In cost-benefit analyses, hip arthroplasty is ranked as one of the most successful medical and surgical treatments available (1 , 2 ). Hip arthroplasty has been shown to give arthrosis patients a considerable reduction of pain as well as improved functional capability, at a lower cost than other forms of treatment (3 ). Nevertheless, the frequency of primary total hip arthroplasties varies significantly in different countries (4 ). These can be difficult to compare, however, since they often date from different periods of time and have been estimated without taking demographic differences into account. Previous reports have indicated that the number of hip arthroplasties caused by rheumatic disorders is declining (5 – 8 ), which may be due to improved drugs. On the whole, however, there is a clearly increasing frequency of hip arthroplasty in the Western countries. This has been documented for all the Nordic countries (9 ), Australia (10 ), Canada (11 , 12 ) and the US (13 ). Within some countries, considerable regional differences in age-standardized rates have been documented (12 , 14 – 17 ). In Norway, the prevalence of primary hip arthroplasty within different regions is unknown. Hence, the purpose of this study was to investigate whether the prevalence of primary total hip arthroplasty varied according to place of residence. We therefore compared the frequency of hip arthroplasties in the population according to place of residence, distributed among the regional health enterprises and county of residence in the years 1989 – 2008. We also wished to focus on changes in the prevalence over time.
Material and methods
In September 1987, the Norwegian Arthroplasty Register (18 ) started to record information on primary and revision hip arthroplasties (19 , 20 ). This information is submitted to the register on paper (not electronically) by orthopaedic surgeons in all hospitals in Norway that perform this type of surgery. Compared to the number of operations reported to the Norwegian Patient Register, a total of 97 per cent of all primary hip arthroplasties were reported to the register in the period 1999 – 2002 (21 ). Prior to 1989, the registration procedure did not comprise all hospitals. In this study we therefore included primary hip arthroplasties performed in the period from January 1989 to December 2008, with the exception of operations performed on patients with no known place of residence (n = 37). The study was based on information on hip arthroplasties reported to the arthroplasty register. Information on the patients’ last known address as of 31 December 2008 was collected from the National Population Register, and demographic data were supplied by Statistics Norway (22 ).
Analyses
The annual unadjusted rate of primary total hip arthroplasties for the years 1989 – 2008 and for the periods 1989 – 90, 1991 – 95, 1996 – 2000, 2001 – 05 and 2006 – 08 was estimated as the number of operations divided by the population mean (average of the population at the beginning and at the end of the year) in Norway. Subdivisions have to the largest extent possible been undertaken in five-year intervals in order to allow for a comparison with other studies. The annual age and gender standardized rate was estimated by way of direct standardization (23 ), using the gender and age distribution (0 – 39, 40 – 49, 50 – 59, 60 – 69, 70 – 79, 80+ years) in Norway as the standard population. The annual rate for each period of time was also estimated for various diagnoses that may lead to arthroplasty, for each of the regional health enterprises and counties, as well as for sub-groups defined by gender and age for the primary operation. Even though different countries may have widely differing age distributions, the choice of a common standard population may still provide comparable figures for prevalence. An annual age-standardized rate was therefore also established on the basis on an estimated European standard population (24 ). Rate ratio (RR) values for the changes in prevalence over time and between geographical areas were estimated using Poisson regression, with the period 1991 – 95 as reference. These analyses were adjusted for gender and age, which were grouped as described above. P values less than 0.05 were assumed to be statistically significant. Analyses were undertaken with the aid of the software package SPSS 16.0 (SPSS Inc., Chicago, IL).
Results
During the period of study, a total of 112,514 hip arthroplasties were performed in Norway. We observed an increase over time in the annual age and gender standardized frequencies of primary total hip arthroplasties (fig 1), from 109 operations per 100,000 inhabitants in 1991 – 95 to 140 in 2006 – 08 (RR = 1.28; 95 % confidence interval 1.25 – 1.30) (e-tab 1). This increase was at its strongest during the second half of the 1990s, and since then, only minor changes have been observed (fig. 1). The highest frequency was observed in 2003, with 157 operations for each 100,000 of the population. The differences between the standardized and the unadjusted rate remained minor during the entire period of study (fig 1).
