Impact statement
Antenatal anxiety affects up to 49% of pregnant women in Pakistan and is associated with preterm birth, low birthweight, postpartum depression, and adverse child health outcomes. Despite this burden, scalable mental health interventions for pregnant women remain limited in low and middle income countries. This study evaluates the economic implications of the Happy Mother, Healthy Baby (HMHB) program, a low intensity cognitive behavioral intervention delivered by trained non specialists within routine antenatal care.
The analysis estimated that HMHB costs PKR 10,641 per mother to deliver. Women receiving the intervention had somewhat higher third trimester healthcare costs, particularly for provider visits and total expenditures, although these differences were small and time limited. Postpartum healthcare costs for children in the HMHB group tended to be lower for provider visits, hospital stays, and overall expenditures, suggesting potential downstream savings through improved child health. However, these trends should be interpreted cautiously, as they do not establish significant cost reductions. Overall, the findings provided no evidence of substantial increases in medical expenditure.
These findings have important implications for policy and practice. Integrating HMHB into routine antenatal care could expand access to psychological support in Pakistan and similar resource constrained settings, where mental health specialists are scarce and stigma limits service uptake. Delivery by trained non specialist providers, supported through structured supervision, offers a practical approach to implementation. Opportunities to improve economic efficiency include reviewing administrative and supervisory costs without compromising intervention quality.
By indicating that psychological support can be delivered without substantial additional healthcare expenditure, this study contributes to the evidence supporting investment in perinatal mental health. Wider implementation of HMHB may improve maternal wellbeing and support healthier child development, although these broader benefits require further evaluation. Future research should examine long term cost effectiveness, implementation sustainability, and adaptation across diverse health systems. HMHB represents a promising approach to addressing unmet perinatal mental health needs in resource limited settings.
Introduction
Pregnancy can be a stressful event and may lead to mental health challenges including anxiety and depression. Stress, anxiety and depression have been independently associated with adverse pregnancy outcomes (Dunkel Schetter and Tanner, Reference Dunkel Schetter and Tanner2012; Hasanjanzadeh and Faramarzi, Reference Hasanjanzadeh and Faramarzi2017). These adverse outcomes include, but are not limited to, preterm births (Dunkel Schetter and Tanner, Reference Dunkel Schetter and Tanner2012), low birthweight (Rahman et al., Reference Rahman, Bunn, Lovel and Creed2007) and increased risk of respiratory illness early in the child’s life (Beijers et al., Reference Beijers, Jansen, Riksen-Walraven and De Weerth2010). Previous research has also established antenatal anxiety to be a predictor of postpartum anxiety and depression (Coelho et al., Reference Coelho, Murray, Royal-Lawson and Cooper2011). The prevalence of mental health challenges in pregnant women remains high, with an estimated 10%–25% of pregnant women suffering from an anxiety disorder at some point during their pregnancy (Shahhosseini et al., Reference Shahhosseini, Pourasghar, Khalilian and Salehi2015). Studies in low- and middle-income countries (LMICs) have estimated that antenatal anxiety affects 34.4% of expectant mothers (Bedaso et al., Reference Bedaso, Adams, Peng and Sibbritt2021). This high prevalence also exists in Pakistan, where anxiety rates have been found to be as high as 49% in pregnant women (Waqas et al., Reference Waqas, Raza, Lodhi, Muhammad, Jamal and Rehman2015). These findings emphasize the importance of identifying expectant mothers who experience anxiety during pregnancy and providing them with effective interventions to combat these challenges both during and after pregnancy.
The Happy Mother, Healthy Baby (HMHB) program is a low-intensity psychosocial intervention delivered by trained non-specialist providers that incorporates core cognitive behavioral principles. Adapted from the WHO Thinking Healthy Program (World Health Organization, 2015), HMHB incorporates behavioral activation, healthy thinking, problem-solving, stress management and family engagement to address antenatal anxiety in a culturally appropriate manner. The program aims to reduce antenatal anxiety, thereby reducing postpartum anxiety and depression and ultimately promoting better outcomes for both mothers and their children (Surkan et al., Reference Surkan, Hamdani, Huma, Nazir, Atif, Rowther, Chaudhri, Zafar, Mullany, Malik and Rahman2020). There is a shortage of attention to interventions for treating mental health challenges in pregnant women, especially in LMICs, including Pakistan, where mental health treatment may be stigmatized despite the high prevalence of mental health challenges (Bhamani et al., Reference Bhamani, Arthur, Van Parys, Letourneau, Wagnild, Premji, Asad and Degomme2022) and where there is a treatment gap and shortage of mental health professionals (Choudhry et al., Reference Choudhry, Khan and Munawar2023). Mental health interventions, like HMHB, that can be incorporated into primary health care services can help to ensure accessibility and affordability of mental health care, leading to increased utilization (World Health Organization, 2018). Previous evaluations of the HMHB intervention carried out among a population of pregnant women in Pakistan with at least mild anxiety demonstrated reductions in maternal anxiety and depression as well as improvements in maternal functioning, mother-infant outcomes and child development (Nisar et al., Reference Nisar, Xiang, Perin, Malik, Zaidi, Atif, Rahman and Surkan2024; Siebach et al., Reference Siebach, Perin, Malik, Atif, Zaidi, Rahman and Surkan2024; Surkan et al., Reference Surkan, Malik, Perin, Atif, Rowther, Zaidi and Rahman2024; Siebach et al., Reference Siebach, Park, Mansoor, Atif, Zaidi, Rahman, Malik and Surkan2025), however, the cost of implementing the intervention remains unknown.
