Benefit Design and Consumer Information: Results From a Randomized Trial

ABSTRACT

Objectives: To test whether consumer inertia in choosing primary care clinics (PCCs) in a tiered total cost of care system can be reduced by supplying tailored information on clinic cost directly to consumers during open enrollment.

Study Design: Randomized controlled trial among individuals with employment-based health insurance. Members were randomly assigned by their zip code of residence, with 100 zip codes assigned to the treatment group and 100 to the control group.

Methods: During the open enrollment period for calendar year 2020, members in the treatment group received emails providing enhanced tier information on the most popular PCCs in their zip code. We ran regression analyses to evaluate the marginal effect of the targeted intervention on consumer choice.

Results: The intervention had only a small marginal effect on choice of PCC. Before the intervention, nearly 85% of consumers were selecting PCCs in the 2 tiers with the lowest cost sharing, suggesting that tiering alone may already have addressed the problems of poor information and distorted prices inherent in many health insurance designs.

Conclusions: Providing members with enhanced tier information on local PCCs had a limited marginal effect on consumer choice, suggesting that informational interventions alone are insufficient to overcome existing inertia or, potentially, that consumers were already adequately informed through the tiered total cost of care benefit design.

Am J Manag Care. 2026;32(4):219-225. https://doi.org/10.37765/ajmc.2026.89918

Takeaway Points

A randomized controlled trial found that reducing information barriers had a limited impact on consumer selection of tiered-network primary care clinics (PCCs) among individuals with employment-based health insurance.

• Prior to intervention, nearly 85% of members selected a PCC in the 2 tiers (of 4) with the lowest consumer cost sharing.
• The intervention had only a small marginal effect on PCC choice. However, members with family coverage were more likely to move to a lower tier, possibly driven by higher potential cost savings at the household level.
• The intervention’s limited impact may be due to effective preexisting incentives for efficient choice, employee inattention, and hassle costs.

Health care costs have been rising faster than wages for many years. In 2024, family coverage premiums represented approximately 32% of median household income and large deductibles added significantly to total exposure to out-of-pocket spending on health care services.^1,2

Initiatives designed to improve the efficiency of the US health care delivery system generally fall into 2 broad categories: provider-oriented payment reforms and consumer-oriented reforms. Provider payment reforms include shared savings programs, accountable care organizations, capitation, and bundled payments. Unfortunately, provider payment reforms have produced limited reductions in spending or improvements in quality of care.^3-5 CMS has tested many provider-oriented alternative payment models. The results, thus far, have been mixed.^6,7

Provider payment reforms generally are invisible to consumers, whereas consumer-oriented reforms rely on consumer choices to bring competitive pressure to bear on health plans and providers. Consumer-oriented approaches include large deductibles, price transparency, narrow networks, preferred provider organizations (PPOs), reference pricing, and tiered cost sharing.^8-19 These initiatives vary in the information and incentives they provide to consumers. Large-deductible plans give consumers an incentive to shop for lower-cost providers but not necessarily any information on where to find them. Price transparency and reference pricing initiatives provide consumers with information on unit prices (fees) but not necessarily on provider practice styles that affect the quantity of health care services. Narrow networks and PPOs may consider only unit prices. Many health insurance benefit designs thus leave consumers facing 2 forms of market failure: poor information and distorted prices.

The problem of inertia (ie, consumers continuing to make the same choice over multiple time periods despite significant changes in their options) has been the subject of intense analysis in recent literature.^20-33 In general, these studies have found that overcoming inertia in consumer-oriented health reforms is difficult and that past attempts to improve consumer information have produced mixed results at best.^34,35

Drake et al identified 3 sources of inertia in choice of health plans and provider networks: tastes for provider continuity, inattention, and hassle costs.⁹ However, inertia is not always inefficient. Continuity of care is associated with better health outcomes.^36,37 Inattention to alternative choices can be rational if the expected gains from making an alternative choice are less than the expected cost of switching. Hassle costs include search costs, the cost of learning the details of new alternatives, risk aversion regarding unknown attributes of a new choice, and simple paperwork issues.^24,26-28

However, some inertia certainly is inefficient and can discourage innovation, new market entry, and competition among existing suppliers.⁹ In this study, we analyzed inertia in the context of a tiered cost-sharing health insurance benefit design that addresses both poor information and distorted prices. The purpose of this study was to test the marginal effect of better consumer information in that environment.

