Need to measure outcome after discharge in surgical audit

Surgical outcomes research: a progression from performance audits, to assessment of administrative databases, to prospective risk-adjusted analysis - how far have we come?

PURPOSE OF REVIEW

This review focuses on how the assessment of surgical quality and safety has evolved from individual performance audits and morbidity and mortality reviews, to assessment of large administrative databases, to the current practice of prospective risk-adjusted analysis by a National Surgical Quality Improvement Program for children's surgical care. This evolution follows the natural availability of surgical outcome data and a national call for improved hospital care safety and quality.

RECENT FINDINGS

Two new advances in children's surgical care include the comparative use of national health record data compiled in administrative datasets and the use of a risk-adjusted assessment of children's surgical morbidity and mortality as assessed by a newly developed National Surgical Quality Improvement Program for children's operative care. The value and application of these two datasets are presented.

SUMMARY

The evolution of the assessment of surgical quality and safety will equip the surgeon with an optimal array of outcome assessment tools to assure the best in surgical quality and safety for the pediatric patient.

Outcomes and prediction of hospital readmission after intestinal surgery1 1No competing interests declared

BACKGROUND

Clinical care pathways reduce postoperative stay after major bowel operations. Concerns about unexpected early readmissions and delays in diagnosis of complications remain unanswered. The objectives of this study were determination of readmission rate and outcomes for patients undergoing intestinal operations.

STUDY DESIGN

Patients readmitted (PR) within 30 days of discharge after intestinal operations were compared with patients who were not readmitted (NR). Variables that might predict readmission were evaluated.

RESULTS

Of 553 patients, 56 (10.1%) were readmitted after 10 days (interquartile range [IQR] 4.5 to 15.5 days). PR and NR groups had similar age, gender, diagnosis, preoperative comorbidities, and index operations. Discharge hemoglobin level, white cell count, antibiotic use, or presence of stoma did not affect readmission. PR had a greater frequency of steroid use (p = 0.03) during index admission. Median length of stay for the index hospitalization was 5 days (IQR 4 to 8 days) for the NR and 6 days (IQR 4.8 to 9 days) for the PR group (p = 0.049). Duration of readmission was 4 days (IQR 2 to 9 days) in the PR group, with equal total median length of stay identical for PR and NR patients with complications (median 12 days). Clinical outcomes for PR patients and NR patients with complications were similar.

CONCLUSIONS

Early readmission is an unpredictable sequel of major bowel operations; it does not correlate with shorter hospital stay. Identification of unpredictable complications after discharge that require later invasive intervention does not adversely affect clinical outcomes. Readmission within 30 days of a patient who has attained standardized discharge criteria may not be a valid indicator of poor quality of care.

Readmissions after colorectal surgery cannot be predicted

INTRODUCTION

Readmission after discharge from the hospital is an undesirable outcome. In an attempt to prevent unplanned readmissions after abdominal or perineal colon resection, we proposed to identify risk factors associated with return to the hospital.

METHODS

Study participants consisted of 249 patients who were operated on from July 1, 1996, to March 30, 1998. All patients who were readmitted within 90 days of discharge from the hospital after surgery were evaluated for the study. A retrospective review of charts was performed to assess whether readmission within 90 days was a direct consequence of the recent operation (unplanned related readmission). These patients were compared with a control group consisting of patients who were never readmitted or who were readmitted with an unrelated problem.

RESULTS

Of the 249 patients, 59 (24 percent) were readmitted within 90 days of discharge from the hospital. Twenty-two (9 percent) were unplanned related readmissions. Ten patients were readmitted with unrelated emergencies, and 27 patients were readmitted electively. In the unplanned related group, there was no correlation between age, gender, admission diagnosis, activity status, or postoperative length of stay and the likelihood of readmission. Patients with multiple chronic medical problems or those who developed postoperative complications did not have a higher readmission rate. Patients with ulcerative colitis or those who underwent abdominoperineal resection or total/subtotal colectomy had a higher incidence of readmissions, although the difference was not significant. The mean interval between discharge from the hospital and readmission with a related complication was 19 days. Small-bowel obstruction was the most common reason for readmission, and all cases resolved with conservative management. Mean length of stay during all readmissions was 8 days.

CONCLUSION

The incidence of unplanned related readmissions 90 days after abdominal or perineal colon resection is 9 percent, and these readmissions could not be predicted from the postoperative course. Because 82 percent of unplanned readmissions occurred within 30 days, this time frame is suitable for computerized comparative analysis.

