Novel Oral Poliovirus Vaccine 2 Safety Evaluation during Nationwide Supplemental Immunization Activity, Uganda, 2022.

Given its enhanced genetic stability, novel oral poliovirus vaccine type 2 was deployed for type 2 poliovirus outbreak responses under World Health Organization Emergency Use Listing. We evaluated the safety profile of this vaccine. No safety signals were identified using a multipronged approach of passive and active surveillance.

Epidemiological description of a protracted cholera outbreak in Tonj East and Tonj North counties, former Warrap State, South Sudan, May-Oct 2017.

BACKGROUND: On 18th May 2017, State Ministry of Health of former Warrap State received a report from Tonj East County of an outbreak of acute watery diarrhoea and vomiting in Makuac payam. We conducted this investigation to confirm the causative organism and describe the epidemiology of the outbreak in order to support evidence-based control measures. METHODS: We defined a suspected case as a resident of Tonj East or Tonj North County with sudden onset of acute watery diarrhoea and vomiting between May 1 and October 15, 2017. A probable case was defined as a suspected case with a positive rapid test for Vibrio cholerae; a confirmed case was a probable case with a positive stool culture for V. cholerae. We conducted systematic case finding by visiting health facilities and villages in the affected payams. We reviewed patient records from 1 May 2017 to 15 October 2017, to identify suspected cholera case-patients. We conducted a descriptive epidemiologic study, examining the distribution of the cases. We computed the attack rates by age, sex, and payam of residence. Case fatality rate was calculated as the ratio of the total number of suspected cholera death to the total number of cholera case-patients. We conducted an oral cholera vaccination campaign after the peak of the outbreak to control and prevent the spread to other payams. RESULTS: We identified 1451 suspected cholera cases between May and October 2017. Of these, 81% (21/26) had a positive rapid diagnostic test for V. cholerae; out of the 16 rectal swabs transported to the National Public Laboratory, 88% (14/16) were confirmed to be V. cholerae O1 serotype Inaba. The epidemic curve shows continuous common source outbreak with several peaks. The mean age of the case-patients was 24 years (Range: 0.2-75y). The clinical presentations of the case-patients were consistent with cholera. Males had an attack rate of 9.9/10000. The highest attack rate was in ≥30y (14 per 10,000). Among the six payams affected, Makuac had the highest attack rate of 3/100. The case fatality rate (CFR) was 3.0% (44/1451). Paliang and Wunlit had an oral cholera vaccination coverage of ≥100%, while 4 payams had a vaccination coverage of < 90%. CONCLUSION: This was a continuous common source cholera outbreak caused by V. cholerae 01 sero type Inaba. We recommended strengthening of the surveillance system to improve early detection and effective response.

Preference and uptake of different community-based HIV testing service delivery models among female sex workers along Malaba-Kampala highway, Uganda, 2017.

INTRODUCTION: Female Sex workers (FSW) and their clients accounted for 18% of the new HIV infections in 2015/2016. Special community-based HIV testing service delivery models (static facilities, outreaches, and peer to peer mechanism) were designed in 2012 under the Most At Risk Populations Frame work and implemented to increase access and utilization of HIV care services for key populations like female sex workers. However, to date there is no study that has been done to access the preference and uptake of different community-based HIV testing service delivery models used to reach FSW. We assessed preference and uptake of the current community-based HIV testing services delivery models that are used to reach FSW and identified challenges faced during the implementation of the models. METHODS: We conducted a cross-sectional study design using quantitative (interview with the health workers in facilities providing services to female sex workers and interviews with FSWs) and qualitative (interviews with Ministry of Health staff, health workers, district health team members, program staff at different levels involved in delivery of HIV care services, FSWs and political leaders to assess for the enabling environment created to deliver the different community-based HIV testing services to FSWs along the Malaba-Kampala highway. Malaba - Kampala high way is one of the major high ways with many different hot spots where the actual buying and selling of sex takes place. We defined FSWs as any female, who undertakes sexual activity after consenting with a man for money or other items/benefits as an occupation or as a primary source of livelihood irrespective of site of operation within the past six months. We assessed the preference and uptake of different community based HIV testing services delivery model among FSWs based on two indicators, i.e., the proportion of FSWs who had an HIV Counseling and Testing (HCT) in the last 12 months and the proportion of FSWs who were positive and linked to care. RESULTS: Overall, 86% (390/456) of the FSWs had taken an HIV test in the last 12 months. Of the 390 FSWs, 72% (279/390) had used static facilities, 25% (98/390) had used outreaches, and 3.3% (13/390) used peer to peer mechanisms to have an HIV test. Overall, 35% (159/390) of the FSWs who had taken an HIV test were HIV positive. Of the 159, 83% (132/159) were successfully linked into care. Ninety one percent (120/132) reported to have been linked into care by static facilities. Challenges experienced included; lack of trust in the results given during outreaches, failure to offer other testing services including hepatitis B and syphilis during outreaches, inconsistent supply of testing kits, condoms, STI drugs, and unfriendly health services due to the infrastructure and non-trained health workers delivering KP HIV testing services. CONCLUSIONS: Most of the FSWs had HCT services and were linked to care through static facilities. Community-based HIV testing service delivery models are challenged with inconsistent supply of HIV testing commodities and unfriendly services.. We recommended strengthening of all HIV testing community-based HIV testing service deliverymodels by ensuring constant supply of HIV testing/AIDS care commoditiesoffering FSW friendly services, and provision of comprehensive HIV/AIDS health care package.