STUDY SETTING AND INTERVENTION

Study Setting

The setting for our study was the Minnesota State Employee Group Insurance Program (SEGIP). Since 2002, SEGIP has covered more than 130,000 state employees and their dependents.³⁸

During open enrollment each November, SEGIP employees choose a primary care clinic (PCC) for the upcoming calendar year. The PCC is a “gatekeeper” that is held responsible for directing patients to specialists, hospitals, and pharmacies. Thus, the PCC has both the responsibility and the authority necessary to influence the total cost of care.

Each year, PCCs are placed into 1 of 4 cost-sharing tiers based on the PCC’s total risk-adjusted per capita cost of health care over the past 2 years.³⁹ PCCs in higher tiers have higher cost sharing, and the cost-sharing differentials among tiers are substantial. For example, during the observation period for this study (2019-2020), the annual first-dollar deductible for families ranged from $500 in tier 1 to $1500 in tier 4, and inpatient hospital cost sharing ranged from a $10 co-pay to 25% coinsurance (details in eAppendix A [eAppendices available at ajmc.com]). Cost-sharing differentials and premiums between tiers are higher for employees in family plans compared with those in single plans, but employee out-of-pocket premiums are the same across all tiers.

Three health plans administered claims for SEGIP members during the observation period. A PCC can be offered by more than 1 health plan and placed into different tiers by different health plans. Thus, the members designate their desired PCC-plan combination during open enrollment.

Behavioral economics literature suggests 2 approaches to influencing consumer choices: economic incentives, such as consumer cost sharing, and choice architecture, referring to the way choices are structured and presented.^40-42 The SEGIP system is appealing from both perspectives. Information on expected total cost of care is summarized for SEGIP members through the PCC’s tier, and the varying levels of cost sharing result in members being responsible for a portion of the marginal differences in total cost of care among PCCs in different tiers. Tiering also gives PCCs a strong incentive to maintain a competitive level of total cost of care.⁴³Additional information about the SEGIP system is available in Dowd et al (2021),¹⁵ McDonald et al,⁴³ Dowd et al (2022),⁴⁴ and Huang et al.⁴⁵

The Consumer Information Intervention

The information intervention in this study focused on reducing inattention and search costs. At the time of the study, 85% of members selected clinics in tiers 1 or 2, but even so, search costs were not inconsequential for SEGIP members. They could use SEGIP’s PCC directory website to look up tiering information for every PCC in the state, but PCCs are not grouped by location. Also, the PCC directory showing the PCC tier by health plan was found on a separate webpage.

Members were randomly assigned based on their zip code of residence, with 100 zip codes assigned to the treatment group and 100 to the control group. To maximize reach, we chose zip codes with the highest member density, representing 80% of all members and tending to have more PCC choices. A map of the treatment and control zip code assignments is included in eAppendix B.

During open enrollment for the 2020 calendar year (October 30-November 20, 2019), members in the treatment group received an email reminding them to check whether their current PCC would be changing its tier, alerting them to significant cost differences among tiers, and especially highlighting the tiers and tier changes among PCC-plans that accounted for 90% of those chosen in the prior year in the member’s zip code of residence. A sample of the treatment group email is shown in eAppendix C. Design of the email followed the principles of Johnson et al,⁴⁶ including reducing the number of alternatives, partitioning the options, using clear labeling, and providing customized information. Members in the control group received the usual communication from SEGIP, reminding them to make their PCC-plan selection based on information available on the statewide website.

Improved information about PCC-plan choices could generate a wide variety of outcomes. Static changes include discovery of a more desirable PCC-plan or confirmation that their current choice is optimal. Dynamic changes include discovery that their current PCC-plan is moving to a higher tier in the coming year, leading the member to choose another PCC-plan or perhaps resigning themselves to the higher tier but possibly then choosing a different PCC-plan in the higher tier.

METHODS

Quantitative Analysis

After comparing descriptive statistics, we examined the main effect of the intervention on changes in the chosen PCCs or health plans as well as the treatment effect, controlling for several characteristics of beneficiaries, including age, sex, women of childbearing age (18-45 years), the tier of the member’s 2019 PCC-plan, whether the member chose single or family coverage, and the health plan in which the member was enrolled in 2019. Although randomization should have balanced all those additional variables, we found some statistically significant differences due in part to our large sample sizes. Single vs family coverage is important because family members can choose different PCCs but all family members must choose PCCs offered by the same health plan. The member’s 2019 health plan is important because one of the health plans is both an insurer and a care system with multiple PCCs.