Figure 1 Frequency of primary total hip arthroplasties per 100,000 inhabitants, by year of operation. Frequencies are unadjusted, standardized according to the gender and age distribution in Norway 1989 – 2008, and standardized according to the European age distribution (European standard population, ESP).
E -table 1 Gender and age standardized frequency 1 for primary total hip arthroplasties per 100,000 inhabitants, by period of time
No. of prostheses
1989 – 1990
1991 – 1995
1996 – 2000
2001 – 2005
2006 – 2008
RR ² (95 % CI) 2006 – 08 vs. 1991 – 95
Total
112,514
116
109
121
143
140
1.28 (1.25 – 1.30)
Gender
Men
34,479
79
76
80
96
99
1.30 (1.26 –1.35)
Women
78,035
146
137
154
181
173
1.27 (1.24 –1.30)
Age
0 –39
2,056
4.3
4.1
4.2
4.2
4.4
1.08 (0.94 – 1.24)
40 – 49
4,438
36
33
32
38
42
1.27 (1.16 – 1.39)
50 – 59
13,783
142
117
124
151
155
1.33 (1.25 – 1.40)
60 – 69
31,456
391
366
401
433
435
1,18 (1.14 – 1.23)
70 – 79
43,921
602
576
632
786
758
1.32 (1.28 – 1.36)
80+
16,860
312
332
400
500
458
1.37 (1.31 – 1.44)
Diagnosis ³
Idiopathic coxarthrosis
81,510
77
74
85
109
109
1.46 (1.43 – 1.50)
Rheumatoid arthritis
3,390
4.0
4.2
4.0
3.7
3,1
0.74 (0.66 – 0.83)
Sequelae femoral neck fracture
11,690
16
14
14
12
10
0.70 (0.66 – 0.75)
Sequelae congenital dysplasia
8,203
10
8.9
8.4
9.6
9,5
1.05 (0.98 – 1.13)
Sequelae congenital dysplasia with dislocation
815
2.3
1.4
0.7
0.4
0,4
0.31 (0.23 – 0.41)
Sequelae Perthes’ disease/ epiphysiolysis
1,480
1.6
1.6
1.6
1.8
1,8
1.12 (0.95 – 1.31)
Ankylosing spondylitis
476
0.5
0.5
0.6
0.5
0,5
0.82 (0.61 – 1.11)
Fracture
1,103
0.3
0.3
0.7
3.0
3,0
9.60 (7.44 – 12.0)
Osteonecrosis of the femoral head
1,597
0.6
0.9
1.5
3.6
3,6
3.99 (3.37 – 4.17)
Other
2,572
2.3
2.4
2.7
3.1
3,1
1.16 (1.02 – 1.31)
[i]
The age-adjusted frequency of hip arthroplasties was lower for men than for women, totalling 99 per 100,000 men and 173 for each 100,000 women. The frequency of hip arthroplasties increased over time for all age groups over 40 years, but was least marked in the age group 60 – 69 years. The increase in frequency was stronger for hip prosthetics implanted because of idiopathic coxarthrosis (RR = 1.46) than what was observed for the total material (e-tab 1).
County of residence and regional health enterprises
An increase in frequency was evident in all the regional health enterprises and counties (tab 2). In our comparison of prevalence in 1991 – 95 and 2006 – 08, we found the greatest increase among inhabitants of the Helse Nord health region (RR = 1.49; 95 % CI: 1.40 – 1.59). On the other hand, Helse Nord has had the lowest frequency throughout the entire period of study (fig 2). At the county level, we found the smallest increases among inhabitants of Rogaland (RR = 1.13; 1.05 – 1.21) and Telemark counties (RR = 1.13; 1.03 – 1.24), and the largest increase in Nordland county (RR = 1.86; 1.70 – 2.04) (tab 2).
