Type: Package
Title: Simplified Lot Quality Assurance Sampling Evaluation of Access and Coverage (SLEAC) Tools
Version: 0.1.3
Description: In the recent past, measurement of coverage has been mainly through two-stage cluster sampled surveys either as part of a nutrition assessment or through a specific coverage survey known as Centric Systematic Area Sampling (CSAS). However, such methods are resource intensive and often only used for final programme evaluation meaning results arrive too late for programme adaptation. SLEAC, which stands for Simplified Lot Quality Assurance Sampling Evaluation of Access and Coverage, is a low resource method designed specifically to address this limitation and is used regularly for monitoring, planning and importantly, timely improvement to programme quality, both for agency and Ministry of Health (MoH) led programmes. SLEAC is designed to complement the Semi-quantitative Evaluation of Access and Coverage (SQUEAC) method. This package provides functions for use in conducting a SLEAC assessment.
License: GPL (≥ 3)
Depends: R (≥ 4.2.0)
Imports: cli, methods, parallel, parallelly, stats
Suggests: covr, knitr, rmarkdown, spelling, testthat (≥ 3.0.0), tibble
Encoding: UTF-8
Language: en-GB
LazyData: true
RoxygenNote: 7.3.3
URL: https://nutriverse.io/sleacr/, https://github.com/nutriverse/sleacr
BugReports: https://github.com/nutriverse/sleacr/issues
Config/testthat/edition: 3
VignetteBuilder: knitr
NeedsCompilation: no
Packaged: 2026-01-31 21:40:03 UTC; ernestguevarra
Author: Ernest Guevarra ORCID iD [aut, cre, cph], Mark Myatt ORCID iD [aut, cph]
Maintainer: Ernest Guevarra <ernest@guevarra.io>
Repository: CRAN
Date/Publication: 2026-02-03 14:30:02 UTC

Simplified Lot Quality Assurance Sampling Evaluation of Access and Coverage (SLEAC) Tools

Description

In the recent past, measurement of coverage has been mainly through two-stage cluster sampled surveys either as part of a nutrition assessment or through a specific coverage survey known as Centric Systematic Area Sampling (CSAS). However, such methods are resource intensive and often only used for final programme evaluation meaning results arrive too late for programme adaptation. SLEAC, which stands for Simplified Lot Quality Assurance Sampling Evaluation of Access and Coverage, is a low resource method designed specifically to address this limitation and is used regularly for monitoring, planning and importantly, timely improvement to programme quality, both for agency and Ministry of Health (MoH) led programmes. This package provides functions for use in conducting a SLEAC assessment.

Author(s)

Maintainer: Ernest Guevarra ernest@guevarra.io (ORCID) [copyright holder]

Authors:

References

Myatt, Mark et al. 2012. Semi-Quantitative Evaluation of Access and Coverage (SQUEAC)/Simplified Lot Quality Assurance Sampling Evaluation of Access and Coverage (SLEAC) Technical Reference. Washington, DC: FHI 360/FANTA.

See Also

Useful links:

Calculate case-finding effectiveness

Description

Calculate case-finding effectiveness

Usage

calculate_cf(cin, cout)

Calculate r_out

Description

Calculate r_out

Usage

calculate_rout(cin, cout, rin, k = 3)

Calculate treatment coverage

Description

Calculate treatment coverage

Usage

calculate_tc(cin, cout, rin, k = 3)

Calculate post-stratification weights

Description

Calculate post-stratification weights

Usage

calculate_weights(pop_df, u5, p)

Check coverage data for post-stratification weighted estimation

Description

Check coverage data for post-stratification weighted estimation

Usage

check_coverage_data(cov_df)

Arguments

cov_df

A data.frame() of stratified coverage survey data.

Value

A message or an error on whether cov_df is structured appropriately for post-stratification weighted estimation.

Check coverage distribution

Description

Check coverage distribution

Usage

check_coverage_homogeneity(cov_df, k = 3, p = 0.05)

Arguments

cov_df

A data.frame() of stratified coverage survey data to get overall coverage estimates of. cov_df should have a variable named cases_in for number of SAM or MAM cases in the programme found during the survey, cases_out for number SAM or MAM cases not in the programme found during the survey, and rec_in for children recovering from SAM or MAM who are in the programme found during the survey. A final required variable should be one that contains identifying geographical information corresponding to the location from which each row of the survey data was collected from.

k

Correction factor. Ratio of the mean length of an untreated episode to the mean length of a CMAM treatment episode

p

Minimum p-value to test statistic. Default is 0.05.

