OUP user menu

Appropriateness of phacoemulsification in Spain

Nerea Fernández de Larrea, Juan Antonio Blasco, Urko Aguirre, Susana García, Belén Elizalde, Gemma Navarro, Sabina Pérez
DOI: http://dx.doi.org/10.1093/intqhc/mzp052 31-38 First published online: 6 December 2009


Objective To evaluate the appropriateness of phacoemulsification procedures performed in four Spanish regions, applying criteria developed by means of RAND/UCLA methodology.

Design Prospective observational study.

Setting Seventeen public teaching hospitals in four regions of Spain.

Participants Patients on waiting list to undergo cataract extraction by phacoemulsification.

Intervention Cataract surgery by phacoemulsification.

Main Outcome Measure Level of appropriateness of each intervention, according to criteria developed by means of the RAND/UCLA appropriateness methodology.

Results Among the 5442 analysed patients the indication of phacoemulsification was appropriate in 69.6%, inappropriate in 7.3% and uncertain in 23.0%. Presence of ocular comorbidity, lack of cataract-induced visual function limitation, anticipated postoperative visual acuity of <0.5, preoperative visual acuity of >0.1 and high surgical complexity were associated with inappropriateness.

Conclusions Some clinical characteristics, when present, make it especially important to obtain a careful assessment of the risks and benefits of surgery. Consideration of these characteristics may improve the appropriateness of phacoemulsification.

  • appropriateness
  • phacoemulsification
  • cataract surgery
  • RAND/UCLA appropriateness methodology


The appropriateness of medical procedures is important at every level of the health-care system, from policy-makers [1] to clinicians deciding whether to perform a procedure. Identifying and avoiding inappropriate procedures may reduce the number of adverse events for the patient and may slow the increase in costs. In the mid-1980s, RAND Corporation and the University of California at Los Angeles (UCLA) developed the RAND/UCLA appropriateness methodology [2] as an instrument to evaluate the appropriateness of medical and surgical procedures. The rationale supporting this methodology is that physicians and surgeons must make decisions about when to apply medical and surgical procedures, even though robust scientific evidence about risks and benefits is frequently incomplete [35]. The RAND/UCLA appropriateness methodology combines the best available scientific evidence with the collective judgement of experts to yield explicit criteria regarding the appropriateness of medical or surgical procedures. According to the RAND/UCLA definition, a procedure is appropriate when ‘the expected health benefit exceeds the expected negative consequences by a sufficiently wide margin that the procedure is worth doing, excluding monetary cost’.

Cataracts are a frequent ocular pathology, with a prevalence that reaches some 15–30% in developed countries [69]. Phacoemulsification is the surgical technique of choice in most cases for cataract correction [1012]. It has been shown to provide clear benefits in terms of patient quality of life and in the prevention of some of the more common problems of old age, such as falls [7, 8, 10, 1315]. As a result, cataract surgery is the most frequently performed of all surgical interventions in the industrialized world [6, 7, 13]. In the Spanish National Health System this type of surgery is usually performed as ‘major ambulatory surgery’, and represents the most frequent surgical procedure in this category (about 31% of all interventions) [16].

Using the RAND/UCLA appropriateness methodology, in 2005 our group established a set of appropriateness criteria for phacoemulsification [17]. These explicit criteria were applied in the present work to assess the appropriateness of phacoemulsification procedures performed in four regions of Spain with the goal of identifying characteristics associated with overuse of phacoemulsification.


Development of appropriateness criteria

This study made use of previously established criteria regarding the appropriateness of phacoemulsification. These criteria were developed by means of the RAND/UCLA methodology, as has been described previously [17]. Briefly, a review was made of the literature and 765 clinical scenarios identified representing the majority of settings in which phacoemulsification is considered. A panel of experts comprising 12 ophthalmologists (including a mix of those who perform cataract extraction and those who do not) rated the appropriateness of this technique in each of these clinical scenarios using a scale ranging from 1 (least appropriate) to 9 (most appropriate). The 765 scenarios were classified into ‘appropriate’, ‘uncertain’ or ‘inappropriate’ [2, 17]. An additional round of analysis was performed for the scenarios in which phacoemulsification was deemed ‘appropriate’ in order to identify the scenarios under which the intervention would also be considered necessary. ‘Necessity’ means that the intervention is not only appropriate but crucial and that it would be improper care not to recommend it in the given clinical situation [18, 19]. Each scenario was rated for necessity using a 9 point scale. Scenarios considered appropriate that also had a median necessity rating between 7 and 9 and without disagreement among the panellists were considered ‘necessary’. Otherwise, they were classified as ‘appropriate’. A complete description of these definitions can be found in a previous publication [17].

