Evaluation of a Comprehensive School Health Programme in Zambia

The proportion of learners who sought formal care (any government health centre or hospital, or the school health room). Equals one if reported at least once during the health diary observation period.

The investigators will calculate this as the proportion of "serious" illness episodes for which a child seeks care, at the child level. In other words, the number of serious illness episodes with care divided by the total number of serious illness episodes, for all children with at least one serious illness episode.

A serious illness episode is one during which a child's ill-health symptoms are severe enough to require medical care, either immediately or the next day. This classification is based on WHO tools (c-IMCI and IMAI): Care is needed when (i) a child experiences any red-flag symptom (e.g. difficulty breathing, convulsions); (ii) has symptoms suggestive of certain conditions (e.g. measles, malaria); (iii) has experienced mild symptoms for a long time (coughing for 14 days, etc).

During an illness episode, a child will be considered to seek medical care if they visit (i) the health room (treatment arm) or (ii) a government clinic or hospital (any arm).

The program's effect on school attendance is assessed using data collected through repeated, unannounced spot checks at each school. The primary attendance measure will be based on the sample of children enrolled in the study at baseline and confirmed to still be enrolled in their original school at the time of the visit. In this sample, the attendance rate, measured at each attendance spot check, will be the proportion of those children confirmed present in class on the day of the visit.

Additional measure of attendance, focusing on students in grades 1, 3, 5, and 7 at the time of the spot check. This register-based measure is defined as the proportion of children on the register who are present on the day of the spot check and confirmed by school staff as enrolled in the selected class on the visit day.

Testing for the presence of P. falciparum, non-falciparum species, or a mixed infection. The tests are performed using a rapid diagnostic test (RDT) on the sample of children taking part in the health diaries (testing will occur at the end of the period).

Haemoglobin concentration measured using a HemoCue photometer in the sample of children taking part in the health diaries (testing will occur at the end of the period). Anemia severity categories will be based on the WHO classification which depends on the age and sex of the child. Haemoglobin is first adjusted for elevation using the WHO (2024) continuous adjustment, defined in g/L as 0.0056384 times elevation in metres plus 0.0000003 times elevation squared, evaluated at each school's recorded GPS elevation and subtracted from the measured value. Moderate-to-severe anaemia is considered when adjusted haemoglobin is below 11.0 g/dL, a threshold common to all age bands represented in the sample.

Incidence of diarrhoea - carer report of the child experiencing three or more loose or watery stools.

An indicator for incidence of haematuria (blood in urine) reported by the caregiver.

The share of days on which the child was reported to have any symptom, over the eight week symptom diary recording period (0 days if no symptoms recorded).

Proportion of completed health-diary days on which the child reported one or more symptoms. Computed per child as (diary days with ≥1 reported symptom) ÷ (completed diary days). Defined for all children in the diary sub-sample; no conditioning on illness status.

Mean number of consecutive symptom-days per illness episode, among episodes in children who reported illness during the diary window. Estimated from a discrete-time hazard model for symptom resolution: episodes still ongoing at the end of the observation window contribute their observed (censored) days rather than being dropped; episodes already symptomatic on the first diary day are excluded (start date unobservable / left-truncated). Reported as a restricted mean over the eight-week observation window.

As for the all-episodes duration outcome, restricted to episodes classified as serious.

A serious illness episode is one during which a child's ill-health symptoms are severe enough to require medical care, either immediately or the next day. This classification is based on WHO tools (c-IMCI and IMAI): Care is needed when (i) a child experiences any red-flag symptom (e.g. difficulty breathing, convulsions); (ii) has symptoms suggestive of certain conditions (e.g. measles, malaria); (iii) has experienced mild symptoms for a long time (coughing for 14 days, etc).

Mean consecutive symptom-days per serious episode, estimated from the same discrete-time hazard model, with ongoing episodes contributing observed (censored) days and day-1 episodes excluded as left-truncated. Reported as a restricted mean over the eight-week window.

