Abstract
Objective
To examine sociodemographic factors associated with having unmet needs in medications, mental health, and food security among older adults during the COVID‐19 pandemic.
Data Sources and Study Setting
Primary data and secondary data from the electronic health records (EHR) in an age‐friendly academic health system in 2020 were used.
Study Design
Observational study examining factors associated with having unmet needs in medications, food, and mental health.
Data Collecting/Extraction Methods
Data from a computer‐assisted telephone interview and EHR on community‐dwelling older patients were analyzed.
Principle Findings
Among 3400 eligible patients, 1921 (53.3%) (average age 76, SD 11) responded, with 857 (45%) of respondents having at least one unmet need. Unmet needs for medications were present in 595 (31.0%), for food in 196 (10.2%), and for mental health services in 292 (15.2%). Racial minorities had significantly higher probabilities of having unmet needs for medicine and food, and of being referred for services related to medications, food, and mental health. Patients living in more resource‐limited neighborhoods had a higher probability of being referred for mental health services.
Conclusions
Age‐friendly health systems (AFHS) and their recognition should include assessing and addressing social risk factors among older adults. Proactive efforts to address unmet needs should be integral to AFHS.
Keywords: aging, healthy place index, neighborhood resources, social determinants of health, unmet needs
What is known on this topic
Both residential geographical area and person‐level sociodemographic factors are associated with having unmet needs for important services.
Little is known about the association of geographical area resource limitations with unmet needs and the probability of receiving services to address unmet needs after accounting for individual factors.
What this study adds
Individual sociodemographic and health risk factors are associated with having unmet needs in what matters to older adults (medications, mental health, and food security).
Geographical area resource limitations are associated with the probabilities of receiving services to address mental health needs after accounting for the individual‐level factors.
Higher levels of anxiety were associated with having unmet needs for medications and the probability of being referred for services to address food insecurity after accounting for other factors.
1. INTRODUCTION
The age‐friendly health systems (AFHS) framework incorporates four elements (What Matters, Medication, Mentation, and Mobility) known as the 4 M's to guide the quality care of older adults. 1 AFHSs face challenges posed by the ongoing COVID‐19 pandemic, which has exacerbated multiple barriers to health care for vulnerable community‐dwelling older adults. Studies suggest that older adults were more likely to forgo or postpone health services due to fears of contracting the virus. 2 , 3 They may experience more difficulties in meeting their basic needs for food, medicine, and mental health. While medicine and mental health have become established services of AFHS, 4 discussion regarding food insecurity is sparse in the existing AFHS literature and more connected to the public health system 5 or community‐based organizations. 6 As food‐insecure older adults have been documented to report fair or poor health status 7 and have limitations in activities of daily living comparable to those of food‐secure older adults 14 years older, 8 assessing food security can be important for AFHS. Yet, it is currently not a standard part of becoming AFHS‐recognized.
We are an academic health system with three of our component entities (a multispecialty medical center, a critical intensive care unit, and a Medicine for Seniors service) recognized as level 2 AFHS by the Institute for health improvement. In order to identify high‐risk, community‐dwelling older adults with unmet needs and refer them for follow‐up services 9 during the COVID‐19 pandemic, we proactively contacted older patients soon after the California stay‐at‐home order was implemented. We were particularly interested in understanding the characteristics of older adults that were associated with high needs for services during the pandemic. 10 , 11 , 12 Given the reported association of geography‐based social vulnerability, we also explored the relationship between a patient's address mapped onto the Health Place Index (HPI) and their needs for services. 13 , 14 We acknowledge that all health systems are facing challenges with respect to social determinants of health, not just those that are AFHS‐recognized. We hope to contribute new insights to AFHS and other health systems.
2. METHODS
Using a computer‐assisted telephone interview approach, nursing trainees 9 called patients meeting the following criteria: (1) served by the health system's primary care providers between April and December 2020, (2) belonged to the health systems' Medicare Accountable Care Organization or a Medicare Advantage plan, (3) community‐dwelling inclusive of those living in communal environments such as assisted living facilities, residential care facilities, and (4) had a risk score of 11 or greater. This risk score was customized by the health system to capture a patient's general health risk‐based several dimensions: patient age, previous hospitalizations or emergency department visits, chronic conditions (including cancer, heart failure, chronic kidney disease, hypertension, chronic obstructive pulmonary disease, cardiovascular disease, diabetes, on dialysis, depression, schizophrenia, bipolar disorder, substance dependence, substance abuse, HIV, and major organ transplant), and empanelment in primary care. Except for primary care empanelment, all the aforementioned factors additively contribute to the risk score (i.e., a higher score indicates higher risk). One point is subtracted from the aggregate if the patient has a designated primary care provider.
