A methodology to estimate the potential to move inpatient to one day surgery

In the last two decades, surgery carried out in outpatient settings has grown much more strongly than inpatient operations [1–3]. Day surgery is generally defined as a non emergency procedure undertaken during the period of a normal working day (not exceeding 12 hours) in hospitals or other facilities set up for this purpose. The concept does not account for minor surgery under local anesthesia usually conducted in physicians' offices [4, 5]. The increase of day surgery was driven by innovations in surgical and anesthetic techniques, together with financial incentives and patient expectations. For many procedures, day surgery appears as an effective and efficient approach, offering several advantages to patients as well as surgeons and hospital managers [6, 7]. While cost savings have been the primary incentive to day surgery development, the benefits in term of enhanced social and emotional recovery are important: less nosocomial infections, quicker return to mobility, high satisfaction in centers where the outpatient procedure has become standard practice [8, 9].

The proportion of surgery performed as a day case varies greatly between countries, however [1, 10]. For the same surgical procedure, there are large variations between countries and among hospitals in the same country [5, 11, 12]. Day surgery is not yet a common practice in many developed countries, with France [1, 10, 13, 14], Italy [1, 15], and Switzerland [16, 17], suggesting a large potential for growth.

A significant number of scientific studies deal with the safety issues and outcomes of specific surgical interventions performed as day surgery. Although interesting, this literature does not provide a complete overview and therefore cannot be used to define the field of day surgery properly.

Schematically, there are two main approaches to defining a domain: definition in extension, i.e. exemplified by a list of items, and definition in intention, based on explicit a priori criteria [18]. Some extensive lists of procedures have been proposed, mainly for reimbursement purposes [15, 19, 20]. These lists are useful because they give some evidence of the feasibility of such procedures during stays of less than 24 hours. Their main shortcoming is that they generally reflect local practices.

A definition in intention has been recently proposed by Dexter and al, based on the principle that only surgical procedures with low anesthetic complexity can be performed in ambulatory care [21]. This complexity, defined by the anesthesia workload apart from anesthesia time, is measured by the American Society of Anesthesiologists' Relative Value Guide (ASA RVG), a scale which is updated annually to reflect changes in surgical practice. For instance knee arthroscopy has three ASA RVG units, whereas a prostatectomy or a laparoscopic cholecystectomy has seven and a heart transplant 20. Practically, the authors set the maximum ASA RVG threshold for deciding if a procedure should be performed or not in an ambulatory facility at seven. Such an approach does not depend too much on local practices, but is based on a theoretically coherent point of view. As the ASA RVG is related to the widely used American Medical Association Current Procedural Terminology (CPT) codes, this approach can be compared to most empirical lists published in the USA [22]. Unfortunately, the RVG is not currently used in other countries and its adaptation to ICD-9-CM classification has several shortcomings, probably due to the fact that only frequent interventions have been extensively analyzed [21]. Many procedures performed commonly as one-day surgery are lacking (e.g. hand, foot, testicle, nerve surgery) and several eligible operations cannot be performed routinely in outpatient units (e.g. vaginal reconstruction, excision of anus, skin transplant). Moreover, anesthetic criteria are disputable to delineate day surgery, because some high risk or painful operations are associated with a relatively low anesthetic workload; for instance lyses of peritoneal adhesion (ASA RVG of 6). We think that a definition in intention should be based on the invasion degree of the surgical organ wound. Its immediate consequences, such as blood loss, postoperative pain and difficulty to recover physiological functions, are probably more accurate to determine length of stay than factors related to anesthesia [23].

We therefore proposed a new theoretical model, based on a priori criteria, allowing the setting of a list of procedures qualified for day surgery. The application of this model to all ICD-9-CM surgical procedures was then compared to a consensual list of one day surgery procedures established for Switzerland [20].

The decision to provide ambulatory surgery, or not to do so, does not depend only on the selected procedure but also on coexisting medical conditions, estimating the potential shift from inpatient to outpatient needs to add patient related criteria to surgical criteria. Two countries were compared: Switzerland where there are as yet few financial incentives to an outpatient shift [24] and the USA, where these incentives are stronger [2].

Although most health professionals and health authorities no longer dispute the benefits of day care surgery in terms of patient outcomes and costs, there is no consensus on the surgery that ought to be performed as day care. Most actual procedures lists are based on adoption curves driven by financial incentives. We propose a comprehensive list, which has strong content validity and allows the routine measure of unnecessary surgical hospitalizations.

The coherence and the content validity of the proposed list are ensured by the classification algorithm, grounded on the surgical route, the complexity of the surgical deed and the mass of injured tissue. Several procedures (see Table 4, 25, 27, 30), although featured in the Swiss reimbursement list, continue to be undertaken as inpatient surgery in many developed countries. However, due to cost containment purposes, there is a growing shift from inpatient to ambulatory settings. For instance, outpatient PTCA, performed for the first time by Kiemerneij and al in 1997, is now considered both safe and feasible for a large part of the routine candidate population [32, 33]. Similarly, although many centers routinely admit patients for overnight observation after endoscopic therapy of choledocholithiasis, six hours are sufficient to detect significant complications [34, 36].

