Relation between education and
the amount of healthcare used
Bachelor-thesis June 2018
Els Hovinga
Faculty of Economics and Business
10509526
University of Amsterdam
Bachelor-thesis
Supervisor: Melvin Vooren
Economics
2
Statement of originality
This document is written by Student Els Hovinga who declares to take full responsibility for the
contents of this document.
I declare that the text and the work presented in this document is original and that no sources
other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of
Abstract
Healthcare is the biggest expense in the Netherlands and it keeps rising. Spending a high amount of
money on healthcare is obvious but there is not an infinite supply, so there has to be done
something to limit these costs. Prevention is the most efficient way to do this and this thesis will
focus on the relation between education and the amount of healthcare used, where education can
be seen as a form of prevention.
The relationship between education and health can be understood in the following way;
receiving education gives someone more knowledge and this includes knowledge on a healthy
lifestyle which can be an incentive for people to make healthier choices. Multiple papers show that
more education leads to better health behaviors. With more education people smoke less, drink less
alcohol, have a lower BMI and wear their seatbelts more often. These health behaviors account for
a substantial part of amount of healthcare used, so health behaviors work as an intermediate step
between education and healthcare use. But something to keep in account is the possibility of a
third unobserved variable, for example socioeconomic status, childhood health or ability. Without
this unobserved variable it is possible to observe people with high education and good health or
vice versa even though the education has no causal effect and it is just the third variable that
causes both.
The data used for the analysis is obtained from the Dutch National Health Survey and
provided 9165 respondents. Three linear regressions are performed with the different dependent
variables; the number of contacts with a medical professional per year; the number of
simultaneous present chronic diseases; and the self-reported health score. The independent
variable is education measured in years. Age, gender and migration background have been added
as control variables.
The results from the regressions show that the correlations between education and the
dependent variables are negative, as expected. One additional year of education predicts a .422
decrease in number of contacts per year, a .096 decrease in number of simultaneously present
chronic diseases, and a . 035 decrease in the self-reported health score (which is an improvement).
A limitation of the study is the possibility that I have excluded the people with the lowest
schooling due to indistinctness in the survey. Secondly, the 𝑅𝑅
2is very low, and therefore does the
model not explain much of the variance in the dependent variables. There also is a small problem
with the reliability of the answers of some of the respondents, but overall do the analyses confirm
the hypothesis of the thesis and is the relation between education and the amount of healthcare
used proved. Further research can be focused on using more comprehensive models with additional
variables.
4
Table of contents
Statement of originality ... 2
Abstract ... 3
Table of contents ... 4
Introduction ... 5
Literature review ... 6
Health behavior ... 6
Third variable ... 7
Causality ... 8
Data and methods ... 9
Data ... 9
Variables ... 9
Empirical approach ...13
Other tools ...14
Results ... 15
Descriptives...15
Correlations...15
Coefficients ...15
Discussion and conclusion ... 17
Discussion ...17
Conclusion ...18
References ... 19
Appendix 1 – SPSS-output ... 22
INTRODUCTION
Healthcare is the biggest expense for the Netherlands and it keeps rising. In 2000 the amount
spent on healthcare was 46,5 billion euro (10.37% of the GNP), and in 2016 it had raised to 96,1
billion euro with what it covered almost 14 percentage of the GNP (Centraal Bureau voor de
Statistiek, b; Government, 2015). For comparison, the expenses on education are 23,2 billion
euro in 2000 (5.2% of the GNP) and 42,9 billion (5.6% of the GNP) in 2016 (Centraal Bureau
voor de Statistiek, a).
But it is for obvious reasons that we spend so much on healthcare, health is one of the
most important things in life. According to the psychological theory of Maslow: “Maslow’s
hierarchy of needs”, every individual needs to achieve certain needs and some take precedence
over others where it is necessary to first fulfill the needs on the bottom of the pyramid and work
your way up to the top. After physical needs (food, water and warmth) follows security which
includes health. Only after acquiring this, people move on to the need of love, friendship,
respect, status, accomplishment, freedom and last self-actualization (Maslow, 1943). So, this
confirms the need of spending on healthcare but unfortunately there is not an unlimited stock of
money.
Aside from being the biggest expense in the Netherlands, the healthcare costs are also
the fastest rising expenses of the Dutch government (Government, 2015). The last couple of
years the government has achieved to reduce the growth to a lower percentage than the
economic growth. But still, it is the biggest expense and not expected to narrow down on itself.
There are numerous ways to diminish these costs. One option is to lower the salaries of health
care professionals or to reduce quality along the lines of taking less time for patients. Also, the
choice of health care system can make a difference. A really important way to reduce the costs is
prevention, because these are often small investments with a much greater reduction of costs in
the future. Prevention is even more efficient if you can focus on specific groups who are at high
risk.
This bachelor thesis will focus on the effect of schooling on the amount of health care
used, where schooling can be seen as a form of prevention. The expectation is that more
education will lead to better health and lower healthcare costs. If the hypothesis turns out to be
true, it can be advisable to invest more in education and find ways to lengthen the years of
schooling of certain individuals or groups. The correlation is expected because it is known that a
low social economic status, low income or a background of migration has influence on health
and the amount of health care used. These factors (low social economic status, low income or
migration background) are also linked to a lower level of education in many papers. Our results
will confirm the hypothesis and show a negative correlation between years of education and
healthcare use measured in more than one way.
In the second paragraph, the necessary theoretical background will be presented in the
literature review. After that the paper will continue with the materials and method in which the
acquired data will be described in detail with the accompanied methods used to perform the
analyses in the thesis. Paragraph four shows the results from the regressions and the final
paragraph includes a discussion on the obtained results together with the conclusion.
6
LITERATURE REVIEW
In this thesis I try to determine the relation between education and health. A good start is to
understand the determinants of health and take a look at the model of the demand for health by
Grossman from 1972. In his paper Grossman (1972) develops a model where “health can be
viewed as a durable capital stock that produces an output of healthy time” (Grossman, 1972). A
prediction of the model is that if education increases the efficiency with which gross
investments in health are produced, the more educated people would demand a larger optimal
stock of health. It can be compared with the labour market. The more educated someone is, the
more productive he is in the labour market and likely to also enhance non labour market
activities like their own health. Therefore, you can assume that more educated people are
healthier.
Health behavior
The relation between education and health can also be understood in an easier way. Receiving
more education gives someone more knowledge and this includes knowledge on a healthy
lifestyle which can be an incentive for people to make healthier choices. These choices can be
about numerous of things; time, goods, diet, smoking, activity.
To extend the theories about health behavior, Cutler and Lleras-Muney (2010) have
done an extensive research on the effect of schooling on health behaviors. They examine the
following behaviors: smoking, diet/exercise, alcohol use, illegal drugs, automobile safety,
household safety, preventive care, and care for people with chronic diseases. Their results show
that education has a positive effect on health behaviors but that the impact is greater at higher
levels of education, rather than lower levels of education (Cutler & Lleras-Muney, 2010).
However, they found that after 10 years of education there is an overall linear
relationship between years of education and smoking. The base level of 10 years of education is
not only recognizable in the relation with smoking. After 10 years of education Cutler and
Lleras-Muney (2010) found a significant increase in the percentage of people who always wear
their seatbelt in the car, a significant decline in the number of days per year in which people had
five or more alcoholic drinks and also the number of people who do vigorous activities
decreases fast after 10 years of education. The results of education and BMI are less obvious, but
still present (Cutler & Lleras-Muney, 2010). For use of illegal drugs Cutler and Lleras-Muney
(2010) found that high educated people report that they are more likely to ever try drugs but
seem better at quitting and controlling their consumption in comparison to low educated
people.
Health behaviors are obviously strongly related to health and the amount of healthcare
used. Smoking is reported to remain with 18.1% the main cause of preventable deaths in the US
(Mokdad, Marks, Stroup, & Gerberding, 2004). Poor diet and physical inactivity account for
15.2%, and also the consumption of alcohol and vehicle crashes are in the top list (Mokdad et al.,
2004).