Despite the high prevalence of antenatal anxiety in low- and middle-income countries, there is limited evidence on the cost-effectiveness of scalable mental health interventions in these settings. By evaluating costs to implementors and patients and comparing the intervention to standard of care, public health policymakers and interventionists can understand the feasibility and the cost-effectiveness of the intervention. Understanding the cost associated with such interventions and the resulting cost-effectiveness is vital for making policy decisions regarding the implementation of health services.
The aim of this study is to evaluate the costs associated with the HMHB program, which has proven effective at improving outcomes for mothers and children within the population of interest, compared to the standard of care. This information can be used to inform decision-making for the future implementation and utilization of the HMHB program.
Methods
Participants
Between April 2019 and October 2022, a phase three, two-arm, single-blind, individual randomized clinical trial was conducted in Rawalpindi, Pakistan (clinicaltrial.gov identifier: NCT03880032) (Surkan et al., Reference Surkan, Hamdani, Huma, Nazir, Atif, Rowther, Chaudhri, Zafar, Mullany, Malik and Rahman2020). Women were sampled from a public tertiary care facility affiliated with Rawalpindi Medical University, Holy Family Hospital (HFH). These 1,200 pregnant women were recruited from the outpatient Department of Obstetrics and Gynecology. During a woman’s initial routine antenatal visit, eligible women were approached for screening and consent. The inclusion criteria were as follows: gestational age ≤ 22 weeks, age ≥ 18 years, ability to speak Urdu, residence ≤ 20 km from Holy Family Hospital and intent to remain in the area through delivery. Non-specialist staff with a master’s degree conducted a secondary eligibility screening. The Structured Clinical Interview for the DSM (SCID) (First et al., Reference First, Williams, Karg and Spitzer2016) to rule out Major Depressive Disorder and Hospital Anxiety and Depression Scale (HADS) (Zigmond and Snaith, Reference Zigmond and Snaith1983) were administered as part of this screening process. A SCID-trained psychiatrist trained the non-specialist staff, achieving good inter-rater reliability. Previously, the HADS has been translated and validated in Urdu (Zigmond and Snaith, Reference Zigmond and Snaith1983; Mumford et al., Reference Mumford, Tareen, Bajwa, Bhatti and Karim1991). Inclusion criteria for the secondary screening were as follows: absence of clinical depression or other serious medical conditions and evidence of at least mild anxiety (a score ≥ 8 on the HADS anxiety subscale). The following criteria excluded potential participants: a diagnosis of a major depressive episode, suicidal ideation, self-report of present or past psychiatric disorders (e.g., bipolar disorder, schizophrenia) or psychiatric care (e.g., anxiolytic medications, psychotropic drugs) and significant physical health problems such as hepatitis and tuberculosis.
Four Institutional Review Boards approved this study. They included Rawalpindi Medical University, the Human Development Research Foundation, the Johns Hopkins Bloomberg School of Public Health and a US NIMH-appointed Data Monitoring and Safety Board. Written informed consent was provided prior to screening and data collection.
Procedures
Trial participants were randomly assigned to one of two treatment arms, HMHB intervention or enhanced routine care, which served as a control group. Randomly permuted blocks of size 4, 8, 12 and 16 were created by the trial statistician, using a pseudo-random-number generator for arm assignment. This was done before enrolment. Envelopes containing each step of the sequence from the assignment list, which was printed in order, were stored and numbered in order with a seven-digit code. Assignment of an eligible woman to the intervention or control group was recorded by pulling and opening the next envelope available after consent to participation was provided. The process was repeated until a sample size of 1,200 (600 in each arm) was reached.
The intervention consisted of six core one-on-one sessions of HMHB (and up to six booster sessions, as needed) delivered to women in the intervention arm (Atif et al., Reference Atif, Nazir, Zafar, Chaudhri, Atiq, Mullany, Rowther, Malik, Surkan and Rahman2020). Weekly sessions were given immediately after enrollment for the first five sessions, with the sixth delivered in the third trimester. Between the fifth and sixth sessions, booster sessions were given to some women (mean number of booster sessions 1.5, SD = 0.78), as deemed necessary by non-specialized providers (NSP). One-on-one sessions were provided at the hospital by non-specialist providers who had completed a two-year bachelor’s degree followed by 2-year non-clinical master’s level training in psychology but had no formal clinical training or licensure to practice as therapists. Before delivering the intervention, providers completed 42–50 h of classroom training followed by supervised field training. Provider competency was assessed using the Enhanced Assessment of Common Therapeutic Factors (ENACT) tool (Kohrt et al., Reference Kohrt, Jordans, Rai, Shrestha, Luitel, Ramaiya and Patel2015), and only providers demonstrating competency delivered the intervention. Throughout the trial, randomly selected intervention sessions were assessed using ENACT. Results were used during weekly supervision sessions with a PhD-level psychologist to provide constructive feedback, practice skills and promote intervention fidelity.
