Although cataract surgery is a safe1 and cost-effective2 surgical procedure, it still poses a significant burden on the healthcare system in terms of cost and operative resources due to the frequency with which it is performed.3 Cataract prevalence is predicted to increase dramatically in the coming decades, which will further increase the burden of this procedure on the healthcare system.4 Operating room (OR) time in particular is a scarce and valuable resource, with case delays, complications, and scheduling inefficiency having a direct impact on the revenue generated for the institution. OR time cost per minute in cataract surgery has been estimated at $11.24 USD5; thus, even minor delays can have a substantial financial impact.
Recently, increasing attention has been paid to improving the efficiency of cataract surgery with the goal of decreasing unit cost per surgery and optimizing the use of OR time.6–9 Efforts thus far have primarily focused on the preoperative identification of cataract surgeries at high risk for intraoperative complications. These complications are potentially devastating in terms of visual outcome for the patient and costly in terms of OR delays and requirements for secondary surgeries. Established preoperative cataract surgery risk-stratification systems may correlate with operative time.7 There is also evidence that surgeon-related factors are an important determinant of operative time.10 A number of studies have explored the potential to predict operative time in various surgical procedures.8,11–13 However, to our knowledge, there have been no specific large-scale efforts to predict operative time in cataract surgery.
Predictive modeling of cataract surgery duration may help improve operative scheduling, thereby promoting the efficient use of OR time. The current strategy for operative scheduling relies on surgeon prediction for case duration, a method of unknown accuracy, and a high degree of intersurgeon variability. The objective of this study was to identify relevant patient-related factors and clinical features that are strongly correlated with operative time for cataract surgery using a large-scale retrospective cohort of patients.
Institutional review board approval was obtained through the Massachusetts Eye and Ear Human Studies Committee, and a waiver of patient consent was obtained given its retrospective nature.
All cases of phacoemulsification were reviewed with intraocular lens insertion performed by 10 Massachusetts Eye and Ear attending cataract surgeons in the Comprehensive Ophthalmology Service between January 1, 2014, and December 31, 2014. Current Procedural Terminology codes 66982 (extracapsular cataract extraction with the insertion of intraocular lens prosthesis, complex) and 66984 (extracapsular cataract extraction with the insertion of intraocular lens prosthesis) were used to identify eligible cases. We only included cases in which phacoemulsification was used for intraocular lens extraction.
We excluded all cases which were primarily attended by resident physicians, because of the association of these cases with longer operative times. The Accreditation Council for Graduate Medical Education case log was reviewed and cross-referenced for cases performed by resident surgeons. Cases marked by residents as primary were considered to be resident-performed cases, whereas cases not logged as resident primary surgeries were considered to be attending cases. Combined cataract surgery cases with planned additional procedures (Descemet stripping endothelial keratoplasty, glaucoma procedures, or pars plana vitrectomy [PPV]) were also eliminated from the analysis.
Four members of the study team reviewed the medical records of all eligible cases. To ensure that data were extracted in a standardized way, all 4 reviewers received training from a study investigator. Baseline data were collected at the preoperative visit closest to the date of surgery.
Operative time data recorded by OR staff at the time of surgery included the date of the surgical case, the time of the day of the beginning and end of the case, operative time (excluding time under direct control of the anesthesiologist), and the total time the patient spent in the OR. Scheduled case duration was provided as standard practice preoperatively by attending surgeons for OR scheduling. Predicted case duration was provided in 15-minute increments as 30, 45, or 60 minutes. Operative time data were merged with the existing clinical data from chart review. The order of the surgical case in the OR schedule was determined using the start time of the case compared with other surgical cases performed that day by the same attending surgeon.
Patient-related, surgeon-related, and other clinical factors were identified that could potentially be associated with operative time based on the previously published reports and experience of our group. Variables included in the analysis were as follows: age (greater than or less than 90 years),14,15 sex,14 body mass index (greater than or less than 30),16 first or second eye, operative eye, advanced cataract,17–20 iris hooks/Malyugin ring (Microsurgical Technology) (poor pupillary dilation), use of trypan blue, diabetic retinopathy,14 presence of pseudoexfoliation (PXF),21,22 axial length (≤22.4, 22.5 to 25.9, or ≥26 mm),15 anterior chamber (AC) depth (≤2.4, 2.5 to 3.9, or ≥4 mm),21 history of alpha-blocker use,23–27 prior PPV,28 history of glaucoma,14 prior ocular surgery, postgraduate year (PGY) level of the assisting resident trainee (PGY2, PGY3, or PGY4),29 years of attending surgeon experience after residency training,10 attending surgeon identity (coded as a unique numerical identifier),10,30 and order of surgery in the OR schedule.31
All statistical analyses were performed using R (R Core Team, 2017) and SAS 9.4 (SAS Institute). Owing to the highly right-skewed distribution of operative time, the analysis was restricted to cases with operative times within 2 standard deviations above the mean operative time (≤46 minutes).