Similar results are reported in the conclusion of the article from Cutler and
Lleras-Muney (2010), what suggests that “every year of education the mortality risk lowers by 0.3
percentage points, or 24 percent, through reduction in risky behaviors” (Cutler & Lleras-Muney,
2010). Other researchers also look into the effect on health behaviors. They describe that
education can be a way to reduce stress levels and thereby lead to better health behaviors
(Brunello, Fort, Schneeweis, & Ebmer, 2016). Brunello, fort, Schneeweis and
Winter-Ebmer (2016) also state that higher educated people are also more likely to have healthier jobs,
live in healthier neighborhoods, and interact with healthier coworkers or friends which will all
induce better health behaviors.
Another way education can have effect on health through behaviors is as a result of the
relation with income. Education is evidently related to more job opportunities and more job
opportunities will in general create more income. In tons of studies, income is reported to cause
better health outcomes (Pickett & Wilkinson, 2015). Healthy food is usually more expensive
than fast food and money gives people the opportunity to make the healthier choice (Kettings,
Sinclair, & Voevodin, 2009). Also, with lower income there is a possibility of financial stress
which can lead to more health problems (Kahn & Pearlin, 2006).
But despite the fact that all these papers confirm the positive correlation between
education and health, there are a lot of reasons why the differences in education cannot be
causally related to the differences in health.
Third variable
Something mentioned in a lot of papers is that the relationship between education and health
can also be due to an unobserved third variable or more unobserved variables(Pickett &
Wilkinson, 2015). Some possibilities for these variables are the following: socioeconomic status,
childhood health, time preference or depreciation of time and ability.
Socioeconomic status, also known as SES, is the social standing of a group or individual
in relation to others including income, education and occupation (Wikipedia, APA). SES is strong
related with both education and health. Some examples show how low SES leads to lower
education; children with low SES have less access to learning materials and experiences,
including books, computers or tutors to create a positive environment (Bradley, Corwyn,
McAdoo, & García Coll, 2001), students with low SES are less likely to have access to
informational resources about college (Brown, Wohn, & Ellison, 2016). And as for the relation
with health, people with a low-SES background have a higher likelihood of being sedentary and
to have a BMI above 30 (Newacheck, Hung, Jane Park, Brindis, & Irwin, 2003). This can be due to
the possible lack of resources, like playgrounds in the neighborhoods or the accessibility of
healthy food (Chen & Paterson, 2006). In contrast to this evidence, Cutler and Lleras-Muney
(2010) reported that in most of the time SES is used as a control variable in studies instead of
specifically study the effect of SES, and when SES actually is studied in the relation with health
outcomes, most of the time only the low/middle SES are examined and not so much the high SES
(Cutler & Lleras-Muney, 2010).
Childhood health influences the health endowments shaped prior to educational
attainment (Arendt, 2005). If children are born with a condition or if they develop one in the
first years of their life, it can give a different look on health and health behavior even before they
start school. This can also be seen as the backward effect of health on education and that doesn’t
stop after the early childhood years. Not being well during the years where most other people
are involved in education can be a reason why to leave school or decide to not participate in
further education.
8
The ability to be healthy can also be explained as time preference or the depreciation of
time. A preference for the future or a low depreciation of time will cause an individual to invest
current costs for benefits in the future; education and health are both activities that fit these
preferences. This preference for the future acquires a certain ability to not succumb for the
“easy” way of choosing for benefits at present time.
If we don’t take this third unobserved variable into account there is a possibility that we
observe people with low education and low heath or high education and good health, even
though the education has no causal effect on health and it is just one of the other variables that
causes both.
The lower likelihood of having health insurance is reported by (Lochner, 2011) as a
possible explanation for the combination of low education and poor health. This is not of
interest for this paper because the analysis is about the Netherlands and this country does not
face the problem of uninsured people because of the healthcare-insurance law. According to this
law everybody is obliged to have health insurance, and the people who can’t pay for the
premium get benefits. Because of this law, only 0.14% of Dutch citizens did not have health
insurance in 2016 (Nederlandse Zorgautoriteit, 2017).
Causality
But aside from the literature about health behaviors and the possible third variable, the
following articles tried to prove the causality of the education effect.
Despite the amount of studies on the correlation between health and education, it
remains difficult to determine the causal relationship and most papers fail to address the
endogeneity of education.
Arendt (2005) showed that education is significantly related to self-reported health and
body mass index, however, with an instrumental variable the results were far from significant.
(Lleras-Muney, 2005) uses an instrument to examine the influence of education on mortality
with the result that more years of schooling reduces mortality after 35 years, but her results
were not significant too. Also, (Adams, 2002) who uses quarters of birth as an instrument to
relate schooling to self-reported health, receives non-significant results. However, (Spasojević,
2010) finds a significant positive relation from education on self-reported health and BMI.
Using self-reported health (SRH) as outcome has both advantages and disadvantages.
The results are subjective which is a disadvantage, but it can be used as a summary of a lot of
health factors together which is harder to measure with objective variables, what makes it an
advantage.
With all discussed literature, the hypothesis for this thesis is that education will have a
positive effect on health and a negative effect on the amount of healthcare used.
DATA AND METHODS
Data
The data used for the regression has been obtained from the database DANS EASY, which is an
institution from KNAW (Royal Dutch Academy of Sciences) and NWO (The Netherlands
Organization for Scientific Research). The original data come from the CBS (Central Bureau for
Statistics) in the Netherlands from the national health survey from the year 2016. The goal of
the national health survey (Dutch: Gezondheidsenquête) is to get an as complete as possible
overview of the (developments in) health, medical contacts, lifestyle and preventive behavior in
the Netherlands. Every year they take a sample of more than 15,000 people to fill in the survey.
The response rate is between 60-65 percent which leads to a net sample of 9165 in the year
2016.
The survey consists of 575 answers of which I have used 50 for the analysis. For most of
the excluded variables, it is obvious why they are excluded because of their irrelevance with
respect to the regressions. For a complete overview of all variables and their exclusion see
appendix 2. I have downloaded the data into SPSS and deleted all those excluded variables.
Before performing the regressions, I have made some other manipulations to the dataset. First, I
put all variables on the right measure (scale, ordinal or nominal) and I have set all data where
had been answered “not relevant”, “don’t know”, or “don’t want to tell” on missing. The final
variables needed for the analyses are the following: Contacts, ChronDiseases, SRHscore,
EducYears, Female, Age, Western and NonWestern. Not all these needed variables were directly
provided by the survey results. The respondents were guided through a flowchart with
consecutive questions and those different answers therefore had to be combined to form the
final variables.
Variables
Contacts gives an interpretation of the number of contacts with a medical professional per year,
in which also the number of days hospitalized are included assuming someone hospitalized has
once a day contact with the doctor. The variable is made with the sum of “number of contacts
with a GP”, “number of contacts with a specialist”, “number of contacts with a psychologist or
psychiatrist”, “number of days hospitalized” and “number of days of day-intake”. To calculate
“number of contacts with a GP” I used the following questions: “When was your last contact with
a GP?” and “How many times the last 4 weeks”. The second question was only asked if the
respondent answered the previous question with “less than 12 months ago”. I manipulated the
data by the following scheme; answering “>12 months” or “never” ends up with a 0, “0 times
the last 4 weeks” with a 6 (these people report to have been to the doctor this year, but not this
month, so the average will be between 1 and 11 which results in 6, and every other number will
be multiplied by 13 because that will be the total of visits per year is you measure with a time
span of 4 weeks. Because of the large number of respondents, we can assume these numbers
taken as means do not create a big bias. Figure 1 shows how the results are modified to their
final value. The blank boxes indicate the questions, the grey boxes the possible answers and in
red the final code in the dataset. These modifications are also done for “number of contacts with
a specialist”, “number of contacts with a psychologist or psychiatrist”, “number of days
hospitalized” and “number of days of day-intake”. The modifications are shown in Figure 1-5.
