Assessment of the feasibility of a comprehensive patient reported outcome set for postpartum women: the maternal patient-reported outcome set (MOMs Study)

Recruitment

Recruitment for this feasibility study was undertaken in a single centre tertiary university maternity hospital with approximately ~8000 births per annum. The aim was to obtain a sample size of 50 participants who responded to the survey at timepoint 1 which has been shown to be adequate for a feasibility study of PROM⁸. The eligibility criteria for recruitment were inpatients on the postnatal wards after delivery of >500g baby, ≥18 years of age, English speaking and signed informed consent. Exclusion criteria were those <18 years old, no adequate understanding of spoken English and mothers who have experienced a perinatal death. All women who were eligible for recruitment were approached by a clinician recruiter. This was a convenience sample recruited during 8am–4pm on the postnatal wards of Cork University Maternity Hospital (CUMH) between 28th November 2022 and 9th January 2023. Women were given Patient Information Leaflets (PIL) with a QR code link to the first survey. They consented to demographic details and details regarding delivery being extracted from the electronic health record (EHR) at CUMH, and to submitting their survey responses using the QualtricsXM™ platform. The EHR stores data from all pregnancies in maternity services in Ireland, allowing all maternal and newborn information to be stored on one record. The data from the survey and the EHR were linked using a unique patient number.

Statistical analysis

Results were reported descriptively using IBM SPSS and Microsoft Excel. Chi square analysis was used to determine difference between participants’ age, parity, and delivery type across the three timepoints. Results of the mean PQoL scores across the three postpartum timepoints were reported as a response rate, mean score and standard deviation.

PROM

The systematic review our team published in 2022⁴ highlighted the top scoring PROM which were chosen as per the COSMIN guidelines⁹. The PQoL⁶ scored highly in structural validity, internal consistency, reliability and hypothesis testing but did not assess cross cultural validity or responsiveness. The PQoL is a 40-item tool using a 5-point Likert scale, it is freely available and easy to administer with a completion time of 5–10 minutes for patients to score. It covers the domains of HRQoL, mental health, role transition and breastfeeding. The PQoL has been reported to have the ability to discriminate between delivery type, postpartum depression and socioeconomic status¹⁰. The addition of the urinary incontinence tool, the ICIQ-UI SF, improved its domain coverage. The ICIQ-UI SF was chosen as it was the one psychometrically evaluated incontinence tool for use in postpartum women⁷.

The last remaining domain of pain during sexual intercourse as recommended by the ICHOM had no psychometrically assessed measure. Therefore, we included two additional questions on pelvic pain with sexual intercourse as follows: ‘If you are currently sexually active, do you now experience pain with intercourse? If so, was this something that you had experienced before? The full questionnaire can be found in the appendix.

Validity

Content validity as defined by COSMIN as ‘the degree to which the content of a measurement instrument is an adequate reflection of the construct to be measured’¹¹. To assess content validity from a patients perspective we asked respondents if the MOM’s questionnaire ‘captures the most important outcomes that matter or have mattered to you?’. A free text box was also included where they were invited to suggest items to add or to be removed from the total PROM. [Appendix].

Participant’s demographics

Women were asked demographic questions including age, education, parity, type of delivery and delivery complications for mother and baby. These were compared to their demographics recorded on the EHR. Type of delivery among study participants was also compared to total hospital figures to assess selection bias. [Appendix].

Recruitment and demographics

Recruitment was performed by one clinician recruiter between 28th November 2022 and 9th January 2023. We approached 89 postpartum women, of which 86 women consented to enrolment in the MOMs study. One woman subsequently withdrew due to time pressure. Therefore, there was a total of 85 participating in the MOMs study.

The demographics of the women recruited and those who responded at time point 1 (T1), 2 (T2) and 3 (T3) are represented in Table 1. There was no statistically significant difference in demographics across groups using chi squared analysis. Data from Ireland’s Health Service Executive (HSE) Maternity Safety Statement for November 2022 suggested that there were 561 mothers who delivered >500g babies in the hospital that month, with 11.8 % having assisted instrumental deliveries and 38.3% having caesarean sections¹². December 2022 had 514 mothers who delivered >500g babies in CUMH with 14% instrumental deliveries and a rate of 42% for Caesarean section¹³. (eTable 1) When comparing the women recruited for the MOMs study to the delivery type of all the deliveries in the hospital over that time, recruitment captured an appropriate proportion of the differing modes of delivery over this time(eTable1).

