Who has Benefited from Choice in Healthcare?

This report studies and compares three county councils before and after the healthcare choice reform.

The healthcare choice reform in primary care in Sweden began with the introduction of models for free establishment and freedom of choice in an independent initiative by the county councils of Halland, Stockholm and Västmanland. Since the mandatory introduction in 2010 of healthcare choice systems in primary care, the consequences resulting from patients’ freedom to choose their surgery, and the opportunities provided for new ones to be set up, have been recurring themes in the political healthcare debate. The objective of the reform, which is to give patients a greater role in their healthcare, can conflict with the principal of fairness, whereby healthcare should be provided according to need. An important question then arises as to whether or not groups in greater need of healthcare services have been displaced in favour of patients with simpler requirements. In this connection, the importance of the regulation and reimbursement systems in determining the healthcare providers’ priorities has received particular attention.

The effects of the introduction of healthcare choice models for various patient groups can be illustrated using empirical data. Existing analyses in this area are currently limited to individual county councils. There is a lack of comparative studies between county councils in which differences in the design of healthcare choice are analysed using a common method. Reviews of the literature also point to a shortage of knowledge about the correlation between freedom of choice and the distribution of healthcare in general, both in Sweden and internationally.

The purpose of this study is to analyse how healthcare utilisation and accessibility as experienced by the patient have developed – before and after the advent of healthcare choice – in groups with long-lasting and/or care-intensive illnesses, compared with the population as a whole. Developments for a number of patient groups (hereinafter referred to as those ‘needing healthcare’), where there is a previously confirmed diagnosis indicating a greater expected need for primary care, are therefore compared with those for the general population. The groups of patients needing healthcare are thus compared with individuals with both greater and lesser healthcare needs, or with people who have indicated that they do not have a diagnosis where they expect to need healthcare.

Developments in healthcare utilisation and accessibility as experienced by the patient for the various need groups are then compared among three county councils/regions with partially differing healthcare choice models. The study involves Stockholm County Council, Östergötland County Council and Skåne Regional Council. The report thus illustrates differences between the county councils’ management and reimbursement models before and after the introduction of the healthcare choice reform. The study ends by also examining differences in healthcare utilisation for the various need groups between the private and public healthcare sector, as well as looking at trends in outpatient care consumption for the various need groups.

The analyses of developments before and after the introduction of the healthcare choice models in the three county councils are summarised as follows:

• a comparison of primary care consumption between groups needing healthcare and the population as a whole;

• a comparison between the private and public healthcare sectors regarding listing and healthcare utilisation in groups needing healthcare and the population as a whole;

• a comparison of overall consumption of outpatient and specialised healthcare between groups needing healthcare and the population as a whole;

• a comparison of accessibility, as experienced by the patient, between groups needing healthcare and the rest of the population;

• an assessment of the extent to which developments in healthcare utilisation can be attributed to the design of the county councils’ management and reimbursement models.

Finally, the purpose is also to examine the basis on which all county councils will monitor developments in primary care consumption for the various patient groups, and the work currently being done on this issue by the county councils.

IMPLEMENTATION

The analyses are based throughout on a comparison between groups with permanent and/or care-intensive illnesses and groups without any major expected healthcare needs.

In the register-based comparison of healthcare utilisation, a study was made of patients with diabetes, asthma, chronic obstructive pulmonary disease (COPD) or dementia, or patients with at least one diagnosis within a broader group of care-intensive illnesses. All comparisons relate to the period 2006–2011. Developments in the groups with major expected healthcare needs are compared with those in the population. In the comparison of accessibility as experienced by the patient, people are included who have at least one of eight diagnoses indicating a regular need for primary care in the group needing healthcare. The results are based on a patient questionnaire conducted in the autumn of 2012. The group with major expected healthcare needs is compared with the rest of the population.

The analyses of healthcare utilisation among the various need groups are based on developments in visits to surgeries (to see either a doctor or a nurse) and home visits. Five different definitions of groups needing healthcare are thus compared with the corresponding gender and age group for the whole population.

