Impact of HIV/AIDS on household food security and quality of life in Malawi

Summary of evaluation¹

By Shannon Senefeld, Catholic Relief Services

Shannon Senefeld is the Regional Technical Advisor, HIV/AIDS with Catholic Relief Services for the Southern Africa Region

The author would like to acknowledge the contributions of Catholic Relief Services-Malawi and CADECOM Dedza-Malawi to the work reflected in this article.

Double-cropping maize and beans as part of Dezda agri programme

Recent research has suggested that HIV/AIDS in southern Africa is directly linked to increased household food insecurity. The research presented here is derived from an evaluation study of Catholic Relief Services' (CRS) home-based care programme in one rural area of Central Malawi. The project commissioned an evaluation of interventions, including food security and outreach to households affected by HIV/AIDS, ending in September 2004.

Methods and design

A total of 326 households participated in the study. The households were selected using area probability sampling. Twenty villages were selected in total. Of the households that participated, 25 households were eliminated from the final analyses due to incomplete responses or invalidated surveys.

Each head of household was administered a survey in the local language. The survey was comprised of three sections. The first was adapted from the Community Household Survey (CHS), which has been used widely in southern Africa to examine household vulnerability. The CHS is a self-reporting measure that asks respondents to identify their household assets and food security levels. The core components of the CHS remained intact for this evaluation, and included questions on existing food supply and expected food security in times of drought. An additional section was added that asked families knowledge-specific questions related to HIV/AIDS, in order to evaluate the effectiveness of the HIV/AIDS awareness campaigns within the areas.

The second section asked for demographic information, including the age of the head of household and the number of children and orphans within each household. This section specifically requested information on HIV/AIDS within the surveyed households. The final section was a Quality of Life Index (QLI), which was selected due to its comprehensive psychosocial nature. The QLI is a standardised measure, developed by Ferrans and Powers². While this measure has been standardised among various populations around the world, it has not been used before within Malawi. However, psychometrics revealed that among the sample population, the QLI performed similarly to past performance with other populations.

Results

A total of 301 household surveys were analysed in the results. The sample represents 20 villages randomly selected from three deaneries. The average age of the head of household was 45.7 years. Female head of households constituted 40.1% of the households. Approximately 3.4 children were reported, on average, for each household, while an average 5.6 people were reported as residing in the households. Just under half (49%) of households reported that an orphan resided in their homes, but only 23.2% of households reported the presence of AIDS orphans in their households.

Of the sample, only 12.9% of the households reported having someone who was living with HIV/AIDS within their households, but 45.5% reported benefiting from a home-based care programme. While home-based care programmes also cover chronic illnesses other than HIV/AIDS, the primary service delivery is around HIV/AIDS related illnesses. This suggests that the sample either did not know their family members may have been infected, or chose not to reveal their family member's status to interviewers. However, more than 65% of the sample indicated that they knew where they could access voluntary counselling and testing services, and more than 90% indicated that they would welcome and care for a family member who was infected. Further data analyses revealed a negative relationship between households who reported the presence of someone infected with HIV/AIDS and the self-reported willingness to disclose status variable (p<0.05).

More than one-third (37.7%) of households reported selling assets in the last three months. Of the 107 households that reported selling assets, the primary reason for doing so was to meet household food needs (40.3%), followed by the need to meet daily household expenses (29%), and the need to cover hospital and doctor bills (11.2%). On average, households reported that in a non-drought year, they would only be able to meet their household food needs for 4.5 months with their current harvest (SD=2.7).

There was a significant difference (p<0.05) in reported food security between households that reported having a household member living with HIV/AIDS and non-affected households. Arelationship (p<0.01) also emerged between the presence of AIDS orphans within the household and a reduced number of months per year the household could meet its food needs. In addition, the number of months that households could meet their food supply with their current harvest was correlated significantly with the participation of the households in the HBC project (p<0.05).

Analyses demonstrated quality of life was predicted by whether or not the household had sold assets in the previous three months (p<0.001) and whether their current harvest food supply was above or below average supplies (p<0.017). Decreased quality of life scores were also significantly associated with the presence of persons living with HIV/AIDS (PLWHA) (p<0.05) and the presence of orphans (p<0.05) in the household.

Computing a two-way analysis of variance for main effects of asset sales and the presence of an HIV household member on quality of life, demonstrated that there was no interaction effect between asset sales and PLWHA presence on quality of life of the head of household. However, a significant main effect on quality of life emerged within the variable of asset sales.

Limitations

One of the main limitations with most of the research focusing on HIV/AIDS is the inability to appropriately and correctly identify people living with HIV or AIDS. Most research now uses 'chronically ill' as a proxy indicator for HIV infection. Initial analyses from some researchers have indicated that this proxy indicator over-exaggerates the cases of HIV infection. Barrère (2005)³ examined the use of this proxy indicator in a survey in Malawi and found that only 54% of a chronically ill sample was likely to have HIV or AIDS based on national infection rates.

As such, the survey used in this study specifically asked if anyone in the household was infected with HIV or AIDS. A total of 12.9% of the households replied affirmatively. According to the National AIDS Commission in Malawi (2003)⁴, the infection rate among rural households is estimated to be 10-15%. Thus, the 12.9% of affirmative responses would fall as projected within the infection rate estimates. However, it is possible that some households answered negatively to this question due to stigma concerns or lack of knowledge of family members' status, which could result in a higher HIV positive population than the one presented here.

Finally, this is a relatively small sample and cannot be applied broadly to the larger rural population in Malawi in other districts. These data were obtained from a specific geographic location for the purposes of an intervention evaluation, thus the results are not necessarily applicable to a broader area.

Discussion

Sewing classes as part of orphan skills training in CRS programme

The results demonstrate that the presence of PLWHA and OVC (orphans and vulnerable children) at the household level also impact the level of food security, which is highlighted by whether the households have access to sufficient food to meet their needs. This relationship is further compounded when households, lacking this supply, cope with their situation by selling their assets. These variables, in turn, affect the overall quality of life of the heads of households.

The data from this evaluation highlight the need to engage in holistic programming. Traditional mitigation responses, such as simple home-based care programmes, cannot respond to the overall stress placed on the household in terms of food insecurity. Psycho-social interventions will also be hampered without addressing the basic food needs, as quality of life will likely decrease alongside assets and food security. In addition, given that these data demonstrate a link between the presence of OVC in the household and the quality of life and food security of the household, programming will need to address these findings in future OVC programming.

Given these results, programming will need to fully address mitigation of HIV/AIDS at the household level. While many theories have abounded regarding the impact of HIV/AIDS on household food security, this evaluation confirmed this link in rural, central Malawi, while uncovering the psychosocial impact of this relationship on the households.

For further information, contact Shannon Senefeld email:ssenefeld@crssaro.org.zw or shannonsenefeld@yahoo.com

¹Dedza End of Project Evaluation, July 2003.

²Ferrans C, & Powers M (1992). Psychometric assessment of the Quality of Life Index. Research in Nursing and Health, 15, 29-38.

³Barrère, B. (2005). Pre-Test of New HIV Indicators. Presentation from ORC Macro, UNICEF, DHS & USAID. Washington, DC. January 6, 2005.

⁴National AIDS Commission (2003). Estimating National HIV Prevalence in Malawi from Sentinel Surveillance Data: Technical Report: October 2003.