Measuring the degree of appropriateness of performed phacoemulsification procedures

A prospective observational study was conducted between October 2004 and July 2005. Patients became eligible for the study at the time of their inclusion in waiting list for cataract surgery.

Data were collected on the cataract extractions performed by phacoemulsification at 17 public hospitals in four regions of Spain (Andalusia, Catalonia, Madrid and the Basque Country). The sample size was estimated in order to detect differences of between 1 and 5% in the rate of inappropriateness of phacoemulsification (α value of 0.05 and 1 − β = 0.80). Within each centre 350 operations were included in the study.

Patients aged between 18 and 90 years were enrolled. All gave their consent to be included. Patients were excluded for the following reasons: presence of corneal dystrophy, cataract surgery associated with some other type of ophthalmic surgery, bilateral cataracts when the second intervention was performed within 3 months of the first, patients who could not collaborate because of organic pathologies or serious psychiatric disorders or who were illiterate, and those who could not speak Spanish or who had some problem that prevented the questionnaires from being completed. Patients who died or changed address before the intervention, for whom clinical data were lost, or for whom data for over 50% of the variables considered important were not available, were deemed lost.

The patients' clinical data were collected by ophthalmologists or other health-care professionals trained in the extraction of data from clinical records and blinded to the study goals. A form was designed for the collection of data, which included sex, age and the seven clinical variables necessary for each intervention to be assigned to one of the 765 clinical scenarios for which appropriateness criteria had been previously established: ocular comorbidity, visual acuity of the eye with the cataract and the contralateral eye, visual function, anticipated surgical complexity of the procedure, laterality of the cataract and anticipated postoperative visual acuity [17]. A data collection manual was elaborated providing clear definitions of each variable. The data collection form and abstraction manual were refined during a pilot study prior to the beginning of the study.

Sociodemographic data (educational status, work status, civil status and living condition) were retrieved from the questionnaire completed by patients prior to the surgery.

The main study outcome was the proportion of surgical interventions in each of the four appropriateness categories (necessary, appropriate, uncertain and inappropriate). The secondary outcome was the prevalence of clinical scenarios encountered in clinical practice, among all of the scenarios developed and rated by the panel of experts.

This study was approved by the ethics committees of the participating centres. The confidentiality of data was maintained in compliance with the Spanish Law of Data Protection (O.L. 15/1999).

Statistical analysis

The unit of analysis was the patient. When two operations had been performed in the same patient during the study period, only the first operation was included in the analysis.

Descriptive analysis was performed and categorical variables were summarized as frequencies and percentages and continuous variables as means and standard deviations (SD). 95% confidence intervals (CI) were calculated for the appropriateness categories.

We compared the characteristics of included patients and those that were excluded because they did not match any of the scenarios or because of lack of the necessary information to assign them to a clinical scenario. The relationship between appropriateness and clinical variables was analysed. Continuous variables were compared by means of the analysis of variance and categorical variables with the chi-square test. P-values <0.05 were considered statistically significant.

Finally, frequencies of the most frequent scenarios in each level of appropriateness were calculated. All computations were developed using the SAS statistical software for Windows v. 9.1.


During the observation period a total of 6112 patients who met the inclusion criteria underwent phacoemulsification. Of these patients, 215 were classified into 100 clinical scenarios that were excluded from our study. Most of the excluded scenarios were deemed to be rare or very infrequent in clinical practice, such as patients with bilateral cataracts having a higher visual acuity in the eye to be operated than in the contralateral eye and patients with a visual acuity of ≤0.1 in both eyes and no difficulties in their daily living activities (ADL). There were 455 patients excluded because they lacked information on a variable required to assign them to a clinical scenario. Patients that did not match any of the clinical scenarios had a higher mean age than those included in the study, but there were no other statistically significant differences in the comparison of sociodemographic characteristics (Table 1). The final study sample was composed of 5442 patients. 58.5% of them were women and mean age 73.2 years (SD 8.9).