Absolute visit rates, and the relative development in healthcare utilisation before and after the introduction of the healthcare choice reform, are presented. The various healthcare choice models of the county councils are also discussed in relation to the pattern of consumption and how this is expected to affect healthcare utilisation. For example, there are differences when balancing between reimbursement per listed patient (capitation), per visit to each category of personnel (payment per visit) and target-related reimbursement.

WHAT ARE THE MOST IMPORTANT RESULTS?

  • Most of the comparisons show that people with major expected healthcare needs are not being displaced in terms of the absolute number of visits following the introduction of the healthcare choice reform.

An assessment of the displacement effect can be based on how absolute healthcare utilisation has decreased or increased before and after the reform for groups needing healthcare. The assessment can also be based on how their share of healthcare utilisation has developed in relation to the population (a relative comparison). Using a weighted7 comparison of primary care utilisation, we can then establish that no absolute displacement effects have occurred for the majority of groups needing healthcare following the introduction of healthcare choice.

Figure 1 shows the change, expressed as a percentage, in the number of weighted primary care contacts for various diagnosis and population groups before and after the introduction of the healthcare reform. The reform was introduced in Stockholm on 1 January 2008. Therefore, the period 2008–2011 after the reform is compared with the period 2006–2007 before the reform. In Skåne and Östergötland, the reform was introduced in 2009 (in Skåne on 1 May 2009 and in Östergötland on 1 September 2009), and therefore the comparison instead is between the period 2010–2011 after the reform and the period 2006–2008 before the reform.

Developments ‘nationally’ refer to a comparison between the average number of weighted healthcare contacts per inhabitant for the period 2009–2011 and that for the period 2006–2008.

The results show that the weighted primary care consumption is increasing in all groups except people with a diagnosis of dementia in Stockholm and Östergötland and people aged 65–84 in Östergötland. For patients with a diagnosis of dementia, the probable explanation is that, because of the course of the illness, this patient cohort is cared for in care homes, in the municipalities and in other institutions. Insofar as the consumption of physical visits reflects the opportunities for contact with primary care, accessibility has therefore improved for most of the groups needing healthcare and for the population as a whole.

  • Since the introduction of healthcare choice, the population in many respects has increased its healthcare utilisation to a greater extent than is the case with people with permanent healthcare needs.

It is above all Skåne that shows the greatest weighted increase in visits in the population as a whole compared with people needing healthcare. On the other hand, in the majority of cases the increase in Östergötland is greater among the groups needing healthcare. In Stockholm County, the increase was greater in half the cases. This is mainly due to the fact that, in certain cases, the increases in home visits by doctors and visits by nurses are greater for the groups needing healthcare under these county councils.

In Skåne, the increase in visits in an overwhelming proportion of the cases has been greater in the population as a whole than in the groups with major expected healthcare needs. This applies to surgery visits to doctors, home visits by doctors and surgery visits to nurses. This may be interpreted as a displacement as far as relative change is concerned, but one where people needing healthcare have a clearly higher healthcare utilisation. A certain degree of support may be lent to this interpretation by the fact that, compared with Stockholm and Östergötland, the results of the patient questionnaire indicate that a larger proportion of patients in Skåne with major expected healthcare needs have indicated that accessibility has deteriorated following the introduction of healthcare choice. However, the change between levels is a marginal one, and the groups needing healthcare still show a distinctly higher level of primary care consumption under all county councils.

Under all three county councils, the population has accounted for a greater proportion of the increase in visits to doctors in the primary care service than have groups permanently needing healthcare.      

In Stockholm and Östergötland, the number of home visits by doctors is increasing significantly more in the groups of those needing healthcare than in the population in general. For Östergötland, a small decrease has been noted in the number of home visits among the population. However, in a weighted comparison of the number of visits to the doctor – both surgery and home visits – the population under all county councils has accounted for a greater proportion of the increase in visits to doctors than have groups permanently needing healthcare.

Looking only at surgery visits to doctors in primary care, the results under all county councils point to the majority of the increase in visits being attributable to the population as a whole rather than those groups with major expected healthcare needs. The clear increase observed in Stockholm in the proportion of visits to doctors in cases of simpler diagnoses, compared with those of people with major healthcare needs, further underlines this trend. This is additionally supported by the general increase in the number of non-specific diagnoses in primary care. However, it is difficult to say whether this development is desirable or not. It may be a case of previous accessibility problems for the population, which the low frequency of visits to doctors internationally suggests.