10
Figure 1 - Modifications to form number of contacts with a GP
Figure 2 - Modifications to form number of contacts with a specialist
Figure 3 - Modifications to form number of contacts with a psychologist or psychiatrist
Figure 5 - Modifications to form number of days of day-intake
Figure 6 - Modifications to form ChronDiseases
The variable ChronDiseases measures the total number of simultaneous present chronic
diseases. I have chosen for this as one of the dependent variables because having a chronic
disease usually leads to a regularity of healthcare consumption and therefore can give a suitable
interpretation of the amount of healthcare used. There is a lot of confusion what the definition
of a chronic disease is. (Bernell & Howard, 2016) try to list the various definitions used by
notable institutions like the World Health Organization (WHO), the Center for Disease Control,
MedicineNet and Wikipedia. The conclusion of Bernell and Howard’s paper is that there is not
one clear definition on chronic diseases, whereas in this paper I have chosen to use the
definition from the WHO: “Noncommunicable diseases (NCDs), also known as chronic diseases, are
not passed from person to person. They are of long duration and generally slow progression. The
four main types of noncommunicable diseases are cardiovascular diseases (like heart attacks and
stroke), cancers, chronic respiratory diseases (such as chronic obstructed pulmonary disease and
12
asthma) and diabetes.” (World Health Organization, 2014) All “long-term” diseases from the
survey fit the definition and are therefore all included. To measure the variable, I have coded all
these 31 questions about the presence of a specific disease 0 if the answer is “no” and 1 if the
answer is “yes”. I have summed up these answers to a total number of synchronically present
chronic diseases. There are two questions for which respondents needed to fill in an
explanation: “What other form of cancer?” and “What other chronic disease?” I have checked all
these answers manually and if they filled in only one condition it was coded a 1, and if they filled
in more than one, this number was added to the final variable. Figure 6 shows the process of
how I computed this variable.
I have added the variable SRHscore to perform a third regression with this as the
dependent variable. The reason for this is the high number of papers in the literature review
with the SRH-score as their dependent variable. This extra regression gives the possibility to
compare the results with the evidence from previous studies better. The self-reported health
score is a single question in the survey where they have to report on a scale from 1 to 5 where 1
is the best health someone could experience and 1 the worst state of health. I have set the
variable on measure scale.
EducYears is the most important variable as it measures education which is my
independent variable. I have chosen the finished level of education over the conducted (but not
finished) level of education. The data that comes from this question is bundled into groups of
different levels of finished level of education with the following values; 1: “primary education”,
2: “VMBO, MBO-1 or the first three years of HAVO”, 3: “HAVO, VWO or MBO”, 4: “bachelor”, 5:
“master or PhD”. Because this is an ordinal variable, I have adjusted this into a scale variable
measuring years of schooling. I have computed the years of schooling for all different levels of
education from data that I obtained from the site of the ministry of education, culture and
science from the Dutch government (Dutch Government, ).
Table 1 - Average years of schooling to complete a level of education
Education level
Total years of schooling
Primary school
8.5
VMBO
12.7
HAVO
13.8
VWO
14.7
MBO-1
13.7*
MBO
15.7
HBO-bachelor
17.9
WO-bachelor
18.7
HBO-master
18.7
WO-master
20
*There is no number of the average duration of MBO-1, only the official duration of 1 year which I
used here instead.
I have calculated the years of schooling per possible answer with these numbers. Since there are
multiple options per answer and we don’t know the proportions of these groups, it will just be
estimation where I assume an equal distribution. For example, I assume that 50% of the
respondents who answered “4” have an HBO-bachelor’s degree and 50% of the respondents
who answered “4” a WO-bachelor’s degree.
Table 2 - Average years of schooling for the ordinal groups in dataset
Ordinal number in dataset
Total years of schooling
1
8.5
2
12.7
3
15.4
4
18.3
5
19.4
From the variable gender I have made the dummy variable Female. The variable for
migration background has three answer options and I have changed this into two dummy
variables. One dummy for a western migration background (Western) and one for a
non-western migration background (NonWestern).
The data for Age in the dataset is grouped as an interval variable. Because all groups are
approximately the same size (5-9 years), I have put it on measurement scale instead of ordinal.
After I had constructed all the needed variables, I have performed three standard linear
regressions. The independent variables were for all the three regressions the same: EducYears
and Female, Age, Western and NonWestern for control. The three different dependent variables
were Contacts, ChronDiseases, and SRHscore.
Empirical approach
In this section the regression model will be explained. The regressions are linear regressions
and the assumptions are that they are unbiased and consistent.
The regressions have the following form:
1: 𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶
= 𝛽𝛽0+ 𝛽𝛽1𝐸𝐸𝐸𝐸𝐸𝐸𝐶𝐶𝐸𝐸𝐸𝐸𝐶𝐶𝐸𝐸𝐶𝐶 + 𝛽𝛽2𝐹𝐹𝐸𝐸𝐹𝐹𝐶𝐶𝐹𝐹𝐸𝐸 + 𝛽𝛽3𝐴𝐴𝐴𝐴𝐸𝐸 + 𝛽𝛽4𝑊𝑊𝐸𝐸𝐶𝐶𝐶𝐶𝐸𝐸𝐸𝐸𝐶𝐶 + 𝛽𝛽5𝑁𝑁𝐶𝐶𝐶𝐶𝑊𝑊𝐸𝐸𝐶𝐶𝐶𝐶𝐸𝐸𝐸𝐸𝐶𝐶 + 𝜀𝜀𝑖𝑖2: 𝐶𝐶ℎ𝐸𝐸𝐶𝐶𝐶𝐶𝑟𝑟𝑖𝑖𝐶𝐶𝐸𝐸𝐶𝐶𝐶𝐶𝐸𝐸𝐶𝐶
= 𝛽𝛽0+ 𝛽𝛽1𝐸𝐸𝐸𝐸𝐸𝐸𝐶𝐶𝐸𝐸𝐸𝐸𝐶𝐶𝐸𝐸𝐶𝐶 + 𝛽𝛽2𝐹𝐹𝐸𝐸𝐹𝐹𝐶𝐶𝐹𝐹𝐸𝐸 + 𝛽𝛽3𝐴𝐴𝐴𝐴𝐸𝐸 + 𝛽𝛽4𝑊𝑊𝐸𝐸𝐶𝐶𝐶𝐶𝐸𝐸𝐸𝐸𝐶𝐶 + 𝛽𝛽5𝑁𝑁𝐶𝐶𝐶𝐶𝑊𝑊𝐸𝐸𝐶𝐶𝐶𝐶𝐸𝐸𝐸𝐸𝐶𝐶 + 𝜀𝜀𝑖𝑖3: 𝑆𝑆𝑅𝑅𝑆𝑆𝐶𝐶𝐶𝐶𝐶𝐶𝐸𝐸𝐸𝐸
= 𝛽𝛽0+ 𝛽𝛽1𝐸𝐸𝐸𝐸𝐸𝐸𝐶𝐶𝐸𝐸𝐸𝐸𝐶𝐶𝐸𝐸𝐶𝐶 + 𝛽𝛽2𝐹𝐹𝐸𝐸𝐹𝐹𝐶𝐶𝐹𝐹𝐸𝐸 + 𝛽𝛽3𝐴𝐴𝐴𝐴𝐸𝐸 + 𝛽𝛽4𝑊𝑊𝐸𝐸𝐶𝐶𝐶𝐶𝐸𝐸𝐸𝐸𝐶𝐶 + 𝛽𝛽5𝑁𝑁𝐶𝐶𝐶𝐶𝑊𝑊𝐸𝐸𝐶𝐶𝐶𝐶𝐸𝐸𝐸𝐸𝐶𝐶 + 𝜀𝜀𝑖𝑖With:
- 𝛽𝛽
0; the constant
- 𝛽𝛽
1𝐸𝐸𝐸𝐸𝐸𝐸𝐶𝐶𝐸𝐸𝐸𝐸𝐶𝐶𝐸𝐸𝐶𝐶 ; the variable for education measured in years of schooling
- 𝛽𝛽
2𝐹𝐹𝐸𝐸𝐹𝐹𝐶𝐶𝐹𝐹𝐸𝐸; the dummy control variable for gender
- 𝛽𝛽
3𝐶𝐶𝐴𝐴𝐸𝐸; the control variable for age
- 𝛽𝛽
4𝑊𝑊𝐸𝐸𝐶𝐶𝐶𝐶𝐸𝐸𝐸𝐸𝐶𝐶 and 𝛽𝛽
5𝑁𝑁𝐶𝐶𝐶𝐶𝑁𝑁𝐸𝐸𝐶𝐶𝐶𝐶𝐸𝐸𝐸𝐸𝐶𝐶; the dummy control variables for migration background
- 𝜀𝜀
𝑖𝑖; the error term with an expectancy of 0
The statistical hypothesis is the following:
- 𝑆𝑆0:
𝛽𝛽1= 0
14
Other tools
The statistical program I have used to perform the manipulation of the data and the regressions
is IBM SPSS Statistics version 24. The SPSS output can be found in appendix 1. All the papers I
have used for the literature review are found with Google Scholar and through references. To
make sure that the literature I have worked with is of sufficient level, I checked if all the papers
referred to are listed on the Tinbergen Institute Journal list.
