Prevalence of Trachoma in 47 Administrative Districts of Zambia: Results of 32 Population-Based Prevalence Surveys

ABSTRACT

Purpose: A number of previous administrative-district-level baseline trachoma prevalence estimates in Zambia required verification. We used methodologies and systems for trachoma surveys considered to represent international best practice in order to generate reliable estimates of the prevalence of trachoma.

Methods: Between March 2016 and July 2017, we undertook 32 population-based prevalence surveys covering 47 administrative districts. In each of the 32 evaluation units (EUs), we selected 31 households in each of 24 clusters. In selected households, trained, certified graders examined all residents aged 1 year and above for evidence of trachomatous inflammation—follicular (TF) and trichiasis. In eyes that had trichiasis, the presence or absence of trachomatous scarring (TS) was recorded, and the subject was asked about previous trichiasis management recommendations from health workers.

Results: Five EUs (encompassing seven administrative districts) had prevalence estimates of trichiasis+TS unknown to the health system in ≥15-year-olds of ≥0.2%, and require public-health-level implementation of trichiasis surgery services. Eleven EUs (encompassing 16 administrative districts) had TF prevalence estimates in 1–9-year-olds of ≥5%. Intervention with the A, F and E components of the SAFE strategy for trachoma elimination is required for nearly 1.5 million people.

Conclusion: Trachoma is a public health problem in some parts of Zambia. The Ministry of Health will continue to partner with other stakeholders to implement the multi-sectoral SAFE strategy. Consideration should be given to re-surveying other suspected-endemic administrative districts in which surveys using older methodologies returned TF prevalence estimates ≥5%.

KEYWORDS:

• Zambia
• trachoma
• trichiasis
• epidemiology
• population-based prevalence survey

Introduction

Zambia was an enthusiastic supporter of the 1998 adoption of World Health Assembly Resolution 51.11, which called for the global elimination of blinding trachoma.¹ The country continues to support that goal, through active participation in the World Health Organization (WHO) Alliance for Global Elimination of Trachoma by 2020,^2,3 and development of strategic plans to eliminate trachoma domestically. This article describes data generation exercises carried out in order to facilitate that planning process.

Trachoma is a neglected tropical disease⁴ caused by the ocular biovar^5,6 of the intracellular bacterium Chlamydia trachomatis. It is common in populations that have inadequate access to water and sanitation.^7–10 In Zambia, as elsewhere, such people tend to live in remote and rural areas, and to be very poor.^11–15

Ocular C. trachomatis is transmitted in eye and nose secretions via fingers, fomites (such as face towels and clothing) and eye-seeking flies,⁷ particularly between members of the same household.^16,17 Infection may be associated with active (inflammatory) trachoma, which often meets the criteria for trachomatous inflammation—follicular (TF) and/or trachomatous inflammation—intense (TI), signs defined within the WHO simplified trachoma grading scheme.¹⁸ Both ocular C. trachomatis infection and active trachoma are more common and more intense in pre-school-age children,^19,20 with immunological factors²¹ and reduced exposure patterns possibly responsible for their lower prevalence in older individuals. Repeated episodes of infection and associated inflammation are needed for the development of significant conjunctival scarring (TS) and for the trachomatous trichiasis (TT) that, in some individuals, supervenes.²² Mathematical modelling suggests that development of these two signs may require more than 100 and 150 C. trachomatis infections, respectively.²³

Blindness from trachoma is prevented using the SAFE strategy,^24–27 which includes surgery for TT, antibiotics to clear infection, and facial cleanliness and environmental improvement to reduce transmission.²⁸ The S component of SAFE should be offered to anyone with TT. The A, F and E components of SAFE are administered to whole populations in which the TF prevalence in 1–9-year-olds is ≥5%. Programmatic planning for public-health-level approaches for reducing both the prevalence of TT and the prevalence of TF relies on prevalence estimates of these signs, which should be generated through population-based surveys.²⁹

In 2012, just prior to the launch of the Global Trachoma Mapping Project (GTMP),³⁰ prevalence surveys were undertaken in each of 65 administrative districts across all 10 provinces of Zambia (Figure 1). These surveys used a variety of approaches,³¹ specifically cluster sampling to generate population-based prevalence estimates according to WHO guidelines,³² and the then-newly-proposed integrated threshold mapping (ITM) methodology.³³ Some of the prevalence estimates that these surveys produced differed markedly from pre-survey expectations. In particular, all nine surveyed administrative districts of Copperbelt Province, which was not historically understood to be trachoma-endemic, had estimates of TF prevalence in 1–9-year-olds that exceeded 10% [unpublished Ministry of Health data]; the near-absence of trichiasis in adults examined in these districts as part of the same surveys could be interpreted as a pointer to recent introduction of trachoma to this population, or a need to reconfirm the TF prevalence estimates. The National Blindness Prevention Committee therefore recommended implementation of a further tranche of surveys in selected districts of Zambia.