Figure 2 Frequency of primary total hip arthroplasties per 100,000 inhabitants, by place of residence in a regional health enterprise and year of operation standardized according to the gender and age distribution in Norway 1989 – 2008.
Table 2 Gender and age standardized frequency 1 for primary total hip arthroplasties per 100,000 inhabitants, by place of residence and period of time
Regional health enterpriseCounty
No. of prostheses
1989 – 1990
1991 – 1995
1996 – 2000
2001 – 2005
2006 – 2008
RR² (95 % CI) 2006 – 08 vs. 1991 – 95
Helse Øst³
39,125
111
104
120
137
134
1.28 (1.24 – 1.32)
Østfold
6,433
113
111
111
131
136
1.23 (1.13 – 1.32)
Akershus
10,220
107
105
127
140
133
1.24 (1.17 – 1.32)
Oslo
10,463
81
88
107
120
120
1.34 (1.26 – 1.43)
Hedmark
6,355
144
120
138
158
162
1.36 (1.25 – 1.47)
Oppland
5,654
154
117
130
152
135
1.16 (1.06 – 1.27)
Helse Sør³
24,678
129
120
132
155
147
1.23 (1.18 – 1.28)
Buskerud
6,612
132
115
127
155
148
1.26 (1.17 – 1.37)
Vestfold
5,770
114
11
132
142
145
1.30 (1.20 – 1.42)
Telemark
5,153
128
127
133
164
144
1.13 (1.03 – 1.24)
Aust-Agder
2,992
132
133
140
164
172
1.29 (1.15 – 1.44)
Vest-Agder
4,151
146
119
136
158
140
1.18 (1.07 – 1.31)
Helse Vest
21,795
120
115
121
144
141
1.22 (1.17 – 1.28)
Rogaland
8,175
126
118
122
144
135
1.13 (1.05 – 1.21)
Hordaland
10,444
108
110
119
142
140
1.27 (1.20 – 1.35)
Sogn og Fjordane
3,176
148
127
127
156
164
1.31 (1.17 – 1.46)
Helse Midt-Norge
16,824
117
116
118
149
152
1.31 (1.25 – 1.37)
Møre og Romsdal
6,315
115
111
110
147
143
1.29 (1.19 – 1.39)
Sør-Trøndelag
6,561
116
109
116
141
154
1.40 (1.30 – 1.51)
Nord-Trøndelag
3,948
122
139
135
169
165
1.19 (1.08 – 1.31)
Helse Nord
10,092
101
86
104
134
129
1.49 (1.40 – 1.59)
Nordland
5,213
93
71
95
135
132
1.86 (1.70 – 2.04)
Troms
3,488
112
112
123
140
128
1.15 (1.03 – 1.28)
Finnmark
1,391
112
91
94
119
116
1.28 (1.08 – 1.52)
[i]
Even though the differences between the regional health enterprises (fig 2) and the counties (tab 2) have decreased considerably over time, some variations remain. In the period 2006 – 08, the Helse Midt health region stood out, with 152 operations for every 100,000 inhabitants (tab 2). This represented a statistically significant higher frequency than in all other regional health enterprises, with the exception of Helse Sør. With regard to counties, there were still some major differences during the final period, even between counties within the same regional health enterprise (tab 2).
Discussion
The information on hip arthroplasties is based on data reported to the Norwegian Arthroplasty register. Data are reported directly by the surgeon, who usually fills in the form submitted to the register immediately after the operation. Even though this reporting of operations to the register is voluntary, we have reason to believe that the register is near-complete. Comparisons with information submitted to the Norwegian Patient Register (NPR) show that the number of primary hip arthroplasties reported to the Norwegian Arthroplasty Register constituted 97 % of the number of prosthetics reported to the patient register (21 ). This completeness of reporting has also been confirmed at the patient level in several hospitals (25 – 27 ). The validity of the reported information has also been investigated in one hospital and showed very satisfactory results with regard to the variables reviewed: side (left, right), type of operation (primary, revision) and the date of operation (25 ).