HMHB was designed for pregnant women who, during the early and mid-stages of their pregnancies, experienced at least mild symptoms of anxiety. HMHB was based on considerable formative research conducted before the trial so that culturally appropriate and pregnancy-specific topics could be incorporated into the intervention (Atif et al., Reference Atif, Nazir, Zafar, Chaudhri, Atiq, Mullany, Rowther, Malik, Surkan and Rahman2020). It employed the core principles and strategies of the Thinking Healthy Program (THP) (World Health Organization, 2015), an evidence-based intervention for perinatal depression that executes its core principles through three sequential steps: identifying ways of thinking, altering unhelpful ways of thinking and practicing healthy behaviors. The goal of HMHB was to improve the personal well-being of recipients as well as promote bonding with the baby and social support (e.g., family) (Atif et al., Reference Atif, Nazir, Zafar, Chaudhri, Atiq, Mullany, Rowther, Malik, Surkan and Rahman2020). To help convey crucial health messages, intervention strategies included behavioral activation, guided discovery, stress management and take-home exercises. These intervention materials employed culturally relevant illustrations and scenarios that participants could relate to their own problems.
Reimbursement for transportation and ultrasounds was provided to participants in both the treatment and control groups. Facilitation in receiving care (e.g., expedited check-ups) and appointment reminders were also provided to both groups. Additionally, the control group also received antenatal checkup visits as usual.
Client Services Receipt Inventory (CSRI)
The Client Services Receipt Inventory (CSRI) data collection tool was utilized to record utilization of services and costs incurred by participants (Beecham and Knapp, Reference Beecham, Knapp and Thornicroft2001). CSRI data were assessed at two points: in the third trimester and at six weeks postpartum (responses pertaining to maternal and childcare were recorded separately). Participants were asked to recall services used and costs incurred over the previous 6 months both in the third trimester and to recall spending since the last interview at 6 weeks postpartum. Service utilization and costs incurred were asked to be recalled from the time of birth for childcare. CSRI data were collected from participants via an interview to inform costs including provider visits, hospital stays, testing costs and medication costs. Information on provider visits included the number of visits, the fee for the visit, average wait time, average time with the provider and travel costs/time. Total cost of the stay was collected for the hospital stay category. Testing costs were approximated by the cost of the test and other costs associated with testing, including travel. Medication costs included the total reported cost spent on all medications in the previous 6 months for mothers and the previous 3 months for children. Time costs were calculated using the average hourly wage in Pakistan (120.6 PKR/h), as reported by the International Labour Organization, as a valuation metric or conversion factor (International Labour Organization, 2021). All costs were measured in Pakistani rupees.
Statistical methods
Statistical analyses were conducted to compare costs between the treatment and control arms for each subsection: provider visits, hospital stays, testing and medications, totals for each time period (third trimester, postnatal maternal and postnatal child) and totals overall.
The total costs for visits were calculated by adding up the total average cost per service used (visit fees and other costs associated with time spent and travel) and multiplying that total by the number of times the service was utilized. The costs of the reported services used were summed to calculate the total visit costs. Hospital stays were totaled by adding the total expenditures reported for each admission reported. Medication costs were also totaled as the sum of the costs of each medication reported. Testing costs were totaled as the sum of the costs of all reported tests, which included test costs and other expenses such as travel. The costs of each subsection were summed to get the total costs for each time period (third trimester, postnatal maternal and postnatal child). The total costs of each time period were summed to generate an overall total for each participant. These costs were inflation-adjusted according to the annual inflation rates in Pakistan to achieve final values reported in 2025 Pakistani rupees.
Analyses were performed by calculating and comparing the average costs of the treatment and the control groups. Two-tailed Welch’s T-tests were used to test the statistical significance of differences in costs between the two groups.
The primary analysis compared the average costs of the treatment and control groups, including all participants. A series of secondary analyses were performed following the main analysis. To explore bias from those with no expenditures, an analysis was conducted comparing the total costs only for those with non-zero expenditures. Bias due to incomplete data was explored by performing a sensitivity analysis that only included participants who had complete data, those who had non-missing data for the third trimester, maternal postpartum and child postpartum. Another analysis was performed using those with complete data but excluding those in that sample who had expenditures equal to zero.
Maternal education level is an important variable of interest in this study as previous research has suggested that higher levels of educational attainment are associated with greater utilization of antenatal care (Wang et al., Reference Wang, Frasco, Takesue and Tang2021). Additionally, those with higher levels of education tend to have greater health literacy, which allows them to better understand the importance of antenatal care and to better understand the recommendations made by health providers (Gordon Singh and Aiken, Reference Gordon Singh and Aiken2017). Differences in maternal education, therefore, could influence maternal and child healthcare utilization and expenditures. Accordingly, subgroup analyses were conducted among participants with complete data, comparing differences in average expenditures between the treatment and control groups for those mothers with an education of middle school or below and for those mothers with an education above middle school.
