How COVID-19 affected food systems, health service delivery and maternal and infant nutrition practices: Implications for moving forward in Kenya

This implementation research study sought to examine the impact of the COVID-19 pandemic on maternal and infant nutrition practices, and related aspects of health and food systems in Nairobi and Uasin Gishu Counties, Kenya. The study triangulated in-depth interviews with 16 pregnant women, 31 lactating women (including COVID-19 positive), 10 facility health workers, 10 community health volunteers, 6 focus group discussions (FGDs) with food vendors, 4 FGDs and 15 stakeholder interviews with government and implementing partners. Trends from Kenyan Health Information System indicators (i.e., exclusive breastfeeding and initiation of breastfeeding, antenatal care) were also examined. During the COVID-19 pandemic, a decline in attendance of antenatal care, and maternity facilities was observed, and corroborated by Kenyan Health Information System data. Lack of clarity among health workers on COVID-19 breastfeeding guidance and fear of COVID-19 infection early in the pandemic were key drivers of early infant formula use, mother–child separation following delivery and delayed initiation of breastfeeding. Most women exclusively breastfed due to Government of Kenya restrictions in movement. Unemployment and job loss was linked to food insecurity and worsened by increased food prices and limited social protection measures. In response, pregnant and lactating women resorted to skipping meals and reducing quantity and variety of foods consumed. Efforts to build forward from COVID-19 in Kenya should include facility and community health education to prevent disruptions in breastfeeding and to support maternal dietary intake, and in the provision of targeted social protection measures alongside other multisectoral interventions (i.e., psychosocial support) for Kenyan pregnant and lactating women. This implementation research study used a mix of qualitative methodologies (i.e., in‐depth interviews [IDIs], focus group discussions, key informant stakeholder interviews) triangulated with quantitative data from the Kenya Health Information System (KHIS). In the process of selection of study sites, first, the research team identified the counties with the highest burden of COVID‐19 cases, with the Ministry of Health Kenya (MoH Kenya, 2021). Nairobi and Uasin Gishu Counties were selected to better understand the variation in the impact of COVID‐19 on health services, food systems and maternal and infant nutrition practices in both rural and urban areas in relation to GoK COVID‐19 restrictions (i.e., lockdowns). Nairobi County is primarily urban (i.e., Nairobi City), with at least one GoK lockdown during the COVID‐19 pandemic. Uasin Gishu County is comprised of rural, agricultural communities who cultivate maize and wheat, beans, Irish potatoes and horticultural crops such as passion fruits, coffee, macadamia nuts and avocadoes (County Government of Uasin Gishu, 2018) and did not experience a COVID‐19 lockdown during the pandemic. In Nairobi County, Embakasi East and Kibra subcounties were identified as study sites, to ensure inclusion of informal settlements and middle‐class settlements, which were hardest hit by the COVID‐19 pandemic. Specifically, Kibra subcounty had the third highest cases of COVID‐19 in the county and houses the largest informal settlement in the country. In Uasin Gishu County, Ainabkoi and Turbo subcounties accounted for almost half of the COVID‐19 infections within the county. Fieldwork occurred between August and September 2021 and comprised a mix of in‐person and virtual data collection. Pregnant women in their second or third trimester, and lactating women 18–49 years of age with at least one child 0–23 months of age, were residents of study communities, and identified with the support of community health volunteers (CHVs) in study sites. COVID‐19‐positive lactating women 18–49 years of age who had at least one child 0–23 months of age were identified through a Nairobi County list of clients with a confirmed polymerase chain reaction COVID‐19‐positive test and were approached via phone for possible study participation (see Table 1). Overview of study sites, participants and methods, by county in Kenya, n = 92 Abbreviations: BF, breastfeeding; CHV, community health volunteer; FGD, focus group discussion; HW, facility health workers; IDI, in‐depth interview; KII, key informant interview. In‐person IDIs were conducted with pregnant and lactating women on topics, such as the impact of COVID‐19 on health‐seeking behaviours, receipt and quality of health services and dietary practices, perceptions, beliefs and related behaviours related to breastfeeding, maternal nutrition and complementary feeding. Food frequency questionnaires (FFQ) were also administered to pregnant women, 15–49 years of age and lactating mother‐child pairs with children 0–23 months of age and lactating COVID‐19‐positive women with children 0–23 months of age. During FFQ, women were probed on all foods consumed in the last 24 h before the interview, as well as foods consumed on a weekly basis. IDIs were carried out with CHVs, who offer maternal, child health and nutrition counselling/promotional information or referral services and facility health workers, providing MCH services. The IDIs explored the perspectives of health workers at facility and community level on the content, type and quality of breastfeeding counselling, extent of use and promotion of breastmilk substitutes, psychosocial support to breastfeeding women, adherence to Ministry of Health (MoH) COVID‐19 breastfeeding guidance. In‐person focus group discussions (FGDs) were also conducted with food vendors who sell cereals, legumes, fruits, vegetables and pulses in open‐air markets examining the effect of COVID‐19 on food systems. Both virtual and in‐person FGDs were conducted with county/subcounty stakeholders on MIYCN, which ranged from 3 to 4 persons. At the national level, virtual key informant interviews (KIIs) were conducted with national level nutrition‐specific stakeholders, nutrition‐sensitive stakeholders at national and county level (n = 8), and county implementing