Scientometric Analysis of the Top 50 Most-Cited Joint Arthroplasty Papers: Traditional vs Altmetric Measures

Abstract

Background

Alternative scientometric measures have introduced a novel view of the scientific literature. This study aimed to identify the top 50 most-cited recent articles in the field of knee and hip arthroplasty, characterize their traditional and alternative scientometric measures, and determine the relationship between traditional and alternative scientometric measures.

Material and methods

The 50 most-cited articles with the term “arthroplasty” in the title that were published between 2015 and 2019 were retrieved from the Scopus database. Alternative scientometric parameters such as Altmetric Attention Score (AAS) from Altmetrics bookmarklet (Altmetrics.com) were retrieved. Scientometric variables such as journal impact factor, first author H-index, and keywords were also extracted.

Results

The 50 most-cited papers accrued 7955 total citations, with a mean of 159.10 ± 56.4 citations per article. The overall mean AAS across the papers was 63.4 ± 164.8. The mean first author’s H-index was 23.8 ± 18.9. Papers published in 2017 and 2018 had a significantly higher mean AAS than those published in 2015 and 2016 (35.1 vs 22.5, P = .009). Citation count was weakly correlated with the AAS (correlation coefficient = 0.379, P = .009). Also, AAS had significant correlations with the journal’s impact factor (P < .001).

Conclusion

We found that the AAS was highest in more recently published papers, while citation count had the opposite trend. The AAS was significantly correlated with the journal’s impact factor and citation count, but the correlation is weak. This suggests that the alternative scientometric measures are complementary to, and not substitutes for, complement traditional measures such as citation count and impact factor.

Background

Traditional scientometric measures, such as citation count, journal impact factor, and the H-index, are extensively used to quantify the distribution and impact of the published scientific literature. While such metrics have proven useful, several drawbacks limit their validity and reliability [1,2]. Furthermore, there is a substantial lead time between the publication of a study and the traditional scientometric measures, meaning that there is a delay of 12-24 months from the time a study is published to the time that such metrics detect the utilization and impact of the study [3].

Alternative scientometric measures have been introduced to complement the traditional measures and minimize the lead time [4,5]. These alternative measures are designed to provide a different view of the effect of scientific publications, namely the social media attention and distribution. The alternative scientometric measures have emerged in response to the ubiquitous use of social media and online forums and nontraditional news outlets by the authors, their audience, and the laypeople. The Altmetric Attention Score (AAS) was developed by the Altmetric Institute (Altmetric.com) and introduced in 2010, as a tool to evaluate the attention garnered by an article on online platforms [3,5]. AAS includes Twitter, Facebook, Google Plus, LinkedIn, news outlets, scientific blogs, Wikipedia, Reddit, policy documents, patents, YouTube, Publons and PubPeer, and online reference managers Mendeley and CiteULike. AAS is an overall weighted index of how widely an article is mentioned on social media.

Joint arthroplasty is one of the most common elective surgical procedures globally, and the volume of primary and revision total joint arthroplasty (TJA) has skyrocketed in recent decades. It is important to know the implications of the rapid increase of joint arthroplasty volume for current research in this area [[6], [7], [8]]. The joint arthroplasty literature is an ever-expanding field with a considerable number of published studies and active researchers. The social media outreach of the current arthroplasty literature is largely unknown. Therefore, the purpose of this study was to identify the top 50 most-cited recent articles in the field of joint arthroplasty and characterize their traditional and alternative scientometric measures. We also sought to determine the relationship between traditional and alternative scientometric measures. We hypothesized that the citation count and AAS would not show a strong correlation in the arthroplasty literature, similar to previously reported data in other fields.

Material and methods

Study selection

The Scopus database was searched for articles published in English between January 2015 and December 2019 that had the term “Arthroplasty” in their title. The 50 most-cited papers were selected from this. The Scopus database was selected because it provides a broader and more comprehensive list of journals, papers, and citation numbers. Two reviewers (S.P.M. and M.S.) screened this list of papers to exclude studies not related to hip or knee arthroplasty. Conflicts were resolved by consensus. The same reviewers extracted the Altmetric data subsequently into an excel sheet. This study did not include patient data and was exempt from institutional review board approval.

Data extraction

Publication year and citation count, as well as the journal’s impact factor, first author’s H-index, and keywords, were extracted from the Scopus database and academic-accelerator.com. The citation density, defined as the number of citations per year, was also calculated. Each article was reviewed, and the main topic and study population were identified. The Altmetric data were extracted from the Altmetric bookmarklet (Altmetrics.com) and included AAS and social media coverage. ResearchGate interest score was also recorded for each article from researchgate.net. Data were entered into a Microsoft Excel sheet. All data were extracted during week 1 (May 01, 2022, to November 01, 2022) to minimize missed data. Level of evidence was extracted as same as mentioned in the included studies. If the study does not mention the level of evidence, Journal of Bone & Joint Surgery guidelines were used to assign level of evidence [9].

