Is Population Health a Reasonable Medical Problem? Historical and Geographic Variation in Disease, Mortality and Infrastructure, Part II: Factors in Current World Health

Is Population Health a Reasonable Medical Problem? Historical and Geographic Variation in Disease, Mortality and Infrastructure, Part II:  Factors in Current World Health

The 194 WHO member states can be analyzed for health/disease precursors and health/disease outcomes, employing data from “World Health Statistics 2013”. Using the measure of gross national income per capita as a general indicator of development and resource availability, it is possible to relate some state attributes to the ultimate bottom line in all matters of health: Mortality. This data-set contains eight separate mortality variables, plus country-by-country male and female life expectancies.

Between potential financial resources (suggested by ‘gross national income per capita’, a measure that does not account for inequality) and the several mortality rates are possible mediating variables, some of which are used for this analysis. These include measures of health infrastructure: safe water, improved sanitation (sewage disposal), measures of childhood (under)nourishment, and government expenditures targeted specifically at health issues. Vaccination rates are also available. The effects of vaccination on mortality rates will be explored.


Zero Order Relations among Population Income, Physician Density, Sanitation and Mortality


Income Group Physician Density % Pop. Using Improved Drinking Water Sources % Pop. Using Improved Sanitation Age-Standard-ized Mortality Rates from Communi-cable Causes Mortality Children <5 years: Diarrhea Mortality Children <5 years: Pneumonia Mortality Children <5 years: Malaria
Low Income 5.1 67 37 636 12 18 11
Lower Middle 7.8 87 47 233 11 19 6
Upper Middle 17.8 93 74 125 4 14 0
High Income 27.1 99 100 31 1 4 1

Source: World Health Statistics 2013 [1], World Health Organization, multiple tables passim.

The table above gives four income clusters among the 194 member states of the World Health Organization, and provides a selection of numerical representations of prevailing conditions relevant to health and mortality. The table shows the generally expected relationships. And according to reasonable expectations, as national per capita income increases, physician density and (age-adjusted) mortality declines. [2] Further, the health related infrastructure is also increasing along with male and female life expectancy (not shown), while mortality declines. (Also not shown: Vaccination rates tend to increase with national income and physician density.)
These expected findings, while suggestive, do not really answer the questions posed in the opening paragraph in Part I of this writing. In sum, those questions asked if medicine could have a positive effect on public health. The surface impressions from such tables as the one above would seem to answer in the affirmative, but there is an underlying story. We shall find that the relations between increasing ‘physician density’ and lowered mortality as shown in the table, are precisely spurious. The same will be shown of vaccination rates. Sanitary infrastructure and nutrition drive decreased pathogen borne mortality.



A starting note: Analysis of this WHO dataset provides a view of the relationships between each of the three variablesfly ‘gross national income per capita’, ‘per capita spending on health’ and ‘government spending on health per capita’ and each of five different measures of overall and childhood mortality. Controlling each of those relationships for ‘physician density’ tends uniformly to reduce significantly or reverse the ameliorative effects of that spending. Too much money goes to create the physician infrastructure and less goes to the complex net comprising sanitary infrastructure. Increasing the numbers of physicians in a population may not be a positive resource for population health. On the other hand, targeted spending increasing the number of nurses and midwives would appear from this analysis to offer better mortality outcomes, perhaps especially in lowering infant mortality and under 5 years mortality. [4] ‘Physician density’ is explored further immediately below.


‘Physician Density’ and Mortality Rates (controlling for sanitation, drinking water and nourishment)
Increasing physician density is part of the whole process of history that is labeled development; there is an increasing division of labor in which service sector professional jobs increase in proportion to all jobs. Similarly, the numbers and proportions of lawyers, architects, social workers and security guards have increased. The increasing proportions of each of the non-medical professions would be positively correlated with decreasing mortality; clearly, those correlations are spurious. We must look outside the medical frame (any presumed medical effects) to understand the epidemiological transition of the past, and that currently occurring in developing countries. As development continues, there also occurs a concomitant increase in chronic, non-communicable diseases. The medical profession watched the decline of pathogen-borne diseases and then it watched the increase in diseases of abundance (cancer, diabetes, etc.). It watched both the demographic transition and the parallel epidemiological transition. The profession was not causal in either.

