Open Access

Underestimation of prevalence of raised blood sugar from history compared to biochemical estimation: support for the WHO rule of halves in a population based survey in Eritrea of 2009


Received: 2 June 2015

Accepted: 6 November 2015

Published: 24 November 2015


To ascertain the prevalence of diabetes mellitus from history and biochemical estimation so as to attest the WHO Rule of halves in a lean population. A population based national survey on diabetes mellitus was carried out in 2009. History and fingerpicks blood analysis were examined according to recommended procedures of the WHO STEPwise approach and the WHO recommended automated machine to compare the two modalities of estimating diabetes prevalence. Over 6000 people with a response rate of 95 % and a prevalence of raised blood glucose of 5.0 %. The prevalence from history of raised blood sugar was 2.2 %. Less than half (47 %) of the persons with high blood glucose were aware of their status with less than half on treatment. Of those on treatment less than half (30 %) were well controlled. Prevalence of raised fasting blood glucose was more than double that estimated from history, with less than half of the people aware of their status and of those on treatment nearly half are under good control. The underestimation of the disease through history supports the WHO rule of halves and calls for the use of biochemical tests when estimating prevalence of diabetes in the general population or at least doubling the rate from history alone.


Diabetes mellitus, which was insignificant in its prevalence in early 1900 (Cook 1901) and was further estimated to be insignificant till the 1960s, is rapidly increasing. The global diabetes disease burden increased from 153 million in 1980 to 347 million in 2008. If the trend continues, by 2030, the projected number of people with diabetes will be 552 million which is nearly 1 in 10 adults (Unwin et al. 2011).

In 2010, 12.1 million people were estimated to be living with diabetes in Sub-Saharan Africa and this was extrapolated to increase to 23.9 million by 2030 (Sicre et al. 2009). The prevalence rates range from 2.5–8 % in the rural and urban communities of West Africa and 1–12 % in the rural and urban communities in East Africa (Mohan et al. 2013). Various studies on trends of diabetes mellitus in the African region demonstrate that there is a dramatic increase in the prevalence affecting both rural and urban communities as well as men and women equally, as in Cameroun whereby type 2 diabetes increased tenfold between 1994 and 2003 (Hanley 2007; Mufunda et al. 2006; Hall et al. 2011; Usman et al. 2006).

The increase in diabetes is proposed to be the result of the demographic change (increase in older age group), urbanization, and associated changes in risk factors such as alcohol use, tobacco smoking, obesity and physical inactivity (Kaufman et al. 1999).

The raised fasting blood glucose (RFBG) level and the coexistence with other non-communicable diseases and related syndromes predispose one to the vital organ damage (Codreanu et al. 2012). It is estimated that 50 % of people with diabetes die of cardiovascular disease; heart disease and stroke being the most common whereas about 2 % of people become blind and roughly 10 % develop severe visual impairment after 15 years of living with uncontrolled diabetes. Among people with diabetes the overall risk of dying is at least double the risk of their peers without diabetes (Mills 2015). The African Region has the highest proportion of undiagnosed diabetes mellitus (78 %). (Unwin et al. 2011).

The rule of halves was first proposed for non-insulin dependent diabetes in the USA in 1947 (Wilkerson and Krall 1947) and confirmed in the UK in 1964 (Sharp et al. 1964). The rule states that approximately half of most common chronic disorder is not detected, and half of those detected are not treated, and that half of those treated are not controlled. The rule of halves still holds true for diabetes (Hayes and Harries 1984) and other chronic diseases validating (Smith et al. 1990; Speight et al. 1983) the WHO rule of halves. Based on these previous studies, it was assumed that the rule of halves could provide a good extrapolation on the actual estimates of the prevalence of diabetes (Hayes and Harries 1984) and related organ damage as gathered from history and being on medication.

The aim of this study was to provide a more accurate estimate of high blood sugar and comparing with estimates from history of the disease and validate the rule of halves for diabetes prevalence in Eritrea.


Study population

A cross sectional population based survey using the WHO STEP 3 method was conducted to determine the prevalence of the common risk factors of the major NCDs was conducted in 2009 in all the six administrative Zobas (Regions) of Eritrea. Persons in the age group 25–74 years participated.

