The study was conducted in four hospitals consisting of three county referral hospitals (Kiambu, Thika, and Gatundu), and PCEA Kikuyu orthopedic (faith-based) selected for providing corrective and rehabilitative outpatient health services to children born with MESBDs in the county. The three county referral hospitals were purposively selected being the only public hospitals providing these services in the county, whereas PCEA Kikuyu orthopedic hospital (faith-based) was selected by simple randomization using sealed envelopes between two faith-based hospitals for providing the same services in the county. A hospital-based cross-sectional study design was adopted to generate cost data from prevalence-based local morbidity data gathered using an ingredient approach to estimate the economic costs of corrective and rehabilitative outpatient health services from health care providers’ perspectives. This was however the best choice of study design for measuring the unit economic cost of health services as an attribute of the population, and thus provided a snapshot of the burden associated with the ‘silent’ public health problem and allowed for generalization of the study results in similar hospital settings in the region. Even though incidence data were readily available and easily accessible for the costing activity, prevalence-based data were excessively preferred to improve the accuracy of the profiled cost data and the estimation of the unit economic costs. This was an economic evaluation study, therefore was reported as per the CHEERS (checklist for consolidated health economic evaluation reporting standards) guidelines ²⁷ .

The study population consisted of all children aged under five years old born to resident women of Kiambu County between January 1 ^st, 2014, and December 31 ^st, 2018, Cases were defined as live births with at least one clinically obvious major external structural birth defect referenced/or described by assistant occupational therapists and/or orthopedic surgeons and presented to the occupational therapy clinics for care from January 1 ^st, 2018, to December 31 ^st, 2018. Caregivers of children born with MESBDs were likely to seek outpatient corrective and rehabilitative health services at the study hospitals whether the children were born in or out of the county. Thus, the eligibility criterion defined above could minimize systemic bias and ensure the reliability of the study results.

The data for cost drivers were gathered retrospectively from health care providers’ perspective for a one-year time horizon between January 1 ^st, 2018, and December 31 ^st, 2018 for purposes of maintaining similar currency conversion, and inflation rates. Discounting for differential timing did not suffice in this study because the value of the resources considered for the analysis were categorized only as current costs.

The total (annual) economic costs were calculated for the defects (349 cases) for computing the unit economic costs as an average of the total costs expressed in Kenya Shillings (KES). The unit economic costs were calculated by dividing by the total annual costs by the number of cases using the following formula: -

Unit economic costs ($) = Total economic costs (KES) Total number ​ of cases

Further, the unit economic costs were converted to United States Dollars ($) at an existing exchange rate of KES 98.00 equivalent to $1.00 in December 2018 using the following formula: -

Unit economic costs ($) = Total economic costs (KES) KES 98.00

This study assumed that the existing currency exchange rate of KES 98.00 in 2018 reflected the Purchasing Power Parity (PPP) globally. Thus, the unit economic costs for corrective and rehabilitative services for the defects were inflated to the U.S Dollar ($) Consumer Price Index (CPI) from January 2018 to December 2018 ³³ .

Before data collection began, the Principal Investigator (PI) recruited and trained four nursing graduates as research assistants (RAs) to ensure that the data abstraction process that spanned from August 1 ^st, 2019 to September 30 ^th, 2019 was carried out in a standardized manner. We adopted an ingredient approach to retrospectively gather the prevalence-based data on caseloads for the cost analysis of corrective and rehabilitative outpatient health services. The cost ingredients of bracing, tendonectomy, and casting interventions quantified consisted of caseloads (morbidity data) by type of MESBDs, the number of braces, the number of bracings, the number of bracing review visits, the number of casting materials, the number of castings, the number of casting review visits, the number of tendonectomies, and the number of tendonectomies review visits considered as direct recurrent costs; and staff emoluments, staff-time, building space for rental, and utility charges categorized as overhead recurrent costs. The prevalence-based morbidity data were retrospectively drawn from outpatient occupational therapist registers; described as medical records containing information on health services provided to children with major external structural birth defects. The information captured in these registers includes dates of clinic visits, outpatient numbers, names of the patients, patients’ age, residence, diagnoses, and therapeutic interventions, among others. Following a predefined inclusion criterion, 349 cases were determined, and associated cost drivers identified, measured, and valued (quantified) using prevailing market prices and entered in a predetermined secondary data abstraction tool. On the other hand, staff-time for the hospitals’ executives comprising the medical superintendents, chief nursing officers, orthopedic surgeons, and health administrative officers were gathered through face-to-face enquires from the occupational therapists being the closest proxies for the officers mentioned above. The ingredient technique and prevalence-based data were chosen for possibly of generating detailed and improved accuracy of the profiled costing data. The data gathered comprised the following: -

Caseloads/morbidity data: Following a predefined exclusion criterion, 349 cases of MESBDs were considered for the cost analyses.

