Meat Inspection Findings at Kiserian Abbatoir, Kajiado County, Kenya
A baseline analysis of digitised ante- and post-mortem inspection records
Abstract
Background. Abattoir meat inspection is both a food-safety control and a passive surveillance system for animal and zoonotic diseases. Kiserian Slaughterhouse in Kajiado County, Kenya, is the pilot site for digitised meat inspection recording on the KABS platform. This analysis establishes a baseline of inspection findings and evaluates the quality of the digital records.
Methods. We analysed all routine ante- and post-mortem inspection records submitted for Kiserian Slaughterhouse from 02 June 2026 to 28 October 2026. Record-level prevalences were estimated with Wilson 95% confidence intervals (CI) and animal-level condemnation rates per 1,000 animals slaughtered with exact Poisson CIs. Negative binomial regression with a log(animals slaughtered) offset estimated incidence rate ratios (IRR) for organ condemnation by species and animal source. Kiserian was compared with the national meat inspection dataset (3,365 reports, 141 abattoirs, 17 counties) using species-adjusted negative binomial models and species-standardised condemnation ratios. Data quality was assessed for completeness, internal consistency and reporting continuity.
Results. 269 inspection records covered 19,735 animals presented and 19,558 approved for slaughter (Cattle, Goats, Sheep by record frequency). Whole-carcass condemnation was rare (0.36 (0.14–0.74) per 1,000 slaughtered). Organ condemnation was reported in 50.6% (44.6–56.5) of records. The liver condemnation rate in cattle was 55.4 (47.0–64.9) per 1,000, and was lower in goats (IRR 0.22 (0.15–0.34)) and sheep (IRR 0.20 (0.13–0.31)). Echinococcus (hydatid) cysts were cited in 73.0% (64.5–80.0) of liver condemnations and liver fluke in 59.8% (51.0–68.1). No Cysticercus bovis was detected among 2,797 cattle carcasses. Compared with national data, Kiserian had lower species-adjusted liver (IRR 0.22 (0.17–0.27)) and lung (IRR 0.09 (0.07–0.12)) condemnation rates, but a much higher share of liver condemnations attributed to hydatid cysts (73.0% vs 36.2%). 91.8% of records carried a reporting date, and records were logged on 92 of 149 calendar days.
Conclusions. Whole-carcass condemnation at Kiserian is low, but organ condemnation, driven by hydatid cysts and liver fluke and concentrated in cattle, points to sustained parasite transmission in the source herds. Organ condemnation rates are lower than national levels, but hydatid cysts make up an unusually large share of liver condemnations, which singles Kajiado out for echinococcosis control. The digital system already supports rate estimation. Mandatory reporting dates, count validation and consistent daily logging would make it suitable for trend surveillance.
Keywords: meat inspection; abattoir surveillance; echinococcosis; fascioliasis; organ condemnation; Kenya; digital surveillance
Key messages for decision-makers
- Hydatid disease is the leading driver of liver losses. It is cited in 73% of liver condemnations, and in 90% of those in goats. Hydatidosis is a zoonosis maintained by dogs, which calls for a joint veterinary–public health (One Health) response in the source communities.
- Cattle carry most of the organ condemnation burden. Cattle livers are condemned at about 4 times the goat rate and lungs at about 12 times, so cattle supply chains should be the priority for parasite control.
- Carcass-level food safety risk is low. Whole-carcass condemnation is 0.04% of animals slaughtered, ante-mortem losses are below 0.1%, and no C. bovis has been detected.
- Kiserian condemns fewer organs than the national average, but more of them for hydatid cysts. Species-adjusted liver and lung condemnation rates are 78% and 91% lower than nationally, yet hydatid cysts account for 73% of liver condemnations against 36% nationally.
- The data system works but has gaps. 8.2% of records are undated, 23 have counts that don’t reconcile, and there was a 20-day gap with no records. Fixing these is low-cost and would turn the system into a reliable trend-surveillance tool.