Bivariate analyses were conducted using the t test and analysis of variance to determine which factors were independently associated with the mean operative time. Multivariate linear regression with backward elimination was then used to identify factors that were significantly associated with operative time, after adjusting for other factors included in the model. A liberal threshold was set for statistical significance (P < .2) to include predictors that were found to be both strongly and moderately associated with operative time in the initial model. Variables were removed from the model one at a time, until all but 2 variables were significantly associated with the outcome at α = 0.20. Although not significantly associated with operative time after adjusting for other predictors, we retained presence of PXF and history of PPV in the final model because of their generally accepted association with case complexity and increased risk for complication.21,22,32–34
Two independent models were developed with respect to the surgeon performing the surgery: excluding the identity of the attending surgeon but replaced with the surgeon's years of experience and PGYs of training of the surgeon's trainee (Model 1) and using surgeon identity (Model 2). These models were developed for 2 primary reasons. First was the ability to generalize the model to other institutions in which surgeons will differ. Second was to quantify the effect of the surgeon independent of their or their trainee's experience.
Of the 1931 cataract surgeries reviewed, 1349 cataract surgeries in 1072 unique patients were included in the final analysis. Case characteristics are summarized in Table 1. The mean operative time was 22.1 ± 7.8 minutes. The range of mean operative times associated with each attending spanned nearly 10 minutes, from 17.2 to 27 minutes for the surgeon with the shortest and longest operative times, respectively. Of the included patients, 5 experienced posterior capsular tear and 4 of these underwent anterior vitrectomy. One additional patient underwent anterior vitrectomy in which no posterior capsular violation was noted intraoperatively.
Preoperative and intraoperative clinical characteristics associated with operative time in the bivariate analysis are summarized in Table 2. Clinical characteristics associated with longer operative time in both Model 1 and Model 2 were male sex, body mass index greater than 30, first-eye surgery, left operative eye, advanced cataract, use of iris hooks, use of Malyugin ring, use of trypan blue, history of diabetic retinopathy, and shorter axial length (<22.5 mm). Additional significant predictors of longer operative time in Model 1 were more advanced level of trainee experience and less attending experience after residency training. Additional significant predictors of longer operative time in Model 2 were shallow AC depth (<2.5 mm) and intersurgeon variability.
Scheduled case duration provided by the surgeon was compared with the operative time and total time in the OR (Table 3). Most cases were scheduled for 30, 45, or 60 minutes. The analysis of variance demonstrated a significant difference in the operative time and total case time based on scheduled case duration (P < .001 for both). The shorter operative time and total OR time were associated with cases scheduled for 30 minutes compared with 45 or 60 minutes (P < .001). There was no significant difference between 45-minute and 60-minute scheduled cases.
The R2 value for Model 2 (R2 = 0.42) was significantly higher than that calculated for Model 1 (R2 = 0.23), that is, more of the variability in operative time was accounted for in Model 2. The most significant difference between these models was attending and resident experience in Model 1 replaced with surgeon identity in Model 2, suggesting intersurgeon variability independent of experience is an important predictor of operative time.
Variables not significant enough for inclusion in the regression models or dropped from the models during backward selection are presented in Table 4. History of alpha-blocker use, although significantly correlated with longer operative time in the bivariate analysis, was dropped from both Model 1 and Model 2 during the selection process. Advanced age and case number of the day were not independently associated with operative time.
Notably, although PXF and history of PPV were associated with prolonged case time in the bivariate analysis, neither reached significance in the multivariate analysis. A high degree of correlation was found between PXF and use of ancillary devices including trypan blue, Malyugin ring, and iris hooks. Thus, when we controlled for these factors, PXF was no longer significantly associated with increased operative time. Both prior PPV and history of PXF were forced into the final models because of their generally accepted association with complicated cataract surgery.
In this study, bivariate and multivariate analyses were used to identify preoperative clinical characteristics affecting operative time in cataract surgery and to build a predictive model of cataract surgery operative time. Although much of the published literature has examined risk factors for complications in cataract surgery, there is only limited information on factors affecting operative time.7,10,30,31,35,36 More accurate prediction of operative time for cataract surgery has the potential to save OR resources and efficiently use surgeon and patient time.