AAS contribution

AAS is calculated as the weighted sum of the article’s mentions across different media platforms, with a larger AAS indicating a higher level of social media attention. Table 1 summarizes the platforms contributing to the AAS and their corresponding coefficient value. Unlisted platforms do not contribute to the AAS [10].

Table 1.

Media platforms contributing to the Altmetric Attention Score and their respective weight coefficient.

Platform	Weight per mention
News media	8
Blogs	5
Wikipedia	3
Policy documents	3
Twitter	1
F1000/Publons/PubPeer	1
Open syllabus	1
Google+	1
Facebook	0.25
YouTube	0.25
Reddit	0.25

Statistical analysis

Quantitative variables are reported as mean ± standard deviation, and categorical variables are reported as frequency. Continuous variables are compared between 2 groups by the Mann-Whitney test and between >2 groups by Kruskal-Wallis Test due to the nonparametric nature of the data. Data normality was assessed by observing the histogram with a normality curve. For discovering the correlation between continuous variables, the Spearman correlation test was used. The level of significance was set at P < .05. SPSS version 22.0 for windows (IBM, Armonk, NY) was used for all the statistical analyses.

Results

After excluding shoulder [[11], [12], [13], [14], [15], [16]] and disc [10], the 50 most-cited recently published papers were retrieved. All of them were in English, and all were available for a full-text review. Bibliographic data of the selected articles are represented in Table 2. The alternative metrics data are shown in Table 3.

Table 2.

List of the 50 most-cited articles in Scopus, 2015-2019 with “Arthroplasty” in the title.

