Abstract
BACKGROUND:
Cost-related nonadherence (CRN), that is, not taking medication as prescribed to save money, may remain disproportionately high among individuals with multiple chronic conditions, particularly during periods of economic stress, such as the COVID-19 pandemic. However, the impact of economic hardship 3 years after the pandemic on CRN levels among individuals with multiple chronic conditions is still largely unknown.
OBJECTIVE:
To examine changes in CRN prevalence in 2020 (pandemic) and 2021 to 2023 (post-pandemic years 1, 2, and 3) relative to 2019 (pre-pandemic) among adults with multiple chronic conditions in the United States.
METHODS:
This is a repeated cross-sectional study using data from the National Health Interview Survey, 2019-2023. Our study sample included 27,413 US adults aged 18 to 64 years with 2 or more of any of 14 chronic conditions and who were prescribed medication. CRN (dependent variable) is a binary measure with values “1” if respondents endorsed 1 of the 4 cost-saving behavior questions—not purchasing medicine refills, delaying refills, splitting pills, or skipping doses to save money—and “0” otherwise. Analyses include survey-weighted CRN prevalence estimates by year and linear probability models assessing prevalence changes in 2020-2023 relative to 2019, overall, and by multiple chronic conditions subgroups (2, 3, and ≥4 conditions).
RESULTS:
The overall CRN prevalence in 2019 was 18.9%, 16.7% in 2020, 13.5% in 2021, 14.5% in 2022, and 15.5% in 2023. CRN decreased in all years relative to 2019 but only significantly by 2.2% (P = 0.001) in 2021 and by 1.4% (P = 0.049) in 2022. The subgroup analysis shows variation in these results, with a significant reduction in CRN in 2021, relative to 2019, limited to those who reported 3 chronic conditions.
CONCLUSIONS:
Fewer adults with multiple chronic conditions reported CRN 1 and 2 years after the pandemic relative to the pre-pandemic in the United States, but those with 4 or more conditions remain vulnerable after the pandemic.
Plain language summary
Adults in the United States with 2 chronic conditions or more reported less often not adhering to medication as prescribed to save money 1 and 2 years after the COVID-19 pandemic started in 2020 compared with 2019 pre-pandemic year. The report of this cost-saving nonadherence behavior was trending to pre-pandemic levels by 2023, with elevated reports, particularly among individuals with 4 or more conditions.
Implications for managed care pharmacy
The reduction in cost-related nonadherence (CRN) during the post-pandemic years shows the potential for government financial stability measures to maintain access to care for individuals with multiple chronic conditions. However, as these programs expired and CRN levels trend to pre-pandemic levels, these individuals may benefit from strategies to address medication affordability, including improved access to pharmacy benefits. Other strategies at the provider level (eg, screening and discussion of lower-cost options) can help mitigate CRN in this population.
The affordability of medications for treating chronic conditions has been a concern long before the COVID-19 pandemic. 1 , 2 The pandemic exacerbated economic hardship and increased inflation, affecting the affordability of many goods, including medication. 3 , 4 As a result, individuals may not take medication as prescribed because of financial reasons, a term referred to as cost-related nonadherence (CRN). 5 CRN can occur if individuals stop or delay filling a prescription, resulting in a treatment gap, or reduce medication intake to avoid cost. 2 , 6
Individuals with multiple chronic conditions often report CRN and are at higher risk of worsened patient health outcomes, such as increased emergency department visits and mortality. 7 – 10 CRN among adults with chronic conditions is still prevalent, 5 , 7 but improvements have been made in recent years with the implementation of programs such as Medicare Part D and the Affordable Care Act. 11 – 13 These programs increased the affordability of and access to medications, resulting in CRN declining among some groups (eg, older adults and the uninsured). 12 , 14 However, the impact of economic hardship 3 years after the pandemic on CRN levels among individuals with multiple chronic conditions is still largely unknown.
A recent report on CRN showed that in 2021, approximately 8% of adults in the United States who had prescription medication did not take it as prescribed to reduce cost. 15 During this period following the pandemic onset in 2020, the prevalence of other financial stability measures related to CRN (eg, food insecurity and problems paying medical bills) 16 decreased in the general population. 17 , 18 These changes may be attributed in part to the enactment of economic rescue policies at the federal level (ie, Coronavirus Aid, Relief, and Economic Security Act and Family First Coronavirus Response Act) and other state and local efforts to mitigate the impact of the pandemic on the health care access and cost for individuals. 17 , 18 The Family First Coronavirus Response Act, for example, enhanced coverage stability and access to care 19 , 20 by increasing Medicaid funding and requiring states to retain individuals continuously enrolled and expand coverage enrollment for the newly eligible. 21 , 22
Despite these improvements, it is still largely unknown whether adults with multiple chronic conditions continued reporting cost-saving behaviors at disproportionally higher rates during the pandemic and post-pandemic years relative to the pre-pandemic year. New information about the potential changes in CRN after 2020 can inform future pandemic responses, supporting continuous access to care and medication to manage chronic conditions during times of economic hardship. In this study, we examined whether CRN changed yearly among US adults aged 18 to 64 years with multiple chronic conditions in 2020 (pandemic) and 2021-2023 (post-pandemic) relative to 2019 (pre-pandemic).