There was a limited correlation between surgeon-planned and actual utilization of OR time. We found no significant difference in the operative time or total OR time between cases with the scheduled duration of 45 minutes and 60 minutes. This contrasted with a significantly shorter operative time and total OR time (both by a mean difference of approximately 6 minutes) for cases scheduled for 30 minutes. These differences in the total case length, albeit small on a case-by-case basis, can cause significant cumulative delays over the course of the full day in the OR.
Perhaps not surprisingly, the strongest predictor of operative time in our analysis was related to the identity of the attending surgeon. Furthermore, a large degree of the overall variability in operative time was explained by the identity of the surgeon as evidenced by the increased R2 value in Model 2 (R2 = 0.42) as compared with Model 1 (R2 = 0.23). This suggests that nearly 20% of the variability in operative time is accounted for by a “surgeon factor” (ie, surgeon identity alone) and that this variability is not accounted for by the surgeon's years of experience or by the level of the surgeon's trainee. Although the inclusion of attending surgeon identity lessens this model's generalizability to other institutions, we have included an alternative analysis with variables applicable to all centers.
In Model 1, both attending years of experience and PGY level of the resident serving as the first assistant in the surgery were significantly associated with operative time. Attending surgeons had a significantly and progressively shorter operative time with greater years of experience. Resident trainees demonstrated the opposite effect with increasing PGY associated with longer operative times. This is likely due to the greater involvement of more senior residents in cataract cases. Although primary resident surgeries were excluded from the analysis, assistant cases were not. Whether a case was resident assisted, and the degree to which it was, is not reliably recorded in our institution. Therefore, this variable was not included in the analysis but likely has an impact on operative times.
Predictably, ancillary instruments (iris hooks, Malyugin ring, or trypan blue) used during surgery were highly correlated with longer operative time in both models. For the management of poor pupillary dilation, use of iris hooks was associated with longer operative time than Malyugin ring. Nderitu et al.36 similarly reported longer operative time with pupillary expansion devices with additional operative time of 14 minutes for iris hooks and 4 minutes for pupil expansion ring. Usefulness of these data is dependent on the predictability of their use preoperatively and the identification of poor pupillary dilation as a preoperative factor. Poor pupillary dilation is often but not always noted in the preoperative period. Similarly, an obscured anterior capsule necessitating the use of trypan blue may be noted in the preoperative period. Intraoperative use of trypan blue is a technically facile procedure, which itself is unlikely to add a significant amount of time to the surgical procedure. The obscuration of the anterior capsule complicating capsulorhexis is likely the true operative time predictor in this instance. Circumstances in which use of ancillary instruments is not predictable, some unforeseen variability in operative time is unavoidable.
Advanced cataract defined as brunescent, mature, Morgagnian, or white was associated with prolonged operative time. Mature and hypermature cataracts are known risk factors for complicated surgery. The Muhtaseb et al.,17 Buckinghamshire,18 and Habib et al.19 criteria all include advanced cataract in their scoring systems for risk of complication in cataract surgery. Achiron et al.7 correlated higher risk surgeries by the Muhtaseb scoring system with increased operative time. It is reasonable to suspect that more advanced cataracts would therefore be associated with longer operative times because of increased case difficulty. Although not a direct measure of operative time, there is also a reported association of nuclear density with the total phacoemulsification time.37,38
Interestingly, left operative eye was associated with longer operative time in both models. Although we are not certain of the reason for this difference, it may be due to the anatomical difficulty of a right-handed surgeon operating through superior wounds having to cross the nasal bridge for access. All surgeons in this cohort were right handed, so the subgroup analysis could not be conducted.
Limitations of this study were its retrospective nature and inclusion of cases from a single academic center. Owing to the inclusion of cases with assisting residents, there was a variable degree of trainee involvement, which we could not control for. Factors that are difficult to measure, such as day-to-day variation in surgeon performance, degree of patient cooperation, and equipment malfunctions, were not taken into account. The inclusion of cases such as patients with PXF syndrome and eyes with a history of PPV undoubtedly injected a degree of variability in operative time that may limit the accuracy of the predictive model. However, we aimed to include variables that are commonly collected in clinical practice, making our model practical and applicable in an average operative suite. Furthermore, there is a great deal of coordination at the system level and efforts by the OR staff, which contribute to the efficient utilization of operative resources. This study did not examine these factors, although they are likely an important element to predicting operative time in any surgical procedure. We would predict these factors to be highly institution dependent and, therefore, may require institution-specific data to be addressed.