The distribution of the cost variables was assessed using measures of skewness and the Shapiro–Wilk test. Given that healthcare expenditure data are commonly right-skewed and bootstrap methods are recommended for the analysis of skewed healthcare cost data in randomized trials (Barber and Thompson, Reference Barber and Thompson2000), an additional sensitivity analysis was performed using bias-corrected and accelerated (BCa) bootstrap confidence intervals with 10,000 resamples to evaluate the robustness of statistical inference. The bootstrap analyses estimated uncertainty around the same arithmetic mean cost differences that were evaluated in the corresponding analyses. Bootstrap analyses were performed for the primary analyses, non-zero expenditure analyses, complete-case analyses and education-stratified subgroup analyses.
Statistical analyses were conducted using Stata (StataCorp, Version 19, College Station, TX).
Results
Overall sample analysis
Table 1 shows the comparison of costs in the HMHB and control groups for all participants for whom data were available. The average maternal health care costs for the HMHB group were higher than those of the control group for all categories in the third trimester and postpartum periods, though few of these differences were statistically significant. Provider visit costs and total costs were significantly higher for the HMHB group at the 5% level in the third trimester, and medication costs were marginally significant at the 10% level for third-trimester and postpartum mothers. The postpartum child data showed lower costs for the HMHB group for provider visits, hospital stays and overall costs, though none of these results were statistically significant. BCa bootstrap sensitivity analyses produced conclusions that were fully consistent with the primary Welch’s t-tests, suggesting that these findings were robust to the skewed distribution of the cost data.
Average patient costs in Pakistani rupees: all data
Table 1. Long description
The table is organized into four main sections.
1. Third trimester T T health care costs:
- Cost for provider visits: H M H B Mean 6,517.76; Control Mean 5,584.56; p-value 0.01.
- Cost for Hosp stays: H M H B Mean 397.92; Control Mean 214.90; p-value 0.39.
- Cost for medications: H M H B Mean 3,028.45; Control Mean 2,584.59; p-value 0.09.
- Cost for testing: H M H B Mean 2,659.21; Control Mean 2,572.61; p-value 0.73.
- Total T T health care costs: H M H B Mean 12,603.34; Control Mean 10,956.66; p-value 0.02.
2. Postpartum maternal P P M health care cost:
- Cost for provider visits: H M H B Mean 7,872.73; Control Mean 7,316.99; p-value 0.18.
- Cost for Hosp stays: H M H B Mean 14,891.65; Control Mean 14,857.83; p-value 0.99.
- Cost for medications: H M H B Mean 4,880.48; Control Mean 4,246.85; p-value 0.08.
- Cost for testing: H M H B Mean 4,325.87; Control Mean 3,793.13; p-value 0.11.
- Total P P M health care costs: H M H B Mean 31,970.73; Control Mean 30,214.79; p-value 0.49.
3. Postpartum child P P C health care cost:
- Cost for provider visits: H M H B Mean 586.00; Control Mean 599.76; p-value 0.93.
- Cost for Hosp stays: H M H B Mean 987.22; Control Mean 1,527.47; p-value 0.39.
- Cost for medications: H M H B Mean 555.75; Control Mean 482.29; p-value 0.57.
- Cost for testing: H M H B Mean 307.91; Control Mean 258.16; p-value 0.68.
- Total P P C health care costs: H M H B Mean 2,436.88; Control Mean 2,867.69; p-value 0.61.
4. Total health care costs:
- Total health care costs: H M H B Mean 38,932.06; Control Mean 34,927.40; p-value 0.11.
Complete-case and nonzero expenditures analyses
The analysis of the expenditures among those with non-zero expenditures showed a similar pattern to the results of the previous analysis, in which the control group had lower average expenditures for the third trimester and postpartum mother surveys, though few of these differences were statistically significant. The results of the postpartum child survey indicated lower expenditures for provider visits, hospital stays, medications and overall expenditures; however, none of these differences were statistically significant (data not shown).
The results of the analysis of only those with complete data are reported in Table 2. Average health care costs in this analysis were lower for the control group compared to the HMHB group for all categories in the third trimester and for postpartum mothers except for testing costs for the third trimester, though this difference was not statistically significant. Only differences in third trimester provider visit costs and total costs, as well as medication costs for both the third trimester and postpartum mothers, were statistically significant. All of these costs were greater for the HMHB group. The costs of hospital stays and overall costs for the postpartum children were lower for the HMHB group compared to the control group, though these differences were not statistically significant. Overall, the analysis provides similar findings to the analysis using the complete set of observations.
Average patient cost in Pakistani rupees: complete data
Table 2. Long description
A table with 9 columns: Category, H M H B (N, Mean, StDev), Control (N, Mean, StDev), and Difference (Diff, p-value). Data is divided into four sections.
1. Third trimester health care costs:
- Cost for provider visits: H M H B Mean 6,428.19; Control Mean 5,362.36; p-value 0.0049.
- Cost for Hosp stays: H M H B Mean 425.61; Control Mean 212.68; p-value 0.43.
- Cost for medications: H M H B Mean 3,055.60; Control Mean 2,497.99; p-value 0.06.
- Cost for testing: H M H B Mean 2,562.24; Control Mean 2,692.98; p-value 0.69.
- Total 3rd trimester costs: H M H B Mean 12,471.63; Control Mean 10,766; p-value 0.02.
