• Open Access

The intention to switch health insurer and actual switching behaviour: are there differences between groups of people?

Authors


Michelle Hendriks
NIVEL
PO Box 1568
3500 BN Utrecht
E-mail: m.hendriks@nivel.nl

Abstract

Background  Several western countries have introduced managed competition in their health care system. In the Netherlands, a new health insurance law was introduced in January 2006 making it easier to switch health insurer each year.

Objective  The objective was to measure people’s intention to switch health insurer and actual switching behaviour. We also examined whether some groups were less inclined to switch health insurer and/or had more difficulty to exert their intention to switch.

Design  In October 2006, members of three Dutch panels indicated whether they intended to switch health insurer during that year’s open enrolment period. In the beginning of 2007, the same people were asked whether they indeed switched health insurer.

Results  Only 1% intended to switch health insurer. Women, older people, lower educated people, people who were insured for a longer period and people who reported a bad or moderate health were less inclined to switch health insurer. The amount of switching was higher among individuals who intended to switch (31%) than among individuals who did not know whether they would switch (7%) and individuals with no intention to switch (2%). Among those who intended to switch health insurer, women and people who reported a good health switched health insurer more often. The years of enrolment were also associated with actual switching behaviour.

Discussion and Conclusions  We might have to temper the optimistic expectations on enhanced choice. Future research should determine why people do not switch health insurer when they intend to and which barriers they experience.

Introduction

Several western countries have introduced some form of managed competition in their health care system.1,2 In Germany, Switzerland and the USA, for instance, health insurers have to compete for the favour of the insurants.3–6 The idea behind introducing a regulated market is that the freedom to switch will stimulate health insurers to compete by offering health care of high quality, good service and sharply priced health insurers. If consumers decide to change health insurer, they should ideally choose the best and/or cheapest health insurer, thereby enhancing the efficiency of the health care system.

The Dutch government has also introduced more managed competition by enacting a new health insurance law in January 2006. See Box 1 for the most important changes in and characteristics of the Dutch health insurance system. The Dutch government has denoted the right to choice as one of seven patients rights.7 Correspondingly, an important change is that all consumers have a free choice between insurance companies during annual open enrolment periods (November–January). Health insurers are obliged to accept all applicants for the basic package and are no longer allowed to select favourable risks or to differentiate the premium and conditions of the basic package according to (proxies for) risk.8 Before, only those with public insurance (about 67% of the Dutch population with an income below a certain maximum) could switch health insurer freely. The largest part of the privately insured (85%) could be denied by health insurers and a smaller part (15%) had no choice between health insurers whatsoever. Health insurers also have the possibility to contract health care selectively and to offer collective arrangements against a reduced nominal premium. These reforms are supposed to lead to a more demand-driven health care system that is cheaper and of higher quality.9

Box 1 The Dutch health insurance system after the insurance reform of 1 January 2006

Health insurance law

  • • Introduced on 1 January 2006
  • • Abolition of distinction between private (33% of the population) and public insurance (67% of the population with an income below a certain maximum)
  • • Insurance under private law with public limiting conditions
  • • Obligation for every citizen to take health insurance
  • • Risk adjustment

Insurance policy

  • • Free choice between health insurers
  • • Basic insurance package which is identical for everybody and health insurers are obliged to accept everybody against the same premium and conditions
  • • Choice between in-kind and restitution policy
  • • Health insurers have the possibility to contract health care selectively
  • • Possibility for citizens to take out complementary insurance (not necessarily with same health insurer as basic package). Health insurers are not obliged to accept everybody
  • • Choice of deductible (min. €100, max. €500); from 2008 obligatory deductible of at least €155
  • • No-claim premium restitution; abolished in 2008
  • • Collectives (via work or other) get premium reduction up to 10%

For the overall aims of managed competition to be achieved, it is important that some of the individuals who are dissatisfied with their health insurer or who believe they can find a better health insurer indeed switch to another health insurer. Not all insured who are dissatisfied have to switch to let the insurance system work. Hirschman has distinguished two ways for consumers to express their discontent with an organization: exit and voice.10 The voice option (i.e. expressing one’s complaints without leaving) is important for health insurers as it provides them with information on what has to be improved. The insurance companies also need the money from those who stay to accomplish improvement in their service.10 This implies that people should at least feel free to switch health insurer and should not perceive or experience barriers when intending to switch or doing so.