Factors contributing to measles transmission during an outbreak in Kamwenge District, Western Uganda, April to August 2015.

BACKGROUND: In April 2015, Kamwenge District, western Uganda reported a measles outbreak. We investigated the outbreak to identify potential exposures that facilitated measles transmission, assess vaccine effectiveness (VE) and vaccination coverage (VC), and recommend prevention and control measures. METHODS: For this investigation, a probable case was defined as onset of fever and generalized maculopapular rash, plus ≥1 of the following symptoms: Coryza, conjunctivitis, or cough. A confirmed case was defined as a probable case plus identification of measles-specific IgM in serum. For case-finding, we reviewed patients' medical records and conducted in-home patient examination. In a case-control study, we compared exposures of case-patients and controls matched by age and village of residence. For children aged 9 m-5y, we estimated VC using the percent of children among the controls who had been vaccinated against measles, and calculated VE using the formula, VE = 1 - ORM-H, where ORM-H was the Mantel-Haenszel odds ratio associated with having a measles vaccination history. RESULTS: We identified 213 probable cases with onset between April and August, 2015. Of 23 blood specimens collected, 78% were positive for measles-specific IgM. Measles attack rate was highest in the youngest age-group, 0-5y (13/10,000), and decreased as age increased. The epidemic curve indicated sustained propagation in the community. Of the 50 case-patients and 200 controls, 42% of case-patients and 12% of controls visited health centers during their likely exposure period (ORM-H = 6.1; 95% CI = 2.7-14). Among children aged 9 m-5y, VE was estimated at 70% (95% CI: 24-88%), and VC at 75% (95% CI: 67-83%). Excessive crowding was observed at all health centers; no patient triage-system existed. CONCLUSIONS: The spread of measles during this outbreak was facilitated by patient mixing at crowded health centers, suboptimal VE and inadequate VC. We recommended emergency immunization campaign targeting children <5y in the affected sub-counties, as well as triaging and isolation of febrile or rash patients visiting health centers.

Risk factors for measles death: Kyegegwa District, western Uganda, February-September, 2015.

BACKGROUND: On 18 August 2015, Kyegegwa District reported eight deaths during a measles outbreak to the Uganda Ministry of Health (MoH). We investigated this death cluster to verify the cause, identify risk factors, and inform public health interventions. METHODS: We defined a probable measles case as onset of fever and generalised rash in a Kyegegwa District resident from 1 February - 15 September 2015, plus ≥1 of the following: coryza, conjunctivitis, and cough. A confirmed measles case was a probable case with measles-specific IgM positivity. A measles death was a death of a probable or confirmed case-person. We conducted an active case-finding to identify measles patients who survived or died. In a case-control study, we compared risk factors between 16 measles patients who died (cases) and 48 who survived (controls), matched by age (±4 years) and village of residence. RESULTS: We identified 94 probable measles cases, 10 (11%) were confirmed by positive measles-specific IgM. Of the 64 probable measles patients aged <5 years, 16 died (case-fatality rate = 25%). In the case-control study, no history of vaccination against measles was found in 94% (15/16) among the case-persons (i.e., measles patients who died) and 54% (26/48) among the controls (i.e., measles patients who survived) (ORM-H = 12; 95% CI = 1.6-104), while 56% (9/16) of case-persons and 67% (17/48) of controls (ORM-H = 2.3; 95% CI =0.74-7.4) did not receive vitamin A supplementation during illness. 63% (10/16) among the case-persons and 6.3% (3/48) of the controls (ORM-H = 33; 95% CI = 6.8-159) were not treated for measles illness at a health facility (a proxy for more appropriate treatment), while 38% (6/16) of the case-persons and 25% (12/48) of the controls (ORM-H = 2.5; 95% CI = 0.67-9.1) were malnourished. CONCLUSION: Lack of vaccination and no treatment in a health facility increased the risk for measles deaths. The one-dose measles vaccination currently in the national vaccination schedule had a protective effect against measles death. We recommended enhancing measles vaccination and adherence to measles treatment guidelines.