Methods
DATA SOURCE AND STUDY SAMPLE
This repeated cross-sectional study used 2019-2023 National Health Interview Survey (NHIS) data from a sample of nationally representative households. 23 The NHIS surveys noninstitutionalized adults residing in the United States to monitor illness, disability, and health care access. We excluded survey years before 2019 because estimates are not comparable owing to methodological changes. 24 In 2019, the survey sample weighting used to generate representative estimates changed from poststratification to a multilevel approach and ranking to adjust for nonresponses and to match the population total. 24 From 2019 to 2023, the total numbers of interviewed adults aged 18 years or older were 31,997, 31,153, 29,482, 27,587, and 29,457, respectively. The sample adult response rates were 59.1% (2019), 48.9% (2020), 50.9% (2021), 47.7% (2022), and 47.0% (2023). 25 This study follows the STROBE reporting guideline for cross-sectional studies. The institutional review board determined that this study did not involve human subjects because the data are publicly available.
We identified adults aged 18 to 64 years who reported having ever been diagnosed with or still having any of 14 chronic conditions: rheumatic (eg, arthritis, rheumatoid arthritis, gout, lupus, or fibromyalgia), diabetes, asthma, chronic obstructive pulmonary disease, anxiety, depression, dementia, angina, coronary heart disease, heart attack, stroke, hypertension, high cholesterol, or cancer. All these conditions were consistently included in the surveys from 2019 to 2023, allowing for better estimates and comparability. We also selected adults who reported having a medication prescribed by a doctor or other health professional 12 months before the interview, as CRN indicators were asked of these respondents exclusively.
STUDY MEASURES
We created a binary measure of CRN if respondents reported at least 1 of 4 questions assessing cost-saving behaviors in the past 12 months: (1) needed prescription medication but did not get it because of cost; (2) delayed filling a prescription to save money; (3) took less medication to save money; and (4) skipped medication doses to save money. We excluded 7 respondents with missing values in all the CRN questions.
Our independent variable was the survey year, where 2019 (pre-pandemic) was the reference year to assess changes comparing the pandemic and post-pandemic years. Individual characteristics (covariates) include age (18-44, 45-54, 55-64), sex (female, male), and US Census region (Northeast, Midwest, South, and West). We accounted for health insurance types (none, private, public, and other, ie, military health insurance and Indian Health Service). Adults who reported having private and Medicaid insurance or 2 different health insurance types were excluded from the analysis. To account for material hardship, we included the ratio of family income poverty threshold (ie, <1.0 [poor], 1.0-1.99 [near poor], 2.0-2.99 [middle income], and ≥3.0 [high income]), which accounts for family size and age structure. 26 Additional covariates include the financial hardship measures associated with CRN: family-level 30-day food security status (food secure vs low/very low food security) 23 and care affordability (ie, needing but not being able to afford medical health care, worried about medical bills, and problems paying medical bills). We also accounted for perceived health status (excellent–very good–good vs fair-poor) and for the type of chronic conditions diagnosed (no vs yes/diagnosis).
STATISTICAL ANALYSIS
Descriptive Analysis.
We described annual CRN by estimating the survey-weighted prevalence and calculating 95% CIs overall and by chronic condition categorization. We used graphs to display trends in CRN prevalence from 2019 to 2023.
Regression Analysis.
We evaluate whether changes in the CRN prevalence after the pre-pandemic era will vary in direction and magnitude across chronic conditions groups. To test this, we used multivariable linear probability models (LPMs) and estimated the survey-weighted slope of CRN (dependent variable with values “0” or “1”) in 2020, 2021, and 2022 relative to 2019, accounting for individual characteristics. We rejected the null hypothesis if the P value was less than 0.05 for the t-test examining the change (slope estimate different from 0). 27 Summary statistics and LPMs were estimated using R 4.3.1, survey package, 27 , 28 and the CRN prevalence by years plotted using ggplot2. 29 We used the “subset” function in our subgroup analysis by chronic condition group to keep the original design for variance estimation. 27
Subgroup Analysis.