2. Postpartum maternal health care cost:
- Cost for provider visits: H M H B Mean 7,825.23; Control Mean 7,470.08; p-value 0.42.
- Cost for Hosp stays: H M H B Mean 15,531.63; Control Mean 14,457.58; p-value 0.68.
- Cost for medications: H M H B Mean 5,068.25; Control Mean 4,328.95; p-value 0.07.
- Cost for testing: H M H B Mean 4,390.78; Control Mean 3,829.23; p-value 0.12.
- Total postpartum maternal costs: H M H B Mean 32,815.89; Control Mean 30,085.85; p-value 0.34.
3. Postpartum child health care cost:
- Cost for provider visits: H M H B Mean 563.15; Control Mean 503.85; p-value 0.69.
- Cost for Hosp stays: H M H B Mean 1,120.16; Control Mean 1,458.83; p-value 0.62.
- Cost for medications: H M H B Mean 616.91; Control Mean 492.30; p-value 0.40.
- Cost for testing: H M H B Mean 320.23; Control Mean 257.46; p-value 0.64.
- Total postpartum child costs: H M H B Mean 2,620.45; Control Mean 2,712.43; p-value 0.92.
4. Total health care costs:
- Total: H M H B Mean 47,907.97; Control Mean 43,564.28; Diff 4,343.69; p-value 0.19.
Sample size N is 309 for H M H B and 307 for Control across all rows.
The results of the analysis of those with non-zero expenditures and complete data also indicated lower expenditures for the control group compared to the HMHB for the third trimester and postpartum mothers, except for testing costs in the third trimester. Again, these results were primarily statistically non-significant. The results for the postpartum children were consistent with the results of the complete response data analysis with the costs of hospital stays and total costs being lower for the HMHB group, but these differences were statistically non-significant. BCa bootstrap sensitivity analyses produced conclusions consistent with the primary analyses for both the complete case and non-zero expenditure analyses.
Subgroup analysis by education level
The results of the subgroup analyses stratified for education level are reported in Tables 3 and 4. For the lower education group, there was no statistically significant difference in costs between the HMHB group and the control group. For the higher education group, third trimester expenditures on provider visits were lower for the control group than the HMHB group by 1,170 rupees (p = 0.02). Total costs for the HMHB were higher than for the control group in the third trimester by 1,778 rupees (p = 0.09). Medication costs were statistically significantly higher for the HMHB group compared to the control group for postpartum mothers by 1,029 rupees (p = 0.06). Overall healthcare costs for the HMHB group were also marginally significantly greater than the control group by 7,151 rupees (p = 0.08).
Average patient cost in Pakistani rupees for participants attaining education of middle school or below
Table 3. Long description
The table is titled Middle school education or below and is divided into three main cost categories plus a final total. The columns are H M H B (N equals 131), Control (N equals 122), and Difference (Diff and p-value).
1. Third trimester (T T) health care costs:
* Cost for provider visits: H M H B Mean 6,149.51; Control Mean 5,210.09; Diff 939.42; p-value 0.09.
* Cost for Hosp stays: H M H B Mean 387.90; Control Mean 48.22; Diff 339.68; p-value 0.37.
* Cost for medications: H M H B Mean 2,981.81; Control Mean 2,455.72; Diff 526.09; p-value 0.25.
* Cost for testing: H M H B Mean 2,283.12; Control Mean 2,414.70; Diff minus 131.58; p-value 0.61.
* Total T T health care costs: H M H B Mean 11,802.34; Control Mean 10,128.72; Diff 1,673.61; p-value 0.11.
2. Postpartum maternal (P P M) health care cost:
* Cost for provider visits: H M H B Mean 7,538.69; Control Mean 7,895.88; Diff minus 357.18; p-value 0.63.
* Cost for Hosp stays: H M H B Mean 14,033.79; Control Mean 15,120.93; Diff minus 1,087.15; p-value 0.79.
* Cost for medications: H M H B Mean 4,333.47; Control Mean 3,947.99; Diff 385.49; p-value 0.50.
* Cost for testing: H M H B Mean 3,902.99; Control Mean 3,343.74; Diff 559.26; p-value 0.33.
* Total P P M health care costs: H M H B Mean 29,808.95; Control Mean 30,308.53; Diff minus 499.59; p-value 0.91.
3. Postpartum child (P P C) health care cost:
* Cost for provider visits: H M H B Mean 481.40; Control Mean 513.39; Diff minus 31.98; p-value 0.89.
* Cost for Hosp stays: H M H B Mean 1,567.77; Control Mean 2,592.46; Diff minus 1,024.70; p-value 0.52.
* Cost for medications: H M H B Mean 721.74; Control Mean 491.73; Diff 230.01; p-value 0.39.
* Cost for testing: H M H B Mean 461.70; Control Mean 366.74; Diff 94.96; p-value 0.75.
* Total P P C health care costs: H M H B Mean 3,232.60; Control Mean 3,964.32; Diff minus 731.72; p-value 0.71.
4. Total health care costs:
* Total health care costs: H M H B Mean 44,843.89; Control Mean 44,401.58; Diff 442.31; p-value 0.94.
