چهار دهه از اقتصاد بهداشت و درمان از طریق یک لنز کتابسنجی

In this paper, we take a bibliometric tour of the last forty years of health economics using bibliographic “metadata” from EconLit supplemented by citation data from Google Scholar and our own topical classifications. We report the growth of health economics (we find 33,000 publications since 1969—12,000 more than in the economics of education) and list the 300 most-cited publications broken down by topic. We report the changing topical and geographic focus of health economics (the topics ‘Determinants of health and ill-health’ and ‘Health statistics and econometrics’ both show an upward trend, and the field has expanded appreciably into the developing world). We also compare authors, countries, institutions and journals in terms of the volume of publications and their influence as measured through various citation-based indices (Grossman, the US, Harvard and the JHE emerge close to or at the top on a variety of measures).

According to Google's Books Ngram Viewer, the terms “health economics” and “Health Economics” started appearing in books only in the 1940s, four decades after the terms “agricultural economics” and “international economics” surfaced. The first two decades of health economics were slow, but in the mid-1960s use of the terms “health economics” and “Health Economics” increased sharply, and the growth has mostly continued since then. In this paper, we take a bibliometric tour of the last forty years of health economics. Our bibliographic “metadata” are drawn from EconLit which dates from 1969 although it includes many earlier classic papers that were reprinted as chapters in collections published after 1969, including for example Arrow (1963). Unlike other databases such as the Social Science Citation Index, SCOPUS, and Medline, EconLit allows for a relatively clean definition of health economics based on EconLit's inclusion criterion (“a substantial economics content”) and a publication's JEL codes. We run risks of omission and commission:there are many non-economics and multi-disciplinary journals that have published many papers by health economists, and some authors may not have chosen a health JEL code despite their paper having a substantial health content; conversely, our net will catch some items that are devoid of economics (and not written by economists), as well as items where health is not the main focus of the publication. In any event, for reasons to be explained below, we had little choice. And our pragmatic definition of the sub-discipline enables us to avoid many esoteric considerations regarding, for example, the emphasis to be placed in characterizing health economics as the topics its practitioners study (e.g. financial aspects of health services) rather than the discipline (viz. economics) characteristically applied to understanding and explaining phenomena in health and health care (Culyer, 1981 and Williams, 1979). Our EconLit metadata include author, title, journal, the year of publication, author's institution, country of focus, keyword, and three-digit JEL code. EconLit does not include any citation data; for 78% of records in our dataset we were able to obtain citation data from Google Scholar. In addition to getting at influence through citations, we also want to say something about the topics that health economists work on. The JEL sub-categories are not especially illuminating1 and the keywords in the EconLit metadata are freely chosen by authors and unsurprisingly are highly heterogeneous. We therefore used our own topical classification system, and assigned a topic to a subset of our records manually. We aggregate our metadata to the author, institution and journal levels, showing which of each have published the most, and which of each have had the greatest impact in terms of citations. We show publication counts for both “country of origin” and “country studied”. We present frequency distributions of the words used in titles and abstracts and keywords. We also show how the subtopic focus has changed over time and varies across the top institutions. Two previous exercises are worth mentioning, although neither comes close in terms of comprehensiveness and ambition. Rubin and Chang (2003) undertook a bibliometric analysis of health economics articles over the period 1991–2000 using EconLit metadata. They also report author rankings (using number of articles and number of pages) but in most other respects their analysis is much more limited, covering a shorter time period, excluding citations, using the limited JEL topic classification system, and omitting many of the cuts of the data that we present. Also of note is the exercise Culyer undertook as background for his four-volume collection of 78 articles on health economics (Culyer, 2006). This was based on a survey of teaching members of the International Health Economics Association (iHEA) in 2004 that asked for their student reading lists, the objective being to ensure representation in the 2006 collection of those articles most frequently cited for student reference by people teaching the subject. His poll of iHEA yielded usable responses from 112 members and 164 articles that were included on at least two reading lists. We use this source as basis for making some intertemporal comparisons between health economics then (2004) and now (2011), and also for comparing the frequency of citations in general (in 2011) with literature recommended by specialist health economics teachers (in 2004). The rest of the paper is organized as follows. Section 2 presents our methods, including the bibliometric methods used to analyze citations and the classification of health economics subtopics. Section 3 introduces our data. Section 4 presents our findings, and Section 5 our conclusions.