We conducted subgroup analyses to assess potential variations in the CRN year comparisons with 3 additional LPMs estimating changes by the multiple chronic condition categorizations. We defined the subgroups with a categorization previously used in CRN studies that groups the count of chronic conditions reported into 3 categories (2 chronic conditions, 3, and ≥4). 5 , 30 To check for the potential of overfitting the subgroup models and complementing the analysis, we ran subgroup analyses excluding the type of chronic condition as adjustment covariates (Supplementary Materials (326.9KB, pdf) , available in online article).
Results
SAMPLE CHARACTERISTICS
Our sample included 27,413 US adults aged 18 to 64 years with chronic conditions who had been prescribed medication and without Medicare or dual health insurance coverage. Approximately 33% to 37% were aged 55 to 64 years, 58% were female, 68% were non-Hispanic White, and approximately 39% were from the South US region (Table 1). Across survey years, most in our sample had a family income above the poverty line and reported having private health insurance. Additionally, approximately 27% to 29% had 2 chronic conditions, 27% to 29% had 3, and 22% to 24% had 4 or more conditions. Hypertension, high cholesterol, depression, anxiety, and rheumatic were the 5 most reported chronic conditions across the survey years.
TABLE 1.
Distribution of the Sample Characteristics for Survey Years 2019-2023
| Characteristics | Survey years | ||||
|---|---|---|---|---|---|
| 2019 (N = 5,578) | 2020 (N = 5,607) | 2021 (N = 5,462) | 2022 (N = 5,101) | 2023 (N = 5,665) | |
| Age, years | |||||
| 18-44 | 1,837 (36.7) | 1,783 (37.9) | 1,902 (38.3) | 1,907 (41.8) | 2,183 (42.6) |
| 45-54 | 1,396 (25.7) | 1,369 (24.9) | 1,335 (25.2) | 1,165 (24.1) | 1,295 (23.5) |
| 55-64 | 2,345 (37.6) | 2,455 (37.1) | 2,225 (36.5) | 2029 (34.1) | 2,187 (33.9) |
| Sex | |||||
| Male | 2,290 (41.8) | 2,336 (42.2) | 2,217 (43.2) | 2,071 (42.3) | 2,245 (41.6) |
| Female | 3,287 (58.2) | 3,270 (57.8) | 3,244 (56.8) | 3,029 (57.7) | 3,417 (58.4) |
| Race | |||||
| White-NH | 3,979 (68.8) | 4,110 (69.4) | 3,845 (69.4) | 3,522 (68.1) | 3,961 (68.3) |
| Black-NH | 666 (12.0) | 583 (10.8) | 588 (11.0) | 575 (10.9) | 580 (10.8) |
| Hispanic | 631 (13.5) | 591 (13.4) | 660 (13.6) | 655 (14.2) | 744 (13.7) |
| Other | 302 (5.7) | 323 (6.4) | 369 (6.1) | 349 (6.8) | 380 (7.1) |
| US regions | |||||
| Northeast | 907 (17.0) | 1,015 (17.6) | 848 (17.2) | 838 (17.0) | 875 (17.2) |
| Midwest | 1,331 (22.8) | 1,377 (22.4) | 1,242 (22.3) | 1,148 (21.5) | 1,339 (23.0) |
| South | 2,140 (38.9) | 1,926 (38.6) | 2,055 (38.8) | 1,914 (38.9) | 2,122 (37.8) |
| West | 1,200 (21.3) | 1,289 (21.3) | 1,317 (21.8) | 1,201 (22.5) | 1,329 (22.0) |
| Poverty/income ratio | |||||
| <1.0 | 770 (13.3) | 599 (11.2) | 642 (11.3) | 604 (10.8) | 620 (10.1) |
| 1.0-1.99 | 874 (16.6) | 775 (16.4) | 783 (14.7) | 775 (16.1) | 892 (15.9) |
| 2.0-2.99 | 863 (16.5) | 833 (15.7) | 861 (15.6) | 760 (15.6) | 843 (15.3) |
| ≥3.0 | 3,071 (53.6) | 3,400 (56.7) | 3,176 (58.3) | 2,962 (57.5) | 3,310 (58.7) |
| Health insurance a | |||||
| Private | 3,837 (67.5) | 4,038 (68.5) | 3,925 (71.5) | 3,575 (69.9) | 4,043 (70.7) |
| Public | 1,023 (18.9) | 902 (17.7) | 968 (17.8) | 1,011 (19.8) | 1,079 (20.3) |
| Other | 235 (3.9) | 247 (4.2) | 223 (3.6) | 229 (4.0) | 230 (3.4) |
| Uninsured | 483 (9.7) | 420 (9.5) | 346 (7.2) | 286 (6.3) | 313 (5.6) |