1 Introduction
Post-mortem inspection at licensed abattoirs is a core control in the farm-to-fork food-safety chain. It is also one of the few sources of systematic, population-level data on the health of slaughtered livestock [1, 2]. Condemnation findings, such as livers rejected for hydatid cysts or fluke, lungs for pneumonia, or whole carcasses for generalised conditions, reflect diseases circulating in the source herds. Several of these are zoonoses of public health importance. In Kenya, meat inspection is governed by the Meat Control Act (Cap. 356) [3], and findings have traditionally been recorded on paper, which limits their timeliness and analytical value.
Kiserian Slaughterhouse, in the peri-urban west of Kajiado County, is the pilot site for digitising meat inspection records on the KABS platform. Each slaughter batch is recorded electronically with its ante-mortem outcomes, carcass and organ condemnations, and the reasons for them.
This report has three objectives:
- to assess the completeness, consistency and continuity of the digital inspection records as a surveillance data source;
- to describe throughput and ante- and post-mortem inspection findings at Kiserian Slaughterhouse during the pilot period; and
- to quantify differences in organ condemnation by species and animal source, and to characterise the aetiology of liver condemnation.
2 Methods
2.1 Study setting
The Kiserian Project digitises meat inspection at two neighbouring slaughter facilities in Keekonyokie Ward, Kajiado West Sub-County, Kajiado County (Figure 1). Both serve the peri-urban market on the south-western edge of Nairobi and draw cattle, sheep and goats mainly from pastoral and agro-pastoral production systems in Kajiado, with occasional consignments from other counties.
- Kiserian Slaughter House, the programme’s pilot site and the subject of this analysis, contributed 269 inspection records covering 19,735 animals between June 2026 and October 2026. Most records were for cattle (38.7%).
- Keekonyokie Slaughter House, about 160 m away, records inspections on the same KABS form. It contributed 185 records covering 4,663 animals between May 2026 and October 2026, mostly cattle (49.2% of records).
The two facilities share the same catchment area. The analyses that follow concern Kiserian Slaughter House; Keekonyokie is described here for context.
2.2 Data source
Inspection data were obtained from the Kiserian Project meat inspection form on the KABS platform. Each record corresponds to one inspection batch of a single species on one day. A batch record captures the number of males and females presented; the numbers dead, rejected, isolated and approved at ante-mortem inspection; the number of whole carcasses condemned; whether organs were condemned, which ones, the number of each and the reasons; partial-condemnation reasons; 17 binary carcass-condition flags; and the county or counties the animals came from. The form was revised during data collection: early submissions have no reporting date or per-organ counts. Both schema versions were harmonised into a single data structure.
2.3 Data processing
Records were restricted to Kiserian Slaughterhouse after free-text facility names had been standardised. Duplicate submissions were removed using the reporting date, facility, location, species and animal counts. Animals presented were defined as the sum of males and females.
2.4 National comparator
For the national comparison, we used the national meat inspection dataset compiled by the Directorate of Veterinary Services covering licensed abattoirs and slaughter slabs across Kenyan counties and was restricted to cattle, sheep and goats. Reports with implausible ante-mortem counts were excluded. Free-text abattoir names were standardised. Variables were harmonised to the Kiserian definitions in Table 1: per-organ counts from the national organ table; liver-condemnation reasons parsed for Echinococcus, fluke and Stilesia hepatica; and the seven carcass conditions recorded on both forms.