Figure 1. Trachoma prevalence surveys undertaken in Zambia, 2012, using either the integrated threshold mapping (ITM) methodology, or a population-based prevalence survey (PBPS) approach. Key to districts: 1. Chibombo; 2. Chisamba; 3. Chitambo; 4. Itezhi tezhi; 5. Kabwe Rural; 6. Kapiri Mposhi; 7. Luano; 8. Mkushi; 9. Mumbwa; 10. Ngabwe; 11. Serenje; 12. Chililbombwe; 13. Chingola; 14. Kalulushi; 15. Kitwe; 16. Luanshya; 17. Lufwanyama; 18. Masaiti; 19. Mpongwe; 20. Mufulira; 21. Ndola; 22. Chadiza; 23. Chipata; 24. Katete; 25. Lundazi; 26. Mambwe; 27. Nyimba; 28. Petauke; 29. Sinda; 30. Vubwi; 31. Chembe; 32. Chienge; 33. Chipili; 34. Kawambwa; 35. Lunga; 36. Mansa; 37. Milenge; 38. Mwansabombwe; 39. Mwense; 40. Nchelenge; 41. Samfya; 42. Chilanga; 43. Chirundu; 44. Chongwe; 45. Kafue; 46. Luangwa; 47. Lusaka; 48. Rufunsa; 49. Shibuyunji; 50. Chama; 51. Chinsali; 52. Isoka; 53. Mafinga; 54. Mpika; 55. Nakonde; 56. Shiwangandu; 57. Chavuma; 58. Ikelenge; 59. Kabompo; 60. Kasempa; 61. Manyinga; 62. Mufumbwe; 63. Mwinilunga; 64. Solwezi; 65. Zambezi; 66. Chilubi; 67. Kaputa; 68. Kasama; 69. Luwingu; 70. Mbala; 71. Mporokoso; 72. Mpulungu; 73. Mungwi; 74. Nsama; 75. Chikankata; 76. Choma; 77. Gwembe; 78. Kalomo; 79. Kazungula; 80. Livingstone; 81. Mazabuka; 82. Monze; 83. Namwala; 84. Pemba; 85. Siavonga; 86. Sinazongwe; 87. Zimba; 88. Kalabo; 89. Kaoma; 90. Limulunga; 91. Luampa; 92. Lukulu; 93. Mitete; 94. Mongu 95. Mulobezi; 96. Mwandi; 97. Nalolo; 98. Nkeyema; 99. Senanga; 100. Sesheke; 101. Shang’ombo; 102. Sikongo; 103. Sioma.

Materials and methods

The survey methodology used was based on that of the GTMP,³⁴ as modified and refined by Tropical Data (www.tropicaldata.org).^35,36 Our approaches were consistent with WHO recommendations for trachoma prevalence surveys.^32,37

Survey teams

Each team was composed of a grader (Ophthalmic Clinical Officer or Ophthalmic Nurse), a recorder (Grade 12 school-leaver), a village guide and a driver. Graders, recorders and team supervisors (ophthalmologists) were trained using the standardized five-day training system detailed in the Tropical Data training manual.³⁸ Only participants who passed stringent tests of competency proceeded to take part in the surveys.³⁴

Sample size

Surveys were powered primarily based on considerations relevant to TF prevalence, with trichiasis prevalence a secondary outcome. Planned sample sizes for each evaluation unit (EU) were consistent with guidance recently published by WHO.³⁹ We sought to have 95% confidence to estimate an expected TF prevalence of 4% with absolute precision of 2%, using a design effect of 2.63, and inflating the result by 20% to account for non-response. This meant that in each EU we needed to include at least the number of households in which 1164 1–9-year-olds would be resident, expecting to examine 970 of them.