The sanitation effects. As a reminder, in this data, zero order correlations (no controls) show strong associations between ‘physician density’ and various measures of diminished mortality. Put simply, as physician density increases, mortality decreases (a ‘negative’ association). However, when there is a control variable for improved sanitation, the strength of the associations increase by an average of about 15%, but the sign of the associations changes to positive. This demonstrates a major effect from improved sanitation on mortality. The measures affected by the control were: overall ‘mortality rate’, the ‘noncommunicable disease mortality rate’, the ‘communicable disease mortality rate’, the ‘infant mortality rate’, ‘the less than 5 years mortality rate’ and the ‘less than 5 years malaria mortality’. That shift represents six of the eight mortality measures. That is, overall, with this ‘improved sanitation’ control, increasing physician density is strongly associated with increased mortality. Apparently, the society that manages its sewage in a sanitary fashion is less likely to die at a given age, and may be controlling mosquito breeding areas better.

The nourishment effects. In controlling for nutritional conditions, the strength of the associations between physician density and diminished mortality declined significantly in five out of eight mortality measures. There is also a significant decline in associational strength for ‘female life expectancy’. A decline in associational strength demonstrates the dependence of the original association on the control variable. Further, after the control procedure, the coefficient sign changed to plus for ‘noncommunicable disease mortality’ and ‘less than 5 years pneumonia mortality rate’; that is, because of the control, as ‘physician density’ increases these two mortality rates also increase. The mortality rate not affected by this control procedure was ‘less than 5 years diarrheal mortality’. This makes sense.

The improved drinking water effects. Improved drinking water has an effect separate from societal financial resources and physician density. Controlling for improved drinking water lowers the disease mortality burden on a population (especially, ‘less than 5 years malarial mortality rate’) and actually has sufficient impact to reverse the signs of the associations between ‘physician density’ and overall ‘mortality’ and ‘communicable disease mortality’.

Vaccination Rates and Mortality Rates
(controlling for sanitation, drinking water and nourishment)

Of the vaccination rates, four vaccines are listed across all 194 WHO member states (Measles, DTP3, HepB3, and Hib3), only the Measles and the DTP3 vaccines have a separable effect on any mortality rates. By comparison, in the U.S. population, among all the vaccines listed in VAERS, these four (plus the pneumococcal vaccines) are linked with the highest childhood vaccine-associated mortality rates.

Increases in vaccination rates have a set of zero order correlations that link higher vaccination rates with lower mortality rates, for all mortality measures. Higher vaccination rates are also associated with increasing male and female life expectancies. All of the coefficients are strong, .7 or higher. These are the expected findings: vaccination saves us.

But what is the result once the above accepted and expected associations are subjected to infrastructural controls? We shall see that the power of vaccines to reduce mortality is only apparent. In fact, the association is spurious.

Improved drinking water effects. When ‘improved drinking water’ is used as a control, the strong associations between vaccination rates and diminished mortality weaken, go toward a correlation of zero, or reverse in sign. The signs are reversed for both the measles vaccination rates and the DTP3 vaccination rates for overall ‘mortality’ and for ‘communicable disease mortality’. Only ‘infant mortality’ and ‘less than 5 years mortality’ remain statistically the same. These two variables are very strongly correlated with ‘infant mortality’ being the dominant factor in ‘less than 5 years mortality’ when regressed against either ‘male life expectancy’ or ‘female life expectancy’. Speculatively, it may be that breast feeding leave infants out of the ‘improved drinking water’ effect.

