The impact of a reimbursement rate reduction on the utilization of antiulcer, antidepressants and antidiabetics in Portugal: A time series analysis

ABSTRACT

Background

According to international research, an increase in the patient share of pharmaceutical expenditure results in a decrease in medicines utilization. In 2010, the Portuguese government reduced the reimbursement rate for certain therapeutic classes. The aim of this study was to evaluate the impact of raising cost to the patient on the utilization of medicines.

Methods

Between January 1996 and December 2015, medicines utilization and cost to the patient per DDD of antihypertensive, antidyslipidemic, antidiabetic, antiulcer and antidepressant medicines were compared. A segmented linear regression of two time series – before and after the change in reimbursement percentage – was used.

Results

During the two time series, there was an increase in the utilization of antidepressants and antiulcer medicines. The consumption growth decelerated following the reduction in reimbursement rate. Cost to the patient decreased for both classes across the two time series, although the trend accelerated during the second.

Conclusion

These findings suggest that a decrease in the reimbursement rate had little impact on the utilization of medicines. It is reasonable to assume that the reduction in percentage covered by the Portuguese National Health System was gradually compensated by the decrease in the absolute amount that patients paid for medicines..

KEYWORDS:

• Pharmaceutical policy
• consumption
• reimbursement
• out-of-pocket
• generics
• co-payment
• patient contribution
• time series analysis

1. Introduction

Governments are the world's largest payers (third-party payer) of healthcare costs through their health-care systems [1,2]. A rise in public expenditure on outpatient medicines led to the implementation of legislative measures aiming to control costs and gain efficiency. Medicines utilization efficiency refers to ensuring that each patient obtains and takes the right medication for their condition in the right formulation, dose, and time frame [3].

One of the most common reforms implemented in Europe between 2010 and 2015 was increasing patient contributions. This was done by reducing the number of citizens exempt from pharmaceutical co-payment and/or decreasing the percentage of pharmaceutical expenses covered by the government or health insurers [4].

Increasing users’ financial responsibility (out-of-pocket) for prescription medicines may help to manage healthcare costs. It may encourage patients to choose more affordable care and also serve as an incentive to discourage unnecessary or inappropriate medicines utilization [5]. However, it may also prevent access to necessary care [6].

The terminology used by different health systems may differ in meaning. Out-of-pocket expenses, meaning those incurred directly by the patient, can include direct patient payment policies such as caps (the maximum number of prescriptions or medicines that are reimbursed), fixed co-payments (patients pay a fixed amount per prescription or medicine), coinsurance (patients pay a percentage of the price), ceilings (patients pay the full price or part of the cost up to a ceiling, after which medicines are free or available at a reduced cost), and tier co-payments (differential co-payments usually assigned to generic and brand medicine) [7]. In Portugal, a coinsurance model is used, and in this study, ‘out-of-pocket’ means the amount paid by patients for medicines at pharmacies that is not covered by the Portuguese National Health System (PNHS) [8]. Unlike in other European countries, there is no out-of-pocket payment in the form of a fixed amount to be paid for a medicine or prescription [9]. Instead, the patient pays the difference between the full price of the medicine and the percentage of the price that is not reimbursed/subsidized by the public payer.

The Pharmacy Retail Price (PRP), or Public Price, is the price of a medicine at the pharmacy level. In general, reimbursement by the PNHS for prescription medicines is defined as a percentage of the PRP, which is set in one of four tiers (15%, 37%, 69%, 90%) [10]. It means the PNHS supports a portion of the PRP, while the patient pays the remainder. The Ministry of Health specifies the pharmacotherapeutic categories that correspond to each tier of reimbursement. The reimbursement tier in which medicines are included is determined by the diseases for which they are indicated. The reimbursement tier rises in accordance with the priority the NHS assigns to the treatment of a certain disease. In addition to this general regime, medicines may be included in special or exceptional reimbursement regimes [10].

In 2010, the Portuguese government reduced reimbursement rates for some therapeutic classes, increasing the patient share, and consequently the cost to the patient. In the general regime, among the top five therapeutic classes, there was a reduction from 69% to 37% for antiulcer medicines and from 95% to 90% for oral antidiabetics. A special regimen for antidepressants was abolished, with the practical consequence of a reimbursement rate reduction from 69% to 37% [11]. Antihypertensives and antidyslipidemic medicines were unaffected by the change in reimbursement policy [11].