Delineation of evaluation units and selection of clusters, households, and individuals

EUs, which were each composed of one or more administrative districts, were framed to encompass populations of roughly 100,000–250,000 people by either taking one administrative district per EU or combining two or more adjacent similar administrative districts. In each EU, 24 clusters (wards) were systematically selected using a probability-proportional-to-ward-size methodology.³² In each selected cluster, 31 households were randomly selected, using compact segment sampling via random draw. In selected households, all residents aged 1 year or above were eligible to participate.

Fieldwork

Fieldwork was completed between March 2016 and July 2017. Provincial and district health offices facilitated community awareness and sensitisation exercises prior to planned survey team visits, using radio messages and community health workers (CHWs). CHWs then served as survey guides. Graders used 2.5× magnifying loupes and sunlight to examine all consenting household residents aged 1 year or above for signs of trachoma. Return visits were arranged to examine residents who were absent at the time of the primary visit.

Data management

Data were entered directly into Android smartphones running the Tropical Data app, a custom-built evolution of the LINKS Android smartphone data collection tool (Task Force for Global Health, Atlanta, GA, USA; https://linkssystem.org).^34,40 At the end of each field day, the recorder uploaded data to the Tropical Data server using a secure, encrypted connection. The Data & Analytics Team checked and cleaned the data while field teams were still in the field; designated health ministry officials reviewed the cleaned data and approved analyses.³⁴ As previously described, these analyses included age standardization of TF prevalence estimates, age- and gender-standardization of trichiasis prevalence estimates, and generation of 95% confidence intervals for each prevalence estimate by bootstrapping, with replacement, the adjusted cluster-level proportions of each sign, over 10,000 replications.³⁴ Owing to current uncertainty over whether it is appropriate to define the trichiasis elimination prevalence threshold counting only those individuals in which TS is found in the same eye as the trichiasis—advanced by some authorities as a possible way to distinguish trachomatous from non-trachomatous disease⁴¹—we present here prevalence estimates for all trichiasis, all trichiasis unknown to the health system, and trichiasis+TS unknown to the health system, each in ≥15-year-olds.

Prevalence categories for TF and trichiasis were provided to the Global Atlas of Trachoma to facilitate planning and global surveillance.^42,43

Ethical considerations

The University of Zambia Biomedical Ethics and Research Committee (reference number 009-03-16) and the London School of Hygiene & Tropical Medicine Research Ethics Committee (6319, 8355) approved the surveys. Provincial and district health offices and local leaders, such as ward councillors and village headmen, were informed and engaged. Survey teams obtained informed verbal consent to proceed from the head of each selected household and informed verbal consent for examination from adults. For examination of minors, the head of the household gave informed verbal consent. Consent was documented in the data collection tool. Examinees with active trachoma were given 1% tetracycline eye ointment to apply to both eyes twice daily for 6 weeks. Examinees with trichiasis were referred to the nearest appropriate health facility for management.

Results

The 32 surveyed EUs had an estimated total population of 5,025,494, of which 98,454 residents (43,987 males, 54,467 females) were enumerated and 91,788 (93%) consenting individuals (39,719 males, 52,069 females) were examined (Table 1) in 23,491 households of 757 clusters. A total of 320 individuals refused to participate, 6334 were absent on the day that a field team visited their household and 12 were not examined for other reasons. Resulting trachoma prevalence estimates are shown in Table 1, alongside comparisons, where available, of previous trachoma prevalence estimates. Figures 2 and 3 display the EU-level prevalence estimates generated by this tranche of mapping.

Table 1. Number of ≥15-year-olds and number of 1–9-year-olds resident, examined, absent and refused in selected households; prevalence of all trichiasis; prevalence of all trichiasis unknown to the health system; prevalence of trichiasis+TS unknown to the health system; and prevalence of trachomatous inflammation—follicular (TF); by evaluation unit, trachoma prevalence surveys, Zambia, 2016–2017, with comparison to prevalence estimates from previous surveys, where available.