Rank	Year	Title and reference	First author and Scopus H-index		Journal	2-y-Impact factor 2020	Study design	Country	Subject	Level of evidence	Cited by
1	2018	Projected volume of primary total joint arthroplasty in the U.S., 2014 to 2030 [7]	Sloan M.	9	Journal of Bone and Joint Surgery, American volume	5.284	Retrospective observational	USA	TJA volume projection	3	452
2	2015	The epidemiology of revision total knee and hip arthroplasty in England and Wales: A comparative analysis with projections for the United States. A study using the national joint registry data set [17]	Patel A.	6	The Bone & Joint Journal	5.082	Retrospective observational	UK	TJA volume projection	3	260
3	2015	Predictors of persistent pain after total knee arthroplasty: A systematic review and meta-analysis [18]	Lewis G.N.	28	British Journal of Anesthesia	9.166	Meta-analysis	New Zealand	Complications	3	242
4	2016	Trends and predictors of opioid use after total knee and total hip arthroplasty [19]	Goesling J.	17	Pain	6.961	Prospective observational	USA	Opioid use	2	238
5	2017	Projected increase in total knee arthroplasty in the United States—an alternative projection model [20]	Inacio M.C.S.	43	Osteoarthritis and Cartilage	6.576	Retrospective observational	Australia	TJA volume projection	3	231
6	2015	Comparative Epidemiology of Revision Arthroplasty: Failed THA Poses Greater Clinical and Economic Burdens Than Failed TKA [21]	Bozic K.J.	64	Clinical Orthopedics and Related Research	4.176	Retrospective observational	USA	Epidemiology	3	229
7	2015	Future projections of total hip and knee arthroplasty in the UK: Results from the UK Clinical Practice Research Datalink [22]	Culliford D.	23	Osteoarthritis and Cartilage	6.576	Retrospective observational	UK	TJA volume projection	3	205
8	2017	Patient Dissatisfaction Following Total Knee Arthroplasty: A Systematic Review of the Literature [23]	Gunaratne R.	2	Journal of Arthroplasty	4.757	Systematic review	Australia	Outcome	3	189
9	2016	Early Results of Medicare's Bundled Payment Initiative for a 90-Day Total Joint Arthroplasty Episode of Care [24]	Iorio R.	49	Journal of Arthroplasty	4.757	Retrospective observational	USA	Health-care economics	3	185
10	2017	Current Epidemiology of Revision Total Knee Arthroplasty in the United States [25]	Delanois R.E.	35	Journal of Arthroplasty	4.757	Retrospective observational	USA	Epidemiology	3	183
11	2016	Patient-related risk factors for periprosthetic joint infection after total joint arthroplasty: A systematic review and meta-analysis [26]	Kunutsor S.K.	41	PLoS ONE	3.24	Meta-analysis	UK	Complications	3	183
12	2017	Current Epidemiology of Revision Total Hip Arthroplasty in the United States: National Inpatient Sample 2009 to 2013 [27]	Gwam C.U.	17	Journal of Arthroplasty	4.757	Retrospective observational	USA	Epidemiology	3	181
13	2017	Fully porous 3D printed titanium femoral stem to reduce stress-shielding following total hip arthroplasty [28]	Arabnejad S.	8	Journal of Orthopedic Research	3.494	Instrumentation	Canada	Instrumentation	NA	177
14	2018	Aspirin or rivaroxaban for VTE prophylaxis after hip or knee arthroplasty [29]	Anderson D.R.	68	New England Journal of Medicine	91.245	Randomized clinical trial	Canada	Thromboembolism	1	168
15	2016	Patient satisfaction after total knee arthroplasty [30]	Choi Y.-J.	5	Knee Surgery & Related Research	NA	Narrative review	South Korea	Outcome	5	166
16	2015	Anterior vs Posterior Approach for Total Hip Arthroplasty, a Systematic Review and Meta-analysis [31]	Higgins B.T.	3	Journal of Arthroplasty	4.757	Meta-analysis	Lebanon	Outcome	3	166
17	2015	Patient-reported outcomes after total and unicompartmental knee arthroplasty: A study of 14 076 matched patients from the national joint registry for England and Wales [32]	Liddle A.D.	22	The Bone & Joint Journal	5.082	Retrospective observational	UK	Outcome	3	165
18	2015	Lifetime medical costs of knee osteoarthritis management in the United States: Impact of extending indications for total knee arthroplasty [33]	Losina E.	78	Arthritis Care & Research	4.794	Retrospective observational	USA	Health-care economics	3	165
19	2015	The clinical outcome of minimally invasive Phase 3 Oxford unicompartmental knee arthroplasty [34]	Pandit H.	50	The Bone & Joint Journal	5.082	Retrospective observational	UK	Outcome	2	154
20	2015	Computer navigation for total knee arthroplasty reduces revision rate for patients less than 65 years of age [35]	De Steiger R.N.	28	Journal of Bone and Joint Surgery—American Volume	5.284	Retrospective observational	Australia	Complications	3	151
21	2015	Pain catastrophizing as a risk factor for chronic pain after total knee arthroplasty: A systematic review [36]	Burns L.C.	9	Journal of Pain Research	3.133	Meta-analysis	Canada	Complications	2	147
22	2015	Current failure mechanisms after knee arthroplasty have changed: Polyethylene wear is less common in revision surgery [37]	Thiele K.	9	Journal of Bone and Joint Surgery—American Volume	5.284	Retrospective observational	Germany	Complications	3	147
23	2016	Discharge Destination After Total Joint Arthroplasty: An Analysis of Postdischarge Outcomes, Placement Risk Factors, and Recent Trends [38]	Keswani A.	14	Journal of Arthroplasty	4.757	Retrospective observational	USA	Outcome	3	142
24	2015	Quantifying the Burden of Revision Total Joint Arthroplasty for Periprosthetic Infection [39]	Kamath A.F.	28	Journal of Arthroplasty	4.757	Retrospective observational	USA	Epidemiology	3	142
25	2015	Surgical approach in primary total hip arthroplasty: Anatomy, technique and clinical outcomes [40]	Petis S.	11	Canadian Journal of Surgery	2.089	Narrative review	UK	Outcome	5	140
