Factors affecting delays in first trimester pregnancy termination services in New Zealand
Dr Martha Silva, Senior Research Fellow, School of Population Health, University of Auckland, Health Systems, Private Bag 92019, Auckland 1142, New Zealand; e-mail: firstname.lastname@example.org
Objective: To identify the factors affecting the timeliness of services in first trimester abortion service in New Zealand.
Method: Primary data were collected from all patients attending nine abortion clinics between February and May 2009. The outcome measured was delay between the first visit with a referring doctor and the date of the abortion procedure. Patient records (n=2,950) were audited to determine the timeline between the first point of entry to the health system and the date of abortion. Women were also invited to fill out a questionnaire identifying personal factors affecting access to services (n=1,086, response rate = 36.8%).
Results: Women who went to private clinic had a significantly shorter delay compared to public clinics. Controlling for clinic type, women who went to clinics that offered medical abortions or clinics that offered single day services experienced less delay. Also, women who had more than one visit with their referring doctor experienced a greater delay than those who had a single visit. The earlier in pregnancy women sought services the longer the delay. Women's decision-making did not have a significant effect on delay.
Conclusions: Several clinic level and systemic factors are significantly associated with delay in first trimester abortion services. In order to ensure the best physical and emotional outcomes, timeliness of services must improve.
In New Zealand as elsewhere, abortions are one of the most common surgical procedures. National statistics indicate that approximately one-quarter of all pregnancies in the country end with an abortion, and one in four women undergoes an abortion in her lifetime.1 Although abortion is a safe procedure when conducted under hygienic conditions by a trained provider, the risk for clinical complications increases with gestational age.2–5 Studies find a significant increase in risk of complications starting from the eighth2 or ninth week of pregnancy.4 Clinicians generally acknowledge that once a woman has chosen an abortion, the earlier the procedure, the better the clinical and psychological outcomes for the woman.
A 2009 study in New Zealand, on which this paper is based, found that only 46% of first trimester abortions happened by the tenth week of pregnancy, with an average of 25 days between the first visit with the referring doctor to enquire about an abortion and the date of the abortion procedure.6 This statistic compares unfavourably with Western Australia and countries such as the United States (US) and the United Kingdom (UK), where 71%, 68% and 59.2% of abortions respectively happen before the ninth week of pregnancy.7–9
Given the evidence in favour of reducing the gestational age at termination, many studies have been carried out in various countries to research the factors affecting the timing of abortions. Previous research identifies three main sources of delay: delay relating to the woman, such as delay in recognition of pregnancy or decision making process; delay relating to structural or service factors, such as appointment waiting times or raising money for fees; and delay related to the health system, such as a complicated referral processes.
Most US studies found that age, ethnicity and socioeconomic status are contributing factors to later abortion10–12 and report considerable delay in recognising and testing for pregnancy.13,14 One US study found that delays in suspecting and testing for the pregnancy caused 58% of delay in patients accessing second trimester abortion services.13 A study conducted in England and Wales found that among women accessing second trimester abortions, 71% reported delays in suspecting they were pregnant, 64% reported delays in confirming they were pregnant, 79% reported delays in deciding to have an abortion and 28% reported delays in first approaching a doctor to enquire about an abortion.15
Several studies also highlight the role of structural and logistical issues, such as raising enough money to pay for the procedure.10,13,14,16,17 This is a major concern for low-income women, particularly in the US, where public funding for abortion services is very limited.
Factors associated with the health system, such as referral processes, are cited as influencing delay in studies in the UK. Ingham and colleagues found that 60% of women reported delays in obtaining an abortion, which was attributable to the health service or system.15 Although another UK study points to limited secondary care capacity as a major source of delay (i.e. limited number of appointment slots per clinic), referral procedures can also add delay and be potentially upsetting when women confront an unsupportive primary care physician.18 This qualitative study with women in the UK showed that referring doctors often required women to have more time to think about their decision.18 However, studies also highlight that a woman's decision-making process for an unwanted pregnancy largely happens outside the influence of a medical professional and for many women begins as soon as the pregnancy is suspected.15,18
In an effort to promote high-quality abortion services, the UK Royal College of Obstetricians and Gynaecologists (RCOG) created a guideline for services that address quality-of-care issues including timeliness of services.19 The guideline recommended a maximum of 21 days between the first contact with a referring doctor and an abortion procedure. Finnie and colleagues found that 56% of surveyed women in the north-east of England were taking longer than this.20
This study aimed to identify what factors affect the timeliness of services in first trimester abortions in New Zealand. We hypothesised that factors related to the service characteristics including the type of abortion methods on offer, as well as systemic and personal factors related to the women using the services, would affect the timing in provision of services.