We ran simple linear probability model regressions for ease of interpretation. However, we also ran logit regressions to account for the binary nature of the dependent variables, which produced only minor differences. Following Abadie et al,⁴⁷ we clustered the error terms by zip code, the unit of randomization.

RESULTS

Descriptive Statistics

Comparisons of descriptive statistics for members in the treatment and control groups are shown in Table 1. Statistically significantly fewer members were in tier 1 PCC-plans in the treatment group than in the control group, but the opposite was true for tier 4, although the numerical differences were quite small. Health plan 2, the plan that also is a care system, had slightly fewer members in the treatment group, and health plan 3 had slightly more members in the treatment group.

Approximately 85% of members selected a PCC-plan in the 2 tiers with the lowest cost sharing in 2019. That might be due to the size of PCCs in different tiers, although there were large PCCs in both the higher and lower tiers. It also could be that some of the membership in lower-cost tiers was due to manipulation of PCCs’ tiers by SEGIP. As described below, a sensitivity test that focused on PCC tier changes that were unlikely to result from SEGIP administrative changes found that the results largely were unchanged.

In an average year, 91% of PCC-plan choices were the same PCC-plan chosen by the member in the previous year, indicating a high level of inertia. Using an algorithm developed by Dowd et al,¹⁵ we constructed a choice set of PCC-plans for each zip code. Although approximately two-thirds of the members in the analytic sample could have chosen a PCC-plan combination in a lower tier in their choice set, only 4% of members in both the treatment and control groups could have chosen their preferred PCC in a lower tier in 2019 by changing their health plan.

Regression Results

Data on the control group provide a baseline for changes in PCCs and health plans. The proportions of members in the control group who changed only their PCC or health plan from 2019 to 2020 were 3.8% and 7.1%, respectively. Approximately 2% of the control group changed both their PCC and their health plan, and approximately 91% changed neither.

Table 2 and Table 3 show the effect of the intervention on the treatment group. In each table, we present the effect of the treatment alone, followed by the treatment effect controlling for several explanatory variables. Table 2 shows that the magnitude of the difference in PCC changes between the treatment and control groups shown in Table 1 is diminished when additional explanatory variables such as sex, age group, and health plan are added to the model but that the precision of the estimate is increased. The latter result was the only estimated treatment effect that achieved statistical significance at the .05 level. There were significant effects across both treatment and control groups for the member’s age. Older members were less likely to change their PCC. Members in tier 3 in 2019 were more likely to change their PCC. Members in health plans 2 and 3 were more likely to change PCCs than members in health plan 1, the largest plan.

Table 3 shows that there was no effect of the treatment on the probability that the member changed health plans. Older members were less likely to change health plans, but female members and members with family coverage were more likely to change health plans than male members and members choosing single coverage. The latter result is surprising because changing health plans must be a joint decision among family members. Members in tier 3 in 2019 were more likely to change plans, but members in health plan 2 were less likely to do so, probably reflecting the fact that leaving health plan 2 necessarily involved changing to a new health care system.

We then examined the probability that members moved to a lower tier from 2019 to 2020 (Table 4). Not all tier changes are member initiated. A PCC’s tier can change because other PCCs changed their risk-adjusted cost. SEGIP also adjusts the tiers of some PCCs to ensure that all members have access to a tier 1 or tier 2 clinic within 30 miles of their workplace. The analysis in Table 4 focuses only on member-initiated changes in the member’s PCC tier. We dropped the members whose tier changed without any change in the member’s clinic, health plan, or care system. Also, members in tier 1 in 2019 were dropped from this analysis because they could not move to a lower tier.

Again, some of the explanatory variables were significant, but the treatment effect was not. We also ran a linear probability regression investigating whether consumers moved up a tier or made any change to either clinic or health plan and found no significant effect on the treatment group.

Members with family coverage were more likely to move to a lower tier, possibly because the per-household gains from lower cost sharing were greater for them. Members in tiers 3 and 4 in 2019 who changed their PCC or health plan were much more likely to move to a lower tier in 2020.

We also interacted each demographic variable in Tables 2 and 3 with female sex and ran separate analyses as in Tables 2 and 3 for employees choosing single vs family coverage. Neither specification produced a statistically significant treatment effect.

