Carpal tunnel syndrome (CTS) is the most frequent nerve-entrapment syndrome and a relevant population disabling condition. CTS is most common in women, with an incidence peak in the 50–59 year age group in contrast to a gradually increasing age-related trend in men.1–3 Atroshi et al. reported that the prevalence of “clinically certain” CTS among the Swedish general population from 25 to 74 years old was 3.8% (men: 2.8%; women: 4.6%).4 Other studies report different figures according to distinct diagnostic criteria and different population groups, countries and study year.5–7 Incidence studies also show very different figures,8 from 109 to 4010 new cases per 10,000 person-years. Nonetheless, incidence studies generally show an increasing temporal trend.2,3,8,10–12

In accordance with this high prevalence, the population incidence of carpal tunnel release (CTR) surgery shows figures between 6-7 interventions per 10,000 person-years in European countries13–17 and two-three times higher in Canada and the United States.18–20 Surprisingly, the study of medical practice variations in CTR has stirred up little interest. Only a few papers have examined this topic: a first study in Maine reported CTR incidence rates from 8.2 to 28.7 interventions per 10,000 inhabitants in 21 hospital referral areas in 199319; Mattioli et al. reported rates -always per 10,000 and year- from 5.0 to 13.2 in 7 Italian regions from 1997-2002,1 and Tepper et al., in Canada, report rates between 8.3 and 19.5 in 4 rural/urban settings.20 In Spain, one report in seven areas in Catalonia described CTR rates between 7.6 and 17.2 in 2000,16 and the Spanish Atlas on Medical Practice Variation project showed rates from 0.2 to 13.9 per 10,000 in 105 healthcare areas from 11 regions in 2002 (from 13.1 to 76.2 in people of 65 years and older),17 although the authors note these estimates could be imprecise because part of the CTR surgery was performed in the outpatient setting or referred to private hospitals, making utilization data difficult to obtain.

Ambulatory surgery and referral to private hospitals are largely used in Spain to improve efficiency and reduce elective surgery waiting lists, but it is not clear if these strategies contribute to decreasing small-area variations (it would suggest that areas with less resources use these policies to enhance their population surgery rates) or, on the contrary, to increasing differences (suggesting that areas with higher surgery rates apply these policies in a more intensive way than their counterparts with lower rates). The objectives of this study are to describe CTR surgery rates in the Valencia region, to analyze variability among their 22 healthcare areas and to evaluate the contribution of ambulatory surgery and referrals to private hospitals to the variability found. We begin from the null hypothesis of the absence of systematic variation between healthcare areas and the absence of association between surgery rates and the percentage of ambulatory surgery or referrals.

First, we calculated the standardized age-sex rates (direct method) using the 2001 Spanish census population as the reference population. Second, we estimated the indirect standardized surgery ratio (SSR), the quotient of the observed (yi) to the expected (ei) number of surgery cases, and represented these ratios in healthcare area maps. The expected cases of CTR were derived based on the age-specific rate for 18 five-yearly age groups and the sex stratum in the standard population, which was the population from the 22 healthcare areas under study. More precisely, ei=∑j,knijkRjk, where nijk is the population in area i, age group j and sex stratum k, and Rjk is the age-sex specific rate for the whole region under study. Because SSR are very dependent on the number of residents in each area, and the variability of observed cases is usually higher than the expected variability in a Poisson distribution (extravariability) we used the spatial smoother method proposed by Besag, York and Mollié (BYM model) 27,28 to smooth the SSR, reduce the extravariation and allow us to visualize subjacent patterns.