Casting: Castings used to stabilize the affected feet consisted of Plaster of Paris (POP) bandages, orthopedic cotton bandages, and glycerine valued at local market prices. A set of these materials valued at $3.9 were used to cast two cases of club foot. The study computed the number of castings and the number of revisits after the procedure for all cases of CTEV treated using these strategies.

Braces: Braces consisted of leather foot covers, rubber soles, and metallic rods used to stabilize cases of club foot. Braces were sourced from the local markets as ready-made products, therefore, were valued at $15.31 using prevailing market prices. The study also enumerated the number of braces and the number of revisits after the procedure for all cases of CTEV treated using this strategy.

The number of tendonectomies: This was a procedure performed by surgeons to extend the Achilles tendon in club feet. This is largely an outpatient specialized procedure, therefore, existing market hospital charges for outpatient specialized surgical procedures were used to value the cost of tendonectomies estimated at $51.02. Similarly, the number of tendonectomies and the number of revisits after the procedure were computed for all the cases of CTEV that adopted this treatment strategy.

Emolument for personnel: The personnel comprised assistant occupational therapists and support staff whose emolument were estimated based on the respective schemes for staff with at least ten years of work experience ^{34– 36} . Emoluments for assistant occupational therapist consisted of basic salary, house allowance, commuter allowance, health risk allowance, and health extraneous allowance, whereas, for the support staff comprised basic salary, house allowance, and commuter allowance ^{34– 36} . The monthly salary and benefits for an assistant occupational therapist at “Grade 10’’ was valued at $1,224.50, whereas support staff at “Grade 14’’ was valued at $173.50.

Renting building space: Occupational therapy outpatient clinics were identified within the respective study hospitals whose plinth floor surface areas were measured in square feet and valued based on the existing local market rates for renting building spaces. The total renting space for the four hospitals was estimated at 3,593.63 square feet and valued at $0.37 local market value.

Utility charges: Utilities included electricity; and water and sewerage estimated at $25.51 and $30.61 per month, respectively.

Supervisory staff-time: The staff-time for the medical superintends/directors, orthopedic surgeons, chief nursing officers/directors, and health administrative officers/directors were identified and measured by consensus through face-face inquiries made to the assistant occupational therapists being the closest proxies for the above-mentioned officers. The staff-time for medical directors and orthopedic surgeons was measured as a single specialized medical practitioner’s consultation valued at $20.40 and quantified for five days a week for one calendar year. The staff-time for nursing directors and administrative directors on the other hand was measured as a single general medical practitioner’s consultation valued at $10.20 for five days a week for one calendar year.

We obtained ethical approval from Kenyatta National Hospital (KNH)-University of Nairobi (UoN) Ethics Review Committee (Ref. No: KNH-ERC/A/44). Data collected were de-identified using anonymous codes and entered in a laptop secured by an alphanumeric coded key known only to the PI to maintain confidentiality.

Patient consent for publication: Patients were not directly involved because data was drawn from the medical registers, thus consent was not required.

Case ascertainment, information, and systemic biases were expected in this study; therefore, the PI began by predefining an eligibility criterion (case definitions) for participation in the study and predetermining a secondary data abstraction tool for purposes of reducing case ascertainment biases. On the other hand, information biases were reduced by training the data collectors on secondary data extraction techniques from the outpatient occupational therapy clinics and entering data into the abstraction tools to ensure the process was conducted in a standardized manner. Further, all the registers for the entire one-year study period (2018) were reviewed and listed all the cases of external structural birth defects to reduce ascertainment and information biases in this study. Systemic bias was also reduced by excluding cases of delayed milestones, and/or developmental conditions due to management intervention similarities.