2.5 Outcome definitions
| Outcome | Definition | Estimator |
|---|---|---|
| Ante-mortem mortality | Animals dead on arrival or in lairage, per 1,000 animals presented | Rate, exact Poisson 95% CI |
| Ante-mortem rejection | Animals rejected at ante-mortem inspection, per 1,000 animals presented | Rate, exact Poisson 95% CI |
| Whole-carcass condemnation | Carcasses condemned in full, per 1,000 animals slaughtered (approved) | Rate, exact Poisson 95% CI |
| Organ condemnation (record level) | Share of inspection records reporting at least one condemned organ | Proportion, Wilson 95% CI |
| Organ-specific condemnation | Livers / lungs / kidneys condemned per 1,000 animals slaughtered | Rate, exact Poisson 95% CI; negative binomial IRR |
| Liver condemnation aetiology | Share of records with a liver condemnation reason that cite each cause | Proportion, Wilson 95% CI; Fisher's exact test |
| Carcass condition prevalence | Share of inspection records flagging each condition | Proportion, Wilson 95% CI |
| Female share | Female animals as a share of animals presented | Proportion, Wilson 95% CI |
2.6 Statistical analysis
Proportions are reported with Wilson score 95% CIs [4], and rates per 1,000 animals with exact Poisson CIs. Animals within a batch are not independent, so animal-level CIs are indicative and likely too narrow. The regression models below account for this extra-Poisson variation. Organ condemnation counts per batch were modelled by negative binomial regression [5] with log(animals approved) as an offset. Covariates were species (reference: cattle) and whether the batch included animals from outside Kajiado County; results are reported as IRRs with Wald 95% CIs. Calendar month was tested with a likelihood-ratio test in the subset of dated records. Associations between categorical variables were tested with Fisher’s exact test. Kiserian was compared with the national data using negative binomial models with a log(animals) offset, which gave species-adjusted and species-stratified IRRs. Sensitivity analyses restricted the national data to 2026 and excluded the three highest-volume national abattoirs. Facilities and counties were benchmarked with indirectly species-standardised condemnation ratios: observed counts divided by the counts expected at national species-specific rates, with exact Poisson CIs. Statistical significance was set at α = 0.05. Analyses used R 4.5.0 [6] with the MASS, sf [7] and ggplot2 packages. Administrative boundaries are from geoBoundaries [8] and GADM 4.1 [9].
2.7 Data quality assessment
Data quality was assessed on three dimensions:
- Completeness: the share of records with each field filled in. Organ-detail fields were assessed only on the records where they apply, i.e. those listing that organ as condemned.
- Internal consistency: whether animals presented minus animals approved equals the dead, rejected and isolated, and whether approvals exceed animals presented.
- Continuity: the share of calendar days in the dated period with at least one record, and the longest run of days with no records.
2.8 Ethical considerations
This was a secondary analysis of routinely collected, aggregate animal-health surveillance data. The records contain no personal identifiers of farmers, traders or inspectors.
3 Results
3.1 Data quality and reporting continuity
Of 45 fields, 43 were complete for every record. The exceptions were reporting date (91.7%) and ward (99.3%). Organ-detail fields were complete wherever they applied. Internal consistency was weaker (Table 2). In 23 records (8.6%), the ante-mortem counts did not reconcile, and 6 reported more animals approved than presented. In 6 of the 8 records with any ante-mortem loss, deaths equalled rejections, which suggests that dead animals were also logged as rejected. Records were logged on 92 of 149 days in the dated period (Figure 2). The longest run with no records was 20 days, and no records were logged on a Sunday.
| Dimension | Indicator | Value |
|---|---|---|
| Completeness | Records with a reporting date | 91.8% |
| Fields complete for every record | 43 of 45 | |
| Organ-detail fields complete where applicable | 100% | |
| Consistency | Records with unreconciled ante-mortem counts | 23 (8.6%) |
| Records with approvals exceeding animals presented | 6 | |
| Loss records with deaths equal to rejections | 6 of 8 | |
| Continuity | Days with at least one record | 92 of 149 (61.7%) |
| Longest run of days with no records | 20 days |
3.2 Throughput and animal characteristics
269 inspection records covered 19,735 animals presented for slaughter (Table 3). Goats contributed the most animals and cattle the most records, which reflects differences in batch size (median 25.5 animals for cattle, against 82 for goats and 100 for sheep). Females made up 37–40% of animals presented across species. Monthly throughput is shown in Figure 3.