To assess variability we use the common statistics of small area variation analysis29: extremal quotient (EQ), coefficient of variation (CV), weighted coefficient of variation (CVw), systematic component of variation (SCV), and empirical Bayes (EB) variance component. Because the presence of areas with extreme values, and as in the Spanish Atlas of variation22, we calculated some statistics excluding areas outside the percentiles 5-95 and denoted them as EQ5-95, CV5-95; CVw5-95 and SCV5-95. Figure 1 shows the formulation of statistics. Here DSRi refers to the direct standardized rate per 10,000 inhabitants at the i-th health area, with i=1,…, I, while yi and ei are the observed and expected cases respectively. Finally, we used the Spearman's correlation coefficient to analyze the bivariate associations between the standardized surgery rates and the proportion of surgical schemes used (inpatient, outpatient and private referral) and resource variables in each area (described in the Table 1). All analyses were carried out using STATA® 9 (StataCorp, College Station, Texas) and R (RDC Team. 2004. http://www.R-project.org) except from the SSR smoother, which was carried out with WinBUGS 1.4.

Figure 1.

Formulation of the descriptive statistics.

(0.2MB).

The Valencia Healthcare Agency performed 4,126 CTR surgeries in 2006 (8.18 per 10,000 inhabitants); 2,861 interventions were carried out in public hospitals, 487 (11.8%) with conventional hospital admissions and 2,374 (57.5%) as ambulatory surgery, and 1,265 (30.7%) were referred and carried out in private hospitals, all as ambulatory surgery. Rates of surgery (Table 2 and Fig. 2) ranged from 1.13 to 12.76 per 10,000 inhabitants between areas in the 5 and 95 percentiles. Small area variation statistics, excluded areas outside percentiles 5 and 95, showed a moderate degree of variability, with 2.7 times more interventions between extreme areas. Surgery rates were higher in women (13.1/10,000 woman/year vs. 3.4 in men), but the statistics of variability were similar.

Figure 2.

Standardized rates of carpal tunnel release surgery (Valencia Community, Spain, 2006).

(0.16MB).

Figure 3 shows the geographical distribution of the standardized surgery rates. Several territories, including the city of Valencia, showed non-significant differences with the Autonomous region in global terms, but some areas on the north and on the south coast had lower rates than average, and the central areas had higher rates. By sex, men and women showed similar territorial patterns, but the lower surgery rate in men produces more non-significant areas. Rates in men and women by areas exhibit a strong correlation (r=0.84; r2=0.72; p<0.001). Table 3 shows the bivariate correlations between CTR surgery rates and the proportion of surgical practices and resources used. None of the variables analyzed was associated with surgery rates.

Figure 3.

Standardized surgery ratios of carpal tunnel release surgery by areas (Valencia Community, Spain, 2006).

(0.32MB).

Results of this study show a rate of 8.2 CTR surgeries per 10,000 inhabitants in the Valencia Region, performed in the most part as ambulatory surgery (88.2%). Excluding areas outside 5-95 percentiles variation was moderate (RV5-95=2.7, slightly higher than that observed in the hip fracture hospitalization, the condition commonly used as a pattern of low variability),17 similar for women and men, and was not associated with the proportion between surgical schemes used in each area, or with the availability of resources.

The CTR surgery rate was higher than that reported in previous studies in the Valencia Region (1.1/10,000 inhabitants for 1995-199630 and 5.7/10,000 for 200217), although those studies did not include referrals to the private sector. On the contrary, the CTR surgery rate was lower than that described for Catalonia in 2000.16 These data suggest that CTR surgery rates could have been increasing in Spain in recent years and, also suggest a substantial variability between regions, as reported in the Atlas of Medical Practice Variation project.17 Regarding the role of ambulatory surgery, this surgical scheme seems clearly consolidated for CTR in the Valencia region (4 in 5 CTR surgeries are performed on ambulatory basis). Lower proportion of ambulatory surgery in public hospitals is probably related to the retention of patients with more complex conditions or with more comorbidity, although it is also possible that public hospitals could increase their proportion of outpatient surgery. The higher rate of surgery in women is consistent with their higher prevalence of CT Syndrome.