Finally, we checked for heterogeneous treatment effects by interacting the treatment group variable with each demographic variable in Tables 2 and 3. None of the interactions were statistically significant in either table.

DISCUSSION

We studied an attempt to reduce search costs and inattention within a tiered cost-sharing health insurance benefit design that is intended to address traditional sources of market failure: poor consumer information and distorted prices. The SEGIP tiered cost-sharing design has 3 essential characteristics: It identifies a clinical entity that is held responsible for total cost of care, it tells consumers where to find lower total cost of care, and it shares the savings with consumers who choose lower-cost providers. In that environment, consumers tend to respond by choosing lower-cost providers.⁴³

We designed an email intervention to provide members with easily accessible cost information for the most popular PCC-plans in the member’s zip code, and we tested the intervention using a randomized controlled trial. We found that the intervention had only a small marginal effect on choice of PCC and no effect on the choice of health plan. It may be the case that in the presence of tiered cost sharing, tastes for provider continuity and hassle costs, particularly at the health plan level, might be more important contributors to inertia—results that are consistent with those of Higuera et al.²⁹

Before the intervention, 85% of members were selecting PCCs in the 2 tiers with the lowest cost sharing. This disproportionate share of members choosing PCCs in the lowest 2 tiers could be spurious (eg, the larger PCCs might happen to be lower cost, perhaps due to economies of scale) or may suggest that the tiered cost-sharing design already is addressing the problems of poor information and distorted prices perpetuated by other health insurance benefit designs. This could be significant for policy, as there has been much discussion about the ability of patients to behave as informed consumers due to bounded rationality and other factors.^48-50

Limitations

These analyses are subject to several limitations. First, the intervention could be considered a light-touch intervention—less direct than a phone call, for example. Myerson et al found that interventions such as phone calls are more effective than email or mail.⁵¹ However, it would be very challenging to convey the information about potential choices during a phone call. Second, the intervention was conducted with state employees in a single state and may not generalize to other types of settings. Third, the intervention sought to reduce search costs and improve attention, but it was not designed to affect hassle costs or members’ desire for continuity of care. Fourth, we were unable to include analysis of members’ utilization of health care services in our analysis, which may have influenced the costs or incentives for consumers to make changes. Fifth, the information intervention contained no information on quality-of-care measures at the PCC or health plan level. Quality information from Minnesota Community Measurement (MNCM) was available on the MNCM website but required a separate search. Finally, a more personalized intervention could have produced a greater effect. Given more resources, the information intervention could be not only specific to zip codes but also tailored to members, alerting them to the fact that their current PCC-plan’s tier would be increasing, decreasing, or available in a lower tier in the coming year.

CONCLUSIONS

Several interesting design questions are raised by these findings. First and foremost is the relative effectiveness of provider-oriented vs consumer-oriented incentives for improved efficiency. Our results are consistent with those of Whaley et al, who found that information interventions had a limited effect on consumer choice of providers unless paired with financial incentives (ie, reference pricing).⁵² Inertia often is associated with inefficiency, but when a concerted effort to improve consumer information is added to a health benefit design that already gives consumers clear provider-specific efficiency information and there is very little effect on consumer choice, it becomes more difficult to avoid the conclusion that consumers were adequately informed prior to the information intervention.

Employment-based health insurance generally covers people who are in good health. Relying on cost sharing alone to discipline the market tends to burden employees who use more services. Including premium differentials, however, influences the choices of healthy employees as well. Furthermore, premium differentials help avoid increased spending in lower tiers caused by moral hazard.

Continued research on employment-based health insurance benefit designs is justified given the increasing problem of affordable care and the large number of Americans who access health insurance through their employer.

Acknowledgments

The authors gratefully acknowledge the assistance of Michaela DeGrande, MHA; discussion and insights of Mary Steffel, PhD, and Adeniyi Togun, MD, PhD, MPH; and collaboration with the Minnesota Department of Management and Budget, especially Joshua Fangmeier, MPP, and colleagues at the State Employee Group Insurance Program, and funding support described below. Funding support was provided by the Robert Wood Johnson Foundation, The Patrick and Catherine Weldon Donaghue Medical Research Foundation, and the Minnesota Department of Revenue.

Author Affiliations: RAND Corporation (TM), Arlington, VA; University of Minnesota School of Public Health (KW, TB, BED), Minneapolis, MN; West Virginia University (TYH), Morgantown, WV.