26	2016	Hypoalbuminemia Independently Predicts Surgical Site Infection, Pneumonia, Length of Stay, and Readmission After Total Joint Arthroplasty [41]	Bohl D.D.	37	Journal of Arthroplasty	4.757	Retrospective observational	USA	Complications	3	138
27	2016	Robotics in Arthroplasty: A Comprehensive Review [42]	Jacofsky D.J.	28	Journal of Arthroplasty	4.757	Narrative review	USA	Robotics	5	137
28	2015	What Is the Learning Curve for the Anterior Approach for Total Hip Arthroplasty? [43]	de Steiger R.N.	28	Clinical Orthopedics and Related Research	4.176	Retrospective observational	Australia	Outcome	3	136
29	2015	A Subject-Specific Musculoskeletal Modeling Framework to Predict in Vivo Mechanics of Total Knee Arthroplasty [44]	Marra M.A.	7	Journal of Biomechanical Engineering	2.097	Biomechanical study	Netherland	Biomechanic	NA	136
30	2016	Enhanced recovery after surgery for primary hip and knee arthroplasty: A review of the evidence [45]	Soffin E.M.	11	British Journal of Anaesthesia	9.166	Narrative review	USA	Recovery & rehabilitation	5	135
31	2016	The epidemiology of failure in total knee arthroplasty avoiding your next revision [46]	Khan M.	5	The Bone & Joint Journal	5.082	Narrative review	UK	Complications	5	135
32	2016	Improved accuracy of component positioning with robotic-assisted unicompartmental knee arthroplasty [47]	Bell S.W.	8	Journal of Bone and Joint Surgery—American Volume	5.284	Randomized clinical trial	UK	Robotics	1	134
33	2015	Systematic Review of Patient-specific Instrumentation in Total Knee Arthroplasty: New but Not Improved [48]	Sassoon A.	15	Clinical Orthopedics and Related Research	4.176	Systematic review	USA	Instrumentation	3	134
34	2017	Effect of genotype-guided warfarin dosing on clinical events and anticoagulation control among patients undergoing hip or knee arthroplasty: The GIFT randomized clinical trial [49]	Gage B.F.	57	JAMA—Journal of the American Medical Association	56.272	Randomized clinical trial	USA	Thromboembolism	1	133
35	2015	Cup position alone does not predict risk of dislocation after hip arthroplasty [50]	Esposito C.I.	19	Journal of Arthroplasty	4.757	Retrospective observational	USA	Complications	3	130
36	2016	Preoperative Reduction of Opioid Use Before Total Joint Arthroplasty [51]	Nguyen L.-C.L.	5	Journal of Arthroplasty	4.757	Retrospective observational	USA	Opioid use	3	129
37	2017	Effect of Bundled Payments and Health Care Reform as Alternative Payment Models in Total Joint Arthroplasty: A Clinical Review [52]	Siddiqi A.	8	Journal of Arthroplasty	4.757	Narrative review	USA	Health-care economics	5	128
38	2017	Dislocation of a primary total hip arthroplasty is more common in patients with a lumbar spinal fusion [53]	Buckland A.J.	16	The Bone & Joint Journal	5.082	Retrospective observational	USA	Complications	3	126
39	2015	Rapid recovery protocols for primary total hip arthroplasty can safely reduce length of stay without increasing readmissions [54]	Stambough J.B.	24	Journal of Arthroplasty	4.757	Retrospective observational	USA	Recovery & rehabilitation	3	126
40	2017	Opioid Use After Total Knee Arthroplasty: Trends and Risk Factors for Prolonged Use [55]	Bedard N.A.	25	Journal of Arthroplasty	4.757	Retrospective observational	USA	Opioid use	3	125
41	2016	Validation of the KOOS, JR: A Short-form Knee Arthroplasty Outcomes Survey [56]	Lyman S.	54	Clinical Orthopedics and Related Research	4.176	Retrospective observational	USA	Outcome	3	122
42	2015	Pelvic Tilt in Patients Undergoing Total Hip Arthroplasty: When Does it Matter? [57]	Maratt J.D.	8	Journal of Arthroplasty	4.757	Retrospective observational	USA	Preoperative planning	3	120
43	2015	General compared with spinal anesthesia for total hip arthroplasty [58]	Basques B.A.	33	Journal of Bone and Joint Surgery—American Volume	5.284	Retrospective observational	USA	Anesthesia	3	118
44	2015	The Otto Aufranc Award: Modifiable vs Nonmodifiable Risk Factors for Infection After Hip Arthroplasty [59]	Maoz G.	5	Clinical Orthopedics and Related Research	4.176	Retrospective observational	USA	Complications	4	117
45	2015	In-home telerehabilitation compared with faceto-face rehabilitation after total knee arthroplasty: A noninferiority randomized controlled trial [60]	Moffet H.	28	Journal of Bone and Joint Surgery—American Volume	5.284	Randomized clinical trial	Canada	Recovery & rehabilitation	1	117
46	2016	Patient-reported outcome measures in arthroplasty registries: Report of the Patient-Reported Outcome Measures Working Group of the International Society of Arthroplasty RegistriesPart II. Recommendations for selection, administration, and analysis [61]	Rolfson O.	25	Acta Orthopaedica	3.717	Comment	Sweden	Outcome	5	116
47	2015	Effect of adductor canal block vs femoral nerve block on quadriceps strength, mobilization, and pain after total knee arthroplasty: A randomized, blinded study [62]	Grevstad U.	10	Regional Anesthesia and Pain Medicine	6.288	Randomized clinical trial	Denmark	Outcome	1	115
48	2015	Does varus alignment adversely affect implant survival and function 6 years after kinematically aligned total knee arthroplasty? [63]	Howell S.M.	49	International Orthopedics	3.075	Prospective observational	USA	Outcome	3	114
49	2017	Alignment options for total knee arthroplasty: A systematic review [64]	Rivière C.	16	Orthopedics and Traumatology: Surgery & Research	2.256	Systematic review	UK	Surgery technique	1	113
50	2017	The minimal clinically important difference for Knee Society Clinical Rating System after total knee arthroplasty for primary osteoarthritis [65]	Lee W.C.	4	Knee Surgery, Sports Traumatology, Arthroscopy	4.342	Retrospective observational	Singapore	Outcome	4	113

