No abstract

No abstract

We revisit Kemeny's constant in the context of Web navigation, also known as "surfing". We generalize the constant, derive upper and lower bounds on it, and give it a novel interpretation in terms of the number of links a random surfer will follow to reach his final destination.

No abstract

Severe anaemia in pregnancy is an important preventable cause of maternal and perinatal morbidity and mortality. Different methods of screening for severe anaemia in pregnancy were evaluated in a 2-phased study conducted in Kilifi, Kenya. In phase 1 (in 1994/95), pallor testing was evaluated alone and in addition to raised respiratory/pulse rates: 1787 pregnant women were examined by one of 2 midwives. Sensitivities for detecting severe anaemia (haemoglobin < 7 g/dL) were 62% and 69% and specificities 87% and 77%, respectively for each of the midwives. Addition of high pulse rate increased sensitivity to 77% and 81%, but specificity reduced to 60% and 51%, respectively. In phase 2, following qualitative in-depth work, a screening questionnaire was developed. An algorithm based on screening questions had 80% sensitivity and 40% specificity. Midwife pallor-assessment was conducted following the screening questionnaire. In this phase (conducted in 1997), the midwife performed very highly in detecting severe anaemia, achieving sensitivity of 84% and specificity of 92%. Spending a few minutes asking women questions may have improved the ability to interpret pallor findings. This study demonstrates the value of pallor testing and raises alternative approaches to improving it.