Lung cancer is the leading cause of cancer death worldwide and approximately 60 percent of patients are diagnosed in advanced stage with a 5-year survival rate of approximately 5 percent (Reference Siegel, Naishadham and Jemal1). Lung cancer patient's care is highly complex, involving several health professionals and different units. Clinical practice guidelines recommend multidisciplinary teams to be used for managing lung cancer patients (Reference Coory, Gkolia and Yang2). Furthermore, growing knowledge about cancer biology and treatment does not always directly translate into the optimal benefit for patients: evidence exists that not all patients benefit equally from innovation (Reference Coory, White and Johnson3).
A survey among medical oncology practices in Florida carried out in 2008 identified several areas as target for quality improvement in NSCLC patients; consequently, a successful audit and feedback strategy have been adopted (Reference Tanvetyanon, Corman and Lee4;Reference Tanvetyanon, Lee and Fulp5). In this scenario, integrated care pathways (ICPs) might be an adequate tool for improving patient care management. Currently, ICPs are widely used in hospitals for a structured and detailed planning of care for patients with a specific clinical problem, thus encouraging the enforcement of guidelines into daily clinical practice (Reference Campbell, Hotchkiss, Bradshaw and Porteous6). ICPs adoption could facilitate a systematic and continuous audit of clinical practices through quality indicators that investigate the three dimensions of quality: professional, organizational, and patient-oriented care (Reference Campbell, Braspenning, Hutchinson and Marshall7–Reference Hermens, Ouwens and Vonk-Okhuijsen9). Several studies have been published on ICP in lung cancer, and this appears to be a plausible strategy that addresses both the issue of quality of care and containing the containment of costs (Reference Hermens, Ouwens and Vonk-Okhuijsen9–Reference Ouwens, Hermens and Hulscher15).
We analyzed the ICP for lung cancer patients at the University Hospital of Udine, relying on the results of a previous study published in 2012 (Reference Fasola, Rizzato and Merlo16). We implemented new indicators according to the most recent scientific evidence and we modified some of the previously used indicators on the basis of the former study results.
The aims of the current study were: (i) to reassess the quality of care pathways in lung cancer patients in a university hospital; (ii) to compare the results with those of the 2008 published study (Reference Fasola, Rizzato and Merlo16); (iii) to highlight further room for improvement in clinical practice for lung cancer patients; (iv) to make the use of quality indicators a “standard practice” in the management of lung cancer patients.
MATERIALS AND METHODS
We retrospectively reviewed the electronic medical records of 169 consecutive NSCLC patients who had their first consultation at the Oncology Department of the University Hospital Santa Maria della Misericordia (Udine, Italy) during 2010.
As in our previous study (Reference Fasola, Rizzato and Merlo16), two working groups, were established: (i) a steering committee, consisting of professionals and healthcare researchers from the University Hospital Santa Maria della Misericordia and from Bocconi University (Milan, Italy), involved in study planning; (ii) a professional focus group of specialists involved in the management of NSCLC patients (i.e., medical oncologists, pulmonologists, radiologists, thoracic surgeons, pathologists, nuclear medicine physicians, and radiation oncologists). The flow chart of the ICP drawn in 2008 was reassessed, but no further change was implemented.
Selection of quality-of-care indicators: A total of fifteen indicators (Table 1) (6 more than in 2008) were selected according RAND-modified Delphi metho and taking into account their availability, reproducibility, significance, and measurability (Reference Campbell, Braspenning, Hutchinson and Marshall8).
Table 1. Study Results of 2008Footnote a and 2010 Analyses
Note. *, Significance of the difference between 2010 results and the benchmark values. °, Literature reference. ^, Focus group values as reference. #, Guidelines value as reference.
a Fasola G et al. Adopting integrated care pathways in non small cell lung cancer: from theory to practice. J Thorac Oncol. 2012;8:1283-1290.
Indicators were developed using group-facilitation techniques designed to explore the level of consensus among a group of experts (including patients representatives) and to aggregate judgments into refined agreed opinions.
These indicators were chosen as well relying on the availability of corresponding benchmark, which was derived from: (i) literature evidence, and (ii) international guidelines, published by the National Comprehensive Cancer Network (www.nccn.org), the European Society for Medical Oncology (ESMO; www.esmo.org), and the American Society of Clinical Oncology (www.asco.org); and expert opinion of the hospital's multidisciplinary lung cancer team (n = 10), which comprised medical oncologists, pulmonologists, radiologists, thoracic surgeons, pathologists, nuclear medicine physicians, and radiotherapists).
