Reducing Caregiver Hunger During Pediatric Hospitalization

Cincinnati Children’s Hospital Medical Center (CCHMC) is a large, urban, academic, free-standing pediatric hospital located in metropolitan Cincinnati. Rates of food insecurity range from 19% to 31% across regional counties.⁶  Our improvement efforts took place on a 41-bed general medical unit that serves as the primary unit for patients admitted to CCHMC’s Hospital Medicine (HM) service. Annually, ∼6500 patients are admitted to the HM service, with an average length of stay of 1.6 days. Approximately 60% of the patients admitted to HM are insured by Medicaid; <2% of patients are uninsured. The remainder of the patients have private insurance.

We focused on caregivers of children with Medicaid or no insurance because the difference in the rates of hunger between this group of caregivers was striking (86% of Medicaid/uninsured reported hunger from 96 surveys and 24% of private-pay caregivers reported hunger from 50 surveys; data were collected from August 22, 2019, to October 7, 2019). Even though nearly 1 in 4 caregivers of children with private insurance reported hunger, potential interventions to address hunger in this population would have required developing a screening system to predict which families may need food support during hospitalization. We elected to focus on caregivers of children with Medicaid or no insurance because of the near-universal need for help in this population as well as the disparity. Therefore, 24% was set as the goal for caregivers of children with Medicaid/uninsured in an attempt to eliminate this disparity.

During hospitalizations, caregivers can purchase full meals (main course, 2 sides, and drink) delivered to their room for $6. Only caregivers with a credit card could order meals over the phone in the patient’s room. Alternatively, caregivers could purchase meals with cash or credit card in the cafeteria, which is open 24 hours a day. It typically costs $8 to $10 in the cafeteria for the same amount of food that can be ordered through room delivery. Before initiating this improvement project, caregivers were eligible for meal assistance through a social work referral by physicians or nurses. With social work approval, a caregiver could pick up 2 meal cards per day, each worth $6, from the family resource center (located on the main floor of the hospital some distance from the unit and open only during the day). The caregiver had to go to the cafeteria to use the meal cards; a caregiver could not use the meal cards to purchase food through the room delivery option.

We assembled a multidisciplinary team to assess the current state of caregiver hunger on the unit and to understand and improve processes to reduce hunger in this at-risk population. Our team consisted of nursing and physician unit leaders and representation from bedside nursing, social work, HM attending and resident physicians, research assistants, and food services leadership.

We created a process map of the existing supports for caregivers who reported hunger (Supplemental Fig 5). We also created a key driver diagram (Fig 1) to highlight: (1) what needed to be in place to address caregiver hunger and (2) potential interventions to reduce hunger among caregivers. In addition, we created a Pareto chart to categorize and prioritize reasons for caregiver hunger during hospitalization (Fig 2).

Our key drivers and associated interventions follow.

Our first test of change consisted of education provided to the unit’s bedside nurses. We pursued this test to raise awareness of hunger as an issue for hospitalized families. Education was provided via presentations at staff meetings and through e-mail communication containing instructions on how to connect families with social work if they reported hunger. We also placed signs in patient rooms, instructing caregivers to ask a nurse for help with obtaining food (Supplemental Fig 6, Table 1).

We carried out multiple tests of change designed to enhance caregiver access to food (Table 1). Some interventions were designed to be temporary changes because of a lack of funding for sustainability. These temporary interventions allowed us to evaluate what support might decrease hunger if funds were to become available long term. Intervention 2 was a 1-week test in which caregivers were offered a single meal voucher per day. Meal vouchers allowed families to receive a main course, 2 sides, and drink delivered to the bedside. Intervention 3 was a change to keep nonperishable food items on the unit for newly admitted patients and caregivers who arrived on the unit late at night. Intervention 4 was a temporary test of change to determine if hunger could be addressed through meal vouchers. Intervention 4 ran for 2 weeks on half of the inpatient unit and provided up to 2 caregivers per patient with 2 meal vouchers per day. We list the half of the unit that received the cards as “intervention” and the half of the unit that did not as “control.” Intervention 5 was an adaptation of intervention 4 using meal cards. In contrast to meal vouchers, meal cards are redeemable in the cafeteria for cash-equivalent value. Intervention 5 provided 2 caregivers per patient with 2 meal cards per day (each valued at $7.50). Intervention 6 was a change in the distribution method for these meal cards (from bedside nurse to charge nurse).