RESULTS
Descriptives
The range of Contacts goes from 0 to 2439 with a mean of 15.11, standard deviation of 41.990
and median of 7, which means 50% of all respondents have 7 times or less contact with a
doctor. The mean of Age is 6.56 which corresponds with 44 years. ChronDiseases has a range of
0 to 19, a mean of 1.76, a standard deviation of 2.305 and a median of 1. SRHscore has a mean of
1.98 with a standard deviation of .801 and a median of 2. EducYears has a range of 8.5 to 19.4, a
mean of 14.73, a standard deviation of 3.268 and a median of 15.4, which corresponds with a
MBO level. A little bit more women engaged in the survey. The average age of the respondents is
41 years, and the percentage of respondents with a migration background sums up to 18
percent. The range, mean, standard deviation, modus and 25, 50 and 75 percentiles for all
variables are listed in Table 3.
Table 3 - Descriptive statistics; range, mean, standard deviation, modus and percentiles
Range Mean SD Modus Percentiles Contacts 0-2439 15.11 41.990 0 25: 1, 50: 7, 75: 18 ChronDiseases 0-19 1.76 2.305 0 25: 0, 50: 1, 75: 3 SRHscore 1-5 1.98 .801 2 25: 1, 50: 2, 75: 2 EducYears 8.5-19.4 14.73 3.265 15.40 25: 12.7, 50: 15.4, 75: 18.3 Female 0-1 .51 .500 1 25: 0, 50: 1, 75: 1 Age 1-11 6.56 2.946 9 25: 5, 50: 7, 75: 9 Western 0-1 .09 .284 0 25: 0, 50: 0, 75: 0 NonWestern 0-1 .09 .288 0 25: 0, 50: 0, 75: 0
Correlations
The highest correlation found is between SRHscore and ChronDiseases, which is .605. Also,
between Contacts and SRHscore and between Contacts and ChronDiseases the correlations are
positive with respectively .238 and .235. The correlation between Contacts and EducYears is
-.026, between ChronDiseases and EducYears -.105, and between SRHscore and EducYears -.123.
These and all other correlations are summarized in table 4.
Table 4 - Correlations Variable 1 2 3 4 5 6 7 1 Contacts 1.000 2 ChronDiseases .235** 1.000 3 SRHscore .238** .605** 1.000 4 EducYears -.026* -.105** -.123** 1.000 5 Female .041** .120** .052** -.027* 1.000 6 Age .082** .398** .342** .105** .021* 1.000 7 Western -.005 .018 .010 .023* .025* .021* 1.000 8 NonWestern -.004 -.002 .038** -.102** .007 -.155** -.099**
N= 9165 except the correlations with EducYears (N=7682), *p<.05, **p<.01
Coefficients
The coefficients for all the variables of the three regressions are listed in table 5. 𝑅𝑅
2is the
proportion of the dependent variable which can be explained by all independent variables
together. The 𝑅𝑅
2’s are relatively low; .006 for the first regression with Contacts, .147 for the
second one with ChronDiseases and .101 for the third with SRHscore what might conclude that
16
the model is not a good fit. The 𝑅𝑅
2is actually lower than the correlations between the
dependent variable and EducYears, the independent variable we are interested in. Except for the
coefficients for migration, all values are significant with a p at least less than 1 percent. The
coefficient in the first regression is -.422. This means that one additional year of education
predicts a .422 decrease in number of contacts per year. The beta from regression two shows
that one year of additional schooling will lower the number of chronic diseases with .096.
And the results from regression three is that an extra year of education will lower the
self-reported health with .035 (which is an improvement). To compare this with results of
Arendt (2005), my beta of .035 is lower than the .079 of Arendt. His way of measuring SRH
matches mine as his mean is 1.622 (compared to my 1.98). This suggests that my results may
underestimate the effect of education on health if we compare it with the results from Arendt
(2005).
The null hypothesis can be rejected for all three dependent variables and we can
conclude that for all models
𝛽𝛽1< 0.
Table 5 - Coefficients and R^2
Regression # 1 2 3 𝜷𝜷𝟎𝟎 11.737** (2.904) .385* (.141) 1.763** (.049) 𝜷𝜷𝟏𝟏 -.422** (.160) -.096** (.008) -.035** (.003) 𝜷𝜷𝟐𝟐 3.912** (1.036) .595** (.050) .085** (.017) 𝜷𝜷𝟑𝟑 1.166** (.228) .359** (.011) .100** (.004) 𝜷𝜷𝟒𝟒 -1.377 (1.816) .107 (.088) .035 (.031) 𝜷𝜷𝟓𝟓 .871(1.959) .531** (.095) .257** (.033) 𝑹𝑹𝟐𝟐 .006 .147 .101
*p<.01, **p<.001
DISCUSSION AND CONCLUSION
Discussion
The first limitation of this thesis is the 1483 (16.2%) missing values of EducYears that are
missing because the respondents have reported “unknown”. It is difficult what to do with this. In
the survey there is no option to fill in that someone did not even complete primary school, so I
expect a serious part of these people to have less education than finished primary school. And if
the respondents of “unknown” really didn’t know their education level, it points in the direction
of low or even no education because it can be expected that one is not too stupid to forget his
education. But this cannot simply be assumed and therefore all these results are changed to
missing instead of zero. This can create a bias where I have excluded the people with the lowest
years of schooling. As from the results of Cutler and Lleras-Mundy (2010) and Lleras-Mundy
(2005) we know that 10 years is a critical point for the effect of schooling on health. With our
used data, only the 11.6% who reported to have finished primary school are below this
boundary and the missing values could make this a much bigger proportion and even double it
what would give significant different results.
Another limitation seems to be the low 𝑅𝑅
2. The model I used only explains .6%, 14.7%
and 10.1% of the variation in the dependent variables. But these numbers are in line with other
papers. Brunello et al. (2016) can only explain 2% of the variance with his model for the relation
between health and years of education. (Fletcher, 2015)model (2015) with health as dependent
variable and years of schooling, gender, race, marital status and year of birth as independent
variable gets a 𝑅𝑅
2of .029. And also, the regression from Li and Powdthavee (2015) with years
of education and self-accessed health explains 1.5%. For my regression with SRHscore the 𝑅𝑅
2is
higher (.147). So even though a model with a higher 𝑅𝑅
2would have the preference, with the
existing literature this is sufficient.
