Effectiveness of a training intervention in increasing the use of misoprostol in postabortion care in Malawi: A quasi-experimental study

Objectives The study was conducted to determine effectiveness of a training intervention in increasing use of misoprostol in management of incomplete abortions. Design A quasi-experimental study with training intervention on use of misoprostol in treatment of incomplete abortion. Setting Five secondary-level public hospitals in Malawi, one in urban and four in semiurban settings. Three intervention and two control sites. Participants Records of women treated for first-trimester incomplete abortion from March to May 2020 (baseline) and April to June 2021 (endline). Clinical data were collected from 865 records, 421 before and 444 after the intervention in all study sites. Intervention Three-hour theoretical training sessions for 81 healthcare workers were conducted in July 2020 at the three intervention sites. Primary and secondary outcome measures Proportion of women with incomplete abortion treated with misoprostol before and after the intervention. The proportion of women treated with sharp curettage at the study sites. Results At the intervention sites, there was a significant increase in use of misoprostol from 22.8% (95% CI 17.9% to 28.0%) to 35.9% (95% CI 30.5% to 41.6%) and significant reduction in use of sharp curettage from 48.1% (95% CI 41.9% to 54.3%) to 39.4% (95% CI 35.3% to 42.6%) p<0.01 at baseline and endline, respectively. The use of misoprostol was significantly higher at the intervention sites with OR of 5.02 (95% CI 1.7 to 14.7) p<0.05 compared with control sites at the endline in multivariable models, and there was a difference in the difference of 14.4% ((95% CI 10.4% to 18.2%) p<0.001) between the intervention and control sites after the intervention. Conclusions A training intervention effectively increased the use of misoprostol in the treatment of incomplete abortions. Increasing misoprostol use will make treatment of incomplete abortion cheaper, easier and more easily accessible. Making quality postabortion care accessible to more women may reduce maternal morbidity and mortality. Further training interventions are recommended. This is a quasi-experimental study with non-equivalent control groups. A retrospective chart review using a pre-tested data extraction tool was done to evaluate the use of misoprostol before and after a training intervention. Baseline and endline results from the three intervention and two control sites were compared. Patient treatment records for a period of 3 months before and 3 months to a year after the intervention were reviewed to note the type of management offered to women with incomplete abortions. The study hypothesis was that a training intervention could increase the use of misoprostol by 10%. The primary outcome of the study was to increase the proportion of women managed with misoprostol for treatment of first trimester incomplete abortion. The secondary outcomes were a decrease in the proportion of women treated with sharp curettage at the intervention and control sites and the increase in the proportion of middle-level healthcare workers (nurse/midwives and clinical officers) providing misoprostol at the intervention sites. The study was conducted in the central region of Malawi, in the gynaecological wards of five public hospitals: Bwaila, Salima and Mchinji district hospitals as intervention sites; Ntcheu and Dedza district hospitals as control sites. Bwaila has an average of 40 first-trimester incomplete abortions per month, and the other four hospitals have an average of 20 first-trimester incomplete abortions per month. A list of nine potential sites in the central region was generated, and the sites were visited to ascertain the availability of appropriate clinical population (women with incomplete abortion), adequacy of staff, geographic location, availability of space for research staff, flexibility to allow study interventions and lack of competing demands. All the district hospitals met the selection criteria. Treatment records from women treated for first-trimester incomplete abortions at the study sites were included. Records of women treated for incomplete abortion beyond the first trimester; records missing relevant data such as gestational age or uterine size; and records missing almost all demographic data were excluded (figure 1). Identification of records included in the study. A two-stage sampling technique was used. Since all the district hospitals in the central region met the selection criteria, we did a simple random sampling using the fishbowl method to select three intervention sites and two control sites. Convenience sampling was used to identify the treatment records that were available for data extraction. Data were collected from 865 records. A pretested data extraction tool was used to collect data from 865 treatment records of women who were managed for first-trimester incomplete abortion: 421 before the intervention; and 444 after the intervention (figure 1). Of the 865 records, 285 were from the two control sites (n=139 at baseline; n=146 at endline) and 580 were from the three intervention sites (n=282 at baseline; n=298 at endline). The baseline data were collected between June and July 2020, and client records from March to May were sampled. Endline data were collected in July 2021 (figure 2), and client records from April to June were used. Data collection was carried out by five trained data collectors. Phases of data collection. The intervention was implemented to improve postabortion care in Malawi by increasing the use of misoprostol in the treatment of incomplete abortions. Training on the use of misoprostol in postabortion care (PAC) was conducted in the three study sites. No training was offered to healthcare workers from the two control sites at any point during the study period. Preintervention data were collected in June and July 2020; postintervention data were collected in July 2021. The in-service training sessions were conducted in July 2020. The intervention was implemented for one year. The training sessions targeted physicians (doctors and clinical officers) and nurse-midwives who were involved in the management of women with first-trimester incomplete abortions. Eighty-one healthcare workers in the three intervention sites were trained in the use of misoprostol in PAC. Of the 81 healthcare workers, 28 were clinical officers, 47 were nurse-midwives and 6 were doctors. One session took place at each of the three sites. The training sessions lasted about 3 hours and were mainly theoretical. The in-service training was facilitated by a local consultant in gynaecology together with the research team. The training was offered at each hospital once, at the commencement of the intervention. This was a cheap and efficient way of delivering the content. The healthcare workers were informed on how to examine the patient, in order to determine if medical treatment is appropriate. Any woman with an incomplete abortion and a uterine size of less than 12 weeks’ gestation by examination, or who had expelled the fetus and was experiencing minor bleeding, was given three tablets of misoprostol (600 μg) orally or two tablets (400 μg) sublingually.9 Misoprostol, a prostaglandin E1 analogue, binds to myometrial cells to cause strong myometrial contractions, leading to the expulsion of retained products of conception.24 Pain relief was achieved by the administration of ibuprofen or paracetamol. The women were also given prophylactic antibiotics: metronidazole 400 mg 8 hourly and doxycycline 100 mg 12 hourly for 5 days. Women coming from a distance were offered a hospital stay for four to 6 hours while waiting for the effect of the drug. After the observation period, the women were discharged if they showed no sign of heavy bleeding and/or the retained products had been expelled. Those with severe bleeding were managed by surgical evacuation and observed further until stable. Women experiencing minor bleeding and those living not far from the hospital were treated as outpatients; they could go home and wait for the products to be expelled. The women were given clear instructions to return if they experienced continuous heavy bleeding, a feeling of light-headedness, dizziness, severe abdominal cramps and signs of infection (fever, chills and offensive vaginal discharge). All women were scheduled for a clinical check-up after 1 week to make sure the uterus was empty. The women were offered family planning counselling immediately after treatment; they were advised to abstain until the bleeding stopped. Data from maternal treatment records were collected using Android devices with the use of forms generated by CSPro V.7.0. Data were then synced to a Dropbox account in real time. STATA V.16.0 was used for detailed statistical analysis after data were exported. Basic descriptive analysis was done to determine the distribution of data. Frequency tables were generated to reveal trends in data, check for any potential anomalies and provide descriptive characteristics using proportions with 95% confidence interval (CI). Differences in proportions of women treated with misoprostol for incomplete abortions at baseline and endline, and between the intervention and control group, were compared by using χ2 statistics. Binary logistic regression was used to determine the use of misoprostol at the intervention and control sites after the intervention in multivariable models adjusting for marital status, education, employment, parity and gestational age. A p<0.05 was considered statistically significant. Patients were not involved in the planning of this study, but district health officers responsible for the study sites were involved in order to gain permission to access patient records.