(citation number was retrieved from the Scopus, February 2022).

Table 3.

Altmetric indices for the 50 most-cited articles published 2015-2019.

Rank	First author	Year	Cited by	Research interest score	Altmetric attention score (AAS)	Twitter	Dimensions	Mendeley	Facebook	Policy source	News outlet	Blogs	Patent	Wikipedia	Research highlight platform	Redditors	Google user	CiteULike	Connotea	Video uploaders
1	Sloan M. [7]	2018	452	299.3	85	8	594	512	1			1
2	Patel A. [17]	2015	260	158	Altmetric hasn't picked up any sharing activity around this article yet.
3	Lewis G.N. [18]	2015	242	153.6	Altmetric hasn't picked up any sharing activity around this article yet.
4	Goesling J. [19]	2016	238	137.2	223	26	270	251	3		28				1
5	Inacio M.C.S. [20]	2017	231	141.5	23	13	264	320	1		2
6	Bozic K.J. [21]	2015	229	137.5	5	8	264	261	1
7	Culliford D. [22]	2015	205	129.2	24	3	238	304			2	1
8	Gunaratne R. [23]	2017	189	122.9	27	18	236	252			2
9	Iorio R. [24]	2016	185	98.8	3		214	138		1
10	Delanois R.E. [25]	2017	183	122.1	2	3	225	274
11	Kunutsor S.K. [26]	2016	183	106.4	13	1	215	322		1	1
12	Gwam C.U. [27]	2017	181	122.4	35		238	208			5
13	Arabnejad S. [28]	2017	177	118.2	82	3	199	432	2		8	2	2			1
14	Anderson D.R. [29]	2018	168	98	657	844	198	446	24		10	5		3	1		1
15	Choi Y.-J. [30]	2016	166	102.2	62	52	180	339			2
16	Higgins B.T. [31]	2015	166	99.6	12	8	194	286		1				1
17	Liddle A.D. [32]	2015	165	90.6	35	41	181	155		2
18	Losina E. [33]	2015	165	96.1	12	5	187	258			1
19	Pandit H. [34]	2015	154	84.9	16	23	169	84	2
20	De Steiger R.N. [35]	2015	151	89.2	14	4	176	149	2		1
21	Burns L.C. [36]	2015	147	86.8	12	12	155	228	1				1
22	Thiele K. [37]	2015	147	88	10	1	171	139	1		1
23	Keswani A. [38]	2016	142	76.1	12	3	170	182			1
24	Kamath A.F. [39]	2015	142	82.8	Altmetric hasn't picked up any sharing activity around this article yet.
25	Petis S. [40]	2015	140	97.5	4	4	168	500	2
26	Bohl D.D. [41]	2016	138	80.7	1	2	161	145
27	Jacofsky D.J. [42]	2016	137	93.6	5	4	186	316						1
28	de Steiger R.N. [43]	2015	136	84.5	8	11	164	132
29	Marra M.A. [44]	2015	136	86.9	2	1	151	298	1									1
30	Soffin E.M. [45]	2016	135	82.1	8	10	163	269	1
31	Khan M. [46]	2016	135	85.2	71	97	153	118	2
32	Bell S.W. [47]	2016	134	113.9	5	3	170	107		1
33	Sassoon A. [48]	2015	134	75.8	6	7	151	158	1
34	Gage B.F. [49]	2017	133	41.1	850	191	150	193	3	1	89	8					5			2
35	Esposito C.I. [50]	2015	130	74.3	2	2	148	145
36	Nguyen L.-C.L. [51]	2016	129	76.1	8	13	160	129							1
37	Siddiqi A. [52]	2017	128	90.6	6	2	162	183	2	1
38	Buckland A.J. [53]	2017	126	81.7	99	148	165	155	3
39	Stambough J.B. [54]	2015	126	76.9	405	2	143	158			51
40	Bedard N.A. [55]	2017	125	35.6	5	8	152	101	1
41	Lyman S. [56]	2016	122	79.4	6	9	148	199	1
42	Maratt J.D. [57]	2015	120	65.3	10	1	142	153	1		1
43	Basques B.A. [58]	2015	118	63.5	6	4	127	167		1					1			1
44	Maoz G. [59]	2015	117	66.3	5	2	130	152	1	1
45	Moffet H. [60]	2015	117	88.2	23	17	147	334	1	1		1					1
46	Rolfson O. [61]	2016	116	77.2	1	1	139	161
47	Grevstad U. [62]	2015	115	74.2	11	16	143	175	4
48	Howell S.M. [63]	2015	114	64.7	3		125	126					1
49	Rivière C. [64]	2017	113	103.9	2	2	135	218
50	Lee W.C. [65]	2017	113	72.3	Altmetric hasn't picked up any sharing activity around this article yet.