Multistage cluster sampling was applied with the guide of the WHO Stepwise approach for the NCD risk factors surveillance, adapted to the Eritrean context to reach the sample size of 6400. In the first stage from the sampling frame of 2624 clusters/villages, 162 clusters were selected by applying probability proportional to size sampling method. In the second stage 40 households were selected from each of the selected 162 clusters using linear systematic random sampling.

Data collection

Data was collected using Personal Digital Assistants (PDAs) Hewlett Packard IPAQ model no. 110. Once the questionnaire was adapted it was tested before the actual data collection process, finally the questionnaire was uploaded into PDAs to be used by the data collectors as e-questionnaire.

Blood sample collection and biochemical analysis

Blood samples were collected from the eligible participants in accordance with the recommended procedures of the WHO STEPwise approach manual. Prior to the collection of the blood samples participants were requested to fill in an informed consent form. Those participants who consented to participate in STEP 3 were requested to fast for 12 h prior to the blood sample collection. The blood samples were collected fingertips of the consented participants using Accutrend glucose between 0700 and 1000 h, following the overnight fast. The samples were analyzed in the field using the WHO recommended automated machine called Accutrend GCT (Accutrend GCT; Roche Diagnostics GmbH D-66298, Mannheim, Germany) and the results were expressed in mg/dl.Diabetes was classified as: Plasma venous value ≥ 7 mmol/l (equivalent to ≥ 126 mg/dl) or Capillary whole blood value ≥ 6.1 mmol/l (equivalent to ≥ 110 mg/dl).

Data quality assurance

The data collectors and supervisors were recruited from the College of Health Sciences and Ministry of Health which are experienced in data collection and research, physical measurements and collection of blood.

In order to ensure quality and reproducibility of the survey result, clearly defined standard STEPS survey procedures were observed. In addition, random checks by field supervisors, survey coordinator and re-interviewing the important questions from at least 5 % of the respondents by different interviewers were done.

Ethical issue

Ethical permission was obtained from the Ministry of Health, Eritrea Ethics Committee prior to the commencement of the survey. All participants signed informed consent before enrolment.


The overall prevalence of raised fasting blood glucose was 5.0 % (95 % CI 4.0–6.0 %), and was significantly higher in males (8.3 %) than in females, 4.3 % (Table 1). Tigrigna ethnic group showed the highest prevalence of raised fasting blood glucose (5.5 %), followed by Tigre (2.3 %) while the remaining seven ethnic groups combined was slightly higher (2.8 %) than that of Tigre ethnic group alone (Table 2).
Table 1

Prevalence of raised fasting blood glucose (RFBG) by gender



Prevalence (%)

Confidence intervals at 95 %










 Both sexes




Table 2

Prevalence of raised fasting blood glucose by ethnic group



Prevalence (%)

Confidence intervals at 95 %

Ethnic group













The prevalence of raised fasting blood glucose level increased with age being highest in the age group 55–64 years and the lowest was in the age group 25–34 years. There was a slight decrease of prevalence among the highest age group i.e. 65–74 years (Fig. 1).
Fig. 1

Prevalence of diabetes by age group

Similarly, among the six regions in the country, matching the degree of urbanization; Maekel had the highest prevalence of raised fasting blood glucose (10.2 %) followed by Debub (5.1 %) and Southern Red Sea Zones (3.9 %).

As presented in Table (3), only 2.2 % of the sample had history of raised fasting blood glucose which is less than half of those with raised fasting blood sugar (5.0 %) biochemically. Among those with history of the disease, less than half (16.1 %) were on treatment and less than half (30.0 %) of those on treatment were controlled (Table 3).
Table 3

Knowledge, treatment and control status on RFBG




Confidence intervals at 95 %

Knowledge status on RFBG by history (told to have diabetes)

 Raised blood sugar status in the total sample

  Raised blood sugar status known (raised)