Source of Funding: The Robert Wood Johnson Foundation and The Patrick and Catherine Weldon Donaghue Medical Research Foundation.

Author Disclosures: Dr Dowd reports receiving grants for this research funded by the Donaghue Foundation. The remaining authors report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.

Authorship Information: Concept and design (TM, KW, TB, BED); acquisition of data (TM, KW, BED); analysis and interpretation of data (TM, KW, TYH, TB, BED); drafting of the manuscript (TM, KW, TYH, BED); critical revision of the manuscript for important intellectual content (TM, KW, TB, BED); statistical analysis (TM, TYH, TB, BED); provision of patients or study materials (TM); obtaining funding (TM); administrative, technical, or logistic support (TM); and supervision (TM).

Address Correspondence to: Tim McDonald, PhD, MPP, RAND Corporation, 1200 S Hayes St, Arlington, VA 22202. Email: tmcdonald@rand.org.

REFERENCES

1. Claxton G, Rae M, Winger A, et al. Employer Health Benefits: 2024 Annual Survey. KFF; 2024. Accessed March 9, 2026. https://files.kff.org/attachment/Employer-Health-Benefits-Survey-2024-Annual-Survey.pdf

2. Income, poverty and health insurance coverage in the United States: 2023. News release. US Census Bureau. September 10, 2024. Accessed March 9, 2026. https://www.census.gov/newsroom/press-releases/2024/income-poverty-health-insurance-coverage.html

3. McWilliams JM, Chen AJ. Understanding the latest ACO “savings”: curb your enthusiasm and sharpen your pencils—part 1. Health Affairs Forefront. November 12, 2020. Accessed March 9, 2026. https://www.healthaffairs.org/content/forefront/understanding-latest-aco-savings-curb-your-enthusiasm-and-sharpen-your-pencils-part-1

4. Burns LR, Pauly MV. Transformation of the health care industry: curb your enthusiasm? Milbank Q. 2018;96(1):57-109. doi:10.1111/1468-0009.12312

5. Markovitz AA, Hollingsworth JM, Ayanian JZ, Norton EC, Yan PL, Ryan AM. Performance in the Medicare Shared Savings Program after accounting for nonrandom exit: an instrumental variable analysis. Ann Intern Med. 2019;171(1):27-36. doi:10.7326/M18-2539

6. Parashuram S, Lee W, Rowan K, et al. The effect of next generation accountable care organizations on Medicare expenditures. Health Aff (Millwood). 2024;43(7):933-941. doi:10.1377/hlthaff.2022.01648

7. Bond AM, Civelek Y, Schpero WL, et al. Long-term spending of accountable care organizations in the Medicare Shared Savings Program. JAMA. 2025;333(21):1897-1905. doi:10.1001/jama.2025.3870

8. Desai SM, Hatfield LA, Hicks AL, Chernew ME, Mehrotra A, Sinaiko AD. What are the potential savings from steering patients to lower-priced providers? a static analysis. Am J Manag Care. 2019;25(7):e204-e210.

9. Drake C, Ryan C, Dowd B. Sources of inertia in the individual health insurance market. J Public Econ. 2022;208:104622. doi:10.1016/j.jpubeco.2022.104622

10. Robinson JC, Whaley CM, Brown TT. Association of reference pricing with drug selection and spending. N Engl J Med. 2017;377(7):658-665. doi:10.1056/NEJMsa1700087

11. Robinson JC, Whaley CM, Brown TT. Reference pricing, consumer cost-sharing, and insurer spending for advanced imaging tests. Med Care. 2016;54(12):1050-1055. doi:10.1097/MLR.0000000000000605

12. Whaley CM, Dankert C, Richards M, Bravata D. An employer-provider direct payment program is associated with lower episode costs. Health Aff (Millwood). 2021;40(3):445-452. doi:10.1377/hlthaff.2020.01488

13. Dowd B, Feldman R, Coulam R. Competitive physician prices in fee-for-service Medicare. Health Affairs Forefront. March 18, 2020. Accessed March 9, 2026. https://www.healthaffairs.org/content/forefront/competitive-physician-prices-fee-for-service-medicare

14. Whaley C, Sood N, Chernew M, Metcalfe L, Mehrotra A. Paying patients to use lower-priced providers. Health Serv Res. 2022;57(1):37-46. doi:10.1111/1475-6773.13711