People can experience several barriers (also called switching costs) when switching health insurer.11–15 First of all, health insurers may have the possibility to refuse people that apply for an insurance. In the Netherlands, health insurers cannot refuse individuals applying for the basic insurance package but do have the liberty to refuse individuals applying for complementary insurance (see Box 1). There are also transaction costs, which include both the direct financial costs of enrolment and the time and effort it takes to find relevant information on alternative insurers. In addition, changing health insurer brings along uncertainty about the quality of the service of the new insurer and concerns about continued coverage of pre-existing conditions. All these barriers may discourage people from switching health insurer.

In the year of the transition to the new Dutch health insurance system (2005–2006) as much as 20% of the Dutch population switched health insurers.16,17 This initial high level of switching can probably partly be explained by the fact that the freedom to switch health insurer without the risk of being denied for the basic package was new for people who were privately insured up to 2006 (about 33% of the Dutch population). Moreover, health insurers promised not to refuse any applicants for the complementary insurance removing some important barriers to switch for all consumers. The last few years, only about 3–4% of the Dutch population switched health insurer. This percentage is comparable to the switching rate among publicly insured before 2006 and to the switching rates in Germany (4–5%) and Switzerland (5%).4,18,19 These switching rates are higher than in Belgium and Israel, where only 1% switches health insurer.4

Empirical studies in several countries have revealed that changing health insurers is more likely among younger people, well-educated people and people in relatively good health.11,16,17,19–22 Inconsistent results were found concerning the relationship between switching behaviour and gender, ethnicity and income.11,17,20,21 The differences between groups are often explained in terms of switching costs. Switching costs are likely to be larger for individuals with greater health care needs (i.e. older people and people in poor health, chronically ill people), because these individuals are especially averse to uncertainty about continued coverage of health care.14,23 In addition, the young and healthy probably need less information for their choice between health insurers as they do not use much health care and base their choice on price only.16 The more educated people and people with a higher income might be more willing to incur risk, and higher educated individuals will probably also experience less difficulty to process all the relevant information on the available health insurers, thereby decreasing the transaction costs.24

The low amount of switching and the differences in switching behaviour between groups are, however, difficult to interpret without information on the intention to switch. The fact that (groups of) people do not switch health insurer may indicate that they are satisfied with their insurer, enumeration that they do not perceive substantial differences between insurers, or that they wanted to switch health insurer but eventually decided not to do so.13 Only the latter implicates that there are barriers for switching health insurer that need to be removed.

Instead of comparing individuals who switch health insurer with those who do not switch, it is therefore important to determine how many people want to switch health insurer and whether the intention to switch health insurer results in actual switching behaviour. Another important question is whether some categories of people are less inclined to switch health insurer and/or have more difficulty to exert their intention to switch than others. Using the open enrolment period in the Netherlands of 2006–2007, we answered the following research questions:

  • 1 How many people have the intention to switch health insurer and how many of them do indeed switch health insurer?
  • 2 Are there differences in the intention to switch health insurer and realizing this intention according to characteristics such as age, gender, education level, years of enrolment, subjective health status and having a chronic illness or handicap?

Method

Design

Given the low prevalence of switching behaviour, it was necessary to gather data of as many people as possible. We approached the members of three panels: the Dutch Health Care Consumer Panel (COPA), the National Panel of the Chronically ill and Disabled (NPCD) and the VGZ Insurants Panel (VGZ). Together, these panels have over 18 000 members. Several background variables were known for all the members of the panels. This study involved a two-stage procedure. In October 2006, members of the three panels received a questionnaire containing a question concerning their intention to switch health insurer by the end of the year. In the beginning of 2007, we asked the same members of the panels whether they switched health insurer or not.