People with dementia or those already enrolled in the Complex Case Management (CCM) Program were excluded. The CCM serves eligible patients based on their risk score and history of using emergency department or inpatient services for advanced stages of heart failure, diabetes, and chronic obstructive pulmonary disease. CCM staff conducts ongoing comprehensive assessments of enrollees' health and social needs, which include food security, medications, transportation, and social support. Therefore, those enrolled in CCM were excluded from this outreach. We asked a question near the end of the interview, however, on if they wanted a nurse to follow up with them. Those patients who wanted a follow‐up call were referred to CCM for further assessments and follow‐up.
Applying the inclusion and exclusion criteria resulted in the identification of 3400 eligible patients. Nursing students were trained to follow a standardized script and record patient responses in Qualtrics (Qualtrics, Provo, UT), a survey platform. Details of the interview approach and training of callers are described elsewhere. 9
The interview asked patients about their access to food and medicine and their mental health. Unmet need for food was identified if a patient had less than 3 days of food in their home, reported having financial barriers to buying groceries, or responded “no” to “are you able to access groceries ok?” Unmet need for medications was defined as present if a patient had 30 or less days of medication or responded “yes” to having concerns about financial hardship, polypharmacy, or questions about medication. Unmet need for mental health services was defined as present if a patient reported not having stress management strategies (e.g., doing physical activities or creative activities, playing with pets), or scored 3 or higher on PHQ‐2 or on GAD‐2. 15 Some patients declined to answer PHQ or GAD resulting in missing data and possibly underestimating the unmet needs for mental health. Recognizing this communication challenge, we conducted more coaching and role‐playing during the orientation of nursing students to prepare them better in asking questions about mental health before they started calling patients. Patients identified as having unmet needs were referred to additional services.
The interviewers were trained to also pay attention to how the patient was communicating. In cases where their scores did not match the concerns they were expressing verbally, callers would still refer these patients for additional services. For example, if a patient talked about not being able to get out of bed, even if their PHQ2 score was lower than 3, the patient would be referred to licensed clinical social worker for further assessment. Forty‐six individuals (13, 15, and 18 for follow‐up services for medication, food, and mental health services, respectively) were referred, although they did not meet the thresholds for having unmet needs for those services based strictly on the numeric values of their answers.
From the electronic health records (EHR), we obtained data on patient age, gender, race/ethnicity, health status, insurance, and social vulnerability. Age was measured by years since birth while gender was female versus nonfemale (including male, transgender male/male‐to‐female, transgender female/female‐to‐male, and others). Based on self‐report, the patient's race/ethnicity was measured by binary variables for Asian, Black/African American, Hispanic/Latino, and other (American Indian/Alaska Native, Native Hawaiian or Other Pacific Islander, and other). Health status was measured by the risk score described earlier. Medicare Accountable Care Organization (ACO) status was measured as a binary variable to differentiate members of a Medicare ACO or a Medicare Advantage plan. Social vulnerability was measured by the Healthy Places Index (HPI) percentile, a composite index of community well‐being for California from publicly available census‐tract data on 25 place‐based community characteristics, including economic (e.g., per capita income), education, social, transportation, and health insurance coverage. 13 , 14 California's over 1650 zip codes have been divided into four quarters based on the HPI Index. 13 , 14 Our organization had incorporated the HPI into the EHR, so the information is readily available to clinicians and data analysts. We extracted that data from the EHR and assigned patients to quartiles according to scores on the California Healthy Places Index, 13 , 14 with quartile 1 indicating least advantaged and quartile 4 indicating most advantaged.
Statistical Analysis: Univariate and bivariate analyses and the margins package in R version 4.0.4 were used. Two sets of regression analyses were done. The first set examined the factors associated with the probability of having unmet needs for food, medication, or mental health services. The second set consisted of conditional logistic regression analyses to examine the factors associated with the probability of being referred for services to address each unmet need, conditional upon having unmet needs for food, medicine, or mental health services, respectively. The 41 patients who received a referral for services but did not report having unmet needs were excluded from this analysis. The covariates included age, gender, race/ethnicity, Medicare ACO status, risk score, and HPI quartile. Sensitivity analyses examined the HPI quartile as a covariate to evaluate the degree of model fit. The results supported, including the HPI quartile as a covariate in all models in the regression analyses.
This quality improvement study was approved by the University of California San Diego Aligning and Coordinating Quality Improvement, Research, and Evaluation Committee and was deemed exempt from institutional review board approval.
3. RESULTS
Among the approximately 19,903 patients served by the health system, 3400 met the inclusion criteria and had been attempted for outreach. Among those, 1479 either declined or were unable to be reached after two attempts, and 1921 unique patients completed the interview. Among the latter, 206 were interviewed multiple times due to their requests for follow‐up. This analysis only included responses to the initial interviews.
Table 1 displays characteristics of the respondents (N = 1921) and the underlying covered population (N = 19,903). Unmet needs for medications existed among 595 (31.0%) patients, among whom 98 (16.5%) were referred to nurses or pharmacists for services to address their medication needs. Regarding food, 196 (10.2%) faced food insecurity, among whom 30 (15.3%) were referred to social workers in the health system to improve their access to food. Regarding mental health, 292 (15.2%) had unmet needs, among whom 70 (24%) were referred to social workers to address their mental health needs. While 857 (45%) had at least one unmet need, 73 had unmet needs for medications and food, 131 had unmet needs for medications and mental health, 41 for food and mental health, and 19 had unmet needs in all three areas. All patients with unmet needs were offered the opportunity to talk to a care team member for additional services. Some declined the offer. Only those who accepted the offer were referred for additional services, including being assessed by a member of the CCM team for CCM eligibility assessment.