We presented information to hospital leadership on the scope of the problem, our successes, and our challenges. Specifically, we presented our control charts updated through intervention 4, as well as the estimated costs of the program (detailed in the Results section). By providing evidence of the effectiveness of the proposed solution to hospital leadership, we obtained sustainable financial support from hospital administration, making it a permanent change on the unit (allowing us to implement intervention 5 described previously and in Table 1).

We began collecting baseline data in August 2019. Each weekday, a research assistant reviewed the list of hospitalized patients on the intervention unit to determine insurance status for eligibility. Nurses routinely predict discharge timing for each patient.⁷  If the patient was predicted to be discharged that day, the research assistant asked 1 caregiver per patient questions about experiencing hunger during their child’s hospitalization. We chose to survey caregivers near to time of discharge to ensure caregivers could reflect on their hunger throughout the entire hospitalization. Questionnaires were available in English and Spanish and were administered on paper before the COVID-19 pandemic.

At the onset of the COVID-19 pandemic (March 2020), all research and quality improvement initiatives were paused. Our quality improvement project restarted in November 2020 with several system and logistical changes. Caregivers of patients who tested positive for COVID-19 or were awaiting COVID-19 test results were given free meals in their rooms. We included these caregivers in our data collection because we hypothesized that they may still experience hunger if they received meals for only part of their hospital stay. Data collection changed from in person to by phone to keep research assistants off the inpatient units.

Our primary outcome measure was the percentage of caregivers with self-reported hunger during their child’s hospitalization. We adapted questions from the US Department of Agriculture’s Food Security Survey Module.⁸  Specifically, we defined hunger as a caregiver responding in the affirmative to any of the following questions:

• During this hospitalization, did you or other adults in your household ever cut the size of your meals or skip meals because there wasn’t enough money for food?

• During this hospitalization, did you ever eat less than you felt you should because there wasn’t enough money for food?

• During this hospitalization, were you ever hungry but didn’t eat because there wasn’t enough money for food?

• During this hospitalization, did you ever eat less than you felt you should for some other reason other than money?

These questions were asked on the day of discharge to assess hunger that may have occurred at any time during hospitalization. Additionally, the day-of-discharge questionnaire included questions on contributors to hunger (if present) (eg, food was too expensive) and whether the family received meal assistance during hospitalization (ie, meal vouchers, meal cards) (process measure).

We also estimated programmatic costs. For vouchers distributed by a research assistant, we tracked the number distributed to capture cost ($6 per voucher). For meal cards, we tracked the amount of money redeemed in the cafeteria. For example, if the card had a $7.50 value on it, but the caregiver only purchased $7 worth of food, the cost captured was only $7.

We analyzed our outcome (the percentage of caregivers who reported hunger during their child’s hospitalization) and process measure (the percentage of caregivers who received meal assistance) using a statistical process control charts (P-charts). We tracked initial data with subgrouping of 8 caregivers to collect baseline data quickly (12 baseline points) before testing interventions. After the baseline data collection, we switched to tracking data weekly. Given that we changed our data collection procedures (paper to phone) and with the changes in food access for caregivers of children with COVID-19, we recalculated baseline hunger rates after the data collection restart in November 2020, before initiating further tests of change.

Centerline shifts were based on established rules for special cause variation; specifically, either points outside control limits or ≥8 points above or below the mean line.⁹  Short interventions that ran for 1 to 2 weeks (interventions 2 and 4) were not included in calculation of means or toward or against midline shifts (ie, we skipped these weeks when counting for runs of 8 points) because these interventions were not intended to result in sustained system change.

Finally, we describe the financial cost of the interventions, presenting the average number of cafeteria cards redeemed per month and the average monthly cost.

Our study was deemed exempt by the CCHMC institutional review board.