Abortion services in New Zealand
Abortion services in New Zealand are part of the core, publicly funded, health care services the population is entitled to access. Private abortion services are available but uncommonly used by women other than non-residents. Pregnancies that present a serious danger to the life of a woman, a serious danger to the physical or mental health of a woman, pregnancies resulting from incest or sexual relations with a guardian, pregnancies in women of mental subnormality, and pregnancies presenting fetal abnormality may all be legally terminated.21,22 Women who experience an unwanted pregnancy must first go to a referring doctor, usually a General Practitioner (GP) or a Family Planning doctor (FP) for confirmation of pregnancy, diagnostic tests and referral to an abortion clinic. Two certifying consultants must individually view the woman seeking an abortion and agree that the case fulfills legal grounds for an abortion. At the time of the study, there was no national service guideline stating the recommended maximum number of days between referral and abortion, as there is in other countries.
A mixed methods approach was used to investigate the factors affecting timeliness of first trimester abortion services in New Zealand. Nine of a total of 13 first trimester clinics throughout the country agreed to participate in this study. The first methodology used was an audit of the clinical records of all patients attending participating abortion clinics between 1 February and 30 April 2009 (n=2,950). The nine participating clinics ranged in size from six beds per week for abortion services to 140 beds per week; together they account for about 70% of all abortions in the country. Eight clinics were on the North Island and one on the South Island. All non-participating clinics were similar in structure to at least one participating clinic. Research assistants routinely visited the participating clinics and extracted data from the paper based clinical records. Data were recorded on a data collection sheet. All patients were assigned a study identification number in order to protect their identities. In this way, no identifying information, such as name, address or national health identification number, left the clinic premises.
The second method of data collection was via a questionnaire. Participating clinics invited all women attending their clinics during the study period to respond. They were given a questionnaire to complete in the waiting room and return to a staff member in a sealed envelope. Of the 2,950 women who attended, 1,086 agreed to complete a questionnaire (response rate=36.8%). The responses from women who chose to participate were linked to the audit information by use of a study-generated identification number. This helped ensure that data could be linked while protecting patients’ confidentiality. Women consenting to completing the questionnaire understood that their data would be linked while ensuring their anonymity. Questionnaire respondents were compared to the complete audit sample to see how representative of this sample they were. Women who answered the questionnaire were not significantly different in age from the complete audit sample and they did not differ in the average delay to termination of pregnancy (TOP). However, New Zealand European and Māori ethnicities were overrepresented in the questionnaire sample.
This study received ethical approval by the Multi-Region Ethics Committee of the Ministry of Health of New Zealand (approval number MEC 08/10/120).
Throughout this study, the outcome of interest was total delay, defined as the total number of days between the first contact with the referring doctor and the abortion procedure. In the case of medical abortions, the date of abortion was defined as the date of expulsion of the products of conception. When the date of first contact with referrer was not available, the earliest date available from diagnostic tests, referral letter and date when the appointment for the TOP clinics was booked, was taken as a proxy. Data constituting the outcome of interest were recorded and stored consistently across clinics.
Explanatory factors include clinic factors, sociodemographic and personal factors. Three clinic types were defined: high-capacity public clinics (clinics that had more than two theatre lists per week), low-capacity public clinics (clinics that had less than two theatre lists per week), and private clinics. Other clinic factors included whether the clinic provided medical abortions, whether counselling is mandatory and whether they offer a single day service. Sociodemographic and personal variables obtained from the clinic audits included age, ethnicity, type of referrer, gestational age at scan. Personal variables obtained via the patient questionnaire include timing of the decision to have an abortion, number of doctor consultations, whether a social worker/counsellor was seen, friend and family support, out-of-pocket expenses (in New Zealand dollars), travel time and perceptions of timeliness. Personal variables obtained through the clinic audit and through the voluntary questionnaire are analysed and presented separately.
Bivariate analyses were conducted between all independent variables using chi-square tests of independence to assess the potential for collinearity issues in the ANOVA models. One-way ANOVA and ANCOVA models were used to identify the independent variables that were associated with the dependent variable – time between the first visit with a referring doctor and the abortion procedure. Due to the fact that clinic type was found to be strongly associated with the outcome variable, all one-way ANOVA models controlled for clinic type. Once the effect of individual variables had been identified, factorial univariate ANOVA models were built to test whether the effects of the factors were maintained when controlling for all other factors. Post hoc tests were run using a Bonferroni correction. All analyses were conducted using SPSS for Windows Version 15, with a significance level of 5%.