15. Dowd BE, Huang TY, McDonald T. Tiered cost-sharing for primary care gatekeeper clinics. Am J Health Econ. 2021;7(3):249-360. doi:10.1086/714360

16. Brot-Goldberg ZC, Chandra A, Handel BR, Kolstad JT. What does a deductible do? the impact of cost-sharing on health care prices, quantities, and spending dynamics. Q J Econ. 2017;132(3):1261-1318. doi:10.1093/qje/qjx013

17. Dowd B, McDonald T. Health care affordability: iron triangle or iron curtain? Health Affairs Forefront. February 16, 2023. Accessed March 23, 2026. https://www.healthaffairs.org/content/forefront/health-care-affordability-iron-triangle-iron-curtain

18. McDonald T, Debbarma A, Whaley C, Reid R, Dowd B. Barriers primary care clinic leaders face to improving value in a consumer choice health plan design. Health Aff Sch. 2023;1(6):qxad065. doi:10.1093/haschl/qxad065

19. McDonald T. Beginning With System Transformation in Mind: Toward a Theory of Policy Leadership for Systemic Challenges. RAND Corporation; 2024. Accessed March 23, 2026. https://www.rand.org/pubs/rgs_dissertations/RGSDA3001-1.html

20. Abaluck J, Gruber J. Choice inconsistencies among the elderly: evidence from plan choice in the Medicare Part D program. Am Econ Rev. 2011;101(4):1180-1210. doi:10.1257/aer.101.4.1180

21. Afendulis CC, Sinaiko AD, Frank RG. Dominated choices and Medicare Advantage enrollment. J Econ Behav Organ. 2015;119:72-83. doi:10.1016/j.jebo.2015.07.009

22. Handel BR. Adverse selection and inertia in health insurance markets: when nudging hurts. Am Econ Rev. 2013;103(7):2643-2682. doi:10.1257/aer.103.7.2643

23. Handel BR, Kolstad JT. Health insurance for “humans”: information frictions, plan choice, and consumer welfare. Am Econ Rev. 2015;105(8):2449-2500. doi:10.1257/aer.20131126

24. Bhargava S, Lowenstein G, Snydor J. Choose to lose: health plan choices from a menu with dominated options. Q J Econ. 2017;132(3):1319-1372. doi:10.1093/qje/qjx011

25. Diamond R, Dickstein MJ, McQuade T, Persson P. Insurance without commitment: evidence from the ACA marketplaces. National Bureau of Economic Research working paper 24668. May 2018. Revised April 2021. Accessed March 9, 2026. https://www.nber.org/papers/w24668.pdf

26. Polyakova M. Regulation of insurance with adverse selection and switching costs: evidence from Medicare Part D. Am Econ J Appl Econ. 2016;8(3):165-195. doi:10.1257/app.20150004

27. Heiss F, McFadden D, Winter J, Wuppermann A, Zhou B. Inattention and switching costs as sources of inertia in Medicare Part D. National Bureau of Economic Research working paper 22765. October 2016. Accessed March 9, 2026. https://www.nber.org/papers/w22765

28. Ho K, Hogan J, Scott Morton F. The impact of consumer inattention on insurer pricing in the Medicare Part D program. RAND J Econ. 2017;48(4):877-905. doi:10.1111/1756-2171.12207

29. Higuera L, Carlin CS, Dowd B. Narrow provider networks and willingness to pay for continuity of care and network breadth. J Health Econ. 2018;60:90-97. doi:10.1016/j.jhealeco.2018.06.006

30. Luco F. Switching costs and competition in retirement investment. Am Econ J Microecon. 2019;11(2):26-54. doi:10.1257/mic.20160332

31. Wolf E, Feher A, Ravel K, Menashe I. Comparing the effects of nudges and automatic plan switching on choice errors among low-income marketplace enrollees. Health Aff (Millwood). 2023;42(7):1002-1010. doi:10.1377/hlthaff.2022.01672

32. Marzilli Ericson KM. Consumer inertia and firm pricing in the Medicare Part D prescription drug insurance exchange. Am Econ J Econ Policy. 2014;6(1):38-64. doi:10.1257/pol.6.1.38

33. Samuelson W, Zeckhauser R. Status quo bias in decision making. J Risk Uncertain. 1988;1(1):7-59. Accessed February 24, 2026. https://www.jstor.org/stable/41760530

