In this study, all groups of patients significantly improved in recovery during the first two weeks following the day surgical procedure. However, the orthopaedic patients did not recover to the same extent as the general and gynaecological patients did. An additional finding was that recovery on postoperative day 7 was associated with patients’ HRQoL 30 days following the day surgical procedure. Also, orthopaedic patients had lower HRQoL already before surgery and did not reach the same level 30 days postoperatively as the general and gynaecological patients did.
The finding that orthopaedic patients had the lowest preoperative health condition, represented by the S-PSR baseline score, can mirror that several orthopaedic patients have preoperative discomforts like pain and impaired mobility [19, 31], which in turn may affect the recovery time course. For instance, arthroscopic patients showed a slower recovery process when such preoperative discomforts were considered . Also, a patient’s preoperative expectations on recovery may matter . Experienced pain and ambulation difficulties before surgery can figure into positive or negative postoperative expectations. Relief from longstanding pain through surgery carries positive expectations. On the other hand there can be worries of a cumbersome rehabilitation process. In the present study, the orthopaedic patients reported more pain and reduced mobility on day 7 compared with the other groups of patients. Postoperative symptoms and discomforts are found to have an impact on recovery and activities in daily living . Orthopaedic surgery is painful  and frequently has an effect on mobility , which often results in a protracted recovery period. Our focus was on recovery from a general perspective; thus only some symptoms were analysed on day 7. However, commonly occurring symptoms and discomforts such as pain, mobility, tiredness, drowsiness, sleep and nausea are included in the S-PSR and QoR-23 scales and are thereby regarded in the overall picture.
All groups of patients improved in recovery from the first postoperative day to day 14. However, the orthopaedic patients had lower recovery compared with the others. The orthopaedic patients had a lower starting position, and since no difference in mean changes in scores existed, this group of patients needed more time to reach the same level of recovery as that of the general and gynaecological patients. Many factors influence recovery following day surgery. Patient characteristics (e.g. age, body mass index, comorbidity), type of surgery, type of anaesthesia and social circumstances are such factors . In this study, no difference in age existed. Comorbidity may be regarded in the light of the ASA classification in our sample. Patients in the general surgery group represented a larger proportion of ASA 2 classified patients. Despite this, the orthopaedic patients experienced lower recovery. Regarding type of surgery, arthroscopic (knee and shoulder) patients have previously been reported to experience a protracted recovery period [8, 33]. General anaesthesia is a postoperative risk following day surgery, but its importance as a risk factor ought to be interpreted with caution . In the present sample, the use of general anaesthesia was proportionally low among the orthopaedic patients and thus does not appear to have influenced the lower level of recovery in this group of patients; neither do social factors appear to have done. Thus, identifying patients at postoperative risk in advance seems beneficial. The identification work can be done by nurses through preoperative screening or a pre-admission appointment, and the support may consist of preoperative education  or a close postoperative follow-up at which the management of pain or other clinical management and self-care  are advised.
The orthopaedic patients had lower HRQoL than the general and the gynaecological patients both before and one month following surgery, which is in concordance with Brattwall et al. . In contrast, Suhonen et al.  found that patients perceived HRQoL as high before as well as after the surgical procedure. These divergent results might be due to slightly different samples, even though the majority of the patients had undergone orthopaedic surgery. To increase the knowledge about HRQoL following orthopaedic day surgery, further research is needed.
The results in this study indicate that orthopaedic patients are a vulnerable group in day surgery, who may benefit from a closer contact with the health care following discharge. Increased support may facilitate overall recovery, which in turn may have a positive effect on the patients’ HRQoL. To better prepare patients for their recovery preoperative information, screening of patients and information at discharge are suggested to be priority interventions for day surgery nurses . Recovery differs within the group  which ought to be considered when post-discharge care is planned. Validated and user friendly questionnaires, used in a telephone follow-up, could be suitable to identify patients in need of more attention.
Postoperative recovery was significantly associated with HRQoL, and to our knowledge no previous study has examined this association. When further adjusted for known covariates, ASA class 2, female gender and general surgery were associated with HRQoL. However, even if these covariates were significantly associated with HRQoL 30 days postoperatively, its contribution to the model explanation was minor and did not have any appreciable impact on HRQoL. It seems that the patients’ recovery was still the most important for perceived HRQoL following day surgery. These results indicate that the assessment of recovery one week postoperatively can be used to identify patients with risk of impaired HRQoL one month following their day surgical procedure. An additional advantage is that HRQoL has long been regarded a major predicting factor of patient satisfaction with outcomes of medical services ; in the day surgery context HRQoL might be considered from this perspective as well.
This study contains a number of methodological weaknesses. The sample size was large and consecutively collected; nevertheless, the gynaecological patients were few in number. No sample size calculation was performed before the study. Instead, the power of the regression models was investigated afterwards using the software G*Power 3 . The power of the test (1-β) for the regression models was >0.90 based on a medium effect size (f2=0.15), a significance level (α) of 0.05, 11 predictors and a sample size of 381 patients. Patients scheduled for general surgery declined participation or were missed being asked more often than those who participated in the study. This could possibly have had an effect on the results, especially since the general surgery patients were proportionally more classified in ASA class 2. A problem with prospective studies is that a number of patients drop out during the data collection. In this study, this might have been due to an age-related factor; it is possible that the younger patients experienced a more rapid recovery process and were back in their ordinary life, and therefore did not deem it important to complete the study assignment. The International Association of Ambulatory Surgery recommends follow-ups to be conducted up to about one month after the surgery . Many patients have recovered by that time. More knowledge on recovery and HRQoL after day surgery is needed to estimate if this recommendation will be regarded as an optimal follow-up period. The proportion of men was larger among the non-responders. Possibly, some of them had agreed to participate before reflecting sufficiently on the commitment involved, and therefore decided to withdraw. Some patients did not fill out the questionnaires on all occasions, which resulted in their exclusion from the paired analysis; the varying sample sizes in the different analyses are due to this. However, on account of the large sample size no imputation of missing data was performed.