Strengths of this study include the identification of factors affecting operative time from raw clinical data and limiting reliance on the previously published risk factors for complication. Future work is necessary to validate the operative time factors and to test the models prospectively.
There is a limited correlation between surgeon-planned and actual utilization of OR time. Surgeon factors, iris challenges, and advanced cataract were most strongly correlated with longer cataract surgery duration. The identification of these and other risk factors may assist surgeons in accurate case scheduling and contribute to more efficient utilization of operative resources.
WHAT WAS KNOWN
- A number of studies have sought to identify preoperative risk factors for complications in cataract surgery.
- Far less is known about the factors affecting operative time in cataract surgery and what preoperative clinical features affect operative time.
WHAT THIS PAPER ADDS
- There was a limited correlation between surgeon-planned and actual utilization of operating room time.
- This is the first large-scale effort to identify the patient-related factors that affect operative time in cataract surgery.
- Two predictive models are presented that may inform more accurate utilization of operating room time.
1. Shaw AD, Ang GS, Eke T. Phacoemulsification complication rates. Ophthalmology 2007;114:2101–2102
2. Lansingh VC, Carter MJ, Martens M. Global cost-effectiveness of cataract surgery. Ophthalmology 2007;114:1670–1678
3. Rein DB, Zhang P, Wirth KE, Lee PP, Hoerger TJ, McCall N, Klein R, Tielsch JM, Vijan S, Saaddine J. The economic burden of major adult visual disorders in the United States. Arch Ophthalmol 2006;124:1754–1760
4. Congdon N, Vingerling JR, Klein BE, West S, Friedman DS, Kempen J, O'Colmain B, Wu SY, Taylor HR; Eye Diseases Prevalence Research Group. Prevalence of cataract and pseudophakia/aphakia among adults in the United States. Arch Ophthalmol 2004;122:487–494
5. Taravella MJ, Davidson R, Erlanger M, Guiton G, Gregory D. Time and cost of teaching cataract surgery. J Cataract Refract Surg 2014;40:212–216
6. De Regge M, Gemmel P, Duyck P, Claerhout I. A multilevel analysis of factors influencing the flow efficiency of the cataract surgery process in hospitals. Acta Ophthalmologica 2016;94:31–40
7. Achiron A, Haddad F, Gerra M, Bartov E, Burgansky-Eliash Z. Predicting cataract surgery time based on preoperative risk assessment. Eur J Ophthalmol 2016;26:226–229
8. Edelman ER, van Kuijk SMJ, Hamaekers AEW, de Korte MJM, van Merode GG, Buhre WFFA. Improving the prediction of total surgical procedure time using linear regression modeling. Front Med 2017;4:85
9. Jones JJ, Chu J, Graham J, Zaluski S, Rocha G. The impact of a preloaded intraocular lens delivery system on operating room efficiency in routine cataract surgery. Clin Ophthalmol 2016;10:1123–1129
10. Strum DP, Sampson AR, May JH, Vargas LG. Surgeon and type of anesthesia predict variability in surgical procedure times. Anesthesiology 2000;92:1454–1466
11. Eijkemans MJ, van Houdenhoven M, Nguyen T, Boersma E, Steyerberg EW, Kazemier G. Predicting the unpredictable: a new prediction model for operating room times using individual characteristics and the surgeon's estimate. Anesthesiology 2010;112:41–49
12. Wright IH, Kooperberg PC, Bonar BA, Bashein G. Statistical modeling to predict elective surgery time comparison with a computer scheduling system and surgeon-provided estimates. Anesthesiology 1996;85:1235–1245
13. Laskin DM, Abubaker AO, Strauss RA. Accuracy of predicting the duration of a surgical operation. J Oral Maxill Surg 2013;71:446–447
14. Narendran N, Jaycock P, Johnston RL, Taylor H, Adams M, Tole DM, Asaria RH, Galloway P, Sparrow JM. The Cataract National Dataset electronic multicentre audit of 55,567 operations: risk stratification for posterior capsule rupture and vitreous loss. Eye 2009;23:31–37
15. Fesharaki H, Peyman A, Rowshandel M, Peyman M, Alizadeh P, Akhlaghi M, Ashtari A. A comparative study of complications of cataract surgery with phacoemulsification in eyes with high and normal axial length. Adv Biomed Res 2012;1:67