Value

A named list of 2 lists: one for case-finding effectiveness (cf) and the second for treatment coverage (tc). For each list, the following values are provided:

Examples

check_coverage_homogeneity(survey_data)

Check p value

Description

Check p value

Usage

check_p(p)

Check population data for post-stratification weighted estimation

Description

Check population data for post-stratification weighted estimation

Usage

check_pop_data(pop_df)

Arguments

pop_df

A data.frame() of population data.

Value

A message or an error on whether pop_df is structured appropriately for post-stratification weighted estimation.

Weighted post-stratification estimation of coverage over several service delivery units

Description

Weighted post-stratification estimation of coverage over several service delivery units

Usage

estimate_coverage_overall(cov_df, pop_df, strata, u5, p, k = 3)

estimate_coverage(cov_df, cov_type = c("cf", "tc"), k = 3)

Arguments

cov_df

A data.frame() of stratified coverage survey data to get overall coverage estimates of. cov_df should have a variable named cases_in for number of SAM or MAM cases in the programme found during the survey, cases_out for number SAM or MAM cases not in the programme found during the survey, and rec_in for children recovering from SAM or MAM who are in the programme found during the survey. A final required variable should be one that contains identifying geographical information corresponding to the location from which each row of the survey data was collected from.

pop_df

A data.frame() with at least two variables: strata for the stratification/grouping information that matches the grouping information in cov_df and pop for information on population for the given grouping information.

strata

A character value of the variable name in cov_df that corresponds to the strata values to match with values in pop_df.

u5

A numeric value for the proportion of the population that is under years old.

p

Prevalence of SAM or MAM in the given population.

k

Correction factor. Ratio of the mean length of an untreated episode to the mean length of a CMAM treatment episode

cov_type

Coverage estimator to report. Either "cf" for case-finding effectiveness or "tc" for treatment coverage. Default is "cf".

Value

A list of overall coverage estimates with corresponding 95% confidence intervals for case-finding effectiveness and treatment coverage.

Examples

cov_df <- survey_data

pop_df <- pop_data |>
  setNames(nm = c("strata", "pop"))

estimate_coverage_overall(
  cov_df, pop_df, strata = "district", u5 = 0.177, p = 0.01
)

Calculate the binomial coefficient n-choose-k

Description

Calculate the binomial coefficient n-choose-k

Usage

get_binom_hypergeom(n, k)

Arguments

n

Total population

k

Number of sample drawn from total population

Value

A numeric value of binomial probability

Get low, moderate, and high probabilities per proportion

Description

Get low, moderate, and high probabilities per proportion

Usage

get_classification_probabilities(x)

Arguments

x

A vector of low, moderate, and high classification labelled as 1, 2, and 3 respectively.

Value

A table object of low, moderate, and high classification probabilities

Calculate hypergeometric probability

Description

Calculate hypergeometric probability

Usage

get_hypergeom(k, m, n, N)

Arguments

k

Number of cases in the sample

m

Number of cases in the population

n

Sample size

N

Population size

Value

A numeric value of hypergeometric probability given specified parameters

Calculate cumulative hypergeometric probabilities

Description

Calculate cumulative hypergeometric probabilities

Usage

get_hypergeom_cumulative(k, m, n, N, tail = "lower")

Arguments

k

Number of cases in the sample

m

Number of cases in the population

n

Sample size

N

Population size

tail

A character vector indicating "lower" (default) or "upper" tail

Value

A numeric value of cumulative hypergeometric probability given specified parameters

Calculate estimated number of cases for a condition affecting children under 5 years old in a specified survey area

Description

Calculate estimated number of cases for a condition affecting children under 5 years old in a specified survey area

Usage

get_n_cases(N, u5, p)

Arguments

N

Population for all ages in the specified survey area.

u5

Proportion (value from 0 to 1) of population that are aged 6-59 months.

p

Prevalence (value from 0 to 1) of condition that is to be assessed.

Value

A numeric value of the estimated number of cases in the specified survey area

Examples

## Calculate number of SAM cases in a population of 100000 persons of all
## ages with an under-5 population of 17% and a prevalence of 2%
get_n_cases(N = 100000, u5 = 0.17, p = 0.02)

Calculate number of clusters to sample to reach target sample size

Description

Calculate number of clusters to sample to reach target sample size

Usage

get_n_clusters(n, n_cluster, u5, p)

Arguments

n

Target sample size of cases for the coverage survey.

n_cluster

Average cluster population for all ages in the specified survey area.

u5

Proportion (value from 0 to 1) of population that are aged 6-59 months.

p

Prevalence (value from 0 to 1) of condition that is to be assessed.

