whoari: World Health Organization study on acute respiratory...

In pbreheny/hdrm: High-Dimensional Regression Modeling

World Health Organization study on acute respiratory illnesses

Description

The presentation of an acutely ill young infant presents health workers, especially those in developing countries, with a very difficult problem. Serious infections are the main cause of morbidity and mortality in infants under 3 months of age in these countries, and diagnosing the severing of the illness is rather difficult.

To study this problem, the World Health Organization (WHO) collected data on a number of readily accessible variables such as vital signs, family history, and clinical observations resulting from physical examination. The patients' disease status was later determined based on the course of the disease and various laboratory tests. The goal of the study was to develop a early prediction rule for grading the severity of the disease so that timely treatment could be delivered (and costly but unnecessary treatments avoided).

The WHO study looked at several acute respiratory illnesses in several countries. This data set contains data on pneumonia from the country Ethiopia.

Format

• y: A pneumonia score, measured on the following scale: 1=No disease, 2=Cold/cough, 3=Pneumonia, 4=Severe pneumonia, 5=Life-threatening illness.

• X: Each row of X represents an individual.

  • The following columns are objective clinical measurements:

    • biwt: Birth weight

    • hcir: Head circumference

    • wght: Weight (in grams)

    • lgth: Length (in centimeters)

    • temp: Temperature (in Celsius)

    • hrat: Heart rate

    • age: Age (in days)

    • rr: Adjusted respiratory rate

    • waz: Weight-for-age (z-score)

  • There are also a number of clinical observations, judged on a severity scale with higher numbers denoting increased severity. For example, 0-3 would represent absent/mild/moderate/severe. Below are some of those variables:

    • hvo: Vomit more (0-1)

    • bat: Birth preterm (0-1)

    • hbr: Engorged breasts (0-1)

    • inc: Respiratory distress (0-2)

    • sr1: Respiratory state (0-3)

    • sr2: Respiratory state (0-3)

    • str: Stridor (0-2)

    • att: Attentive (0-2)

    • skr: Skin rash (0-2)

    • hyp: Hypotonia (0-3)

    • nut: Nutritional status score (0-3)

    • oab: Sclerema (0-1)

  • The rest of the variables are described in this table. In the table, hx stands for history, ausc for auscultation, and hxprob for history of problems. The author made an effort to cluster relevant variables together; some variables are listed twice as a result.

Dimensions

• n = 816

• p = 66

References

I obtained the data from http://biostat.mc.vanderbilt.edu/wiki/Main/DataSets. The original reference is:

Harrell F, Margolis P, Gove S, Mason K, Mulholland E, Lehmann D, Muhe L, Gatchalian S and Eichenwald H (1998). Development of a clinical prediction model for an ordinal outcome: the World Health Organization multicentre study of clinical signs and etiological agents of pneumonia, sepsis and meningitis in young infants. Statistics in Medicine, 17: 909-944.

pbreheny/hdrm documentation built on Jan. 17, 2024, 8:53 p.m.