We have focused on the past 40 years. Health economics did, of course, exist before 1971 as we have already noted. Here we reflect on what went on before our starting date and offer some comments on the state of the subject. In his Foreword to Uncertain times: Kenneth Arrow and the changing economics of health care (Hammer, 2003), the 40th birthday Festschrift celebrating Ken Arrow's 1963 classic, Mark Pauly nicely observed that this article “made research in health economics respectable. But it did more than that. It made it interesting.” He went on to identify two attributes of the article that laid the foundations for health economics: “It showed how some behaviors in medical markets could be brought within the purview of standard economic models of competing, maximizing agents … (and) offered an explanation that atypical institutional arrangements in medical care markets are a reaction to special features of this market. … In doing so, it discussed concepts that made (and make) economists attentive but uncomfortable, like trust and morals” (pp. vii–viii). Prior to “Uncertainty” and, indeed, for some while after its appearance, the literature of health economics was exceedingly thin and more descriptive than analytical. The first textbook in health economics (Klarman, 1965) was in part a plea for more attention from economists: “Few economists work actively on the problems of the health field” (p. 10). The bibliography in this book has remarkably few items, by today's standard, by health economists—nor, indeed, by economists of any hue. Much the same was true in 1972, when Victor Fuchs published his celebrated Who shall live? (Fuchs, 1972) whose bibliographies still contain few contributions by economists. Much was made of listing the characteristics of health and health care that seemed to mark it out from other goods and services, often with the implication that the mere recitation of (qualitative) characteristics provided sufficient grounds for public intervention. An early pioneer in this vein was Selma Mushkin (1958) who also, however, laid some of the foundations for health as investment (Mushkin, 1962). Those few economists who did take an interest would scarcely have described themselves as ‘health economists’, typically having much broader research interests: economists like James Buchanan in the US (Buchanan, 1965) and the Jewkeses in the UK (Jewkes and Jewkes, 1961) who gave some ad hoc (and usually critical) attention to the UK's National Health Service. A notable exception was Burton Weisbrod, whose article on human capital (Weisbrod, 1961) was an analytical precursor of Grossman (1972) though not in fact cited by him (Mushkin (1962) was, however, cited). In the early 1960s, as now, the US was dominant. In the UK only two economists spent a substantial amount of their time researching (and none teaching) health economics: Michael Cooper at Exeter and Dennis Lees at Nottingham. None of the aforementioned economists is highly cited in our database. Weisbrod has several items, the highest scoring one coming 80th. Klarman has several, the highest place being well out of the top 1000. Mushkin has but one (a 1999 Spanish lecture) also well out. Buchanan's one and only foray was well out as well, as was Lees. Cooper and the Jewkeses did not appear in the database at all. Sic transit gloria economisti! Another signal talent also appeared on the scene in 1963. This was Martin Feldstein with his article on the application of econometrics to the National Health Service (Feldstein, 1963) which reached its apogee with the publication of his tour de force in 1967 (Feldstein, 1967). This text heralded the birth of health econometrics. Indeed, the book was used in some universities at the time as a text in applied econometrics. Although the article does not feature in our database, his subsequent work on physician services (Feldstein, 1970), hospital cost inflation (Feldstein, 1971) and welfare loss of ‘excess’ insurance (Feldstein, 1973) all pioneered the use of econometric methods—with substantial policy controversy tantalisingly spicing up the mixture. Since then we have charted the course of health economics through its literature. The growth has been astonishing, well outrunning its initial stable-mate, the economics of education. Economists writing on health and health care have plainly become a great deal more specialized than was the case, though some of the most cited contributions have come from economists who would not consider themselves health economists at all. Some topics, such as ‘Health and its value’ and ‘Economic evaluation’, show no clear trend in popularity, while others display a clear trend: the topics ‘Medical insurance’ and ‘Supply of health services’ have lost favor, while the topics ‘Determinants of health and ill-health’ and ‘Health statistics and econometrics’ show a clear upward trend. While not analyzed in our paper, Rubin and Chang's (2003) analysis of a subset of our data confirms our suspicion that the frequency of joint authorships has increased over time; we speculate that much of that trend is due to multi-disciplinary collaborations. The geographic focus of health economics publications has broadened considerably over the last four decades, though central Asia, Africa, and the Middle East remain relatively under-researched by health economists. There are few surprises amongst the most cited names in health economics—though the order is much determined by the choice of index. The ten most cited health economists are Cutler, Gruber, Newhouse, Pauly, Viscusi, Currie, Grossman, Sloan, Wagstaff and van Doorslaer. All citation indices probably tend to discriminate against those who go against the grain, which probably accounts for the relative lack of prominence of economists like Robert Evans, the father of ‘supplier-induced demand’ (Evans, 1974) and numerous well-aimed