Value

A numeric value of the estimated number of clusters to sample to reach target sample size.

Examples

## Calculate number of villages to sample given an average village population
## of 600 persons of all ages with an under-5 population of 17% and a
## prevalence of SAM of 2% if the target sample size is 40
get_n_clusters(n = 40, n_cluster = 600, u5 = 0.17, p = 0.02)

Calculate sample size and decision rule for a specified LQAS sampling plan

Description

Calculate sample size and decision rule for a specified LQAS sampling plan

Usage

get_sample_n(N, dLower, dUpper, alpha = 0.1, beta = 0.1)

get_sample_d(N, n, dLower, dUpper, alpha = 0.1, beta = 0.1)

Arguments

N

Total population size of cases in the specified survey area

dLower

Lower triage threshold. Values from 0 to 1.

dUpper

Upper triage threshold. Values from 0 to 1.

alpha

Maximum tolerable alpha error. Values from 0 to 1. Default is 0.1

beta

Maximum tolerable beta error. Values from 0 to 1. Default is 0.1

n

Sample size

Value

A list of values providing the LQAS sampling plan for the specified parameters. The list includes sample size, decision rule, alpha error and beta error for the specified classification scheme

Examples

get_sample_n(N = 600, dLower = 0.7, dUpper = 0.9)
get_sample_d(N = 600, n = 40, dLower = 0.7, dUpper = 0.9)

Select sampling clusters using systematic sampling

Description

Select sampling clusters using systematic sampling

Usage

get_sampling_clusters(N_clusters, n_clusters)

get_sampling_list(cluster_list, n_clusters)

Arguments

N_clusters

Total number of clusters in survey area.

n_clusters

Number of sampling clusters to be selected.

cluster_list

A data.frame containing at least the name or any other identifier for the entire set of clusters to sample from.

Value

An integer vector for get_sampling_clusters() giving the row index for selected clusters. A data.frame for ⁠[get_sampling_list()]⁠ which is a subset of cluster_list.

Examples

get_sampling_clusters(N_clusters = 211, n_clusters = 35)
get_sampling_list(cluster_list = village_list, n_clusters = 70)

LQAS classifier

Description

LQAS classifier

Usage

lqas_classify_(
  cases_in,
  cases_out,
  rec_in = NULL,
  k = 3,
  threshold = c(0.2, 0.5),
  label = FALSE
)

lqas_classify(
  cases_in,
  cases_out,
  rec_in = NULL,
  k = 3,
  threshold = c(0.2, 0.5),
  label = FALSE
)

lqas_classify_cf(cases_in, cases_out, threshold = c(0.2, 0.5), label = FALSE)

lqas_classify_tc(
  cases_in,
  cases_out,
  rec_in,
  k,
  threshold = c(0.2, 0.5),
  label = FALSE
)

Arguments

cases_in

Number of SAM and/or MAM cases found during the survey who are in the CMAM programme.

cases_out

Number of SAM and/or MAM cases found during the survey who are in the CMAM programme.

rec_in

Number of children recovering from SAM or MAM found during the survey who are in the programme.

k

Correction factor. Ratio of the mean length of an untreated episode to the mean length of a CMAM treatment episode

threshold

Decision rule threshold/s. Should be between 0 and 1. At least one threshold should be provided for a two-tier classifier. Two thresholds should be provided for a three-tier classifier. Default is a three-tier classifier with rule set at 0.2 and 0.5.

label

Logical. Should the output results be classification labels? If TRUE, output classification are character labels else they are integer values. Default is FALSE.

Value

A data.frame() of coverage classifications for case-finding effectiveness and for treatment coverage.

Author(s)

Ernest Guevarra

Examples

lqas_classify(cases_in = 6, cases_out = 34, rec_in = 6)

with(
  survey_data,
  lqas_classify(
    cases_in = cases_in, cases_out = cases_out, rec_in = rec_in
  )
)

Produce misclassification probabilities

Description

Produce misclassification probabilities

Usage

lqas_get_class_prob(x)

Arguments

x

Simulated results data produced by lqas_simulate_test()

Value

A list object of class lqasClass for LQAS misclassification probabilities results

Examples

sim <- lqas_simulate_test(
  pop = 10000, n = 40, dLower = 0.6, dUpper = 0.9, replicates = 5, runs = 5
)

lqas_get_class_prob(x = sim)

Simulate survey data of covered/cases and non-covered/non-cases given a coverage/prevalence proportion

Description

Simulate survey data of covered/cases and non-covered/non-cases given a coverage/prevalence proportion