| Species | Records | Animals presented | Median batch size | Approved for slaughter | Female share, % (95% CI) |
|---|---|---|---|---|---|
| Cattle | 104 | 2,826 | 25.5 | 2,797 | 36.7 (34.9–38.5) |
| Goats | 90 | 8,994 | 82.0 | 9,010 | 38.0 (37.0–39.0) |
| Sheep | 75 | 7,915 | 100.0 | 7,751 | 39.7 (38.6–40.8) |
| All | 269 | 19,735 | 64 | 19,558 | 38.5 (37.8–39.2) |
3.3 Ante-mortem findings
Ante-mortem losses were low across species (Table 4). Overall, 7 animals died before slaughter and 7 were rejected, a mortality rate of 0.35 (0.14–0.73) and a rejection rate of 0.35 (0.14–0.73) per 1,000 animals presented.
| Species | Presented | Dead | Rejected | Isolated | Mortality per 1,000 | Rejection per 1,000 | Isolation per 1,000 |
|---|---|---|---|---|---|---|---|
| Cattle | 2,826 | 2 | 2 | 0 | 0.71 (0.09–2.56) | 0.71 (0.09–2.56) | 0.00 (0.00–1.31) |
| Goats | 8,994 | 2 | 3 | 0 | 0.22 (0.03–0.80) | 0.33 (0.07–0.97) | 0.00 (0.00–0.41) |
| Sheep | 7,915 | 3 | 2 | 6 | 0.38 (0.08–1.11) | 0.25 (0.03–0.91) | 0.76 (0.28–1.65) |
| All | 19,735 | 7 | 7 | 6 | 0.35 (0.14–0.73) | 0.35 (0.14–0.73) | 0.30 (0.11–0.66) |
3.4 Post-mortem findings
3.4.1 Whole-carcass condemnation
7 carcasses were condemned in full, a rate of 0.36 (0.14–0.74) per 1,000 animals slaughtered (Table 5). Monthly rates and their 95% CIs overlapped throughout the study period (Figure 4).
| Species | Slaughtered | Carcasses condemned per 1,000 | Records with organ condemnation, % | Livers per 1,000 | Lungs per 1,000 | Kidneys per 1,000 |
|---|---|---|---|---|---|---|
| Cattle | 2,797 | 0.36 (0.01–1.99) | 75.0 (65.9–82.3) | 55.4 (47.0–64.9) | 28.2 (22.4–35.2) | 2.5 (1.0–5.2) |
| Goats | 9,010 | 0.22 (0.03–0.80) | 35.6 (26.4–45.8) | 13.3 (11.0–15.9) | 2.6 (1.6–3.8) | 0.8 (0.3–1.6) |
| Sheep | 7,751 | 0.52 (0.14–1.32) | 34.7 (24.9–45.9) | 11.2 (9.0–13.8) | 2.3 (1.4–3.7) | 0.3 (0.0–0.9) |
| All | 19,558 | 0.36 (0.14–0.74) | 50.6 (44.6–56.5) | 18.5 (16.7–20.5) | 6.1 (5.1–7.3) | 0.8 (0.5–1.3) |
3.4.2 Organ condemnation
At least one condemned organ was reported in 50.6% (44.6–56.5) of records. Livers and lungs dominated: 18.5 (16.7–20.5) livers and 6.1 (5.1–7.3) lungs were condemned per 1,000 animals slaughtered. Organ condemnation rates were markedly higher in cattle than in small ruminants (Figure 5). After adjustment, the liver condemnation rate was 78% lower in goats and 80% lower in sheep than in cattle, and the lung rate around 92% lower in both (Table 6, Figure 6). Batches containing animals from outside Kajiado (n = 12) did not differ significantly in liver (p = 0.982) or lung (p = 0.430) condemnation rates, although the estimates are imprecise. Calendar month was not associated with liver (likelihood-ratio p = 0.153) or lung (p = 0.068) condemnation.