| Food insecure b | 767 (14.0) | 552 (11.3) | 475 (8.7) | 578 (10.9) | 771 (13.6) |
| Worried about medical bills | 3,086 (56.7) | 3,125 (58.7) | 2,862 (53.0) | 2,622 (52.5) | 3,006 (53.8) |
| Problem paying medical bills | 1,190 (23.0) | 1,065 (21.1) | 892 (16.9) | 847 (17.2) | 988 (18.0) |
| Could not afford medical care | 742 (14.0) | 579 (11.4) | 588 (10.7) | 519 (10.2) | 583 (10.4) |
| Health status | |||||
| Fair-poor | 1,376 (26.4) | 1,245 (23.6) | 1,203 (22.2) | 1,164 (23.3) | 1,297 (23.5) |
| Chronic conditions | |||||
| Rheumatic c | 2,124 (36.8) | 2,081 (34.4) | 1,956 (33.5) | 1,767 (33.4) | 1,908 (32.2) |
| Diabetes | 1,050 (19.2) | 925 (16.8) | 1,004 (18.8) | 889 (17.4) | 950 (16.8) |
| Asthma | 990 (17.9) | 1,002 (18.3) | 1,020 (18.5) | 944 (18.8) | 1,070 (19.2) |
| COPD | 429 (7.5) | 471 (8.4) | 444 (7.6) | 364 (7.0) | 387 (6.6) |
| Depression | 2,460 (44.3) | 2,525 (46.1) | 2,568 (47.1) | 2,458 (48.5) | 2,763 (48.9) |
| Anxiety | 2,275 (40.9) | 2,384 (44.6) | 2,474 (46.2) | 2,377 (48.0) | 2,777 (49.7) |
| Dementia | 24 (0.4) | 29 (0.4) | 34 (0.6) | 22 (0.5) | 38 (0.6) |
| Angina | 130 (2.6) | 113 (1.9) | 133 (2.3) | 100 (1.9) | 139 (2.4) |
| Coronary heart disease | 329 (5.7) | 328 (5.3) | 326 (5.8) | 270 (5.2) | 281 (5.2) |
| Heart attack | 247 (4.2) | 220 (3.7) | 254 (4.6) | 196 (3.8) | 211 (3.8) |
| Stroke | 229 (3.9) | 205 (3.7) | 184 (3.3) | 172 (3.1) | 210 (3.5) |
| Hypertension | 3,217 (57.4) | 3,100 (54.7) | 2,992 (54.2) | 2,757 (52.7) | 3,070 (52.6) |
| High cholesterol | 2,661 (47.0) | 2,744 (47.3) | 2,654 (47.6) | 2,420 (46.5) | 2,773 (48.1) |
| Cancer | 752 (12.4) | 781 (12.2) | 707 (11.8) | 613 (11.3) | 727 (11.8) |
Data are shown as n (%). %: Survey-weighted prevalence (ie, percentages).
Health insurance: analyzed as binary variables for each health insurance type (private public Medicare, none, and other, that is, military health insurance and Indian Health Service). Because some participants may report having Medicare and private insurance or public, not all categories were mutually exclusive.
Family-level 30-day food security status (food secure vs low secure).
Rheumatic, that is, rheumatoid, arthritis, gout, lupus, or fibromyalgia.
COPD = chronic obstructive pulmonary disease; NH = non-Hispanic.
CHANGES IN CRN PREVALENCE
For adults aged 18 to 64 years with 2 or more chronic conditions, the overall prevalence of CRN (Figure 1) (Supplementary Table 1 (326.9KB, pdf) ) in 2019 (pre-pandemic year) was 18.9%, 16.7% in 2020 (pandemic year), and 13.5%, 14.5%, and 15.5% in 2021-2023 (post-years 1 to 3, respectively). The adjusted LPM estimates in Table 2 show the changes in CRN after the pre-pandemic year, overall, and by chronic condition group. In the pandemic year of 2020, no significant changes were assessed in the overall prevalence of CRN relative to the pre-pandemic year. However, the prevalence decreased significantly by 2.2% (P = 0.001) and 1.7% (P = 0.049) in post-pandemic years 1 and 2, respectively, relative to the pre-pandemic year. By 2023, CRN levels were not significantly different from those observed in 2019 (P = 0.214).
FIGURE 1.
CRN Prevalence by Year, Overall, and Chronic Condition Group Among Adults Aged 18 to 64 Years, 2019-2023
Prevalence estimates and 95% confidence bands are survey-weighted. Overall = all sample; 2 = 2 chronic conditions; 3 = 3 conditions; and ≥4 = 4 or more conditions.
CRN = cost-related nonadherence.
TABLE 2.