TABLE 1.
Characteristics of patient participants and covered patient population
| Participants (N = 1921) | Covered patient population (N = 19,903) | ||||
|---|---|---|---|---|---|
| N (%) or Mean (SD) | N (%) or Mean (SD) | N (%) or Mean (SD) | |||
| With unmet needs for medications | Age | 76 (10.5) | Age | 46 (21.1) | |
| Yes | 595 (31.0%) | Gender | Gender | ||
| No | 979 (51.0%) | Female | 995 (51.8%) | Female | 11,004 (55.3%) |
| Missing | 347 (18.1%) | Nonfemale | 926 (48.2%) | Nonfemale | 8899 (44.7%) |
| With unmet needs for food | Race/Ethnicity | Race/Ethnicity | |||
| Yes | 196 (10.2%) | Asian | 135 (7.0%) | Asian | 3489 (17.5%) |
| No | 1368 (71.2%) | Black | 66 (3.4%) | Black | 805 (4.1%) |
| Missing | 357 (18.6%) | Caucasian/White | 1428 (74.3%) | Caucasian/White | 7623 (38.3%) |
| With unmet needs for mental health services | Hispanic/Latino | 132 (6.9%) | Hispanic/Latino | 3351 (16.8%) | |
| Yes | 292 (15.2%) | Other | 160 (8.3%) | Other | 4635 (23.3%) |
| No | 1196 (62.3%) | Medicare ACO | Medicare ACO | ||
| Missing | 433 (22.5%) | Yes | 1595 (83.0%) | Yes | 9968 (50.1%) |
| Referred for medication services | No | 269 (14.0%) | No | 4836 (24.3%) | |
| Yes | 98 (16.5%) | Missing | 57 (3.0%) | Missing | 5099 (25.6%) |
| No | 398 (66.9%) | Risk score | Risk score | ||
| Missing | 99 (16.6%) | Mean (SD) | 13.4 (5.3) | Mean (SD) | 3.2 (4.5) |
| Referred for food security services | Median [Min, Max] | 13 [0, 33] | Median [Min, Max] | 1 [0, 36] | |
| Yes | 30 (15.3%) | Missing | 199 (10.4%) | Missing | 0 (0.0%) |
| No | 146 (74.5%) | HPI percentile (individual patients) | HPI percentile (individual patients) | ||
| Missing | 20 (10.2%) | 1st quartile | 123 (6.4%) | 1st quartile | 1360 (6.8%) |
| Referred for mental health services | 2nd quartile | 264 (13.7%) | 2nd quartile | 2896 (14.6%) | |
| Yes | 70 (24.0%) | 3rd quartile | 574 (29.9%) | 3rd quartile | 6185 (31.1%) |
| No | 186 (63.7%) | 4th quartile | 875 (45.5%) | 4th quartile | 9119 (45.8%) |
| Missing | 36 (12.3%) | Missing | 85 (4.4%) | Missing | 343 (1.7%) |
| PHQ‐2 score | |||||
| At least 3 | 105 (5.47%) | ||||
| Less than 3 | 1309 (68.1%) | ||||
| Missing | 507 (26.4%) | ||||
| GAD‐2 score | |||||
| At least 3 | 145 (7.6%) | ||||
| Less than 3 | 1271 (66.2%) | ||||
| Missing | 505 (26.3%) | ||||
Abbreviations: ACO, accountable care organization; GAD, general anxiety disorder; HPI, healthy place index; PHQ, personal health questionnaire.
Demographically, the average age of participants was 76 (SD 10.5); 52% were female and 48% nonfemale (including male, transgender male/female, and other). The racial composition was 74% White, 7% Asian, 7% Hispanic, 3% Black, and 8% others. Medicare ACO members accounted for 83.0% of the patients. Patients in higher quartiles of HPI were over‐represented than patients in lower HPI quartiles. The 1920 patients lived in 608 unique HPI percentile values: 78 (12.8%) in the first quartile, 132 (21.7%) in the second, 193 (31.7%) in the third, and 205 (33.7%) in the fourth quartile. The larger underlying population was younger, and the average age was 46 (SD21.1), with a higher proportion of Asians (17.5%), Blacks (4.1%), Hispanics (16.8%), and others (23.3%) and a lower proportion of Whites (38.3%). They also had lower risk scores. The distribution across the HPI quartiles was similar between the participants and the larger population. Because the goal of the project was to reach vulnerable older adults rather than a representative sample of the underlying covered population, the dissimilarities in some characteristics did not cause concern.