				2016–2017 (current) survey data
		2008–2012 survey data		≥15-year-olds							1–9-year-olds
Province	Districts included in evaluation unit [district label in Figure 1–3] (estimated evaluation unit population in 2017)	TF prevalence^e, %	Trichiasis prevalence^c, %	Resident	Examined	Absent	Refused	Prevalence of all trichiasis, %^a (95% CI) [NOC]	Prevalence of all trichiasis UTTHS, %^a (95% CI) [NOC]	Prevalence of trichiasis+TS UTTHS, %^a (95% CI) [NOC]	Resident	Examined	Absent	Refused	TF prevalence, %^b (95% CI)
Luapula	Chembe [31], Mansa [36] (209,090)	C: ND; M: 10.5^d	C: ND; M: 1.7^d	1598	1469	118	11	0.02 (0.00–0.05) [2]	0.02 (0.00–0.05) [2]	0.02 (0.00–0.05) [2]	1,318	1,287	30	1	3.8 (1.5–6.6)
	Mwansabombwe [38], Kawambwa [34] (156,662)	M: ND; K: 7.3^d	M: ND; K: 0.7^d	1736	1609	123	4	0.13 (0.02–0.26) [5]	0.07 (0.00–0.18) [2]	0.05 (0.00–0.16) [1]	1,337	1,310	27	0	1.1 (0.4–1.9)
	Milenge [37], Samfya [41], Lunga [35] (275,242)	M: 16.4^d; S: 7.9^d; L: ND	M: 0.8^d; S: 0.5^d; L: ND	1689	1573	109	7	0.07 (0.01–0.13) [5]	0.05 (0.01–0.09) [4]	0.05 (0.01–0.09) [4]	1,377	1,347	30	0	3.7 (2.3–5.4)
Central	Itezhi tezhi [4], Mumbwa [9] (260,188)	I: 35.8^d; M: 4.0	I: 0.5^d; M: 0.0^d	1973	1830	137	6	0.14 (0.05–0.28) [7]	0.09 (0.01–0.21) [5]	0.09 (0.01–0.21) [5]	1,408	1,355	48	5	4.7 (2.0–7.1)
	Kapiri Mposhi [6] (253,706)	13.3^d	6.9^d	1568	1,414	139	13	0.20 (0.06–0.40) [10]	0.12 (0.02–0.26) [7]	0.09 (0.00–0.22) [5]	1,105	1,067	35	3	5.6 (2.8–9.2)
	Mkushi [8] (155,534)	4.0^d	0.0^d	1493	1378	109	5	0.07 (0.00–0.17) [3]	0.05 (0.00–0.13) [2]	0.03 (0.00–0.09) [1]	1,243	1,223	16	4	6.8 (3.6–11.6)
	Kabwe Rural [5] (199,042)	7.2^d	1.2^d	1313	1170	143	0	0.07 (0.00–0.12) [3]	0.07 (0.00–0.12) [3]	0.07 (0.00–0.12) [3]	808	789	19	0	1.2 (0.5–2.0)
	Serenje [11], Chitambo [3] (274,434)	15.3^d	0.5^d	1590	1434	147	9	0.01 (0.00–0.04) [1]	0.00 (0.00–0.00) [0]	0.00 (0.00–0.00) [0]	1,263	1,230	33	0	4.1 (1.5–6.6)
Lusaka	Kafue [45] (137,883)	7.1^d	0.0^d	1646	1517	128	1	0.08 (0.00–0.25) [2]	0.08 (0.00–0.23) [1]	0.08 (0.00–0.23) [1]	1,045	1,009	36	0	1.8 (1.0–3.0)
Muchinga	Chama [50] (125,670)	15.5^d	4.5^d	1263	1178	76	7	0.79 (0.34–1.36) [15]	0.48 (0.17–0.85) [11]	0.30 (0.05–0.64) [7]	975	954	20	1	4.7 (1.7–9.1)
	Nakonde [55] (98,867)	0.8^d	0.7	1564	1435	126	3	0.23 (0.06–0.51) [8]	0.20 (0.03–0.47) [6]	0.02 (0.00–0.04) [2]	1,045	1,010	35	0	3.8 (1.5–6.6)
	Isoka [52], Mafinga [53] (149,372)	I: 20.0^d; M: ND	I: 1.8^d; M: ND	1584	1474	104	6	0.08 (0.00–0.21) [4]	0.08 (0.00–0.21) [4]	0.05 (0.00–0.16) [3]	1,109	1,080	28	1	2.1 (0.5–3.4)
	Chinsali [51], Shiwangandu [56] (149,664)	C: 28.5^d; S: ND	C: 0.6^d; S: ND	1500	1371	120	8	0.28 (0.08–0.57) [7]	0.15 (0.04–0.31) [4]	0.11 (0.00–0.26) [3]	1,119	1,074	45	0	5.0 (2.5–8.4)
	Mpika [54] (19,436)	9.7^e	0.3^e	1446	1297	143	6	0.39 (0.07–0.81) [10]	0.23 (0.04–0.51) [7]	0.21 (0.02–0.51) [6]	1,105	1,064	38	3	3.8 (1.5–6.6)
