoking cessation, and insomnia. Other health plans and state Medicaid organizations have announced coverage for specific PDTs.^7-10 As found by a recent survey, more than 56% of payer respondents affirmed they were developing or were interested in developing an approach for digital health products and DTx coverage in the next 18 months.¹¹

Without an efficient way to differentiate between products payers should evaluate based on the likelihood of providing clinical benefits from the universe of digital health applications, payer organizations will experience infrastructure and staff capacity challenges.¹² Likewise, without clarity regarding which products payers are likely to consider for reimbursement, digital developers face greater uncertainty in product adoption and the incentives for innovation. Generalizable criteria that identify where there is agreement that payers should clinically evaluate a digital health product for potential inclusion on a formulary or coverage similar to a prescription drug could be useful. We also sought to understand areas of disagreement and the corresponding rationale.

Methods

Contemporary issues associated with digital health products were gleaned from the 2021 Academy of Managed Care Pharmacy (AMCP) Partnership Forum on DTx and AMCP committees regarding digital health needs. AMCP staff created 14 hypothetical product profiles varying 6 key characteristics to elucidate product characteristics warranting evaluation. These characteristics included (1) the type of digital health product (eg, digital health application, DTx, or PDT), (2) purpose (eg, screening, monitoring, or treating a disease), (3) whether the FDA reviewed the product, (4) if a prescription was required, (5) the clinical condition treated (eg, bipolar disorder, asthma, or cancer), and (6) the product’s expected place in therapy (ie, an adjunct to or in place of existing treatments) (see Supplementary Exhibit 1 ^{(101.7KB, pdf)} , available in online article).

We used a 4-step, modified Delphi approach to identify aggregate opinions on which hypothetical product profiles payers should further evaluate for formulary coverage or reimbursement. This included an initial round of voting by experts, group discussion, a second round of voting, and a final group discussion. Before the second vote, experts were presented with anonymized results from voting in round 1, initial discussion comments, and their individual responses. This modified Delphi approach has been used as an iterative method to collect and aggregate informed judgments from experts, provide greater clarity on areas of agreement, and identify where uncertainty remains.^13-15

The AMCP DTx Advisory Group Payer Evaluation Subgroup served as experts for the modified Delphi panel. The Advisory Group included senior pharmacy and medical directors actively involved with their organizations’ DTx efforts. These 16 individuals were employed as PBMs (5), by regional health plans (5), as employer benefit consultants (1), as subject matter experts related to digital health technologies (3), or by DTx or biopharmaceutical manufacturers (2) (see the Acknowledgments).

Advisory group members rated on a scale from 1 (strongly disagree) to 9 (strongly agree) the statement that a “payer (ie, health plan or PBM) should evaluate this product for inclusion on their formulary or coverage similar to a prescription drug.” Panelists considered evaluation to be conducted by any payer evaluation body (eg, a pharmacy and therapeutics committee, a health technology assessment group, or an innovation center). They were asked to remain agnostic to the benefit category (eg, pharmacy, medical, or digital health benefits) because these may differ by organization and line of business. All voting was done using SurveyMonkey or meeting polls. Discussions occurred between May 2022 and September 2022 and were transcribed for thematic analysis.

Quantitative agreement was determined using a 2-step process. We first calculated median appropriateness ratings and grouped scores into terciles (7-9 agree, 4-6 neither agree nor disagree, and 1-3 disagree). Next, we calculated quantitative agreement based on 75% within the same tertile. Disagreement was assessed if there were 4 responses at either end of the terciles. Indeterminate agreement was concluded if neither criterion was met. We presented final ratings as individual profile results by terciles and quantitative agreement as determined by median scores and the distribution of scores. For key product characteristics (eg, FDA review vs no FDA review), we presented the mean and SDs of the individual median scores and thematically summarize the discussions.

Results

Across the 14 product profiles, median scores ranged from 1 (strongly disagree) to 9 (strongly agree) with the need to evaluate the product profile. Four profiles (29%) reached quantitative agreement for evaluation. There was quantitative disagreement (eg, scores were in both the upper and lower terciles) on the need to evaluate 5 product profiles (36%, median scores 2-8). The need to review was indeterminant, with panelists neither agreeing nor disagreeing for 5 product profiles (36%, median scores 1-8) (Figure 1). The discussions identified the following thematic considerations (Table 1).

FIGURE 1.