Improved sanitation effects: Measles vaccine. ‘Improved sanitation’ is applied as a control variable in the correlation between measles vaccination rates and mortality. This control causes significant declines in coefficient strength to occur in ‘infant mortality rate’, ‘less than 5 years mortality’ and the ‘less than 5 years malarial mortality rate’. The ‘improved sanitation’ control variable also causes the strength of the associations to decline to near zero for overall ‘mortality’, and ‘communicable disease mortality’. The coefficients decline and approximate zero for both male and female life expectancies. The associations change signs to positive for ‘noncommunicable disease mortality’ and for ‘less than 5 years pneumonia mortality’. That is, for these two measures, increasing mortality is associated with increases in vaccination rates. The association stays unchanged in strength and sign for ‘less than 5 years diarrheal mortality’.

Improved sanitation effects: DPT3 vaccine. For the DPT3 vaccination rates where ‘improved sanitation’ is used as a control, the signs of the associations switch to positive with overall ‘mortality’, ‘noncommunicable disease mortality’, ‘communicable disease mortality’, and ‘less than 5 years mortality. This indicates that as DPT3 vaccination rates increase, mortality in these three areas also increases. The coefficients for all other mortality measures decline to approximate zero, except ‘less than 5 diarrheal mortality’ which remains unchanged in strength and sign.

The nourishment effects: Measles vaccine. Measles vaccination rates when controlled by an indicator of poor nourishment switch the coefficient signs to positive for the coefficients of ‘noncommunicable disease mortality’, ‘communicable disease mortality’ ‘less than 5 years diarrheal mortality’, and ‘less than 5 years pneumonia mortality’. This make sense given that the measles vaccine is known to cause pneumonia and bowel infections. The signs of other measures of mortality remain negative but the strengths of their coefficients decline significantly by 50% or more.

The nourishment effects: DTP3 vaccine. Using an indicator of poor nourishment as a control for the measurement of the relationship between DPT3 vaccination rate and mortality causes all signs to switch to positive while retaining robust coefficient strength. Thus, in extant conditions of malnourishment, increasing the DPT3 vaccination rate switches all eight indicators of mortality to positive signs. That is, the higher vaccination rates lead to higher death rates, as measured by overall ‘mortality’, ‘noncommunicable disease mortality’, ‘communicable disease mortality’, ‘infant mortality’, ‘less than 5 years mortality’, ‘less than 5 years diarrheal mortality’, ‘less than 5 years pneumonia mortality’ and ‘less than 5 malarial mortality’. Furthermore, the uncontrolled positive relationships between the DPT3 vaccination rates and both male and female life expectancies decline to approximate zero. This vaccine seems to act like the straw that broke the camel’s back.

Using the three infrastructural health measures. When all three control variables are applied simultaneously, the result shows that increasing vaccination rates (measles, DTP3) leads to increasing mortality rates in each of the eight measures of mortality; further, vaccine associated deaths are sufficiently powerful to produce a negalight-banner-dirty-festive-mediumtive effect on male and female life expectancy.

Summary of Variables

No controls (zero order relationships only):

-Income per capita

-Health spending

-Physician Density

-Population sanitary and nourishment conditions

– Vaccination rates

As each of these variables increases, all measured mortality declines, and life expectancy increases in strong associations.

Controlling for ‘improved drinking water,’ ‘improved sanitation,’ and ‘(under) nourishment)’:

-Physician density 

-Vaccination rates

As each variable increases, all measured mortality increases, and life expectancy tends to decrease.


Populations bolstered by the availability of good water, proper sewage disposal, and sufficient nourishment expect their children to live. Vaccination does kill children in these wealthier populations, but not as many — or as obviously. The medical profession, with all its cultural authority, actively denies serious vaccine-associated adverse effects, including death. This makes public debate over vaccine safety, effectiveness, and informed consent all but impossible.