A full description of the study data has been published elsewhere.6 Only 38.6% of women participating in this study had a delay of 21 days or less, as recommended by RCOG (data not shown).
Of the nine participating clinics, two were high-capacity public clinics, six were low-capacity public clinics and one was private. Four clinics provided medical terminations of pregnancy (MTOPs), and six clinics made counselling a mandatory part of their service. Six clinics offered a same day service, whereas the remaining three clinics organised their services as a two day process, not necessarily on consecutive days.
Table 1 shows the average total delay to TOP procedure in relation to the different clinic factors. The effect of each clinic factor on delay was tested individually and subsequently added to a full model to control for the other clinic factors. Clinic type was significantly associated with delays to TOP, with the private clinic averaging 13.9 days delay, while the low-capacity public clinics and the high-capacity public clinics averaged around 26 days (F(2,2946)=247.521, p<0.001). Controlling for clinic type, clinics that did not offer MTOP services had significantly longer delays, with an additional two days of delay compared to clinics who did offer MTOPs to their patients (F(1,2945)=54.854,p<0.001). Clinics who offered counselling to women but did not require it in order to proceed with services had an average delay of 22.6 days, compared to a significantly longer delay of 25.7 days for clinics that made counselling mandatory for all patients (F(1,2944)=69.534, p<0.001). Similarly, clinics that offered a same day service had significantly shorter delays to procedure than clinics who required multiple visits (F(1,2945)=4.401, p=0.036). When all clinic factors were added to the full model simultaneously, to control for the interaction between factors, the only variable that ceased to be significantly associated with delay was mandatory counselling. Together the clinic factors accounted for just over one-fifth of the variation in total delay (R2=0.225).
Table 1. Sample description and average days delay between the first visit with referring doctor and TOP procedure, by clinic factors.
|Clinic Type|| || || || || |
|High-capacity Public||2||1741||26.7 (10.2)||<0.001||<.001|
|Low-capacity Public||6||352||25.9 (10.1)|| || |
|Private||1||857||13.9 (7.9)|| || |
|MTOPs available|| || || || || |
|No||5||813||26.4 (9.9)|| || |
|Counselling compulsory|| || || || || |
|No||3||755||22.6 (9.5)|| || |
|Single day service|| || || || || |
|No||5||2041||26.4 (10.3)|| || |
Looking at personal factors associated with delay (see Table 2), we see that over half the sample was under 25 years of age and New Zealand European. About 23% of women were Māori and 12% were Pacific. Almost all women were referred by either a GP (75%) or a Family Planning clinic (17.4%). Almost half of all women had a scan before the sixth week of pregnancy, and another 35% had a scan by the end of the eight week of pregnancy. The age of the woman seeking an abortion and the gestational age at the time of scan have a statistically significant individual effect on delay. Controlling for clinic type, women twenty years of age or younger experience a delay of 26.1 days, which gradually decreases as age increases, with women 41 years old and above having the shortest delay of 21.4 days (F(5,2930)=3.117,p=0.008). Post hoc tests indicated that women over 36 years old were significantly different to women under 36 years old. Similarly, women with the youngest gestational age at scan had the longest delays to procedure. Women with pregnancies of less than 48 days (or less than 6 weeks) waited an average of 26.6 days, women with pregnancies between 49 and 62 days old (7 weeks – 8 weeks 6 days) waited 25.2 days, women with pregnancies between 63 and 76 days (9 weeks – 10 weeks 6 days) waited 20.4 days and women with pregnancies of more than 77 days (11 weeks onwards) waited 14.6 days (F(3,2933)=29.528, p<0.001). Post hoc tests indicated that all categories were significantly different from each other. When clinic type, age, ethnicity, referrer type and gestational age at scan were added to the model simultaneously, only gestational age at scan was significantly associated with delay (F(3,2431)=8.991, p<0.001). Together the personal factors from the audit data accounted for between one-fifth and one-quarter of the variation in total delay (R2=0.238).