A third limitation involves the range of Contacts that is very large in comparison to other
variables what is due to some major outliers. These exact responses have been looked over and
the reasons they report these high values have the following reasons. The first one reports to be
41 times hospitalized and that the last hospitalization was for 41 days, this would be an extreme
coincidence and it is therefore expected that it was only one hospitalization of 41 days. The
other ones report respectively 6 hospitalizations with the last one of 199 days, and 40
hospitalizations with the last one 60 days. This all adds up very fast and comes above 366 which
would be the most possible contacts counted in days per year. But because I use an approach of
the total number of contacts, these are the top ones, it is expected that there are also ones who
add up below the actual number. There only is no possibility to look if the distribution is
nominal since the lower bound is 0 which also is a regular answer(and even the modus).
Because of the high sample size, I expect this not to create a bias.
The final problem is that I have to rely on the competence of the respondents. The open
answers make it clear that not all of them answer as I would want to. I use the definition of the
WHO to clarify if something is a chronic condition or not, but the people filling in the survey do
not use this definition and make numerous things up. In the open question of which other
disease they have this becomes clear. Some examples of what people have filled in are:
“infection of my toe which is already gone”, “fatigue because of your compelling way of doing
this survey”, “grief”, “fever”, “varicose veins, but you clearly don’t find that an disease”, “first
18
time to the doctor and got a CT scan”, “I have edematous legs, do everything by bike and that is
doable and tram, but it is tough”, “broken heart”, “problems till one year ago”. These are all
answers I don’t want to include in my study but it is not possible to exclude them because this is
the only question where I get more information than just a number, and for all other questions it
is not known how people have dealt with them. But I expect them to not have been filled in as I
would like best. If this harms the results is unclear.
Something else to keep in mind is that the number of contacts with a medical specialist
may not be the best way to measure health. Dutch news articles report that due to the raise in
personal risk and personal contribution, more people refuse to see the doctor. Even though the
general practitioner is always for free, people are scared that they will be referred to a specialist
or have to do tests that affect their personal risk. There has not yet been done research if this
actually happens in the Netherlands since the change in personal risk only happened a couple
years ago. If this would be true, I expect that the actual relation between education and use of
healthcare would be more than the results show because the low-income people are related to
low education.
For this thesis it would have been best if there was also controlled for income, parental
education or childhood health endowment, as the literature review formulated that these are
relevant factors. Unfortunately this was not possible with the data.
Conclusion
The purpose of this thesis was to find a relation between the level of education and the amount
of healthcare used in the Netherlands to get a better understanding of where the rising
healthcare costs come from and to target groups with education as a form of prevention.
One of the main findings is the negative relation between education and both the number of
contacts with a medical professional and the amount of present chronic diseases. Both these
variables give an interpretation of the amount of healthcare used and can therefore be a
determinant of the costs. Also, the negative relation with the self-reported health is proved but
the comparison with the literature is still difficult.
For further research a few recommendations can be made. First of all, it would be interesting to
use more comprehensive models and include more variables because health has many more
determinants than I used in this thesis. Secondly, the direct relationship with healthcare costs
can be worked out.
REFERENCES
Adams, S. J. (2002). Educational attainment and health: Evidence from a sample of older adults.
Education Economics, 10(1), 97-109.
Arendt, J. N. (2005). Does education cause better health? A panel data analysis using school
reforms for identification. Economics of Education Review, 24(2), 149-160.
Bernell, S., & Howard, S. W. (2016). Use your words carefully: What is a chronic disease?
Frontiers in Public Health, 4, 159.
Bradley, R. H., Corwyn, R. F., McAdoo, H. P., & García Coll, C. (2001). The home environments of
children in the united states part I: Variations by age, ethnicity, and poverty status. Child
Development, 72(6), 1844-1867.
Brown, M. G., Wohn, D. Y., & Ellison, N. (2016). Without a map: College access and the online
practices of youth from low-income communities. Computers & Education, 92, 104-116.
Brunello, G., Fort, M., Schneeweis, N., & Winter-Ebmer, R. (2016). The causal effect of education
on health: What is the role of health behaviors? Health Economics, 25(3), 314-336.
Centraal Bureau voor de Statistiek. (a). Overheidsuitgaven. Retrieved from
https://opendata.cbs.nl/statline/#/CBS/nl/dataset/80393ned/table?ts=1529759307156
Centraal Bureau voor de Statistiek. (b). Zorguitgaven stijgen in 2016 met 1,8 procent. Retrieved
from
https://www.cbs.nl/nl-nl/nieuws/2017/20/zorguitgaven-stijgen-in-2016-met-1-8-procent
Chen, E., & Paterson, L. Q. (2006). Neighborhood, family, and subjective socioeconomic status:
How do they relate to adolescent health? Health Psychology, 25(6), 704.
20
Cutler, D. M., & Lleras-Muney, A. (2010). Understanding differences in health behaviors by
education. Journal of Health Economics, 29(1), 1-28.
Dijsselbloem, J. (2015). Miljoenennota 2016.
().https://www.rijksoverheid.nl/documenten/begrotingen/2015/09/15/miljoenennota-2016. Retrieved from
http://www.rijksbegroting.nl/2015/voorbereiding/miljoenennota,kst199382_16.html
Dutch Ministry of Education, Culture and Science. (2014). Key figures 2009-2013, education,
culture and science . (). Retrieved from
https://www.government.nl/documents/reports/2014/08/12/key-figures-2009-2013-ministry-of-education-culture-and-science
Fletcher, J. M. (2015). New evidence of the effects of education on health in the US: Compulsory
schooling laws revisited. Social Science & Medicine, 127, 101-107.
Grossman, M. (1972). On the concept of health capital and the demand for health. Journal of
Political Economy, 80(2), 223-255.
Kahn, J. R., & Pearlin, L. I. (2006). Financial strain over the life course and health among older
adults. Journal of Health and Social Behavior, 47(1), 17-31.
Kettings, C., Sinclair, A. J., & Voevodin, M. (2009). A healthy diet consistent with australian health
recommendations is too expensive for welfare-dependent families. Australian and New
Zealand Journal of Public Health, 33(6), 566-572.
Lleras-Muney, A. (2005). The relationship between education and adult mortality in the united
states. The Review of Economic Studies, 72(1), 189-221.
Maslow, A. H. (1943). A theory of human motivation. Psychological Review, 50(4), 370.
Mokdad, A. H., Marks, J. S., Stroup, D. F., & Gerberding, J. L. (2004). Actual causes of death in the
united states, 2000. Jama, 291(10), 1238-1245.
Nederlandse Zorgautoriteit. (2017). Marktscan zorgverzekeringsmarkt 2017. ().
Newacheck, P. W., Hung, Y. Y., Jane Park, M., Brindis, C. D., & Irwin, C. E. (2003). Disparities in
adolescent health and health care: Does socioeconomic status matter? Health Services
Research, 38(5), 1235-1252.
Pickett, K. E., & Wilkinson, R. G. (2015). Income inequality and health: A causal review. Social
Science & Medicine, 128, 316-326.
Spasojević, J. (2010). Chapter 9 effects of education on adult health in sweden: Results from a
natural experiment. Current issues in health economics (pp. 179-199) Emerald Group
Publishing Limited.
