Brian Altonen, MPH, MS

A WORK IN PROGRESS.

This study of population pyramids demonstrated a certain amount of regularity and predictibility of population health age-gender statistics in relation to health care behaviors, practices and needs.  There is a certain repetitive nature of outcomes that at first made me wonder if there were some inherent and hidden errors in this approach to the analyses.  However, it has already been shown that populations, especially large populations, demonstrate a regression to the mean statistical behavior.  The more people you study, the more likely you are of finding the true curve in these behaviors, to which expectations and average findings can be assigned.

This regression to the means also related to how people behave regarding their health.  Be these activities related to intervention activities such as exercise or dietary change, or deciding to begin a new form of cancer therapy or prescription drug program, there is an age-gender relationship that exists for this decision making process.  Across various disease types, organ systems, needs and demands regarding personal health and hygiene practices, human behavior tends to become more and more typical and expected as the population being surveyed or monitored for health is increased.  This also means that even though sub-population A has one type of chronic, lifelong condition requiring medication group I to engage in the appropriate treatment program, and population B group II medicines to accomplish the same, the behaviors these two different groups with two different chronic lifelong diseases display can often be surprisingly similar.  For this reason, analyses of renal chronic diseases can often mimic results and findings for diabetes or hypertension disease intervention programs.  No matter what the medicine or intervention activity required for preventive health purposes may be, human behavior can often dictate whether or not we chose to perform that recommended preventive health activity.  Be the medicine related to blood pressure levels, reduction in anxiety rates, treatment of chronic pain, people in certain age groups often behave the same across very different levels of ICD types.  This results in a recurring Peak Age for prescription drug use when these drugs appear to work on different populations with similar, not identical, age-gender distributions of the diseases under review.   Due to this recurring pattern in human prescription drug and intervention activities behavior, a series of rules could be defined relating these behaviors to the age-gender distribution patterns.  Together these rules imply there is a centralized theme that may be followed when designing intervention activities and programs–a “golden rule” as to who to engage in these programs due to their degree of involvement and numbers of people engaged in this form of health behavior.

The Golden Rule

The Golden Rule is use to determine whether to use Peak Age or Peak Prevalence to design your programs.

Peak age is use when the focus is on numbers of people and costs.  This methodology serves to monitor cost-related concerns, not risk related risk-related concerns.  It is used when supply and demand are the primarily concern, such as answering the questions

• Do we need to develop intervention programs for 35,000, 50,000 or 100,ooo people?
• Which age range do we want to base our intervention plans upon–the 2-year plan targeting 100,000 people, the 5-year plan targeting 50,000 people, or the 10-year plan targeting 35,000 people?
• Which one is expected to have better results?
• Which age range of patients tends to follow-up on our intervention procedures the best?

The Incidence/Prevalence approach delves into the same public health concern, but is focused on such things as

• How do we prevent the higher costs that might be incurred later in life?
• At what age do our patients produce the highest costs?
• At what age range could prevention begin with the goal of reducing those costs, preferably ten years before the demand for care reaches its highest peak?

Using smoking as an example, we use incidence to design plans devoted to reducing teenage tobacco use, but more cost-centered peak age methodology to reduced the costs incurred due to health care requirements during the later years.  Individuals who pass  the peak age for their age-gender and even ethnicity group are going to cost the system many times more than the same people ten years before.  Even though we try to intervene in these younger groups, these programs are rarely that effective costwise.  As a result, reactive care is engaged in rather than preventive care.  This same logic and methodology may not have the same effect with another dual peak age medical condition, such as asthma.  For asthma, both age peaks have definite programs that can be established in order to reduce future costs for Chronic Asthma, Bronchitis, COPD and other diseases more likely to ensue in chronic disease cases.  In this case, both incidence/prevalence and cost are being managed by using this method of designing your intervention programs.

The “Golden Rule” Peak Ages are when the maximum ‘n‘ is/are seen for a particular ICD, by gender.  Chronic diseases often have age peaks that are slightly to moderately different between genders for many of the chronic diseases.  Old age degenerative diseases (many to most diagnoses related to heart and kidney ICDs, age related neurological degeneration, etc)  are more equally dispersed between genders, although greater numbers of women usually exist due to the greater lifespan of women.

Peak Ages are related to cost and incidence/prevalence in the following ways:

1. Peak Age is when the largest number of people with a particular disease are still alive and in need of health care
2. Once peak age is passed, the numbers of people with the disease begin to reduce at increasing rates, which are  due to the disease itself or to co-morbidities and other events in life
3. Cost is related to peak age in that as age increases following peak age, cost increases as well, typically exponentially, even though there is an decrease in the numbers of individuals with the particular malady due to mortality.  Therefore, increasing costs can still be attached to a particular ICD in spite of increased mortality; these costs are typically due to comorbidities and worsening physiological states.
4. There is a relationship between peak age and peak 1-year incidence prevalence rates that provides us with important insights into understanding the disease,  its impact and progression in life, and the related human behaviors.  It provides us this information in 1-year increments, not 5-year age groups.  It is specific an often needs minimal age adjustments in total N is high enough for comparisons with a base population.
   
5. When peak age is much greater than peak incidence/prevalence age, this normally means there is a delay in  the interventions taking place.  Ideal intervention-peak age relationships should demonstrate nearly equal peak ages in terms of incidence/prevalence and raw case counts.
6. Peak age for disease and any intervention related activity should in theory be equal, although the two generally are not.  Invention activities such as exercise and dietary change never demonstrate a direct correlation with age of diagnosis and presence of the disease.  Interventions related to prescription drug use tend to be delayed compared with diagnosis ages, meaning that individual tend to go on drugless therapy for a short awhile, such as delaying the use of anti-diabetic medications in spite of diabetes diagnosis or delaying the use of anti-hypertensive drugs due to feared side effects.
7. Peak age for a disease and peak age for prescription drug use related to that disease therefore can also be different, for many of the same intervention behavior related reasons.
8. Peak age is useful in determining the age ranges for designing particular intervention programs that are targeted more specifically to special age groups, for example STD concerns impact genders quite differently, with boys and men more in need of interventions than women for some diseases and vice versa.
9. Diseases can have peak ages that greatly differ between genders.  Some diseases demonstrate early onset in men versus women, with women experiencing peak incidence and peak ages much later in life.  Schizophrenia for example peaks in men when they are in their twenties, peaking much later at in smaller amounts for women.
10. Diseases can have peak ages that present in bimodal or polymodal fashion, meaning that 2 or more peaks exist for which 2 or more intervention programs can be designed.  For example, asthma peak ages demonstrate a double peak–childhood and 40 yo adults.  Two programs can be designed for these groups.
11. Peak ages can be related to multiple cause and effect relationships.  A syndrome with four peak ages may have four completely different reasons for diagnosis and documentation.
12. Peak age methodology is very applicable to studying and planning interventions for exceptionally rare cases or conditions with some form of preventable behaviors or activities associated with them.

It is up to the researcher of population pyramid approach to apply this method for a better understanding of how to design intervention programs.  This methodology is as applicable to treating very common conditions like diabetes and asthma, as it is to treating and preventing rare behavioral or physical medical conditions such as determining peak suicide risk age groups for specific forms of suicide, specific types of population at risk for infibulation, peak ages and genders associated with specific forms of recreational activities induced accidents, specific forms of rare childhood diseases in need of detailed screening programs enhancement involving children in the right age group.

Notes

A WORK IN PROGRESS.