Composition of the panels

Dutch Health Care Consumer Panel

The COPA consists of about 2800 members and these members are representative for the general Dutch population aged 18 years or older concerning age and gender. Every 2 years, one third of the panel members is renewed. This renewal ensures that the panel remains a cross-section of the population enumeration, that members do not develop specific knowledge of and attention for health care issues and no ‘questionnaire-fatigue’ occurs. New members for the panel are sampled from the general population. Sampled people receive an information letter about the panel and are called within a week after receiving that letter. If they are interested they receive a questionnaire on background characteristics. When that questionnaire is returned they are considered members of the panel.

National Panel of the Chronically Ill and Disabled

The NPCD is a nationwide longitudinal research programme investigating the consequences of having a chronic disease or long-term physical disability for patients and their families in the Netherlands.25 The panel has about 3800 members and is representative for the population of independently living people of 15 years and older with a (somatic) chronic illness or disability. Chronically ill patients were recruited in Winter 2004/2005 via a nationally representative sample of 35 general practices in the Netherlands. Patients were selected by their GPs on the basis of the presence of a diagnosis of a somatic chronic disease as defined by the Council for Public Health and Health Care of the Netherlands.26 The disabled patients were selected through a screener question on disabilities in two large-scale regular population surveys. Other selection criteria were aged 15 years or older, not terminally ill, and a sufficient mastery of the Dutch language. Panel members take part in the panel during 4 years. Every year a quarter of the chronically ill panel members are renewed, disabled are renewed every 4 years. Patients received written information about the study and participating patients gave informed consent.

VGZ Insurants Panel

The VGZ consists of over 11 500 insurants of one particular Dutch health insurer, VGZ. The aim of the panel is to gather information on consumers’ experiences with and expectations of health care in general and their health insurer in particular. Members for the panel were recruited through an announcement in the magazine of the health insurer. People who responded positively received an information letter about the panel and a questionnaire on background characteristics. When that questionnaire was returned they were considered members of the panel. As the VGZ only existed for 1 year at the time of the study, no renewal of panel members had taken place yet.

The protection of the collected data from the COPA and the NPCD was laid down in privacy regulations. All three panels are registered by the Dutch Data Protection Authority (no. 1262949, no. 1283171 and no. 1309664).

Measurements

At time of registration, panel members filled in a questionnaire on several background variables. These variables included for all panels age, gender, education level and self-reported health status. Of the members of COPA and VGZ, but not of the members of NPCD, we also knew how long they were insured at their health insurer. In addition, we knew that all members of NPCD had one or more doctor-diagnosed chronic illness or long-term physical disability. Of the members of the other two panels it was unknown whether they suffered from a doctor-diagnosed chronic illness or handicap.

In the questionnaires of October 2006, people were asked whether they intended to switch health insurer at the end of the year. Possible answers were no, yes and I don’t know (yet). In the beginning of 2007, we asked the same members of the three panels whether they were still insured at the same health insurer or whether they switched health insurer. Members of COPA and of NPCD also indicated their reasons for switching or not switching.

Statistical analyses

As both members of NPCD and of VGZ are older than a sample of the general Dutch population, we used a weight factor to make the sample representative for the general Dutch population concerning gender (male; female) and age (18–44; 46–64; 65 years and older). Both unweighted and weighted results will be presented. The statistical analyses mentioned below were performed on the weighted results.

We examined the relationship between respondent characteristics and the intention to switch by performing chi-squares tests for the following characteristics: gender, age, education level, years of enrolment and subjective health status. To determine whether the realization of the intention to switch differed according to characteristics of the respondents, we performed a logistic regression analysis on actual switching behaviour for the individuals who intended to switch health insurer with gender, age, education level and self-reported health status as predictor variables. Years of enrolment at the health insurer were not included in this regression analysis, as we did not know how long members of the NPCD were insured at their current health insurer. Instead, we performed a chi-square test to examine separately the relationship between years of enrolment and actual switching behaviour without the members of NPCD included.