LPM Estimates for CRN and Year, Overall and Subgroup Analysis by Chronic Condition Groups, 2019-2023
| Year | Overall (N = 27,413) | 2 CCs (n = 12,495) | 3 CCs (n = 7,624) | ≥4 CCs (n = 7,294) | ||||
|---|---|---|---|---|---|---|---|---|
| Slope (SE) | P value | Slope (SE) | P value | Slope (SE) | P value | Slope (SE) | P value | |
| Unadjusted estimates | ||||||||
| 2019 | — | — | — | — | — | — | — | — |
| 2020 | −2.2 (0.92) | 0.015 | −1.4 (1.26) | 0.268 | −1.3 (1.67) | 0.426 | −4.5 (2.06) | 0.028 |
| 2021 | −5.4 (0.9) | <0.001 | −3.8 (1.17) | 0.001 | −7.7 (1.5) | <0.001 | −6.4 (1.99) | 0.001 |
| 2022 | −4.5 (0.92) | <0.001 | −3.3 (1.16) | 0.004 | −4.1 (1.64) | 0.012 | −7.1 (2.04) | 0.001 |
| 2023 | −3.5 (0.87) | <0.001 | −2.3 (1.11) | 0.040 | −4.4 (1.62) | 0.006 | −5.0 (1.97) | 0.012 |
| Adjusted estimates | ||||||||
| 2019 | — | — | — | — | — | — | — | — |
| 2020 | −0.5 (0.76) | 0.500 | −0.5 (1.06) | 0.662 | −0.1 (1.41) | 0.932 | −0.7 (1.66) | 0.680 |
| 2021 | −2.2 (0.69) | 0.001 | −1.7 (0.96) | 0.080 | −3.6 (1.32) | 0.006 | −1.4 (1.6) | 0.366 |
| 2022 | −1.4 (0.73) | 0.049 | −1.3 (0.94) | 0.154 | −1.3 (1.48) | 0.387 | −1.4 (1.59) | 0.389 |
| 2023 | −0.9 (0.69) | 0.214 | −0.3 (0.93) | 0.734 | −1.4 (1.32) | 0.291 | −1.1 (1.52) | 0.476 |
Survey-weighted multiple LPM estimated the slope (coefficient) adjusted by age, sex, race, US region, poverty income ratio, health insurance type, financial factors (food insecurity, worried about medical bill, and problems paying medical bills), health status, and chronic conditions (rheumatic, ie, rheumatoid, arthritis, gout, lupus, or fibromyalgia; diabetes; asthma; chronic obstructive pulmonary disease; angina; coronary heart disease; heart attack; stroke; hypertension; high cholesterol; anxiety; or depression). Dementia was excluded as a covariate from the models owing to the few cases identified, as shown in Table 1.
CC = chronic condition; CRN = cost-related nonadherence; LPM = linear probability model; slope = LPM slope or estimated change in each year compared with 2019.
SUBGROUP ANALYSIS
By chronic condition subgroups, the CRN prevalence followed the overall estimates, but adults in the 4-or-more-conditions groups reported significantly higher CRN prevalence during the study period than adults with 2 conditions (Figure 1) (Supplementary Table 1 (326.9KB, pdf) ). In the models by condition subgroups (Table 2) (Supplementary Table 2 (326.9KB, pdf) ), among those with 2 conditions, CRN levels decreased, but not significantly, in 2021 only by 1.7% (P = 0.080). For those with 3 conditions, we observed a decrease of 3.6% (P = 0.006) in the CRN prevalence only in 2021. In contrast with the overall estimates, among adults who reported 4 or more conditions, no changes in CRN were found from 2020 to 2023 relative to 2019.
Discussion
The study findings show that the prevalence of US adults aged 18 to 64 years with chronic conditions who did not take their medication as prescribed to reduce cost remained similar at the COVID-19 pandemic onset in 2020 compared with the pre-pandemic 2019 year but decreased subsequently. We observed that changes in CRN in the pandemic and post-pandemic years vary according to the number of chronic conditions reported. We also found that CRN remained disproportionally higher for those diagnosed with 3 or 4 or more chronic conditions during the study years.
Our findings showed that decreases in CRN among individuals with chronic conditions seem to follow the pre-pandemic period reports, with declines attributed in part to the implementation of programs to increase coverage and access to medicines. 11 – 13 In the post-pandemic period, affordability and access to medication may have been impacted by enacted coverage protections that resulted in previously uninsured adults obtaining insurance coverage and more opportunities to afford medication. 31 Our findings also suggest that CRN can be addressed with government programs during times of financial hardship and that the level of financial support seemed appropriate to limit CRN increases among adults with multiple chronic conditions.