The ethics of attribution: the case of health care outcome indicators

The ethical basis of clinical outcomes measurement is a desire to improve care in a way which will increase both clinical effectiveness and value for money-beneficence as well as competence. To date in the U.K., any debate about producing comparative indicators of clinical outcomes has been concerned mainly with the unfairness to individual doctors or clinical teams of judging their performance on this basis. There has been less interest in the prime purpose of such production, which is to increase the accountability and effectiveness of the NHS as a publicly funded service. Rather than working to improve clinical effectiveness and outcomes within clinical services, health authorities which wish to improve outcomes for their populations have been encouraged simply to shift the contract to another provider of care. The key issue on which the ethics of either action rests is the extent to which the attribution of outcome to intervention is valid and reliable and, therefore, that judgements about performance are just and thus ethical. The consequence of unjust judgements may be to increase the inequalities that medical care resource allocation should attempt to reduce.

Hospital readmission after trauma: an analysis of outpatient complications

BACKGROUND

Outpatient complications leading to hospital readmission after hospitalization for trauma have not been examined.

METHODS

A retrospective chart review of all trauma victims admitted to a Level 1 trauma center from January of 1990 to January of 1995 was performed to characterize patients who required readmission after hospitalization for trauma. Risk factors for readmission were determined by stepwise regression analysis.

RESULTS

Of 15,463 trauma admissions, 209 patients (1.4%) required readmission, 84% within 30 days, 71% within 14 days. Reasons for readmission included wound (29%), abdominal (29%), pulmonary (18%), and thromboembolic (19%) complications. Fifty of the patients (24%) readmitted with a complication required an operation. Risk factors for readmission included: operation during first hospitalization (p < 0.0001), penetrating injury (p = 0.0001), and advanced age (p = 0.0001). Injury Severity Score, length of hospitalization, and gender were not independent predictors of readmission.

CONCLUSIONS

Outpatient complications leading to readmission after hospitalization for trauma are not common; however, many are serious and require operative intervention. Because most complications were identified by the second week after discharge, outpatient follow-up visits should be scheduled within 7 to 14 days. Based on our findings, we recommend protocols be established to ensure follow-up for trauma patients, especially those who have had an operation, were victims of penetrating injury, or those > 65 years of age.

Infection registration underestimates the risk of surgical wound infections

Two consecutive bedside prevalence studies of 455 surgical patients were made by the same infection control nurse in 15 surgical and gynaecological departments in eight Danish hospitals. There was a high degree of diagnostic agreement between the prevalence survey and the clinical data. Four point six percent had a deep, and another 4.6%, a superficial surgical wound infection (SWI). Two months after the second survey only one third of these patients had their infection correctly recorded by the routine hospital surveillance of SWI. Better routines need to be developed to secure a valid, reliable and simple registration of relevant infectious wound complications. A follow-up was carried out with self-administered questionnaires in 2976 patients, of whom 1447 (48.6%) responded. This post-discharge survey showed that 15.7% had been treated with antibiotics, because of pus in the wound, 12.4% had to have the wound reopened and 6.6% experienced both these treatments.

Recollected versus contemporary patient reports of pre-operative symptoms in men undergoing transurethral prostatic resection for benign disease

The objective of the study was to investigate the reliability of patients' recollected pre-intervention symptom status and the impact of those symptoms compared with contemporary pre-operative reports, and to test the stability of recollected views. In design (A) a self-completed symptom questionnaire was administered before (contemporary) and 3 months after (recollected) surgery. In design (B) a self-completed symptom questionnaire on recollected pre-operative symptoms was administered 12 and 14 weeks after surgery. Setting (A) comprised the twin consultant urological unit in the Chesterfield and North Derbyshire Royal Hospital NHS Trust, and setting (B) a sample from the National Prostatectomy Audit of 5281 patients. The subjects were 77 consecutive patients scheduled for transurethral resection of the prostate (TURP), and 170 consecutive respondents undergoing TURP. The main outcome measures were the difference in group mean scores for The American Urological Association (AUA) Symptom Index, Impact Index (a score of symptom impact), and 14 constituent questions; association assessed using Pearson's correlation coefficient; agreement assessed using weighted Kappa statistics. Complete paired data sets were available for 58 (75%) men for the Symptom Index, and for 61 (79%) men for the Impact Index. Pre-operative mean Symptom Index scores for contemporary and recollected were similar, as were mean scores for the Impact Index. However, only poor to fair levels of association and agreement were obtained for the Symptom Index (r = 0.6, kappa (w) = 0.3) and Impact Index (r = 0.6, kappa (w) = 0.3). Results for the constituent questions were similar.(ABSTRACT TRUNCATED AT 250 WORDS)