Regarding variation, the range found was similar to that described for areas in Catalonia in 1995-200016 or in Maine at the start of the 1990s,19 but lower than in other studies,22,30 although the variability found in these last studies seems due to the presence of territories with extreme rates and variation statistics would have been similar to the rest of the studies if areas outside the 5-95 percentiles had been excluded. Nonetheless, this moderate variation has relevant repercussions at the population level. For example, the Valencia region would perform 2,306 more CRT surgeries if the rate of the area in the 95 percentile were adopted as the region average rate (and the costs of this increase in surgery rates could be higher than the savings generated by the transfer of all CTR interventions to the ambulatory setting). Nevertheless, regional healthcare services in Spain maintain a substantial waiting list for CTR surgery (about 2.3 people per 10,000 inhabitants in 2006)31 and lowering rates could also have a relevant impact on waiting times.

Although the population resident in each area is the object of different surgical intensity, these differences do not seem to be associated with the avalaibility of orthopedic resources in each departament, nor with the proportion of surgical practices, including referral to private hospitals. The hypothesis that ambulatory surgery increases the rates of surgery has been presented occasionally.32 Our results suggest that higher rates in CTR surgery are not related to higher use of ambulatory surgery or more referrals to the private sector.

This study has several limitations. First, the ecological design affects the causality of the observed relations and the transfer to the individual level. Second, the variation between centres can be affected by the completeness and quality of registers, although CTR is a common intervention and is usually well registered by the Hospital Admission and Documentation services. Third, the data used corresponds to the year 2006, the only one for which we have information on the concerted activity. Since then there have been several changes in the Valencia Health Agency (opening of new hospitals, extension of private concessions, management contracts with indicators of productivity and incentives, etc.) that could modify the results found. Fourth, surgery rates in some areas could be theoretically affected by the degree of attention to non-residents (people living in other communities or other countries) excluded from the study, although the small number of exclusions (less than 1% of the admissions for all causes) makes it unlikely that this effect may have had a major impact on surgery rates. Finally, and more relevant, the hypothesis that ambulatory surgery increases surgery rates requires a longitudinal design and the findings of this cross-sectional study, although suggestive, should not be considered as definitive. This hypothesis has never been formally evaluated and, although it is not supported by our results, other data suggest technological or organizational changes could increase surgery rates in areas with different basal rates. For example, CTR surgery in Catalonia16 increased from 7.6 to 13.2 per 10,000 persons/year between 1996 and 2000, whereas the area with a lower basal rate increased from 5.0 to 7.6, the area with a higher basal rate increased from 11.1 to 15.1, suggesting that increments in surgical rates were rarely used to reach any convergence between territories with possible under and overutilization. In fact, territories with higher rates increased at the same rhythm as those of other territories, sustaining the variability among them.

Medical practice variation studies do not permit us to determine what rate is right,33 but highlights that healthcare services manage very similar populations in very dissimilar ways, and at least raise the question of whether some populations are under or overtreated, with the subsequent problems of quality losses and waste of scarce resources. There is evidence of the effectiveness of CTR surgery in clinical improvement and other endpoints, but the severity threshold to decide between surgery and conservative management remains uncertain.34 Clinical trials in patients with mild symptoms can be useful to clarify these aspects, but the use of functional measures or health-related quality of life instruments could also be interesting to objectify the severity threshold currently used in each territory and to inform practice policies, whether to increase (if in some areas patients suffer unacceptable pain or functional loss conditions) or decrease (if in some areas surgery is performed in mild or inappropriate cases) surgery rates.

P. Rodríguez-Martínez, S. Peiró and M. Gisbert-Grifo conceived the idea, did the data analyses, performed the literature search and wrote the first draft of the manuscript. J. Calabuig-Pérez performed the data extraction. J. Librero, E. Bernal-Delgado, M. Ridao-López, J. Calabuig-Pérez and G. Sanfélix-Gimeno participated in conceiving the study design, prepared the databases, data analyses and interpretation of the data. All authors have been involved in discussed analyses, interpretation, and presentation. All authors have contributed to, seen and approved the final submitted version of the manuscript.

This study is part of the Atlas of Medical Practice Variation in the Spanish National Health System research project, funded by the Instituto de Salud Carlos III at the Spanish Ministry of Health (PI06/1673, PI05/2490, G03/202), the Obra Social y Cultural de Ibercaja, and the Fundación Instituto de Investigación en Servicios de Salud.

None.