Katz 2001.

Study characteristics
Methods	Study design: Prospective controlled study How were missing data handled? women who were LTFU and did not complete the follow‐up questionnaires were excluded from the analysis. Randomisation ratio: N/A Recruitment method: not described but women applied for employment; therefore, we assumed that job adverts were circulated. Sample size justification and outcome used: NR Sampling method: purposive sampling. Women enrolled had applied for part‐time employment in their own or neighbouring communities. Selection was based on results of a reading and writing test, relevant work experience and an interview. Study aim or objective: to evaluate the impact of providing a small income on the HH food expenditures and nutritional status (MUAC) of women employed part‐time in a health project compared to women not employed. Study period: 2 years: February 1993 to January 1995. Unit of allocation or exposure: individuals (women)
Participants	Baseline characteristics Intervention or exposure group: Age: mean, years: 25.2 (SD 6.2) Place of residence: rural area of the Sarlahi District, Nepal Sex: female, %: 100 Ethnicity and language: NR Occupation: NR Education: literacy, n (%): 334 (98.2); ≥ 10 years of schooling, %: 23 SES: likelihood of having HH servants, %: 35.4; likelihood of spending > 4 hours per week fetching firewood, %: 14.9 Social capital: NR Nutritional status: MUAC, mean: 22.8 (SD 2.0) Morbidities: NR Concomitant or previous care: none Control Age: mean, years: 28.9 (SD 7.7) Place of residence: rural area of the Sarlahi District, Nepal Sex: female, %: 100 Ethnicity and language: NR Occupation: NR Education: literacy, n (%): 322 (81.7); ≥ 10 years of schooling, %: 13.2 SES: likelihood of having HH servants, %: 21.1; likelihood of spending > 4 hours per week fetching firewood, %: 24.8 Social capital: NR Nutritional status: MUAC, mean: 23.0 (SD 2.2). Morbidities: NR Concomitant or previous care: none Overall Age: NR Place of residence: rural area Sex: female, %: 100 Ethnicity and language: NR Occupation: NR Education: NR SES: likelihood of having HH servants: OR 2.0 (95% CI 1.5 to 2.9); likelihood of spending > 4 hours per week fetching firewood: OR 0.17 (95% CI 0.07 to 0.38) Social capital: NR Nutritional status: NR Morbidities: NR Concomitant or previous care: none Inclusion criteria: employees based on the results of a reading and writing test, relevant work experience and an interview. Exclusion criteria: NR Pretreatment: women who were hired were significantly younger than those who were not (25.2 years vs 28.9 years) (table 1). They were more likely to be literate (98.2% vs 81.7%; OR 10.8, 95% CI 4.9 to 28.2), to have ≥ 10 years of formal schooling (23.2% vs 13.2%; OR 2.0, 95% CI 1.3 to 3.0), and to have HH servants (35.4% vs 21.1%; OR 2.0, 95% CI 1.5 to 2.9). They were less likely to smoke (2.4% vs 12.4%; OR 0.54, 95% CI 0.36 to 0.80) and to spend > 4 hours per week fetching firewood (14.9% vs 24.8%; OR 0.17, 95% CI 0.07 to 0.38). Those who were hired and those who were not hired were comparable with respect to caste, HH size, and ownership of animals and other HH goods such as radios, watches, bicycles and furniture. Group differences assessed using t‐test for continuous data and Chi2 test for categorical data. Attrition per relevant group: intervention (employed): 9/350 (2.6%) (7 no longer employed, 2 on leave of absence); control (not employed): 125/520 (24%) (2 dead, 2 moved to hired group, 85 no longer in area, 36 were not at initial addresses) Description of subgroups measured and reported: no subgroups reported Total number completed and analysed per relevant group: 341 employed and 395 not employed used for all baseline and follow‐up outcomes, except changes in MUAC (data for 335 employed and 383 not employed). Total number enrolled per relevant group: intervention: 350; control: 520 Total number randomised per relevant group: N/A