Table 2. Percentage distribution of socio demographic and personal factors and average days delay between the first visit with referring doctor and TOP procedure, by personal factors
|Age|| || || || || |
|21–25||28.9|| ||25.1 (10.9)|| || |
|26–30||18.5|| ||24.7 (10.4)|| || |
|31–35||11.7|| ||24.7 (10.4)|| || |
|36–40||8.7|| ||22.3 (9.8)|| || |
|41+||3.3|| ||21.4 (10.7)|| || |
|Ethnicity|| || || || || |
|NZ European||52.5||2950||25.6 (10.0)||NS||–|
|Māori||22.8|| ||25.8 (11.0)|| || |
|Pacific||12.3|| ||26.2 (11.1)|| || |
|Chinese||5.8|| ||20.6 (12.1)|| || |
|Other Asian||7.5|| ||24.6 (10.6)|| || |
|Other||11.6|| ||21.6 (11.0)|| || |
|Referring doctors|| || || || || |
|Family Planning||17.4|| ||25.5 (10.5)|| || |
|Other||6.3|| ||21.2 (12.0)|| || |
|Gestational age at scan (days)|| || || || || |
|49–62||34.6|| ||25.2 (10.2)|| || |
|63–76||12.1|| ||20.4 (9.3)|| || |
|77+||3.7|| ||14.6 (10.6)|| || |
Women's Questionnaire Results
Table 3 shows the distribution of personal factors that were collected via the questionnaire. Two-thirds of women made the decision to have an abortion either before or as soon as they had a confirmed pregnancy, more than two-thirds had more than one consultation with the referring physician, 65% had seen a social worker, and 58% had the full support of their friends and family. Twenty-eight per cent reported that nobody knew about their pregnancy. Three-quarters of the women participating in the study travelled under one hour to get to the clinic and half reported out of pocket expenses of under NZ$50.
Table 3. Percentage distribution of personal and process factors and average days delay between the first visit with referring doctor and TOP procedure, by personal factors
|Timing of decision to have abortion|| || || || || |
|Before pregnancy was suspected||11.2||1082||23.0 (10.5)||NS||–|
|As soon as pregnancy was suspected||26.0|| ||23.9 (9.4)|| || |
|As soon as pregnancy was confirmed||29.2|| ||24.7 (10.1)|| || |
|After pregnancy was confirmed||33.6|| ||25.2 (10.8)|| || |
|Number of doctor consultations|| || || || || |
|2||47.7|| ||24.8 (9.7)|| || |
|3+||20.1|| ||27.8 (11.4)|| || |
|Social worker seen|| || || || || |
|No||31.7|| ||23.2 (10.4)|| || |
|Not yet||3.5|| ||24.1 (9.8)|| || |
|Friend and family support|| || || || || |
|Fully support||57.6||1014||24.2 (10.0)||NS||NS|
|Some support||13.8|| ||27.6 (10.4)|| || |
|Do not support||0.7|| ||26.6 (11.0)|| || |
|Nobody knows||27.9|| ||23.2 (10.3)|| || |
|Travel time to clinic|| || || || || |
|Under 20 minutes||34.5||1019||23.9 (10.5)||NS||–|
|21–60 minutes||41.5|| ||24.7 (9.9)|| || |
|1–2 hours||17.3|| ||25.2 (10.2)|| || |
|Over 2 hours||6.7|| ||24.1 (11.1)|| || |
|Total out of pocket expenses·|| || || || || |
|No reported expenses||14.6||1086||24.5 (0.8)||NS|| |
|Under $50||60.3|| ||25.0 (10.0)|| || |
|$51–$100||14.2|| ||23.5 (10.6)|| || |
|$101–$200||4.4|| ||26.1 (10.4)|| || |
|More than $200||6.5|| ||20.6 (10.8)|| || |
Service factors obtained from the questionnaire that were significantly associated with delay to procedure included the number of consultations with the referring doctor (F(3,1005)=6.839, p<0.001). Women that had only one visit with the referring doctor before being referred to the TOP clinic had an average delay of 22.6 days, and women who had two visits with the referring doctor had an extra two days delay. When women had three or more consultations with the referring doctor they added on average a further three days to the delay. Post hoc tests indicated that all categories were significantly different from each other. Controlling for clinic type, when timing of abortion decision, number of doctor consultations, whether a social worker was seen, friend and family support were simultaneously added to the model, the number of doctor consultations remained significantly associated with delay (F(3,526)=3.026, p=0.029). Together the personal factors from the questionnaire data accounted for just over one-tenth of the variation in total delay (R2=0.129).
Women's perceptions of delay were then assessed to see whether women who wanted to have an abortion sooner actually experienced a greater delay than those who indicated that the time they waited was all right. Women who thought they had waited too long to have an abortion had an average delay of 27 days (F(5,998)=7.644,p<0.001). Women who thought the time they had waited for an abortion was all right had an average delay of 20.9 days. Some women did not know when their procedure would be scheduled at the time they responded to the questionnaire, and these women had an average delay of 30.6 days (data not shown).
