Background: Unintended pregnancies pose substantial risk to mothers and children. In Pakistan, unintended pregnancies account for 46% of all pregnancies. Lack of geographic access to open and well-supplied family planning (FP) centers may be related to the occurrence of such pregnancies, particularly in rural areas. Objective: The objective of this analysis is to determine if geographic access to family planning centers in the Thatta district of Pakistan is related to unintended pregnancy rates among married women. Methods: We conducted a community-based, nested case-control study of 800 pregnant women identified from the database of an active surveillance system, which registers and follows all pregnant women in the catchment area of Thatta district. Women were enrolled during the first trimester; those that reported their pregnancy to be unintended were selected as cases (n = 200), and those whose pregnancies were intended served as controls (n = 600). We defined geographic access as including both the distance of a family planning center from the woman’s home, and availability of personal transportation. Logistic regression was used for analysis. Results: In the multivariate model, neither distance [OR = 1.0; 95% CI (0.95 - 1.05)] nor availability of transportation [OR = 1.14; 95% CI (0.78 - 1.67)] were significantly associated with unintended pregnancy. In fact, women with unintended pregnancies were more likely to be aware of family planning [OR = 2.21; 95% CI (1.23 - 3.97)] and more likely to have been using a contraceptive method before conceiving their index pregnancy [OR = 3.59; 95% CI (1.83 - 7.06)]. Other factors related to unintended pregnancy were older maternal age [OR = 1.13; 95% CI (1.08 - 1.17)], having already had at least one son [OR = 3.13; 95% CI (1.93 - 5.07)]; spousal opposition to contraceptive use, [OR = 3.24; 95% CI (1.89 - 5.56)] and low spousal education level [OR = 1.85; 95% CI (1.08 - 3.18)] as compared to women with intended pregnancy. Conclusion: Lack of geographic access to FP centers is not a risk factor for unintended pregnancy in women from the Thatta district. However, in this population, unintended pregnancies are more common among older women, women having at least one son, and those who have a spouse who does not approve of contraceptive use, and is less educated. Of note, women who reported unintended pregnancy did have knowledge about FP and were more often using contraceptives before they conceived.
Unintended pregnancies have substantial, social, economic and health consequences on families [
In Pakistan, out of about 9 million pregnancies, 46% are believed to be unintended and 54% of these end in induced abortions [
Women who experience unintended pregnancies either have an unmet need for contraception or experience a failure of the contraceptive method [
Lack of access to FP services is considered an important risk factor of unmet need and/or unintended pregnancies [
Since 2008, The Department of Community Health Sciences (CHS) at Aga Khan University (AKU) has maintained a maternal-newborn health (MNH) registry in 14 union councils (UC) of District Thatta [
Pregnant women who were residents of the 14 UCs and were enrolled in MNH registry before 12 weeks of gestation were included in this study. Cases were defined as pregnant women who already had children and did not want more at all, or they wanted to have children but later on in their lives. The controls were pregnant women who reported their current pregnancy as intended. Pregnancy intention was assessed during the enrollment of participants in MNH registry. Geographic access to working FP centers was defined as distance in kilometers from women’s homes to the FP centers and availability of transportation.
200 women who reported unintended pregnancy and 600 women who reported their pregnancy as intended. This sample size was based on proportion of exposure and other known risk factors of unintended pregnancy among controls ranging from 10% - 70% [
Computer generated identification numbers of cases and controls were obtained from data management system of MNH registry and initially 2887 cases and 7368 controls were identified irrespective of any trimester at the time of enrollment. Based on the eligibility criteria of ≤12 weeks of gestation during the index pregnancy, only 300 cases and 735 controls were identified from the MNH registry. Considering the required sample size, only 235 cases and 692 controls were approached for the interview and 35 cases and 92 controls were excluded due to different reasons as shown in
The trained interviewers approached eligible participants by identifying their home addresses from the registry. A separate written informed consent was taken from each participant prior to inclusion in the study. Data about socio-demographic, socioeconomic, fertility-related, contraceptive use, and access-related factors were collected from women by data collectors. Information regarding knowledge about contraceptive methods and FP centers,
home visits and FP services provision by lady health workers (LHW), were also collected from participants. LHW are female community health workers who provide a range of services including antenatal, postnatal, FP, and child health services by making home visits in the community [
The data for geographic access, or the measurement of the distance from patient’s homes to FP centers, was extracted from three sources: information on household address from the MNH registry, GPS data of sampled households and their nearest FP centers in the 14 UC, and through the determination of adigitized road network of the 14 UC of Thatta district by the data collectors.