Factors Associated with Virological Non-suppression among HIV-Positive Patients on Antiretroviral Therapy in Uganda, August 2014-July 2015.

BACKGROUND: Despite the growing number of people on antiretroviral therapy (ART), there is limited information about virological non-suppression and its determinants among HIV-positive (HIV+) individuals enrolled in HIV care in many resource-limited settings. We estimated the proportion of virologically non-suppressed patients, and identified the factors associated with virological non-suppression. METHODS: We conducted a descriptive cross-sectional study using routinely collected program data from viral load (VL) samples collected across the country for testing at the Central Public Health Laboratories (CPHL) in Uganda. Data were generated between August 2014 and July 2015. We extracted data on socio-demographic, clinical and VL testing results. We defined virological non-suppression as having ≥1000 copies of viral RNA/ml of blood for plasma or ≥5000 copies of viral RNA/ml of blood for dry blood spots. We used logistic regression to identify factors associated with virological non-suppression. RESULTS: The study was composed of 100,678 patients; of these, 94,766(94%) were for routine monitoring, 3492(4%) were suspected treatment failures while 1436(1%) were repeat testers after suspected failure. The overall proportion of non-suppression was 11%. Patients on routine monitoring registered the lowest (10%) proportion of non-suppressed patients. Virological non-suppression was higher among suspected treatment failures (29%) and repeat testers after suspected failure (50%). Repeat testers after suspected failure were six times more likely to have virological non-suppression (ORadj = 6.3, 95%CI = 5.5-7.2) when compared with suspected treatment failures (ORadj = 3.3, 95%CI = 3.0-3.6). The odds of virological non-suppression decreased with increasing age, with children aged 0-4 years (ORadj = 5.3, 95%CI = 4.6-6.1) and young adolescents (ORadj = 4.1, 95%CI = 3.7-4.6) registering the highest odds. Poor adherence (ORadj = 3.4, 95%CI = 2.9-3.9) and having active TB (ORadj = 1.9, 95%CI = 1.6-2.4) increased the odds of virological non-suppression. However, being on second/third line regimens (ORadj = 0.86, 95%CI = 0.78-0.95) protected patients against virological non-suppression. CONCLUSION: Young age, poor adherence and having active TB increased the odds of virological non-suppression while second/third line ART regimens were protective against non-suppression. We recommend close follow up and intensified targeted adherence support for repeat testers after suspected failure, children and adolescents.

A large and persistent outbreak of typhoid fever caused by consuming contaminated water and street-vended beverages: Kampala, Uganda, January - June 2015.

BACKGROUND: On 6 February 2015, Kampala city authorities alerted the Ugandan Ministry of Health of a "strange disease" that killed one person and sickened dozens. We conducted an epidemiologic investigation to identify the nature of the disease, mode of transmission, and risk factors to inform timely and effective control measures. METHODS: We defined a suspected case as onset of fever (≥37.5 °C) for more than 3 days with abdominal pain, headache, negative malaria test or failed anti-malaria treatment, and at least 2 of the following: diarrhea, nausea or vomiting, constipation, fatigue. A probable case was defined as a suspected case with a positive TUBEX® TF test. A confirmed case had blood culture yielding Salmonella Typhi. We conducted a case-control study to compare exposures of 33 suspected case-patients and 78 controls, and tested water and juice samples. RESULTS: From 17 February-12 June, we identified 10,230 suspected, 1038 probable, and 51 confirmed cases. Approximately 22.58% (7/31) of case-patients and 2.56% (2/78) of controls drank water sold in small plastic bags (ORM-H = 8.90; 95%CI = 1.60-49.00); 54.54% (18/33) of case-patients and 19.23% (15/78) of controls consumed locally-made drinks (ORM-H = 4.60; 95%CI: 1.90-11.00). All isolates were susceptible to ciprofloxacin and ceftriaxone. Water and juice samples exhibited evidence of fecal contamination. CONCLUSION: Contaminated water and street-vended beverages were likely vehicles of this outbreak. At our recommendation authorities closed unsafe water sources and supplied safe water to affected areas.