Canadian diatribes against his southerly neighbor. The highest scoring institutions (according to the h-index) are Harvard and the World Bank followed by MIT, Berkeley and Chicago, Pennsylvania, and Michigan and York (UK). Harvard is the least specialized of the top institutions in terms of topics studied; the World Bank and York are more specialized. The US dominates health economics, being home to 38 of the top 50 institutions. The UK has seven top-50 institutions, and Canada two. The most influential journal (in terms of aggregate citations) is the Journal of Health Economics followed by Health Economics. The Handbook of Health Economics and the Quarterly Journal of Economics are better cited, however, than either of these two leading journals in terms of the citation rates. Before closing we should remind the reader that our publications are drawn exclusively from EconLit for the reasons explained in Section 3. For the same reasons, we cannot get a systematic assessment of the bias this causes, but what we can do is see—for a subset of authors—how much we are missing of the health economics corpus by focusing exclusively on EconLit. Using Harzing's (2010)Publish or Perish software, which is essentially a user-friendly ‘frontend’ to Google Scholar, we have analyzed the publications and citations of the top 11 health economists in Table 5 (two tied for 10th position). Some of the ‘missing’ publications we unearth are simply in other areas of economics; many are, in fact, in EconLit but we did not include them in our database because they do not have a health JEL code.22 But several are about health. Cutler's fourth most cited publication is an article in Health Affairs on technological change in medicine (Cutler and McClellan, 2001). Sloan's most cited paper is a Northwestern University Law Review article on valuing life and limb in tort (Bovbjerg et al., 1988), and his second and third most cited publications are JAMA articles on medical malpractice ( Hickson et al., 1992 and Hickson et al., 1994). Wagstaff's second and fourth most cited publications are articles in Social Science and Medicine and The Lancet ( Victora et al., 2003 and Wagstaff et al., 1991). Pauly's second and fourth most cited articles are in JAMA and the New England Journal of Medicine ( Hillman et al., 1989 and Naylor et al., 1989). Van Doorslaer's most cited publications include articles in Social Science and Medicine and the Annual Reviews of Public Health ( Wagstaff et al., 1991 and Wagstaff and van Doorslaer, 2000). Chaloupa's highly cited articles include an article in the British Medical Journal (Wakefield et al., 2000), and three of Newhouse's eight most cited articles appeared in the New England Journal of Medicine ( Brennan et al., 1991, Leape et al., 1991 and Manning et al., 1984). Many of these articles probably do belong to the field of health economics. Whether all do is less clear. The controlled trial by Manning et al. (1984) of a prepaid group practice clearly belongs. By contrast, the other two Newhouse papers (Manning et al., 1984) are clearly interdisciplinary in nature, and are described as such in their abstracts. But even if we were to agree that a ‘missing’ paper is part of the health economics corpus, it is not straightforward to decide how influential it is, relative to articles published in EconLit-indexed journals. Different fields have different citation rates reflecting different practices and different readership volumes ( Alonso et al., 2009 and Hirsch, 2005). Medical journals, in particular, have higher citation rates than economics journals. For example, according to Google Scholar, the most cited 1990 JAMA article has been cited 4446 times as of February 17, 2012, while the most cited 1990 AER paper has been cited just 2101 times. So some of the ‘missing’ articles may indeed be cited a lot; but this may reflect—at least in part—the higher citation rates among medical journals. What we can probably safely conclude is that the top 11 health economists—and probably many other health economists too—have also published several articles in medical and law journals that have been influential among a broad audience, and probably among economists too. We leave open the tantalizing question of whether our reliance on EconLit biases our results in such a way as to deny anyone entry into the top ten! Standing back from the bibliometrics, an interesting question might be whether health economics has fulfilled the promise diagnosed by Pauly in his reflections on Arrow's pioneering piece. ‘Respectable’? Apparently so judging from the thorough permeation of policy circles by health economists and the ample flow of research funding (at least to those deemed ‘respectable’!). ‘Interesting’? Certainly, and not least because of the challenges that have had to be faced up to in adapting economists’ ‘standard economic models’. The question cries out “When is the standard ‘purview’ being embraced or rejected?” Our estimate is that, if the standard is taken as that of “competing, maximizing agents”, there is much in health economics that challenges the usefulness of competitive models, or maximizing models, and that uses alternative maximands, especially in normative analyses. Health economics has yet, we think, to face up to many of the ethical dilemmas in health economics—how, for example, to embody in non-arbitrary and non-partisan ways some of the ‘moral’ issues mentioned by Pauly, especially in health care investment decisions and the design of insurance systems. Likewise, applied health economists have not really faced up to the implications, both positive and normative, of the assault by cognitive psychologists and experimental economists on the utilitarian (especially expected utility) foundations of the economics core. But in this they probably cannot be held to be more accountable than applied economists of any kind.