World Health Organization. (2014). Noncommunicable diseases. Retrieved from
http://www.who.int/topics/noncommunicable_diseases/en/
22
Appendix 1 – SPSS-output
Statistics
Contacts ChronDiseases SRHscore EducYears Female Age Western NonWestern
N Valid 9165 9165 9164 7682 9165 9165 9165 9165 Missing 0 0 1 1483 0 0 0 0 Mean 15.11 1.76 1.98 14.7317 .51 6.56 .09 .09 Mode 0 0 2 15.40 1 9 0 0 Std. Deviation 41.990 2.305 .801 3.26521 .500 2.946 .284 .288 Minimum 0 0 1 8.50 0 1 0 0 Maximum 2439 19 5 19.40 1 11 1 1 Percentiles 25 1.00 .00 1.00 12.7000 .00 5.00 .00 .00 50 7.00 1.00 2.00 15.4000 1.00 7.00 .00 .00 75 18.00 3.00 2.00 18.3000 1.00 9.00 .00 .00
Correlations
Contacts ChronDiseases SRHscore EducYears Female Age Western NonWestern
Contacts Pearson Correlation 1 .235** .238** -.026* .041** .082** -.005 -.004
Sig. (2-tailed) .000 .000 .023 .000 .000 .643 .719
N 9165 9165 9164 7682 9165 9165 9165 9165
ChronDiseases Pearson Correlation .235** 1 .605** -.105** .120** .398** .018 -.002
Sig. (2-tailed) .000 .000 .000 .000 .000 .088 .813
N 9165 9165 9164 7682 9165 9165 9165 9165
SRHscore Pearson Correlation .238** .605** 1 -.123** .052** .342** .010 .038**
Sig. (2-tailed) .000 .000 .000 .000 .000 .324 .000
N 9164 9164 9164 7681 9164 9164 9164 9164
EducYears Pearson Correlation -.026* -.105** -.123** 1 -.027* .105** .023* -.102**
Sig. (2-tailed) .023 .000 .000 .019 .000 .043 .000
N 7682 7682 7681 7682 7682 7682 7682 7682
Female Pearson Correlation .041** .120** .052** -.027* 1 .021* .025* .007
Sig. (2-tailed) .000 .000 .000 .019 .048 .018 .519
N 9165 9165 9164 7682 9165 9165 9165 9165
Age Pearson Correlation .082** .398** .342** .105** .021* 1 .021* -.155**
Sig. (2-tailed) .000 .000 .000 .000 .048 .048 .000
N 9165 9165 9164 7682 9165 9165 9165 9165
Western Pearson Correlation -.005 .018 .010 .023* .025* .021* 1 -.099**
Sig. (2-tailed) .643 .088 .324 .043 .018 .048 .000
N 9165 9165 9164 7682 9165 9165 9165 9165
NonWestern Pearson Correlation -.004 -.002 .038** -.102** .007 -.155** -.099** 1
Sig. (2-tailed) .719 .813 .000 .000 .519 .000 .000
24
**. Correlation is significant at the 0.01 level (2-tailed). *. Correlation is significant at the 0.05 level (2-tailed).
Regression 1 with dependent variable “Contacts”
Model Summary
Model R R Square Adjusted R Square Std. Error of the Estimate 1 .078a .006 .005 45.316a. Predictors: (Constant), NonWestern, Female, Western, EducYears, Age
Coefficients
a Model Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta 1 (Constant) 11.737 2.904 4.041 .000 EducYears -.422 .160 -.030 -2.640 .008 Female 3.912 1.036 .043 3.778 .000 Age 1.166 .228 .059 5.126 .000 Western -1.377 1.816 -.009 -.758 .448 NonWestern .871 1.959 .005 .445 .657Regression 1 with dependent variable “ChronDiseases”
Model Summary
Model R R Square Adjusted R Square Std. Error of the Estimate 1 .383a .147 .146 2.196a. Predictors: (Constant), NonWestern, Female, Western, EducYears, Age
Coefficients
a Model Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta 1 (Constant) .385 .141 2.737 .006 EducYears -.096 .008 -.132 -12.410 .000 Female .595 .050 .125 11.865 .000 Age .359 .011 .350 32.600 .000 Western .107 .088 .013 1.212 .226 NonWestern .531 .095 .060 5.596 .00026
Regression 1 with dependent variable “SRHscore”
Model Summary
Model R R Square Adjusted R Square Std. Error of the Estimate 1 .317a .101 .100 .763a. Predictors: (Constant), NonWestern, Female, Western, EducYears, Age
Coefficients
a Model Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta 1 (Constant) 1.763 .049 36.064 .000 EducYears -.035 .003 -.143 -13.116 .000 Female .085 .017 .053 4.897 .000 Age .100 .004 .289 26.200 .000 Western .035 .031 .012 1.149 .251 NonWestern .256 .033 .086 7.761 .000Appendix 2 – Explanation variables
Variable Explanation Included Excluded
because of irrelevance
Excluded because of other reasons + explanation
GEZBASMode Mode x
GEZENQInterviewDatum Datum interview x
GEZGEWCorrectieGewicht CorrectieGewicht x
GEZGEWEindGewicht EindGewicht x
GEZHHBGeslachtOP Geslacht onderzoekspersoon x
GEZVRGBetaaldWerk Afleiding of iemand betaald werk heeft of niet This could give an indication about the SES, but is
found too simplistic to include in the regression GEZVRGAantalUrenWerk Hoeveel uren werkt u in totaal gemiddeld per
week, overuren en onbetaalde uren niet meegerekend?
x GEZVRGAantalUrenWerkSchatting Om hoeveel uur per week gaat het dan ongeveer.
Is dat: x
GEZVRGDuurZwang Zwangerschap: duur zwangerschap x
GEZVRGGeboorteGewicht Gewicht: geboortegewicht x
GEZVRGGeboorteLengte Lengte: geboortelengte x
GEZVRGConsultatiebureau Consultatiebureau: aantal keren het
consultatiebureau bezocht x
GEZVRGMaatspositie Maatschappelijke positie This could give an indication about the SES, but is
suitable for the regression
GEZVRGVasteDienst Maatschappelijke positie: vaste dienst This could give an indication about the SES, but is
suitable for the regression
GEZBASErvarenGezondheid Algemene (ervaren) gezondheid x
GEZVRGLangdAandoening Eén of meer langdurige ziekten of aandoeningen x
GEZVRGBelemAct Activiteitenbeperking: vanwege
gezondheidsproblemen beperkt in activiteiten die mensen gewoonlijk doen
x
GEZVRGZwanger Zwangerschap: op dit moment in verwachting x
GEZVRGDuurZwangOP Zwangerschap: aantal weken zwanger X
GEZVRGBevallen Zwangerschap: bevallen in de afgelopen 2 jaar x
GEZVRGWaarGebKnd Zwangerschap: plaats van de geboorte; thuis, in
het ziekenhuis of ergens anders x
GEZVRGAndGebKnd Zwangerschap: plaats van de geboorte x
GEZVRGKraamZorg Kraamzorg: gebruik van kraamzorg x
28
GEZVRGStoel ADL: item stoel; gaan zitten en opstaan uit een
stoel x
GEZVRGBed ADL: item bed; in en uit bed stappen x
GEZVRGTrap ADL: item trap; de trap op- en aflopen x
GEZVRGEten ADL: item eten; eten en drinken x
GEZVRGKleden ADL: item kleden; aan en uitkleden x
GEZVRGHandenWassen ADL: item handen wassen; het gezicht en de
handen wassen x
GEZVRGVolledigWassen ADL: item volledig wassen; in bad gaan of douchen x
GEZVRGToilet ADL: item toilet; van het toilet gebruik maken x
GEZVRGKamer ADL: item kamer; zich verplaatsen naar een
andere kamer op dezelfde verdieping x
GEZVRGWoning ALD: item woning; de woning verlaten en
binnengaan x
GEZVRGBuiten ADL: item buiten; zich verplaatsen buitenshuis x
GEZVRGHulpADL ADL hulp: normaal gesproken hulp krijgen bij een
of meer ADL-handelingen x
GEZVRGMeerHulp ADL hulp: meer hulp willen bij een of meer
ADL-handelingen x
GEZVRGWelHulp ADL hulp: hulp willen krijgen bij een of meer
ADL-handelingen x
GEZVRGMaaltijd IADL: item maaltijd; maaltijden bereiden x
GEZVRGTelefoon IADL: item telefoon; telefoneren x
GEZVRGBoodschap IADL: item boodschappen; boodschappen doen x
GEZVRGMedicatie IADL: item medicatie; op tijd de juiste medicijnen
innemen x
GEZVRGHuishoudenLicht IADL: item huishouden 1; licht huishoudelijk werk x
GEZVRGHuishoudenZwaar IADL: item huishouden 2; zwaar huishoudelijk
werk x
GEZVRGAdministratie IADL: item administratie: het bijhouden van
geldzaken en dagelijkse administratie x
GEZVRGHulpHA IADL hulp: normaal gesproken hulp krijgen bij een
of meer IADL-activiteiten x
GEZVRGMeerHA IADL hulp: meer hulp willen bij een of meer
IADL-activiteiten x
GEZVRGWelHA IADL hulp: hulp willen krijgen bij een of meer
IADL-activiteiten x
GEZVRGOesoGesprekDrie OESO: item horen; een gesprek kunnen volgen in een groep van 3 of meer personen, zo nodig met hoorapparaat
GEZVRGOesoGesprekEen OESO: item horen; met één andere persoon een gesprek kunnen voeren, zo nodig met
hoorapparaat
x
GEZVRGOesoOgen OESO: item zien; de kleine letters in de krant
kunnen lezen, zo nodig met bril of contactlenzen x
GEZVRGOesoGezicht OESO: item zien; op een afstand van 4 meter het
gezicht van iemand herkennen, zo nodig met bril of contactlenzen
x
GEZVRGOesoDragen OESO: item bewegen; een voorwerp van 5 kilo,
bijvoorbeeld een volle boodschappentas, 10 meter dragen
x
GEZVRGOesoBukken OESO: item bewegen; vanuit staande positie
kunnen bukken en iets van de grond oppakken x
GEZVRGOesoLopen OESO: item bewegen; 400 meter aan een stuk
kunnen lopen zonder stil te staan x
GEZVRGBril Hulpmiddelen zien: dragen van een bril x
GEZVRGLenzen Hulpmiddelen zien: dragen van contactlenzen x
GEZVRGAnderHulpmiddelenZien Hulpmiddelen zien: een ander hulpmiddel hebben
voor zien of lezen dan een bril of contactlenzen x
GEZVRGHoorapparaat Hulpmiddelen horen: bezit hoorapparaat x
GEZVRGgeluidsversterking Hulpmiddelen horen: bezit een speciaal apparaat voor geluidsversterking, bijvoorbeeld voor telefoon of televisie
x GEZVRGHulpmiddelBewegen Maakt u wel eens gebruik van een hulpmiddel om
u voort te bewegen, zoals een stok, kruk, looprek, rollator, rolstoel of scoot(er)mobiel?