We observed CRN continued declining 1 and 2 years after the pandemic years despite the documented economic constraints associated with the pandemic, including the broad public health emergency regulations (eg, stay-at-home orders, business closures, limited air traveling). 3 , 32 CRN reduction in years 1 and 2 after the pandemic may reflect the impact of other economic relief programs such as Coronavirus Tax Relief (eg, child tax credits in 2020 and 2021) and broader distribution of coronavirus relief funds. This reduction in CRN is also consistent with recent reductions in other financial hardship factors, such as food insecurity and problems paying bills, observed in our analysis and recent reports. 17 , 18
Despite these encouraging findings, more than 15% of US adults aged 18 to 64 years with chronic conditions did not take their medication as prescribed to save costs in 2023. Within this group, some remain more vulnerable to this cost-saving behavior. CRN levels among these groups may have been impacted by the increases in out-of-pocket (OOP) costs for prescribed drugs, such as those reported for insulin in 2021-2022. 33
Our findings have several implications for policy and practice. First, during the pandemic years, US economic relief policies (eg, American Rescue Plan’s unemployment compensation or increased Medicaid matching funds) 34 may have attenuated potential increases in CRN levels among adults with chronic conditions. For policymakers, evaluating which programs effectively reduced the overall CRN can inform future government responses to maintain prescribed medication’s affordability during economic hardship. However, it still needs to be seen whether the CRN decline in post-pandemic years is associated with better population outcomes (eg, mortality or emergency department visits) and if these changes continued in 2024 after the pandemic rescue policies ended.
Second, although we showed overall improvement in CRN in the post-pandemic years relative to the pandemic, the high levels among those suggest that the more medically complex individuals remain vulnerable and may need additional assistance to afford medication. For providers, existing programs and interventions may benefit from integrating the patients’ preferences toward their disease management and related costs. 35 , 36 Screening for CRN during patient visits and adding cost-reducing strategies to workflows (eg, reviewing prescription concerns and considering lower-cost alternatives) can help mitigate nonadherence due to financial reasons. 37 New research is needed on the trade-offs adults with multiple chronic conditions make when engaging in CRN according to OOP cost thresholds and expected treatment effects. Such information can help design prescription coverage plans that better respond to their health care needs without limiting the medication effects and preventing them from improving outcomes.
LIMITATIONS
This study has several limitations. First, we used cross-sectional data; thus, we cannot directly assess whether the changes in CRN in the years after the pandemic reflect the implementation of COVID-19 economic relief programs and the sustained economic hardship. Second, we did not consider the specific medications or a broader set of combinations of conditions that may be highly susceptible to CRN (eg, high-cost medication treatments). Still, by focusing on 14 prevalent conditions, we sought to produce generalizable results to inform strategies for addressing CRN more broadly. Third, cost-saving behavior questions are self-reported, which creates the potential for response bias. Fourth, the estimates in this study, where we operationalized CRN by affirmative responses to any of 4 NHIS questions, are not directly comparable to those generated from measures using a different composite of questions. We opted for a CRN measure used in studies with NHIS data to capture more broadly changes in reduced adherence to prescribed medication because of cost. 38 , 39 Finally, we can only evaluate CRN in 2019 and after because of changes in NHIS sampling that year. 24 Nonetheless, we could still detect emerging CRN patterns among multiple chronic condition groups before and after the pandemic.
Conclusions
In the post-pandemic years 2021 and 2022, individuals with multiple chronic conditions experienced an overall reduction in medication cost-saving behaviors relative to pre-pandemic year 2019. During these periods, the enacted governmental financial stability measures supported continuous access to care and medications to manage multiple chronic conditions. Despite this improvement, CRN was trending to pre-pandemic levels by 2023 and individuals with multiple chronic conditions remain vulnerable, particularly those with 4 or more conditions. Our findings support the need for future strategies to identify individuals at risk of CRN and monitor potential changes in population health outcomes. Individuals with multiple chronic conditions may benefit from strategies to address prescription drug affordability, including improved access to pharmacy benefits, particularly during periods of sustained economic hardship.
