Pictorial methods to assess heavy menstrual bleeding in research and clinical practice: a systematic literature review

The first PBAC was introduced in 1990 by Higham et al. and depicted three images (icons) that represented specific brands of feminine items soiled with increasing amounts of blood [29]. The icon scores were 1, 5 and 20 points for towels and 1, 5 and 10 points for tampons. A range of blood volumes produced visually similar stain sizes. Consequently, the score allocated to each icon was related, but not necessarily equal, to the applied blood volume in milliliters. The size of blood clots was compared with the diameter of UK coinage, and the number of flooding events was recorded, although there were no associated scores. Twenty-eight women of undisclosed menstrual status were monitored between one and 3 cycles each, giving 55 patient-assessed cycles in total. The PBAC scores ranged from 5 to 456 (median 121), whereas MBL measured by the AH technique ranged from 2 to 366 (median 74), giving a correlation coefficient (r) of 0.847. When used as a diagnostic test for HMB, a score of ≥100 gave a reported sensitivity and specificity of 86 and 89%, respectively. However, the PBAC progressively underestimated blood volume as menstrual loss increased, which demonstrated an important limitation of the method for women with HMB. The Higham PBAC is shown in Fig. 2a.

This PBAC was reassessed in 1994, although women were allowed to use their usual sanitary materials rather than those initially validated. With a specificity of 52%, it detected more false positives than the Higham study [30]. A year later, a modified pictorial chart was certified for use with specific feminine products. The scoring system was identical to the Higham PBAC, but the icons looked different (Fig. 2b). Based on receiver operating characteristic analysis, a cut-off point of 185 showed the greatest positive and negative predictive values for HMB, and it was concluded that a score of > 185, rather than > 100, should be used as a diagnostic threshold [31]. In 1999, adolescent Nigerian girls tested a novel pictorial chart with four towel icons (Fig. 2c). The sanitary products were not identified. Maximum sensitivity and specificity were achieved at a cut-off score of only 50, which was attributed to the fact that all the study population had normal MBL [32]. In 2000, the validity of the PBAC was challenged when Reid et al. tried to reproduce findings from the Higham pictorial chart on a population of women with self-reported HMB. Based on a specificity of 7.5% and a low correlation (r = 0.47) between PBAC score and AH, it was judged to be an inappropriate method to measure MBL [33].

In 2001, a new version of the PBAC was introduced by Wyatt et al. The menstrual pictogram depicted five icons representing blood loss on towels and four icons for tampons. The method was validated using simulated menstrual fluid (blood/saline, 1:1 ratio) to represent the physiological setting, in which the visible stain typically comprises about 50% blood (J.L. Magnay and K.M. Wyatt, personal communication). Scores for blood clots and extraneous loss were included. Crucially, all scores were quoted as blood volume with a cut-off limit of 80 mL, which made it directly comparable to the AH technique. The pictogram was validated using specified towel and tampon brands with various absorbency ratings (Fig. 2d) [34].

Since these early validation studies, towel and tampon manufacturers have responded to market demand for more comfortable, discrete feminine products with enhanced fluid absorbency. Sanitary products used in the early PBACs became obsolete, and ultra-slim towels containing superabsorbent polymers rapidly gained popularity. In 2011, an updated version of the Higham chart was validated for Always Ultra towels (Proctor & Gamble, Cincinnati, OH, USA), the most popular brand of ultra-slim towel in the USA and UK [38, 39]. Tampons, clots and episodes of flooding were not evaluated so, to minimize extraneous loss, women were asked to use double towels during episodes of HMB. Maximum sensitivity and specificity were achieved at a cut-off score of 150 [35]. Two years later, Larsen et al. revalidated the menstrual pictogram after obtaining identical sanitary wear to that used in the original chart by Wyatt et al. (J.L. Magnay and L. Larsen, personal communication). At a threshold of 80 mL, the method gave a sensitivity of 96% and a specificity of 92% in dichotomizing response to medical treatment in terms of a ≥ 50% or < 50% decrease in MBL. This was the first time a pictorial chart had been endorsed for other than diagnostic purposes; namely, the MBL assessment of women with uterine fibroids treated with an investigational drug [36]. However, the validation was done with an obsolete version of feminine items and cannot be applied to other studies. In 2014, a new menstrual pictogram (SAP-c version) was validated with simulated menstrual fluid for use with Always Ultra towels and a choice of three absorbencies (Fig. 2e). Correction factors were applied to the provisional icon scores to account for the progressive increase in blood fraction with menstrual volume, as demonstrated in Fig. 3 [5, 26, 40]. The cut-off value for HMB was 80 mL.

Several derivative PBACs were detected in our database search but, with the exception of those described above, objective confirmation of their validity was not reported. Most charts referenced the scoring system of Higham et al. and, to a lesser extent, Janssen et al., although several modifications were made. The brand(s) of sanitary products were stated for some PBACs [15, 41‐45] but not for others [46‐65]. The icon illustrations were often different and/or the individual scores were altered, whether by error or by design [41, 43, 46‐49, 51, 52, 55‐57, 60‐63, 66, 67].