Share of correct responses across three menstrual health knowledge items. The item set is the final items retained after in-country piloting for comprehension and relevance, from a pre-specified pool adapted from validated instruments. Computed per respondent as (correct responses) / (items answered), averaged within the relevant sub-sample: girls grade 5 or above who have heard of menstruation. Higher values indicate greater knowledge.

Share of responses indicating good menstrual management across the practices items. Item set is the final 5 items pilot-retained from a pre-specified pool adapted from the same validated instruments. Computed per respondent as (good-management responses) / (items answered), averaged across girls who have reached menarche in the sub-sample. Higher values indicate better practices.

First principal component of a PCA over the menstrual health stigma items (final 4 items pilot-retained from a pre-specified pool adapted from the same validated instruments; construction rule fixed prior to piloting). Loadings estimated on the pooled sample; scores standardised to the control-group mean and standard deviation. Oriented so that higher values indicate lesser stigma; a beneficial programme effect corresponds to a reduction. Constructed for girls grade 5 or above who have heard of menstruation.

Item-response-theory (IRT) ability score in literacy, estimated over per-item responses from the grade-banded exam blocks (lower-, middle-, and upper-grade versions) placed on a common within-domain scale via shared anchor items. Standardised to the control-group mean and standard deviation. Higher values indicate greater literacy ability.

IRT ability score in numeracy, estimated over per-item responses from the grade-banded exam blocks (lower-, middle-, upper-grade versions) linked to a common within-domain scale via shared anchor items. Standardised to the control-group mean and SD. Higher values indicate greater numeracy ability.

IRT ability score in science and health knowledge, estimated over per-item responses from the grade-banded exam blocks (lower-, middle-, upper-grade versions) linked to a common within-domain scale via shared anchor items. Standardised to the control-group mean and SD. Higher values indicate greater ability.

Composite learning score, defined as the mean of the three control-standardised domain scores (literacy, numeracy, science), each an IRT ability score placed on a common within-domain scale across grade bands via anchor items and standardised to the control-group mean and SD. Domains enter with equal weight. Higher values indicate greater overall learning ability.

Students' working memory and attention is measured using the Digit Span task, a widely used short-term memory and working memory assessment. Enumerators orally present a sequence of digits at a rate of one per second, and students are asked to repeat the digits in the same order (Digit Span Forward) and then in reverse order (Digit Span Backward). The task measures students' auditory attention, concentration, and working memory capacity, which are foundational components of general cognitive functioning and predictive of academic achievement. Scores are recorded as the total number of correct sequences reproduced, following standard administration and scoring procedures.

Proportion of children confirmed present in class on the day of the visit, out of the total number of children enrolled at the school at baseline and still alive. This measure effectively treats children who may have transferred to another study school as 'absent', dropped out, or temporarily relocated.

Proportion of children confirmed present in class on the day of the visit either in their original school or in another school to which they have transferred if that school is one of the study schools, out of the total number of children enrolled at baseline and still alive. Analysis will follow Intention-to-Treat principle, treating participants as assigned to their original treatment arm, even if they moved to a different study arm school.

From the diary's daily records, caregivers flag whether the child attended school each day and records the reason for each absent day. The reasons are: illness (1), was menstruating (2), farming (3), school closure or school event (4), bad weather (5), religious reasons (6), market day (7), public holiday (8), fishing (9), mining (10), working with parents (11), no uniform/shoes/books (12), and other (-97). A day missed is considered caused by an illness or menstruation-related absence when the reason is (1) or (2). The sample is restricted to school days by dropping weekends and public holidays. This measure is the share of the child's school days in the diary window lost to health-related absence: health-related absence days divided by total school days observed.

Menstrual health knowledge index (share of correct responses), constructed identically to the girls' secondary-outcome knowledge - same item set and same (correct responses) / (items answered) construction - on the sample of boys who have heard of menstruation. Higher values indicate greater knowledge. Parallels the girls' secondary outcome to assess whether the programme shifts boys' menstrual knowledge.

