QUALITY / TOOLS: QUANTITATIVE   (Q/TQ)

Statistics and Epidemiology


Statistics: "A branch of mathematics dealing with the collection, analysis, interpretations, and presentation of masses of numerical data" ("Webster's New Collegiate Dictionary").

Descriptive statistics are measures used to describe a data set: e.g. mean median, mode, standard deviation.
Inferential statistics is the analysis of data for the purpose of  deducing relationships, predictions, etc.

Expected Value and Decision Trees are managerial tools for asssessing options.

Data: historically, a plural noun. Although so commonly (mis)used as a singular noun that such usage is becoming acceptable in standard English, the most correct use is in its plural form and to use Datum as the singular form.


Epidemiology: "The study of the patterns of determinants and antecedents of disease in human populations. Epidemiology utilizes biology, clinical medicine, and statistics in an effort to understand the etiology (causes) of illness and/or disease. The ultimate goal of the epidemiologist is not merely to identify underlying causes of a disease but to apply findings to disease prevention and health promotion." (AcademyHealth, "Glossary").


See the "epidemiological and statistical terms" section of the AcademyHealth "Glossary" for distinctions between:
    sensitivity and specificity
    type I and type II error
    reliability and validity
    association and causality


Trust is a primary factor in statistics and their analyses. Discretion in selecting and defining categories of data, complexities of collecting the data, and technical sophistication required for analyzing the data all place the direct assessment of the quality of the data and analysis beyond the capacity of most users of the data.
Thus trust in the sources is a primary determinant of the understanding and use of data and their interpretation. Similarly, accuracy and fairness in the development, presentation, and analysis of data are primary determinants of future trust. Trust is an essential concern of both private and public data sources. It is a major determinant of the efficiency and fairness of the markets and of the capacity and success of public policy processes.


Common sense is at least as important as quantitative sophistication in understanding and using data. A short book that makes this point and helps develop sensitivity to some basic issues is:

"Educated Guesses
Making Policy about Medical Screening Tests"
Louise B. Russell
University of California Press (1994)

Russell challenges common presumptions about the value of  various screening tests (e.g. pap smear, PSA, cholesterol level), primarily by raising questions related to opportunity costs (i.e. what is lost by taking the suggested actions), to false positives and false negatives produced by the screens, and to how important the results are to treatments decisions and outcomes. "Patients, clinicians, and payers need to recognize the extent to which the guidelines gloss over or ignore considerations of potentially great importance to them. For patients, the questions have to do with whether a screening test is the best way to spend time, emotional energy, and money to preserve or improve personal health. For doctors and their professional associations, the questions center on the most productive way to spend the ten or fifteen minutes allotted to each patient's appointment and, of course, the impact of the answers on their professional lives. For payers, the issues have to do with how best to spend employers' or taxpayers' money to improve -- of even whether the money would be better spent in alternative ways."

These concerns reflect earlier work by Russell, published as:

"Is Prevention Better Than Cure?"
Louise B. Russell
Brookings Institution (1986)

Russell demonstrates that many claims of financial gains through disease prevention do not prove out under careful analysis. For example, the savings in health care treatments for the person saved from a disease may far exceed the costs of preventing the disease in that person. But the total costs of a prevention program may not exceed the reduced costs of treatments, because most of costs of prevention are incurred for persons who will not contract the disease in any case. She points out that this is not necessarily a reason to drop prevention programs, because even if the costs of prevention exceed their savings, the net costs may be worthwhile for the improvement in length and quality of life.