To get insight in the influence of having a doctor-diagnosed chronic illness or handicap on the intention to switch health insurer and actual switching behaviour, we compared the members of the NPCD with the members of the other two panels using chi-square tests. We did not include this variable in the regression analysis, as panel membership is only a proxy for the presence of a doctor-diagnosed chronic illness or handicap.

Results

Respondents

The October 2006 questionnaire was send to 1298 members of the COPA and 1107 members (response rate = 85%) filled in the questionnaire. Of the NPCD, 2903 members received the questionnaire and 2411 people (response rate = 83%) filled in the questionnaire. A sum of 10 195 members of the VGZ received the questionnaire and 8881 (response rate = 87%) responded. In the beginning of 2007, we asked the 12 399 respondents of the October 2006 questionnaires whether they switched health insurer. A sum of 10 721 individuals (response rate = 87%) answered this question. See Table 1 for the characteristics of the respondents.

Table 1.   Background variables of the respondents
 COPANPCDVGZTotal
(n = 1107)(n = 2411)(n = 8881)(N = 12 399)
  1. Values given are in percentages.

  2. COPA, Dutch Health Care Consumer Panel; NPCD, National Panel of the Chronically ill and Disabled; VGZ, VGZ Insurants Panel.

Gender
 Male54375450
 Female46634750
Age
 18–44 years25121415
 45–64 years48434142
 65 years or older27464543
Education level
 Low education23452730
 Moderate education45404343
 High education32153027
Years of enrolment
 0–2 years2334
 2–5 years875
 5–10 years12108
 over 10 years578062
 Unknown10020
Self-reported health status
 Bad/Moderate22482831
 Good46454545
 Very good/Excellent3282824

Intention to switch

In October 2006, 166 (1%) individuals indicated that they intended to switch health insurer by the end of the year; 10 547 (86%) respondents had no intention to switch, and 1536 (13%) respondents did not know yet whether they wanted to switch health insurer.

The results of the chi-square tests revealed that the intention to switch health insurer was correlated with gender, age, education level, years of enrolment and self-reported health status (see Table 2). Women, older people, people with a lower education, people who were insured at their health insurer for a longer period and people who reported a bad or moderate health were less inclined to switch health insurer by the end of the year than their counterparts. Members of the NPCD (i.e. the chronically ill and disabled) had less intention to switch health insurer than members of COPA and VGZ.

Table 2.   Intention to switch health insurer in percentages (n = 12 249) and chi-squares indicating the relationship between background variables and intention to switch
 UnweightedWeighted for age and gender
No (%)Don’t know (%)Yes (%)No (%)Don’t know (%)Yes (%)χ2
  1. **P < 0.01; ***P < 0.001.

  2. COPA, Dutch Health Care Consumer Panel; NPCD, National Panel of the Chronically ill and Disabled; VGZ, VGZ Insurants Panel.

Gender
 Male861318117217.73***
 Female8612183152
Age
 18–44 years7820377203263.71***
 45–64 years8414284142
 65 years or older91819181
Education level
 Low education90918811285.27***
 Moderate education8514181172
 High education8315280183
Years of enrolment
 0–2 years7819376213112.14***
 2–5 years7721271254
 5–10 years8216279192
 over 10 years8712183152
Self-reported health status
 Bad/Moderate891018613158.05***
 Good8613181162
 Very good/Excellent8415279183
Panel membership
 COPA and VGZ851318117253.86***
 NPCD (chronically ill and disabled)8910187112
Total8613182162

Switching behaviour

Actual switching behaviour was known for 10 721 respondents. Of these respondents, 10 426 (97%) individuals were still insured with the same health insurer in the beginning of 2007; 295 (3%) respondents had switched health insurer.