x GEZVRGHulpmiddelAnatomisch Maakt u wel eens gebruik van een anatomisch
hulpmiddel, zoals orthopedisch schoeisel, een arm- of beenprothese, of een beugel of spalk. Een beugel voor het gebit wordt hier niet onder verstaan?
x
GEZVRGHulpmiddelIncontinentie Maakt u wel eens gebruik van
incontinentiemateriaal, een blaaskatheter, stoma of stomamateriaal voor urine of ontlasting?
x GEZVRGAdhdRusteloos ADHD: item rusteloos; rusteloos gedrag vertonen,
bijna nooit stil kunnen zitten x
GEZVRGAdhdFriemelen ADHD: item friemelen; voortdurend zitten
friemelen en draaien x
GEZVRGAdhdAandachtKort ADHD: item concentratie; zich slechts kort op een
bepaalde bezigheid richten x
30
GEZVRGZindNacht Zindelijkheid s nachts bij kinderen x
GEZVRGDyslexi Woordblindheid of dyslexie bij kinderen This is a disability rather than a disease. Disability is
an umbrella term, covering impairments, activity limitations, and participation restrictions according to the WHO. A disability can be the consequence of a disease but not necessarily.
GEZVRGAutisme Autisme of een aan autisme verwante stoornis,
zoals het syndroom van Asperger, of PDD-NOS bij kinderen
x
GEZVRGVerstHand Verstandelijke beperking bij kinderen This is a disability rather than a disease. Disability is
an umbrella term, covering impairments, activity limitations, and participation restrictions according to the WHO. A disability can be the consequence of a disease but not necessarily.
GEZVRGsuikerziekte Langdurige aandoening: suikerziekte x
GEZVRGBehandelingSuikerziekte Langdurige aandoening: suikerziekte; onder behandeling of controle van de huisarts of specialist voor suikerziekte
The acquired variable of the presence of a chronic disease, in this case diabetes, has already been answered in the question above. This doesn’t add value to our variable.
GEZVRGGebruikInsuline Langdurige aandoening: suikerziekte; gebruik
insuline Idem
GEZVRGGebruikInsuline6maanden Langdurige aandoening: suikerziekte; insuline gaan gebruiken binnen 6 maanden nadat suikerziekte werd vastgesteld
Idem
GEZVRGBeroerte Langdurige aandoening: ooit een beroerte,
hersenbloeding of herseninfarct gehad x
GEZVRGRecentBeroer Langdurige aandoening: in de afgelopen 12
maanden een beroerte, hersenbloeding of herseninfarct gehad
After a cerebrovascular event, people don’t cure. Therefore it is not of interest if their event happened the last past year or earlier.
GEZVRGHartinfarct Langdurige aandoening: ooit een hartinfarct gehad x
GEZVRGRecentHartinf Langdurige aandoening: in de afgelopen 12
maanden een hartinfacrt gehad Idem
GEZVRGRcHartaandoening Langdurige aandoening: in de afgelopen 12 maanden een andere ernstige hartaandoening gehad, zoals hartfalen of angina pectoris
x
GEZVRGAangeborenAand Langdurige aandoening: aangeboren aandoening
bij kinderen x
GEZVRGAangeborenHartaandoening Langdurige aandoening: aangeboren
hartaandoening bij kinderen The previous question covers all congenital conditions.
GEZVRGKankerooit Langdurige aandoening: ooit een vorm van kanker
and finally if they had others. This would therefore double the outcome.