Perform LQAS on simulated data based on specified decision rules

Usage

lqas_simulate_population(proportion, pop)

lqas_simulate_run(proportion, pop, n, dLower, dUpper)

lqas_simulate_runs(
  pop,
  n,
  dLower,
  dUpper,
  pLower = 0,
  pUpper = 1,
  fine = 0.01,
  runs = 50,
  cores = parallelly::availableCores(omit = 1)
)

lqas_simulate_test(
  pop,
  n,
  dLower,
  dUpper,
  pLower = 0,
  pUpper = 1,
  fine = 0.01,
  runs = 50,
  replicates = 20,
  cores = parallelly::availableCores(omit = 1)
)

Arguments

proportion

A numeric value of a coverage/prevalence proportion to simulate on. Values should be between 0 and 1.

pop

Population size from which simulated coverage survey data is to be taken from.

n

Sample size of actual or test coverage data.

dLower

A numeric value for the lower classification threshold proportion. Value should be between 0 and 1.

dUpper

A numeric value for the upper classification threshold proportion. Value should be between 0 and 1.

pLower

Starting proportion for simulations. Default is 0.

pUpper

Ending proportion for simulations. Default is 1.

fine

Granularity of simulated proportions. Default is 0.01.

runs

Number of simulation runs to perform per coverage proportion. Default is 50 runs.

cores

The number of computer cores to use/number of child processes will be run simultaneously.

replicates

Number of replicate LQAS simulations to perform. Default is set to 20 replicates.

Value

A data.frame with 2 variables: id for unique identifier and case for numeric vector of cases and non-cases (1s and 0s)

A data.frame with variable cases for total number of covered/cases, outcome for LQAS outcome, and proportion for the coverage/prevalence proportion being simulated on. For ⁠[lqas_simulate_test()]⁠, a 'sleac“ class object.

Examples

lqas_simulate_population(proportion = 0.3, pop = 10000)

lqas_simulate_run(
  proportion = 0.3, pop = 10000, n = 40, dLower = 0.6, dUpper = 0.9
)

lqas_simulate_runs(
  pop = 10000, n = 40, dLower = 0.6, dUpper = 0.9, runs = 10
)

lqas_simulate_test(
  pop = 10000, n = 40, dLower = 0.6, dUpper = 0.9, runs = 5, replicates = 5
)

plot helper function for lqas_simulate_test() function

Description

plot helper function for lqas_simulate_test() function

Usage

## S3 method for class 'lqasSim'
plot(x, ...)

Arguments

x

An object of class lqasSim produced by lqas_simulate_test() function

...

Additional plot parameters

Value

An LQAS probability of classification plot

Examples

x <- lqas_simulate_test(
  pop = 10000, n = 40, dLower = 0.6, dUpper = 0.9, replicates = 5, runs = 5
)
plot(x)

Population data of districts of Sierra Leone based on 2015 census

Description

Population data of districts of Sierra Leone based on 2015 census

Usage

pop_data

Format

A tibble with 14 rows and 2 columns

Variable Description
district District name
pop Population

Source

https://sierraleone.unfpa.org/sites/default/files/pub-pdf/Population%20structure%20Report_1.pdf

Examples

pop_data

print helper function for lqas_get_class_prob() function

Description

print helper function for lqas_get_class_prob() function

Usage

## S3 method for class 'lqasClass'
print(x, ...)

Arguments

x

An object resulting from applying the lqas_get_class_prob() function.

...

Additional print parameters

Value

Printed output of lqas_get_class_prob() function

Examples

sim <- lqas_simulate_test(
  pop = 10000, n = 40, dLower = 0.6, dUpper = 0.9, replicates = 5, runs = 5
)

x <- lqas_get_class_prob(x = sim)
print(x)

SLEAC survey data from Sierra Leone

Description

SLEAC survey data from Sierra Leone

Usage

survey_data

Format

A tibble with 14 rows and 6 columns

Variable Description
country Country
province Province
district District
cases_in SAM cases found who are in the programme
cases_out SAM cases found who are not in the programme
rec_in Recovering SAM cases in the programme
cases_total Total number SAM cases found

Source

Ministry of Health, Sierra Leone

Examples

survey_data

List of villages in Bo District, Sierra Leone

Description

List of villages in Bo District, Sierra Leone

Usage

village_list

Format

A tibble with 1001 rows and 4 columns

Variable Description
id Unique identifier
chiefdom Chiefdom
section Section
village Village

Source

Ministry of Health, Sierra Leone

Examples

village_list