Table 3 shows the number and percentage of people that switched health insurer as a function of the intention to switch. Actual switching behaviour was known for 117 of the 166 individuals who intended to switch; 65 (39%) individuals were still insured with the same health insurer and thus did not behave as planned and 52 (31%) individuals switched health insurer as planned. The portion of people who switched health insurer was substantially higher in the group who intended to switch (31%) than in the group who did not known yet whether they intended to switch (7%) and the group who had no intention to switch (2%). Fourteen individuals who did not switch health insurer while they did intend to indicated their reasons for not switching. It appeared that they mainly stayed with their health insurer because they nonetheless were satisfied with the premium and/or coverage of the insurance and/or because they did not know what to expect from a new health insurer (not in table).

Table 3.   Switching behaviour as a function of intention to switch
Intention to switchSwitching behaviour
No switching (%)Switched (%)Unknown (%)
No (n = 10 547)9074 (86)164 (2)1309 (12)
Don’t know (n = 1536)1183 (77)72 (7)281 (18)
Yes (n = 166)65 (39)52 (31)49 (30)

The results of the logistic regression analysis revealed that for those who intended to switch health insurer actual switching behaviour was correlated with gender and self-reported health status (see Table 4). No effects were found for age and education level. Women switched health insurer more often than men and people who reported good health switched health insurer more often than people who reported very good to excellent health. We also found differences in switching behaviour depending on the years of enrolment (χ2 = 14.73, P < 0.01). Among those who intended to switch health insurer, people who were insured at their health insurer for 2–5 years or for 10 years or longer switched less often than people who were insured at their health insurer for up to 2 years or for 5–10 years. Respondents who were insured at their health insurer for over 10 years switched health insurer least often. Finally, among those who intended to switch health insurer, we found no differences in actual switching behaviour between the members of NCPD (i.e. the chronically ill and disabled) and the members of COPA and VGZ (χ2 = 0.14; P = 0.71).

Table 4.   Switching behaviour of those who intended to switch in percentages (n = 117) and odd ratios indicating the relationship between background variables and switching behaviour
 UnweightedWeighted for age and gender
No switching (%)Switched (%)No switching (%)Switched (%)Odds ratio 95% CI
  1. 1Not included in the regression analyses because data were not known for all groups

  2. **P < 0.01; ***P < 0.001.

  3. COPA, Dutch Health Care Consumer Panel; NPCD, National Panel of the Chronically ill and Disabled; VGZ, VGZ Insurants Panel.

Gender
 Male663458420.33**0.16–0.66
 Female47534555Reference 
Age
 18–44 years475349511.180.28–4.94
 45–64 years604060400.160.27–3.43
 65 years or older54465248Reference 
Education level
 Low education594158420.740.27–2.05
 Moderate education534751491.010.49–2.10
 High education56445347Reference 
Years of enrolment1
 0–2 years31692971  
 2–5 years56445248  
 5–10 years36643466  
 over 10 years66356832  
Self-reported health status
 Bad/Moderate683264371.080.36–3.20
 Good465442582.91**1.36–6.19
 Very good/Excellent59416238Reference 
Panel membership1
 COPA and VGZ56445347  
 NPCD (chronically ill  and disabled)54465050  
Total56445248  

Discussion

In this study, we investigated how many people want to switch health insurer per year and how this behavioural intention relates to actual switching behaviour. Furthermore, we determined whether the intention to switch health insurer and the realization of this intention differed according to characteristics such as age, gender, education level, years of enrolment, subjective health status and having a chronic illness or handicap. Members of three Dutch panels were asked in October 2006 whether they intended to switch health insurer that year. In the beginning of 2007, the same people indicated whether they switched health insurer or not.

The results revealed that only a small group of people intended to switch health insurer. In line with previous studies, women, older people, people with a lower education level and people facing health issues were less inclined to switch health insurer. In addition, people who were insured at their health insurer for a longer period had less intention to switch health insurer. This is in line with Frank et al. who have shown that people with longer periods of attachment to a particular health insurer were less likely to express an intention to switch health insurer.12

Respondents who intended to switch health insurer indeed switched more often than respondents who did not know yet whether they wanted to switch and respondents with no intention to switch. Other studies have revealed that the number of people who enacted their intention to switch appears to be stable in the Netherlands over the years.27–29 This means that asking people about their intention to switch health insurer can be used to predict the amount of actual switching behaviour. This has indeed been performed successfully in the Netherlands.27–29