Bivariate analyses suggest variations in rates of unmet needs for medication, food, or mental health services across patient age and race/ethnicity (Table 2). For example, those with unmet needs for mental health services were younger, with an average age of 72.8, compared to those without unmet mental health needs, with an average age of 76.1. Blacks (51.7%) and Latinos (45.9%) led in rates of having unmet medication needs compared to Whites (36.1%) and Asians (33.3%). Latinos (22.5%) and Blacks (22.4%) also led in rates of food insecurity compared to Asians (16.2%) and Whites (10.6%). Additional bivariate results suggested that patients in Medicare ACO had a lower rate of having unmet needs for mental health services (18.8%) compared with patients in Medicare Advantage plans (25.3%). Patients whose PHQ‐2 scored 3 or higher had a higher rate of having unmet needs for medications compared to those whose PHQ‐2 score was 2 or lower (37.1%). Patients whose GAD‐2 scored 3 or higher had a higher rate (53.2%) of having unmet needs for medications compared to those with GAD‐2 score at 2 or lower (36.6%). Patients with a GAD‐2 score greater or equal to 3 also had a higher rate of food insecurity (18.1%) compared to those whose GAD‐2 was 2 or lower (11.8%). Furthermore, while HPI was not associated with significant differences in having unmet needs for any of the three services at the conventional level (p < 0.05), those in the first (15.5%) and second (15.6%) quartiles and those missing address data to calculate HPI (17.2%) had higher rates of food insecurity compared to those in the third (13.3%) and fourth (10.2%) quartiles. Those in the first HPI quartile (42.9%) and missing HPI data (48.5%) also had a higher probability of having unmet needs for medicine than those in the other quartiles (2nd 34.0%, 3rd 39.9%, 4th 35.8%). Furthermore, those referred for food insecurity were younger, nonfemale, Black, and Hispanic. The same groups also had higher rates of referrals for mental health services.
TABLE 2.
Bivariate statistics on patient characteristics associated with unmet needs and referral for additional services
| Had Unmet Needs | Referred for Services to Address Unmet Needs | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Medication | Food | Mental Health | Medication | Food | Mental Health | |||||||
| Yes (N = 595) | No (N = 979) | Yes (N = 196) | No (N = 1368) | Yes (N = 292) | No (N = 1196) | Yes (N = 98) | No (N = 398) | Yes (N = 30) | No (N = 146) | Yes (N = 70) | No (N = 186) | |
| N (%) or Mean (SD) | N (%) or Mean (SD) | N (%) or Mean (SD) | N (%) or Mean (SD) | N (%) or Mean (SD) | N (%) or Mean (SD) | |||||||
| Age | 74.8 (11.0) | 76.1 (10.2) | 74.3 (12.1) | 75.8 (10.3) | 72.8 (11.5) | 76.1 (10.2) | 74.3 (10.1) | 75.1 (11.1) | 65.3 (14.4) | 76.5 (10.6) | 69.3 (12.0) | 74.3 (10.9) |
| Gender | ||||||||||||
| Female | 298 (36.3%) | 524 (63.7%) | 113 (13.9%) | 701 (86.1%) | 150 (19.4%) | 623 (80.6%) | 47 (19.2%) | 198 (80.8%) | 10 (9.80%) | 92 (90.2%) | 31 (24.2%) | 97 (75.8%) |
| Nonfemale | 297 (39.5%) | 455 (60.5%) | 83 (11.1%) | 667 (88.9%) | 142 (19.9%) | 573 (80.1%) | 51 (20.3%) | 200 (79.7%) | 20 (27.0%) | 54 (73.0%) | 39 (30.5%) | 89 (69.5%) |
| Race/Ethnicity | ||||||||||||
| Asian | 428 (36.1%) | 759 (63.9%) | 125 (10.6%) | 1051 (89.4%) | 227 (20.3%) | 892 (79.7%) | 67 (18.1%) | 304 (81.9%) | 13 (11.1%) | 104 (88.9%) | 46 (22.8%) | 156 (77.2%) |