Menstrual health stigma index, constructed identically to the girls' secondary-outcome stigma index - first principal component of a PCA over the same stigma items - on the sample of boys who have heard of menstruation. Standardised to the boys' control-group mean and standard deviation; oriented so that higher values indicate lesser stigma, consistent with the girls' index.

Parallels the girls' secondary outcome to assess whether the programme shifts boys' menstrual stigma.

Measured using a short school belonging and climate scale drawn from the Young Lives study, designed to capture children's subjective experiences of school in a simple, age-appropriate way. The tool focuses on core dimensions of belonging-enjoyment, perceived care from teachers and peers, safety, respect, and mutual support-using a consistent three-point response scale ("No, almost never", "Sometimes", "Yes, almost always"). Its brevity makes it suitable for large surveys with primary-school children while still capturing meaningful variation in how supported and included students feel in their school environment. To analyse responses, investigators will construct a simple index using principal component analysis, normalising on the control group for ease of interpretation.

Using the child well-being at school scale developed by the Young Lives study to capture children's emotional experiences while at school in a simple and age-appropriate way. The tool focuses on both positive and negative affective states-see appendix-and uses a consistent three-point response scale ("No, almost never", "Sometimes", "Yes, almost always"). Its simplicity makes it suitable for administration to primary school children. To analyse responses, investigators will construct a simple index using principal component analysis, normalising on the control group for ease of interpretation.

A PCA index z-scored on the control group from the learner survey, combining a conduct checklist (teacher late, absent, or on their phone) with a 5-point Likert engagement battery (teacher cares, makes lessons interesting, offers extra help, motivates), oriented so a higher score is better conduct and engagement

Indicator equal to 1 if child responds yes to any bullying item, from a five item yes/no checklist of acts experienced in the last week attended (mean names, threats, physical harm, exclusion, lies).

Count of acts of bullying, when child responds yes to any bullying item, from a five item yes/no checklist of acts experienced in the last week attended (mean names, threats, physical harm, exclusion, lies).

From a parallel yes/no checklist of acts a teacher used on the child in the past month, spanning verbal (shouting) and escalating physical acts. The primary measure is an indicator for any act and the secondary measures are the count of distinct acts and a physical-punishment indicator restricted to the acts involving physical contact.

Acceptance of corporal punishment, elicited via a teacher discipline vignette (respondents hear a scenario and rate the depicted teacher's reaction). Standardised-mean acceptance-norms index over four items: (i) own acceptability of the reaction; (ii) injunctive social norm - number of five colleagues expected to find it acceptable (0-5); (iii) descriptive social norm - number of five colleagues expected to react likewise (0-5); (iv) own behavioural attitude - self-rated likelihood of acting the same way. Each item is oriented so that higher indicates greater acceptance and standardised to the control-group mean and SD before averaging (standardisation places the differing item scales on a common metric). Higher index values indicate greater acceptance of corporal punishment.

Teacher-reported job satisfaction, summarised across the eight satisfaction items (teacher satisfaction 1-8) as standardised-mean index. Higher values indicate greater satisfaction. Teacher-level.

Emotional-exhaustion subscale of the Maslach Burnout Inventory, the primary burnout indicator, summarised as subscale mean. Higher values indicate greater exhaustion.

Indicator = 1 if the teacher reports intending to leave teaching at the point of the survey. All endline-surveyed teachers. Higher = greater intent to leave.

Indicator = 1 if a teacher randomly sampled at baseline is no longer at their original school at endline (whether they left teaching or moved schools), from tracking of the baseline-sampled cohort rather than the endline survey. Population differs from the survey-based teacher outcomes; measures change over the baseline-to-endline interval. Time Frame: Unit: proportion.