(Altmetric indices were retrieved from the Scopus, February 2022).

Table 4 illustrates the citation and AAS data broken down by year, country of origin, and journal characteristics. Overall, the 50 most-cited papers accrued 7955 total citations, with a mean of 159.10 ± 56.4 (range 113-452) citations per article and 32.7 ± 19.6 citations per year per article. The overall mean AAS across the papers was 63.4 ± 164.8 (median = 10.5). The mean first author’s H-index was 23.8 ± 18.9, and each paper had a mean of 6.5 ± 4.7 authors. Not surprisingly, 24 articles on our list were published in 2015. However, the 2 papers with the highest AAS were published in 2018. Furthermore, papers published in 2017 and 2018 (most recent papers) had a significantly higher mean AAS than those published in 2015 and 2016 (149.5 ± 273.1 vs 29.5 ± 77.7, P = .030). No significant difference was observed between them regarding citation number (175.3 ± 87.0 vs 152.8 ± 38.6, P = .75).

Table 4.

Citation and Altmetric Attention Score (AAS) data broken down by variables of interest.

	Frequency (percent)	Citations (mean ± SD)	AAS (mean ± SD)
Overall	50 (100%)	159.10 ± 56.4	63.4 ± 164.8
Year
2015	24 (48%)	153.2 ± 41.0	29.8 ± 86.4
2016	13 (26%)	150.8 ± 33.9	32.1 ± 61.7
2017	11 (22%)	154.5 ± 39.1	113.1 ± 261.1
2018	2 (4%)	310.0 ± 200.8	371.0 ± 404.5
2019	0	-	-
P valuea		.36	.12
Country
USA	25 (50%)	159.1 ± 56.4	75.1 ± 188.8
UK	9 (18%)	165.4 ± 45.1	21.3 ± 23.0
Australia	4 (8%)	176.8 ± 42.5	18.0 ± 8.6
Canada	4 (8%)	152.3 ± 26.7	193.5 ± 310.5
Others (N = 1 for Denmark, Germany, Lebanon, Netherland, New Zealand, South Korea, Singapore, Sweden)	8 (16%)	150.1 ± 43.0	16.3 ± 22.9
P valuea		.44	.23
Article type
Original	37 (74%)	160.5 ± 62.7	79.1 ± 189.5
Retrospective observational	27 (54%)	166.4 ± 69.0	34.6 ± 82.6
Randomized clinical trial	5 (10%)	133.4 ± 21.2	309.2 ± 411.3
Others (N = 1 for Biomechanical study, comment, instrumentation)	3 (6%)	143.0 ± 31.1	28.3 ± 46.5
Prospective observational	2 (4%)	176.0 ± 87.7	113.0 ± 155.6
Review	13 (26%)	155.0 ± 34.3	19.0 ± 23.2
Narrative review	6 (12%)	140.2 ± 13.3	26.0 ± 31.5
Meta-analysis	4 (8%)	184.5 ± 41.1	12.3 ± 0.57
Systematic review	3 (6%)	145.3 ± 39.2	11.7 ± 13.4
P valueb (comparing review vs original articles)		.57	.81
P valuea (among review article types)		.11	.63
Open access
Yes	4 (8%)	144.8 ± 28.5	420.7 ± 210.4
No	46 (92%)	160.3 ± 58.2	119.5 ± 18.4
P valueb		.72	.73
Site
Knee	22 (44%)	149.9 ± 35.9	17.4 ± 19.2
Knee and hip	17 (34%)	187.4 ± 79.9	129.0 ± 262.4
Hip	11 (22%)	133.8 ± 20.3	57.6 ± 164.8
P valuea		.014∗	.66
Topic
Outcome	13 (26%)	141.4 ± 25.0	16.4 ± 17.5
Complication	10 (20%)	151.6 ± 36.5	25.2 ± 34.9
Epidemiologic	4 (8%)	183.8 ± 35.6	14.0 ± 18.2
TJA volume projection	4 (8%)	287.0 ± 112.3	44.0 ± 35.5
Others (N = 1 for anesthesia, biomechanics, preoperative planning, surgery technique)	4 (8%)	121.8 ± 9.9	5.0 ± 3.8
Recovery & rehabilitation	3 (6%)	126.0 ± 9.0	145.3 ± 225.0
Health-care economics	3 (6%)	159.3 ± 28.9	7.0 ± 4.6
Opioid use	3 (6%)	164.0 ± 64.1	78.7 ± 125.0
Robotics	2 (4%)	135.5 ± 2.1	5.0 ± 0.0
Instrumentation	2 (4%)	155.5 ± 30.4	44.0 ± 53.7
Thromboembolism	2 (4%)	150.5 ± 24.7	753.5 ± 136.5
P valuea		.03∗	.11
Journal
Journal of Arthroplasty	15 (30%)	148.1 ± 25.2	38.1 ± 106.1
Journal of Bone and Joint Surgery	6 (12%)	186.5 ± 130.8	23.83 ± 30.7
Clinical Orthopedics and Related Research	5 (10%)	147.6 ± 46.2	6.0 ± 1.2
Bone and Joint Journal	5 (10%)	168.0 ± 53.7	55.3 ± 37.0
Others	19 (38%)	159.1 ± 56.4	117.0 ± 247.9
P valuea		.90	.09