An increase in cost to the patient is likely to lead to a decrease in the use of essential medicines [12], although most patients are not particularly sensitive to changes in out-of-pocket cost [13].

This paper investigated whether a decrease in reimbursement rates in Portugal resulted in a decrease in the utilization of antiulcer, antidepressant and oral antidiabetic medicines.

2. Methods

2.1. Data

The study examined outpatient medicines utilization (also referred to as ‘consumption’), expressed as DDD per million inhabitants per month, between January 1996 and December 2015.

The monthly utilization and cost to the patient per DDD of antiulcer, antidepressant and oral antidiabetic medicines prior to the new reimbursement rate were compared to the same indicators following the change.

To minimize the effect that potential extraneous factors could have on the researched outcome, antihypertensives and antidyslipidemic medicines were used as comparators, as they were not affected by the decrease in reimbursement rates. Therefore, the study was conducted for the top five therapeutic classes, which were either impacted (antidepressants, antiulcer and antidiabetics) or unimpacted (antihyperlipidemic and antihypertensive) by the new reimbursement policy.

The data in standard units (Counting Units) was obtained from Intercontinental Medical Statistics (IMS), now IQVIA, databases Health Pharmaceutical Information Portugal (IFP) [14–16], and converted into Defined Daily Doses (DDD) [17]. Data provided by IQVIA is collected by audits of retail pharmacies’ estimated sales of registered medicines based on their purchases from wholesalers in Portugal, which reflect the purchases of medicines by the Portuguese population in an outpatient context.

The DDD methodology (WHO Collaborating Centre for Drug Statistics Methodology) is a commonly used technical metric. It provides a measure of exposure in a given population, enabling comparisons over different periods of time. The DDD is assigned to each medicine at the fifth level (chemical substance) ATC classification, and it is the assumed average daily maintenance dose in adults for a medicine used for its main indication [17]. DDD method can be used according to many specific denominators, but we defined our standardized unit as DDD per million inhabitants per month.

2.2. Calculation

1. Therapeutic Classes were organized as follows:

   • Antiulcer (A02B);

   • Antidiabetics (A10N, A10C, A10S, A10H, A10J, A10L, A10M and A10K);

   • Antihyperlipidemic (C10A and C10C);

   • Antihypertensives (C09D, C09B, C09C, C09A, C08A, C07A, and C03A).

2. Converting Counting units in DDDs $\mathrm{Number}\mathrm{of}\mathrm{DDDs}\left(\mathrm{per}\mathrm{strenght}\mathrm{and}\mathrm{month}\right)={\displaystyle \frac{\mathrm{TCU}\times \left(\mathrm{DS}/\mathrm{DDD}\mathrm{conversion}\mathrm{factor}\mathrm{field}\right)\left(\ast \right)}{\mathrm{Population}\left(\mathrm{Portugueses}\mathrm{census}\right)}\times 1,000,000}$

where CU = Counting Unit = the smallest common dose in the form of a product and, for example, for oral solid forms, one tablet or capsule is used, whereas for liquid forms, the ampoule, vial or bottle may be used. TCU = Total CU sold per Dosage Strength and month. DS = Dosage Strength; Population = Portuguese Population during the period; the calculation is multiplied by 1,000,000 to convert the population size to ‘per 1,000,000 population’; (*) Using the searchable version of the ATC index from the Collaborating Center for Drug Statistics Methodology database

1. Number of DDD per chemical substance (ATC5) per month $\sum _{i=1}^{\mathrm{\infty }}\mathrm{Number}\mathrm{of}\mathrm{DDDs}\left(\mathrm{per}\mathrm{strenght}\mathrm{and}\mathrm{month}\right)i$

where i = number of strengths per ATC5.