| Black/African American | 32 (33.3%) | 64 (66.7%) | 16 (16.2%) | 83 (83.8%) | 14 (15.1%) | 79 (84.9%) | 8 (36.4%) | 14 (63.6%) | 2 (16.7%) | 10 (83.3%) | 3 (30.0%) | 7 (70.0%) |
| Caucasian/White | 30 (51.7%) | 28 (48.3%) | 13 (22.4%) | 45 (77.6%) | 9 (16.1%) | 47 (83.9%) | 6 (26.1%) | 17 (73.9%) | 5 (50.0%) | 5 (50.0%) | 7 (77.8%) | 2 (22.2%) |
| Hispanic/Latino | 50 (45.9%) | 59 (54.1%) | 25 (22.5%) | 86 (77.5%) | 20 (18.7%) | 87 (81.3%) | 9 (23.7%) | 29 (76.3%) | 9 (40.9%) | 13 (59.1%) | 9 (52.9%) | 8 (47.1%) |
| Other | 55 (44.4%) | 69 (55.6%) | 17 (14.2%) | 103 (85.8%) | 22 (19.5%) | 91 (80.5%) | 8 (19.0%) | 34 (81.0%) | 1 (6.67%) | 14 (93.3%) | 5 (27.8%) | 13 (72.2%) |
| Medicare ACO | ||||||||||||
| No | 85 (42.1%) | 117 (57.9%) | 29 (14.3%) | 174 (85.7%) | 48 (25.3%) | 142 (74.7%) | 9 (14.8%) | 52 (85.2%) | 10 (35.7%) | 18 (64.3%) | 11 (34.4%) | 21 (65.6%) |
| Yes | 488 (36.8%) | 838 (63.2%) | 159 (12.1%) | 1158 (87.9%) | 236 (18.8%) | 1020 (81.2%) | 89 (21.2%) | 330 (78.8%) | 18 (12.7%) | 124 (87.3%) | 57 (26.3%) | 160 (73.7%) |
| Missing | 22 (47.8%) | 24 (52.2%) | 8 (18.2%) | 36 (81.8%) | 8 (19.0%) | 34 (81.0%) | 0 (0%) | 16 (100%) | 2 (33.3%) | 4 (66.7%) | 2 (28.6%) | 5 (71.4%) |
| PHQ‐2 score | ||||||||||||
| At least 3 | 50 (49.0%) | 52 (51.0%) | 18 (17.3%) | 86 (82.7%) | 12 (29.3%) | 29 (70.7%) | 5 (38.5%) | 8 (61.5%) | ||||
| Less than 3 | 481 (37.1%) | 814 (62.9%) | 157 (12.0%) | 1148 (88.0%) | 69 (17.2%) | 331 (82.8%) | 22 (15.2%) | 123 (84.8%) | ||||
| Missing | 64 (36.2%) | 113 (63.8%) | 21 (13.5%) | 134 (86.5%) | 17 (30.9%) | 38 (69.1%) | 3 (16.7%) | 15 (83.3%) | ||||
| GAD‐2 score | ||||||||||||
| At least 3 | 75 (53.2%) | 66 (46.8%) | 26 (18.1%) | 118 (81.9%) | 16 (27.1%) | 43 (72.9%) | 9 (45.0%) | 11 (55.0%) | ||||
| Less than 3 | 460 (36.6%) | 797 (63.4%) | 150 (11.8%) | 1117 (88.2%) | 65 (16.8%) | 321 (83.2%) | 18 (12.9%) | 121 (87.1%) | ||||
| Missing | 60 (34.1%) | 116 (65.9%) | 20 (13.1%) | 133 (86.9%) | 17 (33.3%) | 34 (66.7%) | 3 (17.6%) | 14 (82.4%) | ||||
| HPI percentile | ||||||||||||
| 1st quartile | 45 (42.9%) | 60 (57.1%) | 16 (15.5%) | 87 (84.5%) | 27 (27.6%) | 71 (72.4%) | 9 (25.7%) | 26 (74.3%) | 3 (25.0%) | 9 (75.0%) | 9 (42.9%) | 12 (57.1%) |
| 2nd quartile | 73 (34.0%) | 142 (66.0%) | 33 (15.6%) | 178 (84.4%) | 39 (19.5%) | 161 (80.5%) | 14 (22.2%) | 49 (77.8%) | 6 (19.4%) | 25 (80.6%) | 14 (41.2%) | 20 (58.8%) |
| 3rd quartile | 190 (39.9%) | 286 (60.1%) | 64 (13.3%) | 417 (86.7%) | 95 (21.0%) | 358 (79.0%) | 29 (18.5%) | 128 (81.5%) | 12 (20.7%) | 46 (79.3%) | 25 (30.5%) | 57 (69.5%) |
| 4th quartile | 255 (35.8%) | 457 (64.2%) | 72 (10.2%) | 633 (89.8%) | 119 (17.6%) | 557 (82.4%) | 44 (20.2%) | 174 (79.8%) | 6 (9.09%) | 60 (90.9%) | 20 (18.5%) | 88 (81.5%) |
| Missing | 32 (48.5%) | 34 (51.5%) | 11 (17.2%) | 53 (82.8%) | 12 (19.7%) | 49 (80.3%) | 2 (8.7%) | 21 (91.3%) | 3 (33.3%) | 6 (66.7%) | 2 (18.2%) | 9 (81.8%) |
| Risk score | 14.1 (5.08) | 13.3 (5.19) | 13.9 (5.32) | 13.6 (5.16) | 14.2 (4.96) | 13.4 (5.19) | 13.6 (5.01) | 14.0 (5.13) | 16.4 (5.36) | 13.1 (5.23) | 13.6 (5.17) | 13.8 (4.72) |
| Missing | 53 (37.1%) | 90 (62.9%) | 18 (13.4%) | 116 (86.6%) | 22 (17.2%) | 106 (82.8%) | 2 (4.7%) | 41 (95.3%) | 4 (25.0%) | 12 (75.0%) | 2 (10.5%) | 17 (89.5%) |
Abbreviations: ACO, accountable care organization; GAD, general anxiety disorder; HPI, healthy place index; PHQ, Personal health questionnaire.