Asterisk indicates that P-value is <.05 and statistically significant.

Bold indicates the significantly different value (P < .05) from other values in the table.

^aKruskal-Wallis test.

^bMann-Whitney test.

The mean and median level of evidence was 3.0 for articles, comprising 6/3/30/2/7 papers with a level 1/2/3/4/5 of evidence, respectively. Regarding 2 popular social media platforms, Twitter and Facebook, a median of 7 (range 0-843) and 1 (range 1-23) post was published, respectively. The median impact factor of journals was 4.8 (range 2.1 to 91.2).

With 25 publications, the United States institutions contributed the greatest number of papers to our list, followed by the United Kingdom [9], Australia [4], and Canada [4]. Figure 1 illustrates the country of origin of the articles assessed in this study.

Figure 1.

Distribution of the country of origin among papers included in this study.

The majority of publications (74.0%) were original research, and retrospective observational studies were predominant with 27 papers. Meta-analyses had the highest mean citation count among all methodologies (184.5 ± 41.1). Studies on knee arthroplasty were more frequent [22], followed by hip and knee [17] and hip arthroplasty [12]. The Journal of Arthroplasty contributed the greatest number of papers [16], followed by the Journal of Bone and Joint Surgery [6], Clinical Orthopedics and Related Research [5], and Bone and Joint Journal [5].

Functional and radiological outcomes [14], complications [11], epidemiologic studies [4], and TJA volume projection [4] were the most frequent subject across the papers. TJA volume projection studies had a significantly higher number of citations among different topics (287.0 ± 112.3, P = .03).

Table 5 summarizes the frequent keywords out of 138 keywords that were used in the papers included in this study.

Table 5.

Keywords used in the papers included in this study (27/50 articles provided keywords, keywords that did not repeat were omitted).

Keywords	Frequency
Total knee arthroplasty	12
Total hip arthroplasty	7
Arthroplasty	7
Epidemiology	6
Knee	4
Total knee replacement	3
Hip	3
Projections	3
Opioid	3
Alignment	2
Chronic pain	2
Incidence rates	2
Infection	2
National Inpatient Sample	2
Osteoarthritis	2
Patient-reported outcomes	2
Readmissions	2
Revision	2
Risk factors	2
Satisfaction	2

Correlations

Using spearman correlation, we found a significant correlation between AAS and citation count. However, the correlation was weak (correlation coefficient = 0.379, P = .009, Fig. 2). Also, AAS had a significant correlation with Twitter (correlation coefficient = 0.601, P < .001), Facebook (correlation coefficient = 0.560, P = .004), news outlets (correlation coefficient = 0.951, P < .001), dimension (correlation coefficient = 0.381, P = .009), and blog (correlation coefficient = 0.82, P = .046) mentions. The citation count was also significantly and strongly correlated with ResearchGate interest score (correlation coefficient = 0.818, P < .001, Fig. 3), but not with the journal’s impact factor (P = .052). AAS was significantly correlated with the journal’s impact factor 2020 (correlation coefficient = 0.547, P < .001, Fig. 4). First author’s H-index was not significantly correlated with AAS or citation number (P = .66 and .44).

Figure 2.

Spearman correlation between citation count and Altmetric attention score (AAS).

Figure 3.

Spearman correlation between citation count and ResearchGate interest score (RIS).

Figure 4.

Spearman correlation between Altmetric attention score (AAS) and impact factor.

Discussion

With the popularity of social media platforms, these outlets have emerged as the primary source of news and information with arguably a larger audience than the traditional news sources. It is not surprising that researchers are also increasingly using social media to promote their research, report their findings, and collaborate with colleagues. Considering the limitations of traditional scientometric measures, alternative metrics have been introduced as complementary measures of the publication impact [66]. A growing number of researchers are cognizant of the alternative metrics, and the frequency of published articles with an Altmetric score is rapidly increasing [67]. In this study, we aimed to analyze the characteristics of the top 50 most-cited articles in the field of arthroplasty to investigate their breadth of dissemination through the traditional and alternative scientometric measures.