Initially, a mean of 86% of all caregivers of Medicaid-insured/uninsured hospitalized children reported that they were hungry during hospitalization (Fig 3). All the caregivers who indicated being hungry during the baseline data collection indicated that they were hungry because of financial reasons. Although many caregivers identified more than 1 factor contributing to their hunger during hospitalization, concern about costs of food was the most common factor identified (Fig 2). The mean line shifted to 94% of surveyed caregivers reporting hunger at the time of this change when signs to raise awareness about hunger were placed in patient rooms (intervention 1). Our tests of providing 1 meal voucher per caregiver per day (intervention 2) and the opening of a food closet with nonperishable food on the unit (intervention 3) were not associated with changes in our primary outcome.

Providing 2 meal vouchers per caregiver per day on half of the inpatient unit (intervention 4) resulted in special cause variation (2 points outside the statistical process control limits), reflecting a significant decrease in the number of caregivers reporting hunger. Of note, the other half of the unit in which meal cards were not provided continued to have high rates of reported hunger without evidence of a change. After intervention 4 concluded, the improvement project was paused because of the COVID-19 pandemic.

We recalculated baseline data after the 8-month pause. This recalculation revealed a new baseline of 81% of caregivers reporting hunger during hospitalization. After obtaining funding to provide 2 meal cards per caregiver per day, each worth $7.50, redeemable in the cafeteria (intervention 5), the rate of hunger fell to 39%. After changing the distribution to charge nurses (intervention 6), the unit’s mean rate of caregiver hunger fell to 15.5%, a rate sustained for 7 months.

In regard to our process measure, the percentage of caregivers reporting meal assistance, the baseline rate of 48% remained stable before pausing because of COVID, with the exception of the period corresponding to tests 2 and 4. During this time, significantly more caregivers reported receiving meal assistance during the distribution of meal vouchers (Fig 4). After restarting the data collection, the baseline rate of assistance was 28%. When the team began providing 2 meal cards per caregiver per day (interventions 5/6), the rate of caregivers reported receiving meal assistance rose to 99% and was sustained.

During intervention 4, the research assistant distributed 234 meal cards to 51 families over a 2-week period, costing $1404. During interventions 5 and 6, an average of 685 cards per month (standard deviation, 107) were redeemed in the cafeteria. The average cost was $4939 per month (standard deviation, $780).

Our multidisciplinary improvement team reduced reported hunger among caregivers of hospitalized children insured by Medicaid or uninsured from a baseline of 86% to 15.5%. The most impactful intervention was providing caregivers with vouchers or cash equivalent cards for 2 meals per day. This study underscores the importance of assessing and addressing caregiver hunger during their child’s hospitalization. Ensuring families have enough food to eat while their child is hospitalized may support caregivers’ capacity to be at bedside, engage in and comprehend education, and, importantly, uphold basic human dignity.

Hospitalizations are stressful and can lead many families to incur significant financial costs.^10,11  The resulting financial strain almost certainly worsens food insecurity within the hospital and in the days following discharge. These costs disproportionately burden those with limited financial means and make transitions from hospital to home even more difficult. Screening for food insecurity during health care encounters provides the opportunity to connect families to community resources.^12,13  Providing direct food support in the hospital may bolster the relationship with families and improve outcomes. For these reasons, the American Academy of Pediatrics recommends that pediatricians screen and identify children at risk for food insecurity and connect these families to resources.¹⁴  The high rates of caregiver hunger highlight the importance of recognizing hospitalization as an important opportunity to discuss hunger and food insecurity. Our rates of caregiver hunger are higher than some other studies^4,15 ; however, we focused on families with Medicaid/uninsured who are more likely to be lower income and, therefore, experience hunger.

Our first intervention relied predominantly on education through staff communication via e-mail and staff meetings. We also placed signs in all patient rooms on the unit with instructions on how to obtain food and encouraged families to ask for help if they were experiencing hunger. Although education is generally a low reliability intervention,^16,17  we noted an increase in the number of caregivers reporting hunger. We hypothesize that the signs may have created an explicit invitation to caregivers to discuss hunger without shame and subsequently empowered families to report their hunger when surveyed by our research assistant.

