1. Introduction
Because personal care and other home and community‐based services (HCBS) are vulnerable to waste, fraud, and abuse, state Medicaid programs use prior authorization (PA) to improve program integrity [1]. However, little is known about Medicaid beneficiaries' experiences with PA for HCBS, even though physicians and consumers associate PA in general with delayed/denied care [2, 3].
We surveyed Virginia Medicaid beneficiaries to examine the timeliness of PA approval for in‐home services. Administrative data suggest that only 3.9% of HCBS PAs in the state were denied since 2020; in the same period, 11.6% of authorizations took longer than 28 days [4]. We examined the association between PA approval delays and beneficiary characteristics, difficulty accessing HCBS, and health plan satisfaction. A novel feature of our survey is that all respondents were enrolled in the state's 1915(c) HCBS waiver and therefore had recent, documented functional and/or medical needs for HCBS, unlike representative surveys of consumers.
2. Methods
The sampling frame consisted of community‐dwelling Virginia Medicaid beneficiaries with at least 6 months of continuous coverage in their Medicaid health plan and the HCBS waiver, and the same Medicare coverage in the prior 6 months (Methods/Table S1). Details on the survey were published previously [5].
Outcomes included: (1) a binary measure of whether, in the past 6 months, the respondent reported delay “in getting approval for services from [their] Medicaid health plan for … care and other in‐home services and conveniences that help with daily activities” (e.g., personal care, adult day care, and skilled nursing care); (2) a binary measure of whether, in the past 6 months, the respondent reported “difficulty accessing care and other in‐home services and conveniences that help with daily activities”; and (3) Medicaid health plan satisfaction ranging from 0 (“worst plan possible”) to 10 (“best plan possible”). Respondent demographic traits and eight measures capturing health, medical, and socioeconomic needs were defined from the survey (Methods S2).
We conducted univariate analyses and chi‐square tests of differences in delayed PA approval for HCBS by respondent trait. Using multivariate regression, we estimated adjusted differences in delayed PA approval (logit models) and examined the association of delayed PA approval with difficulty accessing in‐home care (logit models) and plan rating (OLS) controlling for respondent characteristics. We assessed statistical significance using two‐tailed tests and alpha of 0.05.
3. Results
A total of 2226 surveys of 6867 sent were returned, containing non‐missing data on PA approvals for in‐home services for 2136 respondents. One‐third of respondents reported delayed PA approval for in‐home services; delays did not differ by most demographic traits (Table S2). In univariate analysis, respondents with more health, medical, or socioeconomic needs were more likely to report delayed PA approval than those with less need (Table 1). For example, 40.8% of respondents with five or more chronic conditions reported delayed PA approval, compared to 31.0% of those with four or fewer conditions (p < 0.001); respondents who had problems getting enough food were more likely to report delayed PA approvals than those who did not (41.6% vs. 28.1%; [p < 0.001]). Adjusting for demographic traits, all need measures, and other types of delays, respondents with several measures of high need were significantly more likely to report delayed PA approval for HCBS (Table S3).
TABLE 1.
Differences in delays in PA approvals for HCBS by respondent need and reports of other delays.
| % of Rs in sample a | % of Rs reporting delays in PA approval for in‐home services b | p c | |
|---|---|---|---|
| Poor self‐rated health status | |||
| No | 76.9% | 32.0% | |
| Yes | 21.3% | 35.4% | 0.17 |
| Missing | 1.8% | ||
| 5+ chronic conditions | |||
| No | 80.2% | 31.0% | |
| Yes | 17.0% | 40.8% | < 0.001 |
| Missing | 2.8% | ||
| Six ADL limitations | |||
| No | 75.5% | 31.4% | |
| Yes | 23.0% | 37.2% | 0.02 |
| Missing | 1.5% | ||
| Adverse consequences of unmet need for LTSS | |||
| No | 57.7% | 26.7% | |
| Yes | 29.6% | 47.6% | < 0.001 |
| Missing | 12.6% | ||
| Emergency department visits in the past 6 months | |||
| No | 63.9% | 30.8% | |
| Yes | 35.0% | 36.8% | 0.005 |
| Missing | 1.1% | ||
| Inpatient hospital stay in the past 6 months | |||
| No | 79.2% | 31.5% | |
| Yes | 19.9% | 38.7% | 0.005 |
| Missing | 0.8% | ||
| Counts on two or fewer people for support | |||
| No | 47.3% | 28.0% | |
| Yes | 50.5% | 37.2% | < 0.001 |
| Missing | 2.2% | ||
| Food does not last | |||
| No | 65.2% | 28.1% | |
| Yes | 32.5% | 41.6% | < 0.001 |
| Missing | 2.3% | ||
| Reported delays in PA approvals for both medications and specialist appointments | |||
| No | 78.9% | 18.8% | |
| Yes | 21.2% | 85.0% | < 0.001 |
| Missing | 0.7% | ||
Abbreviations: LTSS, long‐term services and supports; PA, prior authorization; R, respondent.