1. Total DDDs (ATC3) per month $\sum _{i=1}^{\mathrm{\infty }}\mathrm{ni}$where i = number of ATC5 per ATC3, n = number of DDD’s per ATC5

2. Total DDDs (Therapeutic Class) per month $\sum _{i=1}^{\mathrm{\infty }}\mathrm{ni}$where i = number of ATC3 per Therapeutic Class; n = number of DDD’s per ATC3

3. Cost to the patient per DDD per Therapeutic Class per month $\mathrm{Cost}\mathrm{to}\mathrm{the}\mathrm{patient}\mathrm{per}\mathrm{DDD}\left(\mathrm{per}\mathrm{month}\right)={\displaystyle \frac{\mathrm{Total}\mathrm{sales}\mathrm{per}\mathrm{TC}}{\mathrm{Total}\mathrm{DDDs}\mathrm{per}\mathrm{TC}}\times \left(1-\text{\%}\mathrm{Reimbursement}\right)}$

2.3. Statistical analysis

To determine the extent of the impact of policy changes for each therapeutic class considered, a segmented linear regression of two time series was used: with data from the periods before and after the reimbursement rate change [18]. No annual seasonality in the time series was assumed. The difference between the two segments can be quantified by measuring the change in level (intercept) and slope. A change in the level between the pre- and post-intervention segments indicates a drastic change, whereas a change in the slope indicates a shift in the growing/ diminishing trend. The slope variation (which can be related to the discrete second derivative) is also computed, to highlight underlying tendencies when both segments, relative to pre- and post-intervention, show an increasing/decreasing trend.

Microsoft® Excel was used to generate the linear regressions.

3. Results

3.1. Medicines utilization

The utilization of therapeutic classes affected by the reduction in reimbursement percentage did not decrease. Following the change, utilization continued to increase, albeit at a slower rate (Figures 1–6). In terms of the most affected classes, the growth trend for antidepressants decelerated by −26.48% (from 429.5–315.87 DDD/month), while it slowed by −1.15% for antiulcer medicines (from 445.8 to 440.7). However, this deceleration of growth was less relevant than for other classes, such as antihyperlipidemic and antihypertensive medications, for which the reimbursement rate was unchanged by the new policy, and antidiabetics, for which the reimbursement rate was only slightly reduced (Table 1 and Figure 7).

Figure 1. Antidepressant utilization (DDD per million inhabitants per month) – January 2006 to September 2010.

Figure 2. Antidepressant utilization (DDD per million inhabitants per month) – October 2010 to June 2015.

Figure 3. Antiulcer utilization (DDD per million inhabitants per month) – January 2006 to September 2010.

Figure 4. Antiulcer utilization (DDD per million inhabitants per month) – October 2010 to June 2015.

Figure 5. Antidiabetics utilization (DDD per million inhabitants per month) – January 2006 to September 2010.

Figure 6. Antidiabetics utilization (DDD per million inhabitants per month) – October 2010 to June 2015.

Table 1. Equation of the regression line.

Therapeutic class	Medicines utilization (expressed in DDD per million inhabitants per month)
	Pre-intervention	Post-intervention
Antidepressants	y = 429,47x − 14 846 774,78	y = 315,73x − 10 429 238,79
Antiulcer	y = 445,77x − 15 597 963,10	y = 440,65x − 15 671 156,42
Antidiabetics	y = 249,24x − 7 939 238,11	y = 162,67x − 4 645 550,43
Antihyperlipidemic	y = 612,20x − 21 784 935,03	y = 440,58x − 15 131 791,23
Antihypertensives	y = 1 182,84x − 36 640 692,55	y = 300,63x − 1 710 082,91
	Cost to the patient per DDD
Antidepressants	y = −0,00032x + 14,02860	y = −0,00076x + 33,02573
Antiulcer	y = −0,00061x + 25,99313	y = −0,00108x + 46,06152
Antidiabetics	y = 0,00011x − 4,22297	y = 0,00006x − 2,00278
Antihyperlipidemic	y = −0,00124x + 53,52561	y = −0,00076x + 33,45489
Antihypertensives	y = −0,00010x + 5,05205	y = −0,00022x + 9,85963

As previously noted, this slope comparison is analyzed to identify underlying tendencies. Figures 7 and 8 illustrate that whereas therapeutic classes show a positive or negative tendency during the first segment, this rate of growth/decrease is considerably altered in some cases for the second segment.

Figure 7. Medicines utilization growth trend.

Figure 8. Patient’s cost per DDD.