Eastern	Chadiza [22], Vubwi [30] (111,646)	C: 14.4^d; V: ND	C: 0.5^d; V: ND	1461	1363	95	3	0.09 (0.00–0.16) [4]	0.09 (0.00–0.16) [4]	0.03 (0.00–0.08) [1]	1,098	1,060	36	2	1.9 (0.6–4.1)
	Lundazi [25] (354,689)	16.0^d	3.7^d	1494	1382	109	3	0.13 (0.01–0.29) [4]	0.07 (0.00–0.20) [2]	0.07 (0.00–0.21) [2]	1,152	1,120	31	1	4.2 (2.5–6.3)
	Katete [24], Mambwe [26] (284,899)	K: 0.5^d; M: 8.2^d	K: 0.2^d; M: 1.5^d	1696	1518	176	2	0.26 (0.03–0.61) [6]	0.21 (0.01–0.57) [5]	0.21 (0.01–0.55) [5]	899	868	30	1	0.1 (0.0–0.4)
Western	Mulobezi [95], Mwandi [96], Sesheke [100] (118,491)	M: ND; M: ND; S: 4.7^d	M: ND; M: ND; S: 0.0	1482	1408	73	0	0.28 (0.11–0.50) [15]	0.21 (0.05–0.44) [11]	0.16 (0.02–0.37) [9]	1,107	1,090	15	2	7.7 (4.1–11.7)
	Nalolo [97], Senanga [99] (110,395)	N: ND; S: 7.4^d	N: ND; S: 0.0^d	1485	1431	51	3	0.45 (0.15–0.90) [12]	0.43 (0.13–0.88) [11]	0.37 (0.08–0.82) [9]	1,074	1,050	23	0	13.4 (9.7–17.3)
Northern	Chilubi [66] (77,804)	15.6^d	0.7^d	1510	1392	109	9	0.28 (0.02–0.47) [7]	0.28 (0.02–0.46) [7]	0.27 (0.00–0.47) [6]	1,333	1,300	31	2	10.4 (6.9–14.3)
	Kaputa [67], Nsama [74] (119,514)	K: 19.6^d; N: ND	K: 1.7^d; N: ND	1516	1450	55	10	0.17 (0.03–0.30) [6]	0.15 (0.00–0.27) [5]	0.15 (0.00–0.27) [5]	1,293	1,261	32	0	5.6 (3.2–8.9)
	Kasama [68] (127,441)	11.4^d	2.8^d	1466	1323	137	6	0.33 (0.08–0.56) [13]	0.14 (0.05–0.27) [9]	0.13 (0.04–0.26) [8]	1,075	1,045	29	1	5.7 (2.4–9.1)
	Luwingu [69] (114,258)	10.1^d	1.6^d	1604	1434	159	11	0.16 (0.03–0.36) [7]	0.13 (0.02–0.31) [5]	0.11 (0.00–0.29) [4]	1,242	1,221	20	1	5.2 (1.9–9.8)
	Mbala [70] (75,475)	3.2^d	0.9^d	1550	1433	112	5	0.13 (0.03–0.24) [5]	0.13 (0.03–0.24) [5]	0.13 (0.03–0.23) [5]	1,071	1,041	28	2	2.5 (1.1–4.1)
	Mporokoso [71] (95,842)	18.7^d	2.4^d	1500	1377	117	6	0.13 (0.05–0.23) [8]	0.09 (0.02–0.18) [5]	0.09 (0.02–0.18) [5]	1,167	1,134	33	0	5.4 (2.8–7.3)
	Mpulungu [72] (98,043)	4.2^d	1.8	1574	1442	125	7	0.55 (0.04–1.45) [7]	0.13 (0.01–0.34) [5]	0.13 (0.01–0.34) [5]	1,158	1,120	37	1	4.7 (2.7–7.0)
	Mungwi [73] (151,058)	16.5^d	2.3^d	1463	1342	111	10	0.45 (0.20–0.71) [14]	0.35 (0.12–0.59) [10]	0.28 (0.08–0.49) [8]	1,124	1,097	26	1	8.6 (5.0–12.2)
Copperbelt	Chililbombwe [12] (79,326)	17.8^d	0.0^d	1718	1513	193	11	0.01 (0.00–0.02) [1]	0.01 (0.00–0.02) [1]	0.01 (0.00–0.02) [1]	980	906	70	4	0.6 (0.2–1.2)
	Chingola [13] (216,626)	13.7^d	0.0^d	1785	1524	240	21	0.08 (0.00–0.16) [2]	0.04 (0.00–0.12) [1]	0.04 (0.00–0.12) [1]	1,038	959	71	8	0.7 (0.3–1.4)
	Kalulushi [14] (100,381)	14.6^d	0.0^d	1591	1328	243	19	0.06 (0.00–0.17) [3]	0.03 (0.00–0.09) [2]	0.00 (0.00–0.00) [0]	965	922	42	1	1.7 (0.4–3.5)
	Luanshya [16] (152,933)	13.6^d	0.0^d	1523	1364	146	13	0.16 (0.00–0.40) [4]	0.15 (0.00–0.39) [3]	0.10 (0.00–0.25) [2]	856	813	35	8	1.0 (0.3–1.9)
	Lufwanyama [17], Mpongwe [19] (171,883)	L: 11.0^d; M: 10.1^d	L: 0.0^d; M: 0.0^d	1145	1075	55	14	0.14 (0.00–0.39) [2]	0.14 (0.00–0.38) [2]	0.14 (0.00–0.38) [2]	831	807	18	6	1.7 (0.6–3.6)