High resolution 0.6 m quick-bird satellite imagery was acquired from Google Earth Pro to develop geographic information system (GIS) database of transportation system of 14 UCs of the Thatta district. Satellite imagery plays a vital role in raster data handling, a type of geographic information system (GIS) data, by providing an excellent basis from which to then extract relevant geographic data information. The images of 14 UCs were downloaded in multiple steps requiring that individual images be mosaicked to obtain a single composite image of each UC that could be used for geo-rectification, or confirmation of geographic location, in order that individual FP centers would be able to be referenced by their exact pixel address [
Thatta district’s road network includes national highway, paved, and non-paved roads (
After creating the digitalized road network, the data from the GPS survey was imported into Arc GIS version 9.3 to output measures of the distances from patient homes to FP centers. Distance was calculated from household to nearest center, assuming that the women are likely to visit the center most proximate to their home. Moreover, the shortest path with respect to distance conforms to the principal of “nearest neighbor from origin” concept, which depends upon real-world route availability, as shown in
Data was analyzed using IBM SPSS version 19 and STATA version 11. Chi-square and independent t-tests were applied for comparative analysis between cases and controls for categorical and continuous variables, respectively. Logistic regression was done to assess the univariate associations with risk factors, which were characterized by odds ratios (OR) and associated 95% confidence intervals (95% CI).
Multi-collinearity was assessed and checks were made for biological plausible interactions between independent variables. Variables having biological and significant associations (p-value < 0.25) with the outcome (unintended pregnancy) were assessed further in multivariate regression analysis. Variables were included in the final model by assessing their significance at a p-value of <0.05 and likelihood ratio testing. The scale of the continuous variables such as age, road network distance, and number of living sons, were checked for assumption of linearity by quartile analysis. The fit of the final model was tested using the Hosmer-Leme show test. Adjusted odds ratios with their 95% confidence intervals were used for interpreting the results.
This study was reviewed and approved by the Ethical Review Committee of the Aga Khan University Karachi, Pakistan.
Women with unintended pregnancies had a higher mean age of 29.7 years compared to women with intended pregnancies 25.5 years (p < 0.05). More women with unintended pregnancies, and their husbands, had an education level of “completed primary or less” (90% and 66.5%, respectively) as compared to women with an intended pregnancy (82.3% and 60%). Of note, women with unintended pregnancies got married at somewhat
earlier ages (19.5 years ± 3.4 years) as compared to women who had intended pregnancies (20.1 years ± 3.3 years), and the difference was statistically significant (p < 0.05). In addition, women with unintended pregnancies had a higher gravidity, number of living children, and number of sons than women with intended pregnancies (
There was no difference in the calculated distance to FP centers from the homes of the women who reported unintended versus those who reported intended pregnancy. The median road network distance between households and the nearest FP center for women with unintended pregnancies was 0.814 km (IQR: 0.32, 4.6) and 0.810 km (IQR: 0.33, 4.41) for women with an intended pregnancy. In terms of transportation, 34.5% of the women with unintended pregnancies and 35.8% of the controls reported that they had access to their own transport (
A significantly higher proportion of the women with unintended pregnancies had knowledge (62%) about the FP centers and visited these centers (18%) as compared to women in the control group (45.3% and 11.5% respectively). Almost equal numbers of women in both groups mentioned that FP centers are close to their homes (women with unintended pregnancy 55.6% vs. 55.1% women with intended pregnancy), and can be accessed on foot (66.7% of the women with unintended pregnancy and 58.0% of the women with intended pregnancy). The majority of the women in both groups were aware of the LHW program and the availability of door to door contraceptive service provision, with 74.5% of the women with unintended pregnancies and 68.3% of the controls having been visited by a LHW at least once during their married life (mean of 10 years) (
In terms of contraceptive knowledge and practices, 28% of the women with unintended pregnancies and 14.7% of the women with intended pregnancies had knowledge of FP methods and were using contraception before conceiving their index pregnancy. Opposition by the woman’s husband and opposition to contraceptive use by the woman herself were both higher among women with unintended pregnancies than in the control group. Fear of the side effects of contraceptive use was also higher among women with unintended pregnancies, 27.5% than in women with intended pregnancy, 6.8%.