| Outcome | Covariate | IRR (95% CI) | p-value | Dispersion (θ) |
|---|---|---|---|---|
| Liver condemnation | Goats (vs cattle) | 0.22 (0.15–0.34) | <0.001 | 0.76 |
| Sheep (vs cattle) | 0.20 (0.13–0.31) | <0.001 | 0.76 | |
| Animals from outside Kajiado | 1.01 (0.45–2.28) | 0.982 | 0.76 | |
| Lung condemnation | Goats (vs cattle) | 0.09 (0.04–0.19) | <0.001 | 0.21 |
| Sheep (vs cattle) | 0.08 (0.04–0.19) | <0.001 | 0.21 | |
| Animals from outside Kajiado | 0.55 (0.12–2.44) | 0.430 | 0.21 |
3.4.3 Aetiology of liver condemnation
Of the 122 records giving a reason for liver condemnation, 73.0% (64.5–80.0) cited Echinococcus (hydatid) cysts and 59.8% (51.0–68.1) liver fluke. 40 records (32.8%) cited both. The share citing hydatid cysts differed by species, from 63% in cattle to 90% in goats (Fisher’s exact p = 0.013). The share citing fluke did not differ significantly by species (p = 0.176) (Figure 7).
3.4.4 Carcass conditions and partial condemnation
6 of the 17 carcass conditions were recorded at least once (Figure 8). Bruising was the most frequent (7.4% (4.9–11.2) of records); it is a welfare indicator of handling during loading, transport and lairage. Never recorded: Abnormal pigmentation, Arthritis, Ascites, C. bovis, Folliculitis, Myotic disease, Overscalding, Peritonitis, Septic peritonitis, Septic pleurisy, Septicaemia. Partial carcass condemnation was reported in 5.2% of records, most often for fractures (Table 7).
| Reason | Records | Species |
|---|---|---|
| Fractures | 8 | Cattle, Goats, Sheep |
| Multiple abscesses | 2 | Cattle |
| Unspecified (recorded as 'Yes') | 2 | Goats, Sheep |
| Contamination | 1 | Goats |
| Septic pleurisy | 1 | Cattle |
3.4.5 Cysticercus bovis
No C. bovis was detected in any of the 269 inspection records, which cover 2,797 cattle carcasses. The upper 95% confidence limit for prevalence among cattle carcasses, if all were inspected, is 0.13%.
3.5 Animal sources
All records listed Kajiado County as a source of animals. 12 records (4.5%) also included animals from other counties (Table 8). Organ condemnation was reported in 10 of these 12 records (Fisher’s exact p = 0.035); given the small number of records, this association should be confirmed as more data accrue.
| County | Records | Share of records |
|---|---|---|
| Kajiado | 269 | 100.0% |
| Taita Taveta | 6 | 2.2% |
| Garissa | 3 | 1.1% |
| Narok | 2 | 0.7% |
| Machakos | 1 | 0.4% |
3.6 Comparison with national meat inspection data
3.6.1 Comparator dataset
The national comparator comprised 3,365 cattle, sheep and goat reports from 141 abattoirs in 17 counties, reported between August 2024 and August 2026. They covered 1,655,263 animals presented and 1,578,181 slaughtered (Table 9). Kajiado County is not represented in the national dataset, so the comparison is effectively Kiserian against the rest of the reporting counties. National throughput is highly concentrated: the three highest-volume abattoirs account for 75.9% of animals slaughtered. Pooled national rates are therefore dominated by large commercial facilities. The facility- and county-level standardised analyses below address this.
