1 INTRODUCTION

The ongoing spread of the coronavirus disease 2019 (COVID-19) continues to challenge and overwhelm the US healthcare system. Prioritization decisions in response to constrained resources and hospital capacity have resulted in restriction and cancellation of elective medical procedures and the expansion of telehealth services for less-urgent conditions.1-4 Additionally, closure and limited operating hours of many medical practices, fear and uncertainty on the spread of COVID-19 and the financial distress caused by the pandemic's unemployment spells and reductions in working hours, led to substantial reductions in healthcare delivery. These disruptions might result in foregone or delayed medical care that patients would seek otherwise. However, delaying or foregoing care is not a suitable long-term option particularly for high-need individuals, as it may result in delayed diagnoses and treatment and impose adverse health, economic and social effects.

Multiple studies have explored changes in the provision of healthcare services during the first peak of the pandemic. Existing surveys indicate that 20%–50% of individuals reported foregoing or delaying medical care, even for serious chronic conditions.5-7 These estimates are in-line with evidence reporting large reductions in outpatient care, urgent and nonurgent emergency department visits and hospital admissions.8-18 Subsequent pandemic-related adverse short-term outcomes, ranging from mental health distress to mortality, have also been observed, though it is unclear whether lower healthcare utilization in the spring may have resulted in higher uptake of healthcare services during the late summer and fall months.19-22 The decision to forego and delay medical care during the COVID-19 pandemic was driven by many patient-, provider- and system-level factors.1, 23-25 Risk-related concerns and fear of exposure and contracting the virus in healthcare facilities, state lockdown policies and limited access to many medical practices may have further restricted patients' ability to seek medical care.5 To date, though, these studies have focused on healthcare utilization during the first months of the pandemic and were generally location-specific or used small-scale surveys, which limit broader nationwide implications.

This study provides evidence on foregone and delayed care as the pandemic advanced through the end of 2020. To do so, we used nationally representative data from three waves (August, October and December 2020) from the Urban Institute's Household Pulse Survey (HPS) for almost 160,000 nonelderly adults representing 136 million individuals in the United States. The survey includes a wide variety of sociodemographic and health-related characteristics, allowing us to describe the most important drivers associated with foregone and delayed care attributed to the COVID-19 pandemic. We then estimated with multivariable regressions the association between participants' sociodemographic and health-related characteristics and foregone and delayed care to grasp the social determinants of foregone or delayed care in the second half of the pandemic. Our findings can help stakeholders and policy makers to identify the evolving rates and factors associated with foregone and delayed care in the United States.

2 METHODS 2.1 Data

This study was a pooled, cross-sectional, retrospective analysis using publicly available data from the Urban Institute's HPS.26 The survey is a nationally representative sample selected by the Census Bureau to study the impact of COVID-19 on social and economic perspectives. The first phase of the survey began on April 23 and lasted until July 2020. The second phase of the survey took place from August 19 to October 26, 2020, with an updated questionnaire, while the third phase began on October 28, 2020, and data collection continued through October 2021.

2.2 Study population

We combined three waves of the second and third phases of the survey. Since the survey asks questions on foregone and delayed care incidents that occurred during the past 4 weeks, we used the waves corresponding to December 9–21 (Week 21), October 14–26 (Week 17) and August 19–31 (Week 13) in an attempt to avoid overlaps and provide a clear trend in care delivery change in the second phase of the pandemic. We focused on nonelderly adults ages 18–64 across all 50 States and the District of Columbia. Our initial sample included 267,711 survey respondents overall. We excluded respondents who were 65 years or older (63,877) and those with missing key variables of interest (48,099). The final sample included 155,825 respondents representative of the 135,835,598 million individuals in the United States after survey weights for national representation were applied. After applying weights, the distribution and means of the sociodemographic variables were similar to national estimates.27 The data were deidentified, and the study was deemed exempt by the Institutional Review Board at the Texas A&M University, College Station, Texas.

2.3 Foregone and delayed care

The two main outcomes of interest were foregone and delayed medical care attributed to the COVID-19 pandemic. We identified those using two dichotomous questions (yes or no) that were included in the survey, phrased as ‘At any time in the last 4 weeks, did you need medical care for something other than coronavirus, but did not get it because of the coronavirus pandemic?’ and ‘At any time in the last 4 weeks, did you delay getting medical care because of the coronavirus pandemic?’.