It may also be the case that healthcare providers have increased the frequency of visits by concentrating on patients who are relatively easy to treat. It is not possible to provide an answer to the question in this report. Further research and investigation are required.

  • There are minor differences between private and public sector healthcare providers in terms of the healthcare utilisation and listing patterns of the various need groups – except for newly-established private healthcare providers, who have a smaller proportion of listed people with major expected healthcare needs.

There has been a strong growth in the number of private surgeries in Skåne and Stockholm during the period. One of the subsidiary aims of the study was to examine whether there are differences in the distribution of private and public sector healthcare between patient groups needing healthcare and the population in general.

During the first few years following the reform, there was a tendency in Stockholm for the patient groups with major expected healthcare needs to obtain their care at public sector units to a greater extent than was the case in the population. During the past year, however, the difference has become smaller and then disappeared. In Skåne, too, there are only small differences between private and public sector healthcare for patient groups with major expected healthcare needs and the population in general.

As far as listing is concerned, information from Stockholm shows that the listing between groups needing healthcare and the population as a whole has developed in the same way. The listing at private surgeries has increased for both groups to approximately the same extent. However, patient groups with major healthcare needs are listed at newly-established private surgeries to a lesser extent than is the population as a whole.

  • In certain cases, the increase in healthcare utilisation can be explained by increased competition and the increase in the supply of primary care services following the introduction of healthcare choice.

The weighted increase in visits in the population of all ages is greatest in Stockholm, at 16 per cent. In Skåne, too, the increase of 14 per cent is in excess of the general increase nationwide, which is 12 per cent.  For Östergötland, the corresponding development is an increase of 7 per cent.

Primary care consumption is affected by the need and demand for healthcare, as well as the supply of primary care services. For example, it is well established in the scientific literature that changes in healthcare supply have a major effect on developments in consumption. In Stockholm and Skåne, there have been significant new developments of surgeries, while in Östergötland the supply has been relatively constant. A possible explanation in Östergötland’s case is the difficulties in recruiting doctors in sparsely populated areas. In addition, a number of private general practitioners in both Skåne and Stockholm, who were reimbursed according to the national rate in Sweden, have chosen to relinquish their agreements. Instead, they have signed contracts in accordance with the licensing terms of the healthcare choice models. One explanation of the increase in healthcare utilisation among the population in general under those county councils is, therefore, quite simply that the overall supply of primary care services has increased. The changes in supply, along with the encouragement given to patients to make active choices, have also led to significantly greater competition to obtain patients. This can also explain an increase in healthcare utilisation.

  • There are no clear signs that the increase in the number of primary care contacts has relieved the pressure on outpatient care in hospitals.

The increase in visits and healthcare contacts that has taken place in the primary care sector has generally not relieved the pressure on outpatient care in hospitals, although examples of the opposite can be found for certain groups. Generally speaking, the number of visits to private specialists has declined under all three county councils. Thus, the differences among the county councils in terms of the reimbursement based on coverage rate do not indicate corresponding differences in the distribution of overall outpatient care utilisation. This applies both to groups needing healthcare and the population as a whole.

  • More patients were positive than were negative about changes in accessibility since the introduction of healthcare choice. This applies both to patients with major expected healthcare needs and others.

Around half – both of patients with major expected healthcare needs and the rest of the population – experienced no change following the implementation of the healthcare choice reform. A greater proportion of those who have experienced a change were positive.

It is interesting to note that there is broad support for the option to choose a provider in primary care. Seventy-five per cent regarded the option to change healthcare provider as a good thing, while only 2 per cent took a negative view. The results of the patient questionnaire show that patients with major expected healthcare needs have a similar impression of how accessibility to primary care has developed since the reform to the rest of the population. This is in line with the results of healthcare utilisation, where we are unable to find any clear signs of the displacement of groups needing healthcare.

Another interesting observation from the patient questionnaire is that the rest of the population considers geographical proximity to the surgery to be a more important explanation for a change of surgery than do those in groups needing healthcare. In the group of people with a diagnosis that requires healthcare, on the other hand, various quality attributes are considered more important. In addition to this, compared with Skåne and Östergötland, all groups in Stockholm indicate to a greater extent that, following the introduction of healthcare choice, more visits to the surgery are required to meet the medical needs.