Validity

70.2% (n=33) of T1 respondents felt that the MOMs questionnaire captured the outcomes that mattered most to them. This rose to 73.3% (n=33) of the respondents at the second time point. At T3 n=25 (69%) agreed. From the free text section, themes regarding question structure and language emerged with regards to the PQoL’s use of the word “husband”. Some comments arising from this included, “I’m a single mother by choice.” At the 6-week time point themes around activity and feeding emerged: patients asked for the inclusion of ‘question regarding how much exercise the individual is participating in, including pelvic floor’ and commented on the PQoL’s narrow questions on infant feeding. ‘The feeding questions only allow for breast feeding not formula feeding’.

Response rate

Of the 85 women recruited for the study, the response rate at T1 was 59 (69%). Of these, 47 (55%) were complete responses. The response rate at T2 was 57 (67%), with a completed response rate of 45 (53%). At T3, the response rate was 41 (48%), with a complete response rate of 37 (44%). The responses that were not complete were either re submissions from the same user number or those respondents that logged out of the questionnaire at the requests for demographic details before the PROM. No respondent commenced the QoL questionnaires and did not complete the survey.

Survey access and length

Of the first timepoint responses 51/59 women (86.4%) accessed the survey via anonymous QualtricsXM™ link sent to them via email and 8 (13.6%) used the QR code on the patient information leaflet (Table 2). All subsequent responses were accessed through emailed survey link. The median time taken to complete the first survey was 417 seconds (IQR 918–297) (7mins), to complete the 6-week survey was 387 seconds (IQR 581–268) (6.5 mins) and to complete the 12-week survey was 384 seconds (IQR 580–297)(6.4 mins). These completion times included demographics and delivery details as well as the validity questions.

Morbidity reporting

The frequencies of patients’ self-reports of morbidity at the time of delivery, including postpartum haemorrhage (PPH), obstetric anal sphincter injury (OASI), high dependency unit admission (HDU), neonatal unit admission (NNU) is shown in Table 3.

There was good concordance between the clinician electronic health record (EHR) and the patients’ self-reports across HDU and NNU admission. OASI morbidity was reported by one extra patient in time point 2 but otherwise had good concordance throughout.

For PPH (defined as all blood loss over 500ml), there were more EHR records of PPH than women self-reported. Timepoint 1 had 14 PPH recorded in the EHR with only 4 self-reporting the same. Similarly, at timepoint 2, 4 women self-reported PPH, while 12 PPH were recorded by clinicians in the EHR. At timepoint 3, 2 women self-reported PPH, with 9 PPH recorded by clinicians in the EHR. When examining PPH by quantity of blood loss, those women who experienced a massive obstetric haemorrhage (MOH) (defined as blood loss >1,000ml) were more likely to correctly self-report PPH as a morbidity. There were 4 MOH at timepoint 1 with 3 women reporting the same, and this was concordant at timepoint 2 and timepoint 3 with all women having an MOH self-reporting the same.

Mean PQoL scores

The mean (±SD) PQoL scores across the three timepoints are represented in Table 4. T1 126 (±8.3) (n=49), T2 123.4 (±7.8) (n=47), T3 124.1 (±7.5) (n=39). The lowest PQoL score recorded at T1 by delivery type was the OVD ventouse group 122.0 (±10.5; n=3) which subsequently at T3 remained the lowest at n=3 118.3 (±6.4). The lowest scores broken down by morbidity were found in the HDU admission n=1 120 and OASI group n=1 121 at T3.

Principal findings

This study examined the feasibility of recruiting inpatient postpartum women to evaluate the use of an online combination PROM covering postpartum quality of life over three time points up to 12 weeks postpartum. We also assessed validity and completeness of the novel outcome set within the first week postpartum (T1) and at 6 weeks after delivery (T2) and (T3) at 12 weeks post-partum. The secondary objectives were to examine the concordance between patient and clinician reports of delivery complications and to assess response rates for an online data collection system. This study represents an important step in a stepped methodological approach to the development of a comprehensive PROM for postpartum care. We have yielded four principal findings;

Firstly, recruitment on the postnatal wards by a single clinician researcher was very successful with 86 of 89 women approached recruited to the study. The demographics of those recruited represented the demographics of deliveries in the hospital over the same time period, with similar numbers of vaginal deliveries and caesarean sections being represented. Despite the small sample size required for a feasibility study, the response rate was good in comparison to other feasibility PROM studies in the literature with a completed response rate of 55% (n=47) at the first time point, 53% (n=45) at the 6-week time point and 44% (n=37) at the third time point. Our feasibility study achieved a favourable response rate, as comparable feasibility studies have reported a response rate of approximately 30%¹⁴.