Average patient cost in Pakistani rupees for participants attaining education above middle school
Table 4. Long description
The table is titled Above middle school education. It contains 9 columns: Category, H M H B (N, Mean, StDev), Control (N, Mean, StDev), and Difference (Diff, p-value).
* Third trimester (T T) health care costs:
- Cost for provider visits: H M H B Mean 6,633.28; Control Mean 5,462.77; Diff 1,170.50; p-value 0.024.
- Cost for Hosp stays: H M H B Mean 453.36; Control Mean 321.13; Diff 132.23; p-value 0.72.
- Cost for medications: H M H B Mean 3,109.9; Control Mean 2,525.86; Diff 584.03; p-value 0.14.
- Cost for testing: H M H B Mean 2,767.66; Control Mean 2,876.49; Diff minus 108.84; p-value 0.83.
- Total T T health care costs: H M H B Mean 12,964.2; Control Mean 11,186.26; Diff 1,777.94; p-value 0.09.
* Postpartum maternal (P P M) health care cost:
- Cost for provider visits: H M H B Mean 8,036.11; Control Mean 7,189.30; Diff 846.81; p-value 0.12.
- Cost for Hosp stays: H M H B Mean 16,633.98; Control Mean 14,020.13; Diff 2,613.85; p-value 0.44.
- Cost for medications: H M H B Mean 5,609.01; Control Mean 4,580.18; Diff 1,028.83; p-value 0.07.
- Cost for testing: H M H B Mean 4,749.77; Control Mean 4,149.39; Diff 600.37; p-value 0.20.
- Total P P M health care costs: H M H B Mean 35,028.87; Control Mean 29,939.00; Diff 5,089.87; p-value 0.17.
* Postpartum child (P P C) health care cost:
- Cost for provider visits: H M H B Mean 623.31; Control Mean 497.56; Diff 125.76; p-value 0.52.
- Cost for Hosp stays: H M H B Mean 790.74; Control Mean 711.24; Diff 79.50; p-value 0.85.
- Cost for medications: H M H B Mean 539.76; Control Mean 492.67; Diff 47.09; p-value 0.77.
- Cost for testing: H M H B Mean 216.12; Control Mean 185.40; Diff 30.72; p-value 0.74.
- Total P P C health care costs: H M H B Mean 2,169.93; Control Mean 1,886.86; Diff 283.06; p-value 0.68.
* Total health care costs:
- Total health care costs: H M H B Mean 50,162.99; Control Mean 43,012.12; Diff 7,150.87; p-value 0.08.
In both the overall (all data and complete data) and the education-stratified analyses, results show a trend of greater spending on third-trimester and postpartum care for the mothers in the HMHB group but lower levels of spending on postpartum care for the children in the HMHB group. The analyses, however, indicate that most of the differences are fairly small as few of the differences in spending between the groups were statistically significant. The BCa bootstrap sensitivity analyses produced conclusions consistent with those of the primary subgroup analyses.
Costs of implementation
The cost of treatment implementation was also calculated from the perspective of program delivery and included the personnel costs associated with training, supervision, intervention delivery and program oversight. Personnel costs were estimated using monthly salaries and time dedicated to each task to derive an hourly wage rate for trainers, therapists and supervisors. Additional implementation costs also included participant reimbursements for ultrasounds and travel costs, participant time costs, administrative expenses and materials necessary for intervention delivery. Costs associated solely with research activities, including research staff and data collection for evaluation purposes, were excluded as the objective was to estimate the cost of delivering the intervention under routine delivery conditions. These values were calculated on a per-participant basis under the assumption that four therapists would deliver the intervention and be capable of treating 32 mothers at a time, with each mother receiving an average of 7.5 sessions over the course of the intervention. The values used in this calculation are reported in Table 5. The implementation costs for one full treatment totaled 10,641 Pakistani rupees per mother in this setting. Implementation costs would vary based on the setting of the program.
Implementation costs for HMHB in Pakistani rupees
Table 5. Long description
The table is organized into four main sections.
1. Initial training and supervision costs per mother:
- Cost of initial training: 859
- Annual cost of refresher training: 156
2. Session delivery and administrative costs per delivery of one session:
- Cost of session delivery: 312
- Cost of administrative work: 156
- Cost of supervision: 135
- Cost of oversight: 125
- Total cost per session delivery: 728
3. Other resources:
- Travel costs per mother: 2,813
- Mother’s workbook per mother: 750
- Ultrasound costs per mother: 600
- Time costs per mother: 67.5 h
4. Total implementation costs of whole intervention per mother:
- Total training and supervision costs: 1,015
- Total cost of session delivery for whole intervention (7.5 sessions): 5,463
- Total cost of additional resources: 4,163
- Total overall cost for whole intervention: 10,641
Discussion
The findings of the economic analysis of the HMHB program indicate little statistically significant difference in medical spending between women who received the treatment and those who did not. Nonetheless, trends in the comparisons suggest that women in the treatment group may have spent more on medical care in the third trimester and postpartum for themselves but also may have spent less on medical care for the children in the postpartum period. This trend in the data remained even when accounting for incomplete data, those who reported zero expenditures and the mother’s education level. Additionally, the subgroup with complete data also had a greater proportion of individuals with a higher level of education, suggesting that those with higher education were more likely to report complete cost data.