REFERENCES
- 1. Piette JD, Rosland AM, Silveira MJ, Hayward R, McHorney CA. Medication cost problems among chronically ill adults in the US: did the financial crisis make a bad situation even worse? Patient Prefer Adherence . 2011;5:187-94. doi: 10.2147/PPA.S17363 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Zivin K, Ratliff S, Heisler MM, Langa KM, Piette JD. Factors influencing cost-related nonadherence to medication in older adults: A conceptually based approach. Value Health . 2010;13(4):338-45. doi: 10.1111/j.1524-4733.2009.00679.x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Cavallo A. Inflation with Covid consumption baskets. IMF Econ Rev . 2024;72(2):902-17. doi: 10.1057/s41308-023-00213-y [DOI] [Google Scholar]
- 4. American Hospital Association . The Financial Stability of America’s Hospitals and Health Systems Is at Risk as the Costs of Caring Continue to Rise. American Hospital Association; 2023:1-9. Accessed December 5, 2024. https://www.aha.org/system/files/media/file/2023/04/Cost-of-Caring-2023-The-Financial-Stability-of-Americas-Hospitals-and-Health-Systems-Is-at-Risk.pdf [Google Scholar]
- 5. Dusetzina SB, Besaw RJ, Whitmore CC, et al. Cost-related medication nonadherence and desire for medication cost information among adults aged 65 years and older in the US in 2022. JAMA Netw Open. 2023;6(5):e2314211. doi: 10.1001/jamanetworkopen.2023.14211 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Piette JD, Heisler M, Horne R, Caleb Alexander G. A conceptually based approach to understanding chronically ill patients’ responses to medication cost pressures. Soc Sci Med . 2006;62(4):846-57. doi: 10.1016/j.socscimed.2005.06.045 [DOI] [PubMed] [Google Scholar]
- 7. Taha MB, Valero-Elizondo J, Yahya T, et al. Cost-related medication nonadherence in adults with diabetes in the United States: The National Health Interview Survey 2013-2018. Diabetes Care. 2022;45(3):594-603. doi: 10.2337/dc21-1757 [DOI] [PubMed] [Google Scholar]
- 8. Van Alsten SC, Harris JK. Cost-related nonadherence and mortality in patients with chronic disease: A multiyear investigation, national health interview survey, 2000-2014. Prev Chronic Dis . 2020;17:E151. doi: 10.5888/pcd17.200244 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Blanchard J, Madden JM, Ross-Degnan D, Gresenz CR, Soumerai SB. The relationship between emergency department use and cost-related medication nonadherence among Medicare beneficiaries. Ann Emerg Med . 2013;62(5):475-85. doi: 10.1016/j.annemergmed.2013.04.013 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Iuga AO, McGuire MJ. Adherence and health care costs. Risk Manag Healthc Policy . 2014;7:35-44. doi: 10.2147/RMHP.S19801 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Xu WY, Shooshtari A, Jung J. Kyoungrae). Disparities in cost-related drug nonadherence under the Affordable Care Act. J Pharm Health Serv Res. 2019;10(2):177-85. doi: 10.1111/jphs.12295 [DOI] [Google Scholar]
- 12. Kennedy J, Wood EG. Medication costs and adherence of treatment before and after the Affordable Care Act: 1999-2015. Am J Public Health . 2016;106(10):1804-7. doi: 10.2105/AJPH.2016.303269 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Zhang J, Bhaumik D, Meltzer D. Decreasing rates of cost-related medication non-adherence by age advancement among American generational cohorts 2004-2014: A longitudinal study. BMJ Open. 2022;12(5):e051480. doi: 10.1136/bmjopen-2021-051480 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Zhang Y, Baik SH, Fendrick AM, Baicker K. Comparing local and regional variation in health care spending. N Engl J Med . 2012;367(18):1724-31. doi: 10.1056/NEJMsa1203980 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Mykyta L, Cohen R. Characteristics of Adults Aged 18–64 Who Did Not Take Medication as Prescribed to Reduce Costs: United States, 2021. National Center for Health Statistics; 2023. doi: 10.15620/cdc:127680 [DOI] [PubMed]
- 16. Patel MR, Piette JD, Resnicow K, Kowalski-Dobson T, Heisler M. Social determinants of health, cost-related nonadherence, and cost-reducing behaviors among adults with diabetes: Findings from the National Health Interview Survey. Med Care . 2016;54(8):796-803. doi: 10.1097/MLR.0000000000000565 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Kim-Mozeleski JE, Pike Moore SN, Trapl ES, Perzynski AT, Tsoh JY, Gunzler DD. Food insecurity trajectories in the US during the first year of the COVID-19 pandemic. Prev Chronic Dis . 2023;20:E03. doi: 10.5888/pcd20.220212 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Cohen R, Cha A. Problems Paying Medical Bills: United States, 2021. National Center for Health Statistics; 2023. doi: 10.15620/cdc:122191 [DOI] [PubMed]