| Dataset | Reports | Abattoirs | Counties | Period | Animals slaughtered | Species mix (cattle / goats / sheep, % of slaughtered) |
|---|---|---|---|---|---|---|
| Kiserian | 269 | 1 | 1 | Jun 2026 – Oct 2026 | 19,558 | 14% / 46% / 40% |
| National | 3,365 | 141 | 17 | Aug 2024 – Aug 2026 | 1,578,181 | 4% / 78% / 18% |
3.6.2 Ante- and post-mortem rates
Table 10 compares crude rates by species. After adjustment for species mix, Kiserian’s liver condemnation rate was 78% lower than the national rate (IRR 0.22 (0.17–0.27), p <0.001). Its lung condemnation rate was 91% lower (IRR 0.09 (0.07–0.12)) (Figure 9). The gap was largest in cattle (liver IRR 0.22 (0.17–0.28)) and sheep (liver IRR 0.13 (0.08–0.20)). For goats, the liver condemnation rate was lower at Kiserian (IRR 0.28 (0.19–0.44)). Whole-carcass condemnation (IRR 1.44 (0.64–3.23)) and ante-mortem rejection (IRR 1.17 (0.45–3.04)) were not significantly different from national levels. Ante-mortem mortality was lower (IRR 0.14 (0.06–0.35)).
| Species | Source | Records | Mortality | Carcasses condemned | Livers | Lungs | Records with organ condemnation, % |
|---|---|---|---|---|---|---|---|
| Cattle | National | 1,122 | 0.78 (0.57–1.04) | 0.55 (0.38–0.77) | 259.4 (255.3–263.5) | 192.6 (189.1–196.2) | 76.0 (73.4–78.4) |
| Kiserian | 104 | 0.71 (0.09–2.56) | 0.36 (0.01–1.99) | 55.4 (47.0–64.9) | 28.2 (22.4–35.2) | 75.0 (65.9–82.3) | |
| Goats | National | 1,283 | 8.7 (8.6–8.9) | 0.09 (0.08–0.11) | 9.7 (9.6–9.9) | 7.1 (7.0–7.3) | 74.2 (71.7–76.5) |
| Kiserian | 90 | 0.22 (0.03–0.80) | 0.22 (0.03–0.80) | 13.3 (11.0–15.9) | 2.6 (1.6–3.8) | 35.6 (26.4–45.8) | |
| Sheep | National | 960 | 1.3 (1.2–1.4) | 0.25 (0.20–0.32) | 43.4 (42.7–44.2) | 33.4 (32.7–34.1) | 65.2 (62.1–68.2) |
| Kiserian | 75 | 0.38 (0.08–1.11) | 0.52 (0.14–1.32) | 11.2 (9.0–13.8) | 2.3 (1.4–3.7) | 34.7 (24.9–45.9) |
The lower organ condemnation rates at Kiserian were robust to restricting the national data to 2026 and to excluding the three highest-volume national abattoirs (Table 11).
| Analysis | Outcome | IRR (95% CI) | p-value |
|---|---|---|---|
| Main analysis (all national reports) | Liver condemnation | 0.22 (0.17–0.27) | <0.001 |
| Lung condemnation | 0.09 (0.07–0.12) | <0.001 | |
| National reports from 2026 only | Liver condemnation | 0.20 (0.16–0.26) | <0.001 |
| Lung condemnation | 0.08 (0.06–0.11) | <0.001 | |
| Excluding the three highest-volume national abattoirs | Liver condemnation | 0.14 (0.12–0.17) | <0.001 |
| Lung condemnation | 0.06 (0.05–0.08) | <0.001 |
3.6.3 Facility benchmarking
Pooled national rates mix very different facilities, so each abattoir with at least five reports was benchmarked with an indirectly species-standardised liver condemnation ratio (SCR). The SCR is the observed number of condemned livers divided by the number expected at national species-specific rates. Kiserian’s SCR was 0.31 (0.28–0.35), lower than 83% of the 48 national abattoirs benchmarked (Figure 10). Facility SCRs varied about 310-fold, which points to genuine differences in disease burden between catchments, differences in inspection and recording practice, or both.