^aAdjusted for gender and age in 5-year age bands (see text).

^bAdjusted for age in 1-year age bands (see text).

^cSingle letters are used in this column to abbreviate the names of administrative districts shown in the column headed “Evaluation unit”, where administrative district-level surveys (as opposed to surveys covering the whole of the evaluation unit) were previously undertaken

^d2012 estimate.

^e2008 estimate.

CI, confidence interval; ND, no data; NOC, number of cases; TS, trachomatous scarring; UTTHS, unknown to the health system.

Figure 2. Prevalence of trichiasis + trachomatous scarring (TS) unknown to the health system (UTTHS), in ≥15-year-olds, by evaluation unit, trachoma prevalence surveys, Zambia, 2016–2017.

Figure 3. Prevalence of trachomatous inflammation—follicular (TF) in 1–9-year-olds, by evaluation unit, trachoma prevalence surveys, Zambia, 2016–2017.

Eleven EUs (34% of EUs mapped, 16 districts, total population 1,473,707) had TF prevalence estimates in children of ≥5%. Eight EUs (12 districts, total population 986,620) had prevalence estimates of trichiasis unknown to the health system in adults of ≥0.2%, of which five EUs (seven districts, total population 618,204) had prevalence estimates of trichiasis+TS unknown to the health system of ≥0.2%. In these populations, trachoma is a public health problem, and interventions are needed.

Discussion

Control, elimination or eradication of a neglected tropical disease like trachoma are thought to deliver multiple benefits to endemic populations. Besides decreasing suffering from the targeted disease, the reduced morbidity that ensues plus the action of implementing control interventions each improves our collective likelihood of delivering results on a range of sustainable development goals.^44–46 In particular, accessing the remote communities in which trachoma is found establishes a beachhead for universal health coverage. Even without considering these knock-on effects, trachoma elimination is objectively cheap,⁴⁷ cost-effective,^48,49 and likely to result in economic gains that significantly exceed the cost of programme implementation.^50,51

In Zambia, after early pilot work,⁵² trachoma elimination has been underway in earnest since 2007, when the first comprehensive tranche of baseline surveys was initiated. From 2007 to 2012, surveys were conducted using several qualitatively and quantitatively different approaches, including ITM. (Relevant data from 2007 to 2012 are included, as comparators for the results of the 2016–2017 surveys, in Table 1.) Later analyses suggested that the use of ITM carried some risk of district misclassification,²⁹ and in the present work, we reverted to the use of cluster-sampled, population-based surveys. Our highly standardized, quality-controlled and quality-assured³⁶ methodologies are considered to provide highly reliable data⁵³; we believe, on this basis, that the prevalence estimates generated here supersede those produced in 2012.