GEZVRGRecentKanker Langdurige aandoening: afgelopen 12 maanden
kanker gehad It doesn’t matter if the cancer was present this year or last year, it will still be a chronic disease. GEZAFLKankerSrtLeukemie Langdurige aandoening: Leukemie of bloedkanker x
GEZAFLKankerSrtLongkanker Langdurige aandoening: Longkanker x
GEZAFLKankerSrtDarmkanker Langdurige aandoening: Darmkanker x
GEZAFLKankerSrtBorstkanker Langdurige aandoening: Borstkanker x
GEZAFLKankerSrtProstaatkanker Langdurige aandoening: Prostaatkanker x
GEZAFLKankerSrtHuidkanker Langdurige aandoening: Huidkanker x
GEZAFLKankerSrtAnders Langdurige aandoening: Andere kankersoort x
GEZVRGAnderSoortKanker Langdurige aandoening: omschrijving andere
soort kanker x
GEZVRGMigraine Langdurige aandoening: migraine of regelmatig
ernstige hoofdpijn afgelopen 12 maanden x
GEZVRGAstma Langdurige aandoening: astma afgelopen 12
maanden x
GEZVRGCopd Langdurige aandoening: COPD, chronische
bronchitis, longemfyseem afgelopen 12 maanden x
GEZVRGPsoriasis Langdurige aandoening: psoriasis afgelopen 12
maanden x
GEZVRGEczeem Langdurige aandoening: chronisch eczeem
afgelopen 12 maanden x
GEZVRGDarmstoornis Langdurige aandoening: ernstige of hardnekkige
darmstoornissen voor een periode langer dan 3 maanden in de afgelopen 12 maanden
x GEZVRGGewrichtontsteking Langdurige aandoening: chronische
gewrichtsontsteking, zoals ontstekingsreuma, chronische reuma, reumatoïde artritis afgelopen 12 maanden
x
GEZVRGRugaandoening Langdurige aandoening: ernstige of hardnekkige
aandoening van de rug, inclusief hernia afgelopen 12 maanden
x
GEZVRGNekSchouder Langdurige aandoening: andere ernstige of
hardnekkige aandoening van de nek of schouder afgelopen 12 maanden
x GEZVRGElleboogPolsHand Langdurige aandoening: andere ernstige of
hardnekkige aandoening van de elleboog, pols of hand afgelopen 12 maanden
x
GEZVRGAllergie Langdurige aandoening: allergie afgelopen 12
32
GEZVRGHogeBloeddruk Langdurige aandoening: hoge bloeddruk
afgelopen 12 maanden x
GEZVRGVernauwingBloedvaten Langdurige aandoening: vernauwing van de bloedvaten in de buik of de benen, geen spataderen afgelopen 12 maanden
x
GEZVRGDuizeligheid Langdurige aandoening: duizeligheid met vallen
afgelopen 12 maanden x
GEZVRGOnvrijwilligUrineverlies Langdurige aandoening: onvrijwillig urineverlies of incontinentie afgelopen 12 maanden x GEZVRGGewrichtsslijtage Langdurige aandoening: gewrichtsslijtage, artrose
of slijtagereuma van heupen of knieën afgelopen 12 maanden
x
GEZVRGLevercirrose Langdurige aandoening: levercirrose afgelopen 12
maanden x
GEZVRGNieraandoening Langdurige aandoening: nieraandoening
afgelopen 12 maanden x
GEZVRGDepressie Langdurige aandoening: depressie afgelopen 12
maanden x
GEZVRGOverigeAandoening Langdurige aandoening: andere langdurige ziekte
of aandoening afgelopen 12 maanden? x
GEZVRGWelkeOverigeAandoening Langdurige aandoening: welke andere langdurige ziekte of aandoening afgelopen 12 maanden x
GEZVRGGevZenuw MHI: item zenuwachtig voelen
GEZVRGGevPut MHI: item in de put zitten zodat niets u kon
opvrolijken
GEZVRGGevKalm MHI: item kalm en rustig voelen
GEZVRGGevSomber MHI: item neerslachtig of somber voelen
GEZVRGGevGeluk MHI: item gelukkig voelen
GEZVRGVerkoudheid Acute ziekte: verkoudheid, griep, keelontsteking of
voorhoofdsholteontsteking x
GEZVRGBronchitis Acute ziekte: acute bronchitis of longontsteking,
hier wordt níet bedoeld chronische bronchitis x
GEZVRGOorontsteking Acute ziekte: oorontsteking x
GEZVRGNieren Acute ziekte: infectie of ontsteking van de nieren,
blaas of urinewegen x
GEZVRGDiarree Acute ziekte: diarree, hiermee wordt bedoeld
tenminste 3 maal dunne ontlasting binnen 24 uur x
GEZVRGBraken Acute ziekte: braken, hiermee wordt bedoeld
tenminste 3 maal braken binnen 24 uur x
GEZVRGPijn Pijn: pijn in de afgelopen 4 weken x
GEZVRGHuisartsContact Huisarts: contact met de huisarts x GEZVRGHuisartsAantalContact Huisarts: aantal contacten afgelopen 4 weken x GEZVRGZiekenhuisopname Ziekenhuis: een nacht of langer in een ziekenhuis
of kliniek gelegen in de afgelopen 12 maanden x GEZVRGAantalZiekenhuisopname Ziekenhuis: aantal keren in het ziekenhuis of
kliniek gelegen in de afgelopen 12 maanden x GEZVRGAantalNachtenZiekenhuis Ziekenhuis: hoeveel nachten in het ziekenhuis of
kliniek gelegen bij de laatste opname x
GEZVRGDagopname Ziekenhuis: in een ziekenhuis of kliniek
opgenomen voor een dagopname afgelopen 12 maanden
x
GEZVRGAantalDagopname Ziekenhuis: aantal dagopnamen x
GEZVRGSpecialist Specialist: contact met de specialist x
GEZVRGSpecialistaantalContact Specialist: aantal contacten afgelopen 4 weken x GEZVRGMedicijnVoorgeschreven Medicijnen: gebruik medicijnen die waren
voorgeschreven door een arts afgelopen 14 dagen x GEZVRGMedicijnNietVoorgeschr Medicijnen: gebruik medicijnen,
kruidengeneesmiddelen of vitamines die niet waren voorgeschreven door een arts afgelopen 14 dagen
x
GEZVRGPilGebruik Medicijnen: gebruik anticonceptiepil x
GEZVRGTandartsContact Tandarts: bezoek aan de tandarts x
GEZVRGTandartsAantalContact Tandarts: aantal contacten afgelopen 4 weken x
GEZVRGOrthodontistContact Orthodontist: bezoek aan de orthodontist x
GEZVRGMondhygienistContact Mondhygiënist: bezoek aan de mondhygiënist x
GEZVRGFysiotherapeutContact Fysiotherapeut: gebruik van fysiotherapie of
oefentherapie afgelopen 12 maanden x
GEZVRGFysiotherapeutAantalCont Fysiotherapeut: aantal keren fysiotherapie
afgelopen 12 maanden x
GEZVRGPsychischeZorgContact Psycholoog: contact psycholoog, psychiater of
psychotherapeut afgelopen 12 maanden x
GEZVRGPsychischeZorgAantalCont Psycholoog: aantal contacten psycholoog, psychiater of psychotherapeut afgelopen 12 maanden
x GEZVRGAlternatiefGenezerCont Alternatief genezer: contact alternatief genezer
afgelopen 12 maanden x
GEZVRGThuiszorg Thuiszorg: betaalde hulp of zorg ontvangen
vanwege problemen met gezondheid afgelopen 12 maanden
x
34
maanden gebruik gemaakt van degezondheidszorg in het buitenland, zoals een huisarts, specialist, ziekenhuis of tandarts? GEZVRGZorgBuitenlandAantal Zorg Buitenland: Hoe vaak heeft u in de afgelopen
12 maanden gebruik gemaakt van gezondheidszorg in het buitenland?
x
GEZVRGZorgBuitenlandType Zorg Buitenland: Om welke zorg ging het? x
GEZVRGZorgBuitenlandLand Zorg Buitenland: In welk land was dat? x
GEZVRGZorgBuitenlandLandCode Zorg Buitenland: Code land x
GEZVRGZorgBuitenlandReden Zorg Buitenland: Wat is de belangrijkste reden dat u contact heeft gehad met een zorgverlener in het buitenland?
x
GEZVRGGriepprik Griep: griepprik gehad in de afgelopen 12
maanden x
GEZVRGBloeddruk (preventief) onderzoek: Bloeddruk voor de laatste
keer gemeten door een zorgverlener x
GEZVRGCholesterol (preventief) onderzoek: Cholesterolgehalte in uw
bloed voor het laatst gemeten door een zorgverlener
x
GEZVRGBloedsuiker (preventief) onderzoek: Bloedsuikerspiegel voor
het laatst gemeten door een zorgverlener x
GEZVRGOccultBloedTest (preventief) onderzoek: Occult bloedtest voor het
laatst laten uitvoeren x
GEZVRGColonscopie (preventief) onderzoek: Colonoscopie ook wel
inwendig onderzoek van de dikke darm voor het laatst laten uitvoeren
x
GEZVRGMammografie (preventief) onderzoek: Mammografie x
GEZVRGUitstrijkje (preventief) onderzoek: Uitstrijkje voor het laatst
laten maken x
GEZVRGProstaatFrq (preventief) onderzoek: PSA-test: aantal keren een
PSA-test laten doen in de afgelopen 5 jaar x
GEZVRGBloeddonor Bloeddonor x
GEZVRGMantelzorg Mantelzorg geven: Heeft u in de afgelopen 12
maanden mantelzorg gegeven? x
GEZVRGMantelzorgNu Mantelzorg geven: Geeft u deze mantelzorg nu
nog? x
GEZVRGMantelzorgUren Mantelzorg geven: Hoeveel uur mantelzorg geeft u momenteel gemiddeld per week, de reistijd meegerekend?
x
GEZVRGMantelzorgDuur Mantelzorg geven: Hoe lang geeft u al mantelzorg? x