Only 55 (2%) ICD-9-CM codes were added to form the reference list (Table 2, 1–3). About half these procedures were justified by sufficient case records in canton Vaud and did not give rise to scientific refutation because they are all related to easily accessible organs: knee surgery, operations on hands, operations on eye structures (Table 3, rows 2, 9, 14). The feasibility of most other procedures (Table 3, row 21 and Table 4, rows 28, 29) was well supported by the literature review: new pacing system implantation and replacement [36–39], tubo-ovarian laparoscopic procedures [40–42], vascular bypass for dialysis [43, 44]. As regards ambulatory orthopedic surgery (such as hallux valgus, Table 4 row 26), the limitation is often postoperative pain, the most common reason for delayed discharge or unanticipated hospital readmission [45]. However, ambulatory pain management seems increasingly to provide relief equivalent to inpatient management [46].

Most overrides of algorithm criteria leading to procedures exclusion from the reference list were related to a well-documented risk of complications requiring hospital care (Table 2, 4–8). Postoperative hemorrhage, iatrogenic organ injuries or complications requiring reoperation are frequent after endoscopic surgery on sinus, pancreatic ducts and the urinary tract [47–49]. Perforation risks following rigid endoscopy of upper aerodigestive tract (endoscopic larynx procedure like dilation, removal of foreign body) require observation exceeding eight hours [50]. Until recently, prolonged bed rest and overnight hospitalization have been the general practice after non-coronary endovascular procedures. However, the advent of new vascular devices already allows early ambulation. A same evening discharge has been proposed for selected patients after carotid endarteriectomy, since preliminary studies have shown that all severe complications (stroke, neck hematoma) occur very early, allowing restricting observation to eight hours [51]. More studies are nevertheless necessary to make recommendations. Pain management or follow up care require hospital infrastructure or prolonged bed rest after the following procedures: closed reduction of limb fractures, toe amputation, salivary fistula closure, eye evisceration. Postoperative discomfort generally justified hospital care in retinal detachment surgery: limited studies have suggested that selected patients might be treated with ambulatory vitrectomy [52]. Although peri-anal or anal incision can be achieved under local anesthesia, anal surgery (except operation on hemorrhoids) is often performed as an inpatient procedure, due to wound related problems, i.e. pain and discomfort caused by drains or irrigation. Although the advantages of day case surgery have long been established for groin hernias, the repair of umbilical hernia remains an inpatient procedure, probably because the type of repair may be influenced by the size of the defect. However, recent studies have shown the feasibility of most repair of umbilical hernia under local anesthesia as a means of reducing unacceptably long waiting times in some countries [53].

Finally, a few surgical procedures identified which qualified as day cases by the classification algorithm, were neither featured in the Swiss consensual list nor performed as day cases in canton Vaud, and were only discussed in anecdotal literature reports: corneal transplantation [54], large skin grafting [55], reconstruction of external ear [56], testicular surgery [57]. These procedures being associated with an average length of stay greatly exceeding three days they were not retained in the list.

To investigate the use of ambulatory surgery, most studies simply compare with benchmarks the proportion of specific procedures performed on outpatients. This approach is flawed because other criteria than the specific type of procedure performed may limit ambulatory use: the patient's clinical condition, other associated procedures.

The low rate of candidacy for day surgery in USA (about 4%) speaks in favor of the validity of our list of surgical procedures eligible for one day surgery. The high rate of candidacy and the variability among hospitals in Switzerland demonstrate the benefits of monitoring the potential development of one day surgery.

In Switzerland, each canton has its own hospital organization and financing rules. Some cantons (Vaud until the year 2003 for instance) have set strong financial incentives to develop one day surgery (better pay for physicians if surgery was performed in a length of stay of less than 24 hours), while other cantons did not introduce such mechanisms, financing hospitals on a length of stay basis for instance. There is no doubt that differences in resource allocation are probably the main reason for the high variability of rates of candidates for one day care [17].

The study has some limitations that do not invalidate our conclusions, but could serve to further research. First of all, we limited the measure of potential day cases to stays shorter than four days. This choice is grounded on pragmatic considerations, but the limit might be pushed to five or six days, especially in countries having longer lengths of stay. The answer to this question requires a systematic review of medical records by experienced surgeons.

Another question is related to admission circumstances. Day care could also be considered for emergency admissions, especially if they occur in the morning. Restricting the scope of one day treatment to elective referrals is probably a conservative approach. Some emergency admissions could fit all the criteria for elective day care. Most patients with a superficial abscess or minor injuries are indeed treated as inpatients, on daily priority-based emergency lists, with a more or less long pre-operative stay and even a postoperative night in hospital depending of the time of surgery. Some studies have shown the feasibility of running an emergency day surgery service for certain minor surgical emergencies [58]. However, such a procedure would have to be coordinated by a dedicated manager and thus is not widespread.

The exclusion of medical claims of hospitals with poor data quality could be a bias. However, an analysis of such excluded cases showed very similar results.

Our algorithm did not account for patients' residence. Some inhospital day stays could be justified by distance, especially if the patient cannot rely on assistance by relatives. This variable would deserve to be considered when comparing hospital performances, located in different regions (mountain, rural or urban areas). Our algorithm detects hospitalizations that are candidates for one day care. This should not lead us to conclude that all of them should be performed as day surgery. Nevertheless, our results suggest that the rate of candidates should be less than 6%.

The proposed approach allows the monitoring of the rate of inpatients stay candidates for one day surgery. Mainly based on rational and a priori criteria, the analysis showed large variations of rates between hospitals. The methodology can be used for infrastructure planning, resources negotiation and monitoring of resource utilization.