partners. All women, CHVs and facility health providers, and food vendors were selected by purposive sampling in study site communities. Secondary data was extracted from the KHIS MoH database for Nairobi and Uasin Gishu Counties and trends were described for initiation of breastfeeding, EBF indicators, ANC attendance, prepandemic (from March 2019 to February 2020) and post declaration of the COVID‐19 pandemic (March 2020–February 2021). The two‐time frames, that is, prepandemic and pandemic, were selected to allow for comparison in access to services. At the health facility, data are documented daily in the registers of the various service delivery points, that is, ANC, maternity and MCH. This is consolidated monthly and subsequently submitted to either the health facility or subcounty records officer for entry into the KHIS database, which received further checks through monthly and quarterly data review meetings. Initiation of breastfeeding is documented in the maternity register and is calculated as a percentage of infants who are put to the breast within 1 h/total live births in the health facility. EBF is documented in the child welfare clinic (i.e., routine child health services), as a proportion of children 0–5 months who were exclusively breastfed in the last 24 h/children 0–5 months of age who visited the health facility. ANC attendance is documented as a percentage of all pregnant women who visited the health facility as the numerator and/the projected pregnant women as the denominator. Study personnel were trained on research ethics, informed consent, privacy/confidentiality and techniques related to qualitative data collection and all data collection tools were pretested for cultural appropriateness and comprehension. Written informed consent was obtained for participants interviewed in–person, while verbal informed consent was obtained from participants in online interviews and FGDs before audio‐recording interviews and FGDs in Swahili or English. Study procedures were approved in Kenya by the Institutional Ethics Review Committee of Masinde Muliro University of Science and Technology. Subsequently, a research license was granted by National Commission for Science Technology and Innovation per Kenyan guidance for conducting research in‐country. Written approvals to conduct the study were provided by the county governments in Nairobi and Uasin Gishu Counties and subcounty and health facility management teams. Interviews with national and county stakeholders, implementing partners, health workers and COVID‐19‐positive lactating women were conducted and audio recorded in English. Interviews with CHVs, pregnant and lactating women, including food vendors were conducted and audio recorded in Swahili. Subsequently, transcribers conversant with both English and Swahili then transcribed all the interviews verbatim into English. The quality of transcriptions was checked for accuracy and completeness against the audio recording while demographic information and food frequency data were also verified against the data collection forms by B. A., C. G., and J. A. K. The researchers (B. A., J. A. K., L. R., C. G.) conducted a preliminary review of the data by reading a subset of the transcripts to create an initial codebook, which included major themes and subthemes that emerged from the data. The researchers (B. A., J. A. K., L. R., C. G.) then coded a subset of the transcripts and discussed and came to consensus about any discrepancies in coding. Based on this discussion, the codebook was refined and finalized. All transcripts were coded using Dedoose online software. Each subtheme was then summarized, and illustrative quotes were selected (see Table 3). Food frequency data were analysed daily and weekly (<3 times, ≥3 times per week) for pregnant and lactating women, and by study site. Food price data on local foods was compiled by county and provided in Supporting Information File: S1. Quantitative data were downloaded from the KHIS and trends were examined by indicator and study site. Summary of dominant themes, by study participant group, Nairobi (NC) and Uasin Gishu (UGC) Counties ‘Mothers were scared to seek services because they don't know whether they are going to contract that virus at the hospital level or not. So, this affected the services that are being provided. The ANC visits reduced and the mothers come when they are almost due for delivery’. (Stakeholder, national level, NC) ‘When COVID hit last year, we closed our business. So we stayed home. And now you know with business, if you're not working, you are not getting any income. Yeah, so that was a huge challenge, but it did not affect my ability to buy food as my husband would buy it since he did not lose his job’. (COVID‐19+ lactating woman, NC) Interviewee ‘When a mother comes to deliver, if mother has COVID‐19, we will not allow her to be very close or she must mask. She must separate immediately after delivery, so we have to protect mother to mask and do what is needed because the child still needs her’. Interviewer: ‘After how long you will give her the child?’ Interviewee: ‘I don't know’. (Health worker, UGC) ‘Pregnant women consume foods rich in proteins and traditional vegetables because, for example, traditional vegetables helps improve the amount of breastmilk, while ugali also helps. Spinach and cabbage hydrates the body’. (Pregnant woman, NC) ‘During COVID‐19, we eat traditional vegetables mostly and then there is also porridge–a lot of porridge for the child. The other one will be a little bit of protein, not too much. And then there is this lemon mixture that we boil every week and drink. The child drinks it, too, even if we don't know if the doctor accepts that. It's a must for the child to drink, so that at least it makes someone strong, so that when COVID‐19 strikes, at least it won't be easy’. (Lactating woman, NC) ‘Because of financial issues, they [mothers] were eating a poor diet; if its ugali, it would just be ugali and vegetables every day because there was no money to buy good proteins such as beans’. (Health worker, UGC) ‘Even now, it's a miracle if I can eat meat once in a week. Just those foods that have protein, I have reduced’. (Lactating woman, NC)