Round 2 Results: Agreement That Payers Should Evaluate

TABLE 1.

Considerations for Payer Organizations Evaluating Digital Health Products for Coverage and Evaluation

Consideration	Supporting considerations
1. Has the product been reviewed by the FDA?	Was the product cleared, authorized, or approved? Was the product not solely approved based on the predicate device?
	Is there peer-reviewed evidence of clinical benefit?
2. Does the product require a prescription?	Are prescription status and payer coverage part of the market commercialization strategy?
3. Will premium dollars be used to pay for this product?	Is payer coverage part of the commercialization strategy?
	Is the product likely to reduce the total cost of care?
4. Does the product treat rather than track or diagnose a condition?	Is the product intended to be used in place of, or as an adjunct to, standard of care?
	Does the product improve the outcomes or reduce the total cost of care if used to guide treatment options?
5. Is the clinical opportunity cost low?	Is the product likely to have a moderate treatment benefit with limited risk?
	Is the product scalable?
	Are there few available or proven effective treatment alternatives?
	Is there sufficient evidence for evaluation? If not, is there an opportunity to invest in a pilot program to learn about the product use among a subset of plan members?
6. Can the product address population health metrics?	Can it improve the patient or provider experience?
	Do patient outcomes differ by health literacy, digital literacy, language, cultural factors, or demographics?

^{Although no algorithm or threshold for evaluation was recommended, answering “yes” to a greater number of questions aligns with a greater agreement regarding the need to evaluate the digital health product, digital therapeutic, or pr e scription digital therapeutic.}

^{FDA = US Food and Drug Administration.}

CONNOTATION OF FDA REVIEW

FDA clearance, authorization, or approval was a signal to evaluate DTx for some payers. Hypothetical product profiles that referenced FDA review, clearance, or authorization had higher scores than products that did not (7.8 ± 2.1 vs 2.6 ± 1.6, respectively). However, panelists described 2 scenarios in which FDA review may be insufficient. First, products based on a predicate (eg, a legally marketed device) only require equivalence to the previously cleared device. Second, others noted that FDA determinations may be favorable even if the evidence is below an established minimal clinically important difference. As one panelist noted, “We don’t assume that just because a device was authorized or cleared by the FDA, a payer body would just add it to the formulary…But that stamp of approval is a big part of evaluation criteria.”

SIGNIFICANCE OF PRESCRIPTION STATUS

Prescription products were more likely to be evaluated than nonprescription products. Mean scores across product profiles requiring a prescription were 8.5 ± 0.6 compared with 4.2 ± 3.0 without. In the group discussion, panelists noted they were more likely to critically assess a product that requires a prescription or access code from a dispensing pharmacy to activate the digital product. The experts also expressed challenges with staying up to date on applications available on Google Play or the Apple App Store and managing access to consumer-focused digital health tools. As summarized by one panelist, “Health plans do not evaluate, nor do they pay for all direct-to-consumer products.”

IMPORTANCE OF PRODUCT INDICATION AND INTENT

The product’s place in care may also play a role in the decision to evaluate. Products indicated to treat a condition tended to be reviewed more often than product profiles that tracked symptoms or screened for medical conditions (8.0 ± 2.2 vs 7.2 ± 3.3 and 6.8 ± 2.7, respectively). Several panelists noted that diagnostics are typically used only once as directed by the provider; thus, payer evaluation was unnecessary. Products indicated for screening were not likely to be evaluated unless the product aided treatment selection, improved patient outcomes, or reduced the total cost of care.

In addition, understanding if the product was intended for use as an adjunct to other treatments or as a standalone treatment potentially replacing other drugs or medical interventions was important. For example, when treating mental health disorders, does the product serve as an adjunct to cognitive behavioral therapy or directly affect the neuromodulation of chemicals in the brain? Panelists noted the need to evaluate if the product replaced an existing therapy or was used as an adjunct treatment.

UNCERTAINTY AND CLINICAL OPPORTUNITY COST CALCULATIONS

Beyond the individual product characteristics, some panelists described a set of considerations focused on risk and the clinical opportunity cost. As one panelist noted, higher-risk products were more likely to be evaluated than lower-risk products. In addition, when alternative treatments with established benefits exist, the product was more likely to be assessed to ensure the wise use of premium dollars. Products easily scaled or used to address unmet needs were also considered more likely to be reviewed.

