ABSTRACT
BACKGROUND: In this evaluation, we aim to strengthen Routine Health Information Systems (RHIS) through the digitization of data quality assessment (DQA) processes. We leverage electronic data from the Kenya Health Information System (KHIS) which is based on the District Health Information System version 2 (DHIS2) to perform DQAs at scale. We provide a systematic guide to developing composite data quality scores and use these scores to assess data quality in Kenya. METHODS: We evaluated 187 HIV care facilities with electronic medical records across Kenya. Using quarterly, longitudinal KHIS data from January 2011 to June 2018 (total N = 30 quarters), we extracted indicators encompassing general HIV services including services to prevent mother-to-child transmission (PMTCT). We assessed the accuracy (the extent to which data were correct and free of error) of these data using three data-driven composite scores: 1) completeness score; 2) consistency score; and 3) discrepancy score. Completeness refers to the presence of the appropriate amount of data. Consistency refers to uniformity of data across multiple indicators. Discrepancy (measured on a Z-scale) refers to the degree of alignment (or lack thereof) of data with rules that defined the possible valid values for the data. RESULTS: A total of 5,610 unique facility-quarters were extracted from KHIS. The mean completeness score was 61.1% [standard deviation (SD) = 27%]. The mean consistency score was 80% (SD = 16.4%). The mean discrepancy score was 0.07 (SD = 0.22). A strong and positive correlation was identified between the consistency score and discrepancy score (correlation coefficient = 0.77), whereas the correlation of either score with the completeness score was low with a correlation coefficient of -0.12 (with consistency score) and -0.36 (with discrepancy score). General HIV indicators were more complete, but less consistent, and less plausible than PMTCT indicators. CONCLUSION: We observed a lack of correlation between the completeness score and the other two scores. As such, for a holistic DQA, completeness assessment should be paired with the measurement of either consistency or discrepancy to reflect distinct dimensions of data quality. Given the complexity of the discrepancy score, we recommend the simpler consistency score, since they were highly correlated. Routine use of composite scores on KHIS data could enhance efficiencies in DQA at scale as digitization of health information expands and could be applied to other health sectors beyondHIV clinics.
Subject(s)
Data Accuracy , HIV Infections , Humans , Female , Kenya/epidemiology , Retrospective Studies , Infectious Disease Transmission, Vertical/prevention & control , HIV Infections/diagnosis , HIV Infections/epidemiology , HIV Infections/prevention & control , ElectronicsABSTRACT
BACKGROUND: Understanding the magnitude and causes of mortality at national and sub-national levels for countries is critical in facilitating evidence-based prioritization of public health response. We provide comparable cause of death data from Kisumu County, a high HIV and malaria-endemic county in Kenya, and compared them with Kenya and low-and-middle income countries (LMICs). METHODS: We analyzed data from a mortuary-based study at two of the largest hospital mortuaries in Kisumu. Mortality data through 2019 for Kenya and all LMICs were downloaded from the Global Health Data Exchange. We provided age-standardized rates for comparisons of all-cause and cause-specific mortality rates, and distribution of deaths by demographics and Global Burden of Disease (GBD) classifications. RESULTS: The all-cause age-standardized mortality rate (SMR) was significantly higher in Kisumu compared to Kenya and LMICs (1118 vs. 659 vs. 547 per 100,000 population, respectively). Among women, the all-cause SMR in Kisumu was almost twice that of Kenya and double the LMICs rate (1150 vs. 606 vs. 518 per 100,000 population respectively). Among men, the all-cause SMR in Kisumu was approximately one and a half times higher than in Kenya and nearly double that of LMICs (1089 vs. 713 vs. 574 per 100,000 population). In Kisumu and LMICs non-communicable diseases accounted for most (48.0 and 58.1% respectively) deaths, while in Kenya infectious diseases accounted for the majority (49.9%) of deaths. From age 10, mortality rates increased with age across all geographies. The age-specific mortality rate among those under 1 in Kisumu was nearly twice that of Kenya and LMICs (6058 vs. 3157 and 3485 per 100,000 population, respectively). Mortality from injuries among men was at least one and half times that of women in all geographies. CONCLUSION: There is a notable difference in the patterns of mortality rates across the three geographical areas. The double burden of mortality from GBD Group I and Group II diseases with high infant mortality in Kisumu can guide prioritization of public health interventions in the county. This study demonstrates the importance of establishing reliable vital registry systems at sub-national levels as the mortality dynamics and trends are not homogeneous.