Previous evaluations of the HMHB intervention have demonstrated improvements in outcomes for both mother and child, including improvements in maternal anxiety and depression, maternal functioning, mother-infant interactions and birth outcomes (Nisar et al., Reference Nisar, Xiang, Perin, Malik, Zaidi, Atif, Rahman and Surkan2024; Siebach et al., Reference Siebach, Perin, Malik, Atif, Zaidi, Rahman and Surkan2024; Surkan et al., Reference Surkan, Malik, Perin, Atif, Rowther, Zaidi and Rahman2024, Reference Surkan, Park, Sheng, Zaidi, Atif, Osborne, Rahman and Malik2025; Olivas et al., Reference Olivas, Park, Malik, Zaidi, Atif, Rahman and Surkan2025; Park et al., Reference Park, Malik, Batchelder, Zaidi, Atif, Rahman and Surkan2025). Our findings suggest that these previously demonstrated benefits may be achieved without significantly increasing direct medical expenditures, supporting the feasibility for widespread use in LMICs.
Postpartum depression has important economic implications because it is associated with substantial healthcare and societal costs. A recent study modeled on the cohort of women giving birth in Pakistan in 2017 estimated the total cost of perinatal depression and anxiety to be US$16 billion, or US$2,680 per woman giving birth. The largest proportion of these costs was attributed to quality-of-life losses (Bauer et al., Reference Bauer, Knapp, Alvi, Chaudhry, Gregoire, Malik, Sikander, Tayyaba, Waqas and Husain2024). Women who experience postpartum depression are significantly more likely to be unemployed, have a lower per capita household income, incur higher medical costs and be part of a lower socioeconomic group at 12 months postpartum than women without depression (Cleary et al., Reference Cleary, Orangi, Garman, Tabani, Schneider and Lund2020). A 2020 study of the economic cost of maternal depression in low-income women in South Africa found that provider costs were significantly higher for women with depression compared to women without depression (Cleary et al., Reference Cleary, Orangi, Garman, Tabani, Schneider and Lund2020). These studies suggest that interventions that reduce maternal depression may generate broader economic benefits not fully captured by healthcare expenditures evaluated in the present study.
When accounting for the implementation cost of 10,641 Pakistani rupees per mother in this setting, the HMHB program requires a relatively modest additional investment relative to the benefits demonstrated in previous effectiveness studies. It is also worth noting that the costs of administrative work and oversight accounted for about half of the total cost of delivering one session to a mother, suggesting that identifying more efficient models for program management and supervision may improve the affordability of the intervention as health systems consider integration into routine antenatal care. The program’s use of trained non-specialist providers under structured supervision further supports scalability in resource-limited settings where mental health specialists may be scarce. Long-term sustainability, however, will depend on sustained financing, workforce training and availability and integration into existing health system structures. Further implementation research should evaluate these factors, as well as costs within specific health system settings to better understand the long-term sustainability of HMHB as a part of routine antenatal care.
Although most differences we observed were not statistically significant, suggesting there is no clear evidence that the costs differed between groups, the data showed a trend of higher maternal expenditures and lower child expenditures in the HMHB group, which may provide insight into how the intervention influences healthcare utilization. One possible explanation for this is that mothers may have been more likely to seek out and utilize prenatal care when enrolled in the HMHB program, potentially resulting in fewer child health problems postpartum. Use of prenatal care in Pakistan is generally very low, with only 41.7% of women having four or more antenatal visits during their last pregnancy in 2022 (Tasneem and Ozdal, Reference Tasneem and Ozdal2023). Some of the biggest reasons suggested for insufficiency of use are lack of information about antenatal care services, lack of prior experience with antenatal care services and perceived lack of availability of services (Nisar et al., Reference Nisar, Aurangzeb, Dibley and Alam2016). Early introduction to antenatal care while participating in the HMHB program may have resulted in mothers being more informed about and more comfortable in using necessary antenatal care services, as previous research has indicated early initiation of antenatal care to be a significant factor in adequate utilization of antenatal services (Gupta et al., Reference Gupta, Yamada, Mpembeni, Frumence, Callaghan-Koru, Stevenson, Brandes and Baqui2014). Additionally, participation in the program may play a role in the continuation of antenatal care, leading to improved outcomes. Knowledge of both the availability of and the importance of antenatal care in conjunction with knowledge of how to access it may increase the likelihood of expectant mothers to utilize this important service (Okedo-Alex et al., Reference Okedo-Alex, Akamike, Ezeanosike and Uneke2019). While this may lead to higher costs to the mother, it also could result in a healthier pregnancy and delivery of a healthier child who will be less likely to experience medical complications in early life.
Antenatal anxiety has also been associated with adverse birth outcomes, including preterm birth and low birthweight (Ding et al., Reference Ding, Wu, Xu, Zhu, Jia, Zhang, Huang, Zhu, J- and Tao2014). These outcomes may necessitate more early medical care for the infant, resulting in higher costs for long-term developmental delays (Sicuri et al., Reference Sicuri, Bardají, Sigauque, Maixenchs, Nhacolo, Nhalungo, Macete, Alonso and Menéndez2011). Results of our trial show that HMHB was related to reductions in low birthweight for women who received the full intervention (Siebach et al., Reference Siebach, Perin, Malik, Atif, Zaidi, Rahman and Surkan2024). Future longitudinal studies are needed to determine if participation in HMHB has long-term effects and what impact the potential reductions in the cost of medical care in the long run would have.