In addition, panelists noted they needed evidence to evaluate a product. Panelists perceived products with evidence submitted to the FDA or published in a peer-reviewed journal to have greater quality and rigor than “self-published” information or information on a website. Without sufficient evidence, the panelists were uncertain of benefits and were less likely to review these products.

POTENTIAL TO IMPROVE POPULATION HEALTH OUTCOMES METRICS

Finally, the potential for products to improve population health outcomes was a priority for selecting products for evaluation. One of the primary benefits of digital products is the potential benefit to close gaps in care and address health disparities. However, as some panelists noted, these products may have unintended consequences or may exacerbate health inequities if desktop or web access was required.

Products that enhance patient or provider experience, medication adherence, and care coordination were also prioritized for evaluation. For example, one panelist noted that some self-insured employers may be interested in products that support employee engagement, experience, and wellness. Another panelist shared that if a nonprescription digital health product had higher-quality evidence, addressed a common condition (eg, diabetes), and was scalable, it may have the potential to make meaningful improvements in population health outcomes and could be worth premium dollars. Products that address population health outcomes were more likely to be prioritized for evaluation.

Discussion

We sought to identify generalizable criteria for health care decision-makers developing policies and to determine which digital health products (including digital applications, wearables, sensors, diagnostics, DTx, or PDTs) should be evaluated for coverage and reimbursement. Rather than serving as an algorithm for horizon scanning or a threshold for evaluation, these considerations can help payer organizations assess the growing digital health pipeline of apps and therapeutics, a critical step to ensure digital health products do not overwhelm existing payer staff and resources.

We identified areas of agreement and disagreement regarding whether payers would evaluate the 14 hypothetical product profiles. The areas of agreement and disagreement may be related to stakeholder type (eg, PBM vs regional plan); however, we had an insufficient sample size for more granular analysis. Another panelist suggested that the differences may reflect each organization’s approach to DTx. For example, some organizations consider digital health products appropriate for inclusion in the overall formulary. Other organizations offer digital health coverage as a supplemental product that clients can purchase.

The group also discussed the ability to make the considerations actionable across organizations. The group recommended against creating an algorithm or threshold to determine which products to evaluate. This allows flexibility for organizations to tailor the process to meet their organization’s priorities and goals.

LIMITATIONS

This study has several limitations. First, a different mix of experts with different responsibilities (eg, pharmacy vs medical directors) or organizational backgrounds (eg, PBMs vs regional plans) might have rated these product profiles differently. Second, the product profiles were hypothetical and included potential product characteristics not reflected in marketed products (eg, no FDA authorization or clearance but requiring a prescription). Third, panelists voted on whether a product should be evaluated and not whether it should be covered. As several panelists noted, the evaluation of a product does not equate with reimbursement. Fourth, Delphi approaches often reduce disagreement across rounds of voting. Our experience differed and may reflect the nuances of the cases, considerations raised during discussions, or evolving strategies for DTx. Fifth, panelists cautioned against using these findings to create an algorithm or quantitative threshold for identifying products for evaluation. Instead, they highlighted the opportunity for this effort to provide a generalizable and flexible framework to aid consistency across reviewers, therapeutic areas, and points in time as new digital health products are developed. Finally, this effort focused on whether payers should evaluate the product profile. Efforts to describe how they should evaluate the products for evaluation were addressed in later phases of the Advisory Group.

Conclusions

Given the rapid market growth and availability of digital health and DTx options, it can be daunting for health care decision-makers to determine which products to evaluate for potential coverage. A panel of experts identified 6 generalizable characteristics for consideration. Products should be evaluated if they (1) were reviewed by the FDA, (2) required a prescription, (3) were paid for using premium dollars, (4) treated rather than monitored or diagnosed a clinical condition, (5) had a low opportunity cost, or (6) addressed population health metrics. These considerations can serve as a starting point for organizations determining their approach to digital health products. A transparent, credible, and efficient process to differentiate the DTx and digital health products most likely to improve health, reduce costs, and close gaps in care will be needed in a world that will increasingly include hundreds of thousands digital health products. Together these considerations may help payers identify emerging products, allocate resources to better plan design benefits, and accurately forecast rates.

Funding Statement

The AMCP Digital Therapeutics Advisory Group, the Payer Evaluation Subgroup, and the development of the research brief were supported by Akili Interactive, Mahana Therapeutics, metaMe Health, Otsuka, Pear Therapeutics, Pfizer Inc, PhRMA, and Xcenda.