2.4 Independent variables

Covariates included respondents' sociodemographic, economic and health-related characteristics and factors, based on data collected in the survey. Sociodemographic and economic variables included age, gender, race/ethnicity, educational level, health insurance coverage, household income, household size (number of people residing in the household) and difficulty to pay for usual household expenses in the last 7 days. The latter was a mutually exclusive categorical measure obtained through the question ‘In the last 7 days, how difficult has it been for your household to pay for usual household expenses, including but not limited to food, rent or mortgage, car payments, medical expenses, student loans and so on? (Not at all difficult, A little/Somewhat difficult, Very difficult)’.

Health-related characteristics included self-reported health status and mental health problems during the past week. We created an ascending composite scale ranging from 0 to 12 to identify the severity of mental health problems (higher = more severe). This score was derived by combining four questions that asked participants about the frequency of feeling anxious/nervous, not being able to stop or control worrying, having little interest or pleasure in doing things, and feeling down, depressed, or hopeless over the last 7 days (0 = Not at all, 1 = Several days, 2 = More than half of the days, 3 = Nearly every day).

2.5 Statistical analyses

We conducted descriptive analyses to present the study population and stratified by the two outcomes of interest to compare characteristics of individuals with and without foregone and delayed medical care. We tested for statistically significant differences using Pearson's χ2 and Wald tests. We then used two multivariable logistic regressions to estimate the association between respondents' sociodemographic, economic and health-related characteristics and factors and foregone and delayed care. We included indicators for the time when each wave of the survey was rolled-out to control for heterogeneity as the pandemic evolved, and state fixed effects to control for unobserved state-level differences. We also replicated the regressions with the four mental health variables as separate predictors to evaluate the validity of the mental health composite score. Finally, we also conducted separate regressions for the two outcomes and each survey wave (six in total) to assess potential changes in predictors over time. All analyses were adjusted for survey weights. Data management and statistical analyses were conducted using Stata version 16.1 (StataCorp, College Station).

3 RESULTS 3.1 Descriptive analyses

Most participants reported good or better health status (80.7%) and low scores (0–2) on the mental health scale (40.8%) with an average of 4.2 (standard deviation [st.d.] = 0.02) (Table 1). Overall, 41.9% and 17.5% of respondents indicated that they had some difficulty and high difficulty in paying for usual household expenses in the last 7 days, respectively.