Univariate analysis showed that geographic access (distance and personal transportation) was not significantly associated with unintended pregnancy (OR 0.99; 95% CI (0.96 - 1.03) and OR: 1.06; 95% CI (0.76 - 1.48)). Increasing age of the women [OR 1.18 (95% CI: 1.14 - 1.22)], their primary education status [OR 2.09 (95% CI: 0.97 - 4.52)], and the education level of their husbands [OR 1.78 (95% CI: 1.09 - 2.91)], were found to be significantly associated with unintended pregnancy on univariate analysis. Biological factors such as increasing gravidity [OR 1.43 (95% CI: 1.34 - 1.52)], number of living children [OR 1.72 (95% CI: 1.58 - 1.89)], having at least one living son OR 2.15 (95% CI: 1.78 - 2.48), and age at the time of marriage [OR 0.95 (95% CI: 0.90 - 0.99)] were also associated with unintended pregnancy.
Having contraceptive knowledge and utilizing FP methods both were associated with an increased odds ratio of unintended pregnancy [5.26 (95% CI: 2.94 - 9.41)], but interestingly, so was having contraceptive knowledge for women who were not using a contraceptive method [OR 2.91 (95% CI: 1.17 - 4.88)]. Moreover, having a husband who opposed the use of FP methods [OR 2.89 (95% CI: 1.45 - 5.75)], or being afraid of side effects of contraceptive use OR 5.17 [(95% CI: 3.32 - 8.06)], also increased the risk of a woman experiencing an unintended pregnancy. On the other hand, home visits by a LHW [OR 1.35 (95% CI: 0.94 - 1.94)] did not significantly increase the odds of unintended pregnancy.
In the final multivariate model, distance to a FP center [OR 1.00; 95% CI (0.95 - 1.05)] and availability of personal transportation OR: 1.14; 95% CI (0.78 - 1.67) remained insignificant predictors of unintended pregnancy (
Characteristic | Cases | Controls | Unadjusted ORs | ||
---|---|---|---|---|---|
Socio-demographics | |||||
Age | 29.66 ± 4.80* | 25.50 ± 4.89* | 1.18 [1.14 - 1.22]*** | ||
Women’s education Intermediate and above Middle to secondary Primary or less | 180 [90%] 12 [6%] 8 [4%] | 494 [82.3%] 60 [10%] 46 [7.7%] | 1 1.15 [0.42 - 3.04] 2.09 [0.97-4.52]*** | ||
Husband’s Education Intermediate and above Middle to secondary Primary or less | 133 [66.5%] 44 [22%] 23 [11.5%] | 360 [60%] 129 [21.5%] 111 [18.5%] | 1 1.65 [0.94 - 2.90] 1.78 [1.09 - 2.91]** | ||
Women autonomy Yes No | 163 [81.5%] 37 [18.5%] | 438 [73%] 162 [27%] | 1 0.61 [0.41 - 0.91]*** | ||
Socio-economic score | 11.5 [9, 14]† | 12 [9.25, 16]† | 0.97 [0.94 - 1.00]** | ||
Religion Non-Muslim Muslim | 6 [3%] 194 [97%] | 21 [3.5%] 579 [96.5%] | 1 1.17 [0.47 - 2.95] | ||
Fertility and access to FP services | |||||
Age at marriage | 19.48 ± 3.44* | 20.06 ± 3.33* | 0.95 [0.90 - 0.99]*** | ||
Gravidity | 5 [4, 8]† | 2 [1, 4]† | 1.43 [1.34 - 1.52]*** | ||
Number of alive children | 4 [3, 6]† | 1 [0, 3]† | 1.72 [1.58 - 1.89]*** | ||
Number of alive sons No alive son At least one alive son | 26 [13%] 174 [87%] | 300 [50%] 300 [50%] | 1 2.15 [1.87 - 2.48]*** | ||
Distance of FP center from house of women | 0.81 [0.32,4.61]† | 0.81 [0.33,4.41]† | 0.99 [0.96 - 1.03] | ||