These clinicians did not participate in the Delphi survey focus group.
The total number of patients evaluated for each indicator (i.e., the denominator) can vary depending on the indicator's content. All data concern the period of 5 months ahead to 9 months after diagnosis.
Indicators related to the time elapsed were measured as median values and were compared with the median benchmark value. To compare the value of the indicators of the study population with the benchmark t-tests, proportion tests and linear regressions were used. A p-value <0.05 was considered statistically significant.
The study was approved by the local ethic committee and was supported by the hospital administration.
RESULTS
Among the 169 selected patients, 146 were considered eligible; reasons for ineligibility are highlighted in (Supplementary Figure 1). Median age was 67 years (range, 41–87 years). The majority of patients were males (65 percent) and presented with adenocarcinoma histology and advanced disease at diagnosis (52.7 percent). The stage distribution and the tumor histology in study population resembled epidemiological and literature data (18). Detailed patients’ characteristics are summarized in Table 2.
Table 2. Patient's Characteristics
Note. NOS, not otherwise specified.
Overall, seven of fifteen indicators were aligned with the benchmarks (Table 3). The observed median time between the first visit with the pulmonologist and the diagnosis (14 days; range, 0–202) was not statistically different from the 15- to 30-day benchmark value (Reference Lo, Zeldin and Skrastins13;17). Linear regression analysis highlighted that stage IIIB and IV patients had a statistically significant shorter time between the first visit with the pulmonologist and the diagnosis (p = 0.003).
Table 3. Study Results of 2010 Analyses
Note. *, Significance of the difference between 2010 results and the benchmark values. °, Literature reference. ^, Focus group values as reference. #, Guidelines value as reference.
PET, positron emission tomography; NSCLC, non small cell lung cancer; RT, radiotherapy.
The observed median time between the first visit with the pulmonologist and the first oncological referral was just within limits in patients with early stages I–IIIA, 84.5 versus 56–84 days (benchmark), whereas it is perfectly aligned for advanced stages, IIIB and IV: 29 (range = 0–261) versus 28–42 days (benchmark) (Reference Ouwens, Hermens and Termeer12).
Three of four indicators assessing the performance of the pathology department (Table 3) were in line with the benchmark values (18–Reference Edwards, Roberts and McKean23). The percentage of patients dead within 30 days from the last chemotherapy administration was 6 percent (benchmark 20 percent), and the observed percentage of patients that received more than three lines of chemotherapy for advanced disease, 2.7 percent versus 10 percent (benchmark), were in line with literature data (18,Reference Ou and Zell20,Reference Murillo and Koeller24–Reference Asahina, Sekine and Horinouchi31).
The remaining eight indicators values were statistically different from benchmark, as you can see below. The percentage of diagnostic bronchoscopic procedures was 60.7 percent (observed) versus 80–85 percent (benchmark) (see also Table 4) (Reference Slade, Rahman and Stanton32–Reference Alberts34). No extemporary cytology was performed during bronchoscopy, dramatically far below the 100 percent benchmark value (Reference Alberts34). Three indicators focusing on the surgical area were also misaligned: N2 positron emission tomography (PET) surgical candidates undergoing mediastinoscopy (0 percent observed versus 100 percent benchmark value), median time from diagnosis to surgery (58.5 observed days, range = 9–191, versus 21 days expected), and median time from PET to surgery (53.5 observed days, range = 10–171 versus 14 days benchmark value) (Reference Ouwens, Hermens and Termeer12;17;Reference Rivera and Mehta19).
Table 4. Distribution of the Diagnostic Bronchoscopies
Only 42.8 percent of the patients received concomitant chemo-radiotherapy for stage III disease compared with the expected 80 percent value of benchmark (p = 0.007) and only 2 percent of the patients in this series were enrolled in clinical trials (benchmark 5–10 percent) (Reference O'Rourke35–Reference Probstfield and Frye39). Finally, in stage IV disease patients, the length of the interval between the date of the diagnosis and the first day of chemotherapy administration was statistically longer than the benchmark's value (34 days observed versus 21 days expected; p < 0.0001) (Reference Ouwens, Hermens and Termeer12;Reference Rivera and Mehta19).
DISCUSSION
The quality of care is high on the political agenda in most Western countries. However, little is known about how quality-related information translates into actual changes in the clinical care and outcomes (Reference Neuss, Malin and Chan40;Reference Numan, Klomp and Li41).