The high frequency with which caregivers reported hunger before the initiation of our initiative likely reflected inadequacies of the meal support structure for caregivers. Before our work, hungry caregivers could receive a meal card worth $6 in the cafeteria by meeting with a social worker on the unit. The $6 card had to be picked up by the family each day in the family resource center (which is not located near the unit) and was redeemable only by going down to the cafeteria. The amount provided was often of insufficient value to purchase a full meal in the cafeteria. In contrast, families with access to credit cards could order a full meal (ie, main course, 2 sides, and drink) for $6 through the same room service that prepares patient meals. Conceptually, the most intuitive solution would be to provide these same in-room meals to all families for the same price; this approach would likely be preferred at institutions with the capability and capacity to provide bedside meals. However, food service leadership noted at the beginning of our project that we could not increase the number of caregivers receiving in-room meals without risking delays to patient meals. So, the large number of caregivers who required meal assistance could not be fed through the same room-service process. The main cafeteria, however, had adequate capacity to provide meals to caregivers with limited financial means. Without collaboration with food services leaders, we may not have been able to successfully identify ways to adapt interventions that were successful in our system; in this case, providing meal cards with enough money to purchase full meals.

We also benefited from the expertise of the bedside nurses on the study team. During intervention 5 (in which caregivers received 2 meal cards per day), the bedside nurses were excited to distribute cards and did so with excellent fidelity (nearly 100% of caregivers reported receiving meal assistance). However, there was ongoing confusion regarding eligibility and how to use the meal cards once they were received. Ultimately, changing the distribution to a single person per shift (charge nurse) improved efficiencies as well as consistency for families, resulting in a further decline in the number of caregivers reporting hunger.

Importantly, this work took place during the COVID-19 pandemic. Early data indicate that the pandemic has likely worsened food insecurity.¹⁸  Thus, supporting food access to caregivers became even more important during the pandemic. That said, on our restart, the baseline rate of hunger was similar after we restarted data collection (86% vs. 81%). Despite our hospital providing meals to caregivers of children in COVID-19 isolation, fewer caregivers reported receiving meal assistance when we restarted data collection (48% vs 28%); this finding may reflect the impact of altered care processes and increased isolation due to COVID-19 precautions.

Finally, the ability to make this project sustainable required evidence and cost estimates for discussions with hospital leadership. Hospital leaders were motivated to support families and decrease disparities in experiences between our Medicaid and private-pay families. Presenting evidence on the prevalence of the problem, as well as the effectiveness of the intervention, and detailing costs of implementation, allowed us to secure funding to implement a permanent change.

Our work is not without limitations. First, this project took place on 1 unit at a single children’s hospital and may not be generalizable to other units, hospitals, or regions. However, key drivers and highlighted interventions would likely be relevant to other settings. Second, although the percentage of caregivers who reported hunger was extremely high, we may have missed some who did not feel comfortable reporting their experience with hunger during their child’s hospitalization. Furthermore, we did not have a caregiver partner involved in the project, so we may have missed potential ways to mitigate this weakness. Third, we only surveyed caregivers and provided meal assistance to caregivers of children insured by Medicaid or who were uninsured. Although our interventions were successful in eliminating the Medicaid/uninsured to private disparity, the rate of baseline hunger in the group of private-pay caregivers was still close to 1 in 4. Fourth, although it is a strength that we administered surveys in both English and Spanish, several interventions (eg, signs in rooms) were in English. Future work should ensure information on assistance is equitably available regardless of language. Last, we did not examine if in-hospital food support meaningfully changes patient outcomes or parent experiences. Does having in-hospital food allow more caregivers to remain at the bedside? Is the “fog” that families report during hospitalization partially relieved by alleviating hunger? Answers to these questions, which we will consider as we spread our intervention, would add to an already strong case for widespread adoption, providing evidence that stretches beyond the humanitarian appeal of alleviating hunger.

Nearly all caregivers of children hospitalized with Medicaid insurance reported experiencing hunger during hospitalization. Providing support for 2 meals per caregiver per day drastically reduces caregiver hunger. Providing evidence of the prevalence of the problem, as well as the effectiveness of the solution, allowed us to implement sustainable change at our institution.

CCHMC

Cincinnati Children’s Hospital Medical Center

HM

Hospital Medicine