Interventions	Intervention characteristics Intervention or exposure: short‐term part‐time employment for women Food access intervention category: increase buying power Intervention type: income generation Description: women were followed over time to assess the impact of employment on changes in HH food expenditure and MUAC. The women had applied for part‐time employment distributing weekly supplements to married women of childbearing age in their own or neighbouring communities. The job involved weekly visits to the homes of about 100 women to provide supplements, note the occurrences of menses in the previous week, record pregnancy status (not pregnant, pregnant, miscarriage in the previous week, stillbirth in the previous week or live birth in the previous week), and record the receipt of supplements. The women received about NPR 900 (USD 15) per month for an estimated 5 hours of work per week. Duration of intervention period: 2 years: 1993–1995 Frequency: monthly income Number of study contacts: 2: baseline (December 1992 to January 1993), and follow‐up after 2 years Providers: research project staff – Sarlahi Study Group Delivery: how salary was given to women was NR Co‐interventions: among the 341 women who had been employed by the project, 106 (31.1%) reported additional cash employment (the project employment was part‐time). Amounts not known. Unclear whether these women were also receiving nutritional supplements as part of the RCT they were working for. Resource requirements: NR Economic indicators: NR Control group: no intervention Co‐interventions: 36/395 (9.1%) women who had not been employed by the nutrition project reported that they had been employed in jobs for which they were paid some cash. Amounts not known. Unclear whether these women were also receiving nutritional supplements as part of the RCT that was ongoing.
Outcomes	Proportion of HH expenditure on food: weekly food expenditure (NR), food expenditure for different food groups (NR) Anthropometry: MUAC
Identification	Sponsorship source: co‐operative agreement No. DAN 0045‐A‐5094 between the office of Nutrition, US Agency for International Development (USAID), the Center for Human Nutrition (CHN), and the Dana Center for Preventive Ophthalmology (DCPO) at Johns Hopkins University. Country: Nepal Setting: rural area of the Sarlahi District Author's name: Joanne Katz Email: NR Declarations of interest: NR Study or programme name and acronym: N/A Type of record: journal article
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (Selection bias)	High risk	CBA; therefore, no randomisation was done.
Allocation concealment (Selection bias)	High risk	Selection of study participants based on them getting employed. They had to do a reading and writing test, demonstrate relevant work experience and they were interviewed.
Baseline characteristics similar (Selection bias)	Low risk	Quote: "The women who were hired were younger and better educated than those who were not hired, but in other respects the two groups of women were similar. After adjustment for these baseline differences, the change in MUAC was not significantly different between the two groups of women."
Baseline outcome measurements similar (Selection bias)	Low risk	Quote: "… after adjustment for baseline differences between the two groups of women, the difference between the two groups was not significant. Among those households buying specific foods, the expenditure on each item was comparable for households of women who were hired and households of women who were not hired (table 3)."
Blinding of participants and personnel (Performance bias)	Low risk	No blinding carried out but it was unlikely that lack of blinding had an effect on the participant's behaviour.
Blinding of outcome assessment (Detection bias)	High risk	No blinding carried out. It is likely that self‐reports of food purchases and expenditures was influenced by knowledge of allocation. MUAC was unlikely to have been influenced by lack of blinding.
Protection against contamination (Performance bias)	High risk	Quote: "At follow‐up, 36 of the 395 women who had not been employed by the nutrition project (9.1%) reported that they had been employed in jobs for which they were paid some cash. Among the 341 women who had been employed by the project, 106 (31.1%) reported additional cash employment (the project employment was part‐time). However, the amount of cash payments associated with these additional activities was not determined."
Incomplete outcome data (Attrition bias)	High risk	Very different proportion of attrition between the groups: 2.6% for women who were hired compared to 24% among women who were not hired. Missing data were excluded from the analysis and information from those in the control group could have an effect on the outcomes.
Selective outcome reporting (Reporting bias)	Unclear risk	No protocol available. Authors stated in the methods that they would conduct baseline comparisons and conduct an assessment of the impact of employment (changes in expenditure and in MUAC) by fitting a linear regression model that adjusts for baseline differences. There was evidence that authors did these analyses.
Other bias	Unclear risk	Misclassification bias: unlikely. Measurement bias: unclear. Measurement of MUAC or food expenditure is not very well described.