Subject(s)
Developing Countries , Global Burden of Disease , Cause of Death , Child , Female , Global Health , Humans , Infant , Kenya/epidemiology , Male , MortalityABSTRACT
Introduction: We examined the contribution of community health workers as frontline responders for the community-based surveillance in Somalia during the first year of the COVID-19 pandemic for detection of COVID-19 cases and identification of contacts. Methods: We retrieved COVID-19 surveillance data from 16 March 2020 to 31 March 2021 from the health ministry's central database. These data were collected through community health workers, health facilities or at the points of entry. We compared the number of suspected COVID-19 cases detected by the three surveillance systems and the proportion that tested positive using the chi-squared test. We used logistic regression analysis to assess association between COVID-19 infection and selected variables. Results: During the study period, 154,004 suspected cases of COVID-19 were detected and tested, of which 10,182 (6.6%) were positive. Of the notified cases, 32.7% were identified through the community-based surveillance system, 54.0% through the facility-based surveillance system, and 13.2% at points of entry. The positivity rate of cases detected by the community health workers was higher than that among those detected at health facilities (8.6% versus 6.4%; p < 0.001). The community health workers also identified more contacts than those identified through the facility-based surveillance (13,279 versus 1,937; p < 0.001). The odds of COVID-19 detection generally increased by age. Community-based surveillance and health facility-based surveillance had similar odds of detecting COVID-19 cases compared with the points-of-entry surveillance (aOR: 7.0 (95% CI: 6.4, 7.8) and aOR: 7.5 (95% CI: 6.8, 8.3), respectively). Conclusion: The community health workers proved their value as first responders to COVID-19. They can be effective in countries with weak health systems for targeted community surveillance in rural and remote areas which are not covered by the facility-based surveillance system.
Subject(s)
COVID-19 , Humans , COVID-19/epidemiology , Community Health Workers , Somalia/epidemiology , Pandemics , Databases, FactualABSTRACT
BACKGROUND: Accurate mortality data associated with infectious diseases such as coronavirus disease 2019 (COVID-19) are often unavailable in countries with fragile health systems such as Somalia. We compared officially reported COVID-19 deaths in Somalia with COVID-19 deaths estimated using verbal autopsy. METHODS: We interviewed relatives of deceased persons to collect information on symptoms, cause, and place of death. We compared these data with officially reported data and estimated the positive and negative predictive values of verbal autopsy. RESULTS: We identified 530 deaths during March-October 2020. We classified 176 (33.2%) as probable COVID-19 deaths. Most deaths (78.5%; 416/530) occurred at home and 144 (34.6%) of these were attributed to COVID-19. The positive predictive value of verbal autopsy was lower for home deaths (22.3%; 95% CI: 15.7-30.1%) than for hospital deaths (32.3%; 95% CI: 16.7-51.4%). The negative predictive value was higher: 97.8% (95% CI: 95.0-99.3%) for home deaths and 98.4% (95% CI: 91.5-100%) for hospital deaths. Conclusions Verbal autopsy has acceptable predictive value to estimate COVID-19 deaths where disease prevalence is high and can provide data on the COVID-19 burden in countries with low testing and weak mortality surveillance where home deaths may be missed.