16. Mansour AM, Al-Dairy M. Modifications in cataract surgery for the morbidly obese patient. J Cataract Refract Surg 2004;30:2265–2268
17. Muhtaseb M, Kalhoro A, Ionides A. A system for preoperative stratification of cataract patients according to risk of intraoperative complications: a prospective analysis of 1441 cases. Br J Ophthalmol 2004;88:1242–1246
18. Butler TK. Risk stratification and assessment in cataract surgery. J Cataract Refract Surg 2012;38:184
19. Habib MS, Bunce CV, Fraser SG. The role of case mix in the relation of volume and outcome in phacoemulsification. Br J Ophthalmol 2005;89:1143–1146
20. Chakrabarti A, Singh S. Phacoemulsification in eyes with white cataract. J Cataract Refract Surg 2000;26:1041–1047
21. Kuchle M, Viestenz A, Martus P, Handel A, Junemann A, Naumann GO. Anterior chamber depth and complications during cataract surgery in eyes with pseudoexfoliation syndrome. Am J Ophthalmol 2000;129:281–285
22. Drolsum L, Haaskjold E, Sandvig K. Phacoemulsification in eyes with pseudoexfoliation. J Cataract Refract Surg 1998;24:787–792
23. Oshika T, Ohashi Y, Inamura M, Ohki K, Okamoto S, Koyama T, Sakabe I, Takahashi K, Fujita Y, Miyoshi T, Yasuma T. Incidence of intraoperative floppy iris syndrome in patients on either systemic or topical alpha(1)-adrenoceptor antagonist. Am J Ophthalmol 2007;143:150–151
24. Chang DF, Campbell JR. Intraoperative floppy iris syndrome associated with tamsulosin. J Cataract Refract Surg 2005;31:664–673
25. Chadha V, Borooah S, Tey A, Styles C, Singh J. Floppy iris behaviour during cataract surgery: associations and variations. Br J Ophthalmol 2007;91:40–42
26. Gani J, Perlis N, Radomski SB. Urologic medications and ophthalmologic side effects: a review. Can Urol Assoc J 2012;6:53–58
27. Haridas A, Syrimi M, Al-Ahmar B, Hingorani M. Intraoperative floppy iris syndrome (IFIS) in patients receiving tamsulosin or doxazosin-a UK-based comparison of incidence and complication rates. Graefes Arch Clin Exp Ophthalmol 2013;251:1541–1545
28. Diaz Lacalle V, Orbegozo Garate FJ, Martinez Alday N, Lopez Garrido JA, Aramberri Agesta J. Phacoemulsification cataract surgery in vitrectomized eyes. J Cataract Refract Surg 1998;24:806–809
29. Randleman JB, Wolfe JD, Woodward M, Lynn MJ, Cherwek DH, Srivastava SK. The resident surgeon phacoemulsification learning curve. Arch Ophthalmol 2007;125:1215–1219
30. Thanigasalam T, Reddy SC, Zaki RA. Factors associated with complications and postoperative visual outcomes of cataract surgery; a study of 1,632 cases. J Ophthalmic Vis Res 2015;10:375–384
31. Gupta D, Taravati P. Effect of surgical case order on cataract surgery complication rates and procedure time. J Cataract Refract Surg 2015;41:594–597
32. Scorolli L, Scorolli L, Campos EC, Bassein L, Meduri RA. Pseudoexfoliation syndrome: a cohort study on intraoperative complications in cataract surgery. Ophthalmologica 1998;212:278–280
33. Ahfat FG, Yuen CH, Groenewald CP. Phacoemulsification and intraocular lens implantation following pars plana vitrectomy: a prospective study. Eye 2003;17:16–20
34. Braunstein RE, Airiani S. Cataract surgery results after pars plana vitrectomy. Curr Opin Ophthalmol 2003;14:150–154
35. Hosler MR, Scott IU, Kunselman AR, Wolford KR, Oltra EZ, Murray WB. Impact of resident participation in cataract surgery on operative time and cost. Ophthalmology 2012;119:95–98
36. Nderitu P, Ursell P. Iris hooks versus a pupil expansion ring: operating times, complications, and visual acuity outcomes in small pupil cases. J Cataract Refract Surg 2019;45:167–173
37. Gupta M, Ram J, Jain A, Sukhija J, Chaudhary M. Correlation of nuclear density using the lens opacity classification system III versus Scheimpflug imaging with phacoemulsification parameters. J Cataract Refract Surg 2013;39:1818–1823
38. Davison JA, Chylack LT. Clinical application of the lens opacities classification system III in the performance of phacoemulsification. J Cataract Refract Surg 2003;29:138–145