Regression analyses showed racial factors significantly associated with the average marginal effects (ME) of having unmet needs and of being referred for additional services (Table 3). Compared with Whites, Black patients (ME 0.14, 90% CI: −0.0029, 0.29) had 0.14 percentage points higher probability of being at risk of not having enough medications. Those with anxiety (GAD2 > 2) (ME 0.13, 90% CI: 0.04, 0.23) and a higher risk score (ME 0.0059, 90% CI: 0.00069, 0.011) were associated with a slightly higher probability of not having enough medications. Blacks (ME 0.11, 90% CI: −0.01, 0.22), and Hispanics (ME 0.11, 90% CI: 0.02, 0.2) had a significantly higher probability of food insecurity. While older age, Blacks and Hispanics were associated with lower probabilities of having unmet needs for mental health services, higher risk scores were associated with a higher probability of having unmet needs for mental health services.
TABLE 3.
Factors associated with having unmet needs and with being referred for services to address unmet needs
| Having unmet needs (Average marginal effect, 90% CI) | Being referred for services to address unmet needs (Average marginal effect, 90% CI) | |||||
|---|---|---|---|---|---|---|
| Medication | Food | Mental health | Medication | Food | Mental health | |
| Age | −0.00087 | −0.00096 | −0.0044*** | 0.00042 | −0.0043** | −0.0074*** |
| [−0.0036, 0.0018] | [−0.0027, 0.00081] | [−0.0063, −0.0024] | [−0.0036, 0.0044] | [−0.0083, −0.00026] | [−0.013, −0.0019] | |
| Gender: Nonfemale | 0.019 | −0.029 | −0.0096 | 0.027 | 0.15*** | 0.066 |
| [−0.036, 0.073] | [−0.066, 0.0072] | [−0.052, 0.033] | [−0.052, 0.11] | [0.04, 0.26] | [−0.049, 0.18] | |
| Race/Ethnicity: Asian | −0.011 | 0.048 | −0.058 | 0.30** | 0.13 | 0.04 |
| [−0.12, 0.1] | [−0.036, 0.13] | [−0.14, 0.023] | [0.06, 0.54] | [−0.1, 0.37] | [−0.29, 0.37] | |
| Race/Ethnicity: African American | 0.14* | 0.11* | −0.076* | 0.06 | 0.24 | 0.53*** |
| [−0.0029, 0.29] | [−0.01, 0.22] | [−0.16, 0.012] | [−0.14, 0.26] | [−0.067, 0.55] | [0.25, 0.81] | |
| Race/Ethnicity: Hispanic/Latino | 0.073 | 0.11** | −0.066* | 0.057 | 0.084 | 0.12 |
| [−0.037, 0.18] | [0.023, 0.2] | [−0.14, 0.003] | [−0.094, 0.21] | [−0.088, 0.25] | [−0.12, 0.36] | |
|
Race/Ethnicity: Other |
0.029 | 0.039 | −0.029 | 0.034 | −0.14*** | 0.11 |
| [−0.075, 0.13] | [−0.036, 0.11] | [−0.11, 0.048] | [−0.12, 0.18] | [−0.21, −0.077] | [−0.14, 0.37] | |
|
Medicare ACO |
−0.054 | 0.00076 | −0.037 | 0.036 | −0.19** | −0.16 |
| [−0.14, 0.035] | [−0.055, 0.057] | [−0.11, 0.033] | [−0.083, 0.16] | [−0.37, −0.008] | [−0.37, 0.051] | |
|
PHQ‐2 score at least 3 |
0.04 | 0.036 | 0.061 | 0.011 | ||
| [−0.073, 0.15] | [−0.034, 0.1] | [−0.07, 0.19] | [−0.18, 0.2] | |||
|
GAD‐2 score at least 3 |
0.13*** | 0.031 | 0.095 | 0.15* | ||
| [0.04, 0.23] | [−0.03, 0.091] | [−0.019, 0.21] | [−0.0083, 0.31] | |||
|
HPI percentile: 1st quartile |
0.0016 | −0.017 | 0.065 | 0.026 | −0.014 | 0.0071 |
| [−0.11, 0.11] | [−0.079, 0.046] | [−0.031, 0.16] | [−0.14, 0.19] | [−0.18, 0.15] | [−0.2, 0.21] | |
|
HPI percentile: 2nd quartile |
−0.06 | 0.027 | 0.011 | 0.029 | 0.024 | 0.19* |
| [−0.14, 0.021] | [−0.032, 0.085] | [−0.053, 0.076] | [−0.1, 0.16] | [−0.11, 0.15] | [−0.0029, 0.38] | |
| HPI percentile: 3rd quartile | 0.017 | 0.022 | 0.023 | −0.033 | 0.076 | 0.081 |
| [−0.046, 0.079] | [−0.021, 0.065] | [−0.026, 0.071] | [−0.12, 0.053] | [−0.044, 0.2] | [−0.052, 0.21] | |
|
Risk score |
0.0059** | 0.0013 | 0.0039* | −0.0036 | 0.0046 | −0.0043 |
| [0.00069, 0.011] | [−0.0022, 0.0048] | [−0.00019, 0.008] | [−0.011, 0.0043] | [−0.0036, 0.013] | [−0.016, 0.008] | |
| N | 1255 | 1266 | 1344 | 399 | 140 | 233 |
| AIC | 1660.93 | 942.19 | 1327.17 | 404.07 | 94.74 | 271.37 |
| BIC | 1732.82 | 1014.20 | 1389.61 | 459.92 | 135.93 | 312.78 |
| Pseudo R2 | 0.04 | 0.04 | 0.04 | 0.06 | 0.58 | 0.19 |
Note: ***p < 0.01; **p < 0.05; *p < 0.10. Comparison groups were: female, White, Non‐Medicare ACO, PHQ‐2 less than 3, GAD‐2 less than 3, HPI in 4th quartile.