ABSTRACT
BACKGROUND: Estimating cause-related mortality among the dead is not common, yet for clinical and public health purposes, a lot can be learnt from the dead. HIV/AIDS accounted for the third most frequent cause of deaths in Kenya; 39.7 deaths per 100,000 population in 2019. OraQuick Rapid HIV-1/2 has previously been validated on oral fluid and implemented as a screening assay for HIV self-testing in Kenya among living subjects. We assessed the feasibility and diagnostic accuracy of OraQuick Rapid HIV-1/2 for HIV screening among decedents. METHODS: Trained morticians collected oral fluid from 132 preembalmed and postembalmed decedents aged >18 months at Jaramogi Oginga Odinga Teaching and Referral Hospital mortuary in western Kenya and tested for HIV using OraQuick Rapid HIV-1/2. Test results were compared with those obtained using the national HIV Testing Services algorithm on matched preembalming whole blood specimens as a gold standard (Determine HIV and First Response HIV 1-2-O). We calculated positive predictive values, negative predictive values, area under the curve, and sensitivity and specificity of OraQuick Rapid HIV-1/2 compared with the national HTS algorithm. RESULTS: OraQuick Rapid HIV-1/2 had similar sensitivity of 92.6% [95% confidence interval (CI): 75.7 to 99.1] on preembalmed and postembalmed samples compared with the gold standard. Specificity was 97.1% (95% CI: 91.9 to 99.4) and 95.2% (95% CI: 89.2 to 98.4) preembalming and postembalming, respectively. Preembalming and postembalming positive predictive value was 89.3% (95% CI: 71.8 to 97.7) and 83.3% (95% CI: 65.3 to 94.4), respectively. The area under the curve preembalming and postembalming was 94.9% (95% CI: 89.6 to 100) and 93.9% (95% CI: 88.5 to 99.4), respectively. CONCLUSIONS: The study showed a relatively high-performance sensitivity and specificity of OraQuick Rapid HIV-1/2 test among decedents, similar to those observed among living subjects. OraQuick Rapid HIV-1/2 presents a convenient and less invasive screening test for surveillance of HIV among decedents within a mortuary setting.
Subject(s)
HIV Infections , HIV-1 , HIV Antibodies , HIV Infections/diagnosis , Humans , Infant , Kenya/epidemiology , Reagent Kits, Diagnostic , Sensitivity and SpecificityABSTRACT
BACKGROUND: The cause of death (COD) statement is a vital statistic that refers to the disease(s) and process(es) that lead to death. Obtaining accurate COD is valuable for mortality prevention priorities. The statements are formulated using International Classification of Diseases and related health problems, version 10 (ICD-10) system. However, physicians may be unfamiliar with these standards or fail to use them and instead refer to mechanisms or manner of death when stating COD. We present results of an of assessment of quality of COD statements in decedent cases reviewed during a one-month mortuary-based surveillance at Kenyatta National Hospital (KNH) and the City mortuaries in Nairobi, Kenya in 2015. METHODS: Quality elements reviewed were completeness, correctness and order of stating the immediate (ICOD), antecedent, underlying (UCOD), and other significant causes (OSCs) as per the ICD 10 standards, in all deaths reported among adolescents and adults aged 15 years or older at the two mortuaries. COD were assessed for correct sequencing from immediate, antecedent, to underlying compared with autopsy pathology and clinical findings where available. Errors in COD statements were classified as missing or containing incomplete information such as: lack of underlying cause of an injury; incorrect words or statements; presence of more than one competing COD; use of the mechanism of death or anatomic and physiologic processes or signs and symptoms, and or laboratory results as CODs. Pearson's χ-squared test was used to compare proportions. RESULTS: Out of 810, 610 (75.3%) deaths having HIV statuses were abstracted and 356 had at least one COD documented; 114 (32%) females and 242 (68%) males; 239 (67.1%) from KNH and 117 (32.9%) City mortuary. The cases from City mortuary had higher rates of correct statements on 116 (99.1%) ICOD, 90 (89.1%) UCOD, and 40 (81.6%) OSCs, compared to KNH Mortuary; 50 (20.9%), 200 (90.1%) and 62 (76.5%) respectively, p < 0.001. The most common type of errors was incomplete information and citing mechanisms of death as the COD. CONCLUSIONS: In addition to revising national forms to conform to ICD-10, there is a need for periodic training of individuals responsible for completing death certificates. This will improve correctness and completeness of COD in order to provide reliable mortality data in Kenya.