3.2. Cost to the patient per DDD

Following the reduction in reimbursement rates, from October 2010 to June 2015, a pre-existing trend in lowering cost to the patient per DDD for both antidepressants and antiulcer medicines intensified. There was a significant drop in cost to the patient per DDD of antihypertensives and no change in reimbursement rates was verified. Consumption changed little for antihypertensives, despite this decrease in cost. The cost to the patient per DDD of antihyperlipidemic medicines slowed its pace of decline, whereas antidiabetic cost continued to rise, but at a slower rate (Table 1 and Figure 8).

These findings are also evident in an examination of the evolution in cost to the patient per DDD for both antidepressants and antiulcer medicines, over the entire time period. A significant increase in total cost to the patient per DDD could be estimated by October 2010, although this was followed by a decline during the following months, with the amount returning to levels closer to (or lower than in the case of antiulcer medicines) those before the reimbursement change (Figure 9).

Figure 9. Cost per DDD Antidepressant and Antiulcer medicines (January 2006 to June 2015).

4. Discussion

This study showed that the decrease in the reimbursement rate did not reduce the utilization of antidepressants or antiulcer medicines in Portugal. Using the same classification as Sinnott et al [19], both essential (antidepressants) and non-essential (antiulcer) medicines, were evaluated. In Portugal, patients pay a percentage of the medicines’ price, and the patient share of pharmaceutical costs increased for both. The utilization continued to grow although at a slower rate.

Previous studies produced contrasting results, showing that an increase in the cost to the patient was likely to lead to a decrease in the utilization of these therapeutic classes: antihyperlipidemic [20,21], antihypertensives [4,22], non-insulin and insulin antidiabetic medicines [4,23,24], and antidepressants [19]. A systematic literature review on the effects of payments for pharmaceuticals also suggests that the higher the out-of-pocket burden, the fewer prescriptions are filled [25].

Keeping in mind that direct patient payment policies can vary, some of this research focused on health systems where patients pay a fixed amount per prescription or medicine (fixed co-payments), while in the Portuguese model patients pay a percentage of the price (coinsurance) [8]. Whatever the model, a rise in out-of-pocket expenses means a rise in the patient's burden, and any potential links to shifts in medicines utilization can be assessed:

• – In Finland, where there are ceiling mechanisms in place to protect patients from excessive payments, an increase in patient co-payment for each purchased type 2 diabetes medicine was associated with a decrease in consumption [23,24].

• – In Italy, a revision of the payment criteria (limited to high-risk users for primary CVD prevention and diabetic patients for secondary CVD prevention) was associated with an immediate reduction in statin use [20].

• – The implementation of the ‘Euro per prescription’ co-payment in Catalonia resulted in a decrease in consumption of several classes of medicines – including antihypertensives, non-insulin and insulin antidiabetic medicines – particularly among those who had previously received free medicines [4].

• – In Slovakia, where a fixed co-payment is applied when setting the reimbursement rate, changes in the average value of out-of-pocket had a major impact on antihypertensive medicine utilization [22].

• – Prescription medicine co-payments were introduced in Ireland in 2010, and while these were minimal, they were still associated with a decrease in medicine consumption, which was more obvious for less essential than essential medicines. The notable exception was antidepressant medicines (essential medicines), with reductions in adherence of −8.3% [19].

Other studies have also shown that an increase in the patient cost is likely to lead to a decrease in the use of essential medicines [12].

Our data do not support these findings, as consumption in Portugal continued to rise, although at a slower rate than before the reimbursement change. The study results are consistent with Andersson et al., who found that increased co-payments had a minimal effect on pharmaceutical cost and volume [26].

Some studies have demonstrated that patient sensitivity to cost may vary depending on the therapeutic class [27]. According to the present study, the deceleration in growth for both antidepressants and antiulcer medicines was less pronounced than that observed for other classes which had less of a reduction in the reimbursement rate (antidiabetics) or even those whose rate was unchanged by the new reimbursement policy (antihyperlipidemic medicines and antihypertensives). Researchers found that increased cost to the patient had a particularly high influence on the risk of losing treatment adherence among the most inelastic therapeutic groups (cardiovascular and antihyperlipidemic) [28]. We were unable to confirm this difference because the cost to the patient was not increased for either antihyperlipidemic or antihypertensive medicines.