Abbreviations: ACO, accountable care organization; GAD, general anxiety disorder; HPI, healthy place index; PHQ, personal health questionnaire.
Table 3 also displays the factors associated with being referred for additional services. Among them, Asians had 0.30 percentage points higher probability for being referred for medication services (ME 0.30, 90% CI: 0.06, 0.54). A lower probability for food insecurity referral was found among those in Medicare ACO (ME −0.19, 90% CI: −0.37, −0.008), while higher probabilities were found among nonfemales (ME 0.15, 90% CI: 0.04, 0.26), and those with anxiety (GAD2 > 2) (ME 0.15, 90% CI: −0.008, 0.31). A lower probability for mental health services referral was found among older age (ME −0.007, 90% CI: −0.013, −0.0019). Groups with higher probabilities for mental health services referral were Blacks (ME 0.53, 90% CI: 0.25, 0.81) and those living in the 2nd quartile of HPI (ME 0.19, 90% CI: −0.0029, 0.38) when compared with Whites and those living in the 4th quartile of HPI, respectively.
4. DISCUSSION
Overall, 45% of older adults reached by this effort had at least one unmet need. This proactive outreach to identify and address these unmet needs highlighted the imperative for AFHS to integrate social determinants of health in serving its patients, besides having highly ranked medicine for seniors, geriatric emergency departments, and inpatient hospitals. It also uncovered the association between anxiety symptoms and a higher probability of unmet needs for medication and of referral for food insecurity, whereas depression symptoms did not appear to be a significant factor. Had we only asked about depression, we would have missed the influence of anxiety and lost some opportunities. This is an actionable observation for health care providers to recognize unmet mental health needs that include anxiety and to ask about medications or food needs.
Racial minority patients had a significantly higher probability of having unmet needs for medicine, food, and mental health services. They also had a higher probability of being referred for services to address those unmet needs. Furthermore, patients living in areas with lower resources (lower quartiles of HPI distribution) had higher rates of experiencing food insecurity, insufficient medication supplies, and higher probabilities of being referred for mental health services. These findings resonate with a recent report on the CMS accountable health community initiative aimed at addressing the critical gap between clinical care and community services in a health delivery system. 6 Community‐based organizations partnering with a health system identified remarkable health‐related social needs and provided a multitude of support services among Medicare and Medicaid beneficiaries, including food and housing assistance. 6
This study has several limitations. First, the study partially addresses two components of the 4Ms of AFHS 1 related to medications and some mentation (depression, not dementia or delirium). While food security clearly matters to older adults, the study lacks direct measure of their health care wishes and mobility, the other 2Ms. However, these readily available measures, although imperfect, can be used by health systems to improve their age‐friendliness. 1 Second, minority and more disadvantaged patients were under‐represented. It is plausible that a higher proportion of those patients were working during the day and were not able to answer the call. Oversampling of minoritized and disadvantaged patients should be attempted in the future. Third, the sample size was relatively small, and because HPI is derived from census tract data and multiple patients can live in the same tract, variations in HPI may be more limited than variations in smaller geographic units, such as census block group or block. Therefore, it may be reasonable to reduce the confidence interval to 90% from the conventional 95% confidence interval. Furthermore, some patients did not have address data and HPI. They could possibly be patients without stable housing or were unhoused and should be prioritized for more follow‐up. Fourth, persons with dementia were excluded as participation in a telephone interview could be difficult for those patients. Last, this project was conducted in one academic health system (among over 2700 health systems) with age‐friendly recognitions in a large urban area during an unprecedented time – COVID lockdown. Results may not be generalizable.
5. CONCLUSION
Significant disparities existed across gender, race, and community resource groups in their needs for food, medication, and mental health services among vulnerable older adults during the COVID‐19 pandemic. This outreach effort enabled us to identify those with such unmet needs and refer them for additional services that could address select components of the 4Ms.