In numerous cases, a validated PBAC was referenced, but the cut-off threshold was adjusted for no obvious reason [46, 55‐57, 59, 64, 65, 68‐72]. Two reports directly compared the Higham and Janssen scoring systems (HMB > 100 and > 185, respectively) in the same study [73, 74]. In some instances, contemporary towels and tampons were deliberately substituted for their validated obsolete counterparts [16, 42, 75], but the cut-off score was not re-established. For a given volume, the visible stained area on modern sanitary materials is less than on older products [75] and can vary greatly between different brand formulations and absorbency ratings [41, 75‐78]. Four studies directly converted PBAC score to MBL volume [17, 72, 79, 80]. However, this relationship is not linear because a range of blood volumes may be assigned to the same icon category. Three studies used a PBAC to simply assess the pattern of menstruation before and after treatment: the first to determine the length of menstrual period [81], the second to additionally record the number of towels used [82], and the third to measure the decrease in number of towels used at the time of heaviest menstrual loss [83]. None reported measured MBL. Table 2 lists some examples of modified, nonvalidated PBACs.

All included components of a PBAC should be assessed. The value assigned to individual towel and tampon icons should be verified against AH for every selected absorbency category. Allocating blood volumes to each icon (as opposed to an empirical value) allows direct assessment of MBL, as in the menstrual pictogram. Whichever scoring system is used, it must be sufficiently sensitive to detect what are considered to be clinically relevant differences in MBL at the appropriate time point following treatment for HMB. An important issue to consider is the increase in stain area when a feminine towel or tampon is subjected to mechanical load [75]. Compression forces exerted by the wearer may ultimately affect icon selection. Choosing products that show minimal stain spread under pressure would limit this confounding element [5, 86]. Other factors that can affect visible stain area are the rate and composition of menstrual flow, individual anatomy, physical activity and posture. It is difficult to control for these variables.

The PBAC must be tested by a spectrum of women with light, normal and heavy MBL to assess not only its validity as a diagnostic tool for HMB but also its ability to estimate normal and low menstrual volumes. Many women who complain of HMB actually have normal periods, and those with excessive blood loss often experience dramatic reductions in MBL after treatment. The scores should be reproducible and yield the same results in repeated applications [86].

The format of a validated PBAC should not vary during use because it may change a patient’s perception of her MBL and, thus, her score. Because towels and tampons often have distinctive shapes and sizes, icon images need to represent the feminine items used [5, 86]. Successive icons must display clear differences in appearance to avoid indecision or mis-scoring. If a paper-based PBAC is chosen, the chart and menstrual diary should easily fit into a typical daytime-sized handbag and have a neutral, discrete cover. It must have a robust construction to prevent disintegration and allow sufficient room for the number of soiled feminine items to be recorded each day [86]. Prospective daily recording is critical for data collection involving a series of recurrent similar events such as MBL, in which the accuracy of retrospective recall tends to be low [90].

If an electronic design is preferred (e.g. a dedicated eDiary, smartphone application or website), a paper version may still be needed in situations in which telephone or Internet reception is unavailable, the equipment has insufficient battery charge, or it has been lost or damaged. Electronic recording of MBL is not a new concept. In 2002, a menstrual symptometrics device for use by patients incorporated a menstrual pictogram programmed into an Amstrad PDA600 PenPad [91]. It was validated for MBL measurement using the paper-based chart of Wyatt et al. as the reference standard [34]. Electronic data logging is generally preferred by patients and results in fewer missing data than a paper calendar and fewer transcription errors when results are transferred to a computer database [90]. A study showed that 80% of adolescents and young women preferred to use a smartphone application version of the PBAC over a paper-based chart, although the method was limited by mobile technology issues [92]. A second study developed an electronic PBAC (ePBAC) that could be completed using the Internet [93]. Neither chart had been validated against the AH method.

To gain maximum patient compliance, use of the PBAC must be as straightforward as possible. At an early stage, patients’ views should be sought about the method and how it may be improved, and the level of compliance should be tested [86, 92]. Women should be supplied with sufficient feminine materials, based on their expected MBL, and carefully instructed on how to score their soiled items. This includes interpretation of nonstandard or fragmented staining patterns. The importance of prospective recording must be emphasized, and the contact details of a liaison nurse should be available in case of any concerns.

Extraneous MBL is common for women experiencing HMB and may be due to sudden flooding episodes, not changing sanitary products often enough or a combination of both. To limit this event, women should be encouraged to use tampons and towels concurrently, to ensure a fresh tampon is worn when showering/bathing or visiting the toilet and to change their feminine items as frequently as possible at times of heavy menstrual flow. If a woman does not wish to use tampons, she should choose towels with the highest absorbency rating and try to replace them before saturation occurs.