It has been suggested that the complete elimination of cost-sharing, regardless of health status, significantly increases medicines utilization [29] and could lead to overconsumption of medicines [30]. However, our study showed that from October 2010 a significant decrease in cost to the patient per DDD of antihypertensives was accompanied by an almost flat growth trend in utilization.

A study conducted in three European countries found that some specific medical conditions raise the overall out-of-pocket burden, depending on the need for medicines utilization [31]. Among these are heart attacks, high blood pressure, and emotional disorders. This may explain why antiulcer medicines, which are not among the medicines used in the aforementioned disorders, were the least affected by a deceleration in growth after the changes in the reimbursement rate.

When medicines such as antidepressants are relatively inexpensive, even the elderly, the age group with the highest utilization but more commonly low-volume users, do not have high out-of-pocket costs [32]. However, in this case, after reimbursement adjustments, the average out-of-pocket cost per package of antidepressants and antiulcer medicines increased by 7.40 and 9.01 euros, respectively [33].

The change in reimbursement rate caused this peak in patient costs, but the amounts gradually returned to pre-implementation levels. The lower reimbursement rate was compensated by a larger cost reduction during the following months/years. An intensification in the downward trend of patient cost per DDD for antidepressants and antiulcer medicines was observed, when the slopes of the segmented linear regression of the time series were compared.

Governments can use certain measures to minimize the impact of decreasing reimbursement rates on out-of-pocket, such as: administrative price cuts, reduction of pharmacy margins, decreasing the VAT on medicines, promotion of generics use, and delays and hurdles for the introduction of innovative and more expensive medicines [34]. Price reductions combined with a rise in generic medicines’ market share can reduce the amount paid by the user (change in absolute amount vs. change in percentage) [35].

Between 2010 and 2015, the average price of outpatient medicines fell by 21.21% in Portugal [36].

In September 2010, only 1¹ out of 8 antiulcer medicines was patent-protected, and most of the more recent antidepressants² already had generics available [15]. Initiatives were implemented to support the development of generic medicines between 2010 and 2015 [11], and the market share in generics climbed from 17.94% in September 2010 to 41.5% in June 2015 [36,37]. This might partially explain the declining trend in patient spending per DDD.

However, we cannot conclude that without a reduction in cost to the patient per DDD, the medicines utilization would be reduced. Moreover, the cost to the patient for antiulcer and antidepressant medicines took around 2 and 4 years, respectively, to return to pre-reimbursement change levels.

Changes in utilization within each therapeutic class also affected the dynamic in the cost per DDD to the patient. The increase in financial burden may cause patients to use less expensive medicines [38]. For example, the rate of decrease in the cost to the patient per DDD of antihyperlipidemic medicines slowed since molecules (e.g. Rosuvastatin and Pitavastatin) with much higher costs than the class average rose in both consumption and value during the time period. Although data for each molecule was examined, this study focused on the entire therapeutic classes rather than individual molecules [14–16].

5. Conclusions

According to the findings of this study, the legislative changes in reimbursement policy did not result in a decrease in the utilization of antiulcer, antidepressant and oral antidiabetic medicines. Lowering the reimbursement rate temporarily doubled the cost per DDD to the patient, but had little influence on the utilization of antidepressants and antiulcer medicines (or antidiabetic medicines). This conclusion differs from previous research.

The utilization of antidepressants and antiulcer medicines continued to grow, although at a slower rate. Antidepressants and antiulcer medicines’ growth decelerated by −26.48% and −1.15%, respectively. The growth deceleration was, however, smaller than that which was observed for other classes, both those which had no reduction in reimbursement rate (antihyperlipidemic = −28.03% and antihypertensives = −74.58%), and that with only a slight reduction (antidiabetics = −34.73%). The deceleration appears to be due to causes other than an increase in out-of-pocket spending for antidepressants and antiulcer medicines.

During the second time segment, the decrease in the reimbursement rate was further compensated by a larger decrease in cost to the patient per DDD. Price reductions and the increase in generics’ market share may have compensated for the reduction of the reimbursement rate, but these impacts took some time to return cost to the patient to pre-change levels. In any case, legislative changes that increase out-of-pocket costs should consider adjustments in other pillars of the pharmaceutical sector, specifically the generic medicines policy.