  Raised blood sugar status not known or not raised




Raised fasting blood glucose during the survey

 Raised fasting blood glucose at day of the survey

  RFBG among those who knew their status




  RFBG among those who did not know




Control status of blood sugar among diabetics

 Control status of blood sugar among those known to have diabetes

  Blood glucose controlled




  Blood glucose not controlled




Proportion on treatment

 Among those who knew their status

  Raised blood sugar statues known and on treatment




  Raised blood sugar Status known but not on treatment




Control status of blood glucose level

 Among those who were on treatment

  Raised blood sugar controlled




  Raised blood sugar but NOT controlled





This cross sectional study conducted in Eritrea used the WHO STEPwise tool to determine the prevalence of raised blood sugar confirmed that less than half of the sample knew of their diabetes status (diagnosed), less than half of those diagnosed were on treatment and of those on treatment less than half had their disease well controlled. These findings validate the Rule of halves that was demonstrated in other studies. These findings corroborate findings from other studies like that from the Indian one whereby of the total estimated 16 % of diabetics nationwide, 33 % were diagnosed to have diabetes whereas double (68 %) were undiagnosed for the disease (Ranjit et al. 2011).

To determine the exact burden of NCDs and their risk factors in a country, the golden rule is to conduct a population based biochemical study. However, these studies are usually very expensive resulting in absence of adequate data for action. The lack of adequate data and underestimation of these diseases in low and middle income countries necessitates for the need of assessing a cheaper yet effective system or rule (Hart 1992). Rules, such as the WHO rule of halves, can assist a country to project towards the actual prevalence estimates of diseases at reasonable costs that in turn can be translated into appropriate control strategies. (Hart 1992).

Globally, the prevalence of raised blood sugar is generally higher among the males than the females (Mohan et al. 2013), this is not a homogenous finding. Our study had also revealed that the prevalence of raised fasting blood glucose was almost double in males compared to females. This could possibly be related to the higher prevalence of NCD risk factors in males such as high tobacco and alcohol use, low fruit and vegetables consumption among the (World Health Organization 2010; Donnelly et al. 1997).

Behavioral risk factors such as smoking could increase the risk of impaired fasting glucose by decreasing insulin sensitivity (Hilawe et al. 2013; Nakanishi et al. 2000) and in the same token men generally tend to have lower hepatic sensitivity to insulin (Færch et al. 2010) thus resulting in raised fasting blood glucose. Similarly, other studies have shown that unhealthy diet (Faerch et al. 1905) and physical inactivity (Assah et al. 1901) are associated with impaired glucose tolerance although physical inactivity was not pronounced in our survey.

Similar results were reported, one from Tanzanian where the prevalence was higher among males than the females (McLarty et al. 1989). Yet, the trend has not been consistent with conflicting reports from several other African countries, including Mauritius, Seychelles, South Africa and Zimbabwe the trend of rising fasting blood glucose level is higher in females than in males (Mufunda et al. 2006).

There are nine Ethnic groups in Eritrea. The prevalence of fasting raised blood sugar was more than double in Tigrigna ethnic group compared to Tigre and the others. This could mainly be due to genetic contributing factors as well as the high urbanization level among the Tigrigna tribe that mainly live in the Maekel region where Asmara, the capital city is located. Urbanization is observed to influence the living style including the dietary habit, physical exercise, alcohol and tobacco consumption as well as other NCD risk factors constituting a significant factor to the rise of diabetes as witnessed from other studies conducted in Cameroun (Echouffo-Tcheugui and Kengne 2011) and Durban, South Africa (Levitt et al. 1993).

Ethnicity influencing the propensity for chronic diseases diabetes included especially in a multi-ethnic population. For example non-white ethnicities were associated with higher risk than whites for developing diabetes with South Asians having the highest risk (Rosellaa et al. 2012). Yet, similar high prevalence of diabetes mellitus was observed in Cameroon whereby exposure to urban environment was strongly correlated to raised fasting blood glucose level (Sobngwi et al. 2004). Thus, considering the role of ethnicity and its risk factors, further studies are encouraged to clearly define the predisposition in Eritrea.

In conclusion, it is important to do biochemical tests when conducting estimations of diabetes in a population. However, using the rule of halves, the scarce information that is gathered from various other sources including from history and medication, can be extrapolated to provide reliable proxy estimates to the prevalence rates of chronic diseases. This eventually, could enable decision makers in low resource settings to develop appropriate and effective strategies for the control of diabetes and other NCDs in order to mitigate timely any complications that may arise from damage to the vital organs.