3.6.4 Spatial comparison
County-level standardised liver condemnation ratios varied widely (Figure 11 A). Kajiado, represented by Kiserian, sits below the national level (SCR 0.31). The share of liver condemnations attributed to hydatid cysts at Kiserian is higher than in every reporting county with enough data (Figure 11 B).
3.6.5 Aetiology of liver condemnation
Hydatid cysts were cited in 73.0% (64.5–80.0) of liver condemnations at Kiserian, against 36.2% (34.2–38.4) nationally (Fisher’s exact p <0.001). Liver fluke shares were 59.8% and 62.8% respectively (p 0.502). Nationally, the tapeworm Stilesia hepatica was cited in 44.0% (41.8–46.2) of liver condemnations. The Kiserian form offers no Stilesia option, so that cause cannot be measured there (Table 12).
| Group | Cause | Kiserian | National | p-value |
|---|---|---|---|---|
| All species | Echinococcus (hydatid) cysts | 73.0 (64.5–80.0) | 36.2 (34.2–38.4) | <0.001 |
| Liver fluke | 59.8 (51.0–68.1) | 62.8 (60.7–64.9) | 0.502 | |
| Stilesia hepatica | Not on form | 44.0 (41.8–46.2) | – | |
| Cattle | Echinococcus (hydatid) cysts | 62.7 (50.7–73.3) | 69.1 (65.8–72.3) | 0.275 |
| Goats | Echinococcus (hydatid) cysts | 90.0 (74.4–96.5) | 0.0 (0.0–0.5) | <0.001 |
| Sheep | Echinococcus (hydatid) cysts | 80.0 (60.9–91.1) | 40.8 (36.6–45.1) | <0.001 |
3.6.6 Carcass conditions
Of the seven conditions recorded on both forms, 3 were significantly less frequent at Kiserian and 0 significantly more frequent (Figure 12). Notably, septic pleurisy, multiple abscesses, C. bovis (flagged in 6–11% of national reports) were rare or absent at Kiserian.
3.6.7 Animal profile
Females made up 38–40% of the goats and sheep slaughtered at Kiserian, against 9–13% nationally; for cattle the shares were 37% and 35% (Figure 13). Nationally, pregnant females accounted for 0.67% (0.66–0.68) of animals presented and appeared in 33.3% of reports. The Kiserian form does not record pregnancy.
3.6.8 Data quality
Table 13 compares data-quality indicators. Female counts are reported in every Kiserian record, against 71.3% of national records. 8.6% of Kiserian records have unreconciled ante-mortem counts, against 10.4% nationally. Kiserian, however, lacks two fields the national form captures: pregnancy status and Stilesia hepatica as a liver-condemnation reason.
| Indicator | Kiserian | National |
|---|---|---|
| Records with a reporting date | 91.8% | 100.0% |
| Records reporting female counts | 100.0% | 71.3% |
| Records with unreconciled ante-mortem counts | 8.6% | 10.4% |
| Reports excluded for implausible counts | 0 | 235 (6.5%) |
| Pregnancy status captured | No | Yes |
| Stilesia hepatica available as liver reason | No | Yes |
4 Discussion
4.1 Principal findings
This baseline analysis of digitised meat inspection records from Kiserian Slaughterhouse yields three main findings.
- Carcass-level risk is low. Whole-carcass condemnation and ante-mortem losses were both rare, and no C. bovis was detected.
- Organ condemnation is common and species-dependent. At least one organ was condemned in 51% of inspection records, with liver and lung condemnation several-fold higher in cattle than in goats and sheep.
- Two parasites dominate liver condemnation. Hydatid cysts and liver fluke account for most liver condemnations, and hydatid cysts make up the largest share in goats.
4.2 Comparison with national data
Against the national dataset, Kiserian condemns markedly fewer livers and lungs per animal slaughtered, after adjustment for species and in species-standardised facility and county comparisons. The finding held when the national data were restricted to 2026 and when the largest national abattoirs were excluded. There are two plausible explanations, and they are not mutually exclusive:
- a genuinely lower burden of lung and liver disease in Kajiado’s supply herds;
- differences in recording practice between facilities.