The fact that the patient’s reported place of residence was the current place of residence rather than the place of residence at the time of the operation could represent a weakness of the study design. However, in light of the age group in question, there is little reason to assume that many of the patients have moved to other counties or regional health enterprises (28 ).
The frequency of primary hip arthroplasties caused by idiopathic coxarthrosis has increased more than for the material as a whole. This could be due to improved access to surgery, but also to a change in indications for operation. It has been shown that if the operation is postponed and the patient’s condition deteriorates, then the outcome of the operation will be poorer (29 , 30 ). This could probably to some extent explain why a higher number of operations than before are performed on young people with less assumed pain and better functional capability. Nevertheless, we found that the increase was most marked among the oldest patients. The increasing number of elderly people with good general health who are able to undergo hip arthroplasty, better anaesthetics and a general consensus of not using age as a contraindication for an operation, are factors that can explain this increase. It is crucial to investigate whether this change in indication for operation may change the results of hip arthroplasty surgery.
In recent years, we have seen that the frequency of primary hip arthroplasty has remained relatively constant, and one may ask whether existing surgical capacity is sufficient to cover demand for this type of surgery. A study published in 1999 showed that hip arthroplasty capacity in England fell six per cent short of demand (31 ). Comparable frequencies of primary hip arthroplasty have been reported for Norway and England (32 ), and if we assume that the prevalence of hip disorders that require implantation of a prosthesis also is similar, it is reasonable to claim that too few, rather than too many, are offered this type of surgery in Norway. Many studies estimate an increasing need for hip arthroplasty in the years to come, in light of an increasing proportion of elderly people in the population and also because an increasing number of young people (33 ) and elderly people undergo surgery (6 , 33 , 34 ). Further studies ought to be undertaken to clarify the need for future surgical capacity, with regard to Norway as well.
It has been shown that the frequency of primary total hip arthroplasties varies significantly in different countries (4 ). However, a study based on national registry data for the years 1996 – 2000 showed that the annual frequencies of arthroplasties caused by primary coxarthrosis were fairly similar in the Nordic countries (9 ). Differences could nevertheless be observed for groups of individuals. In concurrence with our study, Lohmann et al. (9 ) observed that the proportion with hip arthroplasty was twice as high among Norwegian women as among men, while this proportion was lower in the other Nordic countries, at 1.1 – 1.3.
Similar to the results of our study, considerable regional differences in the frequencies of primary total hip arthroplasties have been found in other countries as well (14 – 16 ). In a Danish study, differences related to diagnoses could not explain the regional variations (15 ). Investigating this issue in Norway is difficult, since generally little is known about the frequency of various diagnoses according to area of residence. Differences in the prevalence of coxarthrosis among various groups of the population have been demonstrated (35 , 36 ), and it is conceivable that this can partly explain the low frequency of arthroplasty operations in some counties. For example, as of 1 January 2009, a total of 22 per cent of the population of Oslo had a non-Western background, compared to 8 per cent in the country as a whole (37 ). Another possible explanation includes the differences in attitudes to this type of operation among various groups of the population, reflected in surveys (38 ). For example, it has been shown that non-European patients have a higher level of discomfort before they undergo surgery, and that they assess this operation as dangerous to a higher extent than patients from a European ethnic background (39 ). Furthermore, studies have revealed a lower frequency of hip arthroplasty in large cities in countries such as England, Denmark, Sweden and Finland (14 , 15 , 17 , 40 ). Attempted explanations refer to how the infrastructure and working conditions are better there than in rural areas, so that it is possible to cope without an artificial joint (14 ). A study from the US showed that low population density was correlated with a high frequency of hip arthroplasty (41 ), while other studies, on the other hand, show that regional differences cannot be explained by population density, or by factors such as density of orthopaedic surgeons, hospital costs and regional gross domestic product (14 , 15 , 17 ). In our study, we have not adjusted for factors such as differences in the extent of hospital coverage and surgical capacity, and further research should also investigate the effects of such factors in the Norwegian context. Geographical differences in the indication for an operation could also be a possible explanatory factor for the variations in frequency according to place of residence (14 ).