Following data collection, filled secondary data abstraction tools were manually checked daily for accuracy and completeness and subsequently entered in a Microsoft Excel spreadsheet (Microsoft Office Professional Plus 2019) by two independent data managers to reduce potential errors. The PI cross-checked and validated the computerized dataset against predetermined data abstraction tools for analyses. Descriptive qualitative categorical variables were summarized in frequency tables, proportions, and percentages to show their distributions, whereas continuous variables were summarised and presented in means (averages). Costs assigned to the direct cost center consisted of; (i) the names and numbers of specific MESBDs (caseloads/morbidity data), (ii) the number of primary and review visits for castings and/or bracings and/or tendonectomies (iii) the number of assistant occupational therapists and their emoluments (iv) the number of support staff and their emoluments, and (v) the floor plinth area for renting. Overheads on the other hand comprised; (i) the number of orthopaedic surgeons and associated staff-time (ii) the number of medical superintendents/directors, and associated staff-time (iii) the number of chief nursing officers/directors and associated staff-time (iv) the number of health administrative officers/directors and associated staff-time, and (v) utility charges for electricity, water and sewerage. All the costs were categorized as recurrent and allocated to the direct cost center whereas overhead costs were shared proportionally among the respective types of MESBDs and allocated to the final costs center for estimation of the economic costs. The unit economic costs were calculated as average costs expressed in U.S Dollars and inflated to the U.S Dollar Consumer Price Index from January 2018 to December 2018 for all types of MESBDs collectively, and individual types of MESBDs. Capital costs did not suffice in this cost study because movable, and fixed capital resources were not considered for valuation because no inventory records existed for furniture, examination couches even though they appeared to have lost more than half of their economic half-lives, and capital donations in kind, whereas motor vehicles, motorcycles, and bicycles were not used either as direct, indirect, or intermediate costs for corrective and rehabilitative health services for the under-fives with MESBDs. The occupational therapy clinics on the other hand, also as fixed capital costs, were exceedingly small portions of the respective hospitals’ floor plinths, hence valued as recurrent costs using prevailing local market prices for building space rental.

Inflation factor and statistical uncertainties: All the resources were categorized as recurrent costs and inflated to the U.S Dollar ($) Consumer Price Index (Calculator) for a one-year time horizon from January 2018 to December 2018 ³³ . This computation adjusted the unit economic costs to purchasing power parity as a factor of inflation, and minimized statistical uncertainties due to cost data scantiness and collection methods ^{32, 33} . Consumer Price Index measures the mean changes in market prices over some time in which consumers pay for goods and services such as health services for MESBDs, thus was preferred because of being the best optimal measure of inflation in this study ³³ .

Of 349 cases 305 (87.39%) comprised congenital talipes equinovarus (CTEV) comprising unilateral congenital talipes equinovarus 300 (85.96%), bilateral congenital talipes equinovarus 3 (0.86%), unilateral congenital talipes equinovarus with germ valgus 1 (0.29%), and congenital talipes equinovarus with spina bifida 1 (0.29%). Additionally, the study observed 35 (10.03%) cases of congenital pes planus (CPP), and 9 (2.58%) cases of neural tube defects (NTD) consisting of hydrocephalus 5 (1.43%), spina bifida 3 (0.86%), and spina bifida with hydrocephalus 1 (0.29%) ( Table 1).

Resource quantification for casting materials costing $3.9 used for two procedures consisted of Plaster of Paris Bandage (7.6cm×2.7m×2pcs), orthopedic cotton bandage (15cm×3m×1pc), and glycerine oil (100 mililitres×1pc) costing $1.84, $1.22, and $0.82, respectively. A set of casting materials valued at $3.9 were used to cast two cases of club foot ( Table 2).

The annual economic costs were estimated and adjusted for inflation factors using the consumer price index calculator (CPI). Of the total unadjusted annual cost for all cases of the observed MESBDs ($392,436.49), almost two-thirds (71.48%) of resource inputs were accounted for by emoluments of occupational therapists, whereas administrative staff-time accounted for about one-quarter (18%) ( Table 3).

Of the total unadjusted economic costs, overhead (shared) costs for these defects consisted of staff-time, staff emoluments, and utilities were estimated at $376,553.05 ( Table 4). Of the total overhead costs ($376,553.05), emoluments for the occupational therapists accounted for three-thirds (74.11%), whereas staff-time costs accounted for 18.72% ( Table 4).

Overheads were allocated proportionally (percentage-based) among the three specific cases: congenital talipes equinovarus (87.39%), congenital pes planus (10.03%), and spina bifida (2.58%) ( Table 5).

Of the total unadjusted annual economic costs for major external structural birth defects; congenital talipes equinovarus was estimated at $343,959.87, whereas congenital pes planus and neural tube defects were estimated at $38,322.97 and $10,153.67, respectively ( Table 6).

Estimation of adjusted economic costs: The unadjusted unit economic costs on the other hand were estimated at $1,124.46 for all the defects, $1,127.74 for congenital talipes equinovarus, $1,094.94 for congenital pes planus, and $ for neural tube defects $1,128.19 ( Table 7). The annual and unit economic costs for all MESBDs observed were adjusted for the inflation factor to $397,765.72. The annual economic costs for congenital talipes equinovarus were estimated at $348,630.80, whereas congenital pes planus and neural tube defects were estimated at $38,843.39, and $10,291.56, respectively ( Table 7). Adjusted unit economic costs on the other hand were estimated at $1,139.73 for all types of MESBDs, $1,143.05 for CTEV, $1,109.81 for CPP, and $1,143.51 for NTD ( Table 7).

The study showed relatively similar unit economic costs of the defects despite wide variations among the caseloads for specific types of MESBDs ( Figure 1).