Our teams mapped a total of 32 EUs covering 47 administrative districts (Table 1). In many of these districts, trachoma is still a disease of public health significance. Particular note is made of the two EUs containing Chilubi (Northern Province) and Nalolo and Senanga (Western Province), which had TF prevalence estimates of >10%. Chilubi is mainly an island administrative district and has different demographic characteristics to other administrative districts of the Northern Region. Nalolo and Senanga are adjacent to other administrative districts with known high burdens of disease [unpublished Ministry of Health data]. For both of these EUs, implementation of the full SAFE strategy is now underway.

In the EUs surveyed here, the prevalence of trichiasis+TS unknown to the health system was generally not markedly lower than the prevalence of trichiasis unknown to the health system. There were only two EUs (Mulobezi, Mwandi and Sesheke of Western Province; and Nakonde of Muchinga Province) in which decisions on whether or not to initiate public-health-level trichiasis surgery interventions would differ using these two metrics. Further data and global policy decisions are awaited.

Zambia is now better placed than ever before to eliminate trachoma. The government is strongly focused on health investment, which is seen to target socioeconomic development by stimulating individual, grass-roots productivity.⁵⁴ Implementation of SAFE, a comprehensive, multi-sectoral strategy, can catalyze development partnerships whilst offering primary, secondary and tertiary prevention against trachoma blindness. For many participating communities, previous opportunities to access quality-assured antibiotics and/or modern surgery will have been limited prior to trachoma programme entry. Bilateral agencies and non-governmental organizations fund interventions against NTDs in Zambia⁵⁵; both political will and partner support are therefore in place. Parallel initiatives that may alleviate poverty and thereby reduce trachoma risk, including foreign direct investment (which seems to have greatest impact in poorer environments⁵⁶), social cash transfer⁵⁷ and rural electrification,⁵⁸ are also being pursued.

In health promotion and disease prevention programmes, community participation and empowerment⁵⁹ and implementation of a range of behaviour change techniques⁶⁰ are key. Zambia’s recent re-commissioning of public health nurse and community health assistant training courses are likely to contribute to future trachoma elimination efforts.

Tremendous progress has been made against trachoma globally in the last few decades.⁶¹ Zambia’s Ministry of Health is encouraged to complete mapping for trachoma in the remaining districts of northwestern province, mobilize resources to implement the SAFE strategy where needed, and continue to lead and coordinate stakeholders keen to assist the country to eliminate trachoma as a public health problem nationwide.

Acknowledgments

The authors are grateful to Girija Sankar and Paul M. Emerson of the International Trachoma Initiative for their support to the Zambia trachoma elimination programme.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

Fieldwork for this study was funded by the Queen Elizabeth Diamond Jubilee Trust and the Government of the Republic of Zambia. Systems for providing epidemiological, training, logistical and data management support to health ministries for the purposes of quality assurance and quality control of trachoma prevalence surveys were developed as part of the Global Tropical Mapping Project (GTMP), then adapted and further refined for Tropical Data (www.tropicaldata.org). The GTMP was funded by the United Kingdom’s Department for International Development (DFID) through the GTMP grant (ARIES: 203145) to Sightsavers; and by the United States Agency for International Development (USAID), through the ENVISION project implemented by RTI International under cooperative agreement number AID-OAA-A-11-00048, and the END in Asia project implemented by FHI360 under cooperative agreement number OAA-A-10-00051. Core support to Tropical Data is provided by DFID, the ENVISION project, the Fred Hollows Foundation, the International Trachoma Initiative, Orbis, the Queen Elizabeth Diamond Jubilee Trust, RTI International, Sightsavers, USAID and the World Health Organization (WHO). AWS was a Wellcome Trust Intermediate Clinical Fellow (098521) at the London School of Hygiene & Tropical Medicine, and is now a staff member of WHO. The authors alone are responsible for the views expressed in this article and they do not necessarily represent the views, decisions or policies of the institutions with which they are affiliated.