Secondly, the initial stage in evaluating validity¹⁵, the preliminary face validity was favourable, as 70% (T1), 73% (T2), and 69% (T3) of respondents expressed agreement that the patient-reported outcome measure (PROM) encompassed all the relevant domains that were personally significant to them.

Third, the concordance between delivery complications self-reported by women and by the healthcare professionals at the time of delivery was high. Women who had required HDU and NNU admission for their child were accurately self-reported. Concordance between the EHR and the patients’ self-reports of OASI was mostly good. However, one episiotomy was incorrectly self-reported as an obstetric anal sphincter injury at time point 2. Interestingly, in relation to postpartum haemorrhage (PPH), our findings revealed that women who experienced PPH during delivery were less inclined to self-report it as a complication. Conversely, those who had a more substantial blood loss exceeding 1000ml (termed major obstetric haemorrhage or MOH) were more likely to acknowledge it as a PPH complication. For instance, at timepoint 1, among the four women identified with an MOH based on their electronic records, three self-reported the same. Similarly, at timepoints 2 and 3, all women who had an MOH self-reported it. Consequently, the underreporting of PPH was more prevalent among participants who encountered a blood loss of less than 1000ml. In relations to the mean PQoL score, this study is not powered to detect changes between group. There is a trend toward a lower PQoL score in the T1 emergency CS and operative delivery ventouse group at T1 future work will evaluate the responsiveness of this tool in a larger sample size.

Finally, the use of QualtricsXM™ software as an interface for women has been found to be appropriate in our study. We observed that the process of emailing survey links yielded better response rates compared to the alternative method of utilizing QR codes on patient information leaflet packs distributed to women. Additionally, the email link served as an efficient means of conducting follow-up assessments during subsequent time points. This approach demonstrates potential for automation if integrated into an electronic healthcare record system, allowing for seamless implementation and data management.

Comparison with existing literature

Previous systematic reviews have failed to identify any comprehensive multi domain tools for a PROM assessing maternity care^1,4. Evaluation of individual domain specific PROMs has been previously published, for example with regards to mental health, postpartum recovery, and incontinence, but no work exists examining a multi domain PROM or a combination of PROMs suitable to universally examine postpartum health and wellbeing^7,16,17. This feasibility study examines the use of a combination postnatal PROM across time periods to assess all domains of postpartum health related quality of life as described by ICHOM work.

Strengths and limitations

This study possesses multiple strengths, the foremost being the systematic and methodical approach employed by using the top recommended PROM for postnatal populations which were chosen based on their psychometric properties and domain coverage as guided by the COSMIN guidelines and ICHOM work on domains of importance to postnatal women^4,5,9.

Second, exploring the validity of this approach from the patient's viewpoint contributes to the existing body of work published by ICHOM. Across the three time points, a significant proportion of women (69–73%) expressed that the combination of tools employed encompassed all the domains that were personally important to them. Moving forward, we plan to expand our research by investigating the emerging themes from the free-text comments on this approach, using a larger sample size.

Third, a notable strength of this study is that the recruited population closely mirrored the general postnatal patients at a tertiary university maternity hospital in terms of delivery type. This representative sampling enhances the extension of our findings to the broader postnatal population in similar healthcare settings.

Fourth, our feasibility study achieved a more favourable response rate, as compared to similar PROM feasibility studies which have reported a response rates of approximately ~30%¹⁴. The response rate in our study was 59 (69%) at T1, 57 (67%) at T2 , and 41 (48%) at T3.

Finally, assessing both the patient self-reported complications and the delivery complication recorded in their EHR extends to a patient’s understanding of what healthcare interventions have been undertaken and the complications they experienced. Future work examining postnatal quality of life after a PPH morbidity should investigate this further.

This study also had limitations that should be noted. First, exclusion criteria included non-English speaking and reading patients, which may have introduced potential language barriers in data collection and interpretation. However, no patient was excluded from the current study based on the inability to read or speak English. Secondly, despite the sample size being adequate for a feasibility study this limited sample may affect the generalisation of the findings. The study was conducted at a tertiary university maternity hospital, which may not represent the broader population of postnatal women. Finally, this study focused on the assessment of outcomes within the first 12 weeks postpartum. Longer-term effects and changes in health-related quality of life beyond this timeframe were not examined, limiting the comprehensive understanding of postpartum experiences to 3 months.