However, a substantial part of the people who intended to switch health insurer did not carry out this intention and stayed insured with the same health insurer. Possibly, these individuals experienced barriers when they tried to switch health insurer. Which barriers (i.e. refusal of the insurant by the health insurer, transaction costs or experienced uncertainty about the quality of the new health insurer) played a role can not be inferred from this study due to the small group that intended to switch health insurer. It is known that in the year of the transition to the new Dutch health insurance system (2005–2006) as much as 20% of the Dutch population switched health insurers.16,17 This high switching rate can be explained in terms of the absence of several (pre-existing) barriers. The freedom to switch health insurer without the risk of being denied was new for people who were privately insured up to 2006 (about 33% of the Dutch population). It is likely that the privately insured people who wanted to switch health insurer for a longer period but could not (especially individuals with health problems) promptly took their chance to switch health insurers from 2005 to 2006. In addition, health insurers promised not to refuse any applicants for the complementary insurance in the transition year making it more attractive to switch health insurer for all consumers. This suggests that the possible refusal of individuals by the health insurers might be an important barrier discouraging people from switching health insurers. This barrier probably will become more important in the Netherlands the upcoming years. Research has shown that health insurers increasingly accept consumers with health risks only under less favourable conditions and they also make it less attractive, if not impossible, for consumers to apply for complementary insurance without taking out the basic package.30

The low level of switching health insurers can also be explained in terms of loyalty. Hirschman argued that especially people who are loyal to their health insurer will use voice rather than exit.10 Possibly, some people decided not to exert their intention to switch but instead choose to be loyal and to express their discontent in other ways, for instance by making a complaint or participating in client councils.

Still another reason why people may not carry out their intention to switch health insurer is that they conclude that the assumed advantages of switching are not that profound or do not exist at all. For instance, low price elasticities could explain why people who intended to switch health insurer eventually did not. Van Dijk et al. found that only 5% of the Dutch people switched to another health insurer when the premium increased with 10%.31 In 2007, the premium for the basic insurance package in the Netherlands varied between 85 and 100 Euros per month, a difference of 17%. On further consideration, people who wanted to switch health insurer might have concluded that the price differences between health insurers were too small and that they could not benefit enough financially by switching health insurer. A similar argument can be used concerning the role of differences between health insurers in provided coverage or free choice of health providers. Based on experiences from the USA, it can be expected that people (especially the elderly and the chronically ill) will not opt for an health insurer with restricted choice of health care providers.32,33 Dutch health insurers have the possibility to contract health providers selectively but are only slowly starting to do so. The current differences in accessibility of health providers might not be substantial enough to persuade people to switch health insurer, but this might change in the near future.

Whether people carried out their intention to switch health insurer did not differ much between groups of people. Among the individuals who intended to switch, men switched less often than women, people with a very good to excellent health switched less often than those with a good health and years of enrolment was correlated with actual switching behaviour with individuals being insured 10 years or longer switching least often. Contrary to previous studies on switching behaviour in general, no effects of age, education level or having a chronic illness or handicap were found on whether or not people switched as they intended. These results suggest that the differences in switching behaviour found in previous studies mainly reflect differences in the intention to switch between groups of people and less the differences in the realization of this intention. Because of this, it is necessary not only to look at actual switching behaviour but also to determine whether people intend to switch health insurer.

The fact that groups of people differ in their intention to switch but not concerning the realization of this intention indicates that perceived barriers may not only discourage people from carrying out their intention to switch health insurer but may even prevent individuals to form an intention to switch. According to the theory of planned behaviour, a behavioural intention is the result of three components: the attitude towards certain behaviour, the social pressure or subjective norm, and perceived behavioural control.34 Perceived behavioural control is the perceived ease or difficulty of performing the behaviour and is determined by beliefs about the presence of factors that may further or hinder performance of behaviour itself (i.e. barriers). In other words, people may be dissatisfied and may want to switch health insurer, but if they perceive barriers to do so this would prevent people to form an intention to switch.