34. Marzilli Ericson KM, Kingsdale J, Layton T, Sacarny A. Nudging leads consumers in Colorado to shop but not switch ACA marketplace plans. Health Aff (Millwood). 2017;36(2):311-319. doi:10.1377/hlthaff.2016.0993

35. Kling JR, Mullainathan S, Shafir E, Vermeulen LC, Wrobel MV. Comparison friction: experimental evidence from Medicare drug plans. Q J Econ. 2012;127(1):199-235. doi:10.1093/qje/qjr055

36. Sabety A. The value of relationships in healthcare. J Public Econ. 2023;225:104927. doi:10.1016/j.jpubeco.2023.104927

37. Wasson JH, Sauvigne AE, Mogielnicki RP, et al. Continuity of outpatient medical care in elderly men. a randomized trial. JAMA. 1984;252(17):2413-2417. doi:10.1001/jama.1984.03350170015011

38. McClure W, McDonald T. Looking back: Minnesota’s growing health reform strategy: a history of health care system reform in Minnesota. Minn J. 2012;29(3):5-8. Accessed March 23, 2026. https://citizensleague.org/wp-content/uploads/2017/07/MNJournalSummer2012.pdf

39. The ACG system: overview. Johns Hopkins. Accessed March 9, 2026. https://pdf4pro.com/amp/download?data_id=356186&slug=the-johns-hopkins-acg-174-system-q-corp

40. Thaler RH, Sunstein CR, Balz JP. Choice architecture. SSRN. Preprint posted online April 2, 2010. doi:10.2139/ssrn.1583509

41. Thaler RH. Nudge, not sludge. Science. 2018;361(6401):431. doi:10.1126/science.aau9241

42. Davis PK, McDonald T, Pendleton-Jullian A, O’Mahony A, Osoba OA. Reforming the teaching and conducting of policy studies to deal better with complex systems. Policy Complex Systems. 2021;7(1):119-139. Accessed March 23, 2026. https://www.rand.org/pubs/external_publications/EP68721.html

43. McDonald T, White KM, Huang TY, Whaley CM, Dowd B. Clinic price reductions in a tiered total cost benefit design. Am J Manag Care. 2021;27(9):e316-e321. doi:10.37765/ajmc.2021.88744

44. Dowd B, Boese T, McDonald T. Tiered cost-sharing health insurance. Minnesota Physician. 2022;35(11). Accessed April 12, 2022. https://www.mnphy.com/0222-cover-one

45. Huang TY, Togun A, Boese T, Dowd BE. Analysis of affordable health care. Med Care. 2022;60(9):718-725. doi:10.1097/MLR.0000000000001755

46. Johnson EJ, Shu SB, Dellaert BGC, et al. Beyond nudges: tools of a choice architecture. Mark Lett. 2012;23(2):487-504. Accessed February 24, 2026. https://www.jstor.org/stable/23259227

47. Abadie A, Athey S, Imbens GW, Wooldridge JM. When should you adjust standard errors for clustering?
Q J Econ. 2023;138(1):1-35. doi:10.1093/qje/qjac038

48. Taylor EA, Carman KG, Lopez A, Muchow AN, Roshan P, Eibner C. Consumer Decisionmaking in the Health Care Marketplace. RAND Corporation; 2016. Accessed March 9, 2026. https://www.rand.org/pubs/research_reports/RR1567.html

49. Sinaiko AD, Bambury E, Chien AT. Consumer choice in U.S. health care: using insights from the past to inform the way forward. The Commonwealth Fund. November 30, 2021. Accessed March 9, 2026. https://www.commonwealthfund.org/publications/fund-reports/2021/nov/consumer-choice-us-health-care-using-insights-from-past

50. Rao P, Fischer SH, Vaiana ME, Taylor EA. Barriers to Price and Quality Transparency in Health Care Markets. RAND Corporation; 2021. Accessed March 9, 2026. https://www.rand.org/pubs/research_reports/RRA1158-1.html

51. Myerson R, Tilipman N, Feher A, Li H, Yin W, Menashe I. Personalized telephone outreach increased health insurance take-up for hard-to-reach populations, but challenges remain. Health Aff (Millwood). 2022;41(1):129-137. doi:10.1377/hlthaff.2021.01000

52. Whaley C, Brown T, Robinson J. Consumer responses to price transparency alone versus price transparency combined with reference pricing. Am J Health Econ. 2019;5(2):227-249. doi:10.1162/ajhe_a_00118