In this scenario, the management of especially complex patients amplifies the risk of suboptimal care. To improve the quality of care, different activities have to be assessed in a reliable way so to become common practice. The efficacy of ICP strategy in lung cancer patients has been tested and validated both in single institutions and in multicenter groups (Reference Hermens, Ouwens and Vonk-Okhuijsen9;Reference von Plessen C10;Reference Ouwens, Hermens and Termeer12–Reference Fasola, Rizzato and Merlo16).
Moreover, treating patients according to evidence-based guidelines has been shown to be a cost-effective strategy for delivering care to NSCLC patients (Reference Neubauer, Hoverman and Kolodziej11). The present study underlines the role of ICP as an essential tool to ensure clinical governance and continuous audit of quality of care. As a matter of fact, our results highlight several areas of good alignment with international guidelines as well as some areas that show room of improvement.
Overall, the Pathology Department exhibited a strong adherence to the ideal ICP given the alignment of the selected indicators to the benchmark values. In fact, despite the well-known intra-tumor heterogeneity, a good consistency between histological diagnosis of the surgical sample and the one of the bronchoscopy biopsy remains valuable for treatment decisions (Reference Gerlinger, Rowan and Horswell42). Moreover, the reassignment of not otherwise specified to the squamous and nonsquamous histotype is key in the treatment strategy.
Both the time interval between the first visit with the pulmonologist and diagnosis and that the one between the bronchoscopy procedure and diagnosis joined well with benchmarks; indeed, for patients with advanced stage (IIIB and IV) the median time was significantly shorter than expected (Table 4). Possible explanations could be the higher easiness in performing tissue biopsies and the priority due to the advanced stage and/or clinical symptoms. Moreover, fewer diagnostic and staging procedures are often required in this setting. The same may be true for the timing between the first visit with the pulmonologist and the first oncology consultation.
The percentage of patients dead within 30 days from the last chemotherapy administration (6 percent) was within the expected values, showing an appreciable 10 percent reduction in comparison with 2008 results (Table 1). This evidence suggests that a palliative care approach was adequately considered. Related benchmarks value are somewhat variable in literature, ranging from 11 percent to 20 percent (Reference Emanuel, Young-Xu and Levinsky25;Reference Earle, Neville and Landrum43;Reference Greer, Pirl and Jackson44).
In our series, the decision to administer a third line therapy was made on a case-by-case basis after considering patient characteristics (age, Performance Status, benefit from previous treatments and high motivation) that could suggest a potential role for an additional treatment. International guidelines indeed do not recommend further treatment after progressing to third line treatment; however, literature data report around 10 percent of patients were treated beyond progression to third line and our performance is well within these limits (Reference Murillo and Koeller24;Reference Zietemann and Duell28;Reference Eccles, Geldart and Laurence29).
Despite a good performance for several indicators, in our hospital there is still large room for improvement, first of all regarding the presurgical procedures and timing for surgery. In our series successfully diagnostic bronchoscopies (a newly introduced indicator) were 20 percent less than the international benchmarks (Reference Slade, Rahman and Stanton32–Reference Alberts34). This difference might be explained both by procedural constraints, partly operator related, and by the lack, at the time of the analysis, of endo-bronchoscopy ultrasound (EBUS) which allows a better accuracy.
Despite the widespread recognition of the added value of extemporary diagnosis, no guideline clearly states this requirement. Therefore, due to a temporary resource shortage in the Pathology Unit, none of these procedures was performed.
None of the twelve patients undergoing surgery with a positive or doubtful presurgical PET for N2 disease underwent mediastinoscopy, compared with 3 percent in the 2008 analysis (Table 1). Although mediastinoscopy is recommended in these cases, the shortage of thoracic surgical sessions and the fact that just one surgeon was confident with this approach, could explain this result. As a result of these findings, some measures have been proposed and adopted to obtain an improvement in diagnosis and staging. The provision of an EBUS to the Pneumology Department, recognized by the ESMO guidelines as a reliable alternative to mediastinoscopy, improved performance in this setting (Reference Vansteenkiste and De Ruysscher D45).
Two other indicators, time between diagnosis and surgery and between the positron tomography, and surgical intervention, significantly worsened in comparison to the 2008 results and fell below benchmark values. These outcomes highlight a persistent difficulty, at the time, in the management of patient flow in the Thoracic Surgery Unit, due both to shortage of operating session and to the length of the preoperative assessment for patients undergoing surgery.