The very large spread of standardised ratios across national abattoirs suggests that recording practice contributes substantially, which argues for harmonised inspection protocols. By contrast, the share of liver condemnations attributed to hydatid cysts at Kiserian is roughly double the national share. Even if Kiserian’s overall condemnation rates are conservative, the echinococcosis signal from Kajiado is strong.
4.3 Performance of the digital recording system
The digital system recorded the core ante- and post-mortem variables reliably: 43 of the 45 form fields were filled in for every record, and the only incomplete fields were reporting date (91.7%) and ward (99.3%). Per-organ counts and reasons were fully recorded wherever they applied, so organ-specific rates and aetiologies can be estimated. Its weaknesses are procedural rather than technical:
- Undated records: 8.2% of records have no reporting date, all from before the date field was introduced.
- Unreconciled counts: 23 records have ante-mortem counts that don’t add up.
- Possible double-counting: dead animals appear to be logged as rejected as well.
- Reporting gaps: there was a 20-day gap with no records, and no Sunday reporting.
These gaps limit trend analysis more than they limit the estimation of rates.
4.4 Strengths and limitations
Strengths. The data are complete for the core variables, and per-organ counts are recorded at batch level. The analysis covers all digitised records from the pilot period, uses exact and score-based interval estimation, and models overdispersion explicitly.
Limitations.
- Record-level data: individual animals cannot be linked to findings, so animal-level CIs ignore clustering within batches.
- Possible reporting bias: findings depend on inspector recording practice and may be under-reported on days with high throughput.
- Limited spatial analysis: almost all animals were sourced from Kajiado, so disease patterns could not be compared between source areas, and the multi-county field doesn’t attribute animals within a batch.
- External validity: this is a single-facility study.
- National comparison: the national dataset covers a longer period (August 2024 to August 2026), uses a different form version, does not include Kajiado, and is dominated by a few large facilities. Rates were therefore compared with species adjustment, standardisation and sensitivity analyses, but residual differences in inspection practice cannot be excluded.
5 Conclusion
Whole-carcass condemnation and ante-mortem losses at Kiserian Slaughterhouse are low, and no C. bovis has been detected. Organ condemnation is the main finding of public health and economic importance. It is driven by hydatid cysts and liver fluke and concentrated in cattle. The digitised inspection records already support robust rate estimation and should be used routinely to guide One Health interventions in the source herds. A few low-cost improvements to data capture would make them suitable for trend surveillance: mandatory dates, count validation, consistent daily logging and a pregnancy field.
Data availability
Inspection records are held on the KABS platform (Directorate of Veterinary Services). The analysis is fully reproducible from the accompanying R and Quarto source, which pulls the records directly from the platform.
References
- Codex Alimentarius Commission. Code of Hygienic Practice for Meat (CAC/RCP 58-2005). Rome: FAO/WHO; 2005.
- World Organisation for Animal Health (WOAH). Terrestrial Animal Health Code. Paris: WOAH.
- Republic of Kenya. Meat Control Act, Cap. 356, Laws of Kenya.
- Wilson EB. Probable inference, the law of succession, and statistical inference. Journal of the American Statistical Association. 1927;22(158):209–212.
- Venables WN, Ripley BD. Modern Applied Statistics with S. 4th ed. New York: Springer; 2002.
- R Core Team. R: A Language and Environment for Statistical Computing. Vienna: R Foundation for Statistical Computing.
- Pebesma E. Simple features for R: standardized support for spatial vector data. The R Journal. 2018;10(1):439–446.
- Runfola D, et al. geoBoundaries: a global database of political administrative boundaries. PLoS ONE. 2020;15(4):e0231866.
- GADM. Database of Global Administrative Areas, version 4.1.