  • People with a higher income indicate to a greater extent that they are more satisfied with developments in primary care availability following the healthcare choice reform.

From an equality perspective, the results of the patient questionnaire indicate that those with a higher income, irrespective of the healthcare diagnosis and the county council to which they belong, tend to be more satisfied with the various aspects of developments in primary care since the introduction of healthcare choice. This indicates a need for more in-depth knowledge of the horizontal fairness of healthcare, i.e. that all people with the same healthcare needs, irrespective of their socioeconomic circumstances, should have the same access to healthcare. The aim is therefore to examine in particular, in a further study of the county councils involved, any differences in healthcare utilisation that may be traced to socioeconomic circumstances.

  • The design of management and reimbursement models is, to an extent, in line with this development.

As well as changes in the supply and demand for primary care services, the design of the reimbursement systems also affects healthcare providers’ incentive to generate healthcare contacts. In the long run, this will also affect developments in consumption. From the design of the study, it is not possible to ascertain the extent to which the reimbursement systems have affected developments in the respective county councils. Instead, in this regard we must be content to discuss how well the results tally with those that we might expect on the basis of the different paths chosen by the county councils.

The county councils’ healthcare choice models show many similarities in terms of the freedom to set up, licensing, freedom of choice and listing. The major difference relates to the reimbursement system8. In relation to the others, Stockholm has a high proportion of variable reimbursement based on visit types, which in turn are differentiated according to professional category. For Stockholm, the variable reimbursement based on visit types amounts to 60 per cent, while it is less than 20 per cent for the other county councils. Stockholm also had a relatively high proportion of variable reimbursement in the years before the introduction of healthcare choice.

The reimbursement per listed individual is also designed differently. In Stockholm, the reimbursement is weighted according to the age of the listed individuals at each surgery. For Östergötland, reimbursement is based not only on age but also on the income circumstances in the surgery’s geographical area. In Skåne, reimbursement is also based on indicators of morbidity in individuals and the socioeconomic circumstances of the listed population. Other differences are that the surgeries in Skåne and Östergötland have greater financial incentives for ensuring that the listed individuals have their outpatient consumption at their own surgery (i.e. a higher proportion of the reimbursement is linked to the surgery’s so-called coverage rate).

The strong increase in visits observed for doctors in Stockholm is in line with the fact that the proportion of the performance-based reimbursement has increased somewhat in conjunction with the reform. The idea that the reimbursement system may have played its part is supported by the continued increase in visits to doctors despite the fact that, from a national perspective, the number of visits per individual was already high before the reform was introduced. The differentiation between the reimbursement to doctors and nurses, where the reimbursement to doctors has been given greater weight, is also in line with the fact that the number of visits to doctors has risen, while the number of contacts with nurses has remained relatively constant.

The small increase in visits in Östergötland is also expected, given that when reimbursement is predominantly based on the number of listed patients there is less incentive to generate healthcare contacts. The increase in the number of primary care providers is also much lower than in Stockholm and Skåne. The differences between developments in healthcare contacts among doctors and nurses are also much smaller than in Stockholm. In Östergötland, the total number of nurses’ visits is increasing somewhat more than the number of visits to doctors. When making a comparison of differences in the numbers of visits in different professional categories, it is also important to point out that these can be traced to various historical traditions among the county councils. They are also affected by the recruiting opportunities of the county councils.

The increase in all healthcare contacts in Skåne cannot, relative to the other county councils, be traced to any strong incentives to generate healthcare contacts as there is no direct visit-related reimbursement system in Skåne. On the other hand, part of the increase probably relates to the fact that the so-called coverage rate is significant, i.e. that a certain accessibility in the form of appointments must be guaranteed for the population, as otherwise reimbursement to the care unit is reduced.

  • It is not currently possible to monitor the effects of the healthcare choice reform on the equality of healthcare provision and report on differences between privately and publicly run units under all of the 21 county councils/regions.

Diagnosis-based and other, more far-reaching, comparisons of the design of healthcare contacts and the distribution at patient level are vital if we are to be able to assess whether the healthcare service is meeting its objectives of equal opportunities in healthcare.