With the purpose of assessing the performance of the Radiotherapy Unit, in this study we introduced and evaluated a new indicator (the percentage of concomitant chemo-radiation in stage III pts) but it was far from the benchmark. Relying on previous literature and international guidelines a benchmark of 80 percent was used; however, a recent study reported a more realistic 40 percent value, considering a better patient selection based on clinical features (Reference Curran, Paulus and Langer46). This confirms that these tools should always be considered a work in progress, that requires a continuous updating of references.
Among the activities of the Medical Oncology Department two QoC indicators did not reach the benchmark, highlighting gaps and potential rooms for improvements. The number of patients enrolled in clinical trials, lower than benchmark value, could be due to the fact that this analysis was performed during a particular time range: indeed, some clinical trials on lung cancer had just been completed while new trials were not yet open to recruitment. On another side, the substantial increase of time interval between diagnosis and first chemotherapy administration in advanced stages, (Table 3), might be explained by the need for repeat CT scan before starting treatment, due to the aging of the previous scan.
Overall, the NSCLC ICP in this university hospital is somewhat in line with guidelines and literature-based evidence. It is remarkable to observe that, as in 2008, the most critical area was the early stage disease; this underlines the need for these ICP to be reshaped as patients with early stage disease have a better prognosis and can be radically cured.
Moreover, some previously reported unsuitableness were again confirmed, demonstrating that working with these tools does not automatically translate into improved clinical activities, as could also be derived from other experience (Reference Tanvetyanon, Lee and Fulp5;Reference Campbell, Hotchkiss, Bradshaw and Porteous6). This confirms that tools themselves are not the guarantee of change and will not achieve maximum success without an actual engagement of the hospital management and an actionable plan (Reference Tanvetyanon, Lee and Fulp5;Reference Campbell, Hotchkiss, Bradshaw and Porteous6).
There may be some barriers to change that would be interesting to investigate in the next future, including: the compliance of professionals, the difficulties in negotiating with general management, and financial or organizational constraints. To address the main shortcomings identified in the study, it is necessary to share the results with general management of the hospital and gain strong commitment by general management itself to identify specific solutions.
Following this experience, some additional proposal were shared and successfully introduced: (i) the presence of pathologists during the bronchoscopic procedures for extemporary diagnosis of the biopsy sample adequacy; (ii) a net increase in the thoracic surgery session number to reduce delays in management and treatment; (iii) a predefined sequence of cardiologic consultation during preoperative assessment of elderly people or patients with heart disease to speed up the presurgical path. The institution of ERAS would probably also help improve patient outcomes.
This study has, of course, strengths and limitations (Supplementary Table 1). One of the main strength is, in our opinion, the full hospital computerization which allowed us to overcome the commonly identified issues in retrieving required information to measure the quality indicators. An ehealth system has been proven successful in supporting integrated care in healthcare pathways and sharing information among centers (Reference Eason and Waterson47–Reference Ellis and Lokai49). A further strength can be considered the experience gained from the previous study, published in 2012.
Among the weakness should be recognized that this is a retrospective, single-center review with a limited sample size. Second, some indicators of the quality of care are lacking, such as perioperative mortality, postoperative length of stay, or the proportion of adjuvant treatment among stage II–III surgical patients, an issue well recorded in literature (Reference Campbell, Braspenning, Hutchinson and Marshall8). Lastly, this analysis does not include patient's perceived quality of care and prehospital evaluation, which, however, has been included in the following study.
While the data compared two different time periods, we did not think there were any other changes that occurred to affect the outcome of the data. Also the 5-year survivals were not recorded and we have no data on whether the delays led to upstaging or worse survival rates.
CONCLUSIONS
Our study demonstrated that the adoption of ICP methodology in a teaching hospital is feasible and enables the assessment of the quality level in healthcare delivery to a specific patient group/population. Relying on the knowledge of this work, the same research team has designed a new prospective study, the third in this series, which has just stopped recruiting cases. Finally, our experience has been shared with the general management and executive staff of the hospital, with the aim of re-engineering the delivery of care focusing on processes, people, and technology.
SUPPLEMENTARY MATERIAL
Supplementary Table 1: https://doi.org/10.1017/S026646231700441X
Supplementary Figure 1: https://doi.org/10.1017/S026646231700441X
CONFLICTS OF INTEREST
The authors have nothing to disclose.