All diagnostic methods have limitations, particularly those with a subjective element. Each PBAC image is allocated a specific score, but patients may assign a range of blood volumes to the same icon category. The choice of icon may be influenced by whether a woman perceives her menstrual loss to be light, normal or excessive. A visual rating system is not as accurate as the AH method, which provides a definitive blood volume and does not rely on individual perception of MBL or errors in selecting or recording the correct icon. Pictorial charts may overestimate low volumes of MBL because whole blood represents a smaller proportion of the visible stain area than at normal and high volumes (Fig. 3). Conversely, the score assigned to the highest icon category can markedly underestimate MBL because of a ceiling effect as a sanitary item becomes fully saturated.

A potential drawback of using modern towels is the difficulty of visualizing the bloodstained area on the upper surface, because fluid is rapidly transferred to the interior region. The SAP-c menstrual pictogram overcomes this problem by scoring the stain on the underside of the towel, where it is clearly visible [5, 86]. Many stain profiles will not conform to the standardized PBAC icons [5, 94]. Stains may be fragmented or a completely different shape, which could potentially result in underestimation or overestimation of MBL, depending on personal perception of the soiled area.

Apart from the diversity in diagnostic accuracy – a fact highlighted by validated PBACs – there is a lack of standardization in measuring post-treatment outcomes. The aim of therapy is to improve a patient’s well-being to an acceptable level, but the relationship between quality of life and a clinically meaningful decrease in MBL remains unresolved. With a single exception, pictorial charts have not been validated as a tool to test the effectiveness of therapy [36], and researchers have had to define their own criteria for treatment success. Some have stipulated a decrease in PBAC score to below the chosen HMB cut-off value [95] or by a certain number of points from the baseline measurement of MBL, e.g. 50 points [15, 96‐98]. Others have advocated a percentage reduction in score; values of 20–50% have been reported [36, 58, 59, 63, 99‐105]. If the treatment goal is amenorrhea (PBAC score < 2), the endpoint is clearly defined [54, 106‐109]. The aim of several reports is just to detect a significant difference in MBL reduction between different treatments for HMB.

The appropriate time point to test efficacy will need to be considered, subject to the type and purpose of therapy. Medical treatments designed to rapidly reduce MBL tend to have short follow-up times of between 1 month and 12 months [44, 110‐113], whereas surgical interventions, such the levonorgestrel intrauterine system and endometrial ablation, are designed for long-term reduction of MBL. Monitoring of menstrual loss has been reported for up to 5 years after surgery at various time intervals, depending on the study [59, 106, 107, 114]. The post-treatment time point(s) for MBL measurement should be clearly defined in the trial design.

The role of PBACs in primary care is currently unclear. The usual procedure to assess MBL combines a woman’s self-perception of her menstrual flow (in terms of light, normal or heavy) with hemoglobin level and/or menstrual markers, such as length of period, number of used sanitary products, number of days of ‘heavy’ bleeding, number of flooding episodes and size of clots passed. Together with quality-of-life issues, this subjective approach is often the only trigger for medical and surgical interventions, although it is unreliable [2, 31, 76].

Measurement of MBL by the AH technique is obviously impractical in this setting, but a simple, semi-objective method would be a useful diagnostic tool. It could clarify the patient’s complaint and also influence the choice and expectations of treatment. Historical research data have shown that some women whose MBL has been shown to be normal can be dissuaded from seeking unnecessary therapy for perceived HMB [3, 115, 116]. A fully validated PBAC is theoretically suitable for this purpose, combined with a daily diary to document responses to specific health-related questions throughout the menstrual cycle. This day-to-day record would have the dual purpose of assessing both symptom severity and the impact of perceived or actual HMB on everyday life, which would comply with the National Institute for Health and Care Excellence (NICE) guidelines to assess quality-of-life issues for women with excessive blood loss. Although NICE guidelines do not recommend routine quantitation of MBL, they do acknowledge that further research into indirect measurements of MBL in primary and secondary care is warranted [6].

Some studies have recommended the use of a PBAC in primary care to increase the diagnostic accuracy of HMB [35, 36, 46], possibly as an adjunct to clinical history and menstrual markers [65, 117, 118]. However, widespread application of this method is currently limited by a lack of expert consensus regarding the choice of pictorial chart, the diagnostic cut-off score for HMB and clear objective criteria in the assessment of post-treatment clinical outcome.

At a practical level, the chart would require at least one menstrual cycle to complete. Results would not normally be available during a first consultation, unless a PBAC had been completed beforehand by prior arrangement with the healthcare center. A dedicated women’s clinic might suit this purpose, in which a nurse practitioner could train patients to use a validated PBAC, distribute the required sanitary protection and address any queries or anxieties [119]. Some women may be discouraged because of the inevitable delay in diagnosis, the rigorous attention to detail required when scoring MBL, the stipulation to change sanitary wear frequently to avoid leakage (which might not always be possible) or the need to use specified and possibly unfamiliar sanitary products. A diagnostic ‘primary care’ PBAC that has been validated for a range of popular sanitary products, particularly those designed for heavy menstrual flow, together with appropriate instruction and support may help to alleviate these concerns. Also, the widespread availability and popularity of smartphones opens the possibility of developing ePBACs that would be acceptable to patients. However, this area has yet to be explored.