Reports the percentage of respondents from the analysis sample with each trait.
Reports the percentage of respondents with that trait experiencing delays in PA approvals for HCBS. Respondents with missing data on a given trait are not included in calculations.
Reports p values from a Pearson chi‐square test of the null hypothesis that the percentage of respondents who report delays in PA approval does not differ for those with or without a given trait. Respondents with missing data on a given trait are not included in calculations.
Compared to respondents who did not report delayed PA approvals for in‐home services, those who reported delays had higher odds of difficulty accessing in‐home services (aOR = 6.74 [95% CI: 4.74–9.50], Table 2). Associations between difficulty accessing in‐home services and delayed PA approval for medications or specialists were not statistically significant. Members reporting delayed PA approval for in‐home services gave their health plans lower ratings (β = −0.21 [95% CI: −0.004 to −0.58], Table 2). Results were robust to sensitivity tests (Tables S4–S9).
TABLE 2.
Adjusted associations between delays in PA approval and difficulty accessing in‐home services and overall health plan rating.
| Explanatory variables | Dependent variable: difficulty accessing in‐home services | Dependent variable: plan rating (0–10) scale | |||
|---|---|---|---|---|---|
| aOR a | 95% CI | β b | 95% CI | ||
| Delays in PA approval for in‐home services | No | Reference | −0.21 | −0.41 to −0.004 | |
| Yes | 6.73 | 4.74–9.53 | |||
| Delays in PA approval for medications | No | Reference | −0.27 | −0.49 to −0.06 | |
| Yes | 0.94 | 0.64–1.37 | |||
| Delays in PA approval for specialist appointments | No | Reference | −0.28 | −0.51 to −0.05 | |
| Yes | 0.84 | 0.57–1.24 | |||
Abbreviations: aOR, adjusted odds ratio; CI, confidence interval; PA, prior authorization.
Reports results from a single logistic regression of the outcome (difficulties accessing in‐home services). The estimation sample includes 1721 observations. Adjusted odds ratios reported, adjusting for age, sex, race, education, marital status, rural residence, the presence of 5+ chronic conditions, six activity of daily living limitations, poor self‐rated health status, adverse consequences of unmet need for long‐term services and supports, emergency department visit in past 6 months, inpatient hospital stay in past 6 months, counts on two or fewer people for supports, and reported that food did not last.
Reports results from a single OLS regression of the dependent variable (plan rating, on a scale of 0–10). The estimation sample includes 1710 observations. Coefficient estimates reported from a model that controlled for age, sex, race, education, marital status, rural residence, the presence of 5+ chronic conditions, six activity of daily living limitations, poor self‐rated health status, adverse consequences of unmet need for long term services and supports, emergency department visit in past 6 months, inpatient hospital stay in past 6 months, counts on two or fewer people for supports, and reported that food did not last.
4. Discussion
Our novel findings on Medicaid HCBS add to evidence from prior national studies that PA burden is generally greater for those with more health needs and that PA delays/denials are associated with reduced access and trust in the healthcare system [3, 6, 7]. While reports of delays may reflect respondents' unobserved traits (e.g., impatience), reporting other types of PA delays was not associated with difficulty accessing in‐home services. Limitations include focusing on one state; like other surveys, delays were subjective and length was not defined [3], and persons experiencing delays may have been more likely to respond.
We provide new cross‐sectional evidence that delayed PA approval for Medicaid HCBS was associated with lower access and health plan ratings. Future work should explore whether access and satisfaction improve following regulations to lessen delays [8] and whether delays can be reduced by streamlining communication between providers, care coordinators, and plans, improving models of care, or using concurrent authorization [9, 10].
Author Contributions
All authors meet the criteria for authorship stated in the Uniform Requirements for Manuscripts Submitted to Biomedical Journals. Jennifer M. Mellor designed the study and made substantial contributions to the acquisition of data, conducted the analysis and interpreted the data, drafted the article, and approved the final version to be published. Peter J. Cunningham and Erin Britton made substantial contributions to the acquisition of the data, revised the article critically for important intellectual content, and gave final approval of the version to be published. Andrew Mitchell and Sandra Dagenhart made substantial contributions to the analysis and interpretation of the data, revised the article critically for important intellectual content, and gave final approval of the version to be published.
Conflicts of Interest
The authors declare no conflicts of interest.
Supporting information
Data S1. Methods S1. Survey procedures.
Table S1. Comparison of survey respondents and all surveyed members.
Methods S2. Construction of variables measuring demographic traits, respondent needs, and other types of PA delays.
Table S2. Differences in delays in PA approvals by demographic traits.
Table S3. Adjusted odds ratio (aOR) from multivariate models of delays in PA approvals for in‐home services for measures of interest.
Table S4. Adjusted odds ratio (aOR) of delays in PA approvals for in‐home services for measures of interest excluding respondents with missing data on demographic traits.