Locus-of-control index over four items:

• "There is little I can do to help a student's learning."
• "Pupils come unprepared from previous grades."
• "Parents do not have the necessary education to help their child be more successful at school."
• "If parents would do more for their children, I could do more." where respondents answer "agree" or "disagree". Summarised as standardised-mean index, oriented so that higher values indicate more locus of control.

Self-reported hours in a typical day spent on health-related programmes in the school, from the teacher time-use module (recall over four mutually exclusive categories: teaching/instructing; planning and grading; health-related programmes; other school responsibilities). Primary time-use quantity; the mechanism-relevant category for a school-health intervention.

Self-reported hours in a typical day spent teaching/instructing, from the teacher time-use module (recall over four mutually exclusive categories: teaching/instructing; planning and grading; health-related programmes; other school responsibilities). Reported alongside health-duty time to capture reallocation away from core instruction.

Prevalence of stunting (chronic undernutrition) among children measured at endline. A child is classified stunted if their height-for-age z-score (HAZ) is more than 2 SD below the reference median, using the WHO Growth Reference for school-aged children and adolescents (de Onis et al. 2007; ages 5-19, sex-specific), with z-scores computed from completed age in months at the date of measurement. HAZ values outside -6 to +6 are treated as biologically implausible and excluded per WHO flagging conventions; the number of excluded observations is reported.

Prevalence of thinness (acute undernutrition; the WHO term for this age group) among children measured at endline. A child is classified thin if their BMI-for-age z-score (BAZ) is more than 2 SD below the reference median, using the same WHO reference. BMI-for-age is used rather than weight-for-age, which is confounded by stunting; BMI-for-age standardises weight against the child's own height. BAZ values outside -5 to +5 are excluded as biologically implausible per WHO flagging; the number excluded is reported.

Using the same haemoglobin measurements collected during the biomarker survey (hbg result, in g/dL), investigators will construct the proportion of children suffering from any anaemia (mild, moderate or severe), using WHO thresholds: A child is classified as anaemic if adjusted haemoglobin falls below 11.5 g/dL at ages 5 to 11, below 12.0 g/dL at ages 12 to 14, below 12.0 g/dL for non-pregnant females aged 15 and above, and below 13.0 g/dL for males aged 15 and above, with age taken at the date of the biomarker round.

Indicator = 1 if a child in the biomarker sample tests positive for malaria and has no treatment recorded in their diary in the past 4 weeks; 0 otherwise, including test-negative children. Defined over all children tested, so the estimate does not condition on infection status. Because this joint indicator reflects both infection prevalence and treatment coverage, it is interpreted as descriptive/mechanistic and read alongside overall infection prevalence, not as a standalone treatment-gap contrast.

Indicator = 1 if a child tests positive and reports no malaria-suggestive symptom (fever, headache, vomiting, lethargy) at the time of the test or in the four preceding days; 0 otherwise, including negatives. Over all children tested; no conditioning on infection status. Descriptive/mechanistic; read alongside overall prevalence.

Share of correct answers on a health-knowledge battery (HN01, HN04, HHK1, HHK2, DS10, HN05) covering nutrition, hygiene, and specific child illnesses. Responses will be scored as correct or incorrect, with "don't know" treated as incorrect, and summed into an additive index.

Relative expected provider quality for the school health room versus the government clinic, from a paired quality-expectation vignette: respondents estimate, for the same 10 hypothetical febrile children, how many would receive a correct diagnosis and how many a full supply of correct treatment at each provider (0-10 each). Relative-trust index = the school-health-room minus government-clinic expectation, averaged across the diagnosis and treatment items (equivalently, the mean of the two paired differences). Positive values indicate greater expected quality at the school health room.

From a takeit-or-leave-it elicitation in which each guardian is offered a randomly assigned annual fee for the school health room [10, 20, 30, 40 or 50 Kwacha - 30 is the real marginal cost of the programme] and states whether they would pay it. For the control group, the same elicitation takes place but the programme description is hypothetical. The willingness to pay measure is the indicator for accepting the offered price.