Study limitations include that CSRI data were self-reported (given there were no available medical records) and recalled retrospectively over a period of 6 months for the mother and 3 months for the child, which could have resulted in inaccurate reporting of costs. The data collection process also introduced the problem of missing data. Some participants did not report a value for some economic measures used in the analysis while others reported that they did not know what their costs had been. This reduced the sample size for analysis significantly from the initial 1,200 to 836 in the overall data analysis in the third trimester. This may have additionally skewed the totals in the overall data analysis, which included anyone who reported values for at least one survey, as they were assumed to have costs of zero for any survey on which they reported they did not know the costs or did not report a value. This highlights loss to follow-up as a significant limitation in the study as approximately 37% of participants did not complete final data collection. The choice to randomize before the intervention was delivered meant that 25% of women assigned to the treatment group never received the intervention. COVID-19, which was concurrent with the study, likely played a role in the loss of follow up as it may have hindered both the participants’ ability and willingness to travel to the hospital to receive care (Rauf et al., Reference Rauf, Zulfiqar, Mumtaz, Maryam, Shoukat, Malik, Rowther, Rahman, Surkan and Atif2021). Lockdowns additionally limited some participants to phone interactions, which may have contributed to loss to follow-up. In addition, measuring all types of medical costs as the sole economic variable may also not accurately represent the impact of the HMHB program since some costs for both the mother and baby were unrelated to the mothers’ mental health or to the trial (e.g., related to preexisting conditions), causing more variance in the data. Increased variance can make it more difficult to find statistically significant differences. This analysis also fails to account for the wider economic and social value that investing in mental health care has on society.
Ultimately, while healthcare costs for mothers in the HMHB group were significantly higher than the control group for provider visits and total costs in the third trimester, the present study did not demonstrate an overall reduction in medical care costs across other cost categories or time points. Prior evaluation of HMHB has shown improved mental health outcomes for participants (Surkan et al., Reference Surkan, Malik, Perin, Atif, Rowther, Zaidi and Rahman2024), and our findings suggest these benefits can be achieved with only a modest and time-limited increase in maternal medical expenditures. Adoption of such a program, which appears to have limited additional costs overall, could be significant in settings like Pakistan where there is a high rate of mental health disorders in conjunction with a cultural stigma surrounding mental health disorders, the combination of which contributes to poor access to mental health services (Shah et al., Reference Shah, Sun, Xu, Jiang and Yuan2022). Our findings of reduced costs for the child’s early medical care could suggest better long-term outcomes for children whose mothers participated in the program, supporting prior research about the impact of antenatal anxiety on birth outcomes (Ding et al., Reference Ding, Wu, Xu, Zhu, Jia, Zhang, Huang, Zhu, J- and Tao2014). The observed trend of increased costs to the mother may result from using more care because the participant feels more motivated and comfortable utilizing the healthcare system, reducing some established barriers to antenatal care in Pakistan. Overall, these findings support consideration of implementing the HMHB program in Pakistan and other similar settings based on the previously demonstrated clinical benefits, which appear to be achievable without showing evidence of corresponding significantly higher medical expenditures.
Open peer review
To view the open peer review materials for this article, please visit http://doi.org/10.1017/gmh.2026.10330.
Author contribution
P.J.S. and A.R. conceived of the study. P.J.S. served as the principal investigator. N.A. supervised the intervention. AM oversaw the data collection in the field. A.Z. managed the data. C.C. conducted the statistical analysis under the supervision of K.D.F. and drafted the manuscript. P.J.S. and K.D.F. edited the manuscript in its entirety. All authors reviewed the manuscript, provided feedback and approved the final version for publication. P.J.S. and K.D.F. denote co-final authorship.
Competing interests
The authors have declared that no competing interests exist.
Data availability statement
The de-identified data that support the findings of this study are not publicly available due to ethical and privacy restrictions. Reasonable requests for de-identified data may be made to the corresponding author, Dr. Abid Malik (abid.malik@hsa.edu.pk), and will be subject to review by the trial investigators and relevant ethics committees, as well as a data sharing agreement. The trial protocol has been published previously (Surkan et al., Reference Surkan, Hamdani, Huma, Nazir, Atif, Rowther, Chaudhri, Zafar, Mullany, Malik and Rahman2020).
Ethics and approvals
This study was approved by four Institutional Review Boards: Rawalpindi Medical University, the Human Development Research Foundation, the Johns Hopkins Bloomberg School of Public Health and the US NIMH-appointed Data Monitoring and Safety Board. Written informed consent was obtained from all participants prior to screening and data collection. The trial was registered at ClinicalTrials.gov (NCT03880032) and conducted in accordance with the Declaration of Helsinki.
Funding statement
This study was supported by the National Institute of Mental Health at the US National Institutes of Health (Grant no. RO1 MH111859 to P.J.S.). The funders had no role in study design, data collection and analysis, decision to publish or preparation of the manuscript.
Open access