An important question is whether it is problematic that certain groups anticipate more barriers for switching health insurer than others. De Jong et al. found that people in general switch health insurer because of premium differences and that only the chronically ill and disabled also valued the content of the insurance package.16 When people who face health issues (i.e. older people and the chronically ill and disabled) feel hindered to switch health insurer, managed competition might revolve more around the premium than the content of the insurance package. As a result older people and the chronically ill and disabled might benefit less from the enhanced choice in the new health care system.13 Policy makers should aim at providing information and equal possibilities for all groups of people to switch health insurer and should stimulate individuals to base their choice of health insurer on both price and quality to make the system perform as intended. Enforcing patient organizations could be a possibility for reaching this goal.

It is unknown how many people have to switch health insurer for managed competition in health care to succeed in any country.13 If the majority of people is satisfied with their current health insurer or decides after a cost-benefit analysis that other health insurers do not provide enough benefits, a low amount of switching health insurer is not alarming. The low amount of switching can also be explained by the knowledge that people have a tendency to leave things as they are because they are, for instance, afraid they are going to regret the choice they make.35 These processes can not be attributed to flaws in the health insurance system. The situation becomes problematic when people feel somehow hindered by the system to switch health insurer when they want to. Future research should investigate why people do not switch health insurer when they intend to do so to get a better insight into the possible barriers people experience when they want to switch health insurer. Policy makers should aim at removing any existing barriers.

Limitations of the study

This study has some drawbacks. For one, the sample is not representative for the Dutch population. Considering the low prevalence of switching health insurer, members of three panels were approached. The largest panel (the VGZ Insurants Panel) consisted of members who were all insured with one particular Dutch health insurer. Compared to the general Dutch population, older people (between 40 and 80 years old), people with a bad to moderate self-reported health status and people who are insured with their health insurer for over 10 years are overrepresented in this panel compared to a sample of a national study on experiences with the service of health insurers. Data for the Dutch population on self-reported health status and years of enrolment with the same health insurer are not available. Although we corrected for the age difference by weighing the results, the other differences might have led to distortions in the results.

Although we approached more than 10 000 people for this study, the number of people intending to switch health insurer shortly before the annual open enrolment period was still low. This made it especially difficult to determine if the characteristics of people who intended to switch were related to actual switching behaviour. This is a problem that is not easy to overcome given the low prevalence of switching between health insurers.

We considered health status an important determinant of the intention to switch health insurer and actual switching behaviour. We, therefore, compared the members of the NPCD with the members of the other two panels assuming that all members of the other panels did not have a doctor-diagnosed chronic illness or disability. Calsbeek et al., however, estimated that about 17% of the Dutch population aged 15 years or older has a chronic illness and/or handicap.36 This means that we probably underestimated the differences between people with and without a chronic illness or disability.

Conclusion

We might have to temper the optimistic expectations of enhanced choice in the health care system.13 In the Netherlands, only a small group switches health insurer per year. This study revealed that only 1% of the respondents intended to switch health insurer and of these individuals a substantial part (39%) did not carry out this intention. Although the realization of the intention to switch did not differ much between groups of people, the intention to switch health insurer did. This might indicate that differences in perceived barriers not only prevent people from switching but also withhold people from forming the intention to switch health insurer. Future research should determine why people do not switch health insurer when they intend to and which barriers they anticipate or experience. This could lead to recommendations for changing the health insurance system and diminish the barriers for switching health insurer.

Sources of funding

The data collection was supported by the Ministry of Social Affairs and Employment for the NPCD, by the Ministry of Health for the COPA and by the health insurer VGZ for the VGZ Insurants panel.

Conflict of interest

In the conduct of the study there were no conflicts of interests.

Acknowledgements

The authors thank the members of the Dutch Health Care Consumer Panel, the VGZ Insurants Panel and the National Panel of the Chronically ill and Disabled for their participation.

Ancillary