Availability of personal transport Yes No | 69 [34.5%] 131 [65.5%] | 215 [35.8%] 385 [64.2%] | 1 1.06 [0.76 - 1.48] | ||
Awareness of FP centre Yes No | 124 [62%] 76 [38%] | 272 [45.3%] 528 [54.7%] | 1 0.51 [0.37 - 0.71]*** | ||
Visited FP centre for availing services Yes No | 36 [18%] 164 [82%] | 69 [11.5%] 531 [88.5%] | 1 0.59 [0.38 - 0.92]*** | ||
Knowledge and family planning service provision by LHWs | |||||
Awareness about LHWs Yes No | 162 [81%] 38 [19%] | 476 [79.3%] 124 [21.7%] | 1 0.9 [0.60 - 1.35] | ||
Door step visit by LHWs at least once No Yes | 51 [25.5%] 149 [74.5%] | 190 [31.7%] 410 [68.3%] | 1 1.35 [0.94 - 1.94]** | ||
Ever discussion about FP with woman by LHW Yes No | 61 [30.5%] 139 [69.5%] | 100 [16.7%] 500 [83.3%] | 1 0.46 [0.31 - 0.66]*** | ||
FP method offered by LHW Yes No | 31 [15.5%] 169 [84.5%] | 46 [7.7%] 554 [92.3%] | 1 0.45 [0.28 - 0.74]*** | ||
Woman took FP method from LHWs Yes No | 28 [14%] 172 [86%] | 44 [7.3%] 556 [92.7%] | 1 0.48 [0.29 - 0.81]*** | |
---|---|---|---|---|
Women used the FP method given by LHWs Yes No | 23 [11.5%] 177 [88.5%] | 36 [6%] 564 [94%] | 1 0.49 [0.28 - 0.85]*** | |
Women were satisfied with FP method given LHW Yes No | 16 [8%] 184 [92%] | 25 [4.2%] 575 [95.8%] | 1 0.5 [0.26 - 0.95]*** | |
Knowledge and family planning service utilization | ||||
Awareness of FP center Yes No | 124 [62%] 76 [38%] | 272 [45.3%] 328 [54.7%] | 1 0.51 [0.37 - 0.70]*** | |
Visited FP center for availing services Yes No | 36 [18%] 164 [72%] | 69 [11.5%] 531 [88.5%] | 1 0.59 [0.38 - 0.92]*** | |
Knowledge and current use of any FP method No knowledge about any family planning method Knowledge with current use Knowledge without current use | 19 [9.5%] 56 [28%] 125 [62.5%] | 157 [26.2%] 88 [14.7%] 355 [59.2%] | 1 5.26 [2.94 - 9.41]*** 2.91 [1.17 - 4.88]*** | |
Husband’s opposition for using contraceptives No Yes | 160 [80%] 40 [20%] | 561 [93.5%] 39 [6.5%] | 1 2.89 [1.45 - 5.75]*** | |
Respondent’s opposition No Yes | 186 [93%] 14 [7%] | 577 [96.2%] 23 [3.8%] | 1 1.18 [0.92 - 3.56]** | |
Fear of side effects or difficulty in using contraceptives No Yes | 245 [72.5%] 55 [27.5%] | 559 [93.2%] 41[6.8%] | 1 5.17 [3.32 - 8.06]*** | |
*Mean ± SD [all such values]. †Median, interquartile range in parenthesis [all such values]. **p-value < 0.25; ***p-value < 0.05. OR: Odd ratio, CI: Confidence interval.
Variable | Adjusted OR [95% CI] |
---|---|
Distance [km] | 1, [0.95 - 1.05] |
Having personal transportation Not having personal transportation | 1 1.14, [0.78 - 1.67] |
Maternal age | 1.13, [1.08 - 1.17] |
Alive sons No alive son At least one alive son | 1 3.13, [1.93 - 5.07] |
Educational status of husband Intermediate and above Middle to secondary Primary or less | 1 1.68, [0.90 - 3.14] 1.85, [1.08 - 3.18] |
Knowledge and current use of any FP method No knowledge Knowledge with current use Knowledge without current use | 1 3.59, [1.83 - 7.06] 2.21, [1.23 - 3.97] |
Husband’s opposition No husband’s opposition Husband’s opposition | 1 3.24, [1.89 - 5.56] |
OR: Odd ratio, CI: Confidence interval.