Subject(s)
Cause of Death , Death Certificates , Quality Control , Adult , Cross-Sectional Studies , Female , HIV Infections/epidemiology , Humans , International Classification of Diseases , Kenya/epidemiology , MaleABSTRACT
BACKGROUND: Death is an important but often unmeasured endpoint in public health HIV surveillance. We sought to describe HIV among deaths using a novel mortuary-based approach in Nairobi, Kenya. METHODS: Cadavers aged 15 years and older at death at Kenyatta National Hospital (KNH) and City Mortuaries were screened consecutively from January 29 to March 3, 2015. Cause of death was abstracted from medical files and death notification forms. Cardiac blood was drawn and tested for HIV infection using the national HIV testing algorithm followed by viral load testing of HIV-positive samples. RESULTS: Of 807 eligible cadavers, 610 (75.6%) had an HIV test result available. Cadavers from KNH had significantly higher HIV positivity at 23.2% (95% CI: 19.3 to 27.7) compared with City Mortuary at 12.6% (95% CI: 8.8 to 17.8), P < 0.001. HIV prevalence was significantly higher among women than men at both City (33.3% vs. 9.2%, P = 0.008) and KNH Mortuary (28.8% vs. 19.0%, P = 0.025). Half (53.3%) of HIV-infected cadavers had no diagnosis before death, and an additional 22.2% were only diagnosed during hospitalization leading to death. Although not statistically significant, 61.9% of males had no previous diagnosis compared with 45.8% of females (P = 0.144). Half (52.3%) of 44 cadavers at KNH with HIV diagnosis before death were on treatment, and 1 in 5 (22.7%) with a previous diagnosis had achieved viral suppression. CONCLUSIONS: HIV prevalence was high among deaths in Nairobi, especially among women, and previous diagnosis among cadavers was low. Establishing routine mortuary surveillance can contribute to monitoring HIV-associated deaths among cadavers sent to mortuaries.
Subject(s)
HIV Infections/mortality , Adolescent , Adult , Aged , Female , Humans , Kenya/epidemiology , Male , Middle Aged , Prevalence , Young AdultABSTRACT
BACKGROUND: A universal health care identifier (UHID) facilitates the development of longitudinal medical records in health care settings where follow up and tracking of persons across health care sectors are needed. HIV case-based surveillance (CBS) entails longitudinal follow up of HIV cases from diagnosis, linkage to care and treatment, and is recommended for second generation HIV surveillance. In the absence of a UHID, records matching, linking, and deduplication may be done using score-based persons matching algorithms. We present a stepwise process of score-based persons matching algorithms based on demographic data to improve HIV CBS and other longitudinal data systems. OBJECTIVE: The aim of this study is to compare deterministic and score-based persons matching algorithms in records linkage and matching using demographic data in settings without a UHID. METHODS: We used HIV CBS pilot data from 124 facilities in 2 high HIV-burden counties (Siaya and Kisumu) in western Kenya. For efficient processing, data were grouped into 3 scenarios within (1) HIV testing services (HTS), (2) HTS-care, and (3) within care. In deterministic matching, we directly compared identifiers and pseudo-identifiers from medical records to determine matches. We used R stringdist package for Jaro, Jaro-Winkler score-based matching and Levenshtein, and Damerau-Levenshtein string edit distance calculation methods. For the Jaro-Winkler method, we used a penalty (Ñ)=0.1 and applied 4 weights (ω) to Levenshtein and Damerau-Levenshtein: deletion ω=0.8, insertion ω=0.8, substitutions ω=1, and transposition ω=0.5. RESULTS: We abstracted 12,157 cases of which 4073/12,157 (33.5%) were from HTS, 1091/12,157 (9.0%) from HTS-care, and 6993/12,157 (57.5%) within care. Using the deterministic process 435/12,157 (3.6%) duplicate records were identified, yielding 96.4% (11,722/12,157) unique cases. Overall, of the score-based methods, Jaro-Winkler yielded the most duplicate records (686/12,157, 5.6%) while Jaro yielded the least duplicates (546/12,157, 4.5%), and Levenshtein and Damerau-Levenshtein yielded 4.6% (563/12,157) duplicates. Specifically, duplicate records yielded by method were: (1) Jaro 5.7% (234/4073) within HTS, 0.4% (4/1091) in HTS-care, and 4.4% (308/6993) within care, (2) Jaro-Winkler 7.4% (302/4073) within HTS, 0.5% (6/1091) in HTS-care, and 5.4% (378/6993) within care, (3) Levenshtein 6.4% (262/4073) within HTS, 0.4% (4/1091) in HTS-care, and 4.2% (297/6993) within care, and (4) Damerau-Levenshtein 6.4% (262/4073) within HTS, 0.4% (4/1091) in HTS-care, and 4.2% (297/6993) within care. CONCLUSIONS: Without deduplication, over reporting occurs across the care and treatment cascade. Jaro-Winkler score-based matching performed the best in identifying matches. A pragmatic estimate of duplicates in health care settings can provide a corrective factor for modeled estimates, for targeting and program planning. We propose that even without a UHID, standard national deduplication and persons-matching algorithm that utilizes demographic data would improve accuracy in monitoring HIV care clinical cascades.