The findings of the study suggest that future efforts of AFHS and health systems, in general, should consider HPI in addition to race and ethnicity. Such targeted approaches can play a role in changing the course of a troubling historical trajectory of inequity. Resource allocation frameworks that increase the chances of older adults from more disadvantaged communities—and particularly those of color—to be provided services to address their unmet needs for all components of the 4Ms can help to reduce inequity and can be one way of mitigating the consequences of past, and in many ways, still ongoing wrongs. 14
FUNDING INFORMATION
University of California San Diego Health System Population Health Services.
ACKNOWLEDGMENT
We are grateful to the nursing students who conducted the outreach phone calls.
Tai‐Seale M, Cheung MW, Kwak J, et al. Unmet needs for food, medicine, and mental health services among vulnerable older adults during the COVID‐19 pandemic. Health Serv Res. 2023;58(Suppl. 1):69‐77. doi: 10.1111/1475-6773.14084
Funding information University of California San Diego Health
REFERENCES
- 1. Fulmer T, Mate KS, Berman A. The age‐friendly health system imperative. J Am Geriatr Soc. 2018;66(1):22‐24. doi: 10.1111/jgs.15076 [DOI] [PubMed] [Google Scholar]
- 2. Splinter MJ, Velek P, Ikram MK, et al. Prevalence and determinants of healthcare avoidance during the COVID‐19 pandemic: a population‐based cross‐sectional study. PLoS Med. 2021;18(11):e1003854. doi: 10.1371/journal.pmed.1003854 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Smolic S, Cipin I, Medimurec P. Access to healthcare for people aged 50+ in Europe during the COVID‐19 outbreak. Eur J Ageing. 2021;11:1‐17. doi: 10.1007/s10433-021-00631-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Lesser S, Zakharkin S, Louie C, Escobedo MR, Whyte J, Fulmer T. Clinician knowledge and behaviors related to the 4Ms framework of age‐friendly health systems. J Am Geriatr Soc. 2022;70(3):789‐800. doi: 10.1111/jgs.17571 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. De Biasi A, Wolfe M, Carmody J, Fulmer T, Auerbach J. Creating an Age‐Friendly Public Health System. Innov Aging. 2020;4(1):igz044. doi: 10.1093/geroni/igz044 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Naeem J, Salazar‐Contreras E, Sundaram V, Wainwright L, Kosel K, Miff S. The Dallas accountable health community: its impact on health‐related social needs, care, and costs. NEJM Catalyst. 2022;3(9):CAT.22.0149. doi: 10.1056/CAT.22.0149 [DOI] [Google Scholar]
- 7. Lee JS, Frongillo EA Jr. Nutritional and health consequences are associated with food insecurity among U.S. elderly persons. J Nutr. 2001;131(5):1503‐1509. doi: 10.1093/jn/131.5.1503 [DOI] [PubMed] [Google Scholar]
- 8. Gundersen C, Ziliak JP. Food insecurity and health outcomes. Health Aff (Millwood). 2015;34(11):1830‐1839. doi: 10.1377/hlthaff.2015.0645 [DOI] [PubMed] [Google Scholar]
- 9. Haley E, Harris V, Agnihotri P, Kwak J, Tai‐Seale M. COVID‐19: a unique opportunity for population health to align with nursing schools to help vulnerable populations at risk of adverse outcomes. Clin J Nursing Care Pract. 2021;5:1‐3. doi: 10.29328/journal.cjncp.1001029 [DOI] [Google Scholar]
- 10. Galea S, Merchant RM, Lurie N. The mental health consequences of COVID‐19 and physical distancing: the need for prevention and early intervention. JAMA Intern Med. 2020;180(6):817‐818. doi: 10.1001/jamainternmed.2020.1562 [DOI] [PubMed] [Google Scholar]
- 11. Brooks SK, Webster RK, Smith LE, et al. The psychological impact of quarantine and how to reduce it: rapid review of the evidence. Lancet. 2020;395(10227):912‐920. doi: 10.1016/S0140-6736(20)30460-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Rajkumar RP. COVID‐19 and mental health: a review of the existing literature. Asian J Psychiatr. 2020;52:102066. doi: 10.1016/j.ajp.2020.102066 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Maizlish N, Delaney T, Dowling H, et al. California healthy places index: frames matter. Public Health Rep. 2019;134(4):354‐362. doi: 10.1177/0033354919849882 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Schmidt H, Weintraub R, Williams MA, et al. Equitable allocation of COVID‐19 vaccines in the United States. Nat Med. 2021;27(7):1298‐1307. doi: 10.1038/s41591-021-01379-6 [DOI] [PubMed] [Google Scholar]
- 15. Kroenke K, Spitzer RL, Williams JBW, Lowe B. An ultra‐brief screening scale for anxiety and depression: the PHQ‐4. Psychosomatics. 2009;50(6):613‐621. doi: 10.1176/appi.psy.50.6.613 [DOI] [PubMed] [Google Scholar]