A proposed One Health approach to control yellow fever outbreaks in Uganda.

Yellow Fever (YF) is an acute viral hemorrhagic disease. Uganda is located within the Africa YF belt. Between 2019 and 2022, the Ugandan Health Authorities reported at least one outbreak of YF annually with an estimated 892 suspected cases, on average per year. The persistent recurrence of this disease raises significant concerns about the efficacy of current response strategies and prevention approaches. YF has been recognized as a One Health issue due to its interrelatedness with the animal and environmental domains. Monkeys have been recognized as the virus primary reservoir. The YF virus is transmitted through bites of infected Aedes or Haemagogus species mosquitoes between monkeys and humans. Human activities, monkey health, and environmental health issues (e.g., climate change and land use) impact YF incidence in Uganda. Additionally, disease control programs for other tropical diseases, such as mosquitoes control programs for malaria, impact YF incidence.This review adopts the One Health approach to highlight the limitations in the existing segmented YF control and prevention strategies in Uganda, including the limited health sector surveillance, the geographically localized outbreak response efforts, the lack of a comprehensive vaccination program, the limited collaboration and communication among relevant national and international agencies, and the inadequate vector control practices. Through a One Health approach, we propose establishing a YF elimination taskforce. This taskforce would oversee coordination of YF elimination initiatives, including implementing a comprehensive surveillance system, conducting mass YF vaccination campaigns, integrating mosquito management strategies, and enhancing risk communication. It is anticipated that adopting the One Health approach will reduce the risk of YF incidence and outbreaks.

Factors associated with acute watery diarrhea among children aged 0-59 months in Obongi District, Uganda, April 2022: A case-control study.

Background: Diarrheal diseases, are major concerns for Ugandan children; persistent in Obongi District despite high rotavirus vaccination rates (2019-2021). The district recorded the country's highest annual acute watery diarrhea (AWD) incidence from 2017 to 2021. Our study, conducted in April 2022, assessed AWD risk factors among 0-59-month-old children in Obongi. Methods: We conducted a 1:2 (193:386) unmatched case-control study. A case was a child (0-59 months) with ≥ 3 loose/liquid stools/day, negative malaria/pneumonia tests, residing in Itula/Parolinya subcounty from 1 to 30 April 2022. Medical records from 10 facilities were reviewed. Simple random sampling identified cases, who were interviewed, and controls were randomly chosen from non-AWD neighboring households. Child health cards provided vaccination details. We used logistic regression to identify factors associated with AWD. Results: Among 193 cases and 386 controls, 104 (54 %) cases and 183 (47 %) controls were male, 58 (30 %) cases and 127 (33 %) controls were aged 12-23 months, 187 (97 %) cases and 369 (96 %) controls had received at least one dose of rotavirus vaccine, 58 (30 %) cases and 120 (34 %) controls treated drinking water. Comorbidity presence (undernutrition, diabetes, HIV) (AOR = 12; CI: 2.5-53), caregiver's unwashed hands post-toilet (AOR = 3.9; CI: 1.2-13), and borehole vs. piped water (AOR = 4.0; CI: 1.7-9.6) linked to AWD. Conclusion: Modifiable factors, including failure of caregivers to wash their hands with soap after visiting toilets and use of borehole water were associated with AWD, suggesting that community sensitization on handwashing at critical times, using clean water and soap, and expanded use of piped water could reduce AWD incidence in this area.

Viruses associated with measles-like illnesses in Uganda.

