Health Technology Assessment (HTA) was originally designed more than 25 years ago with the purpose of assessing the consequences of medical technologies and was primarily based on the synthesis of existing evidence (22). HTA has its beginnings in the United States, with the establishment of the Office of Technology Assessment in 1972, and has since then spread to Canada, Australia, and Europe, with Sweden and the Netherlands as some of the first HTA agents, and later to many other countries around the world (1;3;19). As of today, a total of forty-two HTA institutions from twenty-one countries are members of the International Network of Agencies for Health Technology Assessment (INAHTA), and the number of members of Health Technology Assessment International (HTAi) after 1 year in service is sixteen organizations. These numbers alone indicate a high level of activity in HTA internationally in 2004.

From the beginning, HTA has been defined by its purpose not by its methodology (16) and the primary method of assessment was review of existing evidence. The reviews were later developed into systematic reviews, and consensus methods were used (13). Today, systematic reviews are still the most-often mentioned method of assessment when looking for descriptions of HTA methods in the literature (4;8;14). Primary research is also included as constituting HTA methodology but most often as a supplement to systematic reviews when evidence is lacking in some areas (21;23).

Despite this uniformity in descriptions of the underlying methods of HTA, time-trend comparisons may well be warranted to disclose any general trends in the practical applications of HTA. Some authors have pointed to the need for methodological developments in HTA (19), others have noticed variations in HTA methodology between HTA agents (15). Several authors see such variations as a potential problem (23;26), and the need for international collaboration in dealing with and defining HTA has been stressed (1;18).

International collaboration has taken place in terms of the EUR-ASSESS project in the late 1990s (16), the ECHTA/ECAHT project in early 2000s (4;15), and the INAHTA initiatives in the early 2000s (9). All of these projects have focused on HTA methodology, and attempts have been made to reach consensus about the content of HTAs as well as the methods applied in HTA. Attempts at the latter have had limited success, most likely because HTAs are applied in different countries with different (political) cultures and because HTAs rely on methods from different scientific fields. Some of the participants involved in the aforementioned initiatives conclude that HTA methodology cannot be standardized and that one methodological approach to HTA does not exist (16;26). It is noteworthy that the INAHTA Checklist developed with the purpose of improving the standards of HTA focuses mainly on the content of HTA, whereas methodological issues are almost absent.

Considering the speed of development within the scientific fields that underlie HTA, the methodologies applied in HTA are likely to have changed over the past decades. Within the past 25 years, there has been a considerable growth in HTA activities due both to the expansion in number of HTA agents and to the production levels of these agents (6). The scope of HTA has also widened from primarily addressing effectiveness and safety issues to covering a broader array of issues such as psychological, organizational, ethical, and legal aspects (6;14;23). All of this gives rise to questions like Have the methodological developments in general spread to HTA methodology? Who are the assessors of the increasing number of produced HTAs? And what type of technologies are the subjects of the increasing numbers of assessments?

The objective of this study is to explore a possible evolution in HTA methods, in type of assessed technology and in type of assessors during the period 1989–2002 to disclose the extent to which the original focus of HTA may have changed and to disclose possible further developments in HTAs worldwide. To these authors' knowledge, no such study of the international time-trends of HTA has been published previously in the literature. Only Menon and Topfer (20) focused on time-trends, but the scope of their study was restricted to analysis of 117 government-funded Canadian HTAs. A few other studies exist that focus on HTA methodology, but none of them have included a time perspective. Perry et al. (24) performed a study of HTA reports originating from 103 HTA institutions in 24 countries. The weakness of this study was that it was not based on actual review of the HTA report but relied on the HTA institutions answers to a questionnaire. Furthermore, the authors applied a very broad definition of HTA institutions to also include medical societies, for-profit organizations, and trade associations. Mears et al. (18) also relied on questionnaire data but restricted the respondents to fifty nonprofit and/or government-financed institutions. Poulsen (25) carried out an actual review of HTA reports produced from a sample of HTA institutions applying inclusion criteria similar to these presented in this study, but the time period was limited to 1989–1996, and no time-trend analysis was performed. Finally, García-Altés et al. (7) reviewed sixty-one HTA reports but restricted their sample to four HTA organizations, restricted their study period to 1999–2001, and had no time perspective in their analysis.

RESULTS

The Production of HTA Reports in 1989–2002

Table 1 displays the number of HTAs by institution in the period 1989 to 2002. NCCHTA in the United Kingdom is by far the most productive HTA institution in the sample, publishing almost 30 percent of the total number of HTAs.

Figure 1 shows the number of HTAs by year of publication and illustrates the significant increase in the number of HTAs during the time period 1989–2002. A total of 52 percent of the reports in the sample are published in the period 2000–2002, which means that almost the same share of HTAs in the sample is published in the past 3 years as in the first 11 years of the study period. This development reflects an increase in the number of institutions producing HTA reports as well as an increase in the number of reports produced by institution.

Number of health technology assessment reports by year of publication

Type of Assessed Technology

During the study period, health procedure is consistently the most assessed type of technology (Table 2), constituting more than 50 percent of all assessed technologies in all four time periods. Assessments of medical devices were salient in the early 1990s and constituted more than 30 percent of the assessed technologies in the period before 1996. After this period, the share decreased. Conversely, the fraction of HTAs with a focus on pharmaceuticals was very low in the early time period, but since the year 1996, the fraction has increased to more than 1/4 of all assessed technologies.

Type of Assessors

Over the study period as a whole, the HTAs are equally frequently performed by the HTA institution's own personnel, external partners, and a combination of the two (Table 3). However, over time, a major shift has taken place. During the first 6 years of the study period, almost all of the HTAs were assessed by the HTA institutions' own personnel (around 90 percent). External partners and joint assessments were only sporadically observed, but from 1996, the use of external partners became more common, and from around 1999, joint assessments became frequent, leaving less than 10 percent of the HTAs assessed by their own personnel alone.

Methods of Assessment

Table 4 shows the frequencies of seven different types of methods of assessment applied in the 433 HTAs. The number sums to a total of 682 observations because more than one assessment method could be used in one HTA and/or the same assessment method could be used more than once.

In accordance with the original definition of HTA, Table 4 shows that HTAs still predominantly entail synthesizing existing evidence with literature review as the most common method of assessment. This method counts 58 percent of all assessments and is prevalent in 92 percent of HTAs, a fraction that is relatively stable during the whole study period.

Meta-analysis, representing another form of secondary research is used in only 3 percent of the HTAs and accounts for only 2 percent of all assessments.

Surveys account for 7–10 percent of the individual assessments and are present in 15 percent of the HTAs. The prevalence of this method of analysis appears to be relatively stable over the study period. The first RCT included as part of an HTA was carried out in 1999, but more have been observed in latter years. There is also a tendency toward a greater prevalence of evaluations or evaluations based on modeling, ranging from 0 percent in the first time period to 8 percent in the last time period.

Economic evaluations have been applied from the beginning of the study period and account for 14 percent of the individual assessments in the early time period, but the role of economics in HTAs has not increased over the time periods. Rather, the relative frequency of this type of assessment is characterized by a slight negative trend.

The term other method consists of a range of alternative assessment methods and has decreased over the time period from around 20 percent to around 10 percent, signifying greater uniformity in the methods applied in HTA.