Table 1. Sociodemographic, economic and health-related characteristics and factors by foregone or delayed care in the last 4 weeks because of the coronavirus pandemic Overall Foregone Delayed No Yes p value No Yes p value 135,835,598 99,309,406 36,526,192 87,070,618 48,764,980 73.1% 26.9% 64.1% 35.9% Age: mean 42.6 (0.07)a 42.4 (0.08) 43.1 (0.14) <0.001 42.6 (0.09) 42.7 (0.12) 0.49 Age groups (%) <0.001 <0.001 18–24 9.0 9.5 7.7 9.4 8.2 25–34 22.9 23.1 22.1 22.8 22.9 35–44 22.7 22.4 23.4 22.1 23.7 45–54 21.8 21.5 22.6 21.7 22.1 55–64 23.6 23.5 24.3 24.0 23.1 Race/ethnicity (%) <0.001 <0.001 Non-Hispanic White 62.0 62.7 60.1 62.2 61.6 Non-Hispanic Black 11.0 10.6 12.0 11.2 10.5 Hispanic 17.5 17.3 18.3 5.8 5.2 Non-Hispanic Asian 5.6 6.0 4.4 17.3 18.0 Other 3.9 3.5 5.1 3.5 4.7 Gender (%) <0.001 <0.001 Male 48.4 50.3 43.1 51.1 43.4 Female 51.6 49.7 56.9 48.9 56.6 Income categories <0.001 <0.001 <$50,000 36.9 34.9 42.4 35.8 38.9 $50,000–$99,999 30.7 30.6 31.0 30.7 30.8 ≥$100,000 32.4 34.6 26.6 33.6 30.4 Difficulty paying usual expenses in the last 7 days (%) No difficulty 40.6 47.0 23.4 47.8 27.9 Some difficulty 41.9 39.9 47.3 39.0 47.0 High difficulty 17.5 13.1 29.3 13.2 25.1 Educational level (%) <0.001 0.001 High school or Less 34.9 34.3 33.4 36.1 30.5 College but no degree 21.7 20.7 24.1 20.8 23.0 Associate's/bachelor's 29.4 30.2 29.1 29.3 31.0 Graduate 14.0 14.8 13.4 13.8 15.5 Household sizeb: mean 3.5 (0.01) 3.4 (0.01) 3.6 (0.02) <0.001 3.4 (0.01) 3.5 (0.02) <0.001 Health insurance coverage (%) <0.001 <0.001 Private 61.7 64.5 54.2 63.3 58.9 Medicaid 15.6 13.4 21.7 13.8 18.9 Medicare 2.6 2.4 3.0 2.5 2.8 Veterans Affairs 2.9 2.6 3.7 2.6 3.4 Other 5.2 5.1 5.3 5.3 5.0 Uninsured 12.0 12.0 12.1 12.6 10.9 Health status (%) <0.001 <0.001 Excellent 18.4 21.7 9.5 22.3 11.5 Very good 32.1 34.9 24.7 34.6 27.8 Good 30.2 29.0 33.4 27.7 32.8 Fair 15.6 12.4 24.4 12.2 21.7 Poor 3.7 2.0 8.0 2.2 6.2 Mental health problems scorec: mean 4.2 (0.02) 3.5 (0.02) 6.1 (0.04) <0.001 3.4 (0.02) 5.7 (0.03) <0.001 Mental health problems during the past week (%) <0.001 <0.001 Low (0–2) 40.8 48.2 20.5 50.4 23.6 Medium (3–6) 32.8 31.9 34.1 30.9 36.2 High (≥7) 26.4 19.9 44.4 18.7 40.2 Waves <0.001 <0.001 Wave 1 (August 19–31) 40.6 40.3 41.3 40.0 41.7 Wave 2 (October 14–26) 33.3 34.2 30.9 34.4 31.2 Wave 3 (December 9–21) 26.1 25.5 27.8 25.6 27.1 Note: Authors' analysis of the Urban Institute's Household Pulse Survey data of the waves corresponding to December 9–21 (Week 21), October 14–26 (Week 17) and August 19–31 (Week 13).

Overall, 26.9% and 35.9% of respondents indicated that they forewent or delayed getting medical care at any time in the last 4 weeks because of the COVID-19 pandemic, respectively (Table 1). Compared to those who did not forego or delay medical care, individuals who forewent or delayed getting care were more likely to be women and have Medicaid or Veterans Affairs coverage (p < 0.001 for both). We also observed higher percentages of foregone or delayed medical care among those in the lowest income category (<$50,000) and those who faced any difficulty in paying usual household expenses in the past 7 days compared to those who did not face any issues (p < 0.001) (Table 1, Figure 1). Individuals who reported foregone or delayed care were also more likely to report fair or poor health and have higher scores on the mental health problems scale (p < 0.001 for both delayed and foregone care) (Table 1, Figure 2).

Percentages of individuals who forewent or delayed care in the last 4 weeks because of the coronavirus pandemic by difficulty to pay usual household expenses in the past 7 days and income

Percentages of individuals who forewent or delayed care in the last 4 weeks because of the coronavirus pandemic by health status and by mental health problems

The results of the multivariable logistic regressions are presented in Table 2. Compared to individuals who reported no difficulty paying usual household expenses in the past week, reporting any difficulty to pay usual household expenses increased the adjusted odds of foregoing and delaying medical care, particularly among those with the highest difficulties (forego: adjusted odds ratio [aOR] = 2.51, 95% confidence intervals [CIs] = 2.30–2.75, p < .001; delay: aOR = 2.24, 95% CI = 2.05–2.45, p < 0.001).

Table 2. Predictors of foregone or delayed medical care in the last 4 weeks because of the coronavirus pandemic Foregone Delayed OR 95% CI p value OR 95% CI p value Age groups 18–24 Ref. Ref. 25–34 1.21 1.06–1.38 0.006 1.14 1.02–1.29 0.027 35–44 1.38 1.21–1.57 <0.001 1.27 1.13–1.43 <0.001 45–54 1.45 1.27–1.65 <0.001 1.27 1.13–1.43 <0.001 55–64 1.55 1.36–1.77 <0.001 1.33 1.18–1.50 <0.001 Race/ethnicity Non-Hispanic White Ref.