View this table:
Table 1

Sociodemographic characteristics of the patients

Patients included (n = 5442)aPatients with insufficient informationb (n = 455)aPatients no matching any of the scenarios (n = 215)aP value
Age Embedded Image (SD)73.2 (8.9)73.2 (9.6)75.4 (8.1)0.003
Sex (%)Women58.559.654.40.43
Education (%)Secondary/university15.412.318.10.33
Civil status (%)Married/living with partner61.961.158.00.61
Work status (%)Active worker5.34.550.41
Transitory inability to work1.11.70
Work in the household26.930.922.8
Living condition (%)Living alone18.218.925.90.06
Living with partner79.580.671.3
In a retirement home1.20.62.2
  • aThe number of patients (n) may be lower in some variables due to missing values. b‘Patients with insufficient information’ refers to those which could not be included in the final analysis because of the gathering of insufficient information for a clinical scenario to be assigned.

The majority of the patients (77.0%) had simple cataracts (Table 2). In 60% of all patients the visual acuity of the cataract involved eye was 0.2 to 0.4, and in 70% of those with a non-simple cataract a post-surgery visual acuity of ≥0.5 was anticipated. In 81% of cases the anticipated surgical complexity was low. Of the 5442 first surgeries performed, the use of phacoemulsification was classified as necessary in 39.9% [95% CI: 38.5–41.2], appropriate in 29.8% [95% CI: 28.5–31.0], inappropriate in 7.3% [95% CI: 6.7–8.1] and uncertain in 23.0% [95% CI: 21.9–24.2].

View this table:
Table 2

Distribution of interventions by level of appropriateness according to preoperative characteristics

Number of patients (%)Necessary (%)Appropriate (%)Uncertain (%)Inappropriate (%)
Type of cataract
 Simple4191 (77.0)48.925.721.73.6
 Associated with diabetic retinopathy192 (3.5)10.442.229.717.7
 Associated with some other ocular disease1059 (19.5)9.343.527.020.2
 Unilateral2608 (47.9)34.830.825.09.4
 Bilateral2834 (52.1)44.528.821.25.5
Preoperative visual acuity of the cataractous eye
 ≥0.5814 (15.0)022.652.025.4
 0.2–0.43268 (60.0)40.133.621.84.5
 ≤0.11360 (25.0)
Anticipated postoperative visual acuitya
 ≥0.5876 (70.0)061.222.616.2
 0.2–0.4298 (23.8)
 ≤0.177 (6.2)0050.749.3
Preoperative visual acuity of the contralateral eye
 ≥0.53289 (60.4)28.632.030.19.3
 0.2–0.41948 (35.8)56.427.511.84.3
 ≤0.1205 (3.8)62.914.617.15.4
Anticipated surgical complexity
 Low/no complexity4396 (80.8)40.930.122.66.4
 Medium545 (10.0)35.431.223.310.1
 High501 (9.2)35.125.826.712.4
Visual function
 No impairment589 (10.8)06.552.940.6
 Glare1249 (23.0)6.746.441.15.8
 Recreational difficulties1584 (29.1)45.730.718.84.7
 Difficulties with activities of daily living2020 (37.1)67.425.56.40.6
  • aOnly for patients with diabetic retinopathy or some other associated ocular disease (n = 1243). All P-values are <0.001.

Table 2 shows the degree of appropriateness of the interventions with respect to the patients' clinical variables. The largest difference in the proportion of patients with an inappropriate surgery was seen in the following comparisons: a) between simple cataracts and cataracts associated with other ocular disease; b) between patients with preoperative visual acuity of the cataract involved eye <0.5 and those with visual acuity ≥0.5; c) between patients with anticipated postoperative visual acuity of >0.1 and those with anticipated visual acuity of ≤0.1; d) between low and high anticipated surgical complexity, and; e) between cataracts causing no functional impairment and cataracts causing glare or difficulties with recreational or daily living activities (ADL). Differences in the proportion rated inappropriate were lower in comparisons between unilateral and bilateral cataracts or among the degrees of contralateral visual acuity.