Developments should therefore be monitored over time and compared among the principals responsible. There is consequently a central need for uniform, comparable information among principals. Based on the results of replies from the county councils to our enquiry documents, we consider that it would only be possible to conduct a follow-up study like the one carried out in this report in around half of the county councils.

Experiences from the study conducted, and previous analyses of conditions in the county councils, point to the importance of speeding up the establishment of a health data register that will create opportunities to describe and analyse developments in primary care at a level that is at least equivalent to that in specialised healthcare.

Furthermore, it is essential for county councils to be able to monitor all publicly financed healthcare, i.e. including activities conducted in the private sector.

Experiences from the study conducted show that this has not been possible in all cases. For example, there is currently no information on how one particular central production factor in primary care – the number of practising general practitioners – has developed over time. This information is available only in respect of personnel employed by the county councils. All producers who receive public sector financing must therefore help by contributing comparable information on the structure of healthcare (e.g. material and personal resources), healthcare processes (e.g. healthcare contacts and diagnoses) and healthcare results (e.g. the results from clinical quality registers and patient satisfaction). All in all, there is a pressing need to create the conditions for homogeneity in the information and to create opportunities to exchange data in a way that allows monitoring of developments over time, not just nationally but also at regional level.

Deficiencies in our study and options for future studies

In order to allow a better assessment of whether the increase in the number of visits, as identified following the healthcare choice reform, can primarily be regarded as one driven by supply or demand, it would be desirable to have access to information on first-time and follow-up visits. However, such an analysis is a demanding one and could not be fitted into the timescale of this study. It is also not possible to measure the quality or content of primary care efforts purely on the basis of information about the number of healthcare contacts. For example, there is currently no information about the length of visits, which is part of the development in volume. We have tried to compensate for this situation by examining developments in patients’ perceptions of this aspect following the healthcare choice reform. It is also important to point out that the results may be affected by the fact that the reform was introduced at different times, and that the one in Stockholm had been in use for longer than was the case in the other county councils. In this context, it is also vital to point out that a number of factors, in addition to the design of the healthcare choice models, may also be important to this development. One example is differences in registration, both among county councils and in that registration in general has perhaps changed over time and even improved with the introduction of the system of freedom of choice. In similar studies in future, it will be important to illustrate the situation before the introduction of the reform, as the management of primary care among the county councils had a different appearance for a long period of time.

RECOMMENDATIONS

• It is absolutely fundamental that it should be possible to follow up and compare efforts in primary care. The government should therefore gather its forces around the current development of nationally collected, comparable information on primary care. A possible starting point would be for the existing statutory patient register to be supplemented by information from primary care. The information should include, among other things, the content and duration of visits. It is of particular importance to ensure that the county councils’ need for comparable information on the quality of all publicly financed healthcare should be met, irrespective of whether the activities are conducted in the public or the private sector. The information should include the quality of healthcare in terms of structure, process and results. The prerequisites for this must be created through the necessary statutory changes.

• Developments in visits under the county councils can, to a certain extent, be traced to the design of the reimbursement systems. An important task for the county councils is therefore to strike the right balance between the various targets that the activity aims to stimulate, as different choices of path lead to certain targets being prioritised above others. It is important, therefore, to continue working on the development of need-weighted reimbursement, one that offers the possibility of putting the premium on both high production and equality of healthcare utilisation. Need-weighted, variable reimbursement is an important supplement to need-weighted reimbursement per listed patient. At the same time, it is important that the reimbursement system should be viewed as complementing the other management systems.

FURTHER WORK IN THE AREA BY THE SWEDISH AGENCY FOR HEALTH AND CARE SERVICES ANALYSIS

Although the results in the report do not show any clear signs of displacement of groups needing healthcare, there is a need for a more in-depth analysis of the various forms of healthcare and of possible inequalities based on socioeconomic factors rather than diagnosis of illness. This therefore applies to horizontal fairness to a greater degree than to vertical fairness. The Swedish Agency for Health and Care Services Analysis therefore intends to carry out further work by studying whether people with the same healthcare needs receive the same level of healthcare irrespective of their socioeconomic circumstances.