OBJECTIVES: In this study, we investigated the causes of measles-like illnesses (MLI) in the Uganda national surveillance program in order to inform diagnostic assay selection and vaccination strategies. METHODS: We used metagenomic next-generation sequencing (M-NGS) on the Illumina platform to identify viruses associated with MLI (defined as fever and rash in the presence of either cough, coryza or conjunctivitis) in patient samples that had tested IgM negative for measles between 2010 and 2019. RESULTS: Viral genomes were identified in 87/271 (32%) of samples, of which 44/271 (16%) contained 12 known viral pathogens. Expected viruses included rubella, human parvovirus B19, Epstein Barr virus, human herpesvirus 6B, human cytomegalovirus, varicella zoster virus and measles virus (detected within the seronegative window-period of infection) and the blood-borne hepatitis B virus. We also detected Saffold virus, human parvovirus type 4, the human adenovirus C2 and vaccine-associated poliovirus type 1. CONCLUSIONS: The study highlights the presence of undiagnosed viruses causing MLI in Uganda, including vaccine-preventable illnesses. NGS can be used to monitor common viral infections at a population level, especially in regions where such infections are prevalent, including low and middle income countries to guide vaccination policy and optimize diagnostic assays.

Measles outbreak in Semuto Subcounty, Nakaseke District, Uganda, June-August 2021.

Background: Semuto Subcounty reported rubella/measles outbreaks in January 2020 and June-August 2021. This study investigated the outbreak in 2021 to determine the scope, and the factors associated with transmission. Methods: A probable case was defined as a resident of Semuto Subcounty with acute onset of fever and a generalized maculopapular rash with either cough/cold or red eyes from 1 June to 31 August 2021. A confirmed case was defined as a probable case with a blood sample positive for measles-specific IgM. A village-matched case-control study was conducted with 30 cases and 122 controls (1:4 ratio). A control was defined as an individual aged 6 months-9 years, sampled at random, with no signs or symptoms of measles from 1 June to 31 August 2021, residing in the same village as the matched case. Adjusted Mantel-Haenszel odds ratios (ORMH) and confidence intervals (CIs) were calculated. Results: Of the 30 cases (27 probable and three confirmed), 16 (53%) were male. The subcounty attack rate (AR) was 3.2/1000. Children aged 5-9 years were the most affected (AR 5.0/1000). Twenty-two (79%) cases and 116 (97%) controls had ever received measles vaccine (ORMH 0.13, 95% CI 0.03-0.52). Interaction with symptomatic persons at water collection points (ORMH 4.4, 95% CI 1.6-12) and playing at community playgrounds (ORMH 4.2, 95% CI 1.7-11) increased the odds of infection. Conclusions: Socializing/congregating at water collection points and community playgrounds facilitated the transmission of measles in this outbreak.

Comparing static and outreach immunization strategies and associated factors in Uganda, Nov-Dec 2016.

INTRODUCTION: the government of Uganda aims at reducing childhood morbidity through provision of immunization services. We compared the proportion of children 12-33 months reached using either static or outreach immunization strategies and factors affecting utilization of routine vaccination services in order to inform policy updates. METHODS: we adopted the 2015 vaccination coverage cluster survey technique. The sample selection was based on a stratified three-stage sample design. Using the Fleiss formula, a sample of 50 enumeration areas was sufficient to generate immunization coverages at each region. A total of 200 enumeration areas were selected for the survey. Thirty households were selected per enumeration area. Epi-Info software was used to calculate weighted coverage estimates. facility. RESULTS: among the 2231 vaccinated children aged 12-23 months who participated in the survey, 68.1% received immunization services from a health unit and 10.6% from outreaches. The factors that affected utilization of routine vaccination services were; accessibility, where 78.2% resided within 5km from a health. 29.7% missed vaccination due to lack of vaccines at the health facility. Other reasons were lack of supplies at 39.2% and because the caretaker had other things to do, 26.4%. The survey showed 1.8% (40/2271) respondents had not vaccinated their children. Among these, 70% said they had not vaccinated their child because they were busy doing other things and 27.5% had not done so because of lack of motivation. CONCLUSION: almost 7 in 10 children aged 12-23 months access vaccination at health facilities. There is evidence of parental apathy as well as misconceptions about vaccination.

Investigating an outbreak of measles in Kamwenge District, Uganda, July 2015.