In all, 400 interventions were classified as inappropriate with 62.0% involving cataracts associated with other ocular disease, 61.3% unilateral and 15.5% involving highly complex surgery. With respect to visual acuity, it was ≥0.5 in the cataract-involved eye in 51.8% of patients and ≥0.5 in the contralateral eye in 76.3% of patients. In 59.8% there was no limitation of visual function. Among those inappropriate interventions involving cataracts associated with other ocular disease, the anticipated postoperative visual acuity was ≤0.1 in 15.7% of cases, between 0.2 and 0.4 in 27.0% and ≥0.5 in 57.3%. In contrast, the 2169 interventions classified as necessary involved cataracts associated with other ocular disease in 5.4% of cases. The visual acuity was ≥0.5 in the operated eye in no case and in the contralateral in 43.4%, and the anticipated postoperative visual acuity in those with cataracts associated with other ocular disease was always >0.1. With respect to visual function, 62.8% of necessary interventions were done in patients with difficulties in ADL and none in patients with no impairment. In the case of interventions classified as appropriate, in 31.8% of them patient had difficulties in ADL and the percentage with no impairment was 2.3%. Characteristics of interventions classified as uncertain were intermediate between those of appropriate and those of inappropriate interventions.

Of the 765 clinical scenarios for which the appropriateness of phacoemulsification had been established a priori, 429 (56.1%) were encountered in our study. Table 3 shows the most common scenarios at each level of appropriateness (necessary, appropriate, uncertain and inappropriate). In the most common scenarios the degree to which visual function was limited was the variable most closely associated with the appropriateness of the procedure. In general, phacoemulsification was necessary in cases in which visual function was most affected; the patients involved experienced difficulties with ADL as well as recreational difficulties (e.g. reading or watching television). In those scenarios in which the procedure was deemed appropriate, the patients suffered less impairment of visual function (two of the most common of these scenarios involved glare as the only limitation). Interventions deemed uncertain involved even less loss of visual function or visual acuity, while those deemed inappropriate involved patients whose visual function was not affected and in whom visual acuity in the cataract involved eye was ≥0.5 or they had some other ocular disease associated with the cataract that limited the visual prognosis.

View this table:
Table 3

Description of the three most common scenarios per level of appropriateness

n%Preoperative visual acuity of the contralateral eyePreoperative visual acuity of the cataractous eyeType of cataractLateralityAnticipated surgical complexityVisual function
Necessary (n = 2169)127712.80.2–0.40.2–0.4SimpleBilateralLow/no complexityDifficulties with activities of daily living
223110.70.2–0.40.2–0.4SimpleBilateralLow/no complexityRecreational difficulties
321810.1≥0.50.2–0.4SimpleUnilateralLow/no complexityRecreational difficulties
Appropriate (n = 1620)126616.4≥0.50.2–0.4SimpleUnilateralLow/no complexityGlare
217110.6≥0.50.2–0.4SimpleBilateralLow/no complexityRecreational difficulties
31157.10.2–0.40.2–0.4SimpleBilateralLow/no complexityGlare
Uncertain (n = 1253)115112.1≥0.50.2–0.4SimpleUnilateralLow/no complexityNo impairment
21149.1≥0.5≥0.5SimpleUnilateralLow/no complexityGlare
31068.5≥0.50.2–0.4SimpleBilateralLow/no complexityGlare
Inappropriate (n = 400)16416≥0.5≥0.5SimpleUnilateralLow/no complexityNo impairment
2276.3≥0.5≥0.5SimpleBilateralLow/no complexityNo impairment
3164≥0.50.2–0.4Associated with some other ocular diseaseaUnilateralLow/no complexityNo impairment
  • n, number of interventions assigned to each clinical scenario; %, percentage of interventions of the total at each level of appropriateness.

  • aThe ‘anticipated postoperative visual acuity’ in this scenario was ≥0.5.


