HLTH 3115 Week 10 Program Evaluation Example

Did the Screening Program Reach Anybody? A First-Year Evaluation Framework for Tuberculosis Screening in Nairobi's Informal Settlements, With Activity Counts and Coverage Kept on Separate Lines

Student Name

Walden University

HLTH 3115: Public and Global Health

Instructor Name

Month Day, Year

Did the Screening Program Reach Anybody? A First-Year Evaluation Framework for Tuberculosis Screening in Nairobi's Informal Settlements, With Activity Counts and Coverage Kept on Separate Lines

The Question Before Effect

The program proposed in Week 9 offers chest X-ray screening for tuberculosis to adults in Nairobi's informal settlements, with a planning population of 500,000 adults and a target of screening at least 60% of them each year, including at least 50% of men aged 25 to 34. Its eventual aim is to reduce the prevalence of tuberculosis, as community-wide screening did in a trial in Vietnam (Marks et al., 2019). Before any claim about prevalence can be made, the first-year evaluation must answer a smaller question: did the program reach the people it was designed for? The RE-AIM framework names this dimension reach, meaning how many of the intended people took part and how closely they resemble the rest, and keeps it apart from effectiveness (Glasgow et al., 1999).

Two Kinds of Number

The evaluation keeps two kinds of number apart.

Activity counts record what the program did: X-rays performed, sputum samples tested, people started on treatment. They have no denominator. A program can report 200,000 X-rays and still have screened only a third of its population, or screened the same willing people twice.

Coverage measures record what share of the intended population the program reached: people screened divided by people eligible. They require a denominator, and the denominator must be the population the program set out to serve, not the people who happened to arrive.

Table 1

Activity Counts and Coverage Measures, Kept Separate

Each coverage measure has its own denominator; reporting the left column without the right would describe effort, not reach.

Which Data Could Support Each Claim

Activity Claims

The program's own registers can support activity claims well. Each screening site records every person X-rayed, every sputum sample sent and every positive result. These counts are reliable for what they measure, provided duplicates are removed when the same person attends twice.

Coverage Claims

Coverage claims are harder, because they depend on the denominator. The program will have three possible sources. The first is the community health workers' household listing, made before screening began in each settlement. It counts the actual people the program invited, by age and sex, and is the best denominator for coverage, but it may miss highly mobile residents and people living in rooms that were not listed. The second is the 2019 census count for the settlements' enumeration areas. It is independent of the program and useful as a check, but it is several years old in places where population changes quickly. The third is the planning figure of 500,000 adults, which should not be used as a denominator for any measured result because it was an estimate made before the program began.

The evaluation will report coverage against the household listing and, as a check, against the census count. If the two differ by more than 20%, the report will say so and treat coverage as uncertain.

Yield and Linkage

Two further measures connect screening to the eventual aim. Yield, the number of cases diagnosed relative to the number expected among those screened, tests whether the program found disease where the survey said it would be. Using the 2016 survey's urban adult estimate of 760 per 100,000 (Enos et al., 2018), screening 300,000 adults would be expected to find roughly 2,280 prevalent cases if the settlements resembled the urban average. A yield far below that would suggest either lower prevalence than assumed or a weak screening pathway. Linkage, the share of diagnosed people who started treatment, tests whether finding cases led to anything. Data for linkage would come from matching the program's positive results against the national treatment register.

Representativeness as Well as Size

Coverage is not only a matter of how many were screened but of who. A program could reach 60% of listed adults by screening almost every woman and older adult while missing most young men, the group the national survey found most affected and least often notified. For that reason, the evaluation reports coverage separately by sex and age group, and compares the characteristics of people screened with those of people listed. If young men are underrepresented among those screened, the program has reached its overall target without reaching its main target, and the report should say so before any other result.

What the First-Year Report Can and Cannot Claim

If coverage reaches 60% of listed adults and 50% of listed young men, the report can say that the program reached its intended population in the first year. It cannot say the program reduced tuberculosis. Prevalence would not be measured until the fourth year, and a single round of screening was not what produced the effect in the Vietnam trial; that effect followed three annual rounds (Marks et al., 2019).

If coverage falls short, the report should say so plainly and should not use activity counts to imply success. A statement such as "the program screened 180,000 adults" must be followed by "36% of the 500,000 listed," if that is the figure, so that readers see the reach as well as the effort.

Conclusion

The first-year evaluation asks whether the program reached anybody before it asks whether the program worked. It keeps activity counts and coverage on separate lines, names the denominator for each coverage measure and states which data source supports each claim. The household listing is the strongest denominator, the census a useful check and the planning figure no denominator at all. Only after reach is established can later years ask about effect.

References

Enos, M., Sitienei, J., Ong'ang'o, J., Mungai, B., Kamene, M., Wambugu, J., Kipruto, H., Manduku, V., Mburu, J., Nyaboke, D., Ngari, F., Omesa, E., Omale, N., Mwirigi, N., Okallo, G., Njoroge, J., Githiomi, M., Mwangi, M., Kirathe, D., ... Weyenga, H. (2018). Kenya tuberculosis prevalence survey 2016: Challenges and opportunities of ending TB in Kenya. PLoS One, 13(12), Article e0209098. https://doi.org/10.1371/journal.pone.0209098

Glasgow, R. E., Vogt, T. M., & Boles, S. M. (1999). Evaluating the public health impact of health promotion interventions: The RE-AIM framework. American Journal of Public Health, 89(9), 1322-1327. https://doi.org/10.2105/AJPH.89.9.1322

Marks, G. B., Nguyen, N. V., Nguyen, P. T. B., Nguyen, T.-A., Nguyen, H. B., Tran, K. H., Nguyen, S. V., Luu, K. B., Tran, D. T. T., Vo, Q. T. N., Le, O. T. T., Nguyen, Y. H., Do, V. Q., Mason, P. H., Nguyen, V.-A. T., Ho, J., Sintchenko, V., Nguyen, L. N., Britton, W. J., & Fox, G. J. (2019). Community-wide screening for tuberculosis in a high-prevalence setting. New England Journal of Medicine, 381(14), 1347-1357. https://doi.org/10.1056/NEJMoa1902129