This study assessed sociodemographic factors, clinic and structural factors and personal factors associated with delay in access to first trimester abortion services. Several structural factors were found to be significantly associated with delay, with the type of clinic attended having the largest single effect. Not surprisingly, women who went to a private clinic waited an average of twelve fewer days between the first visit with a referring physician and the abortion procedure compared to public clinics of any size. Controlling for clinic type, women who went to clinics that offered medical abortions or clinics that offered single day services experienced less delay. Also, women who had more than one visit with their referring doctor experienced a greater delay than those who had a single visit. Across all clinics, a systemic factor influencing timeliness of services became apparent. The earlier in pregnancy women started the process to access an abortion service, the longer they had to wait to access services. With relation to the most commonly cited personal factor in international literature,13–15 the timing of the decision-making process did not have a significant effect on delay.
Strengths of this research include the multi-dimensional approach to assessing factors associated with delay. The study design allowed us to assess systemic, clinic and structural issues as well as personal factors to broadly illustrate their impact on timeliness of services. However, this study also had limitations which must be mentioned. The audit methodology did not include data from the referring doctor, so the timing of events before the referral to the abortion clinic was either constructed via the questionnaire, or proxy measures were used. However, we believe that the proxy measures give us a conservative estimate of delay. Secondly, the sample of women answering the questionnaire cannot be considered representative of the audit population. Thirdly, although we know that two-thirds of the women included in this study made the decision to have an abortion before or as soon as they had a positive pregnancy test, we are unable to quantify the delays associated with any further difficulty or fluctuation in decision-making between the first visit to the referring physician and the abortion procedure. Despite these limitations, this study represents the first large-scale assessment of timeliness of abortion services in New Zealand.
Except for the type of clinic visited, most other variables explored in the three models account for a change in delays ranging from two to five days. Reducing delays by even these amounts would be programmatically significant, considering the difference in patient satisfaction with the differing lengths of delay. Taken together, the factors investigated explained a significant proportion of the variance in the delay experienced by women using these clinics. Furthermore, the funding of health services in New Zealand is such that women do not have an option to choose which clinic to go to, unless they are prepared to pay the price for private services. If a woman lives in an area that does not offer medical abortion and has a two-day service structure, she is likely to have to wait longer for her procedure than a woman living in an area that does offer medical abortions and single day services. This suggests that inequities exist in the access to more timely services.
Findings also indicate that services are currently provided according to greatest ability to benefit. Women with higher gestations and therefore increased likelihood of complications are fast-tracked through the system in order to offer first trimester services and avoid the costlier and more complicated second trimester services. However, women identifying early pregnancies are particularly disadvantaged by this. Although many clinics now offer medical abortions, infrastructure and other barriers limit the number of patients that can access this service per clinic. The lack of capacity for many clinics to offer medical abortion services largely contributes to this delay for women with earlier gestations. In addition, many abortion providers in New Zealand prefer conducting procedures after the eighth week of pregnancy to reduce the risk of retained products. Therefore, women with early pregnancies are given delayed appointments.
Findings suggest the need for a more streamlined referral process. Given that the decision-making process is reviewed as part of the certification process, and social workers are an integrated part of many clinic services, women not displaying any clinically concerning conditions do not necessarily benefit from multiple visits with referrers. The New Zealand standards of care for women accessing abortion services23 do not address the timeliness of services from the first contact with the referring doctor to the time of termination. This document only notes that referring doctors should be aware of the timeline and refer as soon as it is evident that the woman is not happy with the diagnosed pregnancy. However, the standards do specify that women must not wait longer than two weeks from the time of referral to the time of the procedure, yet they do not provide a definition for what should be considered time of referral. In the context of this study, time of referral was defined as the time of booking the appointment with the abortion clinic. Under this definition, women were waiting more than two weeks between time of referral.6 In New Zealand, bypassing traditional gatekeepers to allow women to self refer into abortion clinics would require a change in law, but would reduce the number of steps and simplify the process. If the current system is kept, GPs and other referring physicians must be reminded of the importance of streamlining the process for women and referring as soon as the pregnancy is diagnosed and identified as unwanted.
Abortion services attend to women at an emotionally difficult and often stressful time, where delays in service can lead to an increased risk of physical and emotional complications. Therefore, once a woman has decided to terminate a pregnancy, abortion services should strive to provide a timely and high quality service. Systemic factors and clinic level factors affecting delays in services must be systematically reviewed in order to ensure equity among all New Zealand women.