The high prevalence of cataracts, their impact on the quality of life, and the frequency with which cataract surgery is performed, highlight the need to assess the degree to which the use of phacoemulsification is appropriate. Given the important benefits that can be obtained from cataract surgery it should be performed in nearly all cases that meet clinical criteria for necessity, and in the majority of those cases in which it would be considered appropriate. However, despite the relatively low risk of complications with phacoemulsification, it should not be offered to patients who are unlikely to gain sufficient benefit to outweigh the risk involved.

At 7.3%, the percentage of patients in which phacoemulsification was deemed inappropriate was low compared with that recorded for some procedures investigated using this same methodology (e.g. percutaneous transluminal coronary angioplasty) [20], and similar to that recorded for yet others (e.g. hip arthroplasty) [21]. Although this percentage is relatively small compared with the percentage reported for some other conditions, the absolute number of phacoemulsifications performed annually is large, making it important to use strategies to reduce this figure. The scenarios in which phacoemulsification was deemed inappropriate usually involved patients whose visual function was not impaired (239 of the 400 inappropriate cases), who showed only a slight reduction in visual acuity (91 cases), or in whom the expected visual improvement was limited due to the presence of ocular comorbidity (38 cases). Especially careful assessment of patients with these characteristics should be performed in order to avoid futile or harmful interventions.

Surgeon enthusiasm for a technique and patient preferences have been proposed to explain the overuse of surgery in general. These factors might also be applicable to phacoemulsification. Some inappropriate surgeries may be performed because some specialists judge that phacoemulsification carries such a low risk that it can be justified even when the loss of visual acuity is small and the potential benefit is limited. For 74 patients with anticipated postoperative visual acuity higher than the preintervention visual acuity, in which surgery was deemed inappropriate, phacoemulsification may have been selected solely on criteria related to visual acuity, without taking accounting for visual function.

In some cases, such as in populations with a high sociocultural level, patient pressure could lead to the inappropriate use of a procedure.

The literature contains few studies of the appropriateness of phacoemulsification using criteria developed by the RAND/UCLA methodology. However, the overall results obtained—69.6% appropriate and 7.3% inappropriate—are very similar to those reported by Choi et al. [22] (77% appropriate and 7.7% inappropriate) for South Korea, but quite different from the 91 and 2%, respectively reported by Tobacman et al. [23] for the US. These discrepancies might be explained by the surgical technique employed. Tobacman et al. describe the situation in 1990, before the use of phacoemulsification became widespread. In contrast, in the work of Choi et al., performed in 1997, phacoemulsification was used in 88% of the interventions and the present study included only phacoemulsification procedures. The smaller risk associated with phacoemulsification compared with extracapsular cataract extraction (the most common technique used at the time of the Tobacman study) may have lead some specialists to expand the range of patients they consider to have an indication; they may therefore have used the technique to treat patients for whom the anticipated benefits were small and insufficiently guaranteed.

Some 96.5% of the interventions encountered in our study had been contemplated in the original list of 765 clinical scenarios. This finding reflects the exhaustive nature of the scenarios assessed by the panel of experts for the definition of the appropriateness criteria. In addition, the specificity of our list of scenarios was also high. Among the 765 scenarios, 56% were encountered between the recorded interventions, compared with the 11% of the list of scenarios in the study by Tobacman [23].

Our study has some limitations. The appropriateness of phacoemulsification for cataract treatment was assessed using reference criteria developed by a panel of ophthalmologists with recognized expertise in the field, following the RAND/UCLA methodology [17]. The validity of this methodology for the development of appropriateness criteria for medical and surgical procedures has been reported for several interventions [2426], including cataract surgery [27, 28]. Still, the percentages of appropriateness should be interpreted in the context of this methodology. Some of the interventions classified as inappropriate could have been appropriately indicated because of the presence of other individual factors not taken into account in our appropriateness criteria. Data collection was carried out by a variety of personnel and in some cases extracted from clinical records so the quality of data may have been influenced by the experience or motivation of those involved. To try to minimize these problems, however, the persons who collected data were trained and had clear operating guidelines. In addition, the percentage of cases in which sufficient information was available for a clinical scenario to be assigned was high at 93%, indeed much higher than in other studies (43 [22] and 33% [23]). With respect to the external validity of the results, extrapolating them to other regions of Spain or other countries should be undertaken with caution as there may be differences in the services offered that have an effect on the degree of inappropriateness of procedures.