Authors’ contributions

YG supervised the field work and drafted the manuscript. AU conceived the work and coordinated the field work. GM provided oversight and he resources and reviewed the manuscript. AG managed the data and analyzed it for manuscript development. JM conceived the project and supported the implementation and manuscript development. All authors read and approved the final manuscript.

Competing interests

The authors declare that they have no competing interest.

Statement of human and animal rights

All procedures followed were in accordance with the ethical standards of the responsible committee on human experimentation (institutional and national) and with the Helsinki Declaration of 1975, as revised in 2008 (5).

Statement of informed consent

Informed consent was obtained from all patients for being included in the study.

Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (, which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.

Authors’ Affiliations

World Health Oragnization Zambia Office
World Health Organization Eritrea Office
Ministry of Health


  1. Assah FK, Ekelund U, Brage S, Mbanya JC, Wareham NJ (1901) Free-living physical activity energy expenditure is strongly related to glucose intolerance in Cameroonian adults independently of obesity. Diabetes Care 2009(32):367–369. doi: (pmid: 7776)Google Scholar
  2. Codreanu I, Sali V, Gaibu S, Suveica L, Popa S, Perico N, Ene-Iordache B, Carminati S, Feehally J, Remuzzi G (2012) Clinical study prevalence of hypertension and diabetes and coexistence of chronic kidney disease and cardiovascular risk in the population of the Republic of Moldova. Int J Hypertens 2012(951734):8. doi: Google Scholar
  3. Cook AR (1901) Notes on the diseases met with in Uganda, central Africa. J Trop Med 4:175–178Google Scholar
  4. Donnelly R, Molyneaux L, McGill M, Yue DK (1997) Detection and treatment of hypertension in patients with non-insulin-dependent diabetes mellitus: does the ‘rule of halves’ apply to a diabetic population? Diabetes Res Clin Pract 37(1):35–40View ArticleGoogle Scholar
  5. Echouffo-Tcheugui JB, Kengne AP (2011) Chronic non-communicable diseases in Cameroon—burden, determinants and current policies. Glob Health 7:44View ArticleGoogle Scholar
  6. Faerch K, Vaag A, Holst JJ, Hansen T, Jørgensen T, Borch-Johnsen K (1905) Natural history of insulin sensitivity and insulin secretion in the progression from normal glucose tolerance to impaired fasting glycemia and impaired glucose tolerance: the Inter99 study. Diabetes Care 2009(32):439–444. doi: (pmid: 6613)Google Scholar
  7. Færch K, Borch-Johnsen K, Vaag A, Jørgensen T, Witte DR (2010) Sex differences in glucose levels: a consequence of physiology or methodological convenience? The Inter99 study. Diabetologica 53:858–865View ArticleGoogle Scholar
  8. Hall V, Thomsen RW, Henriksen O, Lohse N (2011) Diabetes in Sub Saharan Africa, 1999–2011: epidemiology and public health implications, a systematic review. BMC Public Health 11:564View ArticleGoogle Scholar
  9. Hanley AJ (2007) The effect of family History of Diabetes on risk factor for developing Gestational Diabetes. Clin Endocrinol 67:754–760View ArticleGoogle Scholar
  10. Hart JJT (1992) Rule of halves: implications of increasing diagnosis and reducing dropout for future workload and prescribing costs in primary care. Br J Gen Pract 42:116–119Google Scholar
  11. Hayes TM, Harries J (1984) Randomized controlled trial of routine hospital clinic care versus routine general practice care for tpe II diabetes. BMJ 289:728–730View ArticleGoogle Scholar