This study indicates that geographic access has no measurable impact on the occurrence of unintended pregnancies in the catchment population of Thatta district. This relationship between distance and unintended pregnancy was assessed after categorizing the distance into >2, >5 and >10 kilometers but the association remained insignificant. This might be due to the fact that, healthcare centres providing FP services are not very far from the homes of women as evidenced by geographical data (
Comparison of distance as an indicator of access to FP services with other studies is difficult, as most of the available literature has assessed the impact of geographic access on the use of health services in general. Studies looking at the impact of geographic access on use of FP services vary from country to country. Our findings are consistent with evidence from Nepal and Thailand that demonstrates that distance is not a limiting factor for access to FP services [
This study showed that women with unintended pregnancies were more likely to be older. Targeting this population subgroup with specific programs might reduce unintended pregnancy in this region of Pakistan. These findings were consistent with the study conducted in Nepal [
Another important finding is that almost twice as many women who had unintended pregnancies knew about and were using a modern FP methods before their index pregnancy (28%) as compared to women with intended pregnancies (14%). Women using a FP method were more likely to regard a pregnancy occurring as unintended, consistent with previous international research [
Women whose husbands were opposed to contraceptive use or were less educated were more likely to face unintended pregnancy. Both members in a couple should be equally responsible for FP, but to date the responsibility of Pakistani men has largely been neglected. As a topic for further research, collecting information about men’s attitudes toward and knowledge about use of contraceptives could contribute greatly to an understanding of FP generally, and unintended pregnancy specifically. The quality of a couple’s relationship and their inter- spousal communication has been shown to have an effect on contraceptive use [
An important strength of the study was the enrollment of pregnant women in the first trimester which presumably allowed for more accurate assessment of pregnancy intention, which reduced recall bias. Secondly, a validated questionnaire based on the standard demographic definition of unintended pregnancy, was used [
Lack of geographic access to working FP centers was not found to be associated with unintended pregnancies in Thatta district. However, based on our analysis, we proposed that the subgroup of older women with at least one living son should be targeted for FP services. This program might be focused on immediate postpartum contraception which has been shown to be a very effective time for intervention [
We also propose that the quality of FP programs should be improved with active promotion of LARC methods and use of emergency contraception. Previous analyses suggest that sufficient information about emergency contraception is generally not conveyed effectively [
Further studies need to be conducted with special attention to the role of men, to the causes of method failure and non-use of contraceptive methods, with attention to how they impact both intended and unintended pregnancy rates.
Author declares that there is no financial or non-financial competing interest for this work.
This study was carried out as a thesis work for MSc Epidemiology and Biostatistics offered by the Aga Khan University. As part of the program and as per university policy I was given 100,000 PKR (approximately USD 1000) for carrying out my research work.
This study was reviewed and approved by the Ethical Review Committee of the Aga Khan University Karachi, Pakistan.
Contribution by Dr. Sumera Aziz Ali: Dr. Sumera Aziz Ali had made the proposal, developed the questionnaire, manual of operations, data collection, cleaning of data and did the analysis for this paper. She has substantial contributions to conception and design and acquisition of data or analysis and interpretation of data. She drafted the manuscript which was finalized by contribution from other authors. She also gave the final approval of the version to be published. She is also accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.
Contribution by Dr. Sarah Saleem: Dr. Sarah Saleem gave her feedback during hypothesis formulation and development of research question. She also gave her rigorous feedback for analysis and reviewed the manuscript critically. She had been involved in drafting complete manuscript and revising it critically for all aspects and has given final approval of the version to be published.
Contribution by Dr. Neelofar Sami: Dr. Neelofar Sami gave her feedback during hypothesis formulation, development of research question and questionnaire. She also helped in formation of manual of operation to conduct the study. She gave her feedback for introduction part of the manuscript.
Contribution by Prof. Dr. Mir Shabbar Ali: Prof. Dr. Mir Shabbar Ali advised in planning and execution of GPS survey and development of GIS maps with special reference to transportation infrastructure of the study area.
Contribution by Mohammad Ahmed: Mr. Mohammad Ahmed helped in technical part of GIS; he supervised in marking the GPS coordinates and road mapping for different sub divisions. He also did the analysis on Arc GIS and generated the variable for distance on GIS.
Contribution by Dr. Shafquat Rozi: Dr. ShafquatRozi helped in the analysis for the complete data and gave her regular feedback at every step of analysis.
Contribution by Dr. Robert L. Goldenberg (MD, Professor): Dr. Goldenberg provided his valuable feedback for each heading of the manuscript including introduction, methods, results, discussion and conclusion by reviewing the manuscript critically. He has given final approval of the version to be published.
Contribution by Dr. Margo S. Harrison (MD, MPH): Dr. Harrison gave provided his valuable feedback for each heading of the manuscript including introduction, methods, results, discussion and conclusion by reviewing the manuscript critically. He has given final approval of the version to be published.
Contribution by Dr. Omrana Pasha: Dr. Omrana was main supervisor for this study and she supervised at every step including proposal writing, data collection, data analysis and interpretation of data. She also helped in drafting the manuscript.
SumeraAziz Ali,SarahSaleem,NeelofarSami,MirShabbar,MuhammadAhmed,ShafquatRozi,Robert L.Goldenberg,Margo S.Harrison,OmranaPasha, (2016) Geographic Access to Working Family Planning Centers and Unintended Pregnancies among Married Women: A Community Based Nested Case Control Study. Open Journal of Epidemiology,06,95-108. doi: 10.4236/ojepi.2016.61010