Even though it was not the aim of this study, measuring the underuse of procedures is an important issue related to the appropriateness of health interventions. We have not found studies addressing underuse in cataract surgery, probably due to difficulties related to the identification of patients with undiagnosed cataract that need surgery. This would require the implementation of a study to diagnose cataract at the population level, which would probably be unfeasible. An indirect way to address the underuse of cataract surgery could be to measure the proportion of patients that have been operated in an advanced stage, when suffering an extremely low visual acuity with important limitations in their ADL.

In conclusion, the appropriateness of cataract extractions performed by phacoemulsification in the studied setting is high, although there is an opportunity to further reduce the percentage of cataract procedures rated inappropriate. A reasonable goal would be to reduce the percentage of inappropriate procedures to <5% without underusing the technique. The adaptation of appropriateness criteria to clinical practice guidelines and the production of electronic aids to decision-making could assist in reaching this goal.


This study supported in part by grants from the Fondo de Investigación Sanitaria (PI03/0550, PI03/0724, PI03/0471, PI03/0828, PI04/1577), the Department of Health of the Basque Country and the thematic networks—Red IRYSS—of the Instituto de Salud Carlos III (G03/220). SP is supported by a research contract from the Instituto de Salud Carlos III.


We are grateful for the support of the staff members of the different services, research and quality units, as well as the medical records sections of the participating hospitals and the professionals who participated in the data collection.


The IRYSS Cataract Group included the following co-investigators: Dr. Jesús Martínez-Tapias, Dr. Eduardo Aguayo (Hospital Universitario Virgen de las Nieves, Granada); Dr. Emilio Perea-Milla (Hospital Costa del Sol-CIBER Epidemiología y Salud Pública (CIBERESP), Málaga); Dr. Eduardo Briones (Hospital Universitario Virgen de Valme, Sevilla); Dr. Juan R. Lacalle (Facultad de Medicina, Universidad de Sevilla); Dr. Marisa Baré, (Corporació Sanitaria Parc Taulí, Sabadell); Dr. Elena Andradas (Agencia Laín Entralgo, Madrid); Dr. Inmaculada Aróstegui (Departamento de Matemática Aplicada, UPV-CIBER Epidemiología y Salud Pública (CIBERESP)); Amaia Bilbao, Berta Ibáñez (Fundación Vasca de Innovación e Investigación Sanitarias-(BIOEF)-CIBER Epidemiología y Salud Pública (CIBERESP), Bizkaia); Dr. Txomin Alberdi (Servicio de Oftalmología, Hospital de Galdakao, Bizkaia); Dr. Sandra de Fernando (Servicio de Oftalmología, Hospital de Cruces, Bizkaia); Dr. Javier Mendicute (Servicio de Oftalmología, Hospital de Donostia, Gipuzkoa); Dr. Jose M. Begiristain (Dirección Territorial de Gipuzkoa); Dr. Idoia Garai (Dirección Territorial de Bizkaia); Dr. Josefa Pérez de Arriba (Dirección Territorial de Araba); Dr. Felipe Aizpuru (Unidad de Investigación del Hospital de Txagorritxu-CIBER Epidemiología y Salud Pública (CIBERESP), Araba); Dr. José Ignacio Pijoan (Unidad de Investigación, Hospital de Cruces-CIBER Epidemiología y Salud Pública (CIBERESP), Bizkaia); Dr. Antonio Escobar. (Unidad de Investigación, Hospital de Basurto-CIBER Epidemiología y Salud Pública (CIBERESP), Bizkaia); Dr. Jose M. Quintana, Nerea González, Iratxe Lafuente (Unidad de Investigación del Hospital Galdakao-Usansolo-CIBER Epidemiología y Salud Pública (CIBERESP), Bizkaia).


  • The co-investigators of the IRYSS Cataract Group are listed in Appendix.


View Abstract