Table S5. Adjusted odds ratio (aOR) of delays in PA approvals for in‐home services for measures of interest controlling for type of Medicare coverage.
Table S6. Adjusted odds ratio (aOR) of difficulty accessing in‐home services for measures of interest excluding respondents with missing data on demographic traits.
Table S7. Adjusted odds ratio (aOR) of difficulty accessing in‐home services for measures of interest controlling for type of Medicare coverage.
Table S8. Coefficient estimates for measures of interest from ordinary least squares (OLS) model of overall health plan rating excluding respondents with missing data on demographic traits.
Table S9. Coefficient estimates for measures of interest from ordinary least squares (OLS) model of overall health plan rating controlling for type of Medicare coverage.
Acknowledgments
The authors have nothing to report.
Funding: The Virginia Department of Medical Assistance Services (DMAS) funded the collection of the data used in this study.
References
- 1. Centers for Medicare and Medicaid Services (CMS) Medicaid Integrity Institute , “Vulnerabilities and Mitigation Strategies in Medicaid Personal Care Services,” (2018), https://www.cms.gov/medicare‐medicaid‐coordination/fraud‐prevention/fraudabuseforprofs/downloads/vulnerabilities‐mitigation‐strategies.pdf.
- 2. American Medical Association , “AMA Prior Authorization Physician Survey,” (2023), https://www.ama‐assn.org/system/files/prior‐authorization‐survey.pdf.
- 3. Pollitz K., Pestaina K., Lopes L., Wallace R., and Lo J., “Consumer Problems With Prior Authorization: Evidence From KFF Survey,” (2023), https://www.kff.org/affordable‐care‐act/issue‐brief/consumer‐problems‐with‐prior‐authorization‐evidence‐from‐kff‐survey/.
- 4. Virginia Department of Medical Assistance Services , “MCO Service Authorization Performance Dashboard,” accessed August 1, 2024, https://www.dmas.virginia.gov/data/mco‐service‐authorization‐performance/.
- 5. Mellor J. M., Cunningham P. J., Britton E., Behrens M., Urmi A. F., and Vega V., “Beneficiary Experience of Care by Level of Integration in Dual Eligible Special Needs Plans,” JAMA Health Forum 5, no. 6 (2024): e241383, 10.1001/jamahealthforum.2024.1383. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Karpman M., Haley J. M., Gonzalez D., Morriss S., and Evans C., “Insights From Nonelderly Adults With Disabilities on Difficulties Obtaining Home‐ and Community‐Based Services and Other Health Care,” (2024), https://www.urban.org/research/publication/insights‐nonelderly‐adults‐disabilities‐difficulties‐obtaining‐home.
- 7. Chino F., Baez A., Elkins I. B., Aviki E. M., Ghazal L. V., and Thom B., “The Patient Experience of Prior Authorization for Cancer Care,” JAMA Network Open 6, no. 10 (2023): e2338182. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Centers for Medicare and Medicaid Services , “CMS Interoperability and Prior Authorization Final Rule CMS‐0057‐F,” (2024), https://www.cms.gov/newsroom/fact‐sheets/cms‐interoperability‐and‐prior‐authorization‐final‐rule‐cms‐0057‐f.
- 9. Salzbrenner S. G., Lydiatt M., Helding B., et al., “Influence of Prior Authorization Requirements on Provider Clinical Decision‐Making,” American Journal of Managed Care 29, no. 7 (2023): 331. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Giardino A. P. and Lyn M. A., “Utilization Management, Case Management, and Care Coordination,” in Medical Quality Management: Theory and Practice, eds. Giardino A. P., Riesenberg L. A., and Varkey P. (Cham: Springer, 2021), 139–175. [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data S1. Methods S1. Survey procedures.
Table S1. Comparison of survey respondents and all surveyed members.
Methods S2. Construction of variables measuring demographic traits, respondent needs, and other types of PA delays.
Table S2. Differences in delays in PA approvals by demographic traits.
Table S3. Adjusted odds ratio (aOR) from multivariate models of delays in PA approvals for in‐home services for measures of interest.
Table S4. Adjusted odds ratio (aOR) of delays in PA approvals for in‐home services for measures of interest excluding respondents with missing data on demographic traits.
Table S5. Adjusted odds ratio (aOR) of delays in PA approvals for in‐home services for measures of interest controlling for type of Medicare coverage.
Table S6. Adjusted odds ratio (aOR) of difficulty accessing in‐home services for measures of interest excluding respondents with missing data on demographic traits.
Table S7. Adjusted odds ratio (aOR) of difficulty accessing in‐home services for measures of interest controlling for type of Medicare coverage.
Table S8. Coefficient estimates for measures of interest from ordinary least squares (OLS) model of overall health plan rating excluding respondents with missing data on demographic traits.
Table S9. Coefficient estimates for measures of interest from ordinary least squares (OLS) model of overall health plan rating controlling for type of Medicare coverage.