The present study had some limitations. The increase of co-payment may cause patients to switch to less expensive medicines, but our analysis evaluated therapeutic classes rather than switches among molecules. The classes were chosen based on the change in reimbursement rate among the top five highest pharmaceutical sales, rather than their cost sensitivity representativeness. The cost to the patient per DDD was defined by the wholesale prices, which are those made available by the IQVIA, and does not reflect the exact cost to the patient. Nonetheless, the calculation of the absolute difference in the average out-of-pocket cost before and after the reimbursement adjustment was done based on the estimated public price. Also, while no seasonality in the time series was assumed, we believe this to be a coherent assumption due to the nature of the conditions treated by these drug classes, (and) because this is, in general, chronic medication. In any case, since each segmented linear regression is relative to almost five years, this effect would be softened.

5.1. Implications for health policies, health-care provision and further research

This research followed an analytical approach from a pharmacoepidemiological observatory to measure the effects of a change in medicines financing policy. Notwithstanding the above limitations, our results suggest that decreasing reimbursement rates may have no relevant effect on medicines utilization, while gradually lowering medicine costs.

Differing from previous studies, this research opens the door for discussion as to whether there is a minimum pricing level above which medicines no longer need to be reimbursed. Many medicines used to treat chronic conditions in an outpatient setting, such as antiulcer, antidepressant and oral antidiabetic medicines, are becoming more inexpensive, and an increase in percentage of out-of-pocket may not have a consequential impact on patients’ total cost of care. If a price threshold could be set at which reimbursement might be reduced or eliminated without endangering the utilization of medicines, public funds could be focused on financing other healthcare technologies. The definition of such a price level should be the subject of future research.

Prior research has found that an increase in co-payment causes a decrease in medicines utilization, which can be related to either more responsible consumption or a decrease in access to medicines [39]. Policymakers should analyze the impact of changes in cost to the patient on: (a) access according to elasticities of demand for pharmaceuticals, and (b) responsible use of medicines (RUM) using validated quality indicators (QIs), for the most used therapeutic groups in the territory.

There is data which suggests that low-income individuals are price-sensitive when it comes to spending on prescription medicines [40]. Our study is focused on global consumption, but it is important to explore how vulnerable populations behave since they may require additional support when out-of-pocket increases [41,42].

It is important to study medicines utilization not only because of the direct impact on specific diseases, but also because if fewer patients take the prescribed medicines, the costs associated with the flow in health facilities may increase. Indeed, the medication may be beneficial to patients suffering from other diseases [43]. As ‘Ockham's razor’ highlights that parsimony leads to more accurate results, in general, we chose an approach for analyzing our data that is as simple as possible. But, if reductions in medicines utilization are found, a comprehensive study of the economic impact of the lower utilization should be conducted.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Correction Statement

This article has been republished with minor changes. These changes do not impact the academic content of the article.

Additional information

Funding

The author(s) reported there is no funding associated with the work featured in this article.

Notes on contributors

António Augusto Donato

António Augusto Donato, graduated in Pharmaceutical Sciences by Faculty of Pharmacy, University of Coimbra (1990), MBA by ISEG, University of Lisbon (2004), PhD program at Faculty of Pharmacy, University of Coimbra (ongoing), Member of the Board of Directors of Pharmaceutical Industry, and teaching experience at Faculty of Pharmacy, University of Coimbra.

Daniel Figueiredo

Daniel Figueiredo concluded his PhD in 2020 where he studied and developed mathematical modeling applied to intracellular environments. During the last years, he started working about pharmacoeconomics, with special attention to market access, in AIBILI, Coimbra, Portugal. Also, he currently lectures in University of Aveiro, Portugal.

Francisco Batel-Marques

Francisco Batel Marques holds a PhD in Pharmaceutical Sciences (University of Wales, Cardiff, UK, 1995) and is currently Associate Professor at the School of Pharmacy, University of Coimbra. Drug Safety and Health Technology Assessment have been his areas of research and services providing. He has been mentor of 8 PhD and 18 MSc thesis and author of 54 peer-reviewed full papers.

Notes

1 Esomeprazole

2 Citalopram, Fluoxetine, Fluvoxamine, Maprotiline, Mirtazapine, Moclobemide, Paroxetine, Sertraline, Trazodone, Venlafaxine