Globalization has opened many fronts for disease outbreaks because of the quick movement of people and porous borders around the world. The emergence of zoonotic diseases and other communicable diseases highlights the need for implementation of the Global Health Security Agenda packages if countries are to achieve compliance with International Health Regulations (IHR 2005). Health workforce development is one of the critical components that must be addressed with utmost urgency if gaps in early disease detection and response are to be addressed. In this regard, this case study is based on a measles outbreak investigation in Uganda simulating a real-life outbreak investigation by field epidemiologists and seeks to demonstrate the principles of applied epidemiology outlining the critical steps in outbreak investigations and generation of evidence for decision making. It aims to shore up the health workforce capacity by providing practical training for field epidemiology students and professionals that builds their skills in outbreak investigation. This case study can be completed in less than three hours.

Factors that affect immunization data quality in Kabarole District, Uganda.

INTRODUCTION: Reliable and timely immunization data is vital at all levels of health care to inform decisions and improve program performance. Inadequate data quality may impair our understanding of the true vaccination coverage and also hinder our capability to meet the program objectives. It's therefore important to regularly assess immunization data quality to ensure good performance, sound decision making and efficient use of resources. METHODS: We conducted an immunization data quality audit between July and August 2016. The verification factor was estimated by dividing the recounted diphtheria, pertussis and tetanus third dose vaccination for children under 1 year (DPT3<1 year) by reported DPT3<1 year. The quality of data collection processes was measured using quality indices for the 3 different components: recording practices, storage/reporting, monitoring and evaluation. These indices were applied to the different levels of the health care service delivery system. Quality index score was estimated by dividing the total question or observation correctly answered by the total number of answers/ observations for a particular component. RESULTS: The mean health center verification factor was 87%. Sixty five percent (32/49) of the health centers had consistent data, 27% (13/49) over reported and 4% (2/49) under-reported. Health center 11s and 111s contributed to over-reporting and under-reporting. All the health centers' reports were complete and timely between January and June and from November to December. The mean quality indices for the 3 different componets assessed were; recording practices 66%, storing/reporting 75%, monitoring and evaluation 43%. There was a weak positive correlation between the health center verifaction factor and quality index though this was not statistically significant (r = 0.014; p = 0.92). CONCLUSION: Lower level health centers contributed significantly to the inconsistencies in immunization data; there were wide variation between the quality indices of recording practices, storage/reporting, monitoring and evaluation. We recommended that District Local Governments and Ministry of Health focus on improving data quality at lower levels of health service delivery.

Positive predictive value and effectiveness of measles case-based surveillance in Uganda, 2012-2015.

INTRODUCTION: Disease surveillance is a critical component in the control and elimination of vaccine preventable diseases. The Uganda National Expanded Program on Immunization strives to have a sensitive surveillance system within the Integrated Disease Surveillance and Response (IDSR) framework. We analyzed measles surveillance data to determine the effectiveness of the measles case-based surveillance system and estimate its positive predictive value in order to inform policy and practice. METHODS: An IDSR alert was defined as ≥1 suspected measles case reported by a district in a week, through the electronic Health Management Information System. We defined an alert in the measles case-based surveillance system (CBS) as ≥1 suspected measles case with a blood sample collected for confirmation during the corresponding week in a particular district. Effectiveness of CBS was defined as having ≥80% of IDSR alerts with a blood sample collected for laboratory confirmation. Positive predictive value was defined as the proportion of measles case-patients who also had a positive measles serological result (IgM +). We reviewed case-based surveillance data with laboratory confirmation and measles surveillance data from the electronic Health Management Information System from 2012-2015. RESULTS: A total of 6,974 suspected measles case-persons were investigated by the measles case-based surveillance between 2012 and 2015. Of these, 943 (14%) were measles specific IgM positive. The median age of measles case-persons between 2013 and 2015 was 4.0 years. Between 2013 and 2015, 72% of the IDSR alerts reported in the electronic Health Management Information System, had blood samples collected for laboratory confirmation. This was however less than the WHO recommended standard of ≥80%. The PPV of CBS between 2013 and 2015 was 8.6%. CONCLUSION: In conclusion, the effectiveness of measles case-based surveillance was sub-optimal, while the PPV showed that true measles cases have significantly reduced in Uganda. We recommended strengthening of case-based surveillance to ensure that all suspected measles cases have blood samples collected for laboratory confirmation to improve detection and ensure elimination by 2020.