Not all populations have equal resistance to vaccine damage and vaccine associated death. Sometimes that is because not all populations receive the same vaccinations in the same numbers. Sometimes it is because the immune systems within the nearly two hundred collectivities of the WHO dataset are not equal. But all populations are now subject to vaccinations. Those populations where the extant conditions assault and weaken their immune systems expect the possibility of childhood death. These populations are also vaccinated at rates in the nearly 80% percent range or higher. Unfortunately, these vaccination regimens add measurably to the death toll of those weakened children. It is probable that the cultural authority claimed by the medical profession, and to which we assent, is not legitimate.

In returning to the questions asked by the first paragraph in Part I of this writing, I still believe there is no evidence that, as a healing force, modern medicine could carry the day against the disease density of populations, historical or current. But further, through widespread vaccination, it appears that medicine can negatively affect population mortality. Medicine is a treatment profession working one patient’s ills at a time, for better and for worse. The profession gains its greatest reach into a population through vaccination, yet that approach, for all the braggadocio that surrounds it, leaves us weaker.

In building the artificial environment that is the chief product of development, we trade contact with the natural, physical environment for the chance to inhabit a place that is biologically safer, but we can also pollute and poison that humanly made environment. When we discover these mistakes, we can create policies that attempt to limit or stop them. We must consider such an option with the injectable environmental toxins pushed upon us by the misguided cultural authority of the medical profession.

Endnotes and References Cited
[1] This report is a treasure trove of data for cross-national comparisons, and for data comparing effects of safe water, adequate hygiene, vaccination rate effects, etc. Using the data requires considerable work in transcribing it from PDF format into a form usable in the SPSS statistical package, however.
[2] The availability of nursing and midwifery would have a similar relationship as physician density to various mortalities in a revised table that included that category of personnel. There are more doctors and nurses where conditions overall are better, rather than the presence of doctors and nurses improving overall conditions. This direction of relationship is true within the U.S. and around the world.
[3] Take for example a simple example of improved drinking water. The Carter Center gives out pieces of fabric and says, in effect, “Here filter your water with this before you drink it”. As a result, the horrible guinea worm infection goes away from the world, this is not medical science, it is a brilliant common sense gift from a world that has the concept of filtration to a world that does not: In 1986 the guinea worm infection incidence was 3.5 million (new cases per year); by 2015 there were a mere 22 cases of guinea worm infections in the world. The Guinea worm eradication campaign has averted at least 80 million cases of this devastating disease among the world’s poorest and most neglected people since its inception.
[4] I also exchanged the ‘X1’ variables (the three money variables) with the control variable, making ‘physician density’ the new ‘X1’ variable. If ‘physician density’ is used as “X1” and three money variable are used in seriatim calculations, there was no significant change in the strength of the resulting coefficients and no change of sign on any coefficient, as compared to the zero order correlation between ‘physician density’ and all mortality measures. This result is counterintuitive, of course, but it demonstrates on the one hand, the limited reach of medicine into populations and suggests, on the other hand, the disrupting effects by medicine on population health with increasing reach (physician density). We get further hints of this deleterious impact from studies that show the mortality effects of physicians’ strikes. In four of the strikes, mortality declined, in three there was no change. In two of the Israeli strikes, organized morticians intervened to get physicians back to work, the undertakers were losing too much business. “To find out whether the industrial action was affecting deaths in the country, the Jerusalem Post interviewed non-profit making Jewish burial societies, which perform funerals for the vast majority of Israelis. Hananya Shahor, the veteran director of Jerusalem’s Kehilat Yerushalayim burial society said, “The number of funerals we have performed has fallen drastically.” Meir Adler, manager of the Shamgar Funeral Parlour, which buries most other residents of Jerusalem, declared with much more certainty: “There definitely is a connection between the doctors’ sanctions and fewer deaths. We saw the same thing in 1983 when the Israel Medical Association applied sanctions for four and a half months” . See for example, S.A. Cunningham, et al., “Doctors’ strikes and mortality: A review”, Social Science & Medicine, 67, 2008, pp. 1784-1788.