- 19. Lyu W, Wehby GL. Effects of the Families First Coronavirus Response Act on coverage continuity and access for Medicaid beneficiaries. Inquiry . 2024;61:469580241282052. doi: 10.1177/00469580241282052 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. McIntyre A, Sommers BD, Aboulafia G, et al. Coverage and access changes during Medicaid unwinding. JAMA Health Forum . 2024;5(6):e242193. doi: 10.1001/jamahealthforum.2024.2193 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Lowey NM. Public Law 116-127 - Families First Coronavirus Response Act. 2020. Accessed December 12, 2024. https://www.congress.gov/116/plaws/publ127/PLAW-116publ127.htm
- 22. Nelson DB, Goldman AL, Zhang F, Yu H. Continuous Medicaid coverage during the COVID-19 public health emergency reduced churning, but did not eliminate it. Health Aff Sch. 2023;1(5):qxad055. doi: 10.1093/haschl/qxad055 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Blewett LA, Drew JAR, King ML, et al. IPUMS Health Surveys: National Health Interview Survey, Version 7.3. Published online 2023. doi: 10.18128/D070.V7.3 [DOI]
- 24. Zablotsky B, Lessem SE, Gindi RM, Maitland AK, Dahlhamer JM, Blumberg SJ. Overview of the 2019 National Health Interview Survey Questionnaire Redesign. Am J Public Health . 2023;113(4):408-15. doi: 10.2105/AJPH.2022.307197 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Division of Health Interview Statistics , National Center for Health Statistics. 2023 National Health Interview Survey (NHIS) Survey Description. Centers for Disease Control and Prevention U.S. Department of Health and Human Services; 2024. Accessed February 28, 2024. https://ftp.cdc.gov/pub/Health_Statistics/NCHS/Dataset_Documentation/NHIS/2023/srvydesc-508.pdf#page=28&zoom=100,64,84
- 26. United States Census Bureau . CPS Poverty Tables Footnotes. 2023. Accessed December 16, 2024. https://www.census.gov/topics/income-poverty/poverty/guidance/poverty-footnotes/cps-tables-footnotes.html#:∼:text=Ratio%20of%20income%20to%20poverty,householder%20may%20be%20either%20spouse
- 27. Lumley T. Complex Surveys: A Guide to Analysis Using R . John Wiley; 2010. [Google Scholar]
- 28. R Core Team . R: A Language and Environment for Statistical Computing. Published online 2023. Accessed January 24, 2024. https://www.R-project.org/
- 29. Wickham H, Chang W, Henry L, et al. Create Elegant Data Visualisations Using the Grammar of Graphics. Published online September 10, 2021. Accessed September 9, 2021. https://ggplot2.tidyverse.org/
- 30. Lee MJ, Khan MM, Salloum RG. Recent trends in cost-related medication nonadherence among cancer survivors in the United States. J Manag Care Spec Pharm . 2018;24(1):56-64. doi: 10.18553/jmcp.2018.24.1.56 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Cohen R, Cha A. Health Insurance Coverage: Early Release of Estimates from the National Health Interview Survey, 2022. National Center for Health Statistics. May 2023. National Center for Health Statistics; 2023. doi: 10.15620/cdc:127055 [DOI]
- 32. Office of Human Services Policy . COVID-19 and Economic Opportunity: Unequal Effects on Economic Need and Program Response. Assistant Secretary for Planning and Evaluation; 2021. Accessed October 24, 2023. https://aspe.hhs.gov/sites/default/files/2021-09/low-income-covid-19-impacts.pdf
- 33. Aitken M, Kleinrock M, Pritchett J. The Use of Medicines in the U.S. 2023: Usage and Spending Trends and Outlook to 2027. IQVIA Institute for Human Data Science; 2023. Accessed March 1, 2024. https://www.iqvia.com/insights/the-iqvia-institute/reports-and-publications/reports/the-use-of-medicines-in-the-us-2023
- 34. Text - H.R.1319 - 117th Congress (2021-2022): American Rescue Plan Act of 2021. 2021. Accessed March 4, 2024. https://www.congress.gov/bill/117th-congress/house-bill/1319/text
- 35. Hohmann NS, Hastings TJ, Jeminiwa RN, et al. Patient preferences for medication adherence financial incentive structures: A discrete choice experiment. Res Social Adm Pharm . 2021;17(10):1800-9. doi: 10.1016/j.sapharm.2021.01.018 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Laba TL, Essue B, Kimman M, Jan S. Understanding patient preferences in medication nonadherence: A review of stated preference data. Patient . 2015;8(5):385-95. doi: 10.1007/s40271-014-0099-3 [DOI] [PubMed] [Google Scholar]
- 37. Fiscella K, Venci J, Sanders M, Lanigan A, Fortuna R. A practical approach to reducing patients’ prescription costs. Fam Pract Manag . 2019;26(3):5-9. [PubMed] [Google Scholar]
- 38. Guadamuz JS, Qato DM. Citizenship status and cost-related nonadherence in the United States, 2017-2021. Health Serv Res . 2023;58(suppl 2):175-85. doi: 10.1111/1475-6773.14185 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Lago-Hernandez C, Nguyen NH, Khera R, Loomba R, Asrani SK, Singh S. Cost-related nonadherence to medications among US adults with chronic liver diseases. Mayo Clin Proc . 2021;96(10):2639-50. doi: 10.1016/j.mayocp.2021.02.026 [DOI] [PMC free article] [PubMed] [Google Scholar]