The top 50 most-cited studies in the arthroplasty literature have garnered a total of 7955 citations with a mean of 159.10 ± 56.4 citations per paper. At the same time, the mean AAS was 63.4 ± 164.8, and the median AAS was 10.5. Previous studies in other orthopedic specialties have found a median AAS of 4 to 235, depending on the topic and age of the publication [66,[68], [69], [70], [71]]. In line with other authors, we found a significant but weak correlation between the citation number and AAS across the included papers [[3], [67], [68], [69],[72], [73]]. Several studies have reported a weak correlation between citation count and AAS in orthopedics [68,69] and other fields [[3], [67],[72], [73]], and occasionally nonsignificant correlations [70], suggesting that Altmetrics do not represent the same measure as traditional scientometric measures. Instead, Altmetric measures could act as a complementary tool to inform the social and cultural impact of the literature.

While AAS was weakly correlated with citation count, it has moderate correlation with the journal’s impact factor (correlation coefficient = 0.547, P < .001, Fig. 4). This may stem from the reader’s tendency to share articles from the journals with a higher impact factor. Although a similar trend was found in total knee literature, a correlation between AAS and impact factor was not found in sports science and total hip arthroplasty literature [71,72,74]. High-impact journals may be more active in engaging with social media, or the quality of studies may play a role in this association. Kunze et al. found that high methodologic quality and a lower rate of bias were significant predictors of a higher AAS in the arthroplasty literature [68]. Compared with citation count, AAS is considered a quantitative variable assessing a paper’s more immediate impact attraction on readers and the intended audience [75]. However, since not every author is active on social media, more active authors will generate a higher AAS for their publications [68]. With the continued growth of the social media outreach of the scientific literature, this effect will likely be minimized in the future.

Citation count is classically delayed by 1-2 years after publication, which is not the case with Altmetric measures. Interestingly, we found that recently published papers (2017-2018) had a significantly higher AAS than papers published in 2015-2016 (149.5 ± 273.1 vs 29.5 ± 77.7, P = .030). This shows that not only the Altmetric measures do not suffer from the delay observed in traditional scientometric measures but also the authors could also be more attentive to social media recently [76].

We found that randomized controlled trials had a higher mean AAS (309.2 ± 411.3) although this did not reach statistical significance. Previous authors have also shown that prospective studies attract more interest on social media [66]. Interestingly, review articles, including meta-analyses, had the highest citation count, but not a higher AAS [77]. The Journal of Arthroplasty contributes to most of the included studies (30%), followed by The Journal of Bone and Joint Surgery (12%), Bone and Joint Journal (10%), and Clinical Orthopedics and Related Research (10%). In a scientometric study of arthroplasty researches by O’Neill et al., as high as 10.9% of all arthroplasty research articles from 2001 to 2016 was published in the Journal of Arthroplasty, similar to our study, while Clinical Orthopedics and Related Research and The Journal of Bone and Joint Surgery ranked second and third, respectively [78]. Kunze et al. analyzed the articles published between January and December 2016 in the 5 journals with the highest impact factor [68]. The mean AAS was 8.6, and the mean citation count was 15. AAS was significantly associated with citations (β = 0.16; P < .0001). They also found that publications from North America and studies concerning validity/reliability measurement gained more social media attention, as measured by AAS. They also noted that studies with more citations are more likely to be mentioned on Twitter, Facebook, and news outlets [68].

We acknowledge several limitations to our study. First, we only included articles with the term “arthroplasty” in their title, whereas it is possible that a highly cited publication did not have this term in its title. Also, we only searched through Scopus to minimize redundancies and duplicate citations, with the downside of being less comprehensive. However, we did not aim to do a systematic review. Finally, we included papers published in 2015-2019 in this study. While this may limit the traditional citation count, the alternative scientometric measures are relatively new and are not applicable to older studies. Despite these limitations, we managed to identify the most impactful articles in the field of arthroplasty. The authors suggest further studies to distinguish the most noteworthy articles in other fields of orthopedic surgery to guide future research.

Conclusions

In this study, the top 50 most-cited papers in the field of arthroplasty were identified. Traditional and alternative scientometric measures were extracted and compared between studies. We found that the AAS was highest in more recently published papers, while citation count had the opposite trend. Also, AAS was significantly correlated with the journal’s impact factor (moderate) and citation count (weak). Meta-analyses and volume projection studies garnered the most citations. Additionally, we found a weak correlation between AAS and citation count, which suggests that the alternative scientometric measures are complementary to, and not substitutes for, complement traditional measures such as citation count and impact factor.

Conflicts of interest

The authors declare that there are no conflicts of interest.