  12. Hilawe EH, Yatsuya H, Kawaguchi L, Aoyama A (2013) Differences by sex in the prevalence of diabetes mellitus, impaired fasting glycaemia and impaired glucose tolerance in sub-Saharan Africa: a systematic review and meta-analysis. Bull World Health Organ 91:671–682DView ArticleGoogle Scholar
  13. Kaufman JS, Owoaje EE, Rotimi CN, Cooper RS (1999) Blood pressure change in Africa: case study from Nigeria. Hum Biol 71(4):641–657Google Scholar
  14. Levitt NS, Katzenellenbogen JM, Bradshaw D, Hoffman MN, Bonnici F (1993) The prevalence and identification of risk factors for NIDDM in urban Africans in Cape Town, South Africa. Diabetes Care 16(4):601–607View ArticleGoogle Scholar
  15. McLarty DG, Swai AB, Kitange HM et al (1989) Prevalence of diabetes and impaired glucose tolerance in rural Tanzania. Lancet 1(8643):871–875View ArticleGoogle Scholar
  16. Mills R (2015) Rule of Halves, Sernova Corp. TSX:V-SVAGoogle Scholar
  17. Mohan V, Seedat YK, Pradeepa R (2013) Rising burden of diabetes and hypertension in SE Asia and African region. Int J Hypertens 2013:14Google Scholar
  18. Mufunda J, Chatora R, Ndambakuwa Y, Nyarango P, Kosia A, Chifamba J, Filipe A, Usman A, Sparks VH (2006a) Emerging non-communicable disease epidemic in Africa: preventive measures from the WHO regional office for Africa. Ethn Dis 16(2):521–526Google Scholar
  19. Mufunda J, Chatora R, Ndambakuwa Y, Nyarango P, Chifamba J, Kosia A, Sparks HV (2006b) Prevalence of non-communicable diseases in Zimbabwe results from analysis of data from the national central registry and urban survey. Ethn Dis 16:718–722Google Scholar
  20. Nakanishi N, Nakamura K, Matsuo Y, Suzuki K, Tatara K (2000) Cigarette smoking and risk for impaired fasting glucose and type 2 diabetes in middle-aged Japanese men. Ann Intern Med 133:183–191View ArticleGoogle Scholar
  21. Ranjit I et al (2011) Importance of controlling diabetes early....the concept of metabolic memory, legacy effect and the case for insulinisation. J Assoc Indian Physicians 59(suppl):8–12Google Scholar
  22. Rosellaa LC, Mustard CA, Stukel TA, Corey P, Huxb J, Roos L, Manuel DG (2012) The role of ethnicity in predicting diabetes risk at the population level. Ethn Health 17(4):419–437View ArticleGoogle Scholar
  23. Sharp CL, Butterfield WJH, Keen H (1964) He Bedford survey. Proc R Soc Med 57:193–204Google Scholar
  24. Sicre R, Shaw J, Zimmet (2009) Diabetes atlas, IDF. The global burden: diabetes and impaired glucose tolerance, vol 4. International Diabetes Federation, BrusselsGoogle Scholar
  25. Smith WCS, Lee AJ, Crombie IK, Tunstall-Pedoe H (1990) Control of blood pressure in Scotland: he rule of halves. BMJ 300:981–983View ArticleGoogle Scholar
  26. Sobngwi E, Mbanya JC, Unwin NC, Porcher R, Kengne AP, Fezeu L, Minkoulou EM, Tournoux C, Gautier JF, Aspray TJ et al (2004) Exposure over the life course to an urban environment and its relation with obesity, diabetes, and hypertension in rural and urban Cameroon. Int J Epidemiol 33(4):769–776View ArticleGoogle Scholar
  27. Speight ANP, Lee DA, Hey EN (1983) Under diagnosis and under treatment of Asthma in childhood. BMJ 286:1253–1256View ArticleGoogle Scholar
  28. Unwin N, Whiting D, Guariguata L, Ghyoot G, Gan D (eds) (2011) Diabetes atlas, 5th edn. International Diabetes Federation, BrusselsGoogle Scholar
  29. Usman A, Mebrahtu G, Mufunda J, Nyarang’o P, Hagos G, Kosia A, Ghebrat Y, Mosazghi A, Atanga SJ, Equbamichael MM (2006) Prevalence of non-communicable disease risk factors in Eritrea. Ethn Dis 16:542–546Google Scholar
  30. World Health Organization (2010) Eritrea STEPS on Noncommunicable disease risk factors survey 2010Google Scholar
  31. Wilkerson HL, Krall LP (1947) Diabetes in a New